VERTICAL DISTRIBUTION OF PELAGIC PHOTOSYNTHESIS Moltke Lyngsgaard.pdf · 2014. 1. 20. · PhD thesis by Maren Moltke Lyngsgaard 5 Acknowledgements First of all, I would like to thank

VERTICAL DISTRIBUTION OF PELAGIC PHOTOSYNTHESIS– implications for marine ecosystem dynamics

Because the tiny plants in the ocean (phytoplankton) need light to carry out photosynthesis, it has generally been assumed that it is phytoplankton distribution and activity near the surface that determines ecological structure and function in marine environments. This PhD focusses on the importance of understanding the vertical distribution of phytoplankton and their activity in controlling the dynamics of marine systems. In particular, this PhD provides important new understan-ding of how nutrient enrichment of coastal ecosystems through human activities infl uences the function of marine ecosystems – understanding that can be useful in assessing the success of efforts to reduce impacts of nutrient enrichment. In addition, the PhD provides examples of cases where the vertical structure of phy-toplankton activity can be important in terms of ocean carbon storage and fi sheries.

Vertical distribution of pelagic photosynthesis – implications for m

arine ecosystem dynam

icsM

aren Moltke Lyngsgaard • P

hD thesis • 2013

PhD thesis 2013Maren Moltke Lyngsgaard

VERTICAL DISTRIBUTION OF PELAGICPHOTOSYNTHESIS– implications for marine ecosystem dynamics

ISBN: xxxxxxxxxxxxxxx


Maren Moltke Lyngsgaard

PhD thesis 2013

Title: Vertical distribution of pelagic photosynthesis – implications for marine ecosystem dynamics Subtitle: PhD thesis Author: Maren Moltke Lyngsgaard Institute: Department of Biology (Center for Macroecology, Evolution and Climate)

Publisher: University of Copenhagen Year of publication: 2013

PhD supervisors: Professor Katherine Richardson, Center for Macroecology, Evolution and Climate, Danish Natural History Museum, University of Copenhagen, Denmark

Co-advisor: Professor Stiig Markager, Department of Bioscience – Applied Marine Ecology and Modelling, Aarhus University, Denmark

Co-advisor Michael Olesen Ramböll, Hannemanns Allé 53, DK-2300 Copenhagen S, Denmark

Assessment committee: Professor Ulrich Sommer, GEOMAR, Heimholz-zentrum für Ozeanforschung KielProfessor Marianne Ellegaard, Dep. of Plant and Environmental Sciences, University of Copenha-gen, DenmarkPer Andersen, ORBICON A/S, Aarhus, Denmark

Please cite as: Lyngsgaard, M.M. 2013. Vertical distribution of pelagic photosynthesis – implications for marine ecosystem dynamics. PhD thesis. University of Copenhagen, Denmark. 142 pp.

Layout: Tinna Christensen Cover photo: Peter Bondo Christensen

ISBN:

Number of pages: 142

Printed by:

Circulation:

Data sheet

Content

Acknowledgements 5

Abstract 7

Resumé 10

List of manuscripts 13

Part I: Synopsis 14Plants of the ocean 14Vertical distribution of phytoplankton activity 15Biomass and primary production 16Phenotypic responses to changes in light environment 17Genotypic response 18Eutrophication and primary production 18The PhD work 19The study area 19Quantifying subsurface primary production 21The fi ndings of this PhD work 22Eutrophication 23The importance of increased deep primary production 25Ecology 27Biological pump 29The next step: Ideas and thoughts for future research 30References for synopsis 32

PART II: PAPERS I-V 43

PAPER I Changes in the vertical distribution of primary productionin response to land-based N-loading 45

PAPER II Deep primary production in coastal pelagic systems:Importance for ecosystem functioning 61

PAPER III Seasonal variation in biological parameters in a temperatecoastal area – decoupling of chlorophyll concentration andprimary production 81

PAPER IV Localised upwelling may lead to heterogeneity in theplankton food web in a frontal region of the Sargasso Sea 103

PAPER V Major contribution of diatom resting spores to vertical fl uxin the sub-polar North Atlantic 123

Appendix: Conference contributions 137

Photo: Peter Bondo Christensen.

5PhD thesis by Maren Moltke Lyngsgaard

Acknowledgements

First of all, I would like to thank my primary supervisor, Professor Katherine Richardson for giving me the opportunity to work with marine phytoplankton ecology. Thank you for being a never ending source of inspir ation, both in the scientifi c work and in your way in handling new tasks in life in general. You have an outstanding energy and ability to always turn things around to the positive. I feel that my ideas have been heard and that you have the patience to listen. Even though we have been physically apart, you have always been there for me when I needed it. Our relationship goes back nine years and I thank you for all the ex-citing adventures our relationship has led to as e.g. sailing half way around the world or celebrating our birthday on an American RV in the North Atlantic, not to mention the research cruises up to Greenland. You are the kind of person where things seem to happen around you and I am glad that I have been a part of that.

Next, I would like to thank my secondary supervisor Professor Stiig Markager for always being there for me. I really enjoyed working together with you and found our discussions fruitful and inspiring. I thank you for your patience in my process of learning SAS computation and for making me do it even though I sometimes didn’t want to. I have learnt a lot from you and appreciate that you have made me feel that my work was really important. Furthermore, I thank you for presenting me to the Danish National Research Institute (now Aarhus Univer-sity) where I have enjoyed being a part of the marine group both in Roskilde and in Silkeborg. I also thank Michael Olesen for interesting discussions and a terrifi c time spent during the fi eld work in Aarhus bight. In relation to the fi eld work, I would also like to thank Torben Vang for being a valuable support and helping me in any way possible to achieve my goals for the fi eld campaigns. In the fi rst year of my PhD, I was situated at Aarhus University in the Department of Marine Ecology. Here, I thank Inge Buss for help in the lab, help during preparation of my fi eld work and last, but not least, for always being excellent company.

I further want to thank Jens Brøgger for our many discussions on the PhD pro-cess and invaluable coffee breaks. During my year in Roskilde, I especially en-joyed sharing an offi ce with Helle Knudsen-Leerbech and I would like to thank you for making my time at the NERI more interesting and fun. During the last year of my PhD, I was situated at the Department of Marine Ecology, Aarhus Uni-versity in Silkeborg. First, I would like to thank everyone at the department for welcoming me and making me feel like a part of your group very quickly. I have really enjoyed staying in the corner offi ce and valued the work-related discussions as well as the personal ones. Thanks to Peter Bondo and Michael Bo Rasmussen for our many very important runs in the forest. Without those, I would not have been as effi cient in my work. Apart from my different affi liations I have been col-laborating with several other researchers.

Thanks to Lars Jonasson from the DMI for our many discussions on stratifi ca-tion. Thanks to Jørgen Bendtsen for your help in understanding some physical oceanographic processes in the sea. Following along the same line, I would espe-cially like to thank Morten Holtegaard Nielsen for patiently discussing hydrog-raphy with me and drawing beautiful fi gures for the manuscripts. I would like to thank Eva Friis Møller for teaching me about zooplankton dynamics and for interesting discussions. Thanks to Hans Henrik Jacobsen for letting me use his calculations of phytoplankton carbon biomass and for your help and support when needed. When I consider the more personal guidance in the PhD process,

6 PhD thesis by Maren Moltke Lyngsgaard

I especially thank Anna Neuheimer who has been a source of inspiration on how to balance science and personal life. I fully enjoyed and learnt from our many discussions on how to make the science work for you more than you working for the science. I also thank you for listening when I had things to share and for our many swims followed by good coffees. Thanks to Erik Askov Mousing for being an excellent fellow PhD student with whom I have experienced the value of real team work and table tennis victories in the middle of the North Atlantic. I really enjoyed your unique humor and interesting ideas and I am sorry our plan of bring-ing our families together and studying marine phytoplankton in the tropics never was brought to reality.

Finally, I would like to thank my family and friends for the invaluable support during my PhD work. Thanks to Lykke Cecilie Mose who has offered help when-ever needed and for our valuable time together with our kids. You have helped to increase the quality of my free time which again has motivated me in my work. A big thank you goes to my parents and their respective partners; Mads Moltke Nielsen, Line Heilskov, Kurt Sørensen and Ane Lyngsgaard who have been supporting this PhD work all the way through and helping out in any way they could. Last, but not least, a special thanks goes to my husband Johannes Nørre-lykke, thank you for your invaluable support and ability to make me laugh when I really needed it. You have always respected my work and done everything pos-sible to give me as much time as I needed to fi nish up this PhD work.


Abstract

As phytoplankton photosynthesis is dependent on light, one might assume that all the phytoplankton activity occurs in the surface of our oceans. This assumption was, however, challenged early in the history of biological oceanography when chlorophyll sampling and fl uorescence profi ling showed deep chlorophyll max-ima (DCM) to be a general feature in the ocean. Today, it is generally accepted that DCMs occur in most of our oceans still, despite this empirical knowledge, subsurface primary production is still largely ignored in marine science.

The work included in this PhD examines the vertical distribution of phytoplank-ton activity and its importance to several major themes of societal interest; car-bon transport, fi sheries ecology and eutrophication effects. The main focus of the thesis is on the importance of vertical distribution of primary production to ecosystem functioning and how this vertical distribution pattern is related to ni-trogen loading in order to elucidate potential trajectories in the recovery from eutrophication. In addition, the importance of vertical phytoplankton activity is evaluated in an oligotrophic tropical ocean as well as in a North Atlantic spring bloom in order to gain knowledge on region-specifi c importance of subsurface phytoplankton activity.

The major part of the PhD work was related to the recalculation and analysis of a 15 year long dataset (from 1998 to 2012) of photosynthetic parameters, phytoplankton biomass, chlorophyll, oxygen, nutrients and physical parameters such as salinity, temperature and light. The dataset includes six different stations located in the Bal-tic Sea transition zone made available by the Danish National Aquatic Monitor-ing and Assessment Program (DNAMAP). General patterns were derived from this dataset of vertical and seasonal variation in primary production, and nitrogen load-ing was then related to these patterns. The opportunity to combine long term data with detailed process studies became available through a two week fi eld campaign on one of the stations from the survey programme which was conducted in July 2010. The purpose of the fi eld study was to examine the importance of the vertical phytoplankton activity for the ecosystem functioning in further detail.

Before the initiation of the PhD work I measured, among many variables, pri-mary production and variable fl uorescence in the Sargasso Sea during my four month participation in the Danish circumnavigating research expedition Galathea III (from year 2006-2007). This data was, during this PhD, analysed to elucidate underlying mechanisms for eel larvae distributions. In May 2008, I spent one month measuring variable fl uorescence in the North Atlantic on the RV KNORR as a part of the North Atlantic Bloom Experiment (NABE). These data were used in the PhD study to describe the importance of the vertical distribution of phyto-plankton and their activity for the carbon transport related to the North Atlantic spring bloom. The fi ve manuscripts included in the study describe the main fi nd-ings related to the analysis of these datasets.

In manuscript I, the recalculation of primary production from the survey pro-gramme was used to quantify deep primary production (DPP). These vertical dis-tribution patterns of primary production were then related to nitrogen loading. The effect of reduced nutrient loading to the Danish coastal waters on primary production (PP) has been widely discussed. It is shown here that not only is ni-trogen loading signifi cantly related to total water column PP, but it is even more strongly related to the vertical distribution of PP. This relationship suggests that reduced loading will increase the deep primary production both in magnitude but


also relative to the total water column PP which is suggested to have a positive effect on oxygen conditions in the bottom waters.

Manuscript II goes into further detail with the prediction of increased deep prima-ry production suggested in manuscript I. Here, the results of the fi eld work carried out in Aarhus Bight were used to examine the characteristics of DPP versus PP in the surface layer, assuming that the water column exhibits different environ-ments with respect to light and nutrients for phytoplankton activity. The potential for changes in oxygen production has, previously been ignored in consideration of the oxygen conditions in the bottom waters of this region. However, this study suggests that DPP, in addition to physical processes has a positive effect on the oxygen concentration in the bottom waters. Furthermore, the photosynthetic char-acteristics and species composition of the subsurface phytoplankton community showed signs of an active community generally comprised of larger cells (mainly dinofl agellates) than in surface waters. The sedimentation of organic material related to the subsurface phytoplankton community was enhanced compared to sedimentation in surface waters. The process studies in Aarhus Bight overall sug-gest that reduced nitrogen loading will change ecosystem functioning in a manner likely to reduce the likelihood and/or intensity of hypoxia events.

In manuscript III, the DNAMAP dataset was examined for general distribution patterns in primary production, chlorophyll, nutrients and zooplankton biomass. Global estimates of PP are often based on chlorophyll derived from satellite imagery. Chlorophyll is, however, constantly varying with nutrients and light. Hence, chlorophyll concentration is not a robust indicator of the carbon fl ow within a pelagic system. This study clearly shows that PP and chlorophyll are decoupled during summer time which was mainly attributed to the seasonal variation in the carbon to chlorophyll ratio. In the search for a better proxy for carbon dynamics than chlorophyll concentrations, a parameterisation of the seasonal variation in primary production was developed. The seasonal variation of chlorophyll was clearly different from that of carbon biomass, especially with respect to the magnitude of the spring bloom. This suggests that the spring bloom may have received far more attention than it deserves, as chlorophyll rather than carbon biomass has most often been used to describe phytoplankton biomass.

In manuscript IV, physical and biological parameters were used to elucidate the mystery of eel larvae distribution in the Sargasso Sea. It has long been known that eels spawn in the Sargasso Sea but the mechanisms behind the concentrated eel larvae in frontal systems has not yet been described, since there has been no apparent link between biomass or PP and the larvae. In this study, however, it was shown that localised vertical mixing could be related to the physiologi-cal state of the phytoplankton in these deep frontal zones, potentially making those a better source of nutrients to support the planktonic food web. This study overall emphasizes the importance of the deep primary production dynamics and phytoplankton physiology to the higher trophic levels and carbon fl ow in the Sargasso Sea.

Manuscript V provides another example of the relevance of understanding phy-toplankton distributions and activity below the surface in order to understand the dynamics of the system as a whole. The carbon fl ow related to the North Atlan-tic spring bloom has been studied intensively and it has been assumed that the phytoplankton species occurring in the surface waters are a good indicator of the species likely to contribute to vertical carbon transport. The study presented here, however, showed the presence of resting spores of the diatom species Chae-toceros aff. diadema at great depths. The organic material found in the sediment


traps containing C. diadema resting spores was sinking very fast and contributing disproportionally to the carbon export to deep waters. Overall, this study suggests that the vertical distribution of phytoplankton species and activity is important to the carbon fl ow in our oceans.

The results of this PhD work show that the vertical distribution of phytoplankton and their activity is important for the understanding, dynamics and functioning of pelagic ecosystems. It, thus, emphasizes that future research and modelling exercises aimed at improving understanding of pelagic ecosystems and their role in the global ocean should include a consideration of the vertical heterogeneity in phytoplankton distributions and activity.


Resumé

Eftersom fytoplankton fotosyntese afhænger af lys, kunne man fristes til at tro, at al fytoplankton aktivitet vil koncentrere sig i overfl aden af verdenshavene. Denne antagelse blev tidligt i den biologiske oceanografi s historie kompromitteret, da fl uorosens profi leringer kunne vise, at dybe klorofyl maxima var en alminelig ob-servation i havet. Trods denne viden, bliver den dybe primærproduktion (PP), der er forbundet med disse forekomster, stadig i dag ofte ignoreret i marin forskning.

Denne PhD inkluderer undersøgelser af den vertikale fordeling af fytoplankton aktivitet og dennes betydning for større marine emner af samfundsmæssig inte-resse så som kulstof transport i havet, fi skeri økologi og eutrofi erings effekter på de kystnære farvande. Arbejdet har specielt været fokuseret på betydningen af den vertikale fordeling af PP for det pelagiske økosystem. Koblingen af denne vertikale fordeling og kvælstof udledning er endvidere blevet undersøgt med det formål at belyse mulige effekter af en reduceret kvælstof belastning på det marine økosystem. Desuden er betydningen af vertikal fytoplankton aktivitet for kulstof dynamikken undersøgt i et oligotroft økosystem samt i en forårsopblomstrings situation i Nortatlanten.

Hovedparten af denne PhD relaterer sig til genberegning af PP og analyse af et datasæt med 15 års målinger af fotosyntetiske parametre, fytoplankton biomasse, zooplankton biomasse, klorofyl, ilt, næringssalte og fysiske parametre som salitet og temperatur. Datasættet er en lille del af det danske nationale overvågningspro-gram og inkluderer seks forskellige stationer placeret i de indre Danske farvande. Datasættet blev brugt til at beskrive generelle mønstre og sammenhænge relateret til den vertikale fordeling af fytoplankton aktivitet. Muligheden for at kombinere disse generelle mønstre, udledt fra overvågningsdata, med mere detaljerede pro-ces studier bevirkede, at en to ugers feltkampagne blev udført i Aarhus bugt (en af overvågningsstationerne) i juli måned 2010. Hovedformålet med dette feltarbejde var at undersøge betydningen af denne vertikale struktur i produktionen for øko-systemet i yderligere detaljer, end hvad der var muligt ud fra overvågnings data alene.

Inden begyndelsen af PhD arbejdet deltog jeg fi re måneder i den danske forsk-nings ekspedition Galathea 3. Her målte jeg bl.a. variabel fl uorosens og PP, som under denne PhD er blevet relateret til fødedynamikken for åle larver i Sargasso havet. Dernæst deltog jeg i ”the North Atlantic Bloom Experiment”, hvor jeg brugte en måned på forskningsskibet KNORR med at måle variabel fl uorosens i Nordatlanten. Disse data blev brugt til at beskrive betydningen af den vertikale fordeling af fytoplankton aktivitet for kulstof transporten til dybhavet i forbin-delse med forårsopblomstringen. De følgende fem manuskripter beskriver i store træk resultater og konklusioner fra dette PhD værk.

Genberegningerne af primærproduktion fra overvågningsdata blev brugt til at kvantifi cere den dybe PP. Denne vertikale fordeling af PP blev dernæst undersøgt i relation til kvælstof udledningen til de danske kystnære farvande. Det har længe været diskuteret, om der er en effekt af de reducerede nærings tilførsler fra land til vand. Dette studie viste dog klart at kvælstofudledningen var signifi kant koblet til størrelsen på den årlige PP. Hvad der måske er endnu vigtigere er, at den koblede endnu stærkere til den vertikale fordeling af PP, hvor mindre kvælstof gav mere PP i den dybere del og mindre i overfl aden. Denne kobling indikerer at en redu-ceret kvælstof tilførsel vil medføre, at større dele af den samlede produktion vil foregå længere nede i vandsøjlen, hvilket potentielt kan have en positiv betydning for iltkoncentrationen i bundvandet.


Manuscript II fokuserer i detaljer på betydningen af en større dyb PP som et po-tentielt resultat af reduceret kvælstof tilførsel. Resultaterne er her baseret på det feltarbejde, der blev udført i juli 2010 i Aarhus bugt samt overvågningsdata for stationen i tidsperioden fra 1999-2012. Iltproduktion i bundvandet har generelt været betragtet som værende ubetydelig. Resultaterne i dette studie viste dog, at den dybe PP havde en positiv effekt på iltkoncentrationen i bundvandet, hvilket hentyder til, at hvis den dybe PP øges i relation til PP i overfl aden, vil dette have en positiv betydning for ilten i bundvandet. Desuden viste fytoplankton samfundet i det dybere lag sig at være vidt forskellig fra fytoplankton samfundet i overfl ade-laget, og adskillige parametre såsom sedimentation, vandbevægelse, fotosyntese karakteristik, celle størrelse og kulstof/klorofyl forhold indikerede, at dette dybe lag udviste et økosystem, der adskilte sig fra økosystemet i overfl adevandet. Altså viste dette studie, at en øget dyb PP vil være baseret på et anderledes økosystem end det, man fi nder i overfl aden. Sedimenteringen af organisk materiale fra dette dybe lag var større end sedimenteringen fra materialet produceret i overfl adelaget. Dette indikerer at et økosystem med en øget dyb PP potentielt kan producere mere organisk materiale til bunden ud fra den samme mængde kvælstof tilført.

Det nationale overvågningsdatasæt blev videre undersøgt i manuskript III med henblik på at beskrive generelle mønstre i PP, klorofyl, næringssalte og zooplank-ton biomasse. Globale estimater af PP er ofte baseret på klorofyl, som er beregnet på basis af satellit målinger. Men da klorofyl konstant varierer med lys og næring, er dette ikke et optimalt bud på en proxy for kulstof dynamik i det pelagiske system. Dette studie viser tydeligt, at klorofyl er afkoblet fra PP om sommeren grundet den markante sæsonvariation i kulstof/klorofyl ratioen. Vi udviklede end-videre en parameterisering af sæsonvariationen i PP, da behovet for en bedre pro-xy for kulstof dynamik tydeligvis er nødvendig. Sæsonmønsteret for klorofyl var markant anderledes fra sæsonmønsteret i kulstof, specielt med henblik på størrel-sen af forårsopblomstringen. Studiet tyder da på, at forårsopblomstringen gennem tiden meget vel har fået mere opmærksomhed end berettiget, da det velkendte forårs peak i tempererede områder højst sandsynligt er et artefakt af at have brugt klorofyl i stedet for kulstof til at beskrive fytoplankton biomassen.

Manuskript IV præsenterer resultater fra Galathea 3 datasættet. Her blev fysiske og biologiske parametre brugt til at belyse mysteriet bag fordelingen af ålelar-ver i Sargasso havet. Man har vidst længe, at ålen gyder i Sargasso havet, men mekanismerne bag koncentrationerne af ålelarver i frontsystemer er endnu ikke blevet beskrevet, da der hidtil ikke er fundet en kobling mellem biomasse, PP og ålelarver. Dette studie viste, at vertikal opblanding i det dybe frontsystem kunne øge fytoplanktonets fysiologi gennem øgede mængder af nærringssalte og derved potentielt øge næringsværdien af disse fødeemner. Overordnet understreger stu-diet betydningen af den dybe produktion og fytoplankton fysiologi for de højere led i fødekæden og kulstof dynamikken i Sargasso havet.

Manuskript V viser et godt eksempel på relevansen af fytoplankton aktivitet dybe-re nede i vandsøjlen. Kulstof dynamikken i den nordatlantiske forårsopblomstring er gennem tiden blevet studeret betydeligt, og det har generelt været accepteret, at fytoplankton arter i overfl aden som regel også var de vigtige arter med henblik på transporten af kulstof til dybhavet. Dette studie viste dog en betydelig forekomst af kiselalgen Chaetoceros aff. diadema på store dybder og ikke i overfl aden, samt en betydelig mængde af hvilecyster fra denne art. Hvilecyster af C. diadema viste sig at bidrage betydeligt til kulstof transporten, da dette materiale sank meget hurtigt til dybere dele af havet. Studiet viser dermed, at den vertikale fordeling af fytoplankton arter- og aktivitet kan være af stor betydning for kulstof dynamikken i det pelagiske system.


Tilsammen viser resultaterne fra denne PhD, at den vertikale fordeling af fyto-plankton aktivitet har stor betydning for forståelse, dynamik og funktion af det pelagiske økosystem. Dette understreger derved, at fremtidig forskning og model-lering, med det formål at forbedre vores forståelse af pelagiske økosystemer og deres rolle i det globale hav, burde inkludere overvejelser af den vertikale forskel-lighed i fytoplanktons fordeling og aktivitet.


List of manuscripts

I: Changes in the vertical distribution of primary production in response to land-based N-loading

II: Deep primary production in coastal pelagic systems: Importance for ecosys-tem functioning

III: Seasonal variation in biological parameters in a temperate coastal area– decoupling of chlorophyll concentration and primary production

IV: Explaining heterogeneity in plankton community structure in a frontal re-gion in the southern Sargasso Sea

V: Major contribution of diatom resting spores to vertical fl ux in the sub-polar North Atlantic


SYNOPSIS

PART I: SYNOPSIS

Plants of the ocean

Most life in the blue part of our world is dependent on marine plants. The microscopic plants living in the pelagic part of the oceans, phytoplankton, are indeed captivating in their forms and massive carbon fi xation which approximately amounts to half of the annual total global fi xation, equaling approx. 45-50 Pg yr-1

(Longhurst et al. 1995, Field et al. 1998, Falkowski and Raven 2007) and half of the global oxygen production (Field et al. 1998). As phytoplankton represent the very fi rst link in the marine food chain, an understanding of what limits their growth and photosyn-thetic processes is essential in order to understand the function-ing of marine ecosystems. The photosynthetic process is overall dependent on solar energy, nutrients and carbon. Pelagic carbon

fi xation is commonly described by estimating the amount of carbon that is fi xed by phytoplankton through photosynthesis in a certain amount of time in a specifi ed water volume, the primary production (PP).

Viktor Hensen early in the history of biological oceanography recognized that the tiniest organisms in the ocean were important for its function and referred to them as the “blood of the sea” (see Mills 1989) before ultimately proposing that they should be called plankton. It has generally been assumed that because the plant component of the plankton (phytoplankton) require light that most of their activity must happen at the immediate surface. However, as Smetacek (2001) pointed out, if all plants of the ocean were placed in the surface where light is most abundant, the world would have been very different from the one we know today with a changed heat budget with reduced evaporation hence very little rain-fall on the continents. Despite the fact that we know that phytoplankton are not all concentrated at the surface, it is generally assumed that surface phytoplankton are representative of the processes occurring in the water column as a whole. Phyto-planktons at the surface are also relatively easy to observe and sample.

One reason for this is that, although small, phytoplankton can be detected from out-er space. A distribution of phytoplankton in the surface waters of the global oceans can be seen in Fig. 1 indicated by satellite derived chlorophyll. In general, there is a high concentration of surface water phytoplankton in temperate regions, especially in the northern hemisphere and upwelling zones, including the upwelling along the equator.

In the late 1950s, scientists began to sample chlorophyll below the sur-face and, thereby, describe vertical patterns of chlorophyll distributions

(Steele and Yentsch 1960, Lorenzen 1967). Later, the development of fl uorescence profi ling instruments provided the opportunity to “see” the water column distributions of chlorophyll (Strickland 1968, Kiefer 1973). That chlorophyll peaks were observed so far from the highest light intensities was initially considered a curiosum but over the years, it has become obvious that deep chlorophyll maxima (DCM) are a very common feature of stratifi ed waters and of great importance to

the overall phytoplankton biomass in tropical and sub-tropical seas (see e.g. Sedwick et al. 2005; Benitez-Nelson et al. 2007) as well as being a

seasonal feature in temporal and polar regions (see e.g. Estrada et al. 1993, Holm-Hansen and Hewes 2004; Tremblay et al. 2008; Richardson et al. 1998).

equator.

In tfa

te.

seasHolm-H

List of abbreviations

PP Primary production SPP Surface primary production DPP Deep primary production SCM Subsurface chlorophyll maximum DCM Deep chlorophyll maximum PE Photosynthesis versus irradiance PBL Pycnocline/bottom layer

Figure 1. Sea surface chlorophyll is distributed globally accord-ing to the availability of light and nutrients. This satellite created picture is from a boreal spring. The large blue regions represent oceans where nutrients are de-pleted in surface waters and green areas denote regions with high concentrations of chlorophyll in surface waters. The fi gure is from the Satellite observations in sci-ence education (NASA) website: http://www.ssec.wisc.edu/sose/cu/cu_m2_static3.html.


SYNOPSIS

Vertical distribution of phytoplankton activity

Phytoplankton often faces the dilemma of having one life-necessity reduced such as light in the accomplishment of gaining another life-necessity such as nutrients when they change their position in the water column. An example of a DCM in the Sargasso Sea situated between 100 m and 150 m is shown in Figure 2.

The biological oceanographic community has, for decades, been trying to explain the presence of these DCMs: Variation with depth was described early by Steele and Yentch (1960) for the waters off New York. The inter-specifi c competition for light and nutrients was described through numerical models establishing the role of competition for the placements of deep chlorophyll maxima (DCM) in the sea (Klausmeier and Litchman 2001). Bienfang et al. (1983) did some experiments that suggested that reduced light intensity could induce sinking of the cells in subtropical waters and this was concurrent with physiological changes in the cell where the chlorophyll to carbon ratio was increased (Cullen 1982) making them a better fi t for the low light environment found at depth. Takahashi and Hori (1984) suggested that nutriclines in the South China Sea were important to the depth of the DCMs. Richardson et al. (2000) described a tidal pumping mechanism, whereby fortnightly movements of water masses “pumps” nutrients into the pyc-nocline region in the southern North Sea, thus fueling sub-surface phytoplankton blooms. Lorenzen (1967a) suggested that differential grazing pressure could have an impact on where the DCMs were found in the Bay of California. The presence and placements of a DCM has also been explained by a combination of the above (Longhurst and Glen Harrison 1989, Beckmann and Hense 2007) and, recently, Navarro and Ruiz (2013) argued that DCMs follow specifi c isoclines and that this can be explained by the hysteresis of the water mass. They show how inter-annual variation in fl uorescence profi les from the Bermuda Atlantic time-series study (BATS) seems to fall in place when they are plotted against density instead of depth.

Cullen and Eppley (1981) showed for the southern California Bight a relationship between the DCM position and the nutricline. This has later been shown to be the case locally in a number of studies and globally from our own data from the 2006-7 Galathea 3 expedition which circumnavigated the globe. On the Galathea Ex-pedition, chlorophyll profi les and measurements of nutrient concentrations were

0.020.060.100.140.180.220.260.300.340.380.420.460.50

Distance (km)

Dep

th (m

)

0 1 2 3 4 5 6 7 8 9 10 11-400

-300

-200

-100

0Chl.a µg l-1

Figure 2. Deep chlorophyll max (converted from fl uorescence to chlorophyll) in the Sargasso Sea is situated at 100 m to 150 m depth. The transect was 11 km long and situated along 64° W and from 19°-29° N (see manuscript IV for map and timing of sam-pling).


SYNOPSIS

carried out in most major ocean basins. The general picture that came up from all these measurements was that the placement of the DCMs was tightly coupled to the depth of the nitraclines when those were situated in deep waters (NO3 ≥ 1 μM, see Figure 3) and the DCM was in general found above the nutricline. There was, however, no coupling when the nitracline was shallow or when the DCM was situated very close to the surface. The global dataset suggests that this may be a general feature of the DCM.

Although the presence of DCMs is well documented, they have generally been assumed to have little impact on water column PP. This may be best exemplifi ed by the use of satellite estimated surface chlorophyll concentrations to calculate global PP (Longhurst et al. 1995, Field et al. 1998, Falkowski and Raven 2007). As less than half the variation in PP can be explained by sea surface chlorophyll (Campbell and O’Reilly 1988, Balch et al. 1992) this method is questionable in stratifi ed regions, where the nutricline is deep and where DCMs are important and consistent feature.

A number of studies in various parts of the world have shown that subsurface peaks or DCMs can be important for PP; The North Sea (see Reid et al. 1990; Richardson and Christoffersen 1991; Richardson and Pedersen 1998; Richardson et al. 2000 and Weston et al. 2005); English Channel (see e.g. Holligan et al. 1984; Sharples et al. 2001), the Celtic Sea (see e.g. Hickman et al. 2012), the Baltic Sea transition zone (see e.g. Lund-Hansen et al. 2006; Lyngsgaard et al. submitted), the Baltic Sea (see e.g. Kononen et al. 2003; Lips et al. 2010), the subtropical Atlantic Ocean (see e.g. Veldhuis and Kraay 2004); and the Greenland Sea (see Richardson et al. 2005). Still, a better understanding of the vertical distribution of PP is clearly necessary to improve estimates of global PP and to understand the ecology of pelagic systems.

Biomass and primary productionMore than 90 % of ocean primary productivity is fi xed through phytoplankton photosynthesis and then transferred through food webs and lost through metabo-lism (Lindeman 1942, Duarte and Cebrian 1996). This has led marine biologists to use the phytoplankton carbon biomass in the oceans to set the upper limits on fi shing yields (Walsh 1981, Falkowski et al. 1998, Behrenfeld et al. 2001). The

0 50 100 150 200 2500

50

100

150

200

250

DNO3 (m)

DD

CM (m

)

Figure 3. The depth of the DCM is coupled to the depth of the ni-tracline (NO3 ≥ 1 μM) across the world oceans. Data is from the Galathea 3 expedition in 2006-2007. The fi gure is drawn by Jør-gen Bendtsen (unpublished).


SYNOPSIS

relationship between PP and phytoplankton carbon biomass is, however, not con-stant. In addition, the picture becomes more complicated when including the fact that chlorophyll is not always a good predictor of phytoplankton carbon biomass and that there is not a constant relationship between fl uorescence and chlorophyll (Cullen 1982, Fennel and Boss 2003). Therefore, one must have knowledge on the specifi c characteristics of the PP to biomass ratio, carbon to chlorophyll ratio and light adaptations/acclimatization of the phytoplankton community (or single cell) (Richardson et al. 1983, Lindley et al. 1995, Morán and Estrada 2001) in the area of interest to be able to use phytoplankton carbon, chlorophyll or fl uores-cence to make any predictions on the amount of food available for higher trophic levels in an ecosystem or PP. The varying relationship between the different bio-logical parameters emphasizes the importance of working directly with PP in the process of understanding carbon dynamics in our oceans.

The word plankton is defi ned as free fl oating organisms in lakes and oceans. Some plankton are, however, capable of movement (see Cullen and Horrigan, 1981 for lab experiments and Eppley et al. 1968 for motile phytoplankton in the real en-vironment) and are able to control their positions in the waters column – at least under some turbulence conditions (Maar et al. 2003). In addition, distinct vertical patterns in the distribution of different species has been recorded (see e.g. Mour-itsen and Richardson 2003; Ross and Sharples 2007 and references therein). We have, however, still little understanding of the underlying factors responsible for these vertical distribution patterns.

Phenotypic responses to changes in light environmentOne of the most examined characteristics of a phytoplankton community living below the surface waters may be the ability to acclimatize and adapt to the lower light inten-sities (see e.g. Morán et al. 2001; Richardson et al. 1983) found below the surface wa-ters. The production characteristics of a phytoplankton community can be described by parameters derived from a photosynthesis-irradiance curve (usually referred to as a P-E curve). These parameters can tell us about the potential photosynthetic perfor-mance of the phytoplankton. The P-E curve can, among other parameters, provide us with a maximum photosynthesis rate (Pmax) and a slope that tells us how fast the pho-tosynthetic centers react to increasing light at low light intensities (α). The intercept on the graph where these two parameters cross is the light intensity where photosyn-thesis initially is light saturated (Ik). Subsurface PP is often characterized by display-ing higher alpha values compared to surface water samples (see e.g. Navarro et al. 2006). This is usually interpreted as an indication that the phytoplankton communities in the deeper water layer are more capable of utilizing the low light intensities found in subsurface waters than surface populations.

The way phytoplankton optimizes their light uptake in low-light environments is, among other things, by placing their chlorophyll in the outermost layer to catch as much light as possible. In addition, a strong positive relationship be-tween the C:CHL ratio and irradiance has been found (when nutrients are not growth limiting) (Bannister and Laws 1980, Goldman 1980) which means that phytoplankton can increase their overall amount of chlorophyll to compensate for the lower amount of light received in subsurface waters as opposed to sur-face waters.

That phytoplankton carbon biomass and chlorophyll does not always peak at the same depths was already shown in the 1950s (see Cullen 1982) and multiple ex-planations have been given since then (Steele and Yentsch 1960, Steele 1964, Lerman et al. 1974, Bienfang and Harrison 1984). Fennel and Boss (2003) argued that the theory presented by Steele in 1964 saying that the DCM was created by an increased chlorophyll to carbon ratio must be the most likely explanation


SYNOPSIS

given that sinking is unlikely to be an important factor in oligotrophic oceans where phytoplankton are often motile. The suggested increase in chlorophyll per biomass at low light levels (i.e. photoacclimation) is based on studies showing that the carbon to chlorophyll ratio is inversely related to mean irradiance (Steele 1964, Kiefer and Kremer 1981, Geider et al. 1997). The maximum in phytoplank-ton biomass was suggested to be controlled by loss processes (e.g. sinking of cells, grazing, cell lysis and advection) in addition to light and situated at the general compensation depth (i.e. where the growth rate is balanced by loss pro-cesses). Even though biomass (i.e. carbon or chlorophyll) have been used in mod-els of PP (Longhurst et al. 1995, Behrenfeld and Falkowski 1997, Sathyendranath et al. 2009) there is, clearly, a need to examine this in further detail if we want to explain changes in PP with changes in chlorophyll or carbon biomass.

Genotypic responseThe adaptation to low light intensities within a phytoplankton community is most visible in the species composition. Richardson et al. (1983) suggested on the basis of a literature review that algal classes rank by light preferences (from lowest to highest light intensities) as dinofl agellates < diatoms < chlorophytes. The species composition of the phytoplankton communities living in subsurface waters may, therefore, be expected to change as the community adapts to lower light intensities.

Dinofl agellates show, in general, a higher C:CHL ratio and a lower growth rate than diatoms (Chan 1980). The dinofl agellates are also the group with lowest light pref-erences. As might be expected from the low-light adaptation among dinofl agellates, they often are an important component of DCMs (see e.g. Holligan and Harbour 1977, Karlson et al. 1996, Mouritsen and Richardson 2003, manuscript II). Here, it can be noted that some chlorophyll-containing dinofl agellates can also use or-ganic matter as a source of carbon or nutrients i.e. mixotrophy (Carlsson et al. 1998, Legrand and Carlsson 1998, Jeong et al. 2005, Loureiro et al. 2009). The impor-tance of mixotrophy among dinofl agellates seems to increase along with an increas-ing amount of research being carried out on the subject (see review by Jeong et al. 2010). This is indeed interesting, because if subsurface phytoplankton communities often are dominated by dinofl agellates, this means that the ratio of mixotrophy to autotrophy possibly will change compared to communities where other phytoplank-ton groups are abundant such as diatoms. An altered mixotrophy/autotrophy ratio may have implications for the ecosystem as a whole. Especially when trying to evaluate or estimate the carbon fl ow via PP related to a phytoplankton community, major errors will occur when that community is partly mixotrophic and most prob-ably it will result in underestimations of growth in terms of carbon. Obviously, it is important to examine the vertical distribution of phytoplankton species/groups to understand the vertical distribution of PP in a pelagic system.

Eutrophication and primary productionHuman activities have caused nutrient enrichment in coastal waters in many areas of the world. This activity has increased markedly and continues to increase with the rising amount of people that inhabit the planet (Steffen et al. 2007). The hu-man infl uence through nutrient loading will increase in the future with continued population growth especially in coastal areas coupled with economic growth (Til-man et al. 2001, Millennium_Ecosystem_Assessment 2005).

It is well established that anthropogenic nutrient enrichment of coastal seas can lead to eutrophication effects, including increases in total PP (Nixon 1995, Smith 2003), altered biogeochemical functioning and biological community structure of coastal waters (Cloern 2001) such as increases in phytoplankton biomass (Ryther and Dun-stan 1971, Beman et al. 2005) and changes in community composition which often


SYNOPSIS

leads to harmful algal blooms (Paerl 1988, Heisler et al. 2008) or changes in biomass stoichiometry (Downing 1997, Klausmeier et al. 2004). This increase in organic ma-terial leads, in some cases, to hypoxia or anoxia (see e.g. Rabalais et al. 2002) owing to the oxygen demand in the degradation of this material. Hypoxia has been linked to eutrophication worldwide (Diaz and Rosenberg 2008) and especially in the Baltic Sea including the transition zone (Conley et al. 2002a, Conley et al. 2002b).

The PhD workThis PhD includes fi ve papers that all deal with different aspects of the vertical distribution of primary production and/or phytoplankton species and relate these distributions to ecosystem function or biogeochemical fl uxes. Two of the papers derive from cruises in which I participated in the Sargasso Sea and in the North Atlantic. The major focus of this thesis, however, was the Baltic Sea Transition Zone and three manuscripts in the thesis derive from this work. Background infor-mation and perspectives relating to these three papers are given in the following.

Studies on the importance of subsurface PP for ecosystem structure and function, addressing subjects as oxygen concentration, biogeochemical cycling and trophic interactions, are often based on interpolated single measurements that are made into annual estimates or scaled up from discrete sampling to more general fi g-ures. The purpose of this PhD study was to consider how the vertical distribution of PP and phytoplankton species infl uences pelagic ecology and biogeochemical cycling. The main focus of the PhD was on a reanalysis of monitoring data in the Baltic Sea transition zone (BSTZ) which is an area heavily infl uenced by eu-trophication (see e.g. Conley et al. 2002).

The study areaThe transition zone between the Baltic Sea and the Skagerrak (BSTZ) (see map in Figure 4) is a highly dynamic area with shallow waters (up to 80 m). The shal-lowest and narrowest parts are situated at the Belts where mixing is enhanced, increasing the oxygen concentration in the waters (Bendtsen et al. 2009). The area exhibits a more or less constant pycnocline because it receives high saline waters from the Skagerrak (from North) and above this layer comes the waters from the Baltic Sea which is relatively low in salinity (see Figure 5).

The BSTZ experiences frequent hypoxia (Conley and Josefson 2001) which re-sults in negative ecosystem effects such as death of benthic dwelling organisms (Fallesen et al. 2000), reductions in the depth distribution of macroalgae (Sand-Jensen et al. 1994) and, possibly, increases in reports of harmful algal blooms (Kaas et al. 1999). Whether there is a direct link between eutrophication and oc-currence of harmful algal blooms or if the apparent increase in the occurrence of harmful algal blooms is a natural result of increased phytoplankton production in general can, however, be discussed (Richardson 1997). The consequences of reoccurring hypoxia in Danish waters motivated the government to enact an ac-tion plan for the aquatic environment with the overall goal of reducing nitrogen loading by 50 % and point source phosphorous loading by 80 %.

In conjunction with the fi rst legislation to reduce anthropogenic nutrient loading to coastal waters, a national aquatic monitoring program (Danish National Aquat-ic Monitoring and Assessment Program, DNAMAP) was begun in 1989 (Conley et al. 2002b). The fi gure below (Figure 6) shows how land-based N-loading to the inner Danish waters have been reduced considerably since the early 1990s (Windolf et al. 2012).


SYNOPSIS

PP in marine waters is one of the biological parameters measured in the DNAMAP. Sev-eral studies in the region from before 1998 indicate that the PP taking place in or below the pycnocline contributes substantially to the annual PP (see Richardson and Christof-fersen, 1991; Björnsen et al. 1993).

PP from the monitoring program was, be-fore 1998, based on measurements on the relationship between photosynthesis and light (P-E curves) from one depth in the sur-face only. Thus, before 1998, PP may have been underestimated due to the fact that subsurface PP was not taken into consid-eration. Carstensen et al. (2003) suggested that the underestimation was in the range of 35 %. The procedure for measuring and calculating PP was, therefore, modifi ed in 1998 (Markager 1998) and the estimate of PP was no longer based on measurements from one depth only but on two depths; one in the surface mixed layer (usually at 0-10 meters depth) and one from the layer in or below the pycnocline in the depth of the deep chlorophyll maximum (DCM).

The DNAMAP has opened up a great opportunity to study the general patterns of PP based on a continuous datasets of more than a 1000 PP profi les measured with identical methods and procedures during the years from 1998 to 2012. The verti-cal distribution of PP was examined in this PhD in relation to nitrogen loading, surface water nutrient concentrations, chlorophyll distribution patterns, oxygen concentration in bottom waters, phytoplankton communities and sedimentation patterns. The results of these analyses are presented in manuscripts I, II and III.

Figure 4. The Baltic Sea transition zone (BSTZ) and the location of the six stations included in the studies on survey program data.

Ska

gerr

ak

Nor

ther

nK

atte

gat

Cen

tral

Kat

tega

t

Sou

ther

nK

atte

gat

Gre

at B

elt

Wes

tern

Bal

tic S

ea

St. 905 St. 413 St. 925 St. GB0 m

20 m

40 m

60 m

80 m

O2

halocline

O2

Skagerrak water

Baltic water

Sea bottom

100 km

Figure 5. Schematic illustration showing the outlines of the hy-drodynamics in the Baltic Sea transition zone. Arrows indicate oxygen sources. The fi gure is re-drawn from Hansen and Bendtsen (2013).


SYNOPSIS

Quantifying subsurface primary production

The mixed layer depth (MLD) has, for more than a half century, been considered as a critical parameter linking hydrographic conditions to PP. The relevance of the MLD for PP, was fi rst identifi ed by Sverdrup (1953) who showed that if the MLD was deeper than the critical depth, PP became limited by light and the overall production decreased. Since then, it has been generally accepted that stabilization of the water column is a prerequisite for a high PP and has also shown to be the initiator of vernal production in temperate regions (Townsend and Thomas 2002, Ross and Sharples 2007). The theory presented by Sverdrup was based on the MLD but this does not directly translate to the three-layer water column structure often observed in the BSTZ.

The division of the water column can be done in many different ways (see Thom-son and Fine 2003) and one must be precise on what exactly this division is used for. In this study, the important depth for the analyses was where there was a physical division (density barrier) that made mixing of material (phytoplankton and oxygen) between the layers very limited or unlikely. This was motivated by the desire to address the question; how much does carbon fi xation and oxygen production, in the part of the water column that is not able to directly exchange oxygen with the atmosphere (i.e. the pycnocline/bottom layer (PBL), see below) contribute to bio-geochemical properties of the water column?

Other studies (see e.g. Bendtsen and Hansen et al. 2013) have assumed that signifi cant carbon fi xation and oxygen production is not occurring in the PBL. In this study, the depth of the beginning of the PBL was identifi ed through a potential density criterion (Δρ/Δz > 1 kg m-4). The suitabil-ity of this criterion was tested by manually going through 256 profi les of plots where oxygen, PP, chlorophyll and density could be evaluated in relation to each other (see Figure 7 for an example of how the water column was divided). The process indicated that the density criterion gave a good match to what was expected on the basis of visual examination of the water column profi les in 90 % of the cases. In the remaining 10 % of the cases, the den-sity criterion (Δρ/Δz > 1 kg m-4) overestimated the start-ing depth of the PBL, i.e. placed it approximately 1-2 m deeper than the depth identifi ed by visual inspection of the profi les.

Year

Nitr

ogen

load

(kt N

yea

r-1)

Nitrogen load/runoff (m

g N l -1)

N-loadingN-loading normalised to runoff

1990 94 96 9892 00 02 04 06 08 10 201210

20

30

40

50

60

70

80

4

5

6

7

8

9

10

11Figure 6. Land-based N-loading in kilo tons and land-based N-loading normalized to runoff du-ring the time period from 1990 to 2012. The data is from the Danish National Aquatic Monitoring and Assessment Program.

Density ρ(kg m-3)

Aarhus Bight: date 2/10 2000

10 12 14 16 18 20 22 24

Dep

th, z

(m)

-16

-14

-12

-10

-8

-6

-4

-2

0

Density, ρΔρ/Δz > 1 kg m-4

Surface layer

Pycnocline- andbottom layer

Figure 7. The starting depth of the pycnocline/bottom layer (Δρ/Δz > 1 kg m-4) is noted with the blue stippled line. The densi-ty profi le is shown for reference. This depth divides the water col-umn into two layers; the surface layer and the pycnocline/bottom layer. The profi le is from Aarhus Bight (date: 02/10 2000).


SYNOPSIS

A physical oceanographer might have divided the water column differently, most likely defi ning more than two layers. The difference between the perspectives of a physical oceanographer and a biological oceanographer may be the time scale of which they consider the subject. The physical approach looks at the immediate structure of the water column whereas the ecological perspective has to include longer time periods to evaluate what effect the apparent water column structure has on the biology. An example could be a day with no wind where a weak gradi-ent is observed. This gradient divides the water masses physically but it only takes a little increase in wind before this division is mixed away. The surface layer, is therefore, considered homogenous in the present study of BSTZ waters which focus on the ecology.

The DCM has been defi ned in many different ways in different studies. That the defi nition of DCM differs between studies complicates comparison of results across regions or even within the same region. Out of the 11 published studies on subsurface PP in stratifi ed waters mentioned above, none had the same defi nition of a DCM, 4 of the studies did not even present a defi nition. Some of the studies defi ned the DCM as being where the maximum chlorophyll concentration in rela-tion to the mean value was observed and others used a criteria value i.e. a chl. a conc. > 0.5 mg Chl l-1. Others again used a chlorophyll criteria value relative to the mean for the entire profi le or just surface waters. Hence, it does not seem like there is one way to defi ne a DCM which is a problem when trying to quantify importance of same on a larger scale.

For the analyses presented here, the water column was divided into two layers; one layer ranging from the surface to the beginning of the pycnocline, and another layer ranging from the beginning of the pycnocline and to the bottom of the water column, i.e. PBL (see Figure 7, the blue line). We, then, defi ned all PP occurring in the PBL as being deep PP (DPP) and all PP above this layer to be surface PP. This simple approach to the quantifi cation of DPP in stratifi ed regions gives a ro-bust estimate that is independent on how and if the DCM is defi ned and includes DPP even though a DCM as such is not present.

The fi ndings of this PhD workThe vertical characteristics of phytoplankton production have been examined in three different contexts in this PhD to gain knowledge on how the pelagic system works: 1) Eutrophication in the BSTZ, where eutrophication and hypoxia is a problem of national concern infl uencing fi sheries and the ‘quality’ of the water. Here, it has traditionally been assumed that the link between nutrient enrichment and hypoxia is an increase in the magnitude of organic material being produced leading to greater biological oxygen demand (BOD) in bottom waters. It is shown here that N-loading can be related to the vertical distribution of PP and that this distribution pattern is important in relation to changes in sedimentation of or-ganic material and O2 production in bottom waters (manuscripts I, II and III). 2) Fisheries ecology in the Sargasso Sea, where eels are known to spawn and larvae have been found to concentrate in frontal zones. We still do not know the under-lying mechanisms for the distribution of these eel larvae and previous studies have not been able to demonstrate any production characteristics in these regions that might suggest that food availability/plankton production here provides bet-ter feeding conditions for eel larvae than outside the frontal zones (Andersen et al. 2011, Riemann et al. 2011). In this study, it is shown that, within the frontal zone, itself, there are differences in production characteristics that are driven by physical processes that stimulate deep primary production. 3) Biological pump in the North Atlantic, where it has earlier been shown that the spring bloom where surface water phytoplankton are abundant, contributes to the transfer of carbon


SYNOPSIS

to depth. In the study presented here, it is shown that the most important carbon transporter at that time was a species (Chaeotoceros aff. diadema) which was only a minor component of the phytoplankton community in the surface waters. My own contribution to that study was the Fv/Fm that showed that resting spores maintained the capacity of photosynthesis to great depth and that there were no differences between the depths suggesting very rapid sinking.

Eutrophication The effect of anthropogenic nutrient loading on PP has, to date, only been evalu-ated based on the total water column PP. This approach ignores that there is more than one environment for phytoplankton in the water column with respect to light and nutrients. Danish waters are stratifi ed most of the year, which segregates the bottom waters from the remaining water column. Hence, bottom water nutrients are constrained from being mixed up and into the often nutrient depleted surface waters during summer time. Likewise, oxygen produced in bottom waters will not generally mix into surface waters, where it can exchange with the atmosphere. This gives reason to believe that the PP in the upper and lower parts of the water column may respond differently to nutrient enrichment and that it will make a dif-ference for water column geochemistry where PP occurs.

It is demonstrated here (manuscript I) for the BSTZ that the magnitude of land-based N-loading correlates with the vertical distribution of PP whereby the pro-portion of PP occurring in the surface waters (the waters above the fi rst density difference > 1 kg m-4) increases with increasing nitrogen runoff. At the same time, the magnitude and the proportion of PP occurring in the pycnocline/bottom layer (PBL, the waters below the fi rst density difference > 1 kg m-4) decreases with increased nitrogen loading. This is an interesting observation – not least of which in light of the fact that both in the BSTZ and many other temperate coastal seas, it has been demonstrated that subsurface PP often contributes signifi cantly to total annual PP (Richardson and Christoffersen 1991, Richardson and Pedersen 1998, Richardson et al. 2000, Weston et al. 2005, Hickman et al. 2012).

With the exception of cases in which eutrophication is suspected to be related to harmful algal blooms (HABs), there has generally not been much focus on how nutrient enrichment may change species composition or the structure of plank-tonic ecosystems. The fi ndings in manuscript I, i.e., that land-based nitrogen load-ing changes the vertical distribution of PP and that this most likely will lead to an increase in DPP in the future (assuming a continued reduction in anthropogenic nutrient loading), emphasizes the importance of knowledge and understanding of the species composition and structure of the phytoplankton community in surface and PBL waters.

Thus, the results presented in manuscript I opened further lines of query as to how changes in DPP may affect the ecosystem dynamics as a whole. The following questions were addressed; 1) how do the photosynthetic characteristics here in the PBL compare to those in surface waters? 2) does the species composition of the phytoplankton community in the PBL differ from that in the surface waters?, 3) are there differences in the sedimentation patterns in the two layers that may give rise to altered overall export of organic material to bottom waters? 4) is DPP a signifi cant source to oxygen in bottom waters of the BSTZ?

To answer these questions, a location in Aarhus Bight in the southern Kattegat was chosen for a fi eld study to supplement the data available in the long-term monitoring data (including phytoplankton taxonomy) (see Figure 4 for station


SYNOPSIS

location). The area is stratifi ed most of the year and DPP contributes about 21 % to the annual PP at the Aarhus Bight station (Manuscripts I and II).

The phytoplankton community in the PBL waters exhibited photosynthetic char-acteristics which indicated that they were adapted/acclimatized to this deeper and darker environment (manuscripts I, II and IV). The phytoplankton community in the surface layer was very distinct from the community in the PBL which also was expected on the basis of a previous study carried out at the station (Mouritsen and Richardson 2003). The Mouritsen and Richardson study was, however, carried out on only one day. That the PBL layer was segregated from the water column above was further established by hydrographical parameters which showed that the fl ow speed and direction of the waters in the PBL was completely different from the layer(s) above where the direction of the water in the PBL could be directly opposite that in the surface waters. The deep layer showed a signifi cantly higher frequency of dinofl agellates (including mixotrophic species) than the surface layer commu-nity, which mostly was comprised of diatoms. Furthermore, size fractionated chlo-rophyll determinations showed that the largest phytoplankton were found in the upper part of the PBL where light and nutrients both were presumably suffi cient for phytoplankton growth. Sedimentation rates were greatest in the PBL. Thus, several parameters suggested that the water column was segregated into two systems where the phytoplankton in the PBL exhibited characteristics quite different from those in the surface waters with respect to community composition and cell size. Nutrient availability and light intensities also differed between the two layers.

In addition, the DPP was correlated to the oxygen concentration in the PBL sug-gesting that DPP contributes substantially to the oxygen in this layer. It is esti-mated here that the DPP can contribute 37-44 % of the annual oxygen demand (manuscript II). Shulenberger and Reid argued already back in 1981 that oxygen was produced in situ by phytoplankton living in deeper waters and that this fea-ture was widespread all over the global oceans. A part from the study presented in manuscript II, this has been followed up by several other studies (Jenkins and Goldman 1985, Lehrter et al. 2009, Murrell et al. 2009) where Strom et al. (2010) have even showed that subsurface PP can ameliorate hypoxia.

The results from manuscript II confi rmed our hypothesis, that PP in the PBL is an important source of oxygen here and that it can be based on different phyto-plankton communities than those in the surface waters. In addition, the analysis showed that sedimentation patterns differed in the PBL and surface waters. This suggests that a higher relative PP in the PBL may increase the overall export of organic material to the bottom but at the same time produce oxygen to this deeper layer. These two processes represent two oppositely directed mechanisms with respect to the prediction of future bottom water oxygen concentrations. The oxy-gen consumption associated with material produced in the PBL must, however, be balanced by the oxygen production here. This means that although a greater percentage of the material from the PBL compared to that produced in the surface waters probably reaches the bottom, it will not contribute to a net loss of oxygen from this layer. All of the material coming from the surface layer will, however, lead to net oxygen removal from the PBL. We, therefore, argue that an increase in DPP during times with less nitrogen loading will have a positive effect on the oxygen conditions in bottom waters.

That more PP takes place deeper in the water column also improves the transpar-ency of the upper part of the water column. This water transparency seems to be a key parameter in the process of ecosystem recovery from eutrophication. This relationship between nutrient loading and state of the ecosystem, with light being a key parameter, was previously suggested for a Fjord in Denmark where reduced


SYNOPSIS

N-loading also here resulted in a vertical redistribution of PP with increased con-tribution from macroalgae to the total production (Krause-Jensen et al. 2012).

Effects of the legislation established in an effort to reduce anthropogenic N-loading to coastal waters on the ecosystem has been lively discussed in various Danish me-dia. Recently, it was argued, in a public meeting at the Danish Parliament the 17th of September 2012 (http://www.ft.dk/Folketinget/udvalg_delegationer_kommissioner/Udvalg/Udvalget_for_foedevarer_landbrug_og_fi skeri/ElektroniskDebat.aspx) that Danish land-based N-loading did not have an effect on the production of organic mat-ter or oxygen in the open waters of the BSTZ. This was argued on the basis of model studies that indicated that the N-loading from Denmark was unimportant relative to the N-loading coming from the adjacent waters i.e. the Skagerrak and the Baltic Sea (Erichsen and Møhlenberg 2011).

The study reported here, which is based on actual data collection and not model studies, demonstrates however, that land-based N-loading from Denmark has a clear effect on the PP in the BSTZ and it especially has an infl uence on the verti-cal distribution which again has been shown to be important to the ecosystem as a whole. That previous studies did not fi nd this clear relationship may be because the total water column PP was considered instead of evaluating the PP as being a combination of PBL and surface water PP. Clearly, the N-loading from the waters adjacent to the BSTZ will also infl uence the PP occurring in the BSTZ but this is not examined in this PhD work.

The importance of increased deep primary productionThe BSTZ ecosystem responses suggested by this study to different levels of ni-trogen loading are schematically depicted in Fig. 8. Under both a situation with large and decreased land-based nutrient loading, the water column is strongly stratifi ed (shown by a strong density gradient). The nitrogen loading is high in box A. Here, almost all the PP occurs in the surface layer and light only reaches to a little more than half the depth of the water column. Oxygen in the system comes from mixing atmospheric oxygen into the surface waters, advection from the adjacent waters (e.g. Skagerrak/North Sea) and PP in the surface layer. How-ever, the source of oxygen generated by PP also generates an oxygen demand be-cause of respiration from heterotrophic activity and the phytoplankton themselves (auto- mixo- and heterotrophic). Remineralization of organic matter back into nutrients takes place in the total water column and there is, overall, sedimentation of organic material from all layers of the water column to the bottom sediment. Box B represents a water column where the nitrogen loading is less than in box A and, as a result, the PP takes place deeper down in the water column and light can potentially reach all the way down to the bottom. Oxygen sources are the same as in box A except for an additional source in the waters below the density gradient generated by the DPP. Remineralization is also here taking place in the total water column. Note, however, that there is a higher sedimentation rate of organic mate-rial to the bottom in box B than in box A, which might provide a high quality food source for benthos (see e.g. Josefson et al. 1995).

The change in the vertical distribution of PP into a higher DPP versus surface layer PP has, in this study, been shown to be a probable response to decreased nitrogen loading. With lower nitrogen loading, surface waters are easily depleted of nutrients and mixing of nutrients from the bottom waters and up into the surface layer is con-strained by stratifi cation. Both this vertical redistribution of PP and the decreased loading result in a lower light attenuation which allows the light to reach deeper down in the water column. Establishment of larger vegetation, i.e. macroalgae, is possible when light reaches the bottom (Krause-Jensen et al. 2012) and large preda-


SYNOPSIS

tors such as fi sh better can see their prey (Aksnes et al. 2004, Aksnes 2007). This is indicated in box B by the inclusion of macroalgae and fi sh. Sedimentation rates increase when a higher fraction of the total water column PP is taking place below the density gradient. This is due to relatively larger cells in the DPP and a shorter transport time to the bottom simply giving less time for heterotrophic activity to ‘use’ the material. Hence, the water column described in box B can, potentially, ex-port more organic material to the bottom from less nitrogen loading than in box A. Finally, DPP introduces a source of oxygen into the PBL that is not present when all PP is taking place in surface waters. The overall outcome of increased DPP relative to surface PP with respect to oxygen is believed to have a positive effect in the PBL given that organic material produced in the PBL cannot contribute extra to the BOD without also contributing to the oxygen production.

Most focus in reducing eutrophication effects in coastal waters has been on reduc-ing total water column PP. This study suggests that total PP is not as responsive to changes in land-based nutrient loading as initially expected. This is because when surface water nutrients become exhausted, the water transparency increases and phytoplankton are able to carry out photosynthesis on the bases of the nutrients found below the surface waters. Thus, rather than a reduction in land-based load-ing leading to a reduction in total PP, it leads to a vertical redistribution of PP.

This redistribution of PP will, however, in itself, reduce the probability of hypoxia and anoxia occurring in bottom waters. Taking as a starting point the two sce-narios depicted in Fig. 8, we have two situations with the same total water column PP but differently distributed down through the water column. In box A, 90 % of the total water column PP is occurring in the surface layer, whereas in box B this is only 70 % i.e. 30 % of PP takes place as DPP.

We assume that the amount of material sedimenting is the same for the two sce-narios and that they both have a BOD in the PBL of 58 % (Fossing et al. 2002) of

A B

O2

O2

O2

O2

RR

PPDensityLight

Black arrows:

Blue arrows:

Orange arrows:

R:

Sedimentation

Oxygen in and out

Nitrogen loading

Remineralization

???

Figure 8. Schematic drawing outlining the ecosystem response to changes in nitrogen loading with respect to PP, light, sedimen-tation and oxygen. The question marks are there because we can only speculate about this, the exact link between nitrogen load-ing, vertical distribution of PP and abundance of macroalgae and larger predators (such as fi sh) is still to be examined for coast-al waters.


SYNOPSIS

PP (based on oxygen production occurring only in the surface layer). If the total water column PP is 200 mg C m-2 day-1 this yields a daily oxygen production of 114 mg O2 m-2 day-1 (using a PQ of 1). The BOD for the PBL will, for both sce-narios, then equal 66 mg O2 m-2 day-1 (58 % of 114 mg O2 m-2 day-1) but in the sce-nario outlined in box B, the fi nal oxygen demand is smaller than in box A because of the higher oxygen production occurring in the PBL. Thus, the oxygen demand in the PBL in box B will be approx. 32 mg O2 m-2 day-1 (BOD – DPP) whereas it is 1.7 times higher in the scenario outlined in box A (55 mg O2 m-2 day-1). This example illustrates the potential importance of a vertical redistribution of PP for ameliorating the hypoxia frequently observed in the BSTZ.

EcologyOne of the purposes of the reanalysis of monitoring data conducted here was to gain a better general understanding of PP dynamics in a stratifi ed region. There-fore, efforts were also made to describe the general patterns (seasonally as well as vertically within the water column) in PP in relation to species composition of phytoplankton communities and chlorophyll distribution patterns.

The general distribution of PP in time and space showed characteristic patterns where the highest PP was found during summer months largely supported by a DPP. Thus, nutrient depletion in surface waters was not able to dampen the productivity in this region. The DPP was, on average over the study period, responsible for 17 % of the annual PP at the six different stations (manuscripts I and II). The vertical distribution of PP was tightly coupled to the surface water nutrient concentrations (dissolved inorganic nitrogen and phosphorous, DIN and DIP) and frequency of stratifi cation where low surface nutrient concentrations were coupled to high DPP.

The characteristic seasonal pattern of PP can be described from winter, summer and spring values because it resembles a sinus curve with the top being the peak summer PP. Spring production occurred in February and March and these were the only months where the general seasonal patterns deviated from the sinus curve. Spring production contributed on average less (10–18 %, average = 14 %) to the annual PP than the DPP (6–30 %, average = 17 %) where spring production was a sporadic event and DPP a consistent feature throughout the summer season.

The chlorophyll was seasonally decoupled from PP when looking at average monthly values where the summer season showed high PP and low chlorophyll. As seen from Figure 9, a far stronger relationship was found between the monthly average values of PP and carbon biomass than between the PP and chlorophyll, and the decoupling was mainly attributed the seasonal variance in the carbon to chlorophyll ratio. This further suggests that the spring biomass peak may have received more attention than it has deserved given that the often described spring peak in phytoplankton biomass may, in some regions, have been nothing more than an artifact of evaluating chlorophyll units instead of carbon units.

A decoupling between PP and chlorophyll was also found when looking vertically within the water column. Here, the depth of the peak in PP was evaluated relative to the depth of the peak in chlorophyll. Results showed that the peak in PP was situated above the peak in chlorophyll in 85 % of the profi les. This was suggested to be explained mainly by three mechanisms; photosynthesis-light response, ac-climation/adaptation and prey selective feeding by zooplankton. Potential grazing from copepods and protozooplankton showed that the most intensive impact from grazing on primary producers occurred during summer and it was suggested that this high summer grazing partly was fed by phytoplankton produced in the PBL i.e. DPP contributing up to 50 % of the total water column PP during summer


SYNOPSIS

months. The importance of the spring bloom in the BSTZ was downscaled in relation to copepod grazing, because this production took place several months before the peak in copepod biomass and potential grazing. That no consistent re-lationship was found between PP and chlorophyll in vertical space or time within the comprehensive dataset analyzed emphasizes the importance of evaluating PP rates instead of state variables such as e.g. chlorophyll or carbon biomass when examining the ecosystem functioning.

The purpose of manuscript III was also to provide a quantitative description of the general distribution patterns of PP in this region which could be used to im-prove modeling efforts which include PP. This is currently being done using a 3D coupled hydrodynamic-biogeochemical model. The model is being improved with the use of the general patterns in PP described in manuscript III. The model simulates photosynthetic growth in carbon units instead of nitrogen units which is the case for most dynamic ecosystem models (Neumann 2000, Neumann et al. 2002, Kuznetsov et al. 2008). Therefore, photosynthesis continues although nutri-ents are depleted. Thus, the PP curve forms a shape similar to the light curve. The model study evaluates the importance of nutrient loadings for PP in the BSTZ by conducting scenarios of changes in Danish nutrient loadings and, also here, the importance of using rates rather than state variables to model ecosystem processes is emphasized. Unfortunately, the model is still running and the results can, there-fore, not be presented in this PhD work.

As in the BSTZ, the Sargasso Sea is an area where the vertical distribution of PP is important to the total (Cox et al. 1982, Steinberg et al. 2001, Casey et al. 2007, Rie-mann et al. 2011). Surface waters (i.e. the part of the water column above the main density difference or pycnocline) is nutrient depleted and the only apparent source to nutrients comes from below. The only exception to this being nitrogen fi xation (see e.g. Bruhn et al. 2010 and references therein) which, however, only is an option for a limited number of species. During our examination of the production charac-teristics in the Sargasso Sea, we found heterogeneity among phytoplankton in the DPP showing a varying physiological state (indicated by Fv/Fm) apparently owing to the introduction of nutrients here (see manuscript IV). So, while changes in the BSTZ production patterns occurring over time and space in the surface layer is due to sporadic nitrogen input from land-based loading, we show that sporadic nitrogen input also changes production characteristics in the Sargasso Sea. Only here it is occurring on a fi ne scale within a frontal zone placed in the deep. That this might occur was already suggested by Goldman (1993) who, on the basis of growth stud-

Prim

ary

prod

uctio

n (g

C m

-2 m

onth

-1)

Autotrophic phytoplankton (g C m-2) Chlorophyll a (mg Chl. m-2)0 1 2 3 4 5 6 7

0

5

10

15

20

25

30P/B Carbon P/B Chl. a

0 20 40 60 80 100 120 140

R2=0.43 R2=0.14

Figure 9. Monthly values of PP versus autotrophic carbon (left) and chlorophyll (right) from the Sound, Aarhus- and Aalborg Bight. Both relationships are signifi cant and positive.


SYNOPSIS

ies carried out in the laboratory on large diatoms collected deep in the water column in the Sargasso Sea, argued that a single deep water nutrient upwelling event could potentially lead to a bloom of diatoms that would be quantitatively important for the PP occurring here on an annual basis but undetectable at the surface.

Thus, also in the Sargasso Sea, vertical distribution of PP is an important param-eter to examine in order to understand ecosystem function. The area is quite dif-ferent from the BSTZ in that the anthropogenic input of nutrients to the system is negligible and the separation between surface and deeper waters occurs much deeper in the water column. Nevertheless, where and when nutrient additions oc-cur are critical parameters in understanding system function. Based on this study and previous studies showing that chlorophyll is poorly related to carbon biomass (Cullen 1982, Fennel and Boss 2003, manuscript III), the importance of the DPP (manuscript I, II, III) and the varying physiological state of the phytoplankton at depth (manuscript IV), satellite derived estimates of surface chlorophyll concen-trations must be concluded to be insuffi cient to examine the production dynamics of this oligotrophic system.

Biological pumpThe species distribution and ecology of phytoplankton has been shown to be of importance for carbon export, not only in the Aarhus Bight, but also in other stud-ies (Heiskanen 1993, Godhe et al. 2001, Beaulieu 2002, Fujii and Matsuoka 2006, Pospelova et al. 2010, Salter et al. 2012). The species composition and cell size of phytoplankton has previously been shown to affect the sinking rate of organic material (see e.g. Smayda 1971) and an extreme example of this occurred in the North Atlantic during May, 2008. Manuscript V demonstrates a major particle fl ux event observed from ‘spikes’ in chlorophyll fl uorescence, backscatter and beam attenuation from the ship and four seagliders (Briggs et al. 2011). This con-siderable transport of carbon from the surface to the deep ocean was driven by a high abundance of resting spores from the diatom Chaetoceros aff. diadema found at depth. Interestingly, this species was never a major component of the phytoplankton community in the surface waters. My measurements of variable fl uorescence (Fv/Fm) showed that the phytoplankton from 300 m, 600 m and 750 m depth all were physiologically capable of photosynthesis and showed similar values indicating that the phytoplankton were not degrading during their sinking through the surface layer most likely because of the high sinking rate. Studies of the distribution of species throughout the water column suggested that the other diatoms found at the surface were probably being degraded in the surface and intermediate waters in that the percentage of total diatom carbon made up of C. diadema increased with depth.

It is argued that the C. diadema resting spores found in the sediment traps would have potentially been “good” sinkers in that they contain more carbon than veg-etative cells (French and Hargraves 1980, Kuwata et al. 1993). Thus, diatoms with a life cycle which includes resting spores may be particularly effective in the vertical transport of carbon. It has previously been believed that the forma-tion of resting spores is coupled to nutrient depletion (French and Hargraves 1980, Garrison 1981, Sanders and Cibik 1985, Pitcher 1986, Kuwata et al. 1993, Oku and Kamatani 1999). However, this hypothesis was diffi cult to apply to a system where the resting spores at depth were heavily silicifi ed while surface water silicate was depleted. The data presented in manuscript 5 indicate that the resting spores may have been formed at depths below the euphotic zone, where Si was not limiting.


SYNOPSIS

This study emphasizes that how different phytoplankton species respond to con-ditions at various depths in the water column infl uences carbon dynamics in a region and that there is a need for a better understanding of how the vertical dis-tribution of phytoplankton activity and community species composition impacts on ecosystem functioning.

In summary, knowledge on the understanding of the vertical distribution of phy-toplankton processes (e.g. species distributions, PP, resting spore formation) is essential to understand the dynamics of marine ecology including interests of the society such as carbon storage, fi sheries and eutrophication.

The next step: Ideas and thoughts for future research The results presented in manuscript II gave a brief insight into the highly produc-tive summer situation in Aarhus Bight. The patterns in PP, oxygen concentration and chlorophyll a distribution found during the fi eld campaigns matched the gen-eral patterns derived from the survey data including frequent measurements for 14 years indicating that this was not just a stochastic feature. However, there was no way to tell if the phytoplankton species distribution combined with different sedi-mentation rates found during the fi eld campaigns was typical for the area at this time or represented an exceptional situation. Further studies on the depth-specifi c sedimentation patterns combined with phytoplankton community composition taxonomy are needed to conclude if this is a general mechanism that potentially can occur other places than in the BSTZ. Furthermore, it is also very interesting whether increased DPP, as an effect of reduced nitrogen loadings, has any effect on the benthic fi lter feeders and bottom dwelling fi ltrators. In addition to this, more studies examining the species composition of phytoplankton communities in subsurface waters are also needed to evaluate the importance of mixotrophy for stratifi ed systems, where an increasing proportion of the PP occurring below surface waters can be predicted under conditions of reduced land-based nutrient loading.

Given that the vertical distribution changes in response to changing nutrient load-ings reported here is a novel result, it would be very interesting to apply this ap-proach to other areas negatively affected by anthropogenic nutrient loading. The fi nding introduces hope and encouragement to the process of restoring damaged ecosystems following the negative ecological effects that often accompany nutri-ent enrichment.

In facing the challenges of climate change and global warming we, as a society, have a responsibility to gain knowledge on the current ecosystem dynamics so that we can learn how to predict possible future changes in our oceans. One of the places where this is especially important is polar waters. Here, the effects of global warming can be seen very clearly e.g. the melting ice (Stroeve et al. 2012) and the area, therefore, literally opens op for valuable research on the effects of a changing climate.

The vertical distribution of PP is of great importance in the High Arctic because it is an area with strongly stratifi ed marine waters (Carmack 2007) and light reach-ing below the nutriclines. PP related to subsurface chlorophyll maxima contrib-utes in the order of 50 % to the new production in the High Arctic (Tremblay et al. 2008, Popova et al. 2010) and this number may increase with increasing light to these depths as a result of melting ice (Tremblay and Smith 2007) or decrease if the depth and strength of the seasonal pycnocline deepens (McLaughlin and Carmack 2010).


SYNOPSIS

What happens to the subsurface PP in these highly productive waters with changing stratifi cation? Several studies indicate that a possible response to increased tempera-tures and stratifi cation leading to nutrient depleted surface waters, ultimately, will result in a change in size distribution (Daufresne et al. 2009, Li et al. 2009, Comeau et al. 2011, Klauschies et al. 2012). Warming, in itself, can be predicted to lead to smaller phytoplankton species (Hilligsøe et al. 2011, Mousing et al. 2013). Will changes in size structure of the phytoplankton community lead to reduced carbon transfer effi ciency to the higher levels of the food web as suggested by Li et al. (2009) or Comeau et al. (2011)?. Or will it generate a shift in phytoplankton com-munity composition given that diatoms have been shown to be coupled to nutrient replete waters (Sarthou et al. 2005)?.

As Tremblay et al. (2012) argues, there is a need for a quantifi cation of the DPP in the arctic to understand the present and possibly also the future importance of the subsurface production feature. Given the results described in this study, a change in the vertical distribution of PP may have consequences for other mechanisms in the ecosystem such as sedimentation, species and size composition of the phytoplank-ton community, oxygen concentration and production and, ultimately, the higher trophic levels in the food web. Tremblay et al. 2012 showed that zooplankton graz-ers may be more independent on the timing of the spring bloom in the High Arctic than previously assumed because they are largely supported by the summer subsur-face PP which emphasizes the importance of this vertical distribution for the fi sher-ies yield in the Arctic waters upon which many societies depend on.

This study highlights the importance of vertical patterns in phytoplankton activity for every aspect in marine ecology, e.g. carbon transfer, trophic interactions, oxy-gen budgets, eutrophication trajectories and fi sheries. The study shows, in three different ocean regions exhibiting widely different pelagic systems, that what we see in the surface is a poor indicator of ecosystem dynamics. This calls for im-provements of the remotely estimated global ocean PP which cannot be done from surface chlorophyll unless the vertical dynamics of phytoplankton production is taken into account. Finally, it should be stressed that the importance of vertical distribution patterns should be accepted and emphasized in the general conceptu-alization of marine ecology.


SYNOPSIS

References for synopsis

Aksnes, D.L. 2007. Evidence for visual constraints in large marine fi sh stocks. Limnology and Oceanography 52:198-203.

Aksnes, D.L., J. Nejstgaard, E. Sædberg and T. Sørnes 2004. Optical Control of Fish and Zooplankton Populations. Limnology and Oceanography 49:233-238.

Andersen, N.G., T.G. Nielsen, H.H. Jakobsen, P. Munk and L. Riemann 2011. Distribution and production of plankton communities in the subtropical convergence zone of the Sar-gasso Sea. II. Protozooplankton and copepods. Marine Ecology Progress Series 426:71-86.

Balch, W., R. Evans, J. Brown, G. Feldman, C. McClain and W. Esaias 1992. The remote sensing of ocean primary productivity: Use of a new data compilation to test satellite al-gorithms. Journal of Geophysical Research: Oceans 97:2279-2293.

Bannister, T.T. and E.A. Laws 1980. Modeling Phytoplankton Carbon Metabolism. Pages 243-258 in P. Falkowski, editor. Primary Productivity in the Sea. Springer US.

Beardall, J., E. Young and S. Roberts 2001. Approaches for determining phytoplankton nutrient limitation. Aquatic Sciences 63:44-69.

Beaulieu, S.E. 2002. Accumulation and Fate of Phytodetritus on the Sea Floor. Pages 171-232 Oceanography and Marine Biology, An Annual Review, Volume 40. CRC Press.

Beckmann, A. and I. Hense 2007. Beneath the surface: Characteristics of oceanic eco-systems under weak mixing conditions - A theoretical investigation. Progress in Oceano-graphy 75:771-796.

Behrenfeld, M.J. and P.G. Falkowski 1997. A Consumer’s Guide to Phytoplankton Pri-mary Productivity Models. Limnology and Oceanography 42:1479-1491.

Behrenfeld, M.J., J.T. Randerson, C.R. McClain, G.C. Feldman, S.O. Los, C.J. Tucker, P.G. Falkowski, C.B. Field, R. Frouin, W.E. Esaias, D.D. Kolber and N.H. Pollack 2001. Biospheric Primary Production During an ENSO Transition. Science 291:2594-2597.

Beman, J.M., K.R. Arrigo and P.A. Matson 2005. Agricultural runoff fuels large phyto-plankton blooms in vulnerable areas of the ocean. Nature 434:211-214.

Bendtsen, J., K.E. Gustafsson, J. Söderkvist and J.L.S. Hansen 2009. Ventilation of bottom water in the North Sea-Baltic Sea transition zone. Journal of Marine Systems 75:138-149.

Bendtsen, J. and J.L.S. Hansen 2013. Effects of global warming on hypoxia in the Baltic Sea-North Sea transition zone. Ecological Modelling 264:17-26.

Benitez-Nelson, C.R., R.R. Bidigare, T.D. Dickey, M.R. Landry, C.L. Leonard, S.L. Brown, F. Nencioli, Y.M. Rii, K. Maiti, J.W. Becker, T.S. Bibby, W. Black, W.-J. Cai, C.A. Carlson, F. Chen, V.S. Kuwahara, C. Mahaffey, P.M. McAndrew, P.D. Quay, M.S. Rappé, K.E. Selph, M.P. Simmons and E.J. Yang 2007. Mesoscale Eddies Drive In-creased Silica Export in the Subtropical Pacifi c Ocean. Science 316:1017-1021.

Bienfang, P., J. Szyper and E. Laws 1983. Sinking rate and pigment responses to light-limitation of a marine diatom: implications to dynamics of chlorophyll maximum layers. Oceanologica acta 6:55-62.

Bienfang, P.K. and P.J. Harrison 1984. Sinking-rate response of natural assemblages of temperate and subtropical phytoplankton to nutrient depletion. Marine Biology 83:293-300.

Bjornsen, P.K., H. Kaas and T.G. Nielsen.1993. Dynamics of a subsurface phytoplankton maximum in the Skagerrak. Marine Ecology Progress Series 95:279-294.

Bloesch, J. and N.M. Burns 1980. A critical review of sedimentation trap technique.Schweiz. Z. Hydrol. 1:1-42.

Boesch, D. 2002. Challenges and opportunities for science in reducing nutrient over-enrichment of coastal ecosystems. Estuaries 25:886-900.

Borum, J. and K. Sand-Jensen 1996. Is Total Primary Production in Shallow Coastal Ma-rine Waters Stimulated by Nitrogen Loading? Oikos 76:406-410.


SYNOPSIS

Briggs, N., M.J. Perry, I. Cetinić, C. Lee, E. D’Asaro, A.M. Gray and E. Rehm 2011. High-resolution observations of aggregate fl ux during a sub-polar North Atlantic spring bloom. Deep Sea Research Part I: Oceanographic Research Papers 58:1031-1039.

Bruhn, A., J. LaRoche and K. Richardson 2010. Emiliana huxleyi (prymnesiophyceae): Nitrogen-metabolism genes and their expression in response to external nitrogen sources. Journal of Phycology 46:266-277.

Buesseler, K.O. 1991. Do upper-ocean sediment traps provide an accurate record of par-ticle fl ux? Nature 353:420-423.

Campbell, J.W. and J.E. O’Reilly 1988. Role of satellites in estimating primary produc-tivity on the northwest Atlantic continental shelf. Continental Shelf Research 8:179-204.

Carlsson, P., H. Edling and C. Béchemin 1998. Interactions between a marine dinofl agellate (Alexandrium catenella) and a bacterial community utilizing riverine humic substances. Aquatic Microbial Ecology 16:65-80.

Carmack, E.C. 2007. The alpha/beta ocean distinction: A perspective on freshwater fl ux-es, convection, nutrients and productivity in high-latitude seas. Deep Sea Research Part II: Topical Studies in Oceanography 54:2578-2598.

Carstensen, J., D.J. Conley and B. Müller-Karulis 2003. Spatial and temporal resolution of carbon fl uxes in a shallow coastal ecosystem, the Kattegat. Marine Ecology Progress Series 252:35-50.

Carstensen, J., M. Sánchez-Camacho, C.M. Duarte, D. Krause-Jensen and N. Marbà 2011. Connecting the dots: Responses of coastal ecosystems to changing nutrient con-centrations. Environmental Science & Technology 45:9122-9132.

Casey, J.R., M.W. Lomas, J. Mandecki and D.E. Walker. 2007. Prochlorococcus con-tributes to new production in the Sargasso Sea deep chlorophyll maximum. Geophysical Research Letters 34:L10604.

Chan, A.T. 1980. Comparative physiological study of marine diatoms and dinofl agellates in relation to irradiance and cell size. Relationship between photosynthesis, growth, and carbon/chlorophyll a ratio. Journal of phycology 16:428-432.

Cloern, J.E. 2001. Our evolving conceptual model of the coastal eutrophication problem. Marine Ecology Progress Series 210:223-253.

Comeau, A.M., W.K. W. Li, J.-E. Tremblay, C.E. Carmack and C. Lovejoy 2011. Arctic ocean microbial community structure before and after the 2007 record sea ice minimum PLoS ONE 6.

Conley, D. 1999. Biogeochemical nutrient cycles and nutrient management strategies. Hydrobiologia 410:87-96.

Conley, D.J., J. Carstensen, G. Ærtebjerg, P.B. Christensen, T. Dalsgaard, J.L.S. Hansen and A.B. Josefson 2007. Long-term canges and impacts of hypoxia in Danish coastal wa-ters. Ecological Applications 17:S165-S184.

Conley, D.J., C. Humborg, L. Rahm, O.P. Savchuk and F. Wulff 2002a. Hypoxia in the Baltic Sea and Basin-Scale Changes in Phosphorus Biogeochemistry. Environmental Sci-ence & Technology 36:5315-5320.

Conley, D.J. and A.B. Josefson 2001. Hypoxia, nutrient management and restoration in danish waters. Pages 425-434 Coastal Hypoxia: Consequences for Living Resources and Ecosystems. AGU, Washington, DC.

Conley, D.J., S. Markager, J.H. Andersen, T. Ellerman and L.M. Svendsen 2002b. Coast-al eutrophication and the Danish national Aquatic Monitoring and Assessment Program. Estuaries 25:706-719.

Cox, J.L., P.H. Wiebe, P. Ortner and S. Boyd 1982. Seasonal Development of Subsurface Chlorophyll Maxima in Slope Water and Northern Sargasso Sea of the Northwestern At-lantic Ocean. Biological Oceanography 1:271-285.

Cullen, J.J. 1982. The Deep Chlorophyll Maximum: Comparing vertical Profi les of Chlo-rophyll a. Canadian Journal of Fisheries and Aquatic Sciences 39:791-803.


SYNOPSIS

Cullen, J.J. and R.W. Eppley 1981. Chlorophyll maximum layers of the Southern Cali-fornia Bight and possible mechanisms of their formation and maintenance. Oceanologica acta 4:23-32.

Cullen, J.J. and S.G. Horrigan 1981. Effects of nitrate on the diurnal vertical migration, car-bon to nitrogen ratio, and the photosynthetic capacity of the dinofl agellate Gymnodinium splendens. Marine Biology 62:81-89.

Daufresne, M., K. Lengfellner and U. Sommer 2009. Global warming benefi ts the small in aquatic ecosystems. Proceedings of the National Academy of Sciences 106:12788-12793.

Diaz, R.J. and R. Rosenberg 2008. Spreading Dead Zones and Consequences for Marine Ecosystems. Science 321:926-929.

Dortch, Q., N. Rabalais, R. Turner and G. Rowe 1994. Respiration rates and hypoxia on the Louisiana shelf. Estuaries and Coasts 17:862-872.

Downing, J.A. 1997. Marine nitrogen: Phosphorous stoichiometry and the global N:P cycle. Biogeochemistry 37:237-252.

Duarte, C.M. 2009. Coastal eutrophication research: a new awareness. Pages 263-269 in J.H. Andersen and D.J. Conley, editors. Eutrophication in Coastal Ecosystems. Springer Netherlands.

Duarte, C.M. and J. Cebrian 1996. The fate of marine autotrophic production. Limnology and Oceanography 41:1758-1766.

Eppley, R.W., O. Holm-Harisen and J.D.H. Strickland 1968. Some observations on the vertical migration of dinofl agellates 12. Journal of phycology 4:333-340.

Erichsen, A.C. and F. Møhlenberg 2011. Effekt af Næringssaltsreduktioner i de åbne in-dre danske farvande. Modelscenarier. Danish Hydrological Institution, Copenhagen.

Estrada, M., C. Marrasé, M. Latasa, E. Berdalet, M. Delgado and T. Riera 1993. Variability of deep chlorophyll maximum characteristics in the Northwestern Mediterranean. Marine Ecology Progress Series 92:289-300.

Falkowski, P. and D.A. Kiefer 1985. Chlorophyll a fl uorescence in phytoplankton: rela-tionship to photosynthesis and biomass. Journal of Plankton Research 7:715-731.

Falkowski, P.G. 1984. Physiological responses of phytoplankton to natural light regimes. Journal of Plankton Research 6:295-307.

Falkowski, P.G., R.T. Barber and V. Smetacek 1998. Biogeochemical Controls and Feed-backs on Ocean Primary Production. Science 281:200-206.

Falkowski, P.G. and J.A. Raven 2007. Aquatic photosynthesis. Princeton University Press.

Fallesen, G., F. Andersen and B. Larsen. 2000 Life, death and revival of the hypertrophic Mariager Fjord, Denmark. Journal of Marine Systems 25:313-321.

Fennel, K. and E. Boss 2003. Subsurface maxima of phytoplankton and chlorophyll: Steady-state solutions from a simple model. Limnology and Oceanography 48:1521-1534.

Field, C.B., M.J. Behrenfeld, J.T. Randerson and P. Falkowski 1998a. Primary Production of the Biosphere: Integrating Terrestrial and Oceanic Components. Science 281:237-240.

Fossing, H., B. Thamdrup, S. Rysgaard, H.M. Sørensen and K. Nielsen 2002. Ilt- og nærings-stoffl uxmodel for Århus Bugt og Mariager Fjord. Modelopsætning og scenarier. 147.

French, F.W. and P.E. Hargraves 1980a. Physiological characteristics of plankton diatom resting spores. Marine Biology Letters 1:185-195.

Fujii, R. and K. Matsuoka 2006. Seasonal change of dinofl agellates cysts fl ux collected in a sediment trap in Omura Bay, West Japan. Journal of Plankton Research 28:131-147.

Garrison, D.L. 1981. Monterey Bay Phytoplankton. II. Resting Spore Cycles in Coastal Diatom Populations. Journal of Plankton Research 3:137-156.

Geider, R.J. 1987. Light and Temperature Dependence of the Carbon to Chlorophyll a Ratio in Microalgae and Cyanobacteria: Implications for Physiology and Growth of Phy-toplankton. New Phytologist 106:1-34.


SYNOPSIS

Geider, R.J., H.L. MacIntyre and T.M. Kana 1997. Dynamic model of phytoplankton growth and acclimation: responses of the balanced growth rate and the chlorophyll a:carbon ratio to light, nutrient-limitation and temperature. Marine Ecology Progress Se-ries 148:187-200.

Godhe, A., F. Norén, M. Kuylenstierna, C. Ekberg and B. Karlson 2001. Relationship between planktonic dinofl agellate abundance, cysts recovered in sediment traps and en-vironmental factors in the Gullmar Fjord, Sweden. Journal of Plankton Research 23:923-938.

Goldman, J. 1980. Physiological Processes, Nutrient Availability, and the Concept of Relative Growth Rate in Marine Phytoplankton Ecology. Pages 179-194 in P. Falkowski, editor. Primary Productivity in the Sea. Springer US.

Goldman, J.C. 1993. Potential role of large oceanic diatoms in new primary production. Deep Sea Research Part I: Oceanographic Research Papers 40:159-168.

Grasshoff, K., K. Kremling and M. Ehrhardt 1999. Methods of Seawater Analysis. 3 edi-tion. Wiley-VCH.

Gustafsson, B.G. 2000. Time-Dependent Modeling of the Baltic Entrance Area. 2. Water and Salt Exchange of the Baltic Sea. Estuaries 23:253-266.

Hansen, H.P. and F. Koroleff 1999. Determination of nutrients. Pages 159-228 in K. Grasshoff, K. Kremling and M. Ehrhardt, editors. Methods of seawater analysis, Ger-many.

Hansen, J.L.S. and J. Bendtsen 2013. Parameterisation of oxygen dynamics in the bot-tom water of the Baltic Sea – North Sea transition zone. Marine Ecology Progress Series 481:25-39.

Heiskanen, A.S. 1993. Mass encystment and sinking of dinofl agellates during a spring bloom. Marine Biology 116:161-167.

Heisler, J., P.M. Glibert, J.M. Burkholder, D.M. Anderson, W. Cochlan, W.C. Dennison, Q. Dortch, C.J. Gobler, C.A. Heil, E. Humphries, A. Lewitus, R. Magnien, H.G. Mar-shall, K. Sellner, D.A. Stockwell, D.K. Stoecker and M. Suddleson 2008. Eutrophication and harmful algal blooms: A scientifi c consensus. Harmful Algae 8:3-13.

Hickman, A.E., C.M. Moore, J. Sharples, M.I. Lucas, G.H. Tilstone, V. Krivtsov and P. M. Holligan 2012. Primary production and nitrate uptake within the seasonal thermocline of a stratifi ed shelf sea. Marine Ecology Progress Series 463:39-57.

Hilligsøe, K.M., K. Richardson, J. Bendtsen, L.-L. Sørensen, T.G. Nielsen and M.M. Lyngsgaard 2011. Linking phytoplankton community size composition with temperature, plankton food web structure and sea-air CO2 fl ux. Deep Sea Research Part I: Oceano-graphic Research Papers 58:826-838.

Holligan, P.M. and D.S. Harbour 1977. The vertical distribution and succession of phy-toplankton in the western English Channel in 1975 and 1976. Journal of the Marine Bio-logical Association of the United Kingdom 57:1075-1093.

Holligan, P.M., P.J.l. Williams, D.A. Purdie and R.P. Harris 1984. Photosynthesis, respi-ration and nitrogen supply of plankton populations in stratifi ed, frontal and tidally mixed shelf waters. Marine Ecology Progress Series 17:201-213.

Holm-Hansen, O. and C. Hewes 2004. Deep chlorophyll-a maxima (DCMs) in Antarctic waters. Polar Biology 27:699-710.

Jenkins, W.J. and J.C. Goldman 1985. Seasonal oxygen cycling and primary production in the Sargasso Sea. Journal of Marine Research 43:465-491.

Jeong, H., Y. Yoo, J. Kim, K. Seong, N. Kang and T. Kim 2010. Growth, feeding and ecological roles of the mixotrophic and heterotrophic dinofl agellates in marine plank-tonic food webs. Ocean Science Journal 45:65-91.

Jeong, H.J., Y.D. Yoo, J.Y. Park, J.Y. Song, S.T. Kim, S.H. Lee, K.Y. Kim and W.H. Yih 2005. Feeding by phototrophic red-tide dinofl agellates: fi ve species newly revealed and six species previously known to be mixotrophic. Aquatic Microbial Ecology 40:133-150.


SYNOPSIS

Josefson, A.B., J.N. Jensen, T.G. Nielsen and B. Rasmussen 1995. Growth parameters of a benthic suspension feeder along a depth gradient across the pycnocline in the southern Kattegat, Denmark. Marine Ecology Progress Series 125:107-115.

Jørgensen, B.B. and K. Richardson 1996. Eutrophication in coastal Marine Ecosystems. Coastal and estuarine studies 52:273.

Jørgensen, L., S. Markager and M. Maar 2013. On the importance of quantifying bio-available nitrogen instead of total nitrogen. Biogeochemistry.

Karlson, B., L. Edler, W. Granéli, E. Sahlsten and M. Kuylenstierna 1996. Subsurface chlorophyll maxima in the skagerrak-processes and plankton community structure. Jour-nal of Sea Research 35:139-158.

Kemp, M.W. and W.R. Boynton 1984. Spatial and Temporal Coupling of Nutrient Inputs to Estuarine Primary Production: The Role of Particulate Transport and Decomposition. Bulletin of Marine Science 35:522-535.

Kiefer, D.A. 1973. Fluorescence properties of natural phytoplankton populations. Marine Biology 22:263-269.

Kiefer, D.A. and J.N. Kremer 1981. Origins of vertical patterns of phytoplankton and nu-trients in the temperate, open ocean: a stratigraphic hypothesis. Deep Sea Research Part A. Oceanographic Research Papers 28:1087-1105.

Kiefer, D.A., R.J. Olson and O. Holm-Hansen 1976. Another look at the nitrite and chlo-rophyll maxima in the central North Pacifi c. Deep Sea Research 23:1199-1208.

Kiørboe, T. 1993. Turbulence, Phytoplankton Cell Size, and the Structure of Pelagic Food Webs. Advances in marine biology 29:1-72.

Klauschies, T., B. Bauer, N. Aberle-Malzahn, U. Sommer and U. Gaedke 2012. Climate change effects on phytoplankton depend on cell size and food web structure. Marine Bio-logy 159:2455-2478.

Klausmeier, C.A. and E. Litchman 2001. Algal Games: The Vertical Distribution of Phytoplankton in Poorly Mixed Water Columns. Limnology and Oceanography 46:1998-2007.

Klausmeier, C.A., E. Litchman, T. Daufresne and S.A. Levin. 2004. Optimal nitrogen-to-phosphorous stoichiometry of phytoplankton. Nature 429.

Kolber, Z., J. Zehr and P.G. Falkowski 1998. Effects of growth irradiance and nitrogen limitation on photosynthetic energy conversion in photosystem II. Plant Physiol. 88:923-929.

Kononen, K., M. Huttunen, S. Hällfors, P. Gentien, M. Lunven, T. Huttula, J. Laanemets, M. Lilover, J. Pavelson and A. Stips 2003. Development of a deep chlorophyll maximum of Heterocapsa triquetera Ehrenb. at the entrance to the Gulf of Finland. Limnology and Oceanography 48:594-607.

Krause-Jensen, D., S. Markager and T. Dalsgaard 2012. Benthic and Pelagic Primary Production in Different Nutrient Regimes. Estuaries and Coasts 35:527-545.

Kuwata, A., T. Hama and M. Takahashi 1993. Ecophysiological characterization of two life forms, resting spores and resting cells, of a marine planktonic diatom, Chaetoceros pseudocurvisetus, formed under nutrient depletion. Marine Ecology Progress Series 102:245-255.

Kuznetsov, I., T. Neumann and H. Burchard 2008. Model study on the ecosystem impact of a variable C:N:P ratio for cyanobacteria in the Baltic Proper. Ecological Modelling 219:107-114.

Kaas, H., Ø. Moestrup, J. Larsen and P. Henriksen 1999. Giftige alger og algeopblom-stringer. National Environmental Research Institute, Roskilde, Denmark.

Legrand, C. and P. Carlsson 1998. Uptake of high molecular weight dextran by the dino-fl agellate Alexandrium catenella. Aquatic Microbial Ecology 16:81-86.


SYNOPSIS

Lehrter, J.C., M.C. Murrell and J.C. Kurtz 2009. Interactions between freshwater input, light and phytoplankton dynamics on the Louisiana continental shelf. Continental Shelf Research 29:1861-1872.

Lerman, A., D. Lal and M. Dacey 1974. Stokes’ Settling and Chemical Reactivity of Sus-pended Particles in Natural Waters. Pages 17-47 in R. Gibbs, editor. Suspended Solids in Water. Springer US.

Letelier, R.M., D.M. Karl, M.R. Abbott and R.R. Bidigare 2004. Light driven seasonal patterns of chlorophyll and nitrate in the lower euphotic zone of the North Pacifi c Sub-tropical Gyre. Limnology and Oceanography 49:508-519.

Li, W.K. W., F.A. McLaughlin, C. Lovejoy and E.C. Carmack 2009. Smallest Algae Thrive As the Arctic Ocean Freshens. Science 326:539.

Lindeman, R.L. 1942. The trophic-dynamic aspect of ecology. Ecology 23:399-418.

Lindley, S.T., R.R. Bidigare and R.T. Barber 1995. Phytoplankton photosynthesis para-meters along 140°W in the equatorial Pacifi c. Deep Sea Research Part II: Topical Studies in Oceanography 42:441-463.

Lips, U., I. Lips, T. Liblik and N. Kuvaldina 2010. Processes responsible for the forma-tion and maintenance of sub-surface chlorophyll maxima in the Gulf of Finland. Estua-rine, Coastal and Shelf Science 88:339-349.

Longhurst, A., S. Sathyendranath, T. Platt and C. Caverhill 1995. An estimate of global primary production in the ocean from satellite radiometer data. Journal of Plankton Re-search 17:1245-1271.

Longhurst, A.R. and W. Glen Harrison 1989. The biological pump: Profi les of plankton production and consumption in the upper ocean. Progress in Oceanography 22:47-123.

Lorenzen, C.J. 1967a. Determination of chlorophyll and pheopigments: Spectrophoto-metric equations. Limnology and Oceanography 12:343-346.

Lorenzen, C.J. 1967b. Vertical distribution of chlorophyll and phaeo-pigments : Baja California. Deep Sea Research and Oceanographic Abstracts 14:735-745.

Loureiro, S., E. Garcés, Y. Collos, D. Vaqué and J. Camp 2009. Effect of marine auto-trophic dissolved organic matter (DOM) on Alexandrium catenella in semi-continuous cultures. Journal of Plankton Research 31:1363-1372.

Lund-Hansen, L.C. 2006. Development and dynamics of a coastal sub-surface phyto-plankton bloom in the southwest Kattegat, Baltic Sea. Oceanologia 48:1-14.

Lund-Hansen, L.C., M.H. Nielsen, A. Bruhn, C. Christiansen, T. Vang, P. Casado-Amezua, K. Richardson and L. Santaloria 2008. A consistent high primary production and chlorophyll-a maximum in a narrow strait – Effects of hydraulic control. Journal of Marine Systems 74:395-405.

Lund-Hansen, L.C., M. Petersson and W. Nurjaya 1999. Vertical sediment fl uxes and wave-induced sediment resuspension in a shallow-water coastal lagoon. Estuaries 22:39-46.

Lundsgaard, C., M. Olesen, M. Reigstad and K. Olli 1999. Sources of settling material: aggregation and zooplankton mediated fl uxes in the Gulf of Riga. Journal of Marine Sys-tems 23:197-210.

Markager, S. 1993. Light absorption and quantum yield for growth in fi ve species of ma-rine macroalgae. Journal of Phycology 29:54-63.

Markager, S. 1998. Dark uptake of inorganic14C in oligotrophic oceanic waters. Journal of Plankton Research 20:1813-1836.

Markager, S., J. Carstensen, D. Krause-Jensen, J. Windolf and K. Timmermann 2010. Effekter af øgede kvælstoftilførsler på miljøet i danske fjorde. National Environmental Research Institute, Aarhus University.

Markager, S., W. Vincent and E.Y. Tang 1999. Carbon fi xation by phytoplankton in high Arctic lakes: Implications of low temperature for photosynthesis. Limnology and Oceano-graphy 44:597-607.


SYNOPSIS

Markager, S. and W.F. Vincent 2000. Spectral Light Attenuation and the Absorption of UV and Blue Light in Natural Waters. Limnology and Oceanography 45:642-650.

Markager, S. and W.F. Vincent 2001. Light absorption by phytoplankton: development of a matching parameter for algal photosynthesis under different spectral regimes. Journal of Plankton Research 23:1373-1384.

McLaughlin, F.A. and E.C. Carmack 2010. Deepening of the nutricline and chloro-phyll maximum in the Canada Basin interior, 2003-2009. Geophysical Research Letters 37:L24602.

Millennium Ecosystem Assessment 2005. Ecosystems and Human Well-being: Wetlands and Water. Cambridge University Press, Cambridge, UK.

Mills, E.L. 1989. Biological Oceanography: An early history. 1870-1960. Cornell Uni-versity Press, Ithaca and London.

Moore, C.M., D.J. Suggett, A.E. Hickman, Y.-N. Kim, J.F. Tweddle, J. Sharples, R.J. Geider and P.M. Holligan 2006. Phytoplankton photoacclimation and photoadaptation in response to environmental gradients in a shelf sea. Limnology and Oceanography 51:936-949.

Morán, X.A.G. and M. Estrada 2001. Short-term variability of photosynthetic parameters and particulate and dissolved primary production in the Alboran Sea (SW Mediterra-nean). Marine Ecology Progress Series 212:53-67.

Mouritsen, L.T. and K. Richardson 2003. Vertical microscale patchiness in nano- and microplankton distributions in a stratifi ed estuary. Journal of Plankton Research 25:783-797.

Mousing, E.A., M. Ellegaard and K. Richardson 2013. Global patterns in phytoplankton community size structure – evidence for a direct temperature effect. Marine Ecology Pro-gress Series In press.

Murrell, M.C., J.G. Campbell, J.D. Hagy Iii and J.M. Caffrey 2009. Effects of irradiance on benthic and water column processes in a Gulf of Mexico estuary: Pensacola Bay, Florida, USA. Estuarine, Coastal and Shelf Science 81:501-512.

Maar, M. and J.L.S. Hansen 2011. Increasing temperatures change pelagic trophodynam-ics and the balance between pelagic and benthic secondary production in a water column model of the Kattegat. Journal of Marine Systems 85:57-70.

Maar, M., T.G. Nielsen, A. Stips and A.W. Visser 2003. Microscale Distribution of Zoo-plankton in Relation to Turbulent Diffusion. Limnology and Oceanography 48:1312-1325.

Navarro, G. and J. Ruiz 2013. Hysteresis conditions the vertical position of deep chloro-phyll maximum in the temperate ocean. Global Biogeochemical Cycles:2012GB004396.

Navarro, G., J. Ruiz, I.E. Huertas, C.M. García, F. Criado-Aldeanueva and F. Echevarría 2006. Basin-scale structures governing the position of the deep fl uorescence maximum in the Gulf of Cádiz. Deep Sea Research Part II: Topical Studies in Oceanography 53:1261-1281.

Neumann, T. 2000. Towards a 3D-ecosystem model of the Baltic Sea. Journal of Marine Systems 25:405-419.

Neumann, T., W. Fennel and C. Kremp 2002. Experimental simulations with an ecosystem model of the Baltic Sea: A nutrient load reduction experiment. Global Biogeochemical Cy-cles 16:7-1-7-19.

Nielsen, S., K. Sand-Jensen, J. Borum and O. Geertz-Hansen 2002. Depth colonization of eelgrass (Zostera marina) and macroalgae as determined by water transparency in Danish coastal waters. Estuaries 25:1025-1032.

Nielsen, T.G., B. Løkkegaard, K. Richardson, F.B. Pedersen and L. Hansen 1993. Struc-ture of plankton communities in the Dogger Bank area (North Sea) during a stratifi ed situation. Marine Ecology Progress Series 95:115-131.

Nixon, S.W. 1995. Coastal marine eutrophication: A defi nition, social causes and future concerns. Ophelia 41:199-219.


SYNOPSIS

Oku, O. and A. Kamatani 1999. Resting spore formation and biochemical composition of the marine planktonic diatom Chaetoceros pseudocurvisetus in culture: ecological signifi cance of decreased nucleotide content and activation of the xanthophyll cycle by resting spore formation. Marine Biology 135:425-436.

Paerl, H.W. 1988. Nuisance Phytoplankton Blooms in Coastal, Estuarine and Inland Wa-ters. Limnology and Oceanography 33:823-847.

Perry, M.J., B.S. Sackmann, C.C. Eriksen and C.M. Lee 2008. Seaglider observations of bloom and subsurface chlorophyll maxima off the Washington coast. Limnology and Oceanography 53:2169-2179.

Petersen, D.L.J. and M. Hjorth 2010. Marine områder 2009. NOVANA. Tilstand og udvikling i miljø- og naturkvaliteten. Faglig rapport fra DMU nr. 800, Danmarks Miljøundersøgelser, Aarhus Universitet, Roskilde.

Pitcher, G.C. 1986. Sedimentary fl ux and the formation of resting spores of selected Chaetoceros species at two sites in the southern Benguela System. South African Journal of Marine Science 4:231-244.

Platt, T., C.L. Gallegos and W.G. Harrison 1980. Photoinhibition of photosynthesis in natural assemblages of marine phytoplankton. Journa of Marine Research 38.

Popova, E.E., A. Yool, A.C. Coward, Y.K. Aksenov, S.G. Alderson, B.A. de Cuevas and A.T.R 2010. Control of primary production in the Arctic by nutrients and light: insights from a high resolution ocean general circulation model. Biogeosciences Discuss 7:5557-5620.

Pospelova, V., S. Esenkulova, S.C. Johannessen, M.C. O’Brien and R.W. Macdonald 2010. Organic-walled dinofl agellate cyst production, composition and fl ux from 1996 to 1998 in the central Strait of Georgia (BC, Canada): A sediment trap study. Marine Micro-paleontology 75:17-37.

Rabalais, N., R.E. Turner, Q. Dortch, D. Justic, V. Bierman, Jr. and W. Wiseman, Jr. 2002. Nutrient-enhanced productivity in the northern Gulf of Mexico: past, present and future. Hydrobiologia 475/476:39-63.

Reid, P.C., C. Lancelot, W.W.C. Gieskes, E. Hagmeier and G. Weichart 1990. Phyto-plankton of the North Sea and its dynamics: A review. Netherlands Journal of Sea Re-search 26:295-331.

Richardson, K. 1997. Harmful or Exceptional Phytoplankton Blooms in the Marine Ecosystem. Pages 301-385 in J. H. S. Blaxter and A. J. Southward, editors. Advances in marine biology. Academic Press.

Richardson, K., J. Beardall and J.A. Raven 1983. Adaptation of Unicellular Algae to Ir-radiance: an Analysis of Strategies. New Phytologist 93:157-191.

Richardson, K. and A. Christoffersen 1991. Seasonal distribution and production of phy-toplankton in the Southern Kattegat. Marine Ecology Progress Series 78:217-227.

Richardson, K., S. Markager, E. Buch, M.F. Lassen and A.S. Kristensen 2005. Seasonal distribution of primary production, phytoplankton biomass and size distribution in the Greenland Sea. Deep-Sea research 52:979-999.

Richardson, K., T.G. Nielsen, F.B. Pedersen, J.P. Heilmann, B. Løkkegaard and H. Kaas 1998. Spatial heterogeneity in the structure of the planctonic food web in the North Sea. Marine ecology progress series 168:197-211.

Richardson, K. and F.B. Pedersen 1998. Estimation of new production in the North Sea: consequences for temporal and spatial variability of phytoplankton. ICES Journal of Ma-rine Science: Journal du Conseil 55:574-580.

Richardson, K., B. Rasmussen, T. Bunk and L.T. Mouritsen 2003. Multiple subsurface phytoplankton blooms occurring simultaneously in the Skagerrak. Journal of Plankton Research 25:799-813.

Richardson, K., A.W. Visser and F.B. Pedersen 2000. Subsurface phytoplankton blooms fuel pelagic production in the North Sea. Journal of Plankton Research 22:1663-1671.


SYNOPSIS

Riemann, L., T.G. Nielsen, T. Kragh, K. Richardson, H. Parner, H.H. Jakobsen and P. Munk 2011. Distribution and production of plankton communities in the subtropical con-vergence zone of the Sargasso Sea. I. Phytoplankton and bacterioplankton. Marine Eco-logy Progress Series 426:57-70.

Riley, G.A., H.M. Stommel and D.F. Bumpus 1949. Quantitative ecology of the plankton of the western North Atlantic. Bulletin Bingham Oceanography 12:1-169.

Ross, O.N. and J. Sharples 2007. Phytoplankton motility and the competition for nutri-ents in the thermocline. Marine Ecology Progress Series 247:21-38.

Ryther, J.H. and W.M. Dunstan 1971. Nitrogen, Phosphorus and Eutrophication in the Coastal Marine Environment. Science 171:1008-1013.

Salter, I., A.E.S. Kemp, C.M. Moore, R.S. Lampitt, G.A. Wolff and J. Holtvoeth 2012. Diatom resting spore ecology drives enhanced carbon export from a naturally iron-fertil-ized bloom in the Southern Ocean. Global Biogeochemical Cycles 26:GB1014.

Sand-Jensen, K., S.L. Nielsen, J. Borum and O. Geertz-Hansen 1994. Phytoplankton and macrophyte development in the Danish coastal zone. Danish Environmental Protection Agency, Copenhagen, Denmark.

Sanders, J.G. and S.J. Cibik 1985. Reduction of growth rate and resting spore formation in a marine diatom exposed to low levels of cadmium. Marine Environmental Research 16:165-180.

Sarthou, G., K.R. Timmermans, S. Blain, and P. Tréguer 2005. Growth physiology and fate of diatoms in the ocean: a review. Journal of Sea Research 53:25-42.

Sathyendranath, S., V. Stuart, A. Nair, K. Oka, T. Nakane, H. Bouman, M.-H. Forget, H. Maass and T. Platt 2009. Carbon-to-chlorophyll ratio and growth rate of phytoplankton in the sea. Marine Ecology Progress Series 383:73-84.

Savchuk, O.P. 2005. Resolving the Baltic Sea into seven subbasins: N and P budgets for 199-1999. Journal of Marine Systems 56:1-15.

Sedwick, P.N., T.M. Church, A.R. Bowie, C.M. Marsay, S.J. Ussher, K.M. Achilles, P.J. Lethaby, R.J. Johnson, M.M. Sarin and D.J. McGillicuddy 2005. Iron in the Sargasso Sea (Bermuda Atlantic Time-series Study region) during summer: Eolian imprint, spatiotem-poral variability and ecological implications. Global Biogeochemical Cycles 19:GB4006.

Sharples, J., C.M. Moore, T.P. Rippeth, P.M. Holligan, J. Hydes, N.R. Fisher and J.H. Simpson 2001. Phytoplankton distribution and survival in the thermocline. Limnology and Oceanography 46:486-496.

Shulenberger, E. and J.L. Reid 1981. The Pacifi c shallow oxygen maximum, deep chlo-rophyll maximum and primary productivity, reconsidered. Deep Sea Research Part A. Oceanographic Research Papers 28:901-919.

Siegel, D.A. and A.F. Michaels 1996. Quantifi cation of non-algal light attenuation in the Sargasso Sea: Implications for biogeochemistry and remote sensing. Deep Sea Research Part II: Topical Studies in Oceanography 43:321-345.

Smayda, T.J. 1969. Some measurements of the sinking rate of fecal pellets. Limnology and Oceanography 14:621-625.

Smayda, T.J. 1971. Normal and accelerated sinking of phytoplankton in the sea. Marine Geology 11:105-122.

Smetacek, V. 2001. A watery arms race. Nature 411:745.

Smith, R.C., R.W. Eppley and K.S. Baker 1982. Correlation of primary production as measured aboard ship in Southern California Coastal waters and as estimated from satel-lite chlorophyll images. Marine Biology 66:281-288.

Smith, V. 2003. Eutrophication of freshwater and coastal marine ecosystems a global problem. Environmental Science and Pollution Research 10:126-139.

Standard, D. 1986. Vandundersøgelse 2201. Klorofyl a. Spektrofotometrisk måling i ethanolekstrakt.


SYNOPSIS

Stedmon, C.A., C.L. Osburn and T. Kragh 2010. Tracing water mass mixing in the Bal-tic-North Sea transition zone using the optical properties of coloured dissolved organic matter. Estuarine, Coastal and Shelf Science 87:156-162.

Steele, J.H. 1964. A study of production in the Gulf of Mexico Mar. Res. 3:211-222.

Steele, J.H. 1974. Patchiness. Blackwell, London.

Steele, J.H. and C.S. Yentsch 1960. The vertical distribution of chlorophyll. J. mar. biol. Ass. U.K. 39:217-226.

Steemann, N.E. 1952. The use of radio-active carbon (14C) for measuring organic pro-duction in the sea. J. Cons. int. Explor. Mer. 18:117-140.

Steffen, W., P.J. Crutzen and J.R. McNeill 2007. The Anthropocene: Are Humans Now Overwhelming the Great Forces of Nature. AMBIO: A Journal of the Human Environ-ment 36:614-621.

Steinberg, D.K., C.A. Carlson, N.R. Bates, R.J. Johnson, A.F. Michaels and A.H. Knap 2001. Overview of the US JGOFS Bermuda Atlantic Time-series Study (BATS): a de-cade-scale look at ocean biology and biogeochemistry. Deep Sea Research Part II: Topi-cal Studies in Oceanography 48:1405-1447.

Stoecker, D.K. 1999. Mixotrophy among Dinofl agellates1. Journal of Eukaryotic Micro-biology 46:397-401.

Strathmann, R.R. 1967. Estimating the Organic Carbon Content of Phytoplankton from Cell Volume or Plasma Volume. Limnology and Oceanography 12:411-418.

Strickland, J.D.H. 1968. A comparison of profi les of nutrient and chlorophyll concentra-tion taken from discrete depths and by continuous recordings. Limnology and Oceano-graphy 13:388-391.

Strickland, J.D.H. and T.R. Parsons 1972. A practical handbook of seawater analysis. Bull. Fish. Res. Bd. Can. 167:1-310.

Stroeve, J., M. Serreze, M. Holland, J. Kay, J. Malanik and A. Barrett 2012. The Arctic’s rapidly shrinking sea ice cover: a research synthesis. Climatic Change 110:1005-1027.

Strom, S.L., E.L. Macri and K.A. Fredrickson 2010. Light limitation of summer primary production in the coastal Gulf of Alaska:physiological and environmental causes. Marine Ecology Progress Series 402:45-57.

Suggett, D.J., C.M. Moore, A.E. Hickman and R.J. Geider 2009. Interpretation of fast repetition rate (FRR) fl uorescence: signatures of phytoplankton community structure ver-sus physiological state. Marine Ecology Progress Series 376:1-19.

Sunda, W.G. and D.R. Hardison 2007. Ammonium uptake and growth limitation in ma-rine phytoplankton. Limnology and Oceanography 52:2496-2506.

Sverdrup, H.U. 1953. On conditions for the vernal blooming of phytoplankton. J. Cons. Perm. Int. Exp. 18:237-295.

Takahashi, M. and T. Hori 1984. Abundance of picophytoplankton in the subsur-face chlorophyll maximum layer in subtropical and tropical waters. Marine Biology 79:177-186.

Tang, E.P.Y. 1996. Why do dinofl agellates have lower growth rates? Journal of Phyco-logy 32:80-84.

Thomson, R.E. and I.V. Fine 2003. Estimating Mixed Layer Depth from Oceanic Profi le Data. Journal of Atmospheric and Oceanic Technology 20:319-329.

Tilman, D., J. Fargione, B. Wolff, C. D’Antonio, A. Dobson, R. Howarth, D. Schindler, W.H. Schlesinger, D. Simberloff and D. Swackhamer 2001. Forecasting Agriculturally Driven Global Environmental Change. Science 292:281-284.

Timmermann, K., S. Markager and K.E. Gustafsson 2010. Streams or open sea? Tracing sources and effects of nutrient loadings in a shallow estuary with a 3D hydrodynamic–ecological model. Journal of Marine Systems 82:111-121.


SYNOPSIS

Townsend, D.W. and M. Thomas. 2002. Springtime nutrient and phytoplankton dynamics on Georges Bank. Marine Ecology Progress Series 228:57-74.

Tremblay, J.-É., D. Robert, D. Varela, C. Lovejoy, G. Darnis, R.J. Nelson and A. Sastri 2012. Current state and trends in Canadian Arctic marine ecosystems: I. Primary produc-tion. Climatic Change 115:161-178.

Tremblay, J.-É., K. Simpson, J. Martin, L. Miller, Y. Gratton, D. Barber and N.M. Price 2008. Vertical stability and the annual dynamics of nutrients and chlorophyll fl uore-scence in the coastal, southeast Beaufort Sea. Journal of Geophysical Research: Oceans 113:C07S90.

Tremblay, J.-E. and J.W. Smith 2007. Chapter 8. Primary production and nutrient dynam-ics in polynyas. Pages 239-269 in S. W. and D. Barber, editors. Polynyas Windows to the world. Elsevier Oceanography Series, Amsterdam.

UNESCO 1981. International oceanographic tables. Paris.

Utermöhl, H. 1958. Zur vervollkommung der quantitativen phytoplankton methodik. Mitt. Int. Ver. Theor. Angew. Limnol 9:1-38.

Valderrama, J.C. 1981. The simultaneous analysis of total nitrogen and total phosphorus in natural waters. Marine Chemistry 10:109-122.

Veldhuis, M.J.W. and G.W. Kraay 2004. Phytoplankton in the subtropical Atlantic Ocean: towards a better assessment of biomass and composition. Deep Sea Research Part I: Oceanographic Research Papers 51:507-530.

Walsh, J.J. 1981. A carbon budget for overfi shing off Peru. Nature 290:300-304.

Weston, K., L. Fernand, D.K. Mills, R. Delahunty and J. Brown 2005. Primary pro-duction in the deep chlorophyll maximum of the central North Sea. J. Plankton Res. 27:909-922.

Windolf, J., G. Blicher-Mathiesen, J. Carstensen and B. Kronvang 2012. Changes in nitrogen loads to estuaries following implementation of governmental action plans in Denmark: A paired catchment and estuary approach for analysing regional responses. Environmental Science & Policy 24:24-33.

Windolf, J., H. Thodsen, L. Troldborg, S.E. Larsen, J. Bøgestrand, N.B. Ovesen and B. Kronvang 2011a. A distributed modeling system for simulation of monthly runoff and ni-trogen 878 sources, loads and sinks for ungauged catchments. Journal of Environmental Monitoring 13:2645-2658.

Windolf, J., P. Wiberg-Larsen, J. Bøgestrand, S.E. Larsen, H. Thodsen, R. Bjerring, N.B. Ovesen, A. Kjeldgaard and B. Kronvang 2011b. Vandløb 2010. NOVANA., Aarhus Uni-versitet, DCE-Nationalt Center for Miljø og Energi, http://www.dmu.dk/pub/SR4.pdf

Photo: Mikkel Noe Nygaard Rasmussen.

PART II: PAPERS I-V


Submitted to Limnology and Oceanography

Maren Moltke Lyngsgaard1,2, Stiig Markager1 & Katherine Richardson1,2

1Department of Bioscience, Applied Marine Ecology and Modelling, Aarhus University, DK-4000 Roskilde, Denmark2Center for Macroecology, Evolution and Climate, Danish Natural History Museum, University of Copenhagen, DK-2200 Copenhagen, Denmark

PAPER IChanges in the vertical distribution of primary production in response to land-based N-loading

Photo: Mikkel Noe-Nygaard Rasmussen


ABSTRACT

Anthropogenic N-loading has decreased signifi cantly in the Baltic Sea Transition Zone over the past 2 decades, thus making it possible to investigate responses of primary production (PP) to changing N-conditions. This study demonstrates that the vertical distribution of PP changes as a function of land-based N-loading. A monitoring data set including 1385 water column photosynthesis estimates, where photosynthetic parameters were determined both in the surface water layer and in the pycno-cline/bottom layer (PBL) at 6 stations near the Danish coast between 1998 and 2012 is examined. Total annual PP and surface layer PP (SPP) correlate positively with land-based N-loading from Denmark (p < 0.003). The percentage of annual PP occurring in the PBL (denoted as deep primary produc-tion, DPP) varied annually between 6 % and 30 % (mean = 17 %). The absolute magnitude of the DPP as well as its relative proportion of total water column PP correlates negatively with N-loading (p < 0.009 and p < 0.0003, respectively). Thus, SPP decreases in response to decreased N-loading while DPP increases. Land-based N-loadings also correlate positively with the light attenuation coeffi cient (R2 = 0.39, p < 0.01) which may, in part, explain the response in DPP to changes in N-loading. DPP occurs in active phytoplankton communities acclimated and/or adapted to low light and producing oxygen in the pycnocline/bottom water. This vertical redistribution of PP in response to changes in N-loading has potentially important consequences for understanding and predicting recovery trajectories following anthropogenic eutrophication of stratifi ed coastal marine waters.

Keywords: Primary production, eutrophication, deep primary production, light attenuation, vertical distribution of primary production, response to nitrogen loading.

Changes in the vertical distribution of primary production in response to land-based N-loading

Maren Moltke Lyngsgaard1,2, Stiig Markager1 & Katherine Richardson1,2

1Department of Bioscience, Applied Marine Ecology and Modelling, Aarhus University, DK-4000 Roskilde, Denmark2Center for Macroecology, Evolution and Climate, Danish Natural History Museum, University of Copenhagen, DK-2200 Copenhagen, Denmark


Land-based N-loading and primary production

INTRODUCTION

Primary production (PP) is an important factor in struc-turing marine ecosystems and changes in PP in response to increased nutrient loading have been identifi ed as be-ing responsible for symptoms of eutrophication in coast-al marine systems (e.g. Nixon, 1995; Smith 2003; Clo-ern 2001). After recognising the relationship between anthropogenic nutrient loading and eutrophication, many countries have initiated programs to reduce nutri-ent enrichment (see, for example, Boesch, 2002) in the expectation that marine PP will respond to a reduction in land-based nutrient loading. In some areas, enough data have now been collected to be able to examine the extent to which this expectation has been realised.

The Baltic Sea transition zone (BSTZ), which comprises the Kattegat and the Belt Seas and, thus, forms the con-nection between the Baltic Sea and the Skagerrak/North Sea is such an area. It is a shallow, stratifi ed and temper-ate marine system with dynamic hydrography. Surface salinity varies in the region from 10-14 in the southern part and reaches up to 20-25 in the northern Kattegat (Gustafsson 2000). Bottom water salinity generally var-ies between 32 and 34 and an essentially permanent pycnocline is present. The area can be characterised as a frontal system where the low saline surface water from the Baltic Sea mixes with the more saline waters com-ing from the Skagerrak. The system demonstrates clear estuarine circulation where water transport is mainly driven by the water level difference between the Arkona Sea and the Northern Kattegat and, ultimately, by the freshwater surplus to the Baltic Sea of 559 km2 year-1 (Savchuk 2005). Mixing with surface water, especially in the Little and Great Belts (see Fig. 1), ventilates the bottom water of the BSTZ (Bendtsen et al. 2009).

The seasonal distribution of PP here is typical for tem-perate coastal waters: Elevated production occurs in association with the spring phytoplankton bloom but peak PP occurs during the summer months (Petersen and Hjorth 2010) as is also seen in other temperate es-tuarine regions such as the Chesapeake Bay (Kemp and Boynton 1984). Winter PP is low and limited by light availability.

Anoxia and hypoxia events in the BSTZ became more widespread over the 20th century (Conley et al. 2007). As these were believed to be a consequence of increases in anthropogenic nutrient loading, legislation was estab-lished in the late 1980s to control and reduce nutrient loading (Conley et al. 2002b). Following the establish-ment of this legislation, Denmark has maintained an extensive marine monitoring program. The data set re-sulting from this monitoring program provides a unique

resource for identifying relationships between PP and changing nutrient loadings which may be of potential relevance in understanding eutrophication responses in other coastal areas.

The purpose of this study was to examine these moni-toring data for evidence of a response in annual PP in this region to changing nitrogen conditions. Because it is well known that the BSTZ, as well as other seasonally or semi-permanently stratifi ed areas (Richardson et al. 2003, Lehrter et al. 2009, Strom et al. 2010), are charac-terised by a signifi cant amount of annual PP taking place in association with sub-surface phytoplankton peaks (Cullen 1982, Richardson and Christoffersen 1991, Karlson et al. 1996a), we chose not only to examine total water column PP but also the seasonal and inter-annual variability in the vertical distribution of PP. We hypothesized that PP occurring in the pycnocline and bottom layer (referred to as deep primary production or DPP) may respond differently than PP in the surface wa-ters (SPP) to changes in land-based N-loading.

MATERIALS AND METHODS

Danish National Aquatic Monitoring and Assessment ProgramData from the time period 1998–2012 were obtained from the database of the Danish National Aquatic Monitoring and Assessment Program, MADS (Conley et al. 2002b).

Figure 1. The Baltic Sea transition zone (BSTZ) and the loca-tion of the six stations



This database is publically available (http://www2.dmu.dk/1_viden/2_Miljoe-tilstand/3_vand/4_mads_ny/de-fault_en.asp) and contains physical, chemical and bio-logical data collected in BSTZ waters since the 1980s. Different municipal and public authorities were respon-sible for data collection and sampling was carried out on different ships and using different sampling equipment. However, all procedures were standardised and inter-comparisons of data collected were frequently carried out by the former Danish National Environmental Research Institute (NERI), now Aarhus University. For the param-eters used in this study, the inter-comparisons were super-vised by one of the authors (SM).

Six study stations were selected based on their having a minimum depth of > 10 m and the amount of data available. Study station depths vary from 14 m (Aal-borg Bight), to 51 m in The Sound (Fig. 1). The four sta-tions of intermediate depths are located in Aarhus Bight, the Little Belt and the Great Belt (2 stations). For each station, the following data were extracted: CTD profi le data with 0.2 m resolution for the vertical distribution of temperature, salinity, chlorophyll fl uorescence and PAR light (4π sensor). In addition, measurements of chloro-phyll a and photosynthetic parameters (see below) were extracted for the discrete depths where sampling had been carried out.

Data for total nitrogen (TN) loadings from Denmark to the BSTZ were also taken from the MADS database. The values are based on a 3D MIKE SHE (Windolf et al. 2011a) groundwater resource model validated on measured TN concentrations in Danish streams. The hy-drological area chosen for this study covered the inner Danish waters and ranged from the Kattegat Sea in the North to the Belt Seas in the South.

The daily surface photosynthetic active radiation (SPAR) was determined by averaging measurements made continuously at several different localities in Den-mark within approximately 30-150 km of the sample locations. Daily insolation was estimated using average SPAR (between the different locations) calculated for every half hour and these average values were used to calculate the 24 hour depth-integrated PP for all six sta-tions. This dataset is not publically available.

Only data from 1998-2012 were used. This period was chosen as the protocol for measuring photosynthesis pa-rameters was altered in 1998 so that parameters were measured at two depths rather than one. These two depths were: 1) an integrated “surface” sample from 0 to 10 meters (unless visual examination of the den-sity profi le indicated a major density difference or pyc-nocline within this depth range) collected using a hose/tube inserted to the desired depth or with Niskin bottles

at 1, 5 and 10 meters depth (or to the depth of the pyc-nocline) and 2) the depth of the deep chlorophyll maxi-mum, defi ned as the depth with the highest chlorophyll fl uorescence and where fl uorescence was greater than twice the average for the surface layer (sample collected using Niskin bottle). Chlorophyll a concentrations were also determined on these two samples. The protocol for the monitoring program prescribed that if no deep chlo-rophyll maximum was encountered, the second sample was taken at the depth of 2 % surface light penetration.

In addition to the two depths described above, chloro-phyll a was analysed at standard depths (every 5 m) throughout the water column starting at 1 m below the surface and ending 1 m above the sediment. Chlorophyll a determinations were made by fi ltering samples onto Whatman GF/F or GF 75 Advantec fi lters. Filters were extracted in ethanol (96 %) for 6-20 hours and sam-ples analysed spectrophotometrically according to the method described by Strickland and Parsons (1972) and modifi ed by Danish Standards (1986).

Photosynthetic carbon assimilation was estimated based on the carbon-14 method modifi ed (Markager 1998) af-ter Steemann Nielsen (1952). P-E curves were calculat-ed from incubations made under artifi cial light (Osram HQI-T or high pressure halogen lamps), where the sam-ples were incubated at seven different light intensities for two hours with metal grids providing approx. 35 % light attenuation between each bottle. Thereafter, the samples were fi ltered (GF/F or GF 75 Advantec fi lters) and the carbon incorporation stopped with acid (200 μl 0.1 N HCl). The amount of incorporated carbon-14 in the phytoplankton was determined by liquid scintilla-tion counting.

Light attenuation at each station was determined by es-timating the diffuse light attenuation coeffi cient (Kd) from the CTD-profi le using a deck sensor as reference.

Data Analyses

The division of the water column into two layers

Density (ρ) was calculated from salinity and tempera-ture profi les (UNESCO 1981). These density profi les (from year 1998 to 2012) were used to calculate the depth separating the surface from the pycnocline bottom layer (PBL). This was defi ned as being at the fi rst verti-cal density gradient of > 1 kg m-4 (see Fig. 2a for an ex-ample). For each day with a distinct pycnocline (where this density criterion was met), every depth in the water column was assigned as being in one of two layers, i.e. in the surface layer or in the pycnocline/bottom layer. This gave the possibility of considering the PP occur-ring in the surface layer (surface primary production,



SPP) and pycnocline/bottom layer (deep primary pro-duction, DPP), respectively When the density criterion of Δρ/Δz > 1 kg m-4 was absent from the water column, the total water column PP was considered as being SPP.

256 visual inspections of density profi les were made to confi rm that the density criterion was effective at deter-mining the starting depth of the PBL. The depths found by the visual inspection matched the depths found by the criterion in 90 % of the cases. The density criterion chosen (1 kg m-4) overestimated the starting depth of the PBL, i.e. identifi ed a deeper (1-2 m) depth, in the re-maining 10 %. Thus, our estimate of the fraction of DPP as a percentage of total water column PP (see below) is conservative.

Estimating chlorophyll a from fl uorescenceFluorescence per unit chlorophyll a changed systemati-cally with depth in the dataset. Therefore, a fl uorescence factor (Fchl = F/[Chl]) was calculated whereby the chlo-rophyll concentration recorded in the discrete sample (Chl) and the fl uorometer measurement from the corre-sponding depth (F) were related. Values for Fchl between sampling depths were assigned by linear interpolation. This was done with 0.2 m resolution and the resulting profi le of Fchl(z) was used to estimate a continuous chlo-

rophyll a profi le (Chl(z) = F(z)/Fchl(z)) (see Fig. 2c for an example of an estimated chlorophyll a profi le).

Estimating primary production through the water columnTo estimate water column PP, chlorophyll-specifi c pho-tosynthesis rates were calculated from the photosynthe-sis-irradiance (P-E) curves and chlorophyll a concen-trations for the two samples according to Markager et al. (1998). A light matrix representing the light intensity at 0.2 meter intervals throughout the water column and hourly intervals over the entire day was constructed us-ing the attenuation coeffi cient and the surface light. This light matrix and the photosynthetic characteristics de-rived from the P-E curves, i.e. alpha, Pmax and intercept for the two sampling depths normalised to chlorophyll a concentrations (alphaB, Pmax

B and interceptB), were com-bined to estimate daily primary production at 0.2 meter intervals throughout the water column (see Fig. 2 for an example) for all stations except the Great Belt 1 station, where only P-E curve parameters from the surface layer were available. The volume primary production (mg C m-3 day-1) was calculated as follows:

PPIrradiance

PmaxB

AlphaB

Δρ/Δzρ

Chl. a

10 12 14 16 18 20 22 24

-16

-14

-12

-10

-8

-6

-4

-2

00 0.5 1.0 1.5 2.0 2.5 0 0.1 0.2 0.3 0.4 0.51.5 1.6 1.7 1.8 1.9 2.0

0 5 10 15 20 25 0 10 20 30 40 50 60 70

a b c

PmaxB (g C g-1 Chl h-1)Density, ρ (kg m-3) Primary production, Chl. a × 10

(mg C m-3), (µg Chl l-1)

AlphaB

(g C g-1 Chl mol-1 m3)Δρ/Δz (kg m-4) Irradiance (mol m-3 day-1)

Dep

th (m

)

Figure 2. Example of how parameters were derived for the primary production estimates. Profi les from the 19th of July 2010 in Aarhus Bight of; a) density (ρ), and density difference m-1 (Δρ/Δz, kg m-4), the starting depth of the pycnocline/bottom layer (PBL) was here 6 m defi ned as the fi rst depth where Δρ/Δz > 1 kg m-4, b) interpolation of the chlorophyll-specifi c alphaB and Pmax

B values from 1 and 15 m depths indicated with black dots, c) primary production, chlorophyll a and irradiance. Irradiance is the sum over 24 hours in 20 cm depth intervals.

Volume primary production =Par

PmaxBPmaxB 1–exp × –alphaB × + Intercept



For the 5 stations with two P-E parameter sampling depths, the photosynthetic characteristics obtained from the surface layer sample were applied down to the start-ing depth of the PBL. From this depth and until the depth where the second sample for photosynthesis de-termination had been taken, the applied photosynthetic parameters were derived by linear interpolation (see Fig. 2b). The photosynthetic parameters derived from the second sampling depth were then applied from the depth of that sampling to the depth where the estimated daily depth-specifi c photosynthetic rate was greater than zero. In this manner, estimates of total water column primary production as well as the contribution made by DPP to the total were made for the 1385 days of sam-pling. Data availability varied between stations and, for each of the 6 stations, ranged from 3 to 14 years (Table 1). The sampling frequency at each station ranged from 12 to 49 measurements per year.

The average annual values of PP were calculated for all six stations. Average monthly values were calculat-ed for each station (including all available years) and these were used to fi ll months for the years where there were no measurements. This procedure allowed us to examine the temporal development in primary produc-tion without the infl uence of missing values in the data-set. We used the same approach with respect to missing values for the P-E parameters and the light attenuation coeffi cient (Kd) when analysing their inter-annual and seasonal variations.

Statistical analysesMultiple linear regressions were constructed to examine the degree to which the estimated primary production could be explained by irradiance and nitrogen loading. Annual PP, annual DPP (g C m-2 yr-1), DPP in percentage of total, annual land-based N-loading to the BSTZ from Denmark (kt of nitrogen), irradiance and the light atten-uation coeffi cient (Kd) were normalized to their mean values over the study period. By doing so, the unit of the

coeffi cients for the regressions performed becomes the percentage change in dependent variable (PP, SPP, DPP, DPP in % and Kd) per percentage change in forcing vari-ables (N-loading and surface irradiance).

The annual land-based N-loading was calculated on a monthly basis and summed over a period of consecu-tive months. When relating PP parameters to N-loading, we fi rst regressed the relationships between the param-eters for a given calendar year against total N-loading in the same calendar year. In recognition of the fact that loading might affect PP differently over the season and that there likely will be a lag between loading and PP response, we then tested different periods for load-ing against the annual PP parameters. A systematic ap-proach was used in choosing the periods to be tested. We started with 24 months covering the whole calendar year before and the year of the given annual primary production. We then excluded and included months so that all possible periods having a minimum length of two months were tested. In all, 276 different periods of N-loading were tested against annual PP, SPP and DPP. The periods giving the highest explanatory power (R2) are given in table 2 and an example of the variability in R2 and coeffi cients is shown in Fig. 7.

The irradiance values used in the statistical analyses rep-resent the actual photosynthetic active radiation (SPAR) at the surface from April to December as this period, in all cases, was that with highest predictive value. A simi-lar procedure was applied as above when the N-loading was considered in relation to Kd.

An analysis of temporal development depth separating the surface and PBL was carried out on normalised (to the average depth for the station) depths. When the nor-malised starting depth of PBL was calculated for each station, a linear regression was used on the normalised values as a function of years per station, and then for an average of the six stations.

Table 1. Number of depth-integrated primary production measurements per year at the six locations. There were 1385 primary production measurements from year 1998 to 2012.

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

Aalborg Bight 19 15 17 18 20 24 16 18 18

Aarhus Bight 26 25 25 25 28 25 29 27 25 20 19 20

Little Belt 22 21 24 24 22

Great Belt 1 20 23 23

Great Belt 2 49 49 48 48 39 48 49 45 42 23 19 19 19 17

The Sound 12 18 21 17 22 23 22 22 19 18 20 19



RESULTS

Total and vertical distribution of primary production

The mean annual PP from 1998 to 2012 for the six sta-tions was 189 ± 17.2 g C m-2 yr-1. The lowest production estimates were found in Aalborg Bight, Aarhus Bight and The Sound (155-175 g C m-2 yr-1), while the pro-duction in Great and Little Belts was higher (209-226 g C m-2 yr-1, Fig. 3). The higher values at these stations probably refl ect nutrient inputs from bottom to surface waters through more intense mixing in the narrow straits (Lund-Hansen et al. 2008). The highest water column PP was found during summer and the seasonal pattern in production largely followed irradiance and temperature (data not shown).

When all stations in the entire study period are con-sidered, the DPP, on average, contributed with 17 % of the annual PP. The contribution of the DPP to the to-tal water column production varied from 6 % in Little Belt to 30 % in Aalborg Bight (Fig. 3). The rest of the stations exhibited contributions of DPP to annual pro-duction of between 13 and 23 % (Fig. 3). The highest average monthly contribution from DPP was observed in Aalborg Bight in May, where the DPP contributed 48 % to total water column production (Fig. 3c). DPP was low or, in some cases, non-existent from October to

February, when the phytoplankton community was well mixed throughout the water column and surface irradi-ance was too low to support production in the PBL (Fig. 3). Overall, the importance of DPP was slightly higher in summer from April to August (8-40 % of summer production) than for the year as a whole.

Photosynthetic parametersThe seasonal pattern of the light intensity at which pho-tosynthesis initially is saturated, Ik, (Pmax/alpha, μmol photons m-2 s-1) is shown in Fig. 4a as the average for all six stations. Samples from the surface layer show in-creasing Ik values from January (68 μmol photons m-2 s-1) to a peak in August (128 μmol photons m-2 s-1) fol-lowed by a decline to 78 μmol photons m-2 s-1 in Decem-ber. For the deep samples, the seasonal variation was smaller. Ik increased from 62 μmol photons m-2 s-1 in May to 89 μmol photons m-2 s-1 in November and then dropped to 55 μmol photons m-2 s-1 in January. A t-test showed signifi cantly higher average monthly Ik-values in the surface layer than in the PBL (t = 3.56, df = 22, p < 0.01, n = 24). In addition to the characteristic season-al pattern seen for the Ik-values from the surface layer, a signifi cant and positive relationship was found between Ik-values from the surface layer and the daily surface irradiance (r = 0.28, p < 0.0001, n = 1080). Thus, the Ik data indicate that the phytoplankton communities in and below the surface layer are acclimated to the prevailing light levels they encounter.

PP

(g C

m-2 m

onth

-1)

Month

Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec

0

10

20

30

40

0

10

20

30

40

a. Aalborg Bight c. Aarhus Bight

b. Little Belt d. Great Belt 1

e. The Sound

f. Great Belt 2

30%

6%

23%21%

16% 13%

161

226

175155

213 209

PPDPP

Figure 3. Monthly average values for total water column primary production (PP) and deep primary production for six different stations within the Baltic Sea transition zone (1998-2012). The annual primary production and the contribution (in %) from DPP to PP are shown by values in the upper right corner of every graph.



Not surprisingly, this conclusion is also supported by the general patterns observed for maximum chlo-rophyll-specifi c photosynthesis rate (Pmax

B), as this parameter controls a large part of the variation in Ik. Average monthly values (including data from 15 years) for Pmax

B showed signifi cantly higher values for phytoplankton populations found in the surface layer (Pmax

B = 3.6 ± 1.4 g C g-1 Chl hour-1), than in the PBL (Pmax

B = 2.3 ± 0.4 g C g-1 Chl hour-1) (t = 3.06, df = 22, p < 0.01, n = 24). An exception to this general pattern was observed for the Sound, where Pmax

B from the two depths were similar.

Monthly averages of PmaxB values in the surface layer

for all six stations for the period 1998-2012 exhibited a seasonal pattern where Pmax increased from 2.4 g C g-1 Chl hour-1 in March to 6.4 g C g-1 Chl hour-1 in August and then decreased to 2.2 g C g-1 Chl hour-1 in December (Fig. 4). Pmax values from surface and PBL waters were not signifi cantly different (t-test, p < 0.05) between the months of November and March (Fig. 4). During this period, the water column was often well mixed and the solar irradiance so low that photosynthesis was likely light limited most of the time.

The general pattern in the chlorophyll-specifi c initial slope (alphaB) of the P-E curves exhibited no signifi -cant difference for phytoplankton populations found in the surface layer and those found in the PBL (Fig. 4c). In addition, the variability, expressed as the coef-fi cient of variance (CV), of alphaB was lower than that of Pmax (CV alphaB = 0.16 compared to CV for Pmax

B = 0.42). This is to be expected as alpha is equal to ab-sorption times quantum yield (see e.g. Markager and Vincent 2001) and light absorption is tightly coupled to the chlorophyll content in the cells (see e.g. Mark-ager 1993). A weak seasonal pattern in alphaB values was detected which resembled that found for Ik and Pmax

B. The alphaB values increased from 10.2 g C g-1 Chl mol photoner-1 m2 in March to the peak value in August (14.2 g C g-1 Chl mol-1 m2), and decreased thereafter to 8.2 g C g-1 Chl mol-1 m2 in December. Higher values were found in January and February than in March and April.

Primary production and land-based N-loading The concentration of nitrogen in land-runoff from Den-mark (mmol N l-1) decreased signifi cantly between 1998 and 2012 (linear regression: p < 0.0001, R2 = 0.84, n = 15, Fig. 5) primarily due to the establishment of water quality regulation (Windolf et al. 2012). Although there were large fl uctuations in the absolute loading due to inter-annual variation in runoff, the absolute delivery of nitrogen from Denmark (kt N) also showed a signifi -cant decrease over the study period (linear regression: p < 0.016, R2 = 0.37, n = 15, see Fig. 5).

Total PP, SPP and DPP, as well as the ratio between SPP and DPP to total PP were all signifi cantly correlated to land-based N-loading (Table 2). However, while SPP was positively correlated to N-loading (see Fig. 6), the opposite was true for DPP.

The relationship between SPP and N-loading can be seen in Fig. 6, where both SPP and nitrogen loading are calculated for the calendar year. The relationship is signifi cant with a p-value < 0.01 and a coeffi cient of

Depth 1Depth 2

l k (µm

ol p

hoto

ns m

-2 s

-1)

Pm

axB (g

C g

-1 C

hl. h

our-1

)A

lpha

B (g

C g

-1 C

hl. m

ol-1 m

2 )

Month

Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec

a

b

c

40

60

80

100

120

140

160

1

2

3

4

5

6

7

4

6

8

10

12

14

16

18

20

Figure 4. Seasonal variation of chlorophyll a-specifi c P-E pa-rameters; a) the light intensity at which photosynthesis is ini-tially saturated, Ik, b) maximum chlorophyll-specifi c photo-synthesis rate, Pmax

B and c) the initial slope of the P-E curve, alphaB. Parameters are shown for 1-10 m (depth 1) and for depths in the pycnocline/bottom layer (depth 2). The param-eters are average values of six stations in the Baltic Sea tran-sition zone (1998-2012) and the error bars indicate the differ-ence in average monthly values between stations.



0.47 % per % change in N-loading. Thus, when N-load-ing changed by 1 %, the SPP changed by 0.47 % in the same direction. It is, however, doubtful that N-loading calculated over the calendar year is the best descriptor in terms of relating N-loading to PP, as N-loading late in the calendar year obviously cannot affect the period of peak production during the same year. Instead, it can be argued that nitrogen has a residence time in the sys-tem. This would argue for including N-loading from the previous year when considering PP in a given year especially given the fact that phytoplankton growth is only believed to be nitrogen limited during the summer

period in the region (Conley 1999). Following this line of reasoning, the N-loading during the summer might be most important in controlling PP. We, therefore, tested a number of different loading periods extending back to January of the year prior to the test year in relation to the PP variables.

A change in N-loading was the best predictor of changes in SPP when the nine months from February to Septem-ber of the study year were included (coeffi cient of 0.60; p < 0.001, Table 2). Surface irradiance was also positive-ly related to SPP but the coeffi cient was not signifi cant.

1998 99 02 04 06 08 10 2012

Year

N-lo

adin

g no

rmal

ized

to ru

noff

(µm

ol N

l-1)

N-loading (kt N

year -1)

10

20

30

40

50

0.2

0.3

0.4

0.5

0.6

0.7

0.8

N-loadN-load/runoff

Figure 5. Land-based annual N-loading (open circles, right y-axis) and land-based annual N-loading normalized to run-off (fi lled circles, left y-axis) from 1998 to 2012. The numbers are based on average values of Danish N-loading to the study area.

Table 2. Statistics for multiple linear regressions between primary production and diffuse attenuation coeffi cient (Kd) versus land-based N-loading from Denmark and surface irradiance. Both dependent and independent variables are scaled to the mean value * 100 so the unit of the coeffi cients is % change in dependent variable per % change in independent variable, e.g. % change in PP/% change in N-loading. Coeffi cients and intercepts are given ± standard errors and values in brackets are p-values for signifi -cance. In addition, each dependent parameter is tested for change per year (annual change). Signifi cant changes (p < 0.05) are indicated in bold.

Dependent variable Period for sum of N-loading

Coeffi cient (C1) forN-loading from

Denmark

Period for sum of surface ir-

radiance

Coeffi cient (C2) for surfaceirradiance

R2 Intercept Annual change

Surface primary production February–September

0.60±0.12(0.0002)

April–December

0.98±064(0.15)

0.69 –58±69 –2.0%(0.09)

Deep primary production March*– February

–0.65±0.21 (0.0088) April–December

4.4±1.0(0.0006)

0.72 –275±98 3.6%(0.08)

Total water columnproduction

February– August

0.40±0.10(0.0025)

April–December

1.43±0.60(0.035)

0.56 –82±65 –1.0%(0.33)

Deep primary productionin % of total

March*–February

–0.72±0.14(0.0003)

April–December

3.2±(0.7)(0.0004)

0.80 –144±68 3.6%(0.026)

Diffuse attenuation coeffi cient (Kd) (1998–2012)

November*–March

0.10±0.05(0.0495) – not

signifi cant0.27 89±5 –0.2%

(0.66)

Diffuse attenuation coeffi cient (Kd) (1990–2012)

November*–March

0.09±0.03(0.0121) – not

signifi cant0.39 90±4 –0.9%

(0.017)

Sur

face

prim

ary

prod

ictio

n (%

)

Land-based N-loading (%)40 60 80 100 120 140 160

60

80

100

120

140

160

Figure 6. Changes in annual primary production in the sur-face layer as a function of changes in N-loading. Change in each parameter is expressed as percentage of the average val-ue for the study period (1998-2012) as a whole. A signifi cant linear relationship between the two parameters (p < 0.01 and R2 = 0.42) is noted on the graph with a stippled line.



Figure 7 shows the effect on the R2-value (explained variance) by varying the timing of a nine month loading period. All nine month periods with a mid-point from the 1st of November in the year prior to the study year to the 1st of July of the study year yielded signifi cant (p < 0.05) coeffi cients for the correlation between SPP and N-loading. The value of the coeffi cient varied from 0.42 to 0.60. This suggests that the N-loading from the previous winter and during the period in which phyto-plankton activity is greatest are the most important driv-ers of SPP. Calculating the N-loading over periods with different lengths gave similar results.

The DPP was strongly and positively correlated with surface irradiance (coeffi cient = 4.4 %/%, p < 0.0001) but negatively related to the N-loading (Table 2). In this case, the most signifi cant relationship was found when N-loading was calculated over a 12 month period starting in March of the year prior to the study year to February of the study year (midpoint between August and September of the year prior to the study year). The coeffi cient to N-loading was negative for all periods but only signifi cant for 12 month periods with mid-points between August and October of the year prior to the study year (Table 2) which means that changes in DPP were best predicted by a period of N-loadings occurring prior to the actual DPP.

PP was also signifi cantly and positively related to N-loading (summed from February to August of the study year, i.e. similar to the case for SPP). However, the rela-tionship was weaker than the relationship between load-ing and SPP. This can easily be explained as the total water column PP is the sum of SPP and DPP and these two layers exhibited opposite responses to N-loading.

The most signifi cant model relating PP parameters to N-loading was that describing the fraction of DPP (as a per-centage of total water column production) as a function of N-loading. This indicates that N-loading potentially redistributes PP in the water column so it shifts from the surface layer to the PBL as loadings are reduced. This result cannot be attributed to changes in the depth where the surface and PBL diverge as there was no signifi cant temporal development in changes in the starting depth of the PBL when the stations were analyzed by linear regression separately (p > 0.05) as well as when they were averaged into one value (normalized to the aver-age depth for the study period as a whole) per year (p > 0.05). A possible explanation for the identifi ed negative relationship between DPP and N-loading is the positive relationship found between N-loading and light attenu-ation in the water column. We, therefore, also tested the relationship between N-loading and the diffuse attenua-tion coeffi cient (Kd). A signifi cant positive relationship was found with a coeffi cient of 0.10 % in Kd per percent change in N-loading (Table 2). Kd-data were available back to 1990. Using this longer time series gave the same coeffi cient but with a lower p-value (Table 2).

Temporal development in biological and physical parametersThe temporal trends of the studied parameters in response to changes in land-based N-loading are masked by the large inter-annual variability in runoff. Nevertheless, we also regressed all of the parameters studied against year (Table 2, last column). While the trends expected in light of the overall reduction in nitrogen loadings were found, two more parameters showed a signifi cant temporal de-velopment. The Kd decreased over time (for the time pe-

Month

R2 -v

alue

Coefficient (%

change in PP

/%

change in N-loading)

0

0.2

0.4

0.6

0.8

0

0.2

0.1

0.4

0.3

0.6

0.5

0.7

May

*

Jun*

Jul*

Aug

*

Sep

*

Oct

*

Nov

*

Dec

*

Jan

Feb

Mar

Apr

May Ju

l

Aug

SepJun

Figure 7. An example of the relationships between periods over which the N-loading is calculated, and the resulting R2-values (●) and coeffi cients (○). The month is the mid-point of the period (a star indicates the year before the calendar year where primary production is calculated) and the periods from Table 2 are indicated in bold. The horizontal stippled line indicates when the p-val-ue is below 0.05. Surface primary production versus N-loading calculated over a nine month period.



riod from 1990–2012) and the fraction of DPP to total PP showed an increase from 1998–2012 (see Table 2). No temporal trends were found for PP, SPP or DPP (as absolute carbon values). Thus, there has, over the period 1998-2012, been a signifi cant vertical redistribution of PP in the BSTZ whereby PP has moved from the surface wa-ters to the PBL and water clarity has increased.

DISCUSSION

Primary production in relation to land-based N-loadings This study demonstrates a redistribution of PP in the wa-ter column with decreasing land-based N-loading from Denmark, whereby the proportion of PP taking place be-neath the surface layer increases with decreasing loading. This fi nding is potentially important for the management of eutrophication induced by human activities. From our general understanding of eutrophication (Cloern 2001), such a redistribution of primary production in response to anthropogenic eutrophication might have been predicted. However, to our knowledge, this is the fi rst time it has been demonstrated for a coastal ecosystem on a decadal timescale. That it has not been documented earlier with fi eld observations is likely due to the fact that there are few locations with substantial time series of data describ-ing phytoplankton photosynthetic activity over a period with concurrent reduction in nutrient loading.

The coeffi cient of the regression relating surface layer primary production (SPP) and N-loading (Table 2) in-dicates that SPP changes 0.60 % per percent change in N-loading (using load period from February to Septem-ber of the study year). The intercept for this regression of 58 % indicates that more than half of the production is based on regenerated N or N from other sources than land-based loading (see below). In the case of the DPP, a 1% increase in land-based N-loading leads to a 0.65 % decrease in DPP. An interesting consequence of the fact that opposite responses are seen in the two depth lay-ers to the reduction of N-loading is that the total wa-ter column production is not as sensitive to changes in N-loading as the two individual layers examined inde-pendently. Thus, total water column PP may not be the ideal parameter to use in assessing system responses to changes in nutrient loading.

That we see a greater shift in the distribution of primary production in the water column than in total PP is remi-niscent of the pattern observed for more shallow aquatic systems where an increase in N-loading leads to a re-distribution of primary production from benthic macro-phytes to phytoplankton, i.e. higher loading results in

the production occurring closer to the surface (Borum and Sand-Jensen 1996, Krause-Jensen et al. 2012).

As no temporal development in the starting depth of the PBL was recorded, the most likely explanation for the patterns observed in both cases must be that higher load-ing leads to a reduction in water clarity. This is support-ed by the relationship found in this study between the N-loading and Kd. That lower R2-values are found for Kd (0.27-0.37) compared to PP (0.45 - 0.83) in relation to nutrient loading agrees well with earlier studies where it has been demonstrated that the relationship between N-loading and ecosystem effects becomes weaker when the response of the parameter is not directly coupled to the loading being examined (see e.g. Carstensen et al. 2011; Timmerman et al. 2010).

The best fi ts between land-based N-loading and PP pa-rameters were found when loading was summed over other periods than the calendar year for which the PP was calculated. For SPP and PP, the period best describ-ing the relationship to N-loading (judged on the basis of the value of R2) was February to September of the study year but also the winter months prior to this period were shown to be important (Fig. 7). Thus, loading during the period when PP is high has a large impact on the PP oc-curring. In contrast, both the DPP and Kd-values were better described by N-loading occurring farther back in time (Table 2 and Fig. 6b). Assuming that the negative coeffi cient for N-loading versus DPP derives, ultimate-ly, from the positive effect on light attenuation, both ob-servations would suggest the occurrence of a lag-period from when the loadings reach the marine system and until the effects are observed.

A likely explanation for this is that light attenuation is governed by accumulated organic matter (CDOM or de-tritus) in the system (Siegel and Michaels 1996) which, in turn, is derived partly from PP over the previous pe-riod . It has earlier been shown for this region that less than 30 % (often much less) of the light attenuation is caused directly by chlorophyll (Markager et al. 2010, Timmermann et al. 2010, Krause-Jensen et al. 2012), whereas most is caused by dissolved organic matter, particulate matter, and water itself. Thus, the inter-an-nual relationship between N-loading and Kd might be further strengthened by the dissolved organic and par-ticulate matter that comes with runoff water from land.

Primary production in the pycnocline/bottom layerThe estimates of PP made with the assumptions used here suggest that a signifi cant proportion of the PP in the study area occurs in the PBL. The average contribution of this DPP to the annual PP for the 6 stations ranged



from 6-30 %. This estimate agrees well with previous fi ndings in the region (Richardson and Christoffersen 1991, Richardson et al. 2000, Richardson et al. 2003). Furthermore, the water column PP estimates indicate that photosynthesis in the PBL is a consistent feature during the period April to October.

The chlorophyll-specifi c P-E parameters give support to the conclusion derived from the PP estimates that a photosynthetically active phytoplankton population is found in the PBL. These parameters also suggest that the phytoplankton population in the PBL is physiologi-cally distinct from populations in the surface layer. Both Ik and Pmax

B are lower in samples taken in the PBL than for samples in the surface layer. This suggests that light is a limiting factor in the PBL. The obvious adaptation to low light would be higher alpha-values. However, chlorophyll specifi c alpha-values cannot be used to de-tect this as alpha and chlorophyll content per cell will co-vary. Chlorophyll a concentration within a single phytoplankton cell increases during the process of accli-matization to lower light intensities (e.g. Richardson et al. 1983, Cullen 1982) . Therefore, the lack of difference in alphaB between depths may be due to higher chloro-phyll contents per cell for phytoplankton from depth 2. Unfortunately, cell counts were not made for the deep samples so we could not calculate alpha per cell.

The data do not allow us to differentiate between ad-aptation occurring at the population level (i.e. through changing species composition) and/or as acclimatiza-tion at the cellular level. Given, however, the very dif-ferent light and nutrient conditions found in the surface layer and PBL, respectively, it seems likely that differ-ent species may have populated the two environments. Earlier studies have documented signifi cant vertical heterogeneity in the distribution of phytoplankton spe-cies in the southern Kattegat (Mouritsen and Richardson 2003, Manuscript II). Further support for photosyntheti-cally active populations being found in the PBL is found in the seasonal changes (highest values being found dur-ing summer months when light intensities are greatest) noted for the P-E parameters derived from populations from the PBL.

Other nitrogen sources The fact that we identify such a clear relationship be-tween the land-based N-loading originating from Den-mark and the vertical distribution of PP occurring at 6 stations in the BSTZ may, initially, seem surprising as the BSTZ receives nutrient loading from several sources (advection from surrounding seas, atmospheric deposi-tion and land-based runoff from several countries). Ni-trogen budgets for the area (Jørgensen et al. 2013) show that land-based N-loading only constitute 14 % of the

total nitrogen input and 21% of the bioavailable N in-put and only about half of the land-based loadings come from Denmark. The explanation may possibly be found in the fact that all of the stations in this study are lo-cated close to the Danish coast. As a result, land-based loading from Germany and Sweden largely is processed prior to reaching the sample locations. The German and Swedish loadings were sought included in the regres-sion analyses but gave poorer fi ts (data not shown). Ni-trogen entering the BSTZ from adjacent seas constitutes 58 % of the bioavailable N (Jørgensen et al. 2013) in-troduced to the region. However, these inputs probably constitute a rather constant background that leaves the changes in land-based loading as a likely potential can-didate in generating inter-annual variability in PP. Fur-thermore, some of those inputs may not be reaching the productive surface layer during the productive season (see arguments in Jørgensen et al. (2013).

The strong coupling between local loading and PP means that local load reductions also can be effective in improving the marine environment in the area. Previ-ous studies (as discussed in Duarte 2009) suggest that a return to the previous condition after load reductions is not to be expected based on data for phytoplankton biomass. However, the results in Table 2 do indicate a linear response between PP and N-loading. The relation-ship found in this study between PP and N-loading sug-gests that eutrophication may, indeed, be reversible and that it can be abated by reduction in local N-loading, as phytoplankton PP is the key process starting the cascade of effects in marine nutrient eutrophication.

Oxygen production in the PBLThe vertical redistribution of PP in relation to land-based N-loading demonstrated here for the BSTZ has potential-ly important implications for the production of oxygen in the PBL. Converting the average DPP from this study (17 % of annual production) to oxygen equivalents (O2: CO2 = 1) yields an oxygen production in the PBL of about 89 g O2 m-2 year-1. Jørgensen and Richardson (1996) ex-amined data from the Southern Kattegat and estimated the oxygen demand in the bottom layer to be 202 g O2 m-2 yr-1 Hansen and Bendtsen (2013) argue that 28 % of PP (~54 g C m-2 yr-1 based on data from this study) is remin-eralized in the bottom waters. This would be equivalent to an approximate oxygen demand of 144 g O2 m-2 yr-1. Thus, the present study suggests that oxygen produced by DPP in the BSTZ may compensate for a considerable fraction of the oxygen demand in this layer.

Subsurface PP has also been shown to counteract, al-though not eliminate, hypoxia in other areas where light penetrates into the pycnocline/bottom layer (Lehrter et al. 2009, Murrell et al. 2009, Strom et al. 2010). Stud-



ies on the importance of oxygen production below the surface layer for the Louisiana continental shelf showed that up to 41 % of the bottom water oxygen demand could be resupplied by DPP and, when sediment pro-duction is included, this oxygen contribution would be even higher (Dortch et al. 1994, Lehrter et al. 2009).

Implications for understanding ecosystem effects of anthropogenic nutrient loadingNutrient enrichment is well known to infl uence both the magnitude of the production of organic material in a sys-tem (e.g. Nixon 1995) and light attenuation in the water column (Nielsen et al. 2002). Until now, however, most studies of eutrophication in marine systems have focused on the infl uence of nutrient enrichment on the total water column PP. The current study suggests, however, that for the BSTZ, changes in land-based N-loading infl uence the vertical distribution of PP more than the total magnitude of PP. We see no reason to believe that this should not be the case for other stratifi ed coastal systems as well.

This vertical redistribution of PP is potentially impor-tant for the function of the system as a whole, as pho-tosynthesis occurring in the P BL introduces oxygen into bottom waters and, thus, ameliorates hypoxia. In addition, a change in the vertical distribution of PP as reported here will likely have consequences for the rela-tive occurrence of different phytoplankton species. This, potentially, can have consequences both for food webs and for the amount of organic material reaching bottom waters. How this vertical redistribution may affect eco-system function remains, however, to be quantifi ed.

ACKNOWLEDGEMENTS

This study received funding from grant number 2104-09-063212 and 2104-09-67259 from the Strategic Re-search Council of Denmark. Additional support was received from the Danish National Reseach Foundation via a grant to the Center for Macroecology Evolution and Climate, University of Copenhagen and from the Department of Bioscience, Aarhus University. We thank the Department of Bioscience, Aarhus University for access to the monitoring data and Morten Holtegaard Nielsen for producing the map of the research area.

RERERENCES

Bendtsen, J., K.E. Gustafsson, J. Söderkvist and J.L.S. Hans-en 2009. Ventilation of bottom water in the North Sea-Baltic Sea transition zone. Journal of Marine Systems 75:138-149.

Boesch, D. 2002. Challenges and opportunities for science in reducing nutrient over-enrichment of coastal ecosystems. Estuaries 25:886-900.

Borum, J. and K. Sand-Jensen 1996. Is Total Primary Pro-duction in Shallow Coastal Marine Waters Stimulated by Nitrogen Loading? Oikos 76:406-410.

Carstensen, J., M. Sánchez-Camacho, C.M. Duarte, D. Krause-Jensen and N. Marbà 2011. Connecting the dots: Re-sponses of coastal ecosystems to changing nutrient concentra-tions. Environmental Science & Technology 45:9122-9132.

Cloern, J.E. 2001. Our evolving conceptual model of the coastal eutrophication problem. Marine Ecology Progress Series 210:223-253.

Conley, D. 1999. Biogeochemical nutrient cycles and nutri-ent management strategies. Hydrobiologia 410:87-96.

Conley, D.J., J. Carstensen, G. Ærtebjerg, P.B. Christensen, T. Dalsgaard, J.L.S. Hansen and A.B. Josefson 2007. Long-term canges and impacts of hypoxia in Danish coastal wa-ters. Ecological Applications 17:S165-S184.

Conley, D.J., S. Markager, J.H. Andersen, T. Ellerman and L.M. Svendsen 2002. Coastal eutrophication and the Danish national Aquatic Monitoring and Assessment Program. Estu-aries 25:706-719.

Cullen, J.J. 1982. The Deep Chlorophyll Maximum: Com-paring vertical Profi les of Chlorophyll a. Canadian Journal of Fisheries and Aquatic Sciences 39:791-803.

Dortch, Q., N. Rabalais, R. Turner and G. Rowe 1994. Respi-ration rates and hypoxia on the Louisiana shelf. Estuaries and Coasts 17:862-872.

Duarte, C. M. 2009. Coastal eutrophication research: a new awareness. Pages 263-269 in J.H. Andersen and D.J. Con-ley, editors. Eutrophication in Coastal Ecosystems. Springer Netherlands.

Gustafsson, B.G. 2000. Time-Dependent Modeling of the Baltic Entrance Area. 2. Water and Salt Exchange of the Bal-tic Sea. Estuaries 23:253-266.

Hansen, J.L.S. and J. Bendtsen 2013. Parameterisation of oxygen dynamics in the bottom water of the Baltic Sea – North Sea transition zone. Marine Ecology Progress Series 481:25-39.

Jørgensen, B.B. and K. Richardson 1996. Eutrophication in coastal Marine Ecosystems. Coastal and estuarine studies 52:273.

Jørgensen, L., S. Markager and M. Maar 2013. On the im-portance of quantifying bioavailable nitrogen instead of total nitrogen. Biogeochemistry.



Karlson, B., L. Edler, W. Granéli, E. Sahlsten and M. Kuylen-stierna 1996. Subsurface chlorophyll maxima in the skagerrak-processes and plankton community structure. Journal of Sea Research 35:139-158.

Kemp, M.W. and W.R. Boynton 1984. Spatial and Temporal Coupling of Nutrient Inputs to Estuarine Primary Production: The Role of Particulate Transport and Decomposition. Bul-letin of Marine Science 35:522-535.

Krause-Jensen, D., S. Markager and T. Dalsgaard 2012. Ben-thic and Pelagic Primary Production in Different Nutrient Regimes. Estuaries and Coasts 35:527-545.

Lehrter, J.C., M.C. Murrell and J.C. Kurtz 2009. Interactions between freshwater input, light, and phytoplankton dynamics on the Louisiana continental shelf. Continental Shelf Re-search 29:1861-1872.

Lund-Hansen, L.C., M.H. Nielsen, A. Bruhn, C. Christian-sen, T. Vang, P. Casado-Amezua, K. Richardson and L. San-taloria 2008. A consistent high primary production and chlo-rophyll-a maximum in a narrow strait – Effects of hydraulic control. Journal of Marine Systems 74:395-405.

Markager, S. 1993. Light absorption and quantum yield for growth in fi ve species of marine macroalgae. Journal of Phy-cology 29:54-63.

Markager, S. 1998. Dark uptake of inorganic14C in oligotro-phic oceanic waters. Journal of Plankton Research 20:1813-1836.

Markager, S., J. Carstensen, D. Krause-Jensen, J. Windolf and K. Timmermann 2010. Effekter af øgede kvælstoftil-førsler på miljøet i danske fjorde., National Environmental Research Institute, Aarhus University.

Markager, S. and W.F. Vincent 2001. Light absorption by phytoplankton: development of a matching parameter for algal photosynthesis under different spectral regimes. Journal of Plankton Research 23:1373-1384.


Murrell, M.C., J.G. Campbell, J.D. Hagy Iii and J.M. Caffrey 2009. Effects of irradiance on benthic and water column pro-cesses in a Gulf of Mexico estuary: Pensacola Bay, Florida, USA. Estuarine, Coastal and Shelf Science 81:501-512.

Nielsen, S., K. Sand-Jensen, J. Borum and O. Geertz-Hansen 2002. Depth colonization of eelgrass (Zostera marina) and macroalgae as determined by water transparency in Danish coastal waters. Estuaries 25:1025-1032.

Nixon, S.W. 1995. Coastal marine eutrophication: A defi ni-tion, social causes, and future concerns. Ophelia 41:199-219.

Petersen, D.L.J. and M. Hjorth 2010. Marine områder 2009. NOVANA. Tilstand og udvikling i miljø- og naturkvaliteten. Faglig rapport fra DMU nr. 800, Danmarks Miljøundersøgel-ser, Aarhus Universitet, Roskilde.

Richardson, K., J. Beardall and J.A. Raven 1983. Adaptation of Unicellular Algae to Irradiance: an Analysis of Strategies. New Phytologist 93:157-191.

Richardson, K. and A. Christoffersen 1991. Seasonal distri-bution and production of phytoplankton in the Southern Kat-tegat. Marine Ecology Progress Series 78:217-227.

Richardson, K., B. Rasmussen, T. Bunk and L.T. Mouritsen 2003. Multiple subsurface phytoplankton blooms occurring simultaneously in the Skagerrak. Journal of Plankton Re-search 25:799-813.

Richardson, K., A.W. Visser and F.B. Pedersen 2000. Sub-surface phytoplankton blooms fuel pelagic production in the North Sea. Journal of Plankton Research 22:1663-1671.

Savchuk, O.P. 2005. Resolving the Baltic Sea into seven sub-basins: N and P budgets for 1991–1999. Journal of Marine Systems 56:1-15.

Siegel, D.A. and A.F. Michaels 1996. Quantifi cation of non-algal light attenuation in the Sargasso Sea: Implications for biogeochemistry and remote sensing. Deep Sea Research Part II: Topical Studies in Oceanography 43:321-345.

Smith, V. 2003. Eutrophication of freshwater and coastal marine ecosystems a global problem. Environmental Science and Pollution Research 10:126-139.


Steemann, N.E. 1952. The use of radio-active carbon (14C) for measuring organic production in the sea. J. Cons. int. Ex-plor. Mer. 18:117-140.


Strom, S.L., E.L. Macri and K.A. Fredrickson 2010. Light limitation of summer primary production in the coastal Gulf of Alaska:physiological and environmental causes. Marine Eco-logy Progress Series 402:45-57.

Timmermann, K., S. Markager and K.E. Gustafsson 2010. Streams or open sea? Tracing sources and effects of nutrient loadings in a shallow estuary with a 3D hydrodynamic–eco-logical model. Journal of Marine Systems 82:111-121.

UNESCO. 1981. International oceanographic tables, Paris.

Windolf, J., G. Blicher-Mathiesen, J. Carstensen and B. Kro-nvang 2012. Changes in nitrogen loads to estuaries following implementation of governmental action plans in Denmark: A paired catchment and estuary approach for analysing regional responses. Environmental Science & Policy 24:24-33.

Windolf, J., H. Thodsen, L. Troldborg, S.E. Larsen, J. Bøgestrand, N.B. Ovesen and B. Kronvang 2011. A distrib-uted modeling system for simulation of monthly runoff and nitrogen 878 sources, loads and sinks for ungauged catch-ments. Journal of Environmental Monitoring 13:2645-2658.


In preparation for Marine Ecology Progress Series

Maren Moltke Lyngsgaard1,2, Katherine Richardson1, Stiig Markager2, Morten Holtegaard Nielsen4, Michael Olesen3 & Jesper Philip Aagaard Christensen2

1Center for Macroecology, Evolution and Climate, Danish Natural History Museum, University of Copenhagen, DK-2200 Copenhagen, Denmark 2Department of Bioscience, Marine diversity and experimental ecology, Aarhus University, DK-4000 Roskilde, Denmark3Department of Marine Biology, Copenhagen University, DK-3000 Helsingør, Denmark4Arctic Technology Centre, Department of Civil Engineering, Technical University of Denmark, Kemitorvet, Building 204, DK-2800 Kgs. Lyngby, Denmark

PAPER IIDeep primary production in coastal pelagic systems:Importance for ecosystem functioning

Photo: Maren Moltke Lyngsgaard


ABSTRACT

Aarhus Bight is located within the Baltic Sea transition zone, a more or less continually stratifi ed coastal region separating the Danish, Swedish and German coasts. Using monitoring data collected from1999 – 2012 and data from a two week fi eld campaign at a station in the Bight, this study dem-onstrates that primary production occurring below the surface layer, i.e. in the pycnocline/bottom layer (PBL) contributes signifi cantly to total primary production. The magnitude of this deep primary production (DPP) correlates signifi cantly with oxygen concentrations in the PBL. O2 concentrations in this layer also correlate signifi cantly with overlying water salinity and water transparency. At the time of the fi eld campaign, the phytoplankton community responsible for DPP was very different than that in the surface layer. In the deeper layer, large cells, especially Ceratium spp., dominated while in surface waters, small diatoms, mainly Proboscis alata, dominated. Photosynthetic parameters showed signs of adaptation/acclimatisation to their respective light environments. Sinking rates (based on sediment trap collections) of carbon and nitrogen were highest in the PBL. Thus, this study demon-strates that the vertical distribution of photosynthesis infl uences both the oxygen conditions in the bottom waters and the fate of the organic material produced. Lyngsgaard et al (submitted) have related nitrogen enrichment in this region to changes in the vertical distribution of primary production. Thus, previously unrecognised effects of eutrophication may include changes in the structure of planktonic food webs and element cycling in the water column, both brought about through an altered vertical distribution of primary production.

Keywords: Primary production, light attenuation, vertical distribution, phytoplankton species distri-bution, sedimentation, oxygen concentration

Deep primary production in coastal pelagic systems:Importance for ecosystem functioning

Maren Moltke Lyngsgaard1,2 ([email protected]), Katherine Richardson1, Stiig Markager2, Morten Holtegaard Nielsen4, Michael Olesen3 & Jesper Philip Aagaard Christensen2

1Center for Macroecology, Evolution and Climate. Danish Natural History Museum, University of Copenhagen, DK-2200 Copen-hagen, Denmark 2Department of Bioscience, Marine diversity and experimental ecology, Aarhus University, DK-4000 Roskilde, Denmark3Department of Marine Biology, Copenhagen University, DK-3000 Helsingør, Denmark4Arctic Technology Centre, Department of Civil Engineering, Technical University of Denmark, Kemitorvet, Building 204, DK-2800 Kgs. Lyngby, Denmark


Deep primary production – Importance for ecosystem functioning

INTRODUCTION

Phytoplankton requires light to carry out photosynthesis and, as light is attenuated exponentially with depth in the ocean, we might expect the highest concentrations of phytoplankton in the immediate surface waters. While, indeed, phytoplankton are most abundant in the photic zone, it is far from always the case that the highest con-centrations are found closest to the surface. In stratifi ed water columns, the highest concentrations of phyto-plankton are often found near the interface between the well-lit surface waters and deeper more nutrient replete water layers. These deep chlorophyll maxima (DCM) have been demonstrated to be a consistent feature in per-manently stratifi ed regions, i.e. tropical and sub-tropical seas (see e.g. Sedwick et al. 2005; Benitez-Nelson et al. 2007) and a seasonal feature in many temporal and polar regions (Estrada et al. 1993, Holm-Hansen and Hewes 2004). In regions with very high insolation rates, phy-toplankton may experience photoinhibition and, there-fore, avoid immediate surface waters (see e.g. Platt et al. 1980; Falkowski 1984) but, in general, the DCM is believed to occur where there is an optimal combina-tion of light and nutrients for phytoplankton growth and production (Klausmeier and Litchman 2001, Beckmann and Hense 2007, Ross and Sharples 2007).

Initially, it was assumed that DCMs represented accu-mulations of sinking cells (Riley et al. 1949). However, in recent decades, it has been demonstrated that DCMs are usually comprised of photosynthetically active phy-toplankton and that they can contribute substantially to overall water column productivity with e.g. up to 75 % on a daily basis in the Dogger Bank area (Richardson et al., 1998) and up to 30 % on an annual basis in the Southern Kattegat (e.g. Richardson and Christoffersen, 1991, Lyngsgaard et al. submitted). Signifi cant subsur-face PP has also been shown to be a feature of stratifi ed areas such as the English Channel (see e.g. Holligan et al. 1984; Sharples et al. 2001), the Celtic Sea (see e.g. Hickman et al. 2012), the North Sea (see e.g. Reid et al. 1990; Richardson and Pedersen 1998; Richardson et al. 2000; Weston et al. 2005), Greenland Sea (Richard-son et al. 2005), the Baltic sea transition zone (see e.g. Lund-Hansen et al. 2006; Lyngsgaard et al. submitted), the Baltic Sea (see e.g. Kononen et al. 2003; Lips et al. 2010), and the subtropical Atlantic Ocean (see e.g. Veld-huis and Kraay 2004).

Phytoplankton populations associated with DCMs of-ten exhibit a different physiological response to light than their counterparts in surface waters suggesting that they represent distinct communities there, and that they have adapted/acclimated to lower light conditions than those in surface waters (see e.g. Richardson et al. 1983;

Mouritsen and Richardson 2003; Moore et al. 2006). In principle, this can be acclimatisation, i.e. similar species in both surface waters and DCM but with altered com-position of the photosynthetic units, chlorophyll content and other physiological responses and/or adaptation i.e. selection for different species present in the two layers. Indeed, a number of studies have demonstrated that phy-toplankton species composition can vary signifi cantly between surface waters and the DCM (e.g. Richardson et al 2005; Hickman et al. 2009) as well as vertically within the DCM and surface waters, themselves (Mour-itsen and Richardson 2003).

It has also been shown that photosynthesis occurring in association with DCMs can ameliorate hypoxia in some stratifi ed water systems (Lehrter et al. 2009, Murrell et al. 2009, Strom et al. 2010). Despite these observations, most models describing marine ecosystems do not ac-commodate for potential heterogeneity on the vertical distribution of primary production including their vary-ing light growth relationships. Such models tend to ei-ther ignore the production of organic material and oxy-gen below the pycnocline (Bendtsen and Hansen 2013) or assume that PP is linearly related to light (i.e. with light saturation) and applying only one factor for light affi nity for all phytoplankton cells (Neumann 2000, Maar and Hansen 2011).

Lyngsgaard et al. (submitted) found that deep primary production (DPP), on average, contributes 17% of the PP at six coastal stations in the Baltic Sea Transition Zone (BSTZ) and concluded that the amount of DPP oc-curring annually was a function of land-based nitrogen loading. This observation suggests that, in addition to increasing the amount of organic material produced in a system as shown by e.g. Nixon (1995), anthropogenic nutrient enrichment potentially changes the vertical dis-tribution of PP which, in turn, may change ecosystem function. This calls for a better understanding of how the vertical distribution of photosynthesis infl uences ecosystem processes. The purpose, therefore, of this study was to explicitly describe characteristics of phyto-plankton populations in the surface layer and in the pyc-nocline/bottom layer (PBL) in an effort to compare their potential to contribute to conditions in the water column and the functioning of the planktonic ecosystem. We hy-pothesised that a) the vertical distribution of PP affects the bottom water oxygen conditions in the BSTZ and b) that differences in taxonomic composition between communities in the surface layer and in the PBL may lead to differences in size structure and feeding strate-gies of the phytoplankton community that can infl uence food web effi ciency and sinking patterns.

We test these hypotheses using monitoring data collect-ed between 1999 and 2012 and fi eld studies conducted



in 2010 in the region of the Southern Kattegat known as Aarhus Bight. The Kattegat and Belt Seas form the BSTZ and are consistently stratifi ed during summer months. Hypoxia is a frequent and reoccurring event in this area (Conley et al. 2002b).

MATERIALS AND METHODS

This study is based on data retrieved from the Danish National Aquatic Monitoring and Assessment Program; MAPS (Conley et al. 2002b) and from a fi eld study car-ried out in July 2010 on the Aarhus University research ship, Tyra. The data from MAPS is publically available in MADS, ODA and ODAM databases (see following link for the MADS database: http://www2.dmu.dk/1_viden/2_Miljoe-tilstand/3_vand/4_mads_ny/default_en.asp) and the data analysed in this study were record-ed at 56º09´10´´N, 10º19´20´´E (Fig. 1.). The analyses carried out during the fi eld campaigns were conducted following the same procedures as used in MAPS for all measurements that also were reported in MAPS. All procedures in the monitoring program were described in common technical guidelines (Kaas and Markager 1998) and quality control was ensured by frequent inter-comparisons of data collected by different contributors.

Study areaThe BSTZ can be characterised as a large frontal sys-tem where the low-saline surface waters from the Baltic Sea are mixed with the more saline waters coming from the Skagerrak. In Aarhus Bight, the average surface (0-5 m) salinities ranging over the year from about 18.8-24.4 and average bottom layer (> 12 m) salinities ranging from 25.7-29 (values from 1999 to 2012 extracted from the MADS). The average temperatures vary between 2.9–17.9 °C in the surface and between 3.6-14.2 °C in the bottom water.

CTD profi lesProfi les of salinity, temperature, chlorophyll fl uores-cence, photosynthetic active radiation (PAR) and oxy-gen concentration in Aarhus Bight were extracted from the MADS database for the years 1999-2012. There were between 1 and 10 (average = 4) CTD profi les made in each month. These same parameters were also measured during the fi eld campaigns with a Seabird 911. Both the ship and the instruments used in the fi eld campign have previously contributed to the MAPS-pro-gram. Oxygen concentration and fl uorescence (Wetlab fl uorometer, model no. FLNTURT) were profi led. The profi ling oxygen sensor (SBE 43 with a Clark electrode) was calibrated with measurements made on discrete wa-ter samples using a dissolved oxygen analyser (SiS DO Analyser, ser. No. 8045). The CTD-mounted fl uorom-eter was calibrated with spectrophotometric determina-tion of chlorophyll a from discrete depths (see below).

Division of the water column into 2 layersSalinity and temperature profi les were used to calculate density (ρ) (UNESCO 1981). 613 density profi les made between 1999 and 2012 were used to calculate the start-ing depth of the pycnocline defi ned with a density criteri-on value of 1 kg m-3 per m (=1 kg m-4). Thus, a pycnocline was defi ned as being present when this density criterion was met at a depth > 3 m and > 1 m from the bottom. This depth interval was chosen in order to exclude short-lived stratifi cation that occasionally develops in the immediate surface waters, i.e. when insolation is high. Waters close to the bottom were excluded as there can be CTD noise in very-near bottom waters. The average monthly frequency of there being a pycnocline was calculated as a percent-age of the number of days where sampling was carried out that a pycnocline was recorded.

For every day with a distinct pycnocline, each depth in the water column was assigned as being in one of two layers, i.e. either in the pycnocline/bottom layer (PBL) or in the surface layer. This gave the possibility of considering the PP occurring in each layer; surface

Figure 1. The Baltic Sea – Skagerrak transition zone consisting of the Kattegat and the Belt Sea (Danish Straits). The location of the sampling station in Aarhus Bight is shown with a black dot. An orien-tation map of the Baltic Sea area is included in the top right corner.



PP (SPP) and deep primary production (DPP), i.e. the PP occurring in the PBL. When the density criterion of Δρ/Δz > 1 kg m-4 was absent from the water column, all PP was considered as being SPP.

It is common practice to use density gradient criteria to defi ne the mixed layer depth (e.g. Kara et al. 2000 or Montégut et al. 2004). The value chosen here was not intended to identify the depth of the surface mixed layer, as is common in many studies, but used rather to delin-eate the pycnocline/bottom waters from the remaining water column. It is, therefore, relatively high compared to the density criteria often used to separate water lay-ers. 256 visual inspections of density profi les were made to examine how well the depth defi ned by the density criterion matched the visually determined division of the water column. These indicated a good match in 90 % of the cases. In the remaining 10 % of the cases, the den-sity criterion (Δρ/Δz > 1 kg m-4) overestimated the depth of the PBL, i.e. placed the starting depth approx. 1-2 m deeper than the visual examination suggested. Thus, our estimate of the thickness of the PBL is, if anything conservative.

LightPhotosynthetically active radiation (PAR) was meas-ured with a spherical Licor UW sensor mounted on the CTD rosette during both fi eld campaigns and the survey programme sampling. Surface PAR (SPAR) was meas-ured at the top of the ship at the same time as the CTD casts and the data stored included the fraction of SPAR in the water column. Values for the diffuse light attenu-ation coeffi cient (Kd, m-1) were estimated for each pro-fi le. PP from both the fi eld campaigns and MAPS were calculated with a second set of SPAR data. This dataset was composed of measured values from different land based localities in Denmark within approximately 30-150 km of Aarhus Bight throughout the study period and averaged for every 30 minutes. This SPAR dataset was used to minimize variation in primary production that was caused by SPAR measurements carried out by dif-ferent sensors (in different resolutions).

Water movementObservations of fl ow velocities were made using a 600 kHz WorkHorse Monitor ADCP (Acoustic Dop-pler Current Profi ler) from Teledyne RDI, which was mounted looking downward over the side of the vessel. The cell size depth of the instrument varied between 25 or 50 cm, whereas other primary settings, such as the blanking depth (100 cm), the transducer depth (50 cm) and the ambiguity velocity (3.30 m s-1) were kept con-stant. The data were collected using single pings, and the instrument was set to ping as fast as possible result-

ing in about one record per second. Although anchored, the vessel used for the fi eld surveys was in constant mo-tion due to wind, waves and currents. To account for these movements, the velocity of the vessel (found us-ing the bottom track of the instrument) was subtracted. The fl ow speed of water movement was calculated as a vertical average of the velocity observations observed in the three depth intervals; 3-5 m, 8-10 m and 13-15 m.

Chl. a concentrationDetermination of chl. a concentration in water samples and sediment trap material was carried out by fi ltering through a Whatman GF/F or GF 75 Advantec fi lters and extracting the chl. a in ethanol (96 %) for 6-20 hours. The extract was centrifuged and absorption at 665 nm was de-termined spectrophotometrically before and after acidifi -cation (1 N HCl) according to the method described by Strickland and Parsons (1972) and modifi ed by Danish Standards (1986). Samples from the fi eld campaign were analysed fl uorometrically (TD-700 Turner fl uorometer) according to Lorenzen (1967a) when the concentration was too low for spectrophotometric analysis (sediment traps and size-fractionated samples).

In the fi eld study, chlorophyll a was fractioned by size using glass fi bre fi lters with three different pore sizes: 56 μm (NY56.047), 10 μm (NY10.047) and 0.7 μm (Whatman GF/F or G5 75 Advantec). Fresh sample wa-ter was used on each fi lter size so that no water sample was fi ltered more than once. Thus, we obtained values for chlorophyll a originating in plankton retained on fi l-ters of three different pore sizes. The concentration of chlorophyll a from phytoplankton larger than 0.7 μm but smaller than 10 μm was obtained by subtracting the concentration retained on the 10 μm fi lter from the concentration found on the 0.7 μm fi lter. This procedure was repeated to obtain an estimate for the size fraction 10-56 μm.

A continuous chlorophyll profi le (depth resolution 0.2 m) was constructed from the GF 75 measurements of chlorophyll and profi le data for fl uorescence. Sam-pling depths were 1, 5, 10 and 15 meters for the sur-vey programme and similarly 3-4 depths for the fi eld campaigns (1 m, 14 m and 1 or 2 in between where one was taken in the DCM). Fluorescence per unit chloro-phyll a changed systematically with depth. Therefore, a fl uorescence factor (Fchl = F/[Chl]) was calculated for each depth with a chlorophyll concentration recorded in the discrete samples (Chl). Values for Fchl between sampling depths were assigned by linear interpolation and used to estimate a continuous chlorophyll a profi le (Chl(z) = F(z)/Fchl(z)). In this study, the use of the term deep chlorophyll max or DCM refers to any increase in chlorophyll in the PBL.



A total number of 214 profi les were examined to locate where the DCMs were located in relation to the PBL. DCMs were situated within the PBL in 67 % (inter-an-nual variation; ±12 %) of profi les with a water column division (presence of a PBL as such were found in ap-prox. 45 % of the profi les with an inter-annual variation of ± 23 %). Thus, in 33 % of these cases, the DCM was found above the PBL. DCMs were observed in 75 % (inter-annual variation ± 11 %) of the profi les examined.

Primary productionThe MADS database contains data for photosynthetic carbon uptake in Aarhus Bight from 225 days during 1999-2012 (exclusive 2004 and 2009). Each P-E (rate of photosynthesis versus light intensity) curve was based on seven data points (incubation bottles) and P-E curves were established for two different depths. The P-E pa-rameters, together with CTD and estimated chlorophyll a profi les were used to calculate water column PP for each of the sampling dates as described below. The fre-quency of P-E parameter measurements in the database ranged from 1 to 4 per month.

Measurements of PP were carried out following a modi-fi ed version (Markager 1998) of the carbon-14 technique presented by Steemann Nielsen (1952) both in MAPS and in the fi eld study. Samples were taken from two depths, i.e. the surface layer and deeper in the water col-umn. The surface layer sample in the MADS data was taken from an integrated surface water sample (from 0-10 m – or to the depth of the pycnocline) collected either with a hose submerged to 10 m or by combining and mixing bottle samples collected at selected depths in the surface layer. The surface sample in the fi eld cam-paigns was taken from 1 m. For both the MADS data and the fi eld campaigns, the deeper water sample was taken at the depth of the sub-surface chlorophyll maximum. When a deep chlorophyll maximum was not present (only in the MADS data), the second depth was chosen to be just below the pycnocline (located by visual in-spection of the density profi les). P-E curves were calcu-lated from 14C incubations with artifi cial light where the samples were incubated at seven different light intensi-ties for two hours using the equation in Webb (1974) according to Markager et al. (1999). Grids of metal were placed between the samples to achieve light attenuation (each grid attenuated 35 % of incident light).

To estimate water column PP, the parameters, derived from the P-E curves (alpha, Pmax and intercept) were normalised to chlorophyll a concentrations (measured in the same sample used for measurement of carbon up-take) and extrapolated over the water column. The pa-rameters from the surface sample were assumed to be valid and constant from surface to the depth of the pycn-

ocline. From there, they were interpolated linearly down to the depth of the second sample and, again from here, they were assumed to be constant down to the bottom.

Volumetric values for PP (mg C m-3 h-1) were calculated for each depth (0.2 m resolution) at 30 minute intervals from the chlorophyll specifi c P-E parameters, the con-tinuous chlorophyll profi le, Kd-values and SPAR. A to-tal of 261 depth profi les of PP were calculated; 255 from MADS and 6 from the fi eld campaigns in July 2010. Area rates were calculated as the sum over the day for all positive values of carbon uptake.

Annual PP, based on survey data, was estimated by cal-culating an average monthly rate and then multiplying this by the number of days in the given month. These monthly values were then added to give annual PP. In some cases, data were missing (e.g. winter months). In these cases, a long time average was calculated for the month for which data was missing. Data from other years, i.e., a missing value for November in a given year would be replaced by the average value for November for the study period as a whole.

PhytoplanktonSpecies identifi cation (species greater than about 2 μm) was carried out following the method described by Üter-möhl (1958). Carbon biomass (μg C l-1) was calculated after Strathmann (1967) and the carbon factors used were 0.13 pg C μm-3 for thecate dinofl agellates, 0.11 for athecate dinofl agellates and for diatoms. For diatoms, the plasma volume was calculated as: bio volume - (0.9 · vacuole volume). It was assumed that the plasma vol-ume was 1 μm thick and that the vacuole volume con-tributed 10 % to the carbon biomass.

During the fi eld study, determination of phytoplankton species was carried out at two discrete depths; surface (1 m) and at the depth of the deep PP sample (8-12.2 m). The samples were preserved with acidifi ed Lugol’s solution (app. 2 % in the sample) in brown glass bottles and identifi ed by ORBICON A/S (Aarhus, Denmark). Based on this identifi cation, we were able to character-ize the phytoplankton communities in different parts of the water column.

Variable Fluorescence using fast repetition rate fl uorometry (FRRF)Variable fl uorescence (Fv/Fm), which indicates the po-tential for electron transport in Photosystem II, was mea-sured with a FASTTRACKA II fl uorometer (Chelsea Instruments group Ltd.). The FASTTRACKA was used in its profi ling mode with the dark chamber enabled and a sinking velocity of maximum 0.2 m s-1. A single turnover



protocol with 12 sequences per acquisition each includ-ing 100 saturation fl ashlets was utilised. The sequence interval was set to 100 ms. Fv/Fm was calculated from a saturation phase fi t following Kolber et al. (1998). Only measurements with a qualifying saturation phase were included in the data set.

Pilot studies in our laboratory (unpublished) have shown that the Fv/Fm signal is infl uenced by the light history experienced by the phytoplankton and the Fv/Fm profi les in this study were, therefore, carried out during night.

NutrientsValues for dissolved inorganic nitrogen (DIN) and phos-phorus (DIP) were extracted from the MADS-database. Measurements of nutrients were conducted on samples from standard depths every fi ve meters down through the water column beginning at 1 meter below surface and ending 1 meter above the sediment. An average val-ue for the surface waters (1 m) and one for the bottom waters (10-16.6 m) was used to calculate the monthly concentrations shown in Fig. 2.

The sampling of DIN and DIP on the fi eld campaigns was carried out at 3 to 4 depths; one in surface (1 m), one in bottom waters (14 m) and 1 or 2 additional depths, where one of them was in the depth of the DCM. Both the DIN and DIP from the fi eld campaigns and from the MADS were measured with fl ow injection on a Scalar autoanalyser (Valderrama 1981, Grasshoff et al. 1999, Hansen and Koroleff 1999).

SedimentationIn the fi eld study, sediment traps were deployed on 5 occasions at 6, 10, 12 and 14 m for 24-48 hours. The traps had duplicate cylinders each with a diameter of 5.2 cm. They were used without baffl es and preserva-tives. Before re-suspension of the sedimented matter, the top layer of water in the cylinders was reduced to 650 ml. Aliquots (duplicates from one of two cylinders) were also taken for microscopy, analysis of chlorophyll a, POC and PON determinations. Correction for back-ground concentration of suspended matter in the ambi-ent water was made in calculations of sedimentation rates. One of the cylinders in the trap array was supplied with a petri dish (covering 93 % of the trap bottom) containing a high viscosity solution of 8% poly-acrylic-amide (PAA). This was done to catch and preserve intact particles and aggregates entering the trap (Lundsgaard et al. 1999). After retrieval, the dishes were taken up and the particles were allowed to settle to the bottom. A few drops of formaldehyde were added on the sur-face for long term preservation and the dishes were photographed. In addition, aliquots of suspended and

sedimented particulate matter from the sediment traps was collected on pre-combusted 13 mm Whatman GF/F fi lters and stored at –18 °C and measured for particulate organic carbon (POC) and nitrogen (PON) on a CarboE-lba Eager 200 CHN elemental analyser.

Analysis of oxygen concentration in the pycnocline/bottom layerTo explore the relationship between dissolved oxygen in the PBL during the summer months and different envi-ronmental variables, a partial least square (PLS) regres-sion was done. The PLS regression was used to identify and select a subset of predicting factors from a selec-tion of potential independent variables. The independent variables used in the initial parameter selection were N-loading (ton), P-loading (ton), freshwater infl ow (m3), cubed wind speed (m3 s-3), surface irradiance (μmol photons m-2 s-1), salinity (average of the layer above the pycnocline, no units), water temperature (average of the layer above the pycnocline, °C) and light attenuation coeffi cient, Kd (m-1). Surface irradiance was the same as the one used for PP (described above). The cubed wind speed was taken from the DMI weather station at location Tirstrup (56° 19’ N 10° 38’ E). The remaining parameters were extracted from the MAPS databas-es. Data on freshwater infl ow, N- and P-loading were based on a 3D MIKE SHE groundwater resource model validated on measured TN concentrations in Danish streams. For model details, see Windolf et al. (2011a). The hydrological area chosen for this study covered the inner Aarhus Bight (area 44, see Windolf et al. 2011 for area description).

In recognition of the fact that there may be a lag pe-riod between a change in an environmental variable and a response in the biological system, different time periods were examined in the relationship between the environmental parameters (independent variables) and the oxygen concentrations: The dependent (Y) variable, in this case oxygen concentration, was constrained to a specifi c time period (1 July to 30 September). Twelve different time periods for the independent environmen-tal variables were tested against the dependent variable. Six of these periods ended 1 July, i.e. the start of the time period for the dependent variable and 6 ended 30 September, i.e. the end of the time period for the de-pendent variable. The periods tested were 3, 4, 5, 7, 11 months prior to the end date and from Jan. the 1st of the previous year to either the 1st of July or 30th of Septem-ber. All data were normalised so that they have a mean of zero and a standard deviation of 1.

The last four years in the data set were chosen for cross validation and the combination of a reduction in both the root mean square error (RMSE) for the calibration


set and a reduction in the RMSE for the cross validation set was used to select a proper set of variables for the PLS regression. The PLS correlation was based on la-tent variables which means that the two parameters were not inter-correlated. This was checked initially using Pearson Correlations between parameters. The analysis was done in MATLAB® using a PLS program package from Eigenvector®.

RESULTS

Primary production in Aarhus Bay: 1999-2012

The annual average for total water column PP in Aarhus Bight over the period 1999-2012 was 155 ± 38 g C m-2 yr-1. The highest monthly PP was found during August (24 g C m-2 month-1, see Fig. 2 a). In contrast, the chl. a conc. (integrated over the entire water column) exhibit-ed relatively low values during the months from April to September. The average concentration of chl. a peaked during spring (March) and fall (November) (see Fig. 2 a). Integrated chl. a concentrations were consistently low during the months when PP and DPP were high (see below). On an annual basis, 42 % of the chl. a in the wa-ter column was found in the PBL. Light was, on average (based on 294 profi les), 1.5 % of surface values below 15 m and ranged from 0.37-1.49 umol m-2 s-1 at the start-ing depth of the PBL during the months from April to September where DPP was contributing considerably to the total water column PP.

DPP contributed 21 % (± 6 %) of the annual PP from 1999 to 2012 (see Fig 2 b). The occurrence of a stratifi ed water column with a considerable density difference (i.e., Δρ/Δz > 1 kg m-4) and a distinct PBL was most frequent from April to September. June had the highest frequency of days with a distinct PBL present (96 %, see Fig. 2b). The DPP was relatively constant throughout the summer months; June, July, August (mean 6.5 ± 0.7 g C m-2 month-1, see Fig. 2b). As a percentage of total PP, DPP was most important in June (36 %) and July (34 %).

The P-E parameters showed signifi cantly higher chlo-rophyll-normalized Pmax

B values at1 m than in the PBL (t-test, t = 4.03, p < 0.0001). This difference was also found for the light intensity at which photosynthesis is initially saturated, Ik (t-test, t = 9.22, p < 0.0001). The opposite was true for the αB values. Here, samples from the PBL showed signifi cantly higher values than sam-ples from 1 m depth (t-test, t = –2.64, p < 0.01).

The average DIN-concentration at 1 m was low from March to October and average DIP concentration was low from March to September. The bottom water con-centrations (10-16 m) never went below 2 μmol l-1 and 0.2 μmol l-1 for DIN and DIP, respectively (see Fig. 2c). A signifi cant negative relationship was found between the average monthly PP and the average monthly con-centration of DIN for the entire water column (r = –0.77, p < 0.005, n = 12). A similar relationship was found for water column PP and DIP (r = –0.85, p < 0.001).


PPChl. a

DIN: SDIN: B

DIP: SDIP: B

Freq pycnocline %DPP

Prim

ary

prod

uctio

n(g

C m

-2 m

onth

-1)

Dee

p pr

imar

y pr

oduc

tion

(g C

m-2 m

onth

-1)

DIN

(µm

ol l-1

1 )

Chl. a (g C

hl m-2)

Frequency of a pycnocline (%)

DIP

(µmol l -1)

Month

Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec

0

1

2

3

4

5

6

0

0.3

0.6

0.9

1.2

1.50

30

60

90

120

1500

5

10

15

20

25

30

0

3

6

9

12

150

2

4

6

8

10

a

b

c

Figure 2. Seasonal variation of average month values from 1999 to 2012 for the station in Aarhus Bight. (a) Monthly pri-mary production (PP) and depth-integrated monthly chl. a. values. (b) Deep primary production and frequency of days with a distinct PBL (Δρ/Δz > 1 kg m-4). (c) Average dissolved inorganic nitrogen and phosphate concentration for the sur-face waters at 1 m (DIN: S, DIP: S) and for bottom waters from 10-16.6 m (DIN: B, DIP: B).



Oxygen concentration in the pycnocline/ bottom layerThrough the PLS regression and the variable selection, we found that average salinity (surface layer i.e. above the PBL) from March to September and light attenua-tion coeffi cient (Kd) from June to September explained the greatest proportion of the variability in oxygen concentration in the PBL. Both the light attenuation and salinity increased the model’s ability to predict the average oxygen concentration in the PBL from July to September. The normalized PLS correlation coeffi cient showed that oxygen in the PBL was positively corre-lated with salinity in the surface layer (coeffi cient of 0.59) and negatively correlated with Kd (coeffi cient of –0.38) which means that it was positively correlated to the transparency of the water (Fig 3).

The surface layer salinity and Kd were signifi cantly and negatively correlated (R2 = 0.01, p < 0.05, n = 371) whereas the PBL salinity and Kd showed no relation-ship. This means that the surface layer salinity was most important in the prediction of Kd whereas the PBL sa-linity did not seem to have any explanatory effect on Kd. A linear regression between the average salinity of the surface mixed layer and the average salinity of the PBL, including the months from June – September and years from 1991-2012, showed that they were signifi -cantly and positively related (R2 = 0.38, p < 0.0001, n = 368). This means that changes in PBL salinity were infl uenced by changes in surface water salinity. The sur-face water salinity can, therefore, be used as an indica-tion of water exchange between the surface layer and

the PBL and as an indicator of the water transparency. In addition, the average oxygen concentration in the PBL (July – September, 1999 to 2012) was signifi cantly and positively correlated with the fraction of DPP occurring in this layer during the same time period (Fig. 4).

Field study in Aarhus Bight: 2010Two 5 day campaigns were carried out at the sampling station in July 2010. The water column at this time was characterized by strong stratifi cation with the starting depth of the pycnocline ranging from about 6 to 12 m (Fig. 5a, 5e). There was generally a higher concentra-tion of chlorophyll in the PBL than in the surface layer. In addition, during the fi rst week, there were patches of chlorophyll centered on the 1022 kg m-3 isopycnal i.e. in the upper region of the PBL (Fig. 5b, starting depth of PBL is noted with a stippled line). During the second campaign, chlorophyll patches were concentrated on the 1019 kg m-3 isopycnal also here in the upper region of the PBL (Fig. 5f).

Oxygen saturation was highest in the region of the layer just above and in the starting depth of the PBL (Fig. 5c, 5g). The distribution of oxygen saturation did not spe-cifi cally track the distribution of the chlorophyll con-centration although both were highest in the region of the PBL starting depth which, again, was higher than the concentrations found near the bottom.

TS diagrams (Fig. 5d, 5h) show two district clusters of data points associated with surface and bottom waters, respectively. This suggests that only two end-members were present and indicates that the horizontal gradient in the water mass properties in the area was small. The baroclinic forcing was, however, strong showing fl ows

Pre

dict

ed o

xyge

n (m

g O

2 l-1)

Observed oxygen (mg O2 l-1)

2 3 4 5 6 73

4

5

6

7Salinity: Mar-Sep x (0.59)kd: Jun-Sep x (-0.38)

Figure 3. Partial least square (PLS) regression for prediction of oxygen concentration in the pycnocline/bottom layer (PBL) during late summer in Aarhus Bight during the years 1991 to 2012. Salinity is average salinity in the water above the pyc-nocline from March to September and the average attenuation coeffi cient (Kd) from June to September. The intercept is 3.4 * 10-16. Averages are time weighted and all data are normalized (mean = 0 and SD = 1). The regression is based on 2 latent variables (salinity and Kd).

Oxy

gen

in th

e P

BL

(mg

O2 l

-1)

Percent deep primary production [%]0 20 40 60

0

2

4

6

8

10

Figure 4. Oxygen concentration in the pycnocline/bottom layer, PBL (mg O2 l-1) as a function of deep primary production (%) during the months from July to September from 1999 to 2012. A linear regression of the two parameters shows a signifi cant and positive relationship (r2 = 0.21, p < 0.01)



in different directions in different parts of the water col-umn. Figure 6 shows the observation of the ADCP as a progressive vector plot during sampling periods where direction and fl ow velocity at different depth intervals can be read off the fi gure directly. Typical fl ow speeds were 0.1 m s-1 or less, with a few examples of higher fl ow speeds as e.g. the 26th of July during the night time

where the bottom water velocity was fl owing approx. 0.26 m s-1 in a Southeast direction.

The deepest water layer is interesting in the evaluation of possible resuspension into the deepest sediment traps (14 m) and the depth intervals from the upper water col-umn (3-5 m and 8-10 m) are interesting in the evaluation

1212

1212

1313

1414

1414

15151616 1616

17171818 181819192020

2020

2121 2121

22222222

2

2

3

33

33

4

4

4

4

2.52.5

2.52.5

3.53.5

3.53.5

3.53.5

3.53.5

3.53.53.53.5

2

2

33

33

4.54.54

1.51.5

1.51.5

2.52.5

2.52.5

2.52.52.52.5

3.53.5

1010

1111

1212

1313 1414 151516161717

18181919

20202121

40405050 6060 8080

8080

8080

9090

9090

9090

7070

70707070

6060

6060 6060

7070 5050

5050 5050

8080 8080

8080

9090

9090

9090

9090

7070

7070

6060

60606060

7070

12

12

13

14

14

1516 16

1718 181920

20

21 21

2222

2

2

3

33

33

4

4

4

4

2.5

2.5

3.5

3.5

3.5

3.5

3.53.5

2

2

33

33

4.54

1.5

1.5

2.5

2.5

2.52.5

3.5

10

11

12

13 14 151617

1819

2021

4050 60 80

80

80

90

90

90

70

7070

60

60 60

70 50

50 50

80 80

80

90

90

90

90

70

70

60

6060

70

10 15 20 25 30 10 15 20 25 30Salinity (‰)

13-07 14-07 15-07 16-07 17-07 26-07 27-07 28-07 29-07 30-07

Dep

th (m

)D

epth

(m)

Dep

th (m

)D

ensi

ty –

100

0 (k

g m

-3)

Chl

orop

yll a

(µg

l-1)

Oxy

gen

satu

ratio

n (%

)Te

mpe

ratu

re (°

C)

0

2

4

6

8

10

12

14

160

2

4

6

8

10

12

14

160

2

4

6

8

10

12

14

16

a e

b f

c

d

g

h

5

10

15

20

25

Figure 5. CTD data from the two periods in July 2010 are shown as contour plots (hand drawn) of Sigma-t (a and e), chlorophyll a (b and f) and oxygen saturation (c and g). The starting depth of the pycnocline/bottom layer (PBL) is shown with dashed lines. Temperature and salinity data are shown for the observations made on d) 13 July at 07:30 (circles), 14 July at 09:30 (diamonds) and 16 July at 10:00 (crosses), and h) 26 July at 16:00 (circles), 28 July at 10:00 (diamonds) and 29 July at 14:00 (crosses). In the TS diagrams, contours of constant density are shown by the thin lines.



of the importance of advection for changes in standing stock. The progressive vector plot does not show any general or consistent fl ow direction and velocity. The difference in fl ow direction between the different depth intervals was, however, consistent where the three depth intervals, in many cases, showed completely different directions. This means that even though the density iso-clines (Fig. 5a, 5e) and the TS diagram (Fig. 5d, 5h) showed similar densities in a given depth interval, these could easily fl ow in different directions. The sampling during the 27th of July from 09:20 to 16:45 provides a good example of this situation. Here, the fl ow speeds were similar (0.09-0.11 m s-1) but directions different. The surface water (3-5 m) fl owed east southeast while the bottom water (13-15 m) was fl owing in the opposite direction at the same time that the depth interval of 8-10 m was heading south southeast.

PhytoplanktonOn all sampling dates in the two fi eld campaigns, the taxonomic composition of the phytoplankton popula-tion in the surface layer differed considerably from that found in the PBL. On all but one sampling date (14 July), the diatom, Proboscis alata, dominated in the surface waters while the PBL community was characterized by a high abundance of dinofl agellates, particularly Ceratium spp.

Several species were specifi c to the water layer in which they were found, i.e. never present in the other layer (Table 1). The number of species found exclusively in the PBL (15-18 species) on any given day was higher than the number of species found exclusively at 1 m (7-14 species). In total, 20-26 different species were identi-fi ed from water samples taken at 1 m and 24-36 different

Table 1. Phytoplankton species found exclusively either at 1 m depth or in the PBL (2). The samples analyzed were made on six different days during the period July 13th to 29th 2010.

1. Class 1. Species found at 1 m only 2. Class 2. Species found in the PBL only

Dinophyceae Alexandrium pseudogonyaulax Dinophyceae Ceratium lineatum

Prorocentrum minimum Ceratium longipes/horridum

Nostocophyceae Aphanizomenon sp. Ceratium macroceros

Dinophysis acuminate

Dinophysis norvegica

Protoperidinium oblongum

Diatomophyceae Guinardia fl accida

Chrysophyceae Dictyocha speculum

(+skeleton)

Figure 6. Progressive vector plots of the water masses in the depth intervals; 3-5 m (open circles), 8-10 m (squares) and 13-15 m (diamonds) based on Acoustic Doppler Current Profi ler (ADCP) observations. The plots cover a total of ten periods where ad-ditional water sampling was carried out. For each period the plots start at the solid circle and show the movements of the water masses in time relative to that point. The markers are placed at a time interval of 1 hour, allowing the average fl ow speed to be calculated by measuring the distance in between the symbols.

13-07 10:00-14:3513-07 22:30-14-07 00:45

14-07 09:30-11:2014-07 15:25-16:05

16-07 07:35-09:00

29-07 10:40-13:5028-07 21:00-23:20

28-07 09:55-15:30

27-07 09:20-16:45

26-07 22:10-27-07 00:20

1 km

West EastDistance

Sou

thN

orth

Dis

tanc

e



species from the deep samples. Thus, both species rich-ness (i.e. the total number of species) and the number of species exclusively found in a specifi c layer were high-est for the deep samples.

The photosynthesis characteristics (Table 2) recorded on 1 m samples and in the PBL indicate that the com-munities were segregated. Samples from 1 m showed a higher chlorophyll normalized Pmax than samples taken from the PBL. Phytoplankton from the PBL generally exhibited a similar or greater chlorophyll-specifi c α value (αB) than phytoplankton from the surface waters at 1 m depth.

The Ik values (Table 2) were signifi cantly higher for phytoplankton in 1 m depth compared to phytoplankton from the PBL (DF = 9, t = 9.22, p < 0.0001). The irradi-ance, PAR, reaching the depth of the start of the PBL

was, on average, 0.44-1.94 μmol m-2 sec-1 and total daily irradiance at this depth ranged from 5-22 % of surface values during the fi eld campaigns.

Carbon to chlorophyll ratios were expected to show highest values in samples from 1 m because the content of chlorophyll has been shown to decrease with increas-ing light intensities (Cullen 1982). The values found in this study were, however, highest in the water samples from the DCM during the 13th, 28th and 29th of July. This pattern was not consistent but, in each of the deep sam-ples exhibiting a higher carbon to chlorophyll ratio than that of surface samples, there was a large biomass of the dinofl agellates, Ceratium lineatum, Ceratium tripos, Ceratium longipes and Dinophysis norvegica.

Profi les of variable fl uorescence showed a consistent pattern of lower σPSII values and higher Fv/Fm values in

Table 2. Carbon biomass [μg C l-1] for diatoms, dinofl agellates and “other” (i.e. all remaining phytoplankton groups). The car-bon to chlorophyll ratio (C:Chl) and P-E parameters: Maximum photosynthesis normalized to chlorophyll (Pmax

B, mg C mg-1 Chl h-1) slope of the initial P-E curve normalized to chlorophyll (αB, mg C mg-1 Chl mol-1 m2) and the light intensity at which photo-synthesis initially is saturated (Ik, μmol photons m-2 sec-1) are shown for samples from 1 m depth and from the depth of the PBL on each sampling date. Signifi cant differences (t-test) between values found at the two different layers are marked with a star (p < 0.05) or three stars (p < 0.0001).

Date 13/7 14/7 16/7 27/7 28/7 29/7 Average 1 m

Average In DCM

Signifi cance of diff.

Depth 1 11 1 8 1 10 1 12 1 10 1 9 1 9.8

Diatoms 66 72 127 85 16 66 49 69 35 49 35 68 55

Dinofl agellates 165 13 18 7 222 26 127 1 243 30 546 21 220 *

Other 22 9 13 20 11 19 11 18 11 35 21 20 15

C:Chl 134 99 62 63 75 94 84 94 129 84 162 87 108

N:P 4.6 1.9 7.0 2.1 3.2 0.8 21.2 1.3 2.8 3.4 29.7 1.5 11.4 1.8 ***

PmaxB 4.1 3.8 4.6 2.4 2.9 1.4 4.8 1.8 2.1 1.3 3.7 2.1 P = 0.05

αB 6.4 12.1 7.4 10.7 3.8 3.9 6.5 4.3 2.7 3.9 5.4 7

Ik 176 86 174 63 210 102 203 117 216 93 195.8 92.2 ***

Above PBLIn PBL

σ (nm2)

F v/Fm

0.8 1.2 1.6 2.0 2.4 0.8 1.2 1.6 2.0 2.4 0.8 1.2 1.6 2.0 2.4 0.8 1.2 1.6 2.0 2.40.30

0.35

0.40

0.45

0.50

0.5526-07-201313-07-2013 28-07-2013 29-07-2013

Figure 7. Variable fl uorescence (Fv/Fm) as a function of the effective cross section in photosystem II (σPSII) for the 4 night profi les measured during the July 2010 fi eld study. Filled circles indicate values from the waters above the pycnocline/bottom layer, PBL, and the open circles indicate values from waters in the PBL.



the PBL than in the surface layer (Fig. 7). This supports the conclusion that there are photosynthetically active phytoplankton in the PBL and that the cells here are, generally, larger than those in the surface waters. The higher σPSII in the surface waters than in the PBL in-dicates the presence of smaller cells there than in the deeper waters. Thus, there were no indications that the phytoplankton communities in the PBL were not com-prised of healthy active cells.

SedimentationThe highest sedimentation rates were consistently found in the deepest samples (see Table 3 and Fig. 8). The ve-locity of sinking particulate organic C at 14 m was 2.4–4 times greater than sinking rates at 6 m. The pattern of nitrogen sedimentation mirrored that of carbon and the sinking rate increased by a factor 2.1-3.4 between 6 and 14 m. Thus, we consistently found that the material pro-duced in the surface waters sank more slowly than the material in the PBL (t-test, t = -2.42, df = 12, p = 0.032).

Table 3. The sinking rate (m day-1) for particulate carbon and nitrogen at different depths (m) during the 5 days of sediment trap employments. Average values for the waters in the surface and the pycnocline/bottom layer (PBL) are noted in the two columns on the right. The values of carbon sinking rates above the PBL are signifi cantly lower than the ones found in the PBL (t = –2.42, df = 12, p < 0.05). *Indicates level of signifi cance (P < 0.05).

14th of July 16th of July 27th of July 28th of July 29th of July Surface In PBL

Depth 6 10 12 14 10 12 14 6 10 12 14 6 10 12 14 6 10 12 14 7.3 12.3

C sinking rate 0.4 0.6 0.4 1.6 0.5 0.4 1.1 0.5 0.4 0.7 2.3 0.3 0.2 0.3 1.0 0.4 0.6 0.5 1.0 0.4 0.8*

N sinking rate 0.4 0.6 0.5 0.9 0.4 0.4 0.9 0.6 0.4 0.5 1.9 0.3 0.2 0.3 0.7 0.4 0.6 0.5 0.8 0.4 0.7

>56 µm

0.7-10 µm

10-56 µm

Sedimentation (m day-1)

Primary production (mg C m-3 day-1)

Dep

th (m

)D

epth

(m)

0 0.5 1.0 1.5 2.0 2.5

0 40 80 120 160 0 40 80 120 160

-16

-14

-12

-10

-8

-6

-4

-2

0

-16

-14

-12

-10

-8

-6

-4

-2

0

a b c

d e f

0 0.5 1.0 1.5 2.0 2.5 0 0.5 1.0 1.5 2.0 2.5

0 40 80 120 160

Figure 8. Depth profi les of primary production with pie charts illustrat-ing the distribution of size-fractionated chl. a (%) and sinking rates of carbon (squares). The size-fractionated chl. a values represent the % of the total chloro-phyll in the sample that was retained on a GFF (nominally 0.7 μm), 10 μm and 56 μm fi lter, respectively. Stippled lines indicate the start-ing depth of the pycno-cline/bottom layer.



Size-fractionated chl. a showed that large cells were more abundant in the PBL than in the surface layer (Fig. 8). This result agrees well with the species data reported above, where the larger dinofl agellates were abundant in the PBL and smaller diatoms in the surface layer.

Pictures of trap material from the PBL showed a high number of fecal pellets and of phytoplankton cells of species from the Ceratium genus. No sand grains were observed in the pictures, suggesting that little re-sus-pended material was being collected in the traps.

DISCUSSION

Concerns about eutrophication effects resulting from anthropogenic nutrient enrichment in coastal seas typi-cally focus on changes in total water column primary production, PP, as it is assumed that changes in total PP will directly lead to changes in the amount of or-ganic material degraded in the system and, hence, oxy-gen conditions (see Rabalais et al. 2002). However, this study clearly demonstrates that the fate of the carbon produced via photosynthesis in this stratifi ed coastal system is highly dependent upon the vertical distribu-tion pattern of the PP occurring here.

The infl uence of photosynthesis on oxygen conditions in the pycnocline/bottom watersEverything else being equal, the amount of photosyn-thesis occurring in the PBL will be a function of light penetrating to these depths. Lyngsgaard et al. (submit-ted) demonstrated a signifi cant relationship between surface irradiance and the magnitude of DPP. Analysis of a long-term data series in this study demonstrates that advection processes can infl uence light conditions in the PBL. Oxygen concentrations in this layer are, at least initially, a function of physical processes, i.e. advection and vertical mixing (Bendtsen et al. 2009, Hansen and Bendtsen 2013). Advective processes transport water with a relatively high light attenuation and low salin-ity from the Baltic Sea, and waters with a lower light attenuation and higher salinity from the North Sea/Sk-agerrak compared to that of the Baltic Sea. The light attenuation is dependent upon the concentration of col-oured dissolved organic matter, CDOM (Markager and Vincent 2000) and the Baltic Sea contains more CDOM in general than the North Sea/Skagerrak (Stedmon et al. 2010). The surface water at the study station in Aarhus Bight originates mainly in the Baltic Sea with approx. 60 % and the bottom water approx. 50 %, whereas the remaining part comes from the North Sea/Skagerrak and German Bight (Stedmon et al. 2010).

Thus, when a change in salinity is observed, it can mainly be attributed to 2 mechanisms; 1) how much freshwater the system receives from runoff containing high nutrient concentrations, detritus and DOM and 2) the fraction of Baltic Sea water compared to North Sea/Skagerrak water in the system. Fig. 3 illustrates that, for Aarhus Bight, a considerable amount of the variation in O2 concentration here can be explained by the salinity and water transpar-ency (i.e. attenuation coeffi cient, Kd). As the salinity of the surface water was coupled to the salinity of the PBL, and because the variation in Kd was determined in the surface layer and not in the PBL, the surface layer salinity could then couple the impact of water transparency (con-trolled by advection and runoff from land) and exchange (by advection) of oxygen rich waters from the Skagerrak, to the PBL oxygen concentration.

When there was an increase in oxygen-rich North Sea/Skagerrak water being transported into the Aarhus Bight area (often as a result of increased westerly winds), the infl ow of water from the Baltic Sea decreased resulting in an increased transparency of the surface water. Natu-rally, the relationship between CDOM and salinity is not constant and other processes, such as plankton and lo-cal sources of detritus, will also affect light attenuation. This could explain why the relationship between Kd and salinity, found in this study, is relatively weak. In any case, improved water transparency will in turn stimu-late DPP, since this is assumed to be light limited and it will, therefore, react with a linear positive response to increasing light intensities.

A high DPP was coincident with high oxygen concen-trations in the PBL. This can be seen by the signifi cant positive relationship between the magnitude of DPP and oxygen conditions in the same layer (Fig. 4). The relationship between DPP and O2 concentration in the PBL indicates that DPP makes a quantitatively impor-tant contribution to the oxygen conditions here. DPP is here estimated to contribute approximately 21 % of the total annual PP occurring at this station during the pe-riod 1999 to 2012.

From this, we can estimate an annual oxygen produc-tion in the PBL of 88 g O2 m-2 yr-1 assuming a photosyn-thetic quotient (PQ) of 1 and a total annual PP of 155 g C m-2 yr-1. The oxygen demand in the bottom layer has been estimated to vary from 48.8 % (Bendtsen and Hansen 2013) to 58 % (Fossing et al. 2002) of PP. Using these values, DPP can be estimated to contribute with 37-44 % of the annual oxygen demand in the PBL in Aarhus Bight.

In addition to producing oxygen, DPP will also result in an increase in organic material in this layer. The deg-radation of this material will, of course, contribute to



the respiratory oxygen demand in the BL following the sinking of the material. However, in contrast to organic material produced in surface waters and sinking to the bottom waters, the oxygen necessary for the degradation of material produced in the PBL is compensated for by the oxygen produced by the photosynthesis producing the material. Thus, it becomes important when consider-ing the relationship between PP and bottom water oxy-gen conditions to differentiate this layer from the rest of the water column.

Surface versus deep primary productionThe photosynthetic and optical parameters recorded from the phytoplankton communities retrieved during the July 2010 fi eld study from surface waters and the PBL indicate two actively photosynthesizing popula-tions, each adapted to the specifi c light characteristics of the water layer in which they are found (light accli-matisation is reviewed in for example Richardson et al 1983; Lindley et al. 1995 and Morán and Estrada, 2001). The relatively high Fv/Fm values recorded in the PBL again support the conclusion that the phytoplankton here were physiologically active and capable of active photosynthesis (see Beardall et al. 2001 for relevance of the method).

Direct comparison of variable fl uorescence characteris-tics between the surface waters and the PBL is compli-cated by the fact that cell size and taxonomic composi-tion of the community, in addition to nutrient status, also infl uences both Fv/Fm and the effective cross section in photosystem II. As an example of this, dinofl agellates have shown to generate lower Fv/Fm and higher effec-tive cross section than diatoms (Suggett et al. 2009). As we know that the phytoplankton community was domi-nated by dinofl agellates in the PBL, we attributed the higher Fv/Fm and lower σPSII values found in the PBL to be an indication of an increased nutrient status and a shift in cell size from small to larger cells (Suggett et al. 2009). Thus, our observations from the Fv/Fm confi rm the results found by chlorophyll a fractionation showing a higher proportion of larger cells in the PBL.

The indication by Fv/Fm of more nutrient replete phy-toplankton in the PBL than in the surface layer was con-sistent with the higher DIN and DIP concentrations re-corded in the PBL. The surface waters show, in general, lower nutrient concentrations during summer months than the bottom waters (10-16.6 m) which never showed average DIN and DIP concentrations below 2 and 0.3 μmol l-1, respectively. The low nutrient concentrations of both DIN and DIP in surface waters, compared to bot-tom waters, were considered in this study, likely to be the main reason why a large part of the summer PP took place deeper as DPP.

Comparison of phytoplankton community composition in surface waters and in the pycnocline/bottom layer

Several lines of evidence (i.e. species taxonomy, cell sizes, variable fl uorescence, photosynthetic parameters and water movements) support our initial hypothesis that the water column was segregated into different lay-ers not only physically but also with respect to ecosys-tem functioning.

A number of species were characteristic of one specifi c layer and never encountered in the other layer (Table 1) Such distinct vertical stratifi cation in species distribu-tions where dinofl agellates were abundant in the PBL has been noted earlier in Aarhus Bight (Mouritsen and Richardson, 2003) but that study was made only on a single day.

Size fractionation of chlorophyll a, taxonomic identifi -cation of species and variable fl uorescence showed, in-dependently of each other, that phytoplankton cells were larger in the PBL than in the surface layer. When nutri-ents are scarce as was observed in the surface waters, it is profi table to have a small surface to volume ratio to increase the effi ciency of which nutrients are taken up (Kiørboe 1993). In contrast, this size strategy would be unlikely to be as important in the PBL, where nutrient concentrations were higher. It should be noted, however, that many of the species found in the PBL are known mixotrophs (see Jeong et al. 2010 and Stoecker 1999 and references therein). Thus, the size/nutrient relation-ship demonstrated by Kiørboe (1993) may not strictly apply here.

That mixotrophy may be an important nutritional strat-egy of the community in the PBL may explain the rel-ative high carbon to chl. a ratios found here on some sampling dates. Dinofl agellates have been shown to contribute more to the carbon than the chlorophyll bio-mass because they have a high carbon to chlorophyll ratio (Geider 1987, Tang 1996) and a high degree of mixotrophy has been discovered within this group (see review by Jeong et al. 2010) . For strictly autotrophic communities which depend on their ability to catch and utilize light, these carbon to chlorophyll ratios would have been expected to be lower as decreasing light in-tensities have been inversely related to increases in the amount of chlorophyll per cell (Kiefer et al. 1976, Cul-len 1982, Falkowski and Kiefer 1985, Fennel and Boss 2003). Mixotrophy might also be part of the explanation for why the photosynthesis to biomass ratio, P:B, was considerably lower in the PBL (0.008-0.25, mean = 0.1) than in the surface ( 0.17-0.63, mean = 0.38).



Photosynthetic rates could not explain the increase in phytoplankton carbon biomass observed in the end of the second fi eld campaign in the PBL. Given the altered carbon to chlorophyll ratios and the increased abundance of Ceratium sp., it seems likely that mixotrophy may have contributed to the production of organic material here. That phytoplankton is known to occur in patches rather than being distributed uniformly in ocean waters (Steele 1974) further suggests that advective processes (showing considerable fl ow speeds) were also important and could most likely explain part of the variability in biomass at this station in general.

Water column characteristicsAn oversimplifi cation of the water column hydrography to a two layer system was done in this study because we were interested in looking at the layer(s) where O2 was in exchange with the atmosphere separately from the layer(s) where it was not. Therefore, we included all layers where we did not believe that there was a direct atmospheric O2 exchange together as one (PBL). That the water column has additional layers was insinuated by the fl ow directions of the water. This only further supports our conclusions that phytoplankton activity and species composition vary considerably in different parts of the water column and emphasizes the impor-tance of evaluating the different layers within the wa-ter column separately, especially in strongly stratifi ed ocean regions.

Comparison of sinking characteristics of particulate material produced in different depths of the water columnWe would expect (Smayda 1969) the larger sized phyto-plankton in the PBL to sink more quickly than the smaller cells in the surface layer and, indeed, the patterns of sedi-mentation rates recorded in the fi eld study reported here suggest this to be the case. We recognise that there can be problems that complicate interpretation of sediment stud-ies made in shallow waters (Buesseler 1991). However, we believe that our conclusion of increased sedimenta-tion in the PBL is robust for a number of reasons.

Re-suspension of bottom material into sediment traps would be the greatest concern in interpreting our results. Lund-Hansen and colleagues (1999) suggested that the bottom current velocity must be higher than 0.1 m s-1 to actually lift the organic material from the bottom and Bloesch and Burns (1980) state that it has to be < 0.14 m s-1 in order to avoid re-suspension into the deep sediment traps. The average bottom fl ow speeds were well below 0.14 m s-1 in this study except for the 26th of July during late night where the bottom fl ow speed was 0.26 m s-1.

The high fl ow rates might have affected the sedimenta-tion results presented in Fig. 8d, i.e. overestimation of the deepest sedimentation velocity. However, the gen-eral lack of sand grains in photographed material from all deep sediment traps indicates that re-suspension was not likely to be controlling the sediment trap results.

Implications for ecologyThe large amount of fecal pellets found in the sediment traps placed in this study in the PBL compared to those in the surface waters suggests that the population in the PBL was being actively grazed as previously sug-gested by Nielsen et al. (1993). L arge particle size has been shown to increase sedimentation rates (Smayda 1969) and one can, therefore, predict that the larger phytoplankton cell size found in the PBL as well as the presence of fecal pellets would lead to increased sedi-mentation rates of the material produced in the PBL as opposed to higher up in the water column.

The effi cient transfer of organic material indicated by enhanced sedimentation in the PBL may provide an im-portant food supply for benthic feeders. We note that Josefson and colleagues (1995) showed that the shell growth of Arctica islandica was highest when they were located immediately below the pycnocline as opposed to further away (deeper) from the pycnocline depth, presumably refl ecting the availability of high quality or-ganic matter produced here. Thus, changes in the verti-cal distribution of PP may also have consequences for benthic food webs.

It is often assumed that total water column PP is the most important factor controlling the plankton food web and carbon fl ow in the water column. This study demonstrates, however, that the vertical distribution of PP in the water column is important for both oxygen and carbon dynamics. An analysis of a long-term data set confi rms that the magnitude of DPP along with ad-vection processes is important in controlling oxygen conditions in pycnocline/bottom waters. Furthermore, at the time of the fi eld study, phytoplankton communi-ties in the PBL were dominated by large phytoplankton which most likely together with high grazing may have resulted in relatively rapid transfer of material to the benthos.

This study suggest that there is a need to quantify the seasonal development in the activity and species distribu-tion of phytoplankton communities in the PBL as well as higher up in the water column in order to develop a bet-ter understanding of the factors infl uencing the vertical distribution of PP, sedimentation and trophic strategies.



ACKNOWLEDGEMENTS

This study received funding from grant number 2104-09-063212 and 2104-09-67259 from the Strategic Re-search Council of Denmark. Additional support was received from Center for Macroecology Evolution and Climate, the Natural History Museum of Denmark, Co-penhagen University and Department for Bioscience, Aarhus University. We thank Department for Biosci-ence, Aarhus University for access to the monitor-ing data and the outstanding crew and captain on the research vessel Tyra for invaluable help and support throughout the fi eld campaigns.

References

Beardall, J., E. Young and S. Roberts 2001. Approaches for determining phytoplankton nutrient limitation. Aquatic Sci-ences 63:44-69.

Beckmann, A. and I. Hense 2007. Beneath the surface: Char-acteristics of oceanic ecosystems under weak mixing condi-tions – A theoretical investigation. Progress in Oceanography 75:771-796.

Bendtsen, J., K.E. Gustafsson, J. Söderkvist and J.L.S. Han-sen 2009. Ventilation of bottom water in the North Sea-Baltic Sea transition zone. Journal of Marine Systems 75:138-149.

Bendtsen, J. and J.L.S. Hansen 2013. Effects of global warm-ing on hypoxia in the Baltic Sea–North Sea transition zone. Ecological Modelling 264:17-26.

Benitez-Nelson, C.R., R.R. Bidigare, T.D. Dickey, M.R. Landry, C.L. Leonard, S.L. Brown, F. Nencioli, Y.M. Rii, K. Maiti, J.W. Becker, T.S. Bibby, W. Black, W.-J. Cai, C.A. Carl-son, F. Chen, V.S. Kuwahara, C. Mahaffey, P.M. McAndrew, P.D. Quay, M.S. Rappé, K.E. Selph, M.P. Simmons and E.J. Yang 2007. Mesoscale Eddies Drive Increased Silica Export in the Subtropical Pacifi c Ocean. Science 316:1017-1021.

Bloesch, J. and N.M. Burns 1980. A critical review of sedi-mentation trap technique. Schweiz. Z. Hydrol. 1:1-42.

Buesseler, K.O. 1991. Do upper-ocean sediment traps pro-vide an accurate record of particle fl ux? Nature 353:420-423.



Estrada, M., C. Marrasé, M. Latasa, E. Berdalet, M. Delgado and T. Riera 1993. Variability of deep chlorophyll maximum characteristics in the Northwestern Mediterranean. Marine Ecology Progress Series 92:289-300.

Falkowski, P. and D.A. Kiefer 1985. Chlorophyll a fl uore-scence in phytoplankton: relationship to photosynthesis and biomass. Journal of Plankton Research 7:715-731.

Falkowski, P.G. 1984. Physiological responses of phyto-plankton to natural light regimes. Journal of Plankton Re-search 6:295-307.

Fennel, K. and E. Boss 2003. Subsurface maxima of phyto-plankton and chlorophyll: Steady-state solutions from a simple model. Limnology and Oceanography 48:1521-1534.

Fossing, H., B. Thamdrup, S. Rysgaard, H.M. Sørensen and K. Nielsen 2002. Ilt- og næringsstoffl uxmodel for Århus Bugt og Mariager Fjord. Modelopsætning og scenarier. 147.

Geider, R.J. 1987. Light and Temperature Dependence of the Carbon to Chlorophyll a Ratio in Microalgae and Cyano-bacteria: Implications for Physiology and Growth of Phyto-plankton. New Phytologist 106:1-34.

Grasshoff, K., K. Kremling and M. Ehrhardt 1999. Methods of Seawater Analysis. 3 edition. Wiley-VCH.

Hansen, H.P. and F. Koroleff. 1999. Determination of nutri-ents. Pages 159-228 in K. Grasshoff, K. Kremling and M. Ehrhardt, editors. Methods of seawater analysis, Germany.

Hansen, J.L.S. and J. Bendtsen 2013. Parameterisation of oxygen dynamics in the bottom water of the Baltic Sea – North Sea transition zone. Marine Ecology Progress Series 481:25-39.

Hickman, A.E., C.M. Moore, J. Sharples, M.I. Lucas, G.H. Tilstone, V. Krivtsov and P.M. Holligan 2012. Primary pro-duction and nitrate uptake within the seasonal thermocline of a stratifi ed shelf sea. Marine Ecology Progress Series 463:39-57.

Holligan, P.M., P.J.l. Williams, D.A. Purdie and R.P. Har-ris 1984. Photosynthesis, respiration and nitrogen supply of plankton populations in stratifi ed, frontal and tidally mixed shelf waters. Marine Ecology Progress Series 17:201-213.

Holm-Hansen, O. and C. Hewes 2004. Deep chlorophyll-a maxima (DCMs) in Antarctic waters. Polar Biology 27:699-710.

Jeong, H., Y. Yoo, J. Kim, K. Seong, N. Kang and T. Kim 2010. Growth, feeding and ecological roles of the mixotrophic and heterotrophic dinofl agellates in marine planktonic food webs. Ocean Science Journal 45:65-91.

Josefson, A.B., J.N. Jensen, T.G. Nielsen and B. Rasmussen 1995. Growth parameters of a benthic suspension feeder along a depth gradient across the pycnocline in the southern Katte-gat, Denmark. Marine Ecology Progress Series 125:107-115.

Kiefer, D.A., R.J. Olson and O. Holm-Hansen 1976. Another look at the nitrite and chlorophyll maxima in the central North Pacifi c. Deep Sea Research 23:1199-1208.

Kiørboe, T. 1993. Turbulence, Phytoplankton Cell Size, and the Structure of Pelagic Food Webs. Advances in marine bio-logy 29:1-72.

Klausmeier, C.A. and E. Litchman 2001. Algal Games: The Vertical Distribution of Phytoplankton in Poorly Mixed Wa-ter Columns. Limnology and Oceanography 46:1998-2007.



Kolber, Z., J. Zehr, and P. G. Falkowski. 1998. Effects of growth irradiance and nitrogen limitation on photosynthetic energy con-version in photosystem II. Plant Physiol. 88:923-929.

Kononen, K., M. Huttunen, S. Hällfors, P. Gentien, M. Lun-ven, T. Huttula, J. Laanemets, M. Lilover, J. Pavelson and A. Stips 2003. Development of a deep chlorophyll maximum of Heterocapsa triquetera Ehrenb. at the entrance to the Gulf of Finland. Limnology and Oceanography 48:594-607.

Lehrter, J.C., M.C. Murrell and J.C. Kurtz 2009. Interactions between freshwater input, light, and phytoplankton dynamics on the Louisiana continental shelf. Continental Shelf Re-search 29:1861-1872.

Lindley, S.T., R.R. Bidigare and R.T. Barber 1995. Phyto-plankton photosynthesis parameters along 140°W in the equa-torial Pacifi c. Deep Sea Research Part II: Topical Studies in Oceanography 42:441-463.

Lips, U., I. Lips, T. Liblik and N. Kuvaldina 2010. Processes responsible for the formation and maintenance of sub-surface chlorophyll maxima in the Gulf of Finland. Estuarine, Coast-al and Shelf Science 88:339-349.

Lorenzen, C.J. 1967. Determination of chlorophyll and pheo-pigments: Spectrophotometric equations. Limnology and Oceanography 12:343-346.

Lund-Hansen, L.C. 2006. Development and dynamics of a coastal sub-surface phytoplankton bloom in the southwest Kattegat, Baltic Sea. Oceanologia 48:1-14.

Lund-Hansen, L.C., M. Petersson and W. Nurjaya 1999. Ver-tical sediment fl uxes and wave-induced sediment resuspen-sion in a shallow-water coastal lagoon. Estuaries 22:39-46.

Lundsgaard, C., M. Olesen, M. Reigstad and K. Olli 1999. Sources of settling material: aggregation and zooplankton mediated fl uxes in the Gulf of Riga. Journal of Marine Sys-tems 23:197-210.

Markager, S. 1998. Dark uptake of inorganic14C in oligotrophic oceanic waters. Journal of Plankton Research 20:1813-1836.

Markager, S., W. Vincent and E.Y. Tang 1999. Carbon fi xa-tion by phytoplankton in high Arctic lakes: Implications of low temperature for photosynthesis. Limnology and Ocean-ography 44:597-607.

Markager, S. and W.F. Vincent 2000. Spectral Light Attenu-ation and the Absorption of UV and Blue Light in Natural Waters. Limnology and Oceanography 45:642-650.

Moore, C.M., D.J. Suggett, A.E. Hickman, Y.-N. Kim, J.F. Tweddle, J. Sharples, R.J. Geider and P.M. Holligan 2006. Phytoplankton photoacclimation and photoadaptation in re-sponse to environmental gradients in a shelf sea. Limnology and Oceano-graphy 51:936-949.

Morán, X.A.G. and M. Estrada 2001. Short-term variability of photosynthetic parameters and particulate and dissolved primary production in the Alboran Sea (SW Mediterranean). Marine Ecology Progress Series 212:53-67.

Mouritsen, L.T. and K. Richardson 2003. Vertical micro-scale patchiness in nano- and microplankton distributions in a stratifi ed estuary. Journal of Plankton Research 25:783-797.

Murrell, M.C., J.G. Campbell, J.D. Hagy Iii and J.M. Caffrey 2009. Effects of irradiance on benthic and water column pro-cesses in a Gulf of Mexico estuary: Pensacola Bay, Florida, USA. Estuarine, Coastal and Shelf Science 81:501-512.

Maar, M. and J.L.S. Hansen 2011. Increasing temperatures change pelagic trophodynamics and the balance between pelagic and benthic secondary production in a water column model of the Kattegat. Journal of Marine Systems 85:57-70.

Neumann, T. 2000. Towards a 3D-ecosystem model of the Baltic Sea. Journal of Marine Systems 25:405-419.

Nielsen, T.G., B. Løkkegaard, K. Richardson, F.B. Pedersen and L. Hansen 1993. Structure of plankton communities in the Dogger Bank area (North Sea) during a stratifi ed situa-tion. Marine Ecology Progress Series 95:115-131.

Nixon, S.W. 1995. Coastal marine eutrophication: A defi ni-tion, social causes, and future concerns. Ophelia 41:199-219.

Platt, T., C.L. Gallegos and W.G. Harrison 1980. Photo-inhibition of photosynthesis in natural assemblages of marine phytoplankton. Journal of Marine Research 38.

Rabalais, N., R.E. Turner, Q. Dortch, D. Justic, V. Bierman, Jr. and W. Wiseman, Jr. 2002. Nutrient-enhanced productiv-ity in the northern Gulf of Mexico: past, present and future. Hydrobiologia 475/476:39-63.

Reid, P.C., C. Lancelot, W.W.C. Gieskes, E. Hagmeier and G. Weichart 1990. Phytoplankton of the North Sea and its dynamics: A review. Netherlands Journal of Sea Research 26:295-331.


Richardson, K. and A. Christoffersen 1991. Seasonal distri-bution and production of phytoplankton in the Southern Kat-tegat. Marine Ecology Progress Series 78:217-227.

Richardson, K., S. Markager, E. Buch, M.F. Lassen and A.S. Kristensen 2005. Seasonal distribution of primary produc-tion, phytoplankton biomass and size distribution in the Greenland Sea. Deep-Sea research 52:979-999.

Richardson, K., T.G. Nielsen, F.B. Pedersen, J.P. Heilmann, B. Løkkegaard and H. Kaas 1998. Spatial heterogeneity in the structure of the planctonic food web in the North Sea. Marine ecology progress series 168:197-211.

Richardson, K. and F.B. Pedersen 1998. Estimation of new production in the North Sea: consequences for temporal and spatial variability of phytoplankton. ICES Journal of Marine Science: Journal du Conseil 55:574-580.


Riley, G.A., H.M. Stommel and D.F. Bumpus 1949. Quanti-tative ecology of the plankton of the western North Atlantic. Bulletin Bingham Oceanography 12:1-169.

Ross, O.N. and J. Sharples 2007. Phytoplankton motility and the competition for nutrients in the thermocline. Marine Eco-logy Progress Series 247:21-38.



Sedwick, P.N., T.M. Church, A.R. Bowie, C.M. Marsay, S.J. Ussher, K.M. Achilles, P.J. Lethaby, R.J. Johnson, M.M. Sarin and D.J. McGillicuddy 2005. Iron in the Sargasso Sea (Bermuda Atlantic Time-series Study region) during sum-mer: Eolian imprint, spatiotemporal variability, and ecologi-cal implications. Global Biogeochemical Cycles 19:GB4006.


Smayda, T.J. 1969. Some measurements of the sinking rate of fecal pellets. Limnology and Oceanography 14:621-625.


Stedmon, C.A., C.L. Osburn and T. Kragh 2010. Tracing wa-ter mass mixing in the Baltic–North Sea transition zone using the optical properties of coloured dissolved organic matter. Estuarine, Coastal and Shelf Science 87:156-162.

Steele, J.H. 1974. Patchiness. Blackwell, London.


Stoecker, D.K. 1999. Mixotrophy among Dinofl agellates1. Journal of Eukaryotic Microbiology 46:397-401.

Strathmann, R.R. 1967. Estimating the Organic Carbon Con-tent of Phytoplankton from Cell Volume or Plasma Volume. Limnology and Oceanography 12:411-418.


Strom, S.L., E.L. Macri and K.A. Fredrickson 2010. Light limitation of summer primary production in the coastal Gulf of Alaska: physiological and environmental causes. Marine Ecology Progress Series 402:45-57.

Suggett, D.J., C.M. Moore, A.E. Hickman and R.J. Geider 2009. Interpretation of fast repetition rate (FRR) fl uore-scence: signatures of phytoplankton community structure versus physiological state. Marine Ecology Progress Series 376:1-19.

Sunda, W.G. and D.R. Hardison 2007. Ammonium uptake and growth limitation in marine phytoplankton. Limnology and Oceanography 52:2496-2506.

Tang, E.P.Y. 1996. Why do dinofl agellates have lower growth rates? Journal of Phycology 32:80-84.

UNESCO 1981. International oceanographic tables, Paris.

Utermöhl, H. 1958. Zur vervollkommung der quantitativen phytoplankton methodik. Mitt. Int. Ver. Theor. Angew. Lim-nol 9:1-38.

Valderrama, J.C. 1981. The simultaneous analysis of total nitrogen and total phosphorus in natural waters. Marine Che-mistry 10:109-122.


Weston, K., L. Fernand, D.K. Mills, R. Delahunty and J. Brown 2005. Primary production in the deep chlorophyll max-imum of the central North Sea. J. Plankton Res. 27:909-922.

Windolf, J., H. Thodsen, L. Troldborg, S.E. Larsen, J. Bøgestrand, N.B. Ovesen and B. Kronvang 2011a. A distrib-uted modeling system for simulation of monthly runoff and nitrogen 878 sources, loads and sinks for ungauged catch-ments. Journal of Environmental Monitoring 13:2645-2658.

Windolf, J., P. Wiberg-Larsen, J. Bøgestrand, S.E. Larsen, H. Thodsen, R. Bjerring, N.B. Ovesen, A. Kjeldgaard and B. Kronvang 2011b. Vandløb 2010. NOVANA, Aarhus Univer-sitet, DCE-Nationalt Center for Miljø og Energi, http://www.dmu.dk/pub/SR4.pdf


In preparation for Estuaries, Coastal and Shelf Science

Maren Moltke Lyngsgaard1,2, Stiig Markager2, Katherine Richardson1 and Eva Friis Møller2

1Center for Macroecology, Evolution and Climate, Danish Natural History Museum, University of Copenhagen, DK-2200 Copenhagen, Denmark2Department of Bioscience, Marine Diversity and Experimental Ecology, Aarhus University, DK-4000 Roskilde, Denmark

PAPER IIISeasonal variation in biological parameters in a temperate coastal area – decoupling of chlorophyll concentration andprimary production

Photo: Peter Bondo Christensen


Abstract

The analysis of 1385 primary production (PP) and chlorophyll profi les, made from 1998-2012 at 6 different stations in the Baltic Sea Transition Zone (BSTZ), showed that PP was decoupled from chlo-rophyll a concentration both in time and vertical space. PP was highest during summer, whereas the chlorophyll biomass peaked during spring and fall. This decoupling was mainly attributed to seasonal changes in the carbon to chlorophyll ratio. The use of chlorophyll as a proxy for biomass may have overestimated the importance of the spring production for pelagic carbon fl ow. The contribution from spring to annual chlorophyll biomass ranged from 16-30 % whereas if carbon units were considered, the spring only contributed 8-23 % of production. Vertical structures further showed that PP peaked at shallower depths than chlorophyll in 85 % of the profi les examined. The vertical distribution of PP was tightly coupled to the surface water nutrient concentrations and showed that, when water column PP was highest, i.e. during summer months, a relatively high percentage of PP was taking place in the pycnocline/bottom layer. Zooplankton biomass and potential grazing was analysed for three of the six stations and the deep primary production (DPP) was suggested to be more important for copepod graz-ing than the spring production because copepod biomass peaked during summer. Protozooplankton grazing dominated during spring and fall. A simple model describing the seasonal distribution of PP in the BSTZ was developed in the search for a better proxy of carbon fl ow in dynamic ecosystems.

Seasonal variation in biological parameters in a temperate coastal area – decoupling of chlorophyll concentration and primary production

Maren Moltke Lyngsgaard1,2, Stiig Markager2, Katherine Richardson1 and Eva Friis Møller2

1Center for Macroecology, Evolution and Climate, Danish Natural History Museum, University of Copenhagen, DK-2200 Copenhagen, Denmark2Department of Bioscience, Marine Diversity and Experimental Ecology Aarhus University, DK-4000 Roskilde, Denmark


Seasonal patterns in primary production

1 Introduction

Modelling energy or element fl ow in pelagic marine ecosystems requires inclusion of a primary production (PP) function. Parameterizing this function is, however, a challenge as comprehensive data relating to primary production in a given area are seldom available. As chlo-rophyll concentrations are easier (and less expensive) to obtain than PP, chlorophyll concentration is often the proxy of choice for describing spatial and temporal distribution patterns in the phytoplankton activity (see e.g. Henson et al. 2010). This is not least the case when describing the impact of anthropogenic eutrophication via nutrient enrichment on the development of phyto-plankton biomass. Numerous studies (e.g. Conley et al. 2000; Timmermann et al. 2012; Carstensen et al. 2011 and references therein) assume a direct link between nu-trient input and biomass as represented by chlorophyll concentration. However, any relationship between nu-trient input and phytoplankton biomass would have to operate through PP.

Biomass or standing stock is, of course, infl uenced by a number of processes in addition to PP. Thus, at any given point in time, PP will be a function of gain (growth and advection) and loss (mortality both through natural and grazing processes and advection). Only growth can be ex-pected to be directly correlated to PP so there is no a priori reason to expect the relationship between biomass and PP to be constant and, even less so, that the relationship be-tween PP and chlorophyll should be constant.

The canonical view of phytoplankton physiology is that carbon to chlorophyll varies with light and nutri-ents, both vertically within the water column as well as seasonally (Cullen 1982, Geider 1987, MacIntyre et al. 2002, Fennel and Boss 2003, Palmer et al. 2013). Nev-ertheless, because it is so much easier to collect compre-hensive data sets that relate to chlorophyll concentra-tions via fl uorescence and other optical characteristics of the water column, standing stock of chlorophyll is still commonly used as a proxy for photosynthetic ac-tivity, especially when estimating global PP from satel-lite derived chlorophyll. As satellites only “see” down to the depth where 37 % of surface light is still avail-able (Joint and Groom 2000) an effort has been made to relate surface chlorophyll to PP (Eppley et al. 1985, Behrenfeld and Falkowski 1997b). Global estimates of PP based on satellite derived chlorophyll have shown a relatively poor fi t to in situ 14C measurements of prima-ry production (Campbell et al. 2002, Carr et al. 2006). These global estimates change, however, considerably when the carbon to chlorophyll ratio was considered and yielded higher estimates of PP in the tropical oceans and a higher seasonal variation in the middle and high

latitudes (Behrenfeld et al. 2005). This suggests that, not only is surface chlorophyll a poor predictor for the verti-cal distribution of chlorophyll but, maybe more impor-tantly, for the carbon biomass.

As the goal for estimating primary production, ulti-mately, is to examine the carbon fl ow in the pelagic sys-tems, the relationships between chlorophyll, carbon and PP clearly needs to be better elucidated. Examples of situations where chlorophyll has been decoupled from PP typically arise under high grazing pressure (Clo-ern 1982, Nielsen et al. 1993b, Pomeroy et al. 2006). Whether this decoupling is caused by the light adapta-tion/acclimatization within phytoplankton groups or single cells, respectively, or if the decoupling is due to intensive grazing has not been well addressed.

The aim of this study, therefore, was to describe the sea-sonal patterns in potential grazing, PP, chlorophyll and carbon biomass in order to be able to identify the appro-priate proxy(ies) for examination of carbon fl ow in a pe-lagic marine system. Suffi cient PP data were available to develop a parameterization of the seasonal variation in PP thanks to a comprehensive monitoring program in the Baltic Sea Transition Zone (BSTZ) established in the wake of legislation designed to curb anthropo-genic nutrient enrichment (Conley et al. 2002). While the parameterization developed is specifi c to the BSTZ, this region is in many ways typical of stratifi ed shallow coastal marine systems. Thus, the approach used and general patterns identifi ed are likely relevant elsewhere as well.

2 Materials and methods

2.1 Study areaSix locations were selected from the data base of the Danish National Aquatic Monitoring and Assessment Program (Conley et al. 2002). All study sites were lo-cated in the transition zone between the Skagerrak and the Baltic Sea, the BSTZ (Fig. 1). Only stations with a bottom depth > 10 m and continuous data were included in the analysis. Station depths vary from 14 m (Aalborg Bight), to 51 m in The Sound (Table 1, Fig. 1).

The BSTZ is a highly dynamic region with respect to hydrography and exhibits typical estuarine features. The area receives outfl owing water from the Baltic Sea that has a freshwater surplus of 559 km2 yr-1 (Savchuk 2005) and infl owing water from the Kattegat/Skagerrak driven by westerly winds. The circulation within the BSTZ is mainly driven by the water level difference between the Arkona Sea and the Northern Kattegat. The surface sa-



linity ranges from 10-14 in the southern part and reaches 20-25 in the Northern Kattegat. Bottom water salinity is about 32-34. (Gustafsson 2000). These conditions mean that the water column is almost permanently stratifi ed often exhibiting a density difference higher than 1 kg m-3 per m-1 (Δρ/Δz >1 kg m-4) that segregates the pycno-cline/bottom waters from the remaining water column. Mixing with surface water, especially in the Little and Great Belts (see Fig. 1), ventilates the bottom water (Bendtsen et al. 2009).

2.2 Data extracted from the National database

Data from the Danish National Aquatic Monitoring and Assessment Program, (MAPS) is publically available (http://www2.dmu.dk/1_viden/2_Miljoe-tilstand/3_vand/4_mads_ny/default_en.asp) and contains physical, chemical and biological data collected in BSTZ waters since the 1980s. Different regional authorities were re-sponsible for data collection and sampling was carried out on different ships using different sampling equip-ment. However, all procedures were standardised in technical guidelines (Kaas and Markager 1998) and the former Danish National Environmental Research Insti-tute (NERI), now Aarhus University, carries out work-shops and inter-comparisons of the data collected. The

parameters in this study were under supervision of one of the authors (SM).

The following parameters were directly extracted from the database; dissolved inorganic nitrogen and phos-phorous from surface (0-5 m) and pycnocline/ bottom waters, phytoplankton biomass, sea surface temperature (0-5 m), zooplankton biomass and size whereas param-eters such as PP, depth-integrated chl. a, zooplankton grazing and depth of the surface mixed layer was calcu-lated specifi cally for this study from parameters in the database.

The daily surface photosynthetic active radiation (SPAR) was calculated as 30 minute mean values from continuous measurements at several different localities in Denmark within approximately 30-150 km of the sample locations.

The analyses of chlorophyll a, inorganic nutrient con-centrations and photosynthetic carbon uptake were all carried out on water from the same sample. Photosyn-thetic carbon uptake was measured on samples from two different depths in the water column: 1) an integrated “surface” sample from 0 to 10 meters (unless visual ex-amination of the density profi le indicated a major densi-ty difference or pycnocline within this depth range, then only to this depth) collected using a hose/tube inserted to the desired depth or with Niskin bottles at 1, 5 and 10 meters depth and 2) the depth of the deep chlorophyll maximum, defi ned as the depth with the highest chloro-phyll fl uorescence and where fl uorescence was greater than twice the average for the surface layer (sample collected using Niskin bottle). The DCM sampled was by defi nition found below the surface layer in 67 % of the profi les. When a deep chlorophyll maximum was absent, the deep sample was taken below the surface mixed layer at the depth of 2 % surface light penetra-tion. Chlorophyll a concentrations were also determined on these two samples. In addition to the two depths described above, the chlorophyll a and inorganic nutri-ent concentrations were analysed from standard depths (every fi ve meters) down through the water column starting at 1 meter below surface and ending 1 meter above the sediment.

2.3 P-E parametersPhotosynthetic carbon assimilation was measured based on a carbon-14 method modifi ed (Markager et al. 1999b) after Steemann Nielsen (1952). Photosynthesis versus irradiance (P-E curves) P-E curves were calculated from incubations made under artifi cial light (Osram HQI-T or high pressure halogen lamps), where the samples were incubated at seven different light intensities for two hours with metal grids providing approx. 35 % light at-

Figure 1. The Baltic Sea transition zone (BSTZ) and the location of the six locations



tenuation between each bottle. Thereafter, the samples were fi ltered (Whatman GF/F or GF 75 Advantec fi lters) and the carbon incorporation stopped with acid (200 μl 0.1 N HCl). The amount of incorporated carbon-14 in the phytoplankton was determined by liquid scintilla-tion counting.

The diffuse light attenuation coeffi cient (Kd) was deter-mined by relating in situ light from the CTD, corrected for surface irradiance measured simultaneously at the top of the ship, to depth.

A Mixed Models test in SAS.9.3 was used, without the Random statement, to examine differences between P-E parameters from surface versus DCM waters.

2.4 Chlorophyll aChlorophyll a fl uorescence (F) through the water col-umn was profi led continuously using a CTD-mounted fl uorometer. Fluorescence per chlorophyll a changed systematically with depth in the dataset. Therefore, a fl uorescence factor (Fchl = F/[Chl]) was calculated whereby the chlorophyll concentration recorded in the discrete sample and the fl uorometer measurement from the corresponding depth (F) were related. Values for Fchl for depths between the samples were calculated by linear interpolation between the depths where Fchl had been determined (every 5 meters). This was done with 0.2 m resolution and then Fchl(z) were used to estimate a continuous chlorophyll a profi le (Chl(z) = F(z)/Fchl(z)).

The average monthly depth- integrated chl. a concentra-tions shown in Fig. 3 were calculated based on the years from 1998 to 2012. Every chl. a profi le was depth-inte-grated before monthly average values were calculated.

The chlorophyll a concentration measurements were carried out by fi ltering samples onto Whatman GF/F or GF 75 Advantec fi lters. Filters were extracted in etha-nol (96 %) for 6-20 hours and the extracts were analysed spectrophotometrically according to the method de-scribed by Strickland and Parsons (1972) and modifi ed by Danish Standards (1986).

2.5 Estimating primary production through the water column

To estimate water column PP, chlorophyll-specifi c pho-tosynthesis rates were calculated from the P-E curves and chlorophyll a concentrations. A light matrix repre-senting the light intensities at 0.2 m intervals through the water column at hourly intervals over each day was constructed using the attenuation coeffi cient and the surface light. This light matrix and the photosynthetic characteristics derived from the P-E curves, i.e. alpha,

Pmax and intercept for the two sampling depths normal-ised to chlorophyll concentrations (alphaB, Pmax

B and interceptB), were combined to estimate daily PP at 0.2 meter intervals throughout the water column (see Fig. 6 for an example) for all stations except the Great Belt 1 station, where only P-E curve parameters from the sur-face layer were available.

For the other 5 stations, the chlorophyll specifi c photo-synthetic characteristics obtained from the surface layer sample were applied down to the starting depth of the pycnocline/bottom layer (PBL). From this depth and un-til the depth where the second sample for photosynthe-sis determination had been made, the applied photosyn-thetic parameters were derived by linear interpolation. The photosynthetic parameters derived from the second sampling depth were then applied from the depth of that sampling to the depth where the estimated daily depth-specifi c photosynthetic rate was greater than zero. In this manner, estimates of total water column PP as well as the contribution to the total production from the PP in the PBL (deep PP or DPP) were made for 1385 days of sampling based on the database resulting from the monitoring program. Data availability varied between stations and for each of the 6 stations it ranged from 3 to 14 years. The sampling frequency at each station ranged from 12 to 49 measurements per year.

The average annual PP listed in Table 2 was calculated as follows; fi rst, an average monthly rate was calculated based on all sampling years and then these rates were multiplied with the number of days in the given month. Secondly, these absolute monthly values were added to-gether resulting in an annual carbon assimilated per m2. The spring PP is, in this study, noted as the PP occurring in the months; February and March due to the systemati-cally increased production during these months. The PP profi les from Aarhus Bight (shown in Fig. 6) are aver-age monthly values based on the years from 1999-2012 calculated for every depth (0.2 m interval) from 0-16.6 m. The standard variation is calculated as the inter-an-nual variation.

2.6 Grazing Mesozooplankton samples were obtained using a sub-mersible pump equipped with a 60 μm net lifted through the water at 10 m min-1 (Møhlenberg, 1987). Samples were preserved in 2–3% buffered formalin. Copep-ods and copepodites were identifi ed to species or ge-nus level, and biomass was calculated on the basis of length-carbon regressions from the literature or fi xed carbon values for stages within each genus (Breteler et al. 1982, Berggreen et al. 1988, Hay et al. 1991, Sa-batini and Kiørboe 1994, Satapoomin 1999). Volumes were then calculated assuming 0.12 pg C μm-3 (Hansen



et al 1997). Samples for the determination of ciliate and heterotrophic dinofl agellate biomass were taken as an integrated sample from 0-10 m, fi xed in acid Lugol’s so-lution (3% fi nal concentration), and 50 or 100 mL were examined after 24 h sedimentation using inverted mi-croscopy. Cell volumes were estimated assuming simple geometric shapes and converted to biomass using the general carbon to volume relationship for protist plank-ton (Menden-Deuer and Lessard 2000).

The potential grazing of protozooplankton (ciliates, heterotrophic dinofl agellates), and calanoid copep-ods on phytoplankton was calculated from the general equations on allometric scaling of maximum specifi c ingestion and maximum clearance rates by Hansen et al. (1997). Cyclopoid and harpacticoid copepods were not included since the genera present at the station are considered to graze mostly on non-phytoplankton prey (Atkinson 1995, Koski and Kiørboe 2005) and the bio-mass of these groups were generally less than 10% of the calanoid biomass. The rates were normalized to in situ temperature by applying a Q10 – value of 2.8 (Hansen et al. 1997). The ingestion rates (= grazing rates) were obtained either directly from maximum in-gestion or calculated from the specifi c clearance rates and the concentration of potential food i.e. autotrophic carbon. This was calculated from chlorophyll by using an average monthly carbon to chlorophyll ratio for each station (values from Jacobsen and Markager in prep). The smaller of the two grazing estimates were used for further calculations, supposing that grazing varied with food concentration but could never exceed maximum ingestion. Station specifi c grazing was fi nally calculated by integrating the top 25 m of the water.

2.7 Water column stratifi cation and deep primary production

Salinity and temperature profi les were used to calculate density (ρ) (UNESCO 1981). 1385 density profi les (from year 1998 to 2012) were used to calculate the starting depth of the pycnocline/bottom layer (PBL) which was defi ned as starting at the fi rst density difference of more than 1 kg m-4. For every day with a distinct pycnocline, each depth in the water column was assigned as being in one of two layers, i.e. in the surface or PBL layer. This gave the possibility of considering the PP occurring in each layer; the surface primary production (SPP) i.e. all the PP taking place above the depth of the density cri-teria (the pycnocline) and the deep primary production (DPP) being the PP occurring below this depth i.e. in the PBL layer. When the density criteria of Δρ/Δz > 1 kg m-4 was absent from the water column, the PP was considered only as SPP.

The value of 1 kg m-4 as density criterion was checked by visual inspections of 256 density and chlorophyll fl uorescence profi les. With a criterion of 1 kg m-4, the starting depth of the PBL found by eye matched the depths found by the criteria in 90 % of the cases. For the remaining 10 % of profi les, the density criterion of 1 kg m-4 overestimated the depth of the beginning of the pycnocline i.e. the depth identifi ed by the criterion was deeper than the depth estimated by eye (1-2 m). Thus, our estimates of the absolute and relative values of DPP are conservative.

2.8 NutrientsValues of dissolved inorganic nitrogen (DIN) and phos-phorus (DIP) were extracted from the national database for the time period from 1998 to 2012. Inorganic nu-trients were measured with fl ow injection on a Scalar auto-analyser (Valderrama 1981, Grasshoff et al. 1999, Hansen and Koroleff 1999). The average water concen-trations were calculated as monthly values in the surface from 1 m depth samples and as an average of measured concentrations from waters deeper than 10 m.

2.9 PhytoplanktonSampling of phytoplankton was carried out as an inte-grated measure of the water column down to the depth where 2 % of the surface light remains. Species identi-fi cation was carried out following the method described by Ütermöhl (1958). Carbon biomass (μg C l-1) was cal-culated after Strathmann (1967) and the carbon factors used were 0.13 for thecate dinofl agellates, 0.11 pg C μm-3 for athecate dinofl agellates and diatoms. For dia-toms, the plasma volume was calculated as: bio volume - (0.9 · vacuole volume). It was assumed that the plasma volume was 1 μm thick and that the vacuole volume contributed 10 % to the carbon biomass. The carbon to chlorophyll values were calculated on measurements carried out on the same water samples. These values were then average into monthly average values for each station based on a time period from 1998 to 2012.

3 Results

3.1 Seasonal variation in a stratifi ed coastal ecosystem

The seasonal variation in PP showed a pattern similar to the seasonal variation in surface irradiance and tempera-ture but with a time lag in peak values (see Fig. 2) and a deviation in spring months. The seasonal variation in PP can be described as one half of a sinus curve with a modifi cation in the spring as follows:



Equation 1:

PP = PPwinter + PPamplitude (cos((month – 0.5 – top)/12 × 2 × π – π) + 1)0.5) + PPFeb + PPMar

If month ≠ 2 then PPFeb = 0, if month ≠ 3 then PPMar = 0

The PPwinter is the level of winter PP and the PPamplitude is the annual amplitude of PP. Monthly numbers are from 1 (January) to 12 (December) and the middle of the year is 6.5, i.e., the 1st of July. The value of top adjusts the timing of the maximum in production, e.g. will top = 1 move the peak productivity to the 1st of August (6.5 + 1 = 7.5), and the parameter top should be read as the deviation in months from midyear. The parameters, PPFeb and PPMar, are added to account for spring bloom production. All parameters are fi tted to average monthly values for depth-integrated PP for each station by the procedure Proc NLIN in SAS 9.3. The model fi t to the actual data is shown in Fig. 2 with r2-values ranging from 0.90 (Aalborg Bight) to 0.99 (Great Belt 2). Residuals show positive values in April indicating that, in some years, the spring bloom con-tinues into this month and positive residuals in August and September indicate the presence of a weak bloom in late summer/autumn.

The contribution from spring PP to annual PP ranged from 6 to 14 % when it was considered as the deviation from the sinus curve (sum of PPFeb and PPMar ) and a little higher when it was considered as the entire pro-duction occurring in the months of February and March (10-18 %, see Table 1). The DPP contributed on average 17 % to the total annual PP (results from Lyngsgaard et al. submitted, see Table 2). The DPP was low or, in some cases, non-existent from October to January (see an ex-ample from Aarhus Bight in Fig. 6), when the phyto-plankton community was frequently mixed throughout the water column and surface irradiance too low to sus-tain a DPP. However, from April to August (inclusive), when the frequency of stratifi cation was at its highest, the vertical distribution of primary production showed a consistent pattern of photosynthesis taking place deeper down in the water column.

The seasonal variation of PP, phytoplankton carbon bio-mass, chlorophyll a and grazing from copepods and pro-tozooplankton in the Baltic Sea transition zone showed characteristic patterns of a productive summer season where the organic material produced was quickly trans-ported to the next trophic level in the food web (see Figs. 3 and 4). PP ranged from 155-226 g C m-2 yr-1 (see Table 1) with peaks during the months of July-August (aver-age rate of 27.9 g C m-2 month-1 ± 6.15 g C m-2 month-1

J F M A M J J A S O N DJ F M A M J J A S O N D J F M A M J J A S O N DMonth

PP

(mg

C m

-2 d

ay-1)

-500

0

500

1000

1500

2000

2500

-500

0

500

1000

1500

2000

2500

a. Aalborg Bight

e. The Sound

c. Little Belt

f. Great Belt 2d. Great Belt 1

b. Aarhus BightPP dataPP modelR = 0.90 R = 0.97

R = 0.97

R = 0.96

R = 0.98 R = 0.99

Figure 2. Average monthly primary production from 1998-2012 together with modelled values (see equation 1) shown for each month of the year. The model fi t to data is shown in R values in the upper right corner. Error bars indicate the inter-annual varia-tion of each monthly value. No standard deviation is shown for the modelled values.


between locations). The depth-integrated chlorophyll concentrations did not track PP during summer months. A general pattern of decoupling between PP and chlo-rophyll a conc. occurred from May to September (see Fig. 3). However, for Aalborg Bight and The Sound, the decoupling started a little later and occurred from June to September and from July to September, respectively.

The chlorophyll a conc. displayed an overall peak in March with a mean value of 3.96 g Chl m-2 (± 1.23). Stable but lower values occurred during summer (mean

value = 1.86 ± 0.47 g Chl m-2) for fi ve out of the six locations. The Sound deviated slightly from this pattern showing the highest chlorophyll a concentration in May (2.6 g Chl m-2). An increase in chlorophyll was noted in October and November (mean = 2.49 ± 0.83 g Chl m-2).

The autotrophic phytoplankton carbon biomass (esti-mated from chlorophyll and the carbon to chlorophyll ratio) and chlorophyll were both signifi cantly and posi-tively correlated to PP using linear regression when values from the Sound, Aarhus and Aalborg Bight were


Table 1. Annual average primary production, annual spring, and deep primary production and grazing is presented as g C m-2 yr1. The percentage of total annual primary production is noted in brackets. Standard deviation in the top row shows the inter-annual variation of annual primary production. The depths of the stations, the average starting depth of the pycnocline/bottom layer (PBL) are given in meters and the frequency of stratifi cation (i.e. when PBL was defi ned) in percentage.

Aalborg Bight Aarhus Bight The Sound Little Belt GB I GB II

Annual PP 161 ± 25 155 ± 37 175 ± 53 226 ± 18 213 ± 12 209 ± 34

Annual spring PP: Feb-Mar 26 (16) 28 (18) 18 (10) 32 (14) 28 (13) 31 (15)

Annual DPP 48 (30) 33 (21) 40 (23) 14 (6) 34 (16) 27 (13)

Grazing (impact) 160 (99) 163 (105) 157 (90)

Depth of station 14.0 16.6 51.0 19.5 21.9 35.0

Av. PBL starting depth 7.5 ± 0.9 9.8 ± 0.9 11.4 ± 0.9 11.1 ± 1.2 9.9 ± 1.4 14.3 ± 3.0

Frequency of PBL 54 ± 26 61 ± 28 90 ± 9 59 ± 29 71 ± 23 85 ± 15

Month

Phy

topl

ankt

on c

arbo

n bi

omas

s (g

C m

-2)

Prim

ary

prod

uctio

n (g

C m

-2 m

onth

-1)

Chlorophyll (m

g Chl m

-2)

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep Oct

Nov

Dec Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep Oct

Nov

Dec Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep Oct

Nov

Dec

0

10

20

30

40

0

10

20

30

40

0

50

100

150

200

0

50

100

150

200

c. Aarhus Bight

d. Little belt e. Great Belt I f. Great belt II

a. Aalborg Bight b. The SoundPPChl

C

Figure 3. Average monthly values of total water column primary production, chlorophyll a and phytoplankton carbon for six loca-tions in the Baltic Sea transition zone (phytoplankton carbon only on three stations).



considered together. There was a stronger relationship between carbon biomass and PP (R2 = 0.43, p < 0.001, n = 36) than between chlorophyll and PP (R2 = 0.14, p < 0.05, n = 36). When PP and chlorophyll were considered for each station separately, only 1 out of the 6 stations showed a signifi cant relationship between chlorophyll and primary production (Table 3) whereas all three were

signifi cant when it came to the relationship between car-bon biomass and primary production (Table 2). Phyto-plankton carbon values were only available for 3 of the 6 stations.

The relationship between chlorophyll and PP was far from constant when looking at the vertical distribu-

CopepodsProtozooplankton

Car

bon

biom

ass

(µg

C l-1

)

Gra

zing

(µg

C l-1

day

-1)

Gra

zing

impa

ct (%

)

Month Month Month

Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec

0

10

20

30

40

50

0

10

20

30

40

50

0

10

20

30

40

50

0

20

40

60

80

0

20

40

60

80

0

20

40

60

80

The

Soun

dA

albo

rg B

ight

Aar

hus

Big

ht

a d g

b e g

c f i

0

50

100

150

2000

50

100

150

200

0

50

100

150

200

Figure 4. Zooplankton biomass (left column: a-c), potential grazing (center column: d-f) and total grazing impact on primary production (right column: g-i) for copepods and protozooplankton in Aalborg Bight, The Sound and Aarhus Bight.

Table 2. Pearson correlations between the average monthly values of primary production (n = 12 for each station) and the follow-ing variables; depth-integrated chl. a, autotrophic phytoplankton carbon (derived from the chl. a and C:Chl ratio) and grazing. R values indicate a signifi cant relationship at the p < 0.05 level and a star indicates a p < 0.001 level.

Primary production

Aalborg Bight The Sound Aarhus Bight Little Belt Great Belt 1 Great Belt 2

Chlorophyll R = 0.04 R = 0.72 R = 0.26 R = 0.09 R = 0.26 R = 0.40

Carbon R = 0.60 R = 0.75 R = 0.80

Grazing R = 0.80 R = 0.83* R = 0.85*



tion patterns on individual profi les. The PP maximum was situated above the chlorophyll maximum. A linear regression between the maximum depth of PP and the maximum depth of chlorophyll showed that they were related but that the depth of the maximum PP was 1/3 (parameter estimate = 0.32) of the depth of the chloro-phyll maximum (r = 0.14, p < 0.0001, n = 309). In 46 out of the 309 profi les inspected the PP and chlorophyll maxima were found at the same depth. Thus, for the re-maining 85 % of the profi les the chlorophyll maximum was located deeper than the PP maximum. The profi les analysed were from stations in Aarhus Bight (years: 1999-2003, 2005-2008 and 2010) and Aalborg Bight (years: 2003-2008).

The total potential grazing from copepods and proto-zooplankton peaked from June to August (see Fig. 4). This was concurrent with observations of high PP rates and low chlorophyll concentrations. However, po-tential grazing from copepods and protozooplankton peaked during different seasons. The copepod biomass was highest in May-June and copepod potential grazing highest in June-July. The protozooplankton biomass decreased during summer resulting in two distinctive peaks observed in spring (March-April) and during fall (September-October). The protozooplankton graz-ing showed a constantly high impact from March to October. The onset of potential grazing impact by pro-tozooplankton and copepods occurred in March-April and it decreased to winter levels again in October-No-vember. In carbon units, the annual potential grazing by copepod and protozooplankton combined was es-timated to approx. 160 g C m-2 yr-1 (see Table 1). The average annual grazing impact on the average annual PP ranged from 90-105 % (see Table 1) where graz-ing estimates suggested consumption to be greater than the carbon produced by PP during the months April-October (see Fig. 4g, h, i). The potential grazing from protozooplankton was generally more important than copepod grazing in Aarhus Bight (69 % of total graz-ing), whereas the copepod grazing was most important in Aalborg Bight and The Sound, with 65 % and 56 % of total grazing respectively.

A Pearson correlation carried out for average monthly PP and total potential grazing for each station showed that grazing was signifi cantly and positively correlated to PP in Aalborg and Aarhus Bights (Table 2). In gen-eral, high values of grazing were found where there were high values of PP. The copepod biomass was sig-nifi cantly linearly correlated to DPP on all three stations (Aarhus Bight: r2 = 0.74, p < 0.001, n = 12; Aalborg Bight: r2 = 0.55, p < 0.01, n = 12; The Sound: r2 = 0.78, p < 0.001, n = 12) indicating that while the total biomass of zooplankton was related to PP the copepod biomass, was related to DPP.

3.2 Primary production, light, stratifi cation and nutrients

The average monthly surface photosynthetically active radiation (SPAR) peaked in June at 39.8 mol photons m-2 day-1. The peak in average water temperature (0-5 m) occurred later (August) with a value of 17.9 ± 1.9 oC (see Fig. 5a). SPAR increased markedly from February (6.8 mol photons m-2 day-1) to March (15.9 mol photons m-2 day-1) during the period of the onset of spring pro-duction.

Stratifi cation is a consistent feature in the Baltic Sea tran-sition zone. Frequencies of sampling days with a strati-fi ed water column (Δρ/Δz > 1 kg m-4) ranged from 54 % in Aalborg Bight to 90 % in the Sound (see Table 2). The Sound and Great Belt II stations showed the highest annu-al frequency of stratifi ed waters. This was because these locations exhibited frequencies above 63 % throughout

Irrad

ianc

e, S

PA

R(m

ole

phot

ons

m-2 d

ay-1)

DIN

(µm

ol l-1

)D

IP (µ

mol

l-1)

Temperature (°C

)

Month

Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec

a

b

c DIP > 10 mDIP 1 m

DIN > 10 mDIN 1 m

0

2

4

6

8

0

0.2

0.4

0.6

0.8

1.0

1.2

10

120

10

20

30

40

50

0

5

10

15

20

25SPARTemp.

Figure 5. Average monthly surface irradiance (SPAR) and tem-perature (average of 0-5 m) for the six locations in the Baltic Sea transition zone (a). Dissolved inorganic nitrogen (b) and phosphorous (c) are shown as average monthly values for sur-face waters (1 m) and waters deeper than 10 m.



the year whereas the other locations showed lower fre-quencies during winter months e.g. ranging down to 6 % in December at location Little Belt. The average depth of the beginning of the pycnocline ranged from 7.5-14.3 m with a variation between average monthly values of up to 3 meters. This variation expresses the difference between summer months (April–August) where there was a more stable stratifi cation (frequencies consistently above 75 %) and a shallower depth of the start of the PBL depth. In the remaining months, the stratifi cation was unstable and deeper (see Appendix). The highest seasonal variations in stratifi cation were found in the Belts.

Dissolved inorganic nitrogen (DIN) and phosphorous (DIP) concentrations in the water column showed a simi-lar seasonal development as the PP only with opposite operational signs. DIN and DIP showed high values in January and February when there was low PP and started declining in February when spring production was setting in (Fig. 5b and c). Average monthly surface DIN conc. was < 1 μmol l-1 from May–September and surface DIP

conc. was < 0.1 μmol l-1 from May-August. The DIP con-centration increased again already in August and the DIN two months later in October to winter values of 0.6 μmol l-1 and 6.3 μmol l-1, respectively, in December. The aver-age N:P ratio (DIN:DIP) for the total water column de-creased from an average value of approx. 14 at the begin-ning of the year (January to March) to a value of approx. 3 in September. Thereafter, it again increased to a value of 10 in December (data not shown).

February and March, the months of spring production, show a clear pattern where the PP took place mainly in surface waters (Fig. 6). Already in April, the surface production decreased and the months of May – August exhibited a pattern of increased production in deeper waters. DIN and DIP concentrations in surface waters decreased concurrently with the downward shift in PP. In September and October, the SPP increased again and, from November to January of the next year, the PP de-creased at the same time as SPAR values fell below 5 mol photons m-2 day-1.

Primary production (mg C m-3 day-1)

Dep

th (m

)

0 50 100 150 0 50 100 150 0 50 100 150 0 50 100 150 0 50 100 150 0 50 100 150

0

2

4

6

8

10

12

14

16

0

2

4

6

8

10

12

14

16

Jan Feb Mar Apr May Jun

Jul Aug Sep Oct Nov Dec

30% 35% 54% 81% 95% 94%

83% 88% 71% 58% 26% 20%

Figure 6. Average monthly primary production profi les (black line) at 0.2 m resolution for each month of the year calculated from Aarhus Bight data covering the years 1999-2010. The grey horizontal bars indicate the inter-annual variation of monthly val-ues. The frequency of days with a pycnocline (relative to total sampling days) is shown in percentage for each month. The highest stratifi cation frequencies are noted with bold numbers.



The shift in vertical distribution from surface to DPP was observed to take place concurrently with decreases DIN and DIP concentrations in surface waters (see Fig. 5b and 5c). A signifi cant and positive correlation between the fraction of DPP (relative to PP) and the surface concen-tration of DIN (Pearson correlation: r = 0.66, p < 0.05, n = 12) and DIP (r = 0.80, p < 0.01, n = 12) was found. High fractions of DPP were only found where DIN and DIP surface water concentrations were low (see Fig. 7).

3.3 Phytoplankton in surface watersThe relative distributions of the larger phytoplankton groups in the water column (to the depth of 2 % of sur-face light); diatoms, dinofl agellates, cyanophytes and the remaining groups combined (designated as ‘Oth-er’) in the Sound, Alborg and Aarhus Bights exhibited

similar patterns (see Fig. 8). Diatoms dominated in the spring bloom (February and March) when most of the PP occurred in surface waters (see Fig. 6 for Aarhus Bight) and a peak in relative diatom biomass was again observed in October-November (see Fig. 8) when PP again mainly occurred in surface waters. Dinofl agel-lates were most frequent and sometimes the dominat-ing group in late summer and autumn when the relative and absolute magnitude of DPP increased (see Fig. 6 for Aarhus Bight). The group ‘Other’ (mainly Prym-nesiophytes, Prasinophytes and the Raphidophyte spe-cies Pseudochatonella) exhibited a relatively constant contribution to the carbon biomass throughout the year, however, with lowest contributions during summer and fall. Cyanophytes were virtually absent in Aalborg and Aarhus Bights. Only in the Sound were they present from May to August with a maximum contribution of

Frac

tion

of s

urfa

ce p

rimar

ypr

oduc

tion

(%)

Frac

tion

of s

urfa

ce p

rimar

ypr

oduc

tion

(%)

DIN (µmole l-1) DIP (µmol l-1)0 2 4 6 8 10 12

Apr - OctoberJan - Mar + Nov - Dec

0 0.2 0.4 0.6 0.850

60

70

80

90

100

110

50

60

70

80

90

100

110

Figure 7. The fraction of surface primary production (SPP) plotted as a function of dissolved inorganic nitrogen (DIN) (upper graph) and phosphorous (DIP) (lower graph) concentration in surface waters (1 m). Black dots represent summer months from April-October and white circles winter months from January-March and November -December. The fraction of SPP is signifi cant-ly and positively correlated to DIN (r = 0.66, p < 0.05, n = 12) and DIP (r = 0.80, < 0.01, n = 12). The encircled winter values are all from the Sound that shows the highest winter stratifi cation of all locations analysed.

Diatoms Dinoflagellates Other Cyanophytes

Phy

topl

ankt

on g

roup

frac

tion

(%)

Month

Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec

0

20

40

60

80

100The SoundAalborg Bight Aarhus Bight

Figure 8. Average monthly percentage distribution of the following phytoplankton groups; Diatoms, Dinofl agellates, Cyanophytes and ‘Other’ (including the remaining phytoplankton groups), for the euphotic zone at three locations in the BSTZ.



~30% (August). The Sound is the station closest to the Baltic Sea where blooms of Cyanophytes are common during late summer (see e.g. Kahru et al. 2007).

On an annual basis, most of the phytoplankton carbon came from diatoms (Fig. 8). The overall contribution of diatom carbon biomass to the annual phytoplankton biomass was highest for the Aalborg Bight (65 %) and lowest for the Sound (37 %). In Aarhus Bight, approxi-mately 50 % of the carbon biomass came from diatoms. The annual carbon contribution from the group ‘other’ was highest in the Sound (28 %) and lowest in Aalborg Bight (14 %). Aarhus Bight had 23 % of the carbon bio-mass coming from the group ‘other’. Dinofl agellates contributed 19 – 27 % to the annual carbon biomass. Aarhus Bight showed the highest contribution and the Sound the lowest. Aalborg Bight had 20 % of the carbon biomass coming from dinofl agellates.

3.4 P-E parameters The general patterns in maximum chlorophyll-specifi c photosynthesis rate (Pmax

B) were analysed for each loca-tion separately and they all showed higher Pmax

B values in samples from the surface waters than from the PBL

(mixed models test, Aalborg Bight: t = 7.55, df = 2388, p < 0.0001; Aarhus Bight: t = 5.61, df = 2388, p < 0.0001; Little belt: t = 5.52, df = 2388, p < 0.0001; Great belt II: t = 8.61, df = 2388, p < 0.0001) except for the Sound (t = 1.54, df = 2388, p = 0.12). The Great belt I was excluded from this analysis, as there were only P-E parameters from the surface waters.

The Pmax B values showed a seasonal pattern where Pmax

B increased from low values in January - March to maxi-mum values in July-August (see Fig. 9) which was coinci-dent with maximum temperatures (see Fig. 5a). The Pmax

B values from surface and PBL layers were not signifi cantly different for the months of December-March on all of the locations except for Aalborg Bight where November was an exception (Fig. 9). During these months, temperature and insolation were low and the water column more often well mixed than for the summer months (see Fig. 5a for irradiance and appendix for stratifi cation).

The initial slope of the P-E curve normalized to chloro-phyll, the alphaB, resembled the seasonal pattern observed for Pmax

B with highest values in summer months (May–September) and lowest values during winter months (November-March the following year) (data shown in

Pm

axB (g

C g

-1 C

hl h

our-1

)

Month

Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec Jan

Feb

Mar

Apr

May Jun

Jul

Aug

Sep Oct

Nov

Dec

0

2

4

6

8

12

10

0

2

4

6

8

12

10

The SoundAalborg Bight Aarhus Bight

Great Belt 2Little Belt Great Belt 1

Surface (1 m)

PBL

Figure 9. Average monthly values of Pmax from the surface mixed layer (1 m) and below are shown for six different locations in the Baltic Sea transition zone (1998-2012). The grey standard deviation bars indicate inter-annual variation for each monthly value. The Pmax values are normalized to chlorophyll a concentration.



appendix). No signifi cant differences between alphaB val-ues from surface and PBL waters were found. The inter-annual variation was high for both depths. The variation between years was markedly higher (ranging from ± 2.1-15.1 for surface samples and ± 2.3-55.8 for PBL samples) than the variation between locations (±1.2–5.8 for surface samples and ± 0.8 – 9.7 for PBL samples).

Not surprisingly, the seasonal pattern of the light in-tensity at which photosynthesis initially is saturated, Ik, (Pmax

/alpha, μmol photons m-2 s-1) resembles that of Pmax

B (data not shown). The Ik values from the surface waters (1-10 m) were signifi cantly higher than the Ik values for the waters in the PBL (t-test, p < 0.0001) for all locations except for the Little Belt (t-test, p = 0.08). This difference in Ik, measured on phytoplankton from the surface waters compared to the PBL was driven by the summer months where the water column was strati-fi ed. In contrast, the winter months when mixing was greatest showed little difference between Ik values from the two depths.

4 Discussion

4.1 Seasonal decoupling of chlorophyll and primary production

The main result of this study is the demonstration of how rates and carbon biomasses are strongly coupled for phy-toplankton, zooplankton, PP and potential grazing. In contrast, phytoplankton biomass described as chlorophyll concentrations are largely uncoupled from these parame-ters as chlorophyll demonstrates a distinctive spring peak not seen in the other parameters. A consequence of this is that the importance of particularly the spring bloom, but also the autumn bloom, as observed from chlorophyll concentrations, for ecosystem dynamics may be over-estimated (Fig. 3). As a percentage of the total annual average biomass at stations; The Sound and Aarhus and Aalborg Bight, the spring bloom only contributed 8-23 % (average = 16 %) when it is considered in carbon units, whereas it contributed 16-30 % (average = 25 %) when it was considered in chlorophyll units. The spring chlo-rophyll contribution to the annual chlorophyll was twice that of the carbon contribution in The Sound. This was most likely due to the relatively low contribution from diatoms in February and March at this station compared to the other two stations (Fig. 8). Thus, when biomass is considered in chlorophyll units, it may overestimate the magnitude of the spring (and fall) biomass.

The strong seasonal pattern observed in PmaxB and Ik

was refl ected in the carbon to chlorophyll ratios used in this study which showed concomitant peaks in July and August (see Jacobsen and Markager, in prep, for

data). This seasonal pattern in the monthly PmaxB and

alphaB values will most likely change if the P-E param-eters are normalized to carbon rather than chlorophyll resulting in lower Pmax

B and alphaB during summer and higher parameter values during spring and fall. Unfor-tunately, we did not have specifi c measurements of car-bon for the samples from which we measured the P-E parameters and we could, therefore, not pursue this hypothesis. The ratio of mixotrophy (i.e. phytoplank-ton capable of using other than autotrophic sources of carbon) to autotrophy must, however, be addressed if these P-E parameters are to be normalized to carbon. Doing so is not straight forward as mixotrophy most likely is much more common among phytoplankton than we have previously assumed and the distribution and importance of mixotrophy still is being quantifi ed (see review by Jeong et al. 2010).

The pigment composition varies with the different phy-toplankton groups where e.g. dinofl agellates in gener-al have a higher carbon to chlorophyll ratio than e.g. diatoms (Geider 1987, Tang 1996). Therefore, a sea-sonal decoupling of carbon and chlorophyll can occur because 1) the dynamic pigment composition within phytoplankton cells is highly affected by nutrients and irradiance (Eriksen and Iversen 1995, Henriksen et al. 2002, Domingues et al. 2011) and 2) because the species composition of the phytoplankton community changes through the different seasons of the year in temperate waters (see e.g. Holligan and Harbour 1977, Domingues et al. 2005). Thus, the higher contribution from the dia-tom group to the total phytoplankton carbon biomass during spring can explain the relatively low carbon to chlorophyll ratio found during these months. Overall, this indicates that the species composition of the phyto-plankton community is relevant to the choice of proxy to examine carbon fl ow.

4.2 Vertical decoupling of chlorophyll and primary production

As light varies over seasons, it also varies vertically in the water column. These changing light environments also generates a decoupling between carbon and chlo-rophyll vertically within the water column (Steele and Yentsch 1960, Steele 1964, Lerman et al. 1974, Bien-fang and Harrison 1984, Fennel and Boss 2003). The deep chlorophyll maximum (DCM) was, in this study, located deeper than the PP maximum in 85 % of the PP and chlorophyll profi les. One obvious reason for this vertical decoupling was, however, that the same amount of phytoplankton can assimilate much more carbon when they are situated further uo in the water column where light intensities are higher. Hence, a part of the explanation lies simply in the photosynthetic response to light.



Another reason for the decoupling could, however, be that the carbon to chlorophyll ratio is known to vary with depth as it does with seasons given the inverse re-lationship between chlorophyll per phytoplankton cell and irradiance (Steele 1964, Kiefer and Kremer 1981, Geider et al. 1997). In addition, the species composi-tion of the phytoplankton community in subsurface wa-ters may also alter the carbon to chlorophyll ratio there due to other strategies than light acclimatization. Ear-lier work in Aarhus Bight showed that the phytoplank-ton community composition varied signifi cantly even within a very narrow depth range (Mouritsen and Rich-ardson 2003), and that increased numbers of dinofl agel-lates were found in subsurface waters (Lyngsgaard et al. in prep, see manuscript II, Holligan and Harbour 1977, Mouritsen and Richardson 2003). As many dinofl agel-lates are known to be mixotrophic (see Jeong et al. 2010 and references therein), their food strategy makes them less dependent on chlorophyll, which may explain why this group in general has a higher carbon to chlorophyll ratio than other phytoplankton groups.

Furthermore, grazing can be selective where certain groups or sizes of phytoplankton are preferred over oth-ers (Atkinson 1995, Meyer-Harms et al. 1999). Thus, potentially preferred prey may be eaten fi rst while the remaining may then accumulate deeper down. This may, ultimately, infl uence the carbon to chlorophyll ratio. Zooplankton biomass has been shown to concentrate in subsurface waters in the BSTZ during summer (Nielsen et al. 1993a, Nielsen and Andersen 2002) which further supports the argument that the impact from grazing also may affect the vertical structure of biomass.

Standing stock, in itself, represents an end product of PP after loss processes have been in operation. As discussed above, standing stock measured as chlorophyll is not constantly related to standing stock measured as carbon biomass. Chlorophyll is, therefore, quite often distinct from phytoplankton carbon fi xation (PP). Despite this empirical knowledge, chlorophyll is still widely used in estimations and evaluations of carbon fl ow in the ocean and probably most frequently in global estimates of PP calculated from satellite derived chlorophyll (Eppley et al. 1985, Longhurst et al. 1995, Antoine et al. 1996, Behrenfeld and Falkowski 1997a). Given these argu-ments, a better proxy for PP is clearly needed and we suggest that the simple model describing the seasonal variation, presented in this study, can be used as a bet-ter proxy for carbon fl ow than chlorophyll and allowing validation of dynamic ecosystem modelling of PP. We acknowledge that the parameterization of PP presented here is specifi c for the BSTZ. However, we believe that the parameterization would be similar in many other shallow stratifi ed coastal regions.

4.3 Top-down control on primary producers

The results of PP and estimates of potential grazing presented here support earlier studies over shorter time periods indicating that primary producers are con-trolled by zooplankton grazing in this the region (see e.g. Kiørboe 1993; Markager et al. 1999; Hansen et al. 2002). It is interesting, however, to note that the high-est impact from grazing on PP occurred during summer where the grazing estimates suggested consumption to be greater than the carbon produced by PP during the months April–October (see Fig. 4c, f, i). The maximum clearance rate was most often higher than the maximum ingestion which means that the estimates of potential grazing were dependent on the food concentration avail-able. As zooplankton, in general, are known to be prey selective feeders (see e.g. Atkinson 1995, Meyer-Harms et al. 1999), the food concentration calculated from chlorophyll and the carbon to chlorophyll ratio, most likely overestimated the food concentration. In addition, copepods and dinofl agellates feed on a variety of plank-ton prey, including ciliates (see e.g. Hansen and Calado 1999) which also can explain part of the reason why the potential grazing impact could exceed 100 %.

The DPP contributed signifi cantly (6-30 %) to the annual PP, especially during this stratifi ed summer season from April to September (range from 25-35 %). This suggests that DPP is likely an important food source for copepods given the high copepod biomass found from May to Sep-tember. Considering the fact that copepod biomass peaks in May-June and that spring production occur in February and March, it seems less likely that copepods are as de-pendent on the spring bloom as have been generally as-sumed (Cushing 1972). The spring PP contributes only 13-18 % of the annual PP and this production takes place approx. four months before copepod biomass peaks. As DPP has been shown to contribute signifi cantly to the an-nual primary production in several other regions (Holligan et al. 1984, Richardson and Pedersen 1998, Richardson et al. 2000, Sharples et al. 2001, Kononen et al. 2003, Veldhu-is and Kraay 2004, Weston et al. 2005, Lund-Hansen 2006, Lips et al. 2010) the importance of DPP for zooplankton feeding may very well apply outside the BSTZ.

The potential grazing presented in this study, was an estimate of how much carbon a given zooplankton bio-mass would ingest under specifi c biological and physical conditions. That grazing was not measured directly, of course, has its limitations when it comes to the interpre-tation of the results. However, as this study focuses on the general patterns in production, detailed grazing rates measured on specifi c dates, would provide only very time-specifi c “snapshots” of grazing which might not necessarily contribute to the understanding of the general patterns focused on here. Therefore, we believe that this



potential grazing as estimated from monthly variation in zooplankton biomass, food availability, temperature and size, is a robust indicator of the seasonal variation in rela-tive grazing pressure occurring in this system.

That the seasonal variation in PP show high summer val-ues in temperate regions is not a novel result (see Boyn-ton and Keefe 1982; Kemp and Boynton 1984 and refer-ences therein) but that this high summer PP is supported by DPP and that it is an important food source to higher trophic levels during this season needs more attention.

4.4 Seasonal variation of the vertical structure in primary production

The seasonal variation in primary production was con-trolled by temperature and light in the study region. The chlorophyll-specifi c P-E parameters showed low values during winter (from November to February the follow-ing year) as an indication that the phytoplankton were light limited during this season. Adaptation and accli-matization to low light, whether it occurs in the phyto-plankton community as a whole through changing spe-cies composition or within individual cells, is clear both over seasons, showing high summer and low winter val-ues, and between depths (highest values in the surface) as seen from Ik-values (Fig. 9). This is also known from laboratory and other fi eld studies (e.g. Richardson et al. 1983, Lindley et al. 1995, Morán and Estrada 2001) . P-E parameters measured in the PBL also exhibited a seasonal variation which indicated that the phytoplank-ton community in this layer were alive and active and capable of adapting or acclimating not only to the lower light intensities but also to the variation in light. The P-E parameter Pmax

B is controlled by enzymatic processes and, therefore, primarily dependent upon temperature. The highest Pmax

B and Ik values were, probably there-fore, found during August when the temperature peaked.

That light intensities are coupled to PP has been shown in previous studies, (see Domingues et al. 2011 and refer-ences therein), but this study further suggests that the sea-sonal pattern in the vertical structure of primary production is coupled to the surface nutrient concentrations. Here, the onset of the vernal production clearly came when stratifi ca-tion and surface irradiance increased in February-March. In April, when stratifi cation was frequent and nutrients be-came scarce, the surface photosynthetic activity decreased whereas the photosynthetic activity in the PBL increased. The results presented in this study suggest that there is a certain threshold upon which the vertical distribution of PP becomes dependent on the surface nutrient concentration. This value is suggested to be in the order of 2 μmol l-1 for surface DIN concentrations for the BSTZ. Both N and P seem to be limiting for phytoplankton growth considering the low summer surface values. The N:P ratio, however,

shows that N probably is the most limiting nutrient in the surface waters during summer as previously augmented by Conley (1999) because the DIP increases in late autumn (September-October) where nitrogen remains low. The surface PP increased again as soon as the stratifi cation weakened and nutrients again were available in surface water due to vertical mixing.

The general patterns found in this study indicate that there is a tight coupling between the frequency of the stratifi cation, the surface nutrient concentrations and the vertical distribution of PP. This coupling is believed to be relevant for future quantifi cation of subsurface PP in other stratifi ed regions and may improve, as suggested by Leeuwen and colleagues (2013), the modeling of subsurface PP in the future.

5 Conclusions

This study clearly showed that PP and chlorophyll are decoupled during summer months. This is mainly at-tributed to the seasonal variation in the carbon to chlo-rophyll ratio. In the search for a better proxy for carbon dynamics than chlorophyll concentrations, a param-eterisation of the seasonal variation in primary produc-tion was developed. The seasonal variation of chloro-phyll was clearly different from that of carbon biomass, especially with respect to the magnitude of the spring bloom. This suggests that the spring bloom may have received far more attention than it deserves, as chloro-phyll rather than carbon biomass has most often been used to describe phytoplankton biomass. The DPP was considered more important for the copepod feeding than the spring production due to the consistent contribution coming from DPP during summer months, whereas the spring PP only occurred in February and March, several months prior to the peak in copepod biomass.

Acknowledgements

This study received funding from grant number 2104-09-063212 and 2104-09-67259 from the Strategic Re-search Council of Denmark. Additional support was received from the Danish National Research Founda-tion via a grant to the Center for Macroecology Evolu-tion and Climate, University of Copenhagen and from the Department for Bioscience, Aarhus University. We thank the Department for Bioscience, Aarhus University for access to the monitoring data. We thank Hans Hen-rik Jacobsen for allowing us to use his calculations of phytoplankton carbon biomass and Morten Holtegaard Nielsen for producing the map of the research area.



A ppendix

A1. Starting depth and frequency of the pycnocline/bottom layer (PBL). The grey area marks the months with the highest frequen-cy of days with a stratifi ed water column.

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov DecAarhus Bight Av. PBLstart (m) 11 11 10 9 8 9 9 9 10 10 10 11

STD (m) 3 3 3 3 3 3 3 3 2 2 3 3Freq. PBL (%) 30 35 54 81 95 94 83 88 71 58 26 20

Aalborg Bight Av. PBLstart (m) 8 8 7 7 7 8 8 8 8 8 9 5STD (m) 2 2 3 3 2 2 2 2 3 3 2 2

Freq. PBL (%) 18 31 68 71 76 76 89 81 61 35 31 17

The Sound Av. PBLstart (m) 10 12 12 10 11 11 12 11 12 10 12 13STD (m) 3 4 5 4 4 3 4 3 4 5 6 7Freq. PBL (%) 82 79 87 96 98 98 98 99 99 84 86 75

Little Belt Av. PBLstart (m) 12 12 11 10 9 10 12 12 12 9 13 12STD (m) 2 3 4 2 3 3 4 3 4 3 3Freq. PBL (%) 39 36 52 65 94 89 87 89 69 61 24 6

Great Belt I Av. PBLstart (m) 12 12 10 9 7 9 9 9 9 11 11 11STD (m) 4 6 3 2 2 3 4 3 3 4 3 3Freq. PBL (%) 39 43 70 91 89 97 93 100 77 68 44 44

Great Belt II Av. PBLstart (m) 17 18 15 12 11 11 11 12 14 15 18 19STD (m) 6 6 5 4 3 4 4 3 5 6 6 7Freq. PBL (%) 63 62 85 98 100 100 98 98 90 80 73 68

A2. Monthly average values of alphaB from two depths for each station. Standard deviation is noted in grey and in parentheses and expresses the inter-annual variation.

Location Aalborg Sound Aarhus Little Belt Great Belt I Great Belt IIDepth 1 2 1 2 1 2 1 2 1 1 2January 12.1 7.6 10.8 13.3 6.7 10.8 8.5 14.6 8.8 7.7 6.3

(9.7) (8.0) (4.9) (17.4) (2.6) (13.7) (1.7) (6.1) (1.6) (3.5) (3.2)February 8.8 13.6 7.1 16.8 10.9 8.4 17.4 10.5 12.5 12.5 10.0

(5.8) (9.4) (3.7) (21.8) (8.3) (9.9) (12.1) (3.8) (5.8) (9.0) (4.4)March 7.2 10.9 9.6 9.1 8.2 7.8 10.0 10.0 11.3 9.7 10.1

(2.9) (8.8) (5.8) (4.4) (5.2) (5.3) (2.2) (2.4) (4.2) (3.0) (5.4)April 14.6 10.2 8.8 7.6 6.0 6.8 15.6 9.7 7.4 12.1 11.1

(7.8) (11.2) (5.0) (5.7) (2.9) (5.8) (13.3) (6.0) (1.3) (12.7) (5.5)May 14.8 7.3 11.9 16.2 8.8 12.3 14.6 10.3 10.5 11.4 10.9

10.6 3.7 8.7 16.8 5.1 12.2 4.1 7.8 4.1 2.8 4.5June 17.4 13.7 9.4 20.4 9.4 13.0 11.3 8.8 7.3 13.5 10.6

7.3 8.1 5.1 21.7 4.5 7.5 3.4 3.1 3.4 4.9 6.3July 15.5 13.3 12.9 13.1 8.0 15.9 13.2 11.5 12.1 12.6 10.5

6.1 7.4 11.8 15.0 3.7 11.4 4.9 7.4 2.9 5.1 4.3August 17.7 11.6 10.0 18.2 12.5 9.3 24.3 16.9 12.1 11.7 10.5

7.6 8.1 4.3 15.8 14.6 4.2 10.9 12.9 2.5 2.6 4.0Septemper 13.2 10.4 13.6 20.4 10.9 10.3 15.1 9.9 11.2 10.8 9.3

7.5 4.2 7.5 18.2 6.9 4.5 4.9 2.3 1.0 3.2 3.2October 12.6 9.5 11.0 8.5 9.1 9.5 9.8 10.0 9.3 10.6 10.7

8.0 4.2 6.5 6.2 3.5 5.0 4.6 4.7 1.0 3.3 3.4November 11.4 11.2 13.5 30.3 6.9 7.1 9.5 8.3 10.2 8.3 9.1

15.1 8.3 12.1 55.8 3.1 3.9 1.8 4.8 1.9 2.1 2.6December 6.4 7.5 7.2 7.0 8.2 7.9 11.6 9.0 10.7 8.6 8.9

2.1 5.6 3.0 5.7 4.1 7.1 4.7 4.8 3.6 4.1 3.4



References

Antoine, D., J.-M. André and A. Morel 1996. Oceanic primary production: 2. Estimation at global scale from satellite (Coastal Zone Color Scanner) chlorophyll. Global Biogeochemical Cy-cles 10:57-69.

Atkinson, A. 1995. Omnivory and feeding selectivity in fi ve copepod species during spring in the Bellingshausen Sea, Antarctica. ICES Journal of Marine Science: Journal du Con-seil 52:385-396.

Behrenfeld, M.J., E. Boss, D.A. Siegel and D.M. Shea 2005. Carbon-based ocean productivity and phytoplankton physiol-ogy from space. Global Biogeochemical Cycles 19:GB1006.

Behrenfeld, M.J. and P. Falkowski 1997. Photosynthetic rates derived from satellite-based chlorophyll concentration. Lim-nology and Oceanography 42:1-20.

Bendtsen, J., K.E. Gustafsson, J. Söderkvist and J.L.S. Hans-en 2009. Ventilation of bottom water in the North Sea-Baltic Sea transition zone. Journal of Marine Systems 75:138-149.

Berggreen, U., B. Hansen and T. Kiørboe 1988. Food size spectra, ingestion and growth of the copepod Acartia tonsa during development: Implications for determination of cope-pod production. Marine Biology 99:341-352.

Bienfang, P.K. and P.J. Harrison 1984. Sinking-rate response of natural assemblages of temperate and subtropical phyto-plankton to nutrient depletion. Marine Biology 83:293-300.

Boynton, W.R. and C.W. Keefe 1982. A comparative analysis of nutrients and other factors infl uencing estuarine phyto-plankton production. Pages 69-90 in V.S. Kennedy, editor. Estuarine comparisons. Academic, New York.

Breteler, W.C.M.K., H.G. Fransz and S.R. Gonzalez 1982. Growth and development of four calanoid copepod species under experimental and natural conditions. Netherlands Jour-nal of Sea Research 16:195-207.

Campbell, J., D. Antoine, R. Armstrong, K. Arrigo, W. Balch, R. Barber, M. Behrenfeld, R. Bidigare, J. Bishop, M.-E. Carr, W. Esaias, P. Falkowski, N. Hoepffner, R. Iverson, D. Kiefer, S. Lohrenz, J. Marra, A. Morel, J. Ryan, V. Vedernikov, K. Waters, C. Yentsch and J. Yoder 2002. Comparison of algo-rithms for estimating ocean primary production from surface chlorophyll, temperature, and irradiance. Global Biogeo-chemical Cycles 16:9-1-9-15.

Carr, M.-E., M.A.M. Friedrichs, M. Schmeltz, M. Noguchi Aita, D. Antoine, K.R. Arrigo, I. Asanuma, O. Aumont, R. Barber, M. Behrenfeld, R. Bidigare, E.T. Buitenhuis, J. Campbell, A. Ciotti, H. Dierssen, M. Dowell, J. Dunne, W. Esaias, B. Gentili, W. Gregg, S. Groom, N. Hoepffner, J. Ishizaka, T. Kameda, C. Le Quéré, S. Lohrenz, J. Marra, F. Mélin, K. Moore, A. Morel, T.E. Reddy, J. Ryan, M. Scardi, T. Smyth, K. Turpie, G. Tilstone, K. Waters and Y. Yamanaka 2006. A comparison of global esti-mates of marine primary production from ocean color. Deep Sea Research Part II: Topical Studies in Oceanography 53:741-770.

Carstensen, J., M. Sánchez-Camacho, C.M. Duarte, D. Krause-Jensen and N. Marbà 2011. Connecting the dots: Re-

sponses of coastal ecosystems to changing nutrient concentra-tions. Environmental Science & Technology 45:9122-9132.

Cloern, J.E. 1982. Does the benthos control phytoplankton biomass in south San Francisco Bay? Marine Ecology Prog-ress Series 9:191-202.

Conley, D. 1999. Biogeochemical nutrient cycles and nutri-ent management strategies. Hydrobiologia 410:87-96.

Conley, D., H. Kaas, F. Møhlenberg, B. Rasmussen and J. Windolf 2000. Characteristics of Danish estuaries. Estuaries 23:820-837.



Cushing, D.H. 1972. The production cycle and the numbers of marine fi sh. Pages 213-232, Symp. Zool. Soc., London.

Domingues, R.B., T.P. Anselmo, A.B. Barbosa, U. Sommer and H.M. Galvão 2011. Light as a driver of phytoplankton growth and production in the freshwater tidal zone of a turbid estuary. Estuarine, Coastal and Shelf Science 91:526-535.

Domingues, R.B., A. Barbosa and H. Galvão 2005. Nutri-ents, light and phytoplankton succession in a temperate estu-ary (the Guadiana, south-western Iberia). Estuarine, Coastal and Shelf Science 64:249-260.

Eppley, R.W., E. Stewart, M.R. Abbott and U. Heyman 1985. Estimating ocean primary production from satellite chloro-phyll. Introduction to regional differences and statistics for the Southern California Bight.

Eriksen, N.T. and J.J.L. Iversen 1995. Photosynthetic pig-ments as nitrogen stores in the cryptophyte alga Rhodomonas sp. Journal of Marine Biotechnology 3:193-195.

Fennel, K. and E. Boss. 2003. Subsurface maxima of phy-toplankton and chlorophyll: Steady-state solutions from a simple model. Limnology and Oceanography 48:1521-1534.

Geider, R.J. 1987. Light and Temperature Dependence of the Carbon to Chlorophyll a Ratio in Microalgae and Cyano-bacteria: Implications for Physiology and Growth of Phyto-plankton. New Phytologist 106:1-34.

Geider, R.J., H.L. MacIntyre and T.M. Kana 1997. Dynamic model of phytoplankton growth and acclimation: responses of the balanced growth rate and the chlorophyll a: carbon ratio to light, nutrient-limitation and temperature. Marine Ecology Progress Series 148:187-200.

Grasshoff, K., K. Kremling and M. Ehrhardt 1999. Methods of Seawater Analysis. 3 edition. Wiley-VCH.

Gustafsson, B.G. 2000. Time-Dependent Modeling of the Baltic Entrance Area. 2. Water and Salt Exchange of the Bal-tic Sea. Estuaries 23:253-266.



Hansen, B.W., E. Stenalt, J.K. Petersen and C. Ellegaard 2002. Invertebrate re-colonisation in Mariager Fjord (Den-mark) after severe hypoxia. I. Zooplankton and settlement. Ophelia 56:197-213.

Hansen, H.P. and F. Koroleff. 1999. Determination of nutri-ents. Pages 159-228, in K. Grasshoff, K. Kremling and M. Ehrhardt, editors. Methods of seawater analysis, Germany.

Hansen, P.J., P.K. Bjornsen and B.W. Hansen 1997. Zoo-plankton Grazing and Growth: Scaling Within the $2-2,000-\mum$ Body Size Range. Limnology and Oceanography 42:687-704.

Hansen, P.J. and A. Calado. 1999. Phagotrophic mechanisms and prey selection in free-living dinofl agellates. Journal of Eukaryotic Microbiology 46:382-389.

Hay, S.J., T. Kiørboe and A. Matthews 1991. Zooplankton biomass and production in the North Sea during the Autumn Circulation experiment, October 1987–March 1988. Conti-nental Shelf Research 11:1453-1476.

Henriksen, P., B. Riemann, H. Kaas, H.M. Sørensen and H.L. Sørensen 2002. Effects of nutrient-limitation and irradiance on marine phytoplankton pigments. Journal of Plankton Re-search 24:835-858.

Henson, S.A., J.L. Sarmiento, J.P. Dunne, I. Lima, S.C. Doney, J. John,and C. Beaulieu 2010. Detection of anthropo-genic climate change in satellite records of ocean chlorophyll and productivity. Biogeosciences 7:621-640.

Holligan, P.M. and D.S. Harbour 1977. The vertical distribu-tion and succession of phytoplankton in the western English Channel in 1975 and 1976. Journal of the Marine Biological Association of the United Kingdom 57:1075-1093.

Holligan, P.M., P.J.l. Williams, D.A. Purdie and R.P. Har-ris 1984. Photosynthesis, respiration and nitrogen supply of plankton populations in stratifi ed, frontal and tidally mixed shelf waters. Marine Ecology Progress Series 17:201-213.

Jeong, H., Y. Yoo, J. Kim, K. Seong, N. Kang and T. Kim 2010. Growth, feeding and ecological roles of the mixotrophic and heterotrophic dinofl agellates in marine planktonic food webs. Ocean Science Journal 45:65-91.

Joint, I. and S.B. Groom 2000. Estimation of phytoplankton production from space: current status and future potential of satellite remote sensing. Journal of Experimental Marine Bio-logy and Ecology 250:233-255.

Kahru, M., O.P. Savchuk and R. Elmgren 2007. Satellite mea-surements of cyanobacterial bloom frequency in the Baltic Sea: interannual and spatial variability. Marine Eco-logy Progress Series 343:15-23.

Kiefer, D.A. and J.N. Kremer 1981. Origins of vertical pat-terns of phytoplankton and nutrients in the temperate, open ocean: a stratigraphic hypothesis. Deep Sea Research Part A. Oceanographic Research Papers 28:1087-1105.

Kiørboe, T. 1993. Turbulence, Phytoplankton Cell Size, and the Structure of Pelagic Food Webs. Advances in marine bio-logy 29:1-72.

Kononen, K., M. Huttunen, S. Hällfors, P. Gentien, M. Lun-ven, T. Huttula, J. Laanemets, M. Lilover, J. Pavelson and A. Stips 2003. Development of a deep chlorophyll maximum of Heterocapsa triquetera Ehrenb. at the entrance to the Gulf of Finland. Limnology and Oceanography 48:594-607.

Koski, M. and T. Kiørboe 2005. Benthic life in the pelagic: Aggregate encounter and degradation rates by pelagic harpac-ticoid copepods. Limnology and Oceanography 50:1254-1263.

Kaas, H. and S. Markager 1998. Technical guidlines for ma-rine monitoring National Environmental Research Institute, Denmark.

Leeuwen, S., J. Molen, P. Ruardij, L. Fernand and T. Jickells 2013. Modelling the contribution of deep chlorophyll maxima to annual primary production in the North Sea. Biogeochemis-try 113:137-152.

Lerman, A., D. Lal and M. Dacey 1974. Stokes’ Settling and Chemical Reactivity of Suspended Particles in Natural Wa-ters. Pages 17-47, in R. Gibbs, editor. Suspended Solids in Water. Springer US.

Lindley, S.T., R.R. Bidigare and R.T. Barber 1995. Phyto-plankton photosynthesis parameters along 140°W in the equa-torial Pacifi c. Deep Sea Research Part II: Topical Studies in Oceanography 42:441-463.

Lips, U., I. Lips, T. Liblik and N. Kuvaldina 2010. Processes responsible for the formation and maintenance of sub-surface chlorophyll maxima in the Gulf of Finland. Estuarine, Coast-al and Shelf Science 88:339-349.

Longhurst, A., S. Sathyendranath, T. Platt and C. Caverhill 1995. An estimate of global primary production in the ocean from satellite radiometer data. J. Plankton Res. 17:1245-1271.

Lund-Hansen, L.C. 2006. Development and dynamics of a coastal sub-surface phytoplankton bloom in the southwest Kattegat, Baltic Sea. Oceanologia 48:1-14.

MacIntyre, H.L., T.M. Kana, T. Anning and R.J. Geider 2002. Photoacclimation of photosynthesis irradiance re-sponse curves and photosynthetic pigments in microalgae and cyanobacteria. Journal of Phycology 38:17-38.

Markager, S., T.G. Nielsen, J. Carstensen, D. Conley, K. Dahl, J. Hansen, P. Henriksen, A. Josefson, M.M. Larsen, B. Pedersen, B. Rasmussen, J. Strand, G. Ærtebjerg, H. Fossing, J.S. Lauersen, O. Hertel, H. Skov, L.M. Svendsen, M. Cleemann and G. Pritzl 1999a. Marine områder. Status over miljøtilstanden i 1998. NOVA 2003. National Environ-mental Research Institute.

Markager, S., W. Vincent and E.Y. Tang 1999b. Carbon fi xa-tion by phytoplankton in high Arctic lakes: Implications of low temperature for photosynthesis. Limnology and Oceano-graphy 44:597-607.

Menden-Deuer, S. and E.J. Lessard 2000. Carbon to Volume Relationships for Dinofl agellates, Diatoms and Other Protist Plankton. Limnology and Oceanography 45:569-579.



Meyer-Harms, B., X. Irigoien, R. Head and R. Harris 1999. Selective feeding on natural phytoplankton by Calanus fi n-marchicus before, during and after 1997 spring bloom in the Norwegian Sea. Limnology and Oceanography 44:154-165.

Morán, X.A.G. and M. Estrada 2001. Short-term variability of photosynthetic parameters and particulate and dissolved primary production in the Alboran Sea (SW Mediterranean). Marine Ecology Progress Series 212:53-67.


Nielsen, T. and C. Andersen 2002. Plankton community structure and production along a freshwater-infl uenced Nor-wegian fjord system. Marine Biology 141:707-724.

Nielsen, T.G., B. Lokkegaard, K. Richardson, F.B. Pedersen and L. Hansen 1993a. Structure of plankton communities in the Dogger Bank area (North Sea) during a stratifi ed situa-tion. Marine Ecology Progress Series 95:115-131.

Nielsen, T.G., B. Løkkegaard, K. Richardson, F.B. Pedersen and L. Hansen 1993b. Structure of plankton communities in the Dogger Bank area (North Sea) during a stratifi ed situa-tion. Marine Ecology Progress Series 95:115-131.

Palmer, M.A., G.L. van Dijken, G. Mitchell, B.J. Seegers, K.E. Lowry, M.M. Mills and K.R. Arrigo 2013. Light and nutrient control of photosynthesis in natural phytoplankton populations from the Chukchi and Beaufort seas, Arctic Ocean. Limnology and Oceanography 58:2185-2205.

Pomeroy, L.R., C.F. D’Elia and L.C. Schaffner 2006. Limits to top-down control of phytoplankton by oysters in Chesa-peake Bay. Marine Ecology Progress Series 325:301-309.


Richardson, K. and F.B. Pedersen 1998. Estimation of new production in the North Sea: consequences for temporal and spatial variability of phytoplankton. ICES Journal of Marine Science: Journal du Conseil 55:574-580.


Sabatini, M. and T. Kiørboe 1994. Egg production, growth and development of the cyclopoid copepod Oithona similis. Journal of Plankton Research 16:1329-1351.

Satapoomin, S. 1999. Carbon content of some common tropi-cal Andaman Sea copepods. Journal of Plankton Research 21:2117-2123.

Savchuk, O.P. 2005. Resolving the Baltic Sea into seven sub-basins: N and P budgets for 1991–1999. Journal of Marine Systems 56:1-15.



Steele, J.H. 1964. A study of production in the Gulf of Mexi-co. Mar. Res. 3:211-222.

Steele, J.H. and C.S. Yentsch 1960. The vertical distribution of chlorophyll. J. mar. biol. Ass. U.K. 39:217-226.


Strathmann, R.R. 1967. Estimating the Organic Carbon Con-tent of Phytoplankton from Cell Volume or Plasma Volume. Limnology and Oceanography 12:411-418.


Tang, E.P.Y. 1996. Why do dinofl agellates have lower growth rates? Journal of Phycology 32:80-84.

Timmermann, K., S. Markager and K.E. Gustafsson 2010. Streams or open sea? Tracing sources and effects of nutrient loadings in a shallow estuary with a 3D hydrodynamic–eco-logical model. Journal of Marine Systems 82:111-121.

UNESCO 1981. International oceanographic tables., Paris.

Utermöhl, H. 1958. Zur vervollkommung der quantitativen phytoplankton methodik. Mitt. Int. Ver. Theor. Angew. Lim-nol 9:1-38.

Valderrama, J.C. 1981. The simultaneous analysis of total nitrogen and total phosphorus in natural waters. Marine Chemistry 10:109-122.


Weston, K., M. Fernand, D.K. Mills, R. Delahunty and J. Brown 2005. Primary production in the deep chlorophyll max-imum of the central North Sea. Journal of Plankton Research 27:909-922.


Under revision for Marine Ecology Progress Searies

Katherine Richardson1*, Jørgen Bendtsen2, Jens Tang Christensen3, Mohamed Adjou1, Maren Molke Lyngsgaard1, Karen Marie Hilligsøe3, Jens B. Pedersen3, Torben Vang3 and Morten Holtegaard Nielsen4

1Center for Macroecology, Evolution and Climate, Universitetsparken 15, DK-2100 Copenhagen, Denmark 2ClimateLab, Symbion Science Park, Fruebjergvej 3, Box 98, DK-2100 Copenhagen, Denmark3Department of Bioscience, Aarhus University, Ole Worms Allé 1, DK-8000 Aarhus, Denmark 4Arctic Technology Centre, Department of Civil Engineering, Technical University of Denmark, Kemitorvet, Building 204, DK-2800 Kgs. Lyngby, Denmark

*Corresponding author: [email protected]

PAPER IVLocalised upwelling may lead to heterogeneity in the plankton food web in a frontal region of the Sargasso Sea

Photo: Katherine Richardson


Abstract

Frontal regions are often associated with increased biological activity but the mechanisms leading to this relationship are not well known. This study, which focuses on a frontal region along the Subtropical Convergence Zone (STCZ) in the Sargasso Sea, demonstrates that there may be localized patches within this frontal region that experience elevated vertical mixing and an associated vertical fl ux of nutrients. The fi rst suggestion of this localized vertical mixing in our data was a greater similarity (Jaccard index) between the diatom communities at 10 m and those in the deep chlorophyll maximum (DCM) located at > 120 m than elsewhere in the frontal region. Thorpe-displacements supported the hypothesis of elevated mixing intensities in this region as did vertical mixing rates inferred from stratifi cation and vertical cur-rent shear calculated from ADCP measurements. Combining these mixing estimates with vertical nutri-ent gradients suggests that nutrient fl uxes to the euphotic zone may be an order of magnitude greater here than elsewhere in the frontal region. Several lines of biological evidence suggest that the planktonic food web may be stimulated at these sites of elevated vertical nutrient mixing These include elevated values of primary production, total seston concentrations, and variable fl uorescence, Fv/Fm. In an accompany-ing study (Munk et al. 2010), eel larvae were found to be most abundant in the part of the frontal region where we also found evidence for localized vertical mixing.

Keywords: plankton, eel larvae, Sargasso Sea, mixing, primary production.

Localised upwelling may lead to heterogeneity in the plankton food web in a frontal region of the Sargasso Sea

Katherine Richardson1*, Jørgen Bendtsen2, Jens Tang Christensen3, Mohamed Adjou1, Maren Molke Lyngsgaard1, Karen Marie Hilligsøe3, Jens B. Pedersen3, Torben Vang3 and Morten Holtegaard Nielsen4

1Center for Macroecology, Evolution and Climate, Universitetsparken 15, DK-2100 Copenhagen, Denmark2ClimateLab, Symbion Science Park, Fruebjergvej 3, Box 98, DK-2100 Copenhagen, Denmark3Department of Bioscience, Aarhus University, Ole Worms Allé 1, DK-8000 Aarhus, Denmark 4Arctic Technology Centre, Department of Civil Engineering, Technical University of Denmark, Kemitorvet, Building 204, DK-2800 Kgs. Lyngby, Denmark

*Corresponding author: [email protected]


Plankton food web in the Sargasso Sea

Introduction

“Fronts”, i.e. regions where different water masses meet, have frequently been identifi ed as being important for the distribution of biological processes in the ocean. In some cases, they have simply been identifi ed to form a faunal boundary delineating populations in different water masses (Diekmann and Pietowski, 2002) while, in other studies, greater biomass accumulations and/or enhanced productivity have commonly been reported at these frontal interfaces between the water masses. Usu-ally, however, the mechanisms leading to this increased productivity are not well understood.

The frontal region in the Southern Tropical Convergence Zone in the southern Sargasso Sea is interesting in that patches of American and European eel larvae have, in several studies (e.g. Kleckner et al. (1983), Kleckner & McCleave (1988) & Munk et al. (2010)), been found here. The reasons for this association are, however, not clear. One hypothesis is that this frontal zone might comprise more rewarding feeding grounds for the larvae than surrounding waters. Two studies (Andersen et al., 2011; Riemann et al., 2011) carried out in concomitantly with the 2007 Munk et al. (2010) study failed, however, to identify signifi cant differences in either primary or secondary production between frontal and non-frontal waters, although the Andersen et al. study did suggest a greater abundance of some zooplankton in the frontal region than elsewhere.

Thus, the current state of understanding regarding the distribution of eel larvae in the Sargasso Sea at the time of the Munk et al. (2010) study is that they were only found in the frontal region, that they were patchily dis-tributed and that there was no indication of increased primary or secondary plankton production that would suggest better feeding conditions in the frontal region than elsewhere for eel larvae. The present study further extends the understanding of the Sargasso Sea ecosys-tem at the time of the Munk et al. (2010) study system by examining the frontal region, itself, for potential heterogeneity in physical and biological water column characteristics.

The front in question is located in the subtropical con-vergence zone (STCZ), a transition zone between 25° N and 32° N in the North Atlantic sub-tropical gyre sepa-rating relatively cold waters in the area dominated by the Westerlies from warm waters in the trade wind region (Leetmaa & Voorhis, 1978). Results from the MODE-experiments (MODE group, 1978) have shown that the STCZ region in the Sargasso Sea can be char-acterized by mixing between the cold and warm water masses in the form of north-south oriented plumes with

spatial scales of about 200 km, where smaller scale fron-tal structures with spatial scales of about 10 km appear on the edges of these plumes.

The meso-scale variability associated with the frontal structures is characterized by temperature gradients of about 0.1 °C km-1, depth ranges of the order of 100 m and relatively short temporal scales of about 3-5 days. Satellite derived estimates of frontal structure and loca-tions confi rm a relatively high probability of front oc-currences in the Sargasso Sea region between 25-35° N and 74-55° W, with signifi cant increased frontal activity in the spring period (Ullman et al., 2007).

The presence of an STCZ and an associated eastward fl owing sub-tropical countercurrent are features typical of the northern sub-tropical gyres in both the Atlantic and the Pacifi c. The countercurrent has been found to be associated with interactions between the general wind-driven circulation from the Westerlies and trades and the thermal stratifi cation (Cushman-Roisin, 1984) and mixed layer depth distribution (Kubokawa & Inui, 1999) of the upper ocean in the gyres. High-resolution studies of the small-scale frontal structures in the North-ern Pacifi c STCZ revealed intense mixing and interleav-ing on spatial scales of ~5 km in the upper 150 m of the water column (Shcerbina et al., 2009). The STCZ is, therefore, a consistent feature of the sub-tropical gyres. It is an area characterized by deep small-scale frontal dynamics on relatively small spatial and temporal scales and these dynamics have the potential of affecting bio-geochemical cycling signifi cantly through increased mixing of nutrients.

The present study is based on data collected on the same cruise as the Munk et al. (2010), Andersen et al. (2011) and Riemann et al. (2011) studies discussed above. We hypothesized that the frontal dynamics described above might potentially create heterogeneity in the plankton production characteristics within the frontal region. If this were the case, then comparing characteristics of the frontal region as a whole with non-frontal regions would not necessarily provide a representative description of the potential of this region to meet the feeding needs of higher tropic levels. We conclude that localized vertical mixing leads to enhanced nutrient delivery to the photic zone at specifi c sites within the frontal region and that this nutrient delivery may stimulate the plankton food web here. Although our data do not provide the basis to identify a direct link between this apparent stimulation at the base of the plankton food web and the feeding by eel larvae, we do note that the greatest concentrations of eel larvae recorded by Munk et al (2010) were found in an area of the frontal region where this enhanced nutri-ent mixing is inferred.



Materials and Methods

Study area: The study was part of the Danish Galathea 3 Expedition (www.galathea3.dk) that circumnavigated the globe in 2006-2007. Sampling in the Sargasso Sea took place from 29 March to 10 April, 2007. Multi-dis-ciplinary studies were carried out. However, planning of sampling positions was made with the main purpose of mapping the distributions of eel larvae (Munk et al. 2010). CTD and water collection were generally carried out during daylight hours (Table 1). Three north-south transects along the longitudes 64° W (Transect 1), 67° W (Transect 2) and 70° W (Transect 3) were sampled. Tran-sect positions were chosen on the basis of the positions of SST isotherms prior to the sampling period and all transects provided samples from both in and outside of the frontal area.

Water column characteristics: Conductivity, tempera-ture (two separate C- and T-sensors), and pressure were measured using a Seabird Instruments 911 series. The instruments were attached to a rosette of 12 Niskin bot-tles (30 l), a fl uorometer (SCUFA), light meters (Bio-spherical) and two oxygen sensors (SBE43). All instru-ments were calibrated and data quality controlled before use. Samples were taken from selected depths for oxy-gen calibration (Winkler), and salinity calibration (Por-taSal). Measurements made with the SCUFA-fl uorome-ter mounted with the CTD were subsequently calibrated against chlorophyll a concentrations (r2=0.79, n=17) from water samples (GF/F fi ltered). The fi lters were stored in glass vials with 5 ml 96 % ethanol and frozen at –20° C. Prior to analysis, the samples were extracted for a minimum of 6 h and a maximum of 24 h in dark-ness at room temperature. The samples were measured

on a TD-700 fl uorometer from Turner Designs, which was calibrated against a pure chlorophyll a standard. Chlorophyll concentration was determined after Method 445.0 by the U.S. Environmental Protection Agency as suggested by Turner Designs (Arar & Collins, 1997). Samples for inorganic nutrients, variable fl uorescence, primary production, and phytoplankton species compo-sition were only collected at every second or third CTD station (Fig. 1; Table 1).

CTD data analysis: Pressure measurements were aver-aged with a time constant of 0.15 s, corresponding to a vertical scale of 8 cm for a descent rate of 0.5 m s-1

(i.e. a scan rate of 24 Hz for the SBE9+ instruments in the CTD system). The two conductivity and temperature sensors were only analysed for the downcasts and rever-sals were removed by only considering measurements with a minimum CTD velocity of 0.1 m s-1. Finally, the CTD-data were bin-averaged in intervals of 0.5 dbar.

Determination of inorganic nutrients: At each station, seawater from the DCM and the standard sampling depths of 10, 30, 60, 100, 200 and 400 m was tapped and immediately frozen. The subsequently fi ltered seawater (Millipore Millex-GP Hydrophilic PES0 0.22 mm) was analyzed for nitrate, nitrite, phosphate, ammonium and silicate by wet-chemistry methods according to Grass-hoff (Ed.) et al. (1983) with a SANPLUS System Scalar auto-analyser at the National Environmental Research Institute (NERI), University of Aarhus, Denmark. The detection limits were 0.06, 0.1, 0.04, 0.3 and 0.2 μmol kg-1 for PO4, NO3, NO2, NH4 and silicate, respectively. A detailed description of the distribution of inorganic nutrients is given in Riemann et al. (2011). Therefore, only nutrient information relevant for the current analy-ses is presented here.

CTD and biologyCTD only

10°

15°

20°

25°

30°

35°

40°

-100° -90° -80° -70° -60° -50° -40°

10°C 15°C 20°C 25°C 30°C

Figure 1. Position of stations in the Sargasso Sea where both CTD and biology measurements (bullets) and where only CTD (squares) measure-ments were made. Longitude and Lat-itude are given on the x and y axes, respectively. Background color de-picts satellite derived SST from 31st of March 2007. The rectangle shows the area of study which is shown in detail in Fig. 6.



Primary production was determined using the carbon 14 method modifi ed after Steemann Nielsen (1952). The actual method applied is described in detail in Hilligsøe et al. (2011). Briefl y, samples were collected both from surface waters (10 m) and the DCM and, after the addi-tion of 14C, incubated at a range of photon fl ux densities for 2 hours. In this manner, Photosynthesis (P) versus Light (E) curves were established for both the surface and DCM populations. These were applied to the hourly average of the light estimated to be found at each depth of the water column in 1 meter depth intervals through-out the day (applying the light attenuation coeffi cient

recorded at the specifi c station and assuming average hourly light averages for light incident upon the surface based on the actual light measured in the area during the week of sampling).

The P-E characteristics for the surface population were assumed to apply for populations down to the DCM and the DCM characteristics for all deeper depths. A bio-mass correction was, however, applied using the actual chlorophyll recorded each 1 m depth interval and as-suming the actual amount of photosynthesis occurring to be linearly related to the chlorophyll concentration.

Table 1. Station list for the three transects with positions and time of measurements (UTC). Stations where CTD, diversity, nutri-ents and biology (i.e. primary production (PP), chlorophyll a) were measured are indicated.

Station ID Latitude [ºN] Longitude [ºW] Datedd/mm/yyyy

Time [UTC] CTD and diversity measurements

PP, chl a andnutrients

Transect 1

17-1 19.00 64.00 03/29/2007 12:08 x x17-5 20.50 64.01 03/29/2007 22:48 x17-10 22.04 64.00 03/30/2007 11:04 x x17-12 23.00 64.00 03/30/2007 16:37 x17-15 24.00 64.00 03/30/2007 23:54 x17-18 24.50 64.00 03/31/2007 05:47 x17-22 25.25 64.00 03/31/2007 22:57 x x17-25 26.00 64.00 04/01/2007 06:00 x17-28 26.50 64.00 04/01/2007 13:01 x x17-33 27.00 64.00 04/01/2007 20:39 x17-36 27.33 64.00 04/02/2007 01:33 x x17-43 27.66 64.00 04/02/2007 14:56 x x17-46 28.00 64.00 04/02/2007 19:42 x17-49 28.33 64.00 04/03/2007 19:42 x17-52 28.66 64.00 04/03/2007 06:30 x

Transect 2

17-55 28.50 67.00 04/03/2007 20:28 x x17-57 27.75 67.00 04/03/2007 01:52 x17-59 27.00 67.00 04/03/2007 07:06 x17-61 26.50 67.00 04/04/2007 11:25 x x17-63 26.00 67.00 04/04/2007 16:07 x17-65 25.66 67.00 04/04/2007 20:05 x17-68 25.33 67.00 04/05/2007 00:58 x17-72 25.00 67.00 04/05/2007 08:26 x x17-74 24.50 67.00 04/05/2007 16:00 x

Transect 3

17-79 24.99 70.00 04/06/2007 08:31 x x17-80 25.50 70.00 04/06/2007 11:50 x17-83 26.00 70.00 04/06/2007 18:15 x17-87 26.50 70.00 04/07/2007 02:37 x17-94 27.03 70.00 04/07/2007 16:02 x x17-95 27.50 70.00 04/07/2007 19:08 x17-99 28.00 70.00 04/08/2007 03:02 x17-103 28.99 70.00 04/08/2007 12:36 x x17-107 30.00 70.00 04/08/2007 20:32 x17-111 32.51 70.00 04/09/2007 14:00 x x



Only particulate primary production is considered here. Note that there are minor differences in the magnitude of the total water column primary production estimates reported here and in Riemann et al. (2011) owing to dif-ferences in the method used to fi t the curves to the P-E parameters between the two studies.

Phytoplankton Community Analysis: Samples for phyto-plankton analysis were taken from Niskin bottles closed at the surface (10 m) and the deep chlorophyll maxi-mum (DCM; depths varied between 100 and 140 m). These were preserved in 1 l brown glass bottles with acidifi ed Lugol’s solution (approximately 2 % fi nal con-centration) for taxonomic determination of nano and microplankton. Identifi cation of the organisms present in this size fraction (> ca. 2 μm) was made using quan-titative light microscopy according to Utermöhl (1958). Cell biovolumes were estimated from linear dimensions and calculated using appropriate geometric volume for-mulas. Carbon content was estimated using carbon to volume relationships given in Edler (1979). Analyses were carried out by Orbicon A/S (Aarhus Denmark).

Jaccard Index, J, (Jaccard, 1901) was calculated to de-termine the similarity between the diatom communities at 10 m and the DCM at each sampled station. J meas-ures the similarity between two communities C1 and C2 and is calculated as:

where α is the number of shared species in the com-munities C1 and C2, β is the number of species present in community C1, but absent in C2, and γ is the number of species present in community C2, but absent from C1. J ranges from 0, when the C1 and C2 have no shared species; to 1 when all species are shared. The diatom subcomponent of the plankton community was selected for the similarity analysis as it was this group that was most easily and best identifi ed.

Zooplankton analysis: As a proxy for mesozooplank-ton distributions, the total abundance of calanoid and cyclopoid copepods was determined at selected sta-tions down to 250 m. Samples for determining the ver-tical distribution of meta-zooplankton (>50 μm) were collected using a HYDRO-BIOS MultiNet Midi. The MultiNet consists of 5 nets with a closing option for 4 of the nets. The MultiNet was lowered and hauled with speeds of 18 m min-1 and 10 m min-1, respectively. Sam-ples were fi xed in 4 % formalin and stored at 5 ºC prior to microscopic enumeration of copepods and nauplii.

Thorpe displacements: Thorpe displacements are a meas-ure of the length scale associated with instabilities in a vertical density profi le and it is assumed that the instabil-

ity is due to overturning induced by turbulent eddies. The vertical displacement of water parcels by eddies causes instabilities in the CTD-profi les and displacements are, therefore, representative of the eddy-induced vertical movement of water parcels from an initial stable strati-fi cation. Thus, Thorpe displacements can be related to the intensity of vertical turbulent mixing (Thorpe, 1977). Thorpe displacements (L) were determined by reorder-ing the density profi les by applying a noise level of Δρ = 10-3 kg m-3 (Ferron et al., 1998). According to Galbraith and Kelly (1996), the minimum thickness of resolvable overturn (Lmin) can be estimated as Lmin = 2 g Δρ/ (N2 ρ0), where g is the constant of gravitation and ρ0 is a ref-erence density. A typical stratifi cation frequency for in-termediate depth levels at 150 m depth in the area was about N = 0.008 s-1, corresponding to a minimum resolv-able thickness of 0.3 m. However, the bin-average of the CTD-profi le made a lower limit on the Thorpe scale dis-placements of about 0.5-1 m. Thorpe displacements were subsequently averaged over 5 m intervals.

ADCP-measurements: Current velocities were meas-ured by two downward-looking ADCPs (Acoustic Dop-pler Current Profi ler) from Teledyne RD Instruments: a 600 kHz Workhorse Monitor and a 75 kHz Workhorse Long Ranger. The two ADCPs had ranges of about 50 m and about 600 m, respectively, where the upper part of the water column had to be blanked out. Data were col-lected using single pings and depth cells of 2 m and 16 m, respectively. The velocity observations were refer-enced to earth by subtracting the vessel velocity, found using the bottom track of one or the other ADCP where available or calculated from GPS data, subjected to a 5-second moving average. There was good agreement between the measurements in the depth range 30 - 50 m where the two instruments overlapped. In the analysis below we consider the mixing below 50 m and, there-fore, only data from the 75 kHz sensor are used. In order to examine the velocity structure of the water column, the ADCP data collected at the time of CTD profi ling were averaged during one hour into one vertical veloc-ity profi le. This corresponds to averaging of about 900 independent observations.

Richardson number and vertical mixing: The Richard-son number (Ri) describes the importance of stratifi ca-tion for damping vertical mixing compared to the verti-cal velocity shear which provides energy for turbulent motion. A large Richardson number, therefore, indicates reduced mixing due to a relatively large stability in the water column and vice versa. The stratifi cation is de-termined from the Brunt Väisäla frequency (N2 = -g/ρ ∂ρ/∂z), where g is the gravitational constant, ρ is density and z the vertical coordinate and the gradient is deter-mined continuously over a 10 m interval). The density was determined from CTD-measurements and the ve-

+ βα

α + γJ =



locity shear (S2) was determined from the eastward (u) and northward (v) velocity components measured by the 75 kHz ADCP-sensor (S2 = (∂u/∂z)2 + (∂v/∂z)2). The Richardson number is defi ned as: Ri = N2/S2. Empiri-cal relationships have been established for calculating the vertical turbulent diffusion coeffi cient (kv) from the Richardson number and we apply the KPP-parameteri-sation after Pacanowski and Philander (1981):

The parameterisation implies a minimum value of kv of about 10-5 m2 s-1 for large Ri-numbers, corresponding to typical values for the ocean interior, whereas small Ri-numbers increase the mixing coeffi cient signifi cantly. The KPP-parameterisation has been found to provide a good estimate of the shear-induced vertical mixing inten-sities when compared to other methods (e.g. Cisewski et al, 2005).

Vertical nutrient fl ux: The vertical nutrient fl ux (FC) due to turbulent mixing was estimated as FN = –kv DINz, where DINz was the vertical gradient of dissolved inor-ganic nitrogen (i.e. the sum of nitrate, nitrite and ammo-nium). The vertical gradient was estimated from observed values in a depth range below the turbulent mixed layer. The vertical nutrient fl uxes were converted to carbon units (i.e. new production, sensu Dugdale and Goering, 1967) by applying a Redfi eld C:N-ratio of η=106:16, i.e. the carbon fl ux was determined as FC(N) = η FN.

Variable fl uorescence: Variable fl uorescence (Fv/Fm), which indicates the potential for electron transport in Photosystem II, was determined on discrete samples (3 replicates) taken from selected depths in the water col-umn. Samples were incubated in darkness for minimum 30 min in a thermally insulated container. Variable fl uo-rescence (Fv/Fm) was measured with a FASTTRACKA fl uorometer (Chelsea Instruments group Ltd.). A single turnover protocol with 30 sequences per acquisition each including 100 saturation fl ashlets and 20 relaxation fl ashlets was utilised. The sequence interval was set to 1000 ms. Fv/Fm was calculated from a saturation phase fi t and a relaxation phase fi t following Kolber et al. (1998). Only measurements with a qualifying saturation phase and relaxation phase were included in the data set.

Stable Isotope analyses: Bulk particulate matter (seston) was obtained by vertical hauls of a Multinet (Hydro-Bios, Kiel) with 0.25 m2 opening and 50 μm mesh nets. In each haul, samples were obtained from 5 pre-programmed strata from 400 m depth and up. Samples were split in half on a Folsom plankton splitter and one half was freeze dried for analysis of C and N stable isotopes.

Carbon and nitrogen isotope analyses were performed at the Scottish Crop Research Institute using an automated nitrogen-carbon analyser (ANCA) coupled to a 20/20 isotope ratio mass spectrometer (SerCon Ltd, Crewe, UK). Samples in crimped tin capsules were introduced via a solid autosampler. The elemental analyzer (EA) reactor tubes were comprised of two quartz glass tubes fi lled with chromium(III) oxide and copper oxide held at 1000 °C and reduced copper, held at 620 °C for com-bustion and reduction, respectively. A post-reactor gas chromatography (GC) column was kept at 70 °C for separation of evolved N2 and CO2. The working stand-ard for N analysis was 1 mg leucine prepared by freeze drying 50 μl of a 20 mg ml-1 stock solution into tin cups, and calibrated against ‘Europa fl our’ and IAEA stand-ards N1 and N2. Only nitrogen isotopes are used in this study and data are reported as δ15N relative to air ac-cording to the formula:

δ15N = [(Rsample/Rstandard) – 1] × 1000

where R is the ratio of the heavy to the light isotope (15N/14N).

In order to convert the δ15N of the seston from the differ-ent strata to a δ15N value for seston in the water column as a whole, the isotope signature for each stratum was multiplied by the weight of the seston in that stratum di-vided by the total weight of seston in the water column down to 400 m and the results for all strata summed.

Results

A thermal frontal system characteristic of the STCZ was encountered between about 24.5 and 27° N on all three transects. Transect 1 supplied the best coverage of the frontal region as well as waters to the north and south of the front and characteristics of that transect are illus-trated here (Fig. 2a). The region to the north was charac-terized by considerably colder surface water compared to south of the front, with temperatures at 10 m depth below 21 °C (e.g. stn. 17-43 (27.66° N). In contrast, the temperatures at 10 m between approximately 26-27.33° N (e.g. 17-25 to 17-36)) were between 22.9-23.3 °C. Relatively small salinity changes were observed in the same area with salinities ranging between 36.70-36.78 at 10 m (i.e. SP, PSS-78). A clear sub-surface salinity maximum, corresponding to Subtropical Underwater (STUW) was observed south of the front at intermedi-ate depths, i.e. between 100-200 m. but elevated sub-surface salinities were also noted north of the frontal region (Fig. 2b). Relatively low dissolved oxygen con-centrations in the STUW of around 175 μmol kg-1 were clearly separated from the more well-ventilated subtrop-

kv = + 10-5 m2 s-15×10-3 m2 s-1 + (1 + 5×Ri)2 × 10-4 m2 s-1

(1 + 5×Ri)3



ical mode water (STMW, also referred to as “Eighteen-degree water”) observed below, i.e. in the depth range 200-400 m (Fig. 2c).

There are no commonly accepted criteria for identify-ing the geographic boundaries of a frontal region. For our analyses, we choose to defi ne the frontal region as being there where outcropping isopycnals are in the density range (σt) between 24.5 and 25.5 kg m-3. Here, the water characteristics are relatively constant and characterized by surface temperatures between 22.3 ºC and 25 ºC (at a depth of 10 m) and salinities above 34.48 (Fig. 3). Stations to the north of the frontal region are signifi cantly colder and waters south of the fronts are signifi cantly warmer and less saline. Tem-perature and salinity are, therefore, relatively similar in the frontal region along the three transects. Where there is good data coverage for both frontal and non-frontal stations, all of the collected data are presented here for purposes of comparison. However, the focus of the analyses is on the frontal region, itself, and try-ing to identify and understand heterogeneity in both water column characteristics and biological activity within this region.

In the deep waters encountered in the frontal zone, i.e., below ~200 m, the STMW mass is relatively well ven-tilated with Apparent Oxygen Utilisation (AOU) values below 50 μM (Fig. 4b, d, f). Higher AOU is observed on the more southerly stations along Transect 1 (64° W).

24.525 25.525.5

26 2626

26.5

26.526.5

24.525 25.525.526 26

26

26.5

26.5

26.5

24.525 25.5

25.5

26 26

26

26.5

26.526.5

24.525 25.5

25.5

26 26 26

26.5

26.5

26.5

no. 1 5 10 12 15 18 22 25 28 33 36 43 46 49

Latitude (°N)19 20 21 22 23 24 25 26 27 28

1619202122232426

3636.236.436.636.836.937

165175180185190195200210

0.10.20.30.40.50.6

Dep

th (m

)D

epth

(m)

Dep

th (m

)D

epth

(m)

100

200

300

0

100

200

300

0

100

200

300

0

100

200

300

0

b)

c)

d)

a)

Figure 2. Observations of (a) potential temperature (°C), (b) salinity, (c) oxygen (μmol kg−1) and (d) chlorophyll a (μg l−1) derived from fl uorescence along Transect 1 at 64 °W.

25

24.5

25.5

θ (°

C)

S (psu)36.0 36.2 36.4 36.6 36.8 37.0

18

20

22

24

28

26

Figure 3. Potential temperature (θ) and salinity (S) of CTD surface measurements along the three transects: 64W (circles), 67W (diamonds) and 70W (triangles). Filled symbols show lo-cations with both biology and CTD-measurements. Isopycnals (σt) are shown in intervals of 0.5 kg m-3.



Transects 2 and 3 only traverse the northern part of the frontal area and, thus, do not extend into the region where high AOU values in the STMW were recorded on Tran-sect 1. However, in all other respects, the density struc-ture along the three transects appears to be comparable.

In our initial analysis of the data, we noted that the greatest similarities (Jaccard Index) between the phy-toplankton communities at the surface and the DCM for the diatoms (Fig. 6) were recorded in the frontal region of Transect 1. To test the hypothesis that J might be used as a proxy for mixing in the water column, we compared this index to the temperature difference from 10 m to the DCM. The two parameters were signifi cantly negatively correlated (p<0.05, n=25 (t-test)), suggesting that the Jaccard Index may be used as a proxy for water column stratifi cation (and, therefore, mixing) in this region. As the variability recorded in the Jaccard Index suggested

that vertical mixing intensity within the frontal region might vary, we turned to other methods to estimate the distribution of vertical mixing within this region.

Thorpe displacements recorded in the frontal region along Transect 1, i.e. the same region that we recorded the highest Jaccard indices, were considerably larger than those recorded on Transects 2 and 3. The largest Thorpe displacements along Transect 1 were observed in the surface mixed layer, i.e. upper 100 m, and also in the depth stratum 100-200 m in the frontal area (Fig. 4). Eel larvae (redrawn from Munk et al. 2010) were encountered on all three transects and only in the frontal zone (as defi ned above), i.e. in the region of the well ventilated STMW (Fig. 4a, c, e). The largest eel larvae concentrations were observed in the frontal zone on Transect 1, where Thorpe displacements also were sig-nifi cantly higher than elsewhere.

24

25

25

25

T0 (1 m)T0 (5 m)

18 20 22 24 26 28 30 32 34Latitude (°C)

80

60

40

20

0

300

200

100

0

300

200

100

0

300

200

100

0

Dep

th (m

)D

epth

(m)

Dep

th (m

)N

(m-2)

N (m

-2)

N (m

-2)

00.10.20.3

00.10.20.3

00.10.20.3

f

e

b

d

c

a

Figure 4. Apparent oxygen utilization (colors in μM) and Thorpe displacements (T0) binned in 5 m depth intervals along b) Tran-sect 1 at 64° W, d) Transect 2 at 67° W and f) Transect 3 at 70° W (all transects are shown against latitude, i.e. white areas are outside of the transects). Length scale of T0 is shown in upper right/left corners and colors indicate units of 1 m (light gray) and 5 m (black), respectively. Isopycnals (σθ) are contoured in intervals of 0.5 kg m-3. Abundances of eel-larvae, redrawn from the study of Munk et al. 2010 (N is number of larvae pr. m-2) in the depth interval between 50 -200 m along each transect are shown in a, c and e. Net samples without eel-larvae are marked with a bullet. Two species of eel-larvae are shown by red (A. Anguilla) and blue (A. rostrata).



The suggestion in the distribution of Thorpe displace-ments that mixing intensities in the frontal region on Transect 1 were larger than those encountered in the fron-tal region on the other two transects is supported by the relatively low Richardson numbers (Ri < 10-2) recorded here. These low values result from weak vertical stratifi -cation (i.e. a low Brunt Väisälä frequency) and relatively large vertical velocity shear below the mixed layer. In the upper 120 m, the locations with low Richardson numbers were in good agreement with areas characterized by large Thorpe displacements (Fig. 4, Table 2). The situation is somewhat different for depths below 100 m. These were, in general, characterized by relatively high Richardson numbers which would indicate the occurrence of less turbulent mixing, while the Thorpe displacements indi-cated signifi cant mixing deeper in the water column. We note, however, that current velocities below 100 m were generally less than 1 cm s-1 and that under these condi-tions, deep vertical variations in currents become diffi cult to detect (i.e. a low signal to noise ratio). In contrast to Transect 1, Transects 2 and 3 were, in general, character-ized by relatively large Richardson numbers in the entire water column below 50 m (Table 2). This distribution of Richardson numbers indicates a much lower infl uence from turbulent mixing in the upper ocean in the frontal area on these two transects compared to that encountered on Transect 1.

Turbulent diffusion coeffi cients (kv) were estimated by applying the KPP-parameterisation in the frontal re-gions of the three transects. Low Ri-values at Transect 1 resulted in large diffusion coeffi cients of about 10-3 m2 s-1, whereas relatively low diffusion coeffi cients of about 10-5 m2 s-1 were found on transects 2 and 3. The corresponding vertical nutrient fl uxes (converted into carbon units) were calculated from the vertical nutri-ent gradients and the diffusion coeffi cients. The largest vertical diffusion coeffi cients in the frontal zone were recorded on Transect 1 (Table 2).

At all stations, inorganic nutrients were low in surface waters: Nitrate in the upper 200 m was < 3.64 μmol kg-1 (mean 0.49 μmol kg-1); phosphate < 0.10 μmol kg-

1(mean = 0.03 μmol kg-1) and silicate between 0.45 and 1.22 μmol kg-1 (mean 0.70 μmol kg-1). The nutricline was located at around 200 m over most of the study area. When the 200 m sample is excluded, nitrate con-centrations in surface waters were < 1.69 μmol kg-1

(mean 0.22 μmol kg-1) and phosphate < 0.08 μmol kg-1 (mean 0.03 μmol kg-1). In the strongly stratifi ed waters farthest south, the low nutrient concentrations extended to even deeper waters, i.e. to between 200 and 400 m, than was the case for the rest of the stations along the transect. The vertical gradients of DIN and DIP between the DCM and 200 m depth varied between (-4.23) · 10-2

25

26

25

26

2526

26

18 20 22 24 26 28 30 32 34Latitude (°C)

150

200

100

50

150

200

100

50

150

200

100

50

Dep

th (m

)D

epth

(m)

Dep

th (m

)1

0

-1

-2

-3

-4

-5

Figure 5. Distribution of the Richardson number along a) Transect 1 at 64° W, b) Transect 2 at 67° W and c) Transect 3 at 70° W. Isopycnals (σθ) are contoured in intervals of 0.5 kg m-3. ADCP-measurements binned in 16 m intervals are shown with small dots.



– (-0.98) · 10-2 μM m-1 and (-6.31) · 10-4 – (-1.41) · 10-4

μM m-1, respectively. The weakest nitrate gradient was observed below the DCM at station 28 in accordance with the larger mixing intensity at this station.

A distinct sub-surface deep-chlorophyll a maximum (DCM) was observed on all transects with typical chl a values between 0.2-0.4 μg l-1 located at about 120 m and between the 25-26 kg m-3 isopycnal (σθ) with the highest chlorophyll concentrations in the DCM being observed north of 27.5° N (e.g. Fig. 2d). South of the front and in the frontal region, the DCM was located between 100 and 150 m. Primary production estimates indicated that between 18 and 87 % of the total water column primary production (PP) was generated in these DCM (mean 50 ± 22%). The vertical attenuation coeffi cient, k (Ed=E0 e-kd, where Ed is the irradiance at a given depth (d) and at the surface (E0), respectively) varied between 0.02 and 0.06 in the study area. Thus, between about 0.2 and 13 % of surface light could be expected to penetrate to 100 m. At the stations north of the front on all transects, the depth of the DCM decreased but remained below 50 m.

There was considerable variability in the water column particulate primary production (PP) recorded in the study area as a whole (Table 3), This is not surprising given that at least three different water regions were be-

Table 2. Density and vertical stratifi cation characteristics in the frontal region of the three transects. Density (σθ), Brunt Väisälä frequenzy (N), Thorpe displacements (LT) and velocity speed (V) are averaged in the depth interval between 70 and 120 m and the minimum Richardson number in this depth interval is also shown (min(Ri)). Vertical nutrient gradients of DIN and DIP are deter-mined between the depth of the chlorophyll maximum and 200 m. The average diffusion coeffi cient (<Kv>), based on the KPP-pa-rameterisation, is applied for estimating the corresponding vertical DIN fl ux and, via the Redfi eld ratio, this is converted to car-bon units (Fc(DIN).

Station σθ(kg m-3)

N10-2 (s-1)

LT(m)

V(m s-1)

min(Ri) DINz10-2 (μM m-1)

DIPz10-4 (μM m-1)

<Kv>(10-4 m2 s-1)

FC(DIN)(mg C m2 d-1)

Interval (m) 70-120 70-120 70-120 70-120 70-120 dcm-200 dcm-200 70-120 ~ DCM

Transect 1

17-22 25.402 0.81 0.03 0.19 1.9 –2.66 –6.07 0.13 2,4

17-25 25.408 0.66 1.18 0.11 <10–6 – – 11.1

17-28 25.308 0.36 8.87 0.12 <10–6 –0.98 –4.90 10.3 69,3

Transect 2

17-61 25.988 0.59 0.17 – – –2.13 –6.25 –

17-63 25.796 0.63 0.05 0.10 1.7 – – 0.14

17-65 25.659 0.94 0.11 0.08 8.4 – – 0.11

17-68 25.770 1.0 0.02 0.07 21.4 – – 0.10

17-72 25.637 1.1 0.00 0.06 13.9 –4.23 –3.62 0.11 3,2

Transect 3

17-79 24.934 0.77 0.00 0.09 6.6 –1.49 –1.41 0.12 1,2

17-80 25.201 1.1 0.03 0.06 6.3 – – 0.11

17-83 25.553 1.0 0.04 0.11 31.6 – – 0.10

17-87 25.341 0.90 0.16 0.21 1.3 – – 0.16

17-94 25.450 0.84 0.02 0.10 1.1 –3.90 –6.31 0.22 5,9

0

0.15

0.2

0.25

>0.3

-72° -68° -64° -60°

18°

21°

24°

27°

30°

33°

Figure 6. Jaccard Index of diatoms at the surface (10 m) versus the DCM. The Jaccard index is shown in color. Positions of sta-tions without observations of phytoplankton species are indicat-ed together with isolines of SST (symbols and SST as in Fig. 1).



ing sampled (i.e front and regions both to the north and south). There were too few samples made in total within the frontal region to detect signifi cant differences be-tween the estimates. However, we note that the highest (329 mg C d-1) of the six primary production estimates made in the frontal region (i.e., 2 on each transect) was recorded at station 17-28 on Transect 1, i.e. the same station where the greatest vertical nutrient fl ux was es-timated.

The variable fl uorescence (Fv/Fm) measurements made at the DCM were signifi cantly higher (t test; p< 0.01) than the values from all other depths (Table 3). DCM values were relatively high (> 0.3) at all stations in the frontal region where data were available. However, only at the frontal stations on transect 1 did these values of > 0.3 extend to shallower depths (i.e., Fv/Fm of 0.39 at 30 m at station 17-22 and 0.34 at 60 m at station 17-28). This suggests that the phytoplankton communities

Transect 1

Station Depth(m)

Fv/Fm±SD PP[mg C m-2

day-1]

Int. chlor[mg Chl

m-²]

43 10 0.20±0.03 276 3360 0.26±0.0180 0.15±0.01100 0.15±0.01120

(DCM)0.13±0.01

Transect 2

55 10 0.26±0.03 131 2760 0.37±0.0180 0.42±0.01100

(DCM)0.44±0.01

170 0.31±0.02

Transect 3

103 10 0.27±0.001 470 2660 0.30±0.0183 0.35±0.04200 0.14±0.05

Table 3b). Stations north of frontal zone.

Table 3c). Stations south of frontal zone.

Transect 1

Station Depth(m)


day-1]

Int. chlor[mg Chl m-²]

1 10 0.21±0.05 322 2130 0.22±0.0560 0.21±0.01

10 10 0.24±0.05 197 21

60 0.26±0.04120

(DCM)0.23±0.01

200 0.10±0.003

Table 3a). Frontal zone stations. Mean and standard deviation (SD) of variable fl uorescence, Fv/Fm, at all sampled depths where valid data were recovered. Data for samples from the DCM are in bold. Three replicate samples (from the same Niskin bottle) were made at each depth and three measurements on each sample. Sample positions are given as in the frontal zone (i.e. in the region between the sur-face outcropping of the 24.5 and 25.5 kg m-3 isopycnals (σθ);Table 3a) north (Table 3b) or south (Table 3c) of the frontal zone. Total water column primary production (mg C m-2 d-1), PP and integrated water column chlorophyll (mg m-2), chl a are also noted for each station.

Transect 1

Station Depth(m)


day-1]

Int. chlor[mg Chl

m-²]

22 10 0.25±0.05 130 2430 0.39±0.02140

(DCM)0.48±0.01

200 0.19±0.01

28 10 0.28±0.02 329 2360 0.34±0.01100

(DCM)0.31±0.01

200 0.12±0.01

Transect 2

61 10 0.14±0.01 45 2960 0.15±0.0180 0.24±0.01200 0.07±0.03

72 10 0.19±0.01 184 2760 0.28±0.002130

(DCM)0.31±0.01

200 0.050±0.03

Transect 3

79 10 0.14±0.10 244 2260 0.30±0.02125

(DCM)0.39±0.17

200 0.16±0.01

83 10 0.035±0.0160 0.087±0.01125

(DCM)0.36±0.02

150 0.42±0.01

94 10 0.24±0.001 116 22 60 0.25±0.01

200 0.12±0.07



above the DCM were more nutrient replete in the frontal region of transect 1 than in the frontal region of the other two transects.

The structure of the chlorophyll-containing micro and nanoplankton community was similar at all stations (Ta-ble 4). Flagellates dominated the biomass and dinofl a-gellates were the second most important group. Togeth-er, these two groups comprised over 95% of the biomass at all stations and depths. Mixotrophic dinofl agellates were more abundant than autotrophic both at 10 m and at the DCM (t-test; p < 0.01). Diatoms contributed from 0.1 to 5% of the total biomass at all stations and depths. In all, 44 different diatom morphotypes were identifi ed in the study area.

No signifi cant difference was found in the average abun-dance of calanoid and cyclopoid copepods in the frontal region of Transects 1 and 2. However, the abundance of these organisms was signifi cantly lower in the fron-tal region of Transect 3 than on the other two transects (p<0.05 Kruskal-Wallis rank sum test). The abundance of nauplii appears lower on transect 3 compared to tran-sect 1 and 2 but the differences between the transects are not signifi cant (Kruskal-Wallis rank sum test). There

Table 4. Biomass (μg C l-1) of chlorophyll-containing protists larger than approximately 2 μm. “n” is number of stations. “Sur-face” is 10 m samples. DCM is Deep chlorophyll maximum.

Autotrophic Dino.

Mixotrophic Dino.

Diatoms Trichodesmium Mesodinium rubrum

Flagellates Biomass totals

Tran

sect

1

North Surface (n=6)DCM (n=5)

0.30±0.070.16±0.07

1.01±0.311.45±0.54

0.03±0.030.10±0.06

0.00±0.000.00±0.00

0.06±0.060.01±0.03

4.13±0.956.30±2.37

5.54±1.128.02±2.62

Front Surface (n=3)DCM (n=3)

0.29±0.250.11±0.06

0.95±0.061.48±0.35

0.04±0.050.05±0.05

0.00±0.000.00±0.00

0.00±0.000.00±0.00

6.25±1.242.93±0.87

7.52±1.504.58±0.59

South Surface (n=5)DCM (n=4)

0.25±0.140.12±0.07

0.76±0.321.03±0.25

0.06±0.030.05±0.01

0.07±0.100.10±0.19

0.00±0.000.00±0.00

5.49±1.573.10±1.52

6.63±2.004.40±1.63

Tran

sect

2


0.22±0.090.08

0.80±0.271.13

0.02±0.020.34

0.00±0.000.00

0.02±0.020.00

4.87±0.904.72

5.93±1.046.27


0.21±0.080.16±0.03

0.84±0.361.32±1.02

0.06±0.010.03±0.08

0.33±0.000.00±0.00

0.00±0.020.02±0.00

4.67±1.244.30±0.59

6.11±1.405.84±1.72

South Surface (n=1)DCM (n=1)

0.310.09

1.141.3

0.040.04

0.000.00

0.010.00

5.474.36

6.965.79

Tran

sect

3


0.22±0.150.17±0.06

1.18±0.641.12±0.05

0.05±0.030.07±0.00

0.00±0.000.00±0.00

0.06±0.070.00±0.00

5.96±5.492.98±0.82

7.48±6.024.34±0.93


0.20±0.080.11±0.06

0.83±0.190.95±0.43

0.04±0.010.08±0.02

0.00±0.000.00±0.00

0.00±0.000.04±0.09

3.40±2.322.46±1.34

4.46±2.413.63±1.18

South Surface DCM

__

__

__

__

__

__

__

Table 5. Total seston (mg m-2) to 400m at stations in the frontal region on Transects 1 and 3.

Integrated seston (0-400 m) mg m-2

Transect 1

Stn. 17-22 1627

Stn. 17-28 1318

Transect 3

Stn. 17-80 610

Stn. 17-94 844

Station Depth interval (m)

0-50 50-100 100-200 200-300 300-400

17-1 0.58 1.44 2.65 4.13 4.57

17-10 0.87 1.74 2.35 3.18 3.88

17-22 1.26 1.94 3.07 5.04 5.50

17-28 0.60 1.84 2.43 4.56 5.34

17-43 2.34 3.30 4.14 5.01 5.18

Table 6. Stable isotope abundance (δ15N, ‰) in particulate matter (seston) obtained in vertical hauls of a Multinet (50 μm mesh) along the transect at 64° W.



were two determinations of total seston in the upper 400 m in the frontal region on both Transects 1 and 3. These indicated that the seston concentration in the frontal re-gion of Transect 1 was approximately double that on Transect 3 (Table 5)

The δ15N signature of the seston was determined at 5 sta-tions on Transect 1. Average water column seston δ15N ranged from 1.97 ‰ to 3.64 ‰ along the transect ap-proaching the frontal region from the south and reached 4.46 ‰ in the colder waters north of the frontal region (Stn 17-43). This latitudinal trend for average water col-umn seston δ 15N was signifi cant (p = 0.006; R2 = 0.74). (Table 6).

Discussion

Most studies attempting to explain the association of par-ticular biological features with frontal regions compare conditions in with those outside of the frontal region. At-tempts (Andersen et al. 2011; Reimann et al. 2011) to apply this strategy in an effort to fi nd a possible explana-tion for the observation (Munk et al. 2010) that eel larvae were only found within the frontal region of the STCZ in the Sargasso Sea, were unsuccessful. Andersen et al. (2011) did fi nd an increased biomass of some zooplank-ton groups/species in the frontal region compared to the non-frontal stations but neither study found evidence for increased primary or secondary production at the frontal compared to the non-frontal stations. Using data collect-ed on the same cruise as the three earlier studies (Munk et al. 2010; Andersen et al, 2011; Reimann et al 2011), we adopted a different strategy and examined the frontal region, itself, for evidence of heterogeneity in physical and chemical conditions that might locally impact bio-logical production. Our results suggest that vertical mix-ing conditions and, consequentially, the magnitude of vertical nutrient fl uxes may vary considerably between stations in the frontal region and that local stimulation of the plankton food web may, therefore, occur at the sites within the frontal region where vertical nutrient fl uxes are greatest. Thus, we suggest that the planktonic food web may be stimulated by the introduction of nutrients at specifi c sites in the frontal regions rather than in the frontal zone as a whole.

The location of the front encountered on this study is in accordance with previous satellite derived observations of fronts in the STCZ (Ullman et al., 2007). The salin-ity maximum we observed may be associated with the Subtropical Underwater mass (STUW) which is formed by subduction in a large area across the central subtropi-cal gyres between 20-0° N and with highest salinities observed east of the frontal zone between 40-30° W

(O’Connor et al., 2005). O’Connor et al. (2005) found that the spreading of STUW in the area occurs at density (σθ) levels ranging between 25.6-26.3 kg m-3 in accord-ance with the observed sub-surface salinity maximum and that the average age of the subducted STUW, esti-mated from CFC-tracer distributions, is 1-5 years. The higher oxygen concentrations recorded in the STMW (Fig. 2c) are consistent with previous reports that the STMW is, in general, more ventilated than STUW in the area as, for example, seen from meridional sections of the CFC-12 tracer in the area, where high CFC-12 is observed below the sub-surface salinity maximum of STUW (Joyce et al., 2001).

Evidence for vertical mixing: Several lines of evidence presented here suggest that vertical mixing was greatest in the frontal region where Munk et al. (2010) record-ed the highest concentrations of eel larvae. Firstly, the highest similarities (Jaccard Index; Fig. 6) in the diatom communities at 10 m and the deep chlorophyll maxi-mum (> 100 m) were recorded in this frontal region of Transect 1. The statistically signifi cant relationship we found between the Jaccard Index and the difference in temperature between 10 m and at the DCM suggests that this index may be a reasonable proxy for mixing in this region.

The hydrographic data collected are not of suffi cient spatial and temporal resolution to directly observe indi-vidual meso-scale eddies. However, we examined data from each CTD station for evidence of vertical mixing at that specifi c geographic location. Relatively high verti-cal mixing intensities at intermediate depth levels down to 200 m, in particular at station 17-25, were derived from Thorpe displacements in the frontal zone of Tran-sect 1 (Fig. 4). These were considerably higher than the Thorpe displacements calculated for frontal stations on the other two transects. Estimates of turbulent vertical diffusion coeffi cients from Thorpe displacements have been shown to provide estimates of mixing intensities in accordance with mixing intensities derived from other methods (Cisewski et al., 2005), although they are as-sociated with relatively large uncertainties when based on few vertical CTD-profi les as is the case in this study where we only have simultaneous measurements from two independent CT-sensors.

A vertical diffusion coeffi ecient (kv) can be estimated from the Thorpe-displacements as: kv = η LT

2 N, where η ~0.2 is a mixing effi ciency (Cisewski et al., 2005). Thus, mixing intensity depends strongly on the Thorpe-displacements. From the depth-averaged values above the DCM (70-120 m) of LT and N across the three fron-tal zones, it can be seen that vertical mixing intensities are at least an order of magnitude larger on Transect 1 than on the other two transects (Table 2). This conclu-



sion of signifi cantly larger mixing intensities occurring on Transect 1 is supported by the distribution of the minimum Richardson number between 70-120 m depth (a minimum value of Ri, rather than a vertical average, is more representative for the presence of turbulence in a given depth layer). In general, low Ri-values (i.e. Ri < 0.25) indicate areas where the vertical velocity shear is expected to contribute signifi cantly to turbulent mixing. The lowest Ri-values in this intermediate depth interval are found at stations 25 and 28 on Transect 1 and these two stations are also characterized by the largest Thor-pe-displacements indicating substantial vertical mixing here compared to the other stations in the area.

Vertical nutrient fl uxes in the frontal region: In general, nutrient fl uxes (FDIN and FDIP) from vertical turbulent mixing can be calculated as: FDIN = –kv DINz, (and cor-respondingly for FDIP). Thus, while the deep vertical gradient of DIN between the DCM and 200 m is slightly smaller at Transect 1 when averaged over the frontal re-gion than on the other two transects, the total nutrient fl ux is expected to be an order of magnitude larger in the frontal region on Transect 1 due to the larger mixing intensities (Table 2).

By applying the estimated turbulent diffusion coeffi -cients from the long-term averaged (one hour) velocity shears, the mixing intensity between 70-120 m depth in the frontal region at Transect 1 was two orders of mag-nitude larger than corresponding values from the two other transects. When combining these estimates with vertical nutrient gradients below the DCM, the averaged nutrient fl ux corresponded to 36 mg C m-2 d-1 from the two stations in the frontal region of Transect 1 (Table 2). Thus, a signifi cant fraction of PP could be explained by vertical mixing of nutrients here. Comparison between the vertical gradients of DIN and DIP indicates that ex-cess nitrate for new production (sensu Dugdale and Go-ering, 1967), i.e. DINz/DIPz > 16, is being transported by vertical mixing to the euphotic zone in the frontal zone of Transect 1.

Eddy-induced mixing in the subtropical gyre has ear-lier been found to play an important role for new pro-duction (NP). By forcing a conceptual model of eddy-induced nutrient mixing with satellite observations of the sea surface variability, McGillicuddy et al. (1998) showed that mesoscale eddy activity had a major impact on the nutrient balance and could explain a signifi cant part (up to 40 %) of the N-requirement for sustaining the annual NP in the Saragasso Sea. These fi ndings were supported by eddy-permitting high-resolution modeling where it was shown that eddy-induced mixing was an important source for nutrients in the upper ocean (ac-counting for about 30 % of NP) in the sub- and mid-latitude North Atlantic (Oschlies and Garçon, 1998). A

study of the distribution of phytoplankton, in particular the diazotrophic Trichodesmium spp., in a comparable frontal area in the central north Pacifi c (26° N), showed that sub-mesoscale and mesoscale (~50 km) dynamics could explain carbon export and particle accumulation in the upper ocean although the specifi c eddy-dynamical processes supporting the nutrient transport remain unre-solved (Guidi et al., 2012).

In the study reported here, the similarity of the plankton populations at 10 m and the DCM, the calculated Thorpe displacements and the mixing intensities inferred from the vertical current shear all support the hypothesis that vertical mixing can play a signifi cant role for sustain-ing the PP in the frontal zone along Transect 1, where the abundance of eel larvae was greatest. We note, how-ever, that horizontal mixing associated with small-scale frontal dynamics may also be important for mixing and transport in the region. In a study of the North Pacifi c STFZ, it was shown that interleaving across a frontal zone was very intense within a distance of about 5 km from the front (Shcherbina et al. 2009). Such small scale features are not resolved by the station spacing in this study but such mixing processes may enhance vertical nutrient fl uxes further.

Relatively small Thorpe displacements at intermediate depths were, in general, recorded south of the frontal zone. The southernmost station (stn 17-1; Fig. 4) was characterized by a relatively large mixing intensity and this may be related to the dynamic region associated with the Antilles Current (Lee et al., 1996). Relatively modest mixing intensity was recorded north of the fron-tal zone, except for station 17-36; located just north of the frontal zone defi ned here, where high mixing was observed. Two lines of biological evidence further sup-port the hypothesis that relatively large vertical nutrient fl uxes were occurring in the frontal zone of Transect 1:

Biological activity in frontal regions of elevated vertical nutrient fl ux: Although the data on biological variables collected within the frontal region are, generally, too few to allow statistically signifi cant identifi cation of dif-ferences in plankton production characteristics between the different frontal stations, several lines of evidence suggest heterogeneity in the plankton production pro-cesses within the frontal region and that the elevated vertical nutrient fl uxes implied in the frontal region on Transect 1 infl uence plankton productivity here:

• The highest primary production estimate recorded in the frontal zone was found at station 17-28, i.e. the same station where maximum vertical nutrient fl ux is implied (Table 2).

• Relatively high values of variable fl uorescence, Fv/Fm, (which is considered to be an indicator of the



potential for electron transport in PSII) recorded by Fast Repetition Rate Fluorometry were found at the DCM over the entire study area. While several factors can infl uence variable fl uorescence, it is generally as-sumed that, for dark adapted cells, Fv/Fm, is an indi-cator of nutrient status (see review by Beardall et al. 2001). This suggests that phytoplankton in the DCM are, generally, more nutrient replete than those higher in the water column and, thus, that the phytoplankton here are fuelled by nutrient sources not available for phytoplankton higher in the water column. The only obvious potential nutrient source that would not also be available for phytoplankton higher in the water col-umn would be nutrients from the deep water below. The fact that the relatively high Fv/Fm values record-ed at the two frontal stations on Transect 1 extend to shallower depths (Table 3) than on any of the frontal stations on Transects 2 and 3 suggests that also the phytoplankton above the DCM in the frontal region of this transect are benefi ting from a sub-surface nu-trient source, again lending support to the hypothesis that greater vertical mixing of nutrients is occurring in the frontal region of Transect 1 than on the two other transects. We acknowledge that one of the factors that can infl uence Fv/Fm in addition to nutrient status is species composition (Suggett et al. 2009). However, the structure of the chlorophyll containing plankton community was very similar over the entire study area (Table 4). Although we only present data here for the chlorophyll-containing nano and microplankton (i.e. those plankton that can be observed in the light micro-scope and, > ≈2 μm), Riemann et al (2011) presented fl owcytometer data describing the distribution of pi-coplankton on the same stations. They found Prochlo-rococcus to dominate the picoplankton biomass and to be especially dominant in the DCM. Synecococcus spp. were also present but generally more abundant higher in the water column. They found no signifi cant horizontal patterns in the abundance of picoplankton or the distributions of these groups. They estimated the biomasses of the picoplankton and what they called “larger algae” (i.e. those we consider here) to be on the same order of magnitude at most stations. Thus, changes in community composition would seem unlikely to be responsible for the differences in Fv/Fm we report here.

• There are very few determinations of total seston but the two taken in the frontal region of Transect 1 com-pared to the two taken in the frontal region of Transect 3 suggest that total seston was approximately twice as high on Transect 1 than on Transect 3 (Table 5). The abundance of calanoid and cyclopoid copepods was also signifi cantly lower on Transect 3 than on the other two transects. Thus, while the seston and meso-zoolankton data do not specifi cally identify Transect 1 as having more secondary production than the two

other transects, they do indicate that there are sig-nifi cant differences in production patterns within the frontal region and, thus, support our overall hypoth-esis of heterogeneity in the plankton food web within the frontal region. One can imagine that vertical mix-ing of nutrients may be an ephemeral occurrence in this region. As a response in secondary production to vertical nutrient mixing must operate through the re-sponse of phytoplankton to the nutrient infl ux, a time lag is to be expected in the secondary production re-sponse in relation to the timing of the nutrient infl ux event. Thus, our inability distinguish differences in the abundance of calanoid and cyclopoid copepods between Transects 1 and 2 may be related to the previ-ous history of fl ux events and not the fl uxes estimated at the time of our sampling.

• Also the δ15N distributions associated with seston along Transect 1 support the hypothesis that there may be increased vertical mixing of nutrients into the frontal region compared to the waters to the south. Atmospheric nitrogen has, by defi nition, a δ15N of 0 while that of deep nitrate is about 5. Thus, the relative contribution of these sources of ni-trogen is refl ected in the isotope signature of food web components (Montoya et al., 2002). Although absolute values will be affected by trophic enrich-ment, a higher δ15N will generally indicate a lesser importance of atmospheric nitrogen in fuelling the planktonic food web. Thus, the higher values of δ15N in seston recorded at frontal stations on Transect 1 compared to the more southerly stations on this tran-sect again suggest vertical mixing of nutrients to be occurring here (Table 6).

Phytoplankton community composition: A striking fea-ture in the plankton community data is the dominance of mixotrophic dinofl agellates among the larger organ-isms present (Table 4). This is the case at 10 m but, even more so, at the DCM. Collos et al. (2013) have recently reported on a mixotrophic dinofl agellate that, under cer-tain grazing conditions, can base over 50 % of its growth (measured as carbon accumulation) on non-autotrophic sources. Given the importance of mixotrophic dino-fl agellates in the Sargasso Sea plankton ecosystem, it seems unlikely that carbon accumulation at the base of the food web here is completely reliant on autotrophic processes. Thus, traditional 14C estimates of primary production may, potentially, seriously underestimate the actual carbon available to the plankton food web here.

Implications for understanding the plankton food web in the Sargasso Sea: This study suggests that elevated vertical mixing rates may occur in association with spe-cifi c hydrographic features within the frontal region of the STCZ in the Sargasso Sea. Several different types of biological measurements indicate that there may be



increased primary production in the plankton commu-nities found at the sites where elevated vertical mixing of nutrients is implied. We note that the greatest abun-dances of eel larvae reported by Munk et al. (2010) were found in association with a site within the frontal zone region where this study indicates elevated rates of verti-cal nutrient transport compared to the rest of the frontal region. Whether these specifi c hydrographic features are important for eel larvae ecology and, if so, whether the larvae only survive in or near these features or whether they accumulate (or hatch) in the vicinity of these fea-tures remains to be determined.

Acknowledgements

The present investigation was supported by Nordea Fund, Villum Kann Rasmussen Fund, The Carlsberg Foundation, Knud Petersen Fund, The Danish Expedi-tion Fund, The Danish Research Council for Nature and Universe, The Danish Research Foundation and our re-spective institutions. The data were collected as part of the Galathea3 expedition under the auspices of the Dan-ish Expedition Foundation. This is Galathea3 contribu-tion no. XXX.

References

Andersen, N.G., Nielsen, T.G., Jakobsen, H.H., Munk, P., Riemann, L. (2011). Distribution and production of plank-ton communities in the subtropical convergence zone of the Sargasso Sea. II. Protozooplankton and copepods. Mar. Ecol. Prog. Ser. 426:71-86.

Arar, E.J., Collins, G.B. (1997). Method 445.0 In Vitro De-termination of Chlorophyll a and Pheophytin a in Marine and Freshwater Algae by Fluorescence. U.S. Environmental Pro-tection Agency. (http://www. epa.gov/microbes/m445_0.pdf)

Beardall, J., Young, E., Roberts, S. (2001). Approaches for determining phytoplankton nutrient limitation. Aquat.sci. 63:44-69.

Cisewski, B., Strass, V.H., Prandke, H. (2005). Upper-ocean vertical mixing in the Antarctic polar front zone. Deep-Sea Res. II, 52:1087-1108.

Collos, Y., Jauzein, C., Laabir, M., Vaquer, A. (2013). Dis-crepencies between net particulate carbon production and 13C-labelled bicarbonate uptake by Alexandrium catonella (DINOPHYCEAE): grazing controls the balance between autotrophic and non autotrophic carbon acquisition. J. Phy-col. 49(3):441-446.

Cushman-Roisin, B. (1984). On the maitanance of the sub-tropical front and its associated countercurrent. J. Phys. Oceanogr. 14:1179-1190.

Diekmann, R., Piatkowski, U. (2002). Early life stages of cephalopods in the Sargasso Sea: distribution and diversity re-lative to hydrographic conditions. Mar. Biol. 141(1):123-130.

Dugdale, R.C., Goering, J.J. (1967). Uptake of new and re-generated forms of nitrogen in marine productivity. Limnol. Oceanogr. 12:196-206.

Edler, L. (ed.). (1979). Recommendations for marine biologi-cal studies in the Baltic Sea.Phytoplankton and chlorophyll. The Baltic Marine Biologists Publ. No 5:1-38.

Eppley, R.W., Peterson, B.J. (1979). Particulate organic mat-ter fl ux and planktonic new production in the deep ocean, Nature 282:677-680.

Ferron, B., Mercier, H., Speer, K (1998). Mixing in the Ro-manche fracture zone. J. Phys. Oceanogr. 28:1929-1945.

Galbraith, P.S., Kelley, D.E. (1996). Identifying overturns in CTD profi les. J. Atmos. Oceanic Technol. 13:688-702.

Goericke, R., Welschmeyer, N. (1998). Response of Sargasso Sea phytoplankton biomass, growth rates and primary pro-duction to seasonally varying physical forcing. J. Plankton Res. 20:2223-2249.

Grasshoff, K., Ehrhardt, M., Kremling, K. (1983). Methods of seawater analysis, 2nd ed. Weinheim.

Guidi, L., et al. (2012). Does eddy-eddy interaction con-trol surface phytoplankton distribution and carbon export in the North Pacifi c Subtropical Gyre?, J. Geophys. Res. 117:G02024, doi:10.1029/2012JG001984.

Halliwell, G.R., Peng, G., Olson, D.B. (1994). Stability of the Sargasso Sea Subtropical Frontal Zone. J.Phys. Oceanogr. 24:1166-1183.

Hilligsøe, K.M., Richardson, K., Bendtsen, J., Sørensen, L., Nielsen T.G., Lyngsgaard, M.M. (2011). Linking phytoplank-ton community size composition with temperature, plankton food web structure and sea-air CO2 fl ux. Deep-Sea Res. I 58:826-838.

Jaccard, P. (1901). Étude comparative de la distribution fl o-rale dans une portion des Alpes et des Jura. Bulletin de la Société Vaudoise des Sciences Naturelles 37:547-579.

Joyce, T.M., Hernandez-Guerra, A., Smethie Jr., W.M. (2001). Zonal circulation in the NW Atlantic and Caribbean from a meridional World Ocean Circulation Experiment hydrographic section at 66 ºW. J. Geophys. Res. 106:C10, 22095-22113.

Klecker, R.D., McCleave, J.D., Wippelhauser, G.S. (1983). Spawning of American eel, Anguilla-Rostrata, relative to ther-mal fronts in the Sargasso Sea. Env. Bio. Fish. 9:289-293.

Klecker, R.D., McCleave, J.D. (1988). The northern limit of spawning by Atlantis eels (Anguilla spp.) in the Sargasso Sea in relation to thermal fronts and surface-water masses. J. Mar. Res. 46:647-667.

Kolber, Z.S., Prasil, O., Falkowski, P.G. (1998). Measure-ments of variable chlorophyll fl uorescence using fast repeti-tion rate techniques: defi ning methodology and experimental protocols. BBA-Bioenergetics. 1367:88-106.



Kubokawa, A., Inui, T. (1998). Subtropical countercurrent in an idealized ocean GCM. J. Phys. Oceanogr. 29:1303-1313.

Leetmaa, A., Voorhis, A.D. (1978). Scales of motion in the sub-tropical convergence zone. J. Geophys. Res., 83 C9:4589-4592.

Lee, T.N., Johns, W.E., Zantopp, R.J., Fillenbaum, E.R. (1996). Moored observations of western boundary current variability and thermohaline circulation at 26.5 ºN in the sub-tropical North Atlantic. J. Phys. Oceanogr. 26:962-983.

McGillicuddy, Jr. D.J., Robinson A.R., Siegel D.A., Jannasch H.W., Johnson, R., Dickey, T.D., McNeil, J., Michaels, A.F., Knap A.H. (1998). Infl uence of mesoscale eddies on new production in the Sargasso Sea. Nature 394:263-266.

MODE group. The Mid-Ocean Dynamics Experiment. (1978). Deep-Sea Res. 25:859-910.

Montoya, J.P., Carpenter, E.J., Capone, D.G. (2002). Nitro-gen fi xation and nitrogen isotope abundances in zooplank-ton of the oligotrophic North Atlantic. Limnol. Oceanogr. 47:1617-1628.

Munk, P., Hansen, M.M., Maes, G.E., Nielsen, T.G., Caston-guay, M., Riemann, L., Sparholt, H., Als, T.D., Aarestrup, K., G. Andersen, N.G., Bachler, M. (2010). Oceanic fronts in the Sargasso Sea control the early life and drift of Atlantic eels. Proc. R. Soc. B 277:3593-3599.

O’Connor B.M., Fine, R.A., Olson, D.B. (2005). A global comparison of subtropical underwater formation rates. Deep-Sea Res. 52:1569-1590.

Oschlies, A., Garçon, V. (1998). Eddy-induced enhance-ment of primary production in a model of the North Atlantic Ocean, Nature 394:266-269, doi:10.1038/28373.

Pacanowski, R.C., Philander, S.G.H. (1981). Parameteriza-tion of vertical mixing in numerical models of the tropical ocean. J. Phys. Oceanogr. 11:1443-1451.

Riemann, L., Nielsen, T.G., Kragh, T., Richardson, K. Parner, H., Jakobsen, H.H, Munk, P. (2011). Distribution and production of plankton communities in the subtropical convergence zone of the Sargasso Sea. I. Phytoplankton and bacterioplankton. Mar. Ecol. Prog. Ser. 426:57-70.

Shcherbina, A.Y., Gregg, M.C., Alford, M.H., Harcourt, R.R. (2009). Characterizing thermohaline intrusions in the North Pacifi c Subtropical Frontal Zone. J. Phys. Oceanogr., 39:2735-2756.

Steemann Nielsen, E. (1952). The use of radio-active carbon (C14) for measuring organic production in the sea. Journal du conseil. 18:117-140.

Suggett D.J., Moore, C.M., Hickmann, A.E., Geider, R.J. (2009). Interpretation of fast repetition rate (FRR) fl uores-cence: signatures of phytoplankton community structure ver-sus physiological state. Mar. Ecol. Prog. Ser. 376:1-19

Thorpe, S.A. (1977). Turbulence and mixing in a Scottish Loch. Philos. Trans. Roy, Soc. London, Ser A, 286:125-181.

Ullman, D.S., Cornillon, P.C., Shan, Z. (2007). On the char-acteristics of subtropical fronts in the North Atlantic. J. Geo-phys. Res. 112:C01010, doi:10.1029/2006JC003601.

Utermöhl, H. (1958). Zur Vervollkommnung der quantita-tiven Phytoplanktonmethodik. Mitt. Int. Ver. Theor. Angew. Limnol. 9:1-38.

Voorhis, A.D., Hersey, J.B. (1964). Oceanic Thermal Fronts in Sargasso Sea. J. Geophys. Res. 69:3809-3814 .


Deep-Sea Research I82 (2013) 60–71, http://dx.doi.org/10.1016/j.dsr.2013.07.013.

Photo: Katherine Richardson.

T.A. Rynearsona,*, K. Richardsonb, R.S. Lampittc, M.E. Sierackid, A.J. Poultonc, M.M. Lyngsgaardb, M.J. Perrye

aGraduate School of Oceanography, University of Rhode Island, Narragansett, RI, USAbCenter for Macroecology, Evolution and Climate, University of Copenhagen, Copenhagen, DenmarkcOcean Biogeochemistry and Ecosystems, National Oceanography Centre, Southampton, UKdBigelow Laboratory for Ocean Sciences, East Boothbay, ME, USAeDarling Marine Center, School of Marine Sciences, University of Maine Walpole, ME, vUS

*Corresponding author

PAPER VMajor contribution of diatom resting spores to vertical fl ux in the sub-polar North Atlantic


Running head

Major contribution of diatom resting spores to vertical fluxin the sub-polar North Atlantic

T.A. Rynearson a,n, K. Richardson b, R.S. Lampitt c, M.E. Sieracki d, A.J. Poulton c,M.M. Lyngsgaard b, M.J. Perry e

a Graduate School of Oceanography, University of Rhode Island, Narragansett, RI, USAb Center for Macroecology, Evolution and Climate, University of Copenhagen, Copenhagen, Denmarkc Ocean Biogeochemistry and Ecosystems, National Oceanography Centre, Southampton, UKd Bigelow Laboratory for Ocean Sciences, East Boothbay, ME, USAe Darling Marine Center, School of Marine Sciences, University of Maine, Walpole, ME, USA

a r t i c l e i n f o

Article history:Received 14 January 2013Received in revised form19 July 2013Accepted 29 July 2013Available online 13 August 2013

Keywords:AggregatesCarbon fluxChaetocerosDiatom resting sporesFloating sediment trapsNorth Atlantic spring bloom

a b s t r a c t

The mass sinking of phytoplankton cells following blooms is an important source of carbon to the ocean'sinterior, with some species contributing more to the flux of particulate organic carbon (POC) than others.During the 2008 North Atlantic Bloom Experiment in the Iceland Basin, we examined planktoncommunity composition from surface waters and from sediment traps at depths down to 750 m.Samples collected with neutrally buoyant Lagrangian sediment traps captured a major flux event.Diatoms comprised Z99% of cell flux into the sediment traps, with vegetative cells and resting spores ofthe genus Chaetoceros contributing 50–95% of cell flux. Resting spores of one species, identified asChaetoceros aff. diadema, were dominant, comprising 35–92% of cell flux. The flux of resting sporesranged from 2 to 63 mg C m�2 day�1 and was significantly correlated with POC flux (p¼0.003). Over thecourse of 10 days, the flux of resting spores increased by 26 fold, suggesting that the cells sank en masse,possibly in aggregates. In contrast, vegetative cells of C. aff. diadema sampled from surface waters duringthe period preceding the flux event generally comprised o1% of the diatom community and neverexceeded 5.2%. Resting spores of C. aff. diadema were rarely observed in surface waters but theirconcentrations increased with depth (to 200 m) below the mixed layer. This increase in resting sporeabundance, coupled with increased dissolved silicic acid concentrations at depth, suggest that themorphological changes associated with spore formation may have occurred in the mesopelagic zone,while cells were sinking. The values of variable fluorescence (Fv/Fm) measured on sediment trap materialdominated by resting spores were among the highest values measured in the study area at any depth.This, in combination with the rapid germination of resting spores in ship-board incubations, suggeststhat vegetative cells were not physiologically stressed during spore formation. The degradation-resistant,heavily silicified resting spore valves explain the high relative contribution of C. aff. diadema restingspores to total plankton carbon at depth. These data emphasize the ephemeral nature of organic carbonflux events in the open ocean and highlight how non-dominant species and transient life stages cancontribute more to carbon flux than their more abundant counterparts.

& 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The biological pump, defined as the process by which surface-generated particulate organic carbon (POC) sinks to depth, has astrong influence on global carbon cycling (e.g., Ducklow et al.(2001), Volk and Hoffert (1985)) and atmospheric CO2 concentra-tions (e.g., Kohfeld et al. (2005), Takahashi et al. (2009)). Overall,the biological pump transfers 1–3% of oceanic primary production

to the deep sea and sediments (De La Rocha and Passow, 2007;DeVries et al., 2012). The magnitude of POC that is successfullytransferred from surface layers to the inner reaches of the oceandepends upon several factors, including the aggregation anddisaggregation of particles and organisms, microbial remineraliza-tion, and grazing and fecal pellet production by zooplankton(reviewed in De La Rocha and Passow (2007), Ragueneau et al.(2006)). The biological pump is controlled, in part, by the taxo-nomic and mineral composition of sinking organisms (Legendreand Rivkin, 2002). For example, diatoms and their silica frustulesare thought to contribute to pulses of phytodetritus that in some

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/dsri

Deep-Sea Research I

0967-0637/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.dsr.2013.07.013

n Corresponding author. Tel.: +1 401 874 6022.E-mail address: [email protected] (T.A. Rynearson).

Deep-Sea Research I 82 (2013) 60–71

126

Running head

PhD thesis by Maren Moltke Lyngsgaard

locations can deliver the equivalent of the annual average carbonflux to the benthos within days to weeks (Beaulieu, 2002).

Diatoms represent one of the most productive groups ofphotosynthetic organisms on earth, generating 30–40% of globalmarine primary productivity each year (Mann, 1999; Nelson et al.,1995). The relative dominance of diatoms in many surface phyto-plankton communities is thought to play an important role inregulating the magnitude and efficiency of the biological pump(Buesseler, 1998; Goldman, 1993; Pondaven et al., 2000;Sarmiento, 2006). Even in oligotrophic regions where diatomsare not generally abundant, material of diatom origin frequentlydominates POC recovered from depth (Karl et al., 2012; Schareket al., 1999). A large part of diatom production is generated duringblooms, i.e., periods of rapid growth and accumulation of cells,with the most conspicuous of these blooms occurring in temperateand sub-polar waters during spring (Longhurst, 1998). At the endof these blooms, diatoms can contribute to the biological pump byrapidly sinking out of the surface layer, at rates of 100–150 m day�1 (Billett et al., 1983), and arriving at the seafloor asrelatively intact and sometimes viable cells (Billett et al., 1983;Cahoon et al., 1994; Smith et al., 1996).

Several mechanisms are believed to contribute to the initiation ofdiatom sinking during a bloom's demise, including nutrient limitationand formation of aggregates (reviewed in Smetacek (1985)). At theonset of nutrient limitation, diatoms undergo senescence and losetheir ability to maintain neutral buoyancy (e.g., Waite et al. (1997)).Many bloom-forming diatoms then produce transparent exopolymerparticles (TEP), which increase the probability that colliding cells orchains will stick together upon contact and form aggregates (Alldredgeand Gotschalk, 1989; Kiørboe and Hansen, 1993; Passow andAlldredge, 1995). Although several phytoplankton taxa can produceTEP, it appears that only diatoms are capable of producing the largequantities of TEP that drive aggregation (Alldredge et al., 1993). HighTEP production and aggregate formation tends to occur particularlywhen cells are nutrient limited and under conditions of intermediateturbulence (Alldredge and Gotschalk, 1989; Alldredge et al., 1993;Kiørboe and Hansen, 1993). Sinking POC also is susceptible tosubstantial bacterial degradation, hence both sinking rate and suscept-ibility to remineralization can dramatically influence the amount andquality of organic carbon that reaches mesopelagic waters and below(reviewed in De La Rocha and Passow (2007)).

Here, we determined the species composition of plankton cellsrecovered from both surface waters and neutrally buoyant sedimenttraps deployed in the sub-polar North Atlantic during a springdiatom bloom to examine how species composition and life historystage influence the magnitude and efficiency of POC flux. Thephysiological status of both bulk sediment trap material and indivi-dual cells was examined to better understand the nature and viabilityof sinking cells. We found that diatoms dominated the planktoncommunity both in the sediment traps and surface waters, but thatthe relative abundances of different species varied greatly. Sedimenttraps were dominated by heavily-silicified resting spores that germi-nated rapidly when brought into culture indicating that the POCtransported to depth was labile. Our results allowed us to identifythat a subset of species from the surface phytoplankton communitycombined with life cycle stage (i.e., resting spores) significantlyinfluenced the magnitude and efficiency of POC flux associated withthe North Atlantic spring diatom bloom.

2. Methods

2.1. Sampling overview

All samples were collected during the North Atlantic BloomExperiment cruise on the R/V Knorr, 2–20 May 2008, yearday (YD)

123–141. Shipboard samples were collected in the Iceland Basin(Fig. 1), in conjunction with a passively drifting, mixed-layerLagrangian float (Alkire et al., 2012; Briggs et al., 2011;Mahadevan et al., 2012; Martin et al., 2011).

2.2. PELAGRA sediment traps

Four deployments of neutrally buoyant, Lagrangian sedimenttraps (PELAGRA; Lampitt et al., 2008) were made between YD 126and 137 (Fig. 1, Table 1). The traps collected sinking material atdepths between 140 and 750 m and time periods of 15–72 h, for atotal of 13 independent samples (depth and time). Each PELAGRAtrap had four collection funnels (0.115 m2 each) with attachedcollection cups. As described in Martin et al. (2011), traps weredeployed in isopycnal mode and collection cups were pro-grammed to open 24 h after deployment and to close minutesbefore the trap ascended to the surface. Collection times weredeliberately varied to assess optimal deployment periods (Table 3).For each deployment, both preserved and live trap material wasrecovered. Live material was recovered from collection cups filledwith seawater obtained from below 400 m depth. Preservedmaterial was recovered from collection cups that additionallyhad a final concentration of 0.5% NaCl and 2% formaldehydebuffered with sodium tetraborate (Na2B4O7 �10 H20). After recov-ery, 1 mL of 40% buffered formaldehyde was added to the contentsof each preserved cup.

2.3. Phytoplankton taxa and carbon content in PELAGRA traps

Species composition and abundance for each PELAGRA deploy-ment were determined using 1 mL subsamples from the preservedtraps, a Sedgwick Rafter slide and quantitative light microscopyaccording to Utermöhl (1958). Species were identified followingSunesen et al. (2008) and Tomas (1997). Detailed size measurements

Fig. 1. Location of surface water samples (circles) and four PELAGRA floatingsediment trap deployments (crosses). Shading indicates depth intervals in meters.Map of sampling locations is shown as a shaded box in the inset map of the NorthAtlantic.

Table 1Date, depth and location of PELAGRA sediment trap deployments. Number of trapsdeployed on a given date varies; see Table 3.

Date(2008)

Yearday Deployment Depths (m) Latitude(1N)

Longitude(1W)

5 May 126 1 140, 230 60.82 27.177 May 128 2 160, 340, 620 61.08 26.6512 May 133 3 320, 600, 750 61.20 26.1216 May 137 4 400, 730 61.44 25.88

T.A. Rynearson et al. / Deep-Sea Research I 82 (2013) 60–71 61


Running head

of Chaetoceros spp. resting spores for carbon content estimation weredetermined from 2mL of preserved trap material from deployment3. Samples were filtered onto 13 mm, 0.8 μm polycarbonate filters,rinsed with trace ammonia solution (pH �10) and air dried. A smallportion (0.25 cm2) of the filter was cut from the center, mounted onan aluminum stub and sputter coated with �2 nm gold. A Leo1450VP SEM (Zeiss Inc.) with Smart SEM (V5.1) software automati-cally captured images of consecutive fields of view from a 15�15grid at a magnification of 2500� , providing 225 images per sample.The valve diameter, mantle, epivalve and hypovalve heights weredetermined for 4100 Chaetoceros spp. resting spores using ImageJ(⟨http://rsbweb.nih.gov.ij/⟩) after scale calibration using a 10 μm scalebar pre-set on each SEM image. Two resting spore shapes wereidentified (scalene and oblate spheroids) and their volume wasdetermined appropriately. Because no specific carbon to volume (C:Vol) equation exists for the resting spore of the identified species,carbon content was determined two ways; (1) using the C:Volrelationship determined for vegetative diatoms in Menden-Deuerand Lessard (2000), and (2) using the C:Vol relationship calculatedfor resting spores of the diatom Chaetoceros curvisetus (Kuwata et al.,1993). The percentage of POC flux contributed by resting spores wascalculated using total POC flux that was measured from the samedeployments (Martin et al., 2011).

2.4. Viability of cells in PELAGRA traps

Live PELAGRA material was used to determine variable fluor-escence (Fv/Fm), chlorophyll a fluorescence and resting sporeviability. Fv/Fm, an indicator of maximal quantum efficiency ofPhotosystem II (Maxwell and Johnson, 2000), was measured onlive trap material from the third and fourth PELAGRA deploy-ments. The material was diluted to a chlorophyll a concentrationof o10 mg L�1 with pre-filtered (Whatman GF/F) surface seawater.Three sub-samples were incubated for Z30 min at approximately6 1C in dark bottles covered with aluminum foil to prevent lightexposure. Fv/Fm was measured within the dark chamber of aFASTtracka II fluorometer (Chelsea Instruments Group Ltd.)mounted and secured in the lab, using FASTpro software (edition2230-001-HB-A). A single turnover protocol with 30 sequences peracquisition, each including 100 saturation and 50 relaxationflashlets, was utilized. The sequence interval was set to 100 ms;the PMT eht (extra high tension) and LED light source (excitationpeak of 470 nm) were optimized for each sample and varied from380 to 540 and 60 to 90 V, respectively. Fv/Fm was calculated fromsaturation and relaxation phase fits following Kolber et al. (1998).The first Fv/Fm determination was made on trap material withinseveral hours of retrieval. Samples were incubated at approxi-mately 6 1C in the dark and Fv/Fm determined at regular intervalsover the next 7 days (material from the third deployment) or after48 h (fourth deployment). Differences in Fv/Fm between depthsand deployments were tested using two-tailed t tests with equalvariances, which were determined using F-tests (Zar, 1996). Livematerial from the third deployment was examined for chlorophyllauto-fluorescence with excitation of photosynthetic accessorypigments, 470–490 nm, using an Axioskop microscope (ZeissInc.) equipped for epifluorescence. Images were recorded with aSPOT camera (Diagnostic Instruments, Inc.).

To determine if resting spores in live trap material collectedfrom the third deployment were viable, 3–400 μL aliquots of trapmaterial from 300, 600 and 750 m were added to 1 mL volumes ofsterile f/20 seawater media (Guillard, 1975) in 48 well platesimmediately after trap recovery. Inoculated plates were incubatedat ambient surface seawater temperature of 9 1C and 50 mmolphotons m�2 s�1 on a 16:8 h L:D cycle. Cultures were checkedmicroscopically using an Axioskop microscope (Zeiss Inc.) after 14,

41, 65 and 89 h; germination of resting spores was documented ateach time point with a SPOT camera (Diagnostic Instruments, Inc.).

2.5. Surface water sampling

Water samples for phyto- and microzoo-plankton communitycomposition and size as well as nutrient concentrations weredetermined fromwater samples collected once per day, typically atmidday (Fig. 1, Table 2). Variable fluorescence (Fv/Fm) was alsodetermined on discrete samples taken from 4 depths (5–60 m) atthe stations where phyto- and microzoo-plankton samples werecollected. Later analysis indicated, however, that despite a mini-mum 30 min dark treatment prior to measurements, the Fv/Fmdeterminations made on samples taken in daylight indicated aninfluence of previous light exposure. To compare surface andsediment trap Fv/Fm, we instead used measurements made onsamples taken at 5 m from stations in the study area that werecollected at night. Fv/Fm measurements on surface and sedimenttrap samples were made using the same protocol.

Complete hydrographic data are available from the Biologicaland Chemical Oceanography Data Management Office for theNAB 2008 project ⟨http://osprey.bcodmo.org/project.cfm?flag=view&id=102&sortby=project⟩. Water samples were collected withNiskin bottles mounted on a CTD-Rosette from 2 depths (usually10 and 30 m) that were both within the surface mixed layer (50–100 m) (Briggs et al., 2011). Taxonomic composition of the domi-nant species was similar between depths, hence only data from10 m is presented. On YD 131, samples for microplankton analysiswere taken from 6 depths (5, 30, 80, 120, 200, and 300 m), the onlydeep profile taken for such analysis.

2.6. Phyto- and microzoo-plankton

Samples were preserved in 1 L brown glass bottles withacidified Lugol's solution (approximately 2% final concentration)for taxonomic determination. The composition of the planktoncommunity was determined using quantitative light microscopyaccording to Utermöhl (1958). Analyses were carried out byOrbicon A/S (Århus, Denmark). Identification of Chaetoceros vege-tative cells and spores was based on Jensen and Moestrup (1998)and Rines and Hargraves (1988). To determine cell biovolumes, thelinear dimensions of cells were measured and biovolume calcu-lated using the appropriate geometric volume formula. Carbon tovolume relationships for either diatoms or for all other plankton

Table 2Date, location and start time for CTD profiles from which water samples werecollected for nutrient analyses and cell counts (cast number is the identifier used inthe BCO–DMO database to identify CTD profiles).

Date (2008) Yearday Time Cast no. Latitude (1N) Longitude (1W)

2 May 123 19:28 1 61.16 25.354 May 125 17:14 2 60.85 26.645 May 126 13:21 8 60.79 27.126 May 127 13:34 12 60.91 27.417 May 128 15:57 15 61.07 26.668 May 129 13:52 21 60.63 25.729 May 130 14:23 29 60.63 27.5910 May 131 11:38 35 61.34 26.5711 May 132 11:31 43 61.52 25.9712 May 133 11:25 46 61.44 25.9613 May 134 14:12 62 61.33 25.9814 May 135 13:14 73 61.24 26.2916 May 137 11:26 89 61.58 26.0818 May 139 12:47 105 61.08 25.4019 May 140 11:40 118 61.30 25.7520 May 141 13:19 121 61.49 25.06

T.A. Rynearson et al. / Deep-Sea Research I 82 (2013) 60–7162

128

Running head


taxa (Menden-Deuer and Lessard, 2000) were used to determinemicroplankton carbon content.

2.7. Nutrient analyses

Water samples were collected for nutrient analyses betweenYD 123 and 141 (Table 2). Water from 10 m depth was collecteddirectly from Niskin bottles into acid-cleaned 60 mL polyethylenebottles, pre-rinsed with three aliquots of sample, and frozenunfiltered immediately; samples were stored up to 6 months at�20 1C prior to analysis (Kallin et al., 2011). Samples wereanalyzed for silicic acid and nitrate plus nitrite (hereafter referredto as nitrate) using a Lachat QuickChems 8000 Flow InjectionAnalysis System (Smith and Bogren, 2001; Wolters, 2002). Sam-ples were slowly thawed in the dark at room temperature for 24 hand vigorously vortexed prior to analysis to avoid silica polymer-ization (Gordon et al., 1994). A depth profile (10–200 m) wascollected on YD 131 and analyzed as described above.

2.8. Statistical analyses

Linear regression analysis was used to describe the relationshipbetween resting spore carbon flux and POC flux in SPSS V. 20 (IBM,Inc). Statistical analyses of diatom community composition wereconducted using Primer 6 (Clarke and Gorley, 2006) and an alphaof 0.05. Changes in sediment trap community composition wereanalyzed by examining the four dominant diatom groups insediment traps: C. aff. diadema (vegetative cells and resting sporescombined), all other Chaetoceros spp., Thalassiosira spp., and allother diatom species. These groups were compared to three levelsof cell flux (o100, 100–400 and 4400 cells 106 m�2 day�1) intothe sediment traps. Samples were standardized, transformedusing a Bray–Curtis similarity matrix and compared usingAnalysis of Similarity (ANOSIM). Community composition ofsurface samples was compared in the same way except surfacecommunity compositions were split into two levels of bio-mass (430 mg C m�3 and o30 mg C m�3) and comparisonswere made using class-level taxonomic divisions (e.g., Bacillario-phyceae, Prasinophyceae). Comparisons of surface communitycomposition with environmental variables (nitrate, silicic acid,surface water temperature and salinity) were conducted usingBEST (Biology-Environment and Stepwise) analysis and class-leveltaxonomic divisions.

3. Results

3.1. Downward flux of phytoplankton taxa and POC

In contrast to most deep ocean and PELAGRA samples collectedin the past, the majority of the material collected during this studywas present as identifiable cells. Counts from preserved trapmaterial revealed that diatoms numerically dominated the 13PELAGRA samples. Other microplankton were present (e.g., for-aminifera, dinoflagellates, ciliates) but comprised o0.25% of cellsin the traps. The number of sinking diatoms collected by thePELAGRA cups peaked during the third deployment (retrieved onYD 136), and decreased thereafter (Fig. 2, Table 3). The mostnumerically dominant cells in the sediment trap material for alldeployments were vegetative and resting spores of Chaetocerosspp., which together comprised an average of 81% of cells. Restingspores comprised between 35% and 92% of all phytoplankton cellsin the traps, and 449% in 10 of the 13 samples (Table 3). Thelargest flux of Chaetoceros resting spores occurred on YD 136,when cell flux exceeded 400�106 cells m�2 day�1 at all depths.The number of vegetative cells of Chaetoceros was highly variable

in trap material, ranging from 2% to 40% of cells. Diatoms in thegenus Thalassiosira comprised between 2% and 45% of cells in thetrap, with a higher numerical dominance during the fourthdeployment. All remaining diatom species contributed between3% and 19% of trapped cells, and included the diatoms Thalassio-nema spp. and Pseudo-nitzschia spp.

The majority (495%) of Chaetoceros resting spores wereidentified as Chaetoceros aff. diadema. Although these restingspores were morphologically similar to C. diadema, there wereimportant differences, including a lack of highly branched spines(Fig. 3). In some cases, resting spores were embedded insidevegetative cell frustules, which were also similar to C. diademaexcept for the foramen, which did not match the taxonomicdescriptions of this species. Resting spores of other Chaetocerosspecies whose vegetative cells were abundant in surface waters(see below) were also observed in the traps, but were uncommon(o5% of total spores); the spores of these species (C. compressusand C. laciniosus) have very distinct resting spore morphologies(Tomas, 1997). Other dominant Chaetoceros species identified fromthe surface are not known to form spores (e.g., C. decipiens;(Tomas, 1997)).

Cell size measurements were used to compute cell carboncontent for C. aff. diadema spores; however, two morphologieswere observed in sediment trap material. One was circular incross-section (Fig. 3A) with an average cell diameter of8.471.5 mm (n¼50), and the other was elliptical in cross-section(Fig. 3B) with average dimensions of 14.772.4 mm and 3.1+0.3 mm(n¼50). Aside from these differences, they appeared identical. Thecarbon contents of the two morphologies were 145.670.5 and90.770.4 pg C cell�1, respectively, when the C:Vol relationship ofvegetative diatoms was used (Menden-Deuer and Lessard, 2000).Carbon content of the resting spores was substantively larger(102074.4 and 568.673.3 pg C cell�1, respectively) when the C:Vol relationship determined for C. curvisetus resting spores wasused (Kuwata et al., 1993). Because the two morphologies wereonly visible in SEM images and were not recorded independentlyin light microscope cell counts, we calculated carbon flux using asize-averaged estimate for the carbon content per resting spore.Using the equation for vegetative diatoms, the size-averaged Ccontent per cell was 11870.3 pg C and the average flux contrib-uted by resting spores was 2.4–62.7 mg C m�2 day�1 or 9–64% ofthe POC flux (Table 3). Using the C:Vol relationship for C. curvisetusresting spores, the size-averaged C content per cell was794.373.8 pg C. Using this relationship, POC flux contributed byresting spores was 16.4–424 mg C m�2 day�1 or 59–431% of thetotal measured POC flux.

0

200

400

600

800

1000

127, 140

127, 230

132, 160

132, 160

132, 340

132, 620

136, 320

136, 600

136, 750

139, 400

139, 730

140, 400

140, 730

Yearday of trap recovery (YD), Depth (m)

Other diatom species

Thalassiosira spp.

Chaetoceros spp.

Resting spores-C. aff. diadema

Cel

ls 1

06 m-2

d-1

YDm

Fig. 2. Flux of diatom cells into each sediment trap. All species listed are vegetativecells except where noted. The category “other diatom species” includes allremaining diatoms including the genera Thalassionema and Pseudo-nitzschia.Recorded yearday (YD) is for closure of the trap cup; see Table 3.



Running head

The largest percentage of POC flux generated by resting sporeswas observed during the third deployment. Across all deploy-ments, POC flux was tightly correlated with resting spore flux(p¼0.003, r2¼0.58) and less so with community composition inthe sediment traps (p¼0.047, R¼0.275).

3.2. Viability of cells from PELAGRA traps

Live trap material from the final two deployments was exam-ined shortly after trap recovery for photosynthetic capacity. Thevariance and mean of Fv/Fm measurements of shallow (320 m) vs.deep (600 and 750 m) samples were not significantly different(F-test, p¼0.7; t-test, p¼0.9) from the third deployment (Fig. 4A).The magnitude of Fv/Fm from live trap material was similar to thehighest value recorded for surface waters during the cruise, i.e.,0.43070.018 (n¼2) on YD 131 for discrete water samples col-lected at night from 5 m. Fv/Fm in the incubated trap materialdecreased similarly for all depths over the following days anddropped significantly after one week of incubation to0.37170.006 for the 300 m sample (t-test, po0.001) and to0.37370.003 for the 750 m sample (t-test, po0.001). From thefourth deployment, Fv/Fm measurements were 0.33870.074(300 m) and 0.43170.012 (700 m). There was no significantdifference in the mean or variance of Fv/Fm between depths forthe fourth deployment (F-test, p¼0.21; t-test, p¼0.22). Underphase contrast microscopy, resting spores were clearly pigmented(Fig. 3D). Resting spores from the largest flux event (third deploy-ment) had bright chlorophyll fluorescence under epifluorescencemicroscopy (Fig. 4B).

After 14 h of incubation under low irradiance at in situ surfacewater temperature, resting spores from all depths began togerminate by casting off the spiny resting spore valve (Fig. 5A–C). By 41 h, cells from all depths had cast off the smooth restingspore valve and undergone one cell division (Fig. 5D). Cell divisioncontinued on days 3 and 4, with formation of 4 and 8 cell chains,respectively (Fig. 5E and F). Frequently, resting spore valves wereobserved lying next to 4 and 8 cell chains, indicating that thevegetative cell chains originated from a single resting spore. Afterthe initial 14 h incubation, growth rates of germinated restingspores were rapid, approximately 1 doubling day�1.

3.3. Diatoms in surface waters

Surface community composition varied significantly as a func-tion of carbon biomass (p¼0.002, R¼0.59). High levels of carbon(430 mg C m�3) were associated with diatoms and low levels ofcarbon (o30 mg C m�3) with other phytoplankton, includingprymnesiophytes, dinoflagellates and cryptophytes (Fig. 6). Atstations sampled on YD 123–128 and 133, diatoms were particu-larly abundant and contributed the majority of organic carbon(56–88%) associated with the phyto- and microzoo-planktoncommunities in the upper 10 m of the water column. Total carbonconcentrations of microplankton ranged from 16 to 78 mg C m�3.

The same diatom genera identified as the most abundant insediment traps were also observed in surface waters. Pseudo-nitzschia spp. comprised, on average, 28% of cell abundance insurface waters, in stark contrast to its average abundance of 2.6%in sediment traps (Tables 3 and 4). Average abundances ofThalassionema spp. and Thalassiosira spp. in surface waters (6%,3% respectively) were more similar to their average abundance insediment traps (4%, 11% respectively). Chaetoceros spp. repre-sented 457% of diatom cells (Table 4) and 36–65% of diatom cellcarbon in surface waters on YD 123–128 and 133. On average,Chaetoceros spp. represented 31% of cell abundance in surfacewaters, in contrast to its abundance in the sediment traps (81%).Chaetoceros diadema comprised an average of only 1% of diatomTa

ble

3Su

mmaryof

cellco

mpositionfrom

PELA

GRAfloa

tingsedim

enttrap

s.Dep

loym

entnumbe

rco

rrespon

dsto

that

usedin

Martinet

al.(20

11).Cellfluxe

sof

major

diatom

taxa

areforve

getative

cells

,unless

otherwisenoted

.Carbo

nfluxe

sof

allCh

aetocerosrestingsp

ores

werege

nerated

usingtheC:Vol

conve

rsionof

Men

den

-Deu

eran

dLe

ssard(2000)

andco

mpared

tototalPO

Cfluxfrom

Martinet

al.(20

11).

Dep

loym

ent

Trap

reco

very

(yea

rday

)Duration

oftrap

colle

ction(h)

Dep

th(m

)To

tal

diatoms

Chae

toceros

restingsp

ores

Chae

toceros

spp.

Thalassiosira

spp.

Thalassion

ema

spp.

Pseudo

-nitzschia

spp.

Other

diatom

spp.

TotalPO

Cfluxd

Chae

tocerosresting

sporecarbon

flux

Restingsp

ore

carbon

flux(%)

Fluxas

cells�10

6m�2day�1

mgCm�2day�1

112

716

140

53.2

30.8

12.0

2.2

2.1

3.4

2.7

15.47

0.8

3.67

0.01

23.6

112

715

230

52.0

20.5

20.8

2.1

2.5

2.5

3.6

14.47

0.7

2.47

0.01

16.8

2a13

272

160

32.6

25.2

2.7

1.6

1.8

1.1

0.2

10.27

0.3

3.07

0.01

29.2

2a13

272

160

43.0

37.2

0.8

1.3

2.9

0.5

0.3

13.37

0.7

4.47

0.01

332

132

4834

087

.280

.02.2

2.0

1.6

1.2

0.2

30.67

1.7

9.57

0.02

30.9

213

260

620

29.8

26.2

1.5

0.7

0.9

0.3

0.2

27.27

1.8

3.17

0.01

11.4

313

624

320

469.1

411.0

33.1

11.0

11.0

1.2

1.8

76.27

6.1

48.67

0.12

63.7

313

624

600

788.0

471.0

188.0

68.0

23.0

20.0

18.0

1647

4.1

55.67

0.14

33.9

313

624

750

963.0

531.0

333.0

51.0

13.0

11.0

24.0

1547

4.8

62.77

0.16

40.7

4b13

924

400

196.3

96.3

41.0

21.0

10.0

14.0

14.0

1207

5.8

11.47

0.03

9.5

4b13

924

730

268.1

113.0

46.2

100.0

5.3

1.8

1.8

95.27

4.6

13.37

0.03

14.1

4c14

024

400

157.9

83.3

42.2

17.0

8.0

4.0

3.4

1127

5.4

9.87

0.02

8.8

4c14

024

730

377.1

132.0

59.7

170.0

5.7

3.2

6.5

75.17

3.7

15.67

0.04

20.8

aTw

otrap

sdep

loye

dat

samedep

th.

bFirst24

hco

llection.

cSe

cond24

hco

llection.

dFrom

Martinet

al.(20

11).


130

Running head


cell abundance in surface waters and its maximum contribution at10 m during the cruise was 5.2% of cell abundance. This is in starkcontrast to the average contribution of resting spores of thisspecies to the sediment traps (63%).

During the time period when Chaetoceros was most abundantin surface waters (YD 123–128, 133; Table 4), the most numerousChaetoceros species identified were C. laciniosus, C. compressus, andC. decipiens. They represented, on average, 64% of all Chaetoceroscells. Many Chaetoceros cells in surface waters were small andlightly silicified and thus it was not possible to identify them fromfixed samples; these unidentified cells comprised, on average, 32%of Chaetoceros in surface waters.

3.4. Depth distribution of Chaetoceros aff. diadema

Few resting spores of C. aff. diadema were recorded in theupper 30 m of the water column (maximal number of 4�103 L�1

was observed at 10 m on YD 127). On YD 131, samples werecollected between 5 and 300 m on the only deep taxonomicprofile. No resting spores were recorded at 5 m (Fig. 7A) and at30 m, they constituted just 0.1 mg C m�3. Between 80 and 300 m,resting spore carbon concentrations were an order of magnitudehigher (between 1.1 and 2.2 mg C m�3). The absolute amount ofdiatom cell carbon decreased with depth from 30 to 300 m(Fig. 7B); however, the relative contribution of diatom carbonfrom C. aff. diadema increased with depth. This species comprisedon average o2% of the diatom carbon biomass at the surface, butcomprised 59% and 82%, respectively, at 200 and 300 m.

3.5. Nutrient concentrations

Inorganic nitrogen concentrations were relatively high, remain-ing over 8 mM in surface waters throughout the cruise. In contrast,dissolved silicic acid concentrations were already low at thebeginning of the sampling period (YD 123,o4 mM) and decreasedto o0.3 mM by YD 133 (Fig. 8A). For surface waters, there were nosignificant correlations between changes in community composi-tion and environmental conditions (nitrate, silicic acid, surfacewater temperature and salinity; p¼0.49). Nutrients measuredfrom the deep taxonomic profile on YD 131 increased with depth(Fig. 8b). The exception was at 30 m where there was a decreaserelative to surface waters in both nitrate and silicic acid.

4. Discussion

In May of 2008, a major particle flux event was observed in theNorth Atlantic from optical ‘spikes’ in chlorophyll fluorescence,backscatter, and beam attenuation from the ship and four Seagli-ders (Briggs et al., 2011). During the flux event, the amount of POCrecovered from floating sediment traps increased by over an orderof magnitude (Martin et al., 2011). We determined that phyto-plankton cells were responsible for generating the flux event andthat their abundance in the sediment traps varied by over 30 foldduring the two-week deployment period, highlighting the ephem-eral nature of downward carbon flux during the North Atlanticspring bloom. Here, we show that a subset of species from the

Fig. 3. Scanning electron micrograph images of the most abundant Chaetoceros aff. diadema resting spores representing two morphologies: (A) the circular cross-section and(B) the elliptical cross-section (both from sediment trap recovery on May 15 (YD 136), 600 m). Light micrographs of a C. aff. diadema resting spore: (C) inside a vegetativevalve and (D) free of the vegetative valve (both from sediment trap recovery May 15 (YD 136), 750 m). Scale bars indicate 2 mm (A, B) and 15 mm (C).



Running head

surface phytoplankton community and a transient life cycle stagethat may be highly resistant to degradation (i.e., resting spores)

can dominate vertical flux, influencing the magnitude and effi-ciency of POC flux associated with the North Atlanticspring bloom.

4.1. Chaetoceros resting spores dominate plankton fluxand POC in the traps

During the peak POC flux, collected by the third sediment trapdeployment (recovered YD 136), Chaetoceros resting spores repre-sented 55–88% of all diatom cells (Table 3). Depending on the C:Vol conversion used, resting spores represented up to 64% or 431%of the total POC flux. The C:Vol relationship based on vegetative

Fig. 4. Photosynthetic capacity of sediment trap material and resting spores fromthe third sediment trap recovery, May 15 (YD 136). (A) Variable fluorescence (Fv/Fm)of trap material collected from 300, 600 and 750 m and measured between 1 and7 days after recovery. (B) Chlorophyll a autofluorescence of resting spores collectedfrom 750 m; excitation at 470–490 nm.

Fig. 5. Germination of resting spores from the third sediment trap recovery on May 15 (YD136). After 14 h incubation, spiny resting spore valves (indicated by arrows) werecast off in trap material from 300 m (A), 600 m (B) and 750 m (C). By 41 h (D), the smooth resting spore valve (denoted by arrow) had been cast off and one cell division hadoccurred; example cell from 600 m. Cell division after 65 h (E, 4-cell chain) and 89 h (F, 8-cell chain) of incubation for trap material from 750 m.

0

20

40

60

80

123

125

126

127

128

129

130

131

132

133

134

135

137

139

140

141

mg

C m

3

Yearday

MicrozooplanktonOther PhytoplanktonDiatoms

Fig. 6. Microplankton carbon from 10 m, including all diatoms, all other phyto-plankton larger than 2 mm (including prymnesiophytes, dinoflagellates and crypto-phytes) and all microzooplankton (including ciliates, heterotrophic dinoflagellatesand choanoflagellates).


132

Running head


diatoms (Menden-Deuer and Lessard, 2000) yielded the 64%estimate and was likely conservative, given that Chaetocerosresting spores have been found to contain four to ten times morecarbon than vegetative cells (French and Hargraves, 1980; Kuwataet al., 1993). Resting spores are thought to be more carbon densedue to the lack of a vacuole and the presence of storage com-pounds, including lipids and carbohydrates (Anderson, 1975;Doucette and Fryxell, 1983; Kuwata et al., 1993). Higher carbondensity is reflected in the POC flux estimates (up to 431%) based onthe C:Vol relationship for C. pseudocurvisetus resting spores(Kuwata et al., 1993). The carbon per cell generated using this

approach clearly led to an unrealistic overestimation of restingspore contribution to total POC flux, perhaps due to the fact that itwas based on resting spores of a different species (althoughthe same genus). For example, C. pseudocurvisetus resting sporeshad a cell volume that was nearly seven times smaller than the

Table 4Abundance of diatom cells at 10 m and percentage of cells belonging to each genus included in Table 3. Percent Chaetoceros includes C. aff. diadema.

Year-day Diatom abundance (103 cells L�1) Chaetoceros (%) C. aff. diadema (%) Thalassiosira (%) Thalassionema (%) Pseudo-nitzschia (%)

123 310 75 0.3 0.6 2.9 16125 460 69 0.8 0.0 8.5 17126 540 75 2.6 3.4 2.9 17127 425 57 3.7 1.3 3.4 24128 1126 71 1.0 0.1 5.3 17129 276 17 4.7 0 4.2 17130 164 2 0.6 4.0 0.8 58131 82 11 1.6 0 6.3 20132 19 6 5.2 0 10.5 18133 598 63 0.2 0 13.9 20134 441 35 0.2 0.9 14.4 42135 171 2 0 0 4.6 70137 40 0.5 0 6.4 0 0139 49 22 0.8 2.7 8.0 53140 22 0 0 23.6 0 36141 31 0.3 0 0.0 6.2 27

0 0.5 1 1.5 2

5

30

80

120

200

300

mg C m-3

Dep

th (m

)

0 5 10 15

5

30

80

120

200

300

mg C m-3

Dep

th (m

)

C. aff.diadema

Other Chaetoceros spp.

Thalassiosira spp.

Other diatom spp.

Resting spores Vegetative cells Resting spores inside vegetative frustules

Fig. 7. Depth profile (5–300 m) on May 10 (YD 131). (A) carbon concentration of C.aff. diadema resting spores, vegetative cells, and resting spores found inside ofvegetative frustules, and (B) carbon concentration of all diatoms.

0

50

100

150

200

0 5 10 15

Dep

th (m

)

L-1

0

2

4

6

8

10

12

120 125 130 135 140

mol

es L

-1

Yearday

Fig. 8. Concentrations of nitrate plus nitrite (diamonds) and silicic acid (squares)from (A) 10 m collected from stations listed in Table 2 (no silicic acid measurementavailable for YD 137), and (B) the depth profile on May 10 (YD 131).



Running head

C. aff. diadema spores (253 μm3 vs. 1680 μm3), and they may havethus been more carbon dense. It has been noted that the C:Volrelationship of individual species can deviate significantly fromeach other (Menden-Deuer and Lessard, 2000). Using eitherestimate, the sediment trap results clearly showed that restingspores contributed a significant percentage of the POC flux duringthe North Atlantic spring bloom.

Prior to and during the flux of resting spores into sedimenttraps, high backscattering signals were observed in the watercolumn and interpreted as sinking aggregates (Briggs et al., 2011).Material in the PELAGRA traps was examined microscopicallyimmediately after recovery and a surprising feature was that noaggregates were visible, unlike previous PELAGRA deploymentswhere aggregates were observed (Lampitt et al., 2008; Salter et al.,2007). We conclude that the flux event was mediated by fragileaggregates that lost their identity in the sediment trap cup beforerecovery. Given that resting spores dominated the carbon biomass,it is likely that they were present within the sinking aggregates, ashas been observed previously (Alldredge et al., 1995). Aggregateformation may have been enhanced by transparent exopolymerparticles (TEP), produced by phytoplankton (Passow andAlldredge, 1995) and identified in large quantities at the peak ofPOC flux during this study (Martin et al., 2011). Thus, the peak fluxof resting spores observed here (4500�106 cells m�2 day�1,Fig. 2) may have been accelerated by aggregation and TEPproduction.

The physiological characteristics of resting spores and sedimenttrap material were suggestive of rapid POC flux with a hightransfer efficiency. The Fv/Fm values recorded in the sediment trapmaterial were, for example, similar to those recorded in surfacewaters at the beginning of our study and among the highestrecorded at any station or depth in the study as a whole. The Fv/Fmdecreased over the week following retrieval from the sedimenttraps (Fig. 4A) indicating a loss of photosynthetic capacity overtime. Thus, the fact that Fv/Fm values recorded in the sediment trapmaterial were similar from all depths where sediment traps weredeployed suggests that sedimentation was occurring rapidly at thistime collection. Strong in vivo chlorophyll fluorescence of theresting spores from the traps also indicates that the material wasrelatively fresh, and not degraded (Fig. 4B). This physiologicalevidence is supported by estimates that the sinking rate ofaggregates was �75 m day�1 (Briggs et al., 2011). Furthermore,resting spores collected from down to 750 m depth germinatedrapidly after incubation with nutrients and light, showing thatthese live spores contained high quality, labile POC (Fig. 5). Indeed,the measured transfer efficiency of POC to the sediment traps forthis flux event was 30–126% higher than the predicted 19% (Martinet al., 2011). Transfer efficiencies reported for diatom blooms inthis region of the North Atlantic are amongst the highest mea-sured in the global ocean (Buesseler and Boyd, 2009), suggestingthat the flux of resting spores described here may not represent anisolated event.

Importantly, resting spore formation and flux is not restrictedto the North Atlantic. For example, resting spores of the diatomEucampia antarctica var. antarctica captured by sediment trapsrepresented up to 71% of all sinking diatom cells following anatural iron fertilization event in the Southern Ocean (Salter et al.,2012). Furthermore, it appeared that the flux of sedimentingmaterial was also comprised of relatively un-degraded organicmaterial, suggesting similarly high transfer efficiencies associatedwith resting spore flux in the Southern Ocean. This indicates thatimportant flux events of diatom resting spores may be generatedby multiple genera and in multiple ocean basins. The significantflux of C generated by resting stages has also been observed inother taxa, such as dinoflagellates. For example, large fluxes ofdinoflagellate cysts have been observed in coastal regions of the

Pacific (Fujii and Matsuoka, 2006; Pospelova et al., 2010), theNorth Sea (Godhe et al., 2001), and the Baltic Sea, where one fluxevent contributed about 45% of the maximum POC flux following aspring bloom (Heiskanen, 1993). Together, these observations haveimportant implications for the disproportionate contribution ofparticular taxa and importantly, different life stages, to thebiological pump.

4.2. Comparison between the surface plankton communityand cells collected in sediment traps

Sampling of surface waters took place during the peak of thediatom spring bloom (Alkire et al., 2012). The species compositionat the surface was typical of phytoplankton communities duringthe spring bloom in this region (Moore et al., 2005; Sieracki et al.,1993); phytoplankton chlorophyll and carbon concentrations insurface waters were average for the peak of the North Atlanticspring bloom (Henson et al., 2009). Phytoplankton concentrationwas patchy during the cruise, as observed in ocean color satelliteimagery (cf. Martin et al. (2011)). Chlorophyll a concentrationsmeasured from the ship and by gliders also exhibited patchiness,with concentrations varying by more than five-fold on any givenday (e.g., 0.35–2 mg m3) (Mahadevan et al., 2012). Movement ofthe ship in and out of patches with elevated chlorophyll wasresponsible for the variability in diatom abundance among oursamples. The sampling strategy allowed for random sampling bothinside and outside of patches, and was geographically welldistributed in the study area.

At sinking rates of 75 m day�1 calculated for the flux event(Briggs et al., 2011), the high number of C. aff. diadema restingspores recovered on YD 136 at depths of 320–750 m were mostlikely present in surface waters as vegetative cells until YD 132.Although Chaetoceros spp. dominated surface phytoplankton com-munities, particularly at the beginning of the cruise (YD 123–128),C. aff. diadema never accounted for more than 5.2% of the diatomcells in surface waters before YD 132 (Table 4). This species wasobserved in all samples throughout the study area until YD 135,but only at low concentrations both in and out of the patches. Wehypothesize that a rarely-occurring species can disproportionallycontribute to total organic carbon flux, in this case most likely as afunction of its life history strategy of resting spore formation.

The observation that resting spores were in great abundance insediment traps on YD 136 at all depths, including the shallowesttrap (320 m), supports the hypothesis that C. aff. diadema was arare species at the surface but a disproportionate contributor toPOC flux (Fig. 2). This conclusion gains further support from thefact that the importance of C. aff. diadema as a percentage of totaldiatom carbon increased dramatically with depth (Fig. 7). While C.diadema occurred together with a number of other Chaetocerosspp. in surface waters, it was essentially the only Chaetocerosspecies found at 300 m. This increasing contribution of C. diademacarbon at greater depths is likely a consequence of the highlysilicified valves of resting spores which appear to be moreresistant to the degradation experienced by their vegetativecounterparts (Hargraves and French, 1983; Kuwata andTakahashi, 1990).

An alternative hypothesis is that the spores were derived froma major bloom of C. aff. diadema that occurred prior to our arrivalon site. This hypothesis would require very slow sinking rates onthe order ofo30 m day�1, which is unlikely given sinking rates of�75 m day�1 estimated between YD 123 and 136 (Briggs et al.,2011). Few resting spores were found in shallow traps during thefirst and second deployment, an observation that is also counter toslow sinking of an early C. aff. diadema bloom. Furthermore,chlorophyll a and particle concentrations measured from the floatsand gliders were very low prior to the start of sampling on YD 123,


134

Running head


suggesting low phytoplankton abundance (Alkire et al., 2012;Mahadevan et al., 2012).

4.3. Initiation of spore formation

The appearance of resting spores in the sediment traps over avery short time period suggests that they formed quickly, con-sistent with reports that Chaetoceros can form resting sporeswithin 6–48 h (reviewed in McQuoid and Hobson (1996)). Restingspore formation is a process that involves wholesale morphologi-cal (Hargraves, 1979; Ishii et al., 2011) and metabolic changes fromthe vegetative life stage (Doucette and Fryxell, 1983; French andHargraves, 1980; Kuwata et al., 1993). While our data do not allowidentification of the trigger(s) for spore formation, they do provideinsights into the environmental conditions present before andduring the time when resting spores were identified in the watercolumn. This is important because diatom resting spores have notconsistently been observed in sinking material (e.g., Billett et al.(1983), Smith et al. (1996)) but when present, can have a largeimpact on C flux (this study; Salter et al., 2012).

In laboratory experiments, diatom resting spore formation hasoften been associated with nutrient depletion (French andHargraves, 1980; Garrison, 1981; Kuwata et al., 1993; Oku andKamatani, 1999; Pitcher, 1986; Sanders and Cibik, 1985). Duringthe period preceding the appearance of diatom spores in thesediment traps, surface waters remained nitrate replete but dis-solved silicic acid concentrations declined from �4 to o1 mM(Fig. 8A). While silicic acid deficiency may have provided a triggerfor spore formation, resting spores captured in the deepestsediment traps appeared to be in excellent physiological condi-tion (Fig. 4). When spores were placed in a nutrient enrichedmedium and provided with light, they began to germinatewithin 14 h (Fig. 5). Such observations do not reconcile well withthe assumption that spores were formed from nutrient stressed,vegetative cells.

An argument that challenges silicic acid depletion as a potentialtrigger for spore formation is the fact that spores themselves areheavily silicified. Resting spores of some Chaetoceros speciescontain 3–4 times more silicon than vegetative cells and it appearsthat they can form resting spores only when sufficient silicic acidis present (Kuwata et al., 1993; Kuwata and Takahashi, 1990). Thus,it is difficult to explain the formation of heavily silicified restingspores under dissolved silicic acid concentrations of o1 mM insurface waters. This is consistent with our observations of therebeing no or only very few spores in surface waters at any timeduring the cruise. Instead, concentrations of spores between 80and 300 m were found to be an order of magnitude higher thanthose ever recorded in surface waters (Fig. 7A). Vegetative cellscontaining spores were only observed between 80 and 200 m, andfree spores were observed at 300 m. This observation suggests thatthe morphological changes associated with spore formation mayhave occurred primarily below the euphotic zone at depths wheresilicic acid, needed to form resting spore valves, was found athigher concentrations (Fig. 8B). Alternatively, silicic acid requiredfor production of heavily-silicified resting spores could potentiallyhave come from re-mobilization of silica in the frustules; however,we have no data on frustule density of biogenic silica, either pre orpost spore formation, to disprove this hypothesis. The datagathered here suggest a generalized scenario whereby the mor-phological changes associated with spore formation occurred ascells sank from the surface waters.

4.4. Ecological implications of resting spore flux

The morphological and metabolic changes associated withresting spore formation may ultimately influence their ability to

survive and form subsequent blooms. Unlike vegetative cells,whose organic frustule covering can be degraded by bacteria,making the silica readily soluble (Bidle and Azam, 1999), restingspores are both inherently more resistant to bacterial degradationand heavily silicified, thereby enhancing their survival capability.For example, resting spores of Chaetoceros pseudocurvisetus appearmore resistant to bacterial degradation than vegetative cells(Kuwata and Takahashi, 1990). Resting spores observed here wereheavily silicified, potentially enhancing their survival capability.For example, the resting spores of some Chaetoceros species wereable to germinate after passing through the guts of their copepodpredators (Hargraves and French, 1983; Kuwata and Tsuda, 2005).Furthermore, when resting spores are present, copepods lowertheir filtering rates or actively avoid ingesting the spores suggest-ing that copepods have evolved strategies to actively avoid restingspore ingestion (Kuwata and Tsuda, 2005). Resting spores sampledhere were able to rapidly resume vegetative growth, dividing atrates of �1 doubling day�1. Rapid germination and growth isconsistent with laboratory observations of C. diadema restingspores (Hollibaugh et al., 1981). The high growth rates achievedimmediately following germination may help to explain thedominance of this spore-forming genus in mid to high latitudeblooms. For example, Chaetoceros spp. have been shown todominate spring blooms over multiple years in regions of theBarents and Norwegian Seas (Degerlund and Eilertsen, 2010) aswell as the waters just east of the Iceland Basin (Bresnan et al.,2009). Importantly, the germination of Chaetoceros spp. restingspores has been observed in the field previously, althoughprimarily in upwelling and coastal regions (Garrison, 1981;Pitcher, 1990).

In open ocean environments, the ability of resting spores toreseed surface waters following germination would depend on acombination of survival time, water depth, winter mixing andcirculation. Resting spores of C. diadema can survive for long timeperiods, although the proportion of cells germinating decreasesover time. For example, after approximately 50 days in darkness,85% of C. diadema resting spores germinated in the laboratory(Hollibaugh et al., 1981). Maximal survival time has not beenmeasured, but 10% of C. diadema cells were still able to germinateafter nearly two years in darkness (Hollibaugh et al., 1981). Inaddition to extended survival ability, resting spores must also beresuspended. While that is unlikely in deeper areas of the NorthAtlantic, it may be possible in shallower regions, such as coastalwaters where Chaetoceros is a common diatom genus. For exam-ple, the neutrally buoyant sediment traps used here were recov-ered within 40 km of the Reykjanes Ridge (Fig. 1), which hasdepths of 500–1000 m in this region. Seagliders operating duringthis experiment occasionally observed high backscatter measure-ments below 600 m, indicative of re-suspended sediments andsediment plumes coming from the ridge (Briggs et al., 2011). Deepwinter mixing in the Iceland Basin to depths of 600 m could bringre-suspended spores back to the surface (Backhaus et al., 2003).Re-suspended spores could then be transported by prevailingcurrents to the Iceland Basin and beyond (Pollard et al., 2004).Similar mixing processes in shallow coastal waters off Greenlandand North America could allow resting spores to reseed broadareas of the North Atlantic.

5. Conclusions

The data presented here show that biological propertiesbeyond community composition, such as species identity and lifecycle stage, can play important roles in the flux of POC to depth.On average, 1–3% of POC leaving the surface ends up below thedepth of sequestration, at 1000 m (reviewed in De La Rocha and



Running head

Passow (2007)). The amount of initial POC reaching great depthmay increase significantly due to the formation and flux of highly-silicified, degradation-resistant diatom resting spores. This is seenin the increased efficiency of POC flux observed in this field studyby Martin et al. (2011), with up to 43% of the POC at 100 mtransferred to depths of 750 m or greater. Importantly, the speciesthat formed resting spores and contributed most to POC fluxduring this event was relatively rare in surface waters. Based onsediment cores from the deep ocean, it is clear that resting sporeshave regularly reached the sediments in large quantities overgeologic time (Abelmann et al., 2006; Grimm et al., 1997; Suto,2006). Resting spores have also been captured in large numbersfrom deep sediment traps following natural iron enrichment(Salter et al., 2012). This suggests that the formation and sinkingof resting spores may essentially “magnify” the effect of surfacediatom blooms by transporting proportionally more POC to depththan from blooms where diatoms are not present or do not formresting spores.

Challenges remain in terms of understanding the biogeochem-ical and ecological impacts of resting spore formation. For exam-ple, the massive open ocean flux event of resting spores that wecaptured in this study was ephemeral and the trigger(s) for sporeformation unknown. These characteristics make it a challenge toeffectively observe, quantify and predict flux events generated byresting spores. A more complete understanding of the impacts offlux events will benefit from future work to better characterize thechemical and physiological composition of resting spores and howthey differ from vegetative cells (e.g., Doucette and Fryxell (1983)).For example, without accurate carbon to volume estimates forresting spores, we were unable to put narrow bounds on thecarbon contribution of resting spores to the total POC flux:depending on the carbon-to-biovolume conversion used, ourestimates ranged by over six-fold. This highlights that additionalbasic information about diatom life stage could be used to betterunderstand large-scale events such as the fate of the NorthAtlantic spring bloom.

Acknowledgments

This work would not have been possible without the vision andhard work of C. Lee and E. D'Asaro who designed the NorthAtlantic Bloom study and the assistance of our Icelandic colleagueK. Guðmundsson. We thank E. Kallin for nutrient analyses,I. Cetinić for assistance with figures, the Captain and crew of theR/V Knorr, and numerous students and colleagues who helpedon the cruises. This work was supported by US NSF OCE0727227(to TAR); US NSF OCE0628379, OCE0628107 and US NASANNX08AL92G (to MJP with subcontracts to MES); Danish ResearchCouncil for Nature and Universe and Danish National ResearchFoundation (to KR); UK Natural Environment Research Council (toAJP and RSL).

References

Abelmann, A., Gersonde, R., Cortese, G., Kuhn, G., Smetacek, V., 2006. Extensivephytoplankton blooms in the Atlantic sector of the glacial Southern Ocean.Paleoceanography 21 (1), PA1013.

Alkire, M.B., D'Asaro, E., Lee, C., Perry, M.J., Gray, A., Cetinić, I., Briggs, N., Rehm, E.,Kallin, E., Kaiser, J., González-Posada, A., 2012. Estimates of net communityproduction and export using high-resolution, Lagrangian measurements of O2,NO3

� , and POC through the evolution of a spring diatom bloom in the NorthAtlantic. Deep Sea Research Part I 64, 157–174.

Alldredge, A.L., Gotschalk, C., Passow, U., Riebesell, U., 1995. Mass aggregation ofdiatom blooms: insights from a mesocosm study. Deep Sea Research Part II 42(1), 9–27.

Alldredge, A.L., Gotschalk, C.C., 1989. Direct observations of the mass flocculation ofdiatom blooms: characteristics, settling velocities and formation of diatomaggregates. Deep Sea Research Part A 36 (2), 159–171.

Alldredge, A.L., Passow, U., Logan, B.E., 1993. The abundance and significance of aclass of large, transparent organic particles in the ocean. Deep Sea Research PartI 40 (6), 1131–1140.

Anderson, O.R., 1975. The ultrastructure and cytochemistry of resting cell formationin Amphora coffaeformis (Bacillariophyceae). Journal of Phycology 11, 272–281.

Backhaus, J.O., Hegseth, E.N., Wehde, H., Irigoien, X., Hatten, K., Logemann, K., 2003.Convection and primary production in winter. Marine Ecology Progress Series251, 1–14.

Beaulieu, S.E., 2002. Accumulation and fate of phytodetritus on the sea floor.Oceanography and Marine Biology 40, 171–232.

Bidle, K.D., Azam, F., 1999. Accelerated dissolution of diatom silica by marinebacterial assemblages. Nature 397, 508–512.

Billett, D.S.M., Lampitt, R.S., Rice, A.L., Mantoura, R.F.C., 1983. Seasonal sedimenta-tion of phytoplankton to the deep-sea benthos. Nature 302, 520–522.

Bresnan, E., Hay, S., Hughes, S.L., Fraser, S., Rasmussen, J., Webster, L., Slesser,G., Dunn, J., Heath, M.R., 2009. Seasonal and interannual variation in thephytoplankton community in the north east of Scotland. Journal of SeaResearch 61 (1), 17–25.

Briggs, N., Perry, M.J., Cetinić, I., Lee, C., D'Asaro, E., Gray, A., Rehm, E., 2011. High-resolution observations of aggregate flux during a sub-polar North Atlanticspring bloom. Deep Sea Research Part I 58, 1031–1039.

Buesseler, K.O., 1998. The decoupling of production and particulate export in thesurface ocean. Global Biogeochemical Cycles 12 (2), 297–310.

Buesseler, K.O., Boyd, P.W., 2009. Shedding light on processes that control particleexport and flux attenuation in the twilight zone of the open ocean. Limnologyand Oceanography 54 (4), 1210–1232.

Cahoon, L.B., Laws, R.A., Thomas, C.J., 1994. Viable diatoms and chlorophyllain continental slope sediments off Cape Hatteras, North Carolina. Deep SeaResearch Part II: Topical Studies in Oceanography 41 (4-6), 767–782.

Clarke, K.R., Gorley, R.N., 2006. PRIMER v6: User Manual/Tutorial. PRIMER-E,Plymouth.

De La Rocha, C.L., Passow, U., 2007. Factors influencing the sinking of POC and theefficiency of the biological carbon pump. Deep Sea Research Part II 54 (5),639–658.

Degerlund, M., Eilertsen, H.C., 2010. Main species characteristics of phytoplanktonspring blooms in NE Atlantic and Arctic waters (681–801N). Estuaries andCoasts 33 (2), 242–269.

DeVries, T., Primeau, F., Deutsch, C., 2012. The sequestration efficiency of thebiological pump. Geophysical Research Letters 39 (L13601).

Doucette, G.J., Fryxell, G.A., 1983. Thalassiosira antarctica: vegetative and restingstage chemical composition of an ice-related marine diatom. Marine Biology 78(1), 1–6.

Ducklow, H.W., Steinberg, D.K., Buesseler, K.O., 2001. Upper ocean carbon exportand the biological pump. Oceanography 14 (4), 50–58.

French, F.W., Hargraves, P.E., 1980. Physiological characteristics of plankton diatomresting spores. Marine Biology Letters 1, 185–195.

Fujii, R., Matsuoka, K., 2006. Seasonal change of dinoflagellates cyst flux collected ina sediment trap in Omura Bay, West Japan. Journal of Plankton Research 28 (2),131–147.

Garrison, D.L., 1981. Monterey Bay phytoplankton. II. Resting spore cycles in coastaldiatom populations. Journal of Plankton Research 3 (1), 137–156.

Godhe, A., Noren, F., Kuylenstierna, M., Ekberg, C., Karlson, B., 2001. Relationshipbetween planktonic dinoflagellate abundance, cysts recovered in sedimenttraps and environmental factors in the Gullmar Fjord, Sweden. Journal ofPlankton Research 23 (9), 923–938.

Goldman, J.C., 1993. Potential role of large oceanic diatoms in new primaryproduction. Deep Sea Research Part I 40 (1), 159–168.

Gordon, L.I., Jennings, J.C.J., Ross, A.A., Krest, J.M., 1994. A suggested protocol forcontinuous flow automated analysis of seawater nutrients (phosphate, nitrate,nitrite and silicic acid). The WOCE Hydrographic Program and the Joint GlobalOcean Fluxes Study, WOCE Operations Manual vol. 3: The ObservationalProgram, Section 3.1: WOCE Hydrographic Program, Part 3.1.3: WHP Operationsand Methods, Woods Hole, Massachusetts, pp. 52.

Grimm, K.A., Lange, C.B., Gill, A.S., 1997. Self-sedimentation of phytoplanktonblooms in the geologic record. Sedimentary Geology 110 (3), 151–161.

Guillard, R.R.L., 1975. Culture of phytoplankton for feeding marine invertebrates. In:Smith, W.L., Chanley, M.H. (Eds.), Culture of Marine Invertebrate Animals.Plenum Press, New York, USA, pp. 29–60.

Hargraves, P.E., 1979. Studies on marine plankton diatoms IV. Morphology ofChaetoceros resting spores. Beiheft zur Nova Hedwigia 64, 99–120.

Hargraves, P.E., French, F.W., 1983. Diatom resting spores: significance andstrategies. In: Fryxell, G. (Ed.), Survival Strategies of the Algae. CambridgeUniversity Press, New York, pp. 49–68.

Heiskanen, A.S., 1993. Mass encystment and sinking of dinoflagellates during aspring bloom. Marine Biology 116 (1), 161–167.

Henson, S.A., Dunne, J.P., Sarmiento, J.L., 2009. Decadal variability in North Atlanticphytoplankton blooms. Journal of Geophysical Research: Oceans 114, C04013(doi: 04010.01029/02008JC005139).

Hollibaugh, J.T., Seibert, D.L.R., Thomas, W.H., 1981. Observations on the survivaland germination of resting spores of three Chaetoceros (Bacillariophyceae)species. Journal of Phycology 17 (1), 1–9.

Ishii, K.-I., Iwataki, M., Matsuoka, K., Imai, I., 2011. Proposal of identification criteriafor resting spores of Chaetoceros species (Bacillariophyceae) from a temperatecoastal sea. Phycologia 50 (4), 351–362.

Jensen, K.G., Moestrup, Ø., 1998. The genus Chaetoceros (Bacillariophyceae) in innerDanish coastal waters. Nordic Journal of Botany 18 (1), 88.


136

Running head


Kallin, E., Cetinić, I., Sauer, M., Perry, M.J., 2011. The 2008 North Atlantic BloomExperiment Calibration Report #6, Laboratory Analysis Report, http://data.bco-dmo.org/NAB08/Laboratory_analysis_report-NAB08.pdf.

Karl, D.M., Church, M.J., Dore, J.E., Letelier, R.M., Mahaffey, C., 2012. Predictable andefficient carbon sequestration in the North Pacific Ocean supported bysymbiotic nitrogen fixation. Proceedings of the National Academy of Sciences109 (6), 1842–1849.

Kiørboe, T., Hansen, J.L.S., 1993. Phytoplankton aggregate formation: observationsof patterns and mechanisms of cell sticking and the significance of exopoly-meric material. Journal of Plankton Research 15, 993–1018.

Kohfeld, K.E., Quéré, C.L., Harrison, S.P., Anderson, R.F., 2005. Role of marine biologyin glacial–interglacial CO2 cycles. Science 308 (5718), 74–78.

Kolber, Z., Zehr, J., Falkowski, P.G., 1998. Effects of growth irradiance and nitrogenlimitation on photosynthetic energy conversion in photosystem II. PlantPhysiology 88, 923–929.

Kuwata, A., Hama, T., Takahashi, M., 1993. Ecophysiological characterization of twolife forms, resting spores and resting cells, of a marine planktonic diatom,Chaetoceros pseudocurvisetus, formed under nutrient depletion. Marine EcologyProgress Series 102, 245–255.

Kuwata, A., Takahashi, M., 1990. Life-form population responses of a marineplanktonic diatom, Chaetoceros pseudocurvisetus, to oligotrophication in region-ally upwelled water. Marine Biology 107 (3), 503–512.

Kuwata, A., Tsuda, A., 2005. Selection and viability after ingestion of vegetativecells, resting spores and resting cells of the marine diatom, Chaetocerospseudocurvisetus, by two copepods. Journal of Experimental Marine Biologyand Ecology 322 (2), 143–151.

Lampitt, R.S., Boorman, B., Brown, L., Lucas, M., Salter, I., Sanders, R., Saw, K.,Seeyave, S., Thomalla, S.J., Turnewitsch, R., 2008. Particle export from theeuphotic zone: Estimates using a novel drifting sediment trap, 234Th and newproduction. Deep Sea Research Part 1 55 (11), 1484–1502.

Legendre, L., Rivkin, R.B., 2002. Fluxes of carbon in the upper ocean: regulation byfood-web control nodes. Marine Ecology Progress Series 242, 95–109.

Longhurst, A., 1998. Ecological Geography of the Sea. Academic Press, San Diego.Mahadevan, A., D'Asaro, E., Lee, C., Perry, M.J., 2012. Eddy-driven stratification

initiates North Atlantic spring phytoplankton blooms. Science 337 (6090),54–58.

Mann, D.G., 1999. The species concept in diatoms. Phycologia 38 (6), 437–495.Martin, P., Lampitt, R.S., Perry, M.J., Sanders, R., Lee, C., D'Asaro, E., 2011. Export and

mesopelagic particle flux during a North Atlantic spring diatom bloom. DeepSea Research Part I 58 (4), 338–349.

Maxwell, K., Johnson, G., 2000. Chorophyll fluorescence—a practical guide. Journalof Experimental Botany 51, 659–668.

McQuoid, M.R., Hobson, L.A., 1996. Diatom resting stages. Journal of Phycology 32(6), 889–902.

Menden-Deuer, S., Lessard, E.J., 2000. Carbon to volume relationships for dino-flagellates, diatoms, and other protist plankton. Limnology and Oceanography45 (3), 569–579.

Moore, C.M., Lucas, M.I., Sanders, R., Davidson, R., 2005. Basin-scale variability ofphytoplankton bio-optical characteristics in relation to bloom state and com-munity structure in the Northeast Atlantic. Deep Sea Research Part I 52 (3),401–419.

Nelson, D.M., Tréguer, P., Brzezinski, M.A., Leynaert, A., Quéguiner, B., 1995.Production and dissolution of biogenic silica in the ocean: revised globalestimates, comparison with regional data and relationship to biogenic sedi-mentation. Global Biogeochemical Cycles 9 (3), 359–372.

Oku, O., Kamatani, A., 1999. Resting spore formation and biochemical compositionof the marine planktonic diatom Chaetoceros pseudocurvisetus in culture:ecological significance of decreased nucleotide content and activation of thexanthophyll cycle by resting spore formation. Marine Biology 135 (3), 425–436.

Passow, U., Alldredge, A.L., 1995. Aggregation of a diatom bloom in a mesocosm:the role of transparent exopolymer particles (TEP). Deep Sea Research Part II 42(1), 99–109.

Pitcher, G.C., 1986. Sedimentary flux and the formation of resting spores of selectedChaetoceros species at two sites in the southern Benguela System. South AfricanJournal of Marine Science 4, 231–244.

Pitcher, G.C., 1990. Phytoplankton seed populations of the Cape Peninsula upwel-ling plume with particular reference to resting spores of Chaetoceros (Bacillar-iophyceae) and their role in seeding upwelling waters. Estuarine, Coastal andShelf Science 31 (3), 283–301.

Pollard, R.T., Read, J.F., Holliday, N.P., Leach, H., 2004. Water masses and circulationpathways through the Iceland Basin during Vivaldi 1996. Journal of GeophysicalResearch: Oceans 109 (C4), C04004.

Pondaven, P., Ragueneau, O., Treguer, P., Hauvespre, A., Dezileau, L., Reyss, J.L., 2000.Resolving the ‘opal paradox’ in the Southern Ocean. Nature 405, 168–172.

Pospelova, V., Esenkulova, S., Johannessen, S.C., O'Brien, M.C., Macdonald, R.W.,2010. Organic-walled dinoflagellate cyst production, composition and flux from1996 to 1998 in the central Strait of Georgia (BC, Canada): a sediment trapstudy. Marine Micropaleontology 75 (1-4), 17–37.

Ragueneau, O., Schultes, S., Bidle, K., Claquin, P., Moriceau, B., 2006. Si and Cinteractions in the world ocean: importance of ecological processes andimplications for the role of diatoms in the biological pump. Global Biogeo-chemical Cycles 20 (4), GB4S02.

Rines, J.E.B., Hargraves, P.E., 1988. The Chaetoceros Ehrenberg (Bacillariophyceae)Flora of Narragansett Bay, Rhode Island, USA. Cramer, Berlin/Stuttgart.

Salter, I., Kemp, A.E.S., Moore, C.M., Lampitt, R.S., Wolff, G.A., Holtvoeth, J., 2012.Diatom resting spore ecology drives enhanced carbon export from a naturallyiron-fertilized bloom in the Southern Ocean. Global Biogeochemical Cycles 26(1), GB1014.

Salter, I., Lampitt, R.S., Sanders, R., Poulton, A., Kemp, A.E.S., Boorman, B., Saw, K.,Pearce, R., 2007. Estimating carbon, silica and diatom export from a naturallyfertilised phytoplankton bloom in the Southern Ocean using PELAGRA: a noveldrifting sediment trap. Deep Sea Research Part II 54 (18–20), 2233–2259.

Sanders, J.G., Cibik, S.J., 1985. Reduction of growth rate and resting spore formationin a marine diatom exposed to low levels of cadmium. Marine EnvironmentalResearch 16, 165–180.

Sarmiento, J.L., Gruber, N. Ocean Biogeochemical Dynamics. Princeton UniversityPress, Princeton, NJ.

Scharek, R., Tupas, L.M., Karl, D.M., 1999. Diatom fluxes to the deep sea in theoligotrophic North Pacific gyre at Station ALOHA. Marine Ecology ProgressSeries 182, 55–67.

Sieracki, M.E., Verity, P.G., Stoecker, D.K., 1993. Plankton community response tosequential silicate and nitrate depletion during the 1989 North Atlantic springbloom. Deep Sea Research Part II 40 (1-2), 213–225.

Smetacek, V., 1985. Role of sinking in diatom life-history cycles: ecological,evolutionary, and geological significance. Marine Biology 84 (3), 239–251.

Smith, C.R., Hoover, D.J., Doan, S.E., Pope, R.H., Demaster, D.J., Dobbs, F.C., Altabet,M.A., 1996. Phytodetritus at the abyssal seafloor across 101 of latitude in thecentral equatorial Pacific. Deep Sea Research Part II 43 (4-6), 1309–1338.

Smith, P., Bogren, K., 2001. Determination of nitrate and/or nitrite in brackish orseawater by flow injection analysis colorimeter: QuickChem Method 31-107-04-1-E, Saline Methods of Analysis. Lachat Instruments, Loveland, CO p. 12.

Sunesen, I., Hernandez-Becerril, D.U., Sar, E.A., 2008. Marine diatoms from BuenosAires coastal waters (Argentina). V. Species of the genus Chaetoceros. Revista deBiología Marina y Oceanografía 43 (2), 303–326.

Suto, I., 2006. The explosive diversification of the diatom genus Chaetoceros acrossthe Eocene/Oligocene and Oligocene/Miocene boundaries in the NorwegianSea. Marine Micropaleontology 58 (4), 259–269.

Takahashi, T., Sutherland, S.C., Wanninkhof, R., Sweeney, C., Feely, R.A., Chipman, D.W., Hales, B., Friederich, G., Chavez, F., Sabine, C., Watson, A., Bakker, D.C.E.,Schuster, U., Metzl, N., Yoshikawa-Inoue, H., Ishii, M., Midorikawa, T., Nojiri, Y.,Körtzinger, A., Steinhoff, T., Hoppema, M., Olafsson, J., Arnarson, T.S., Tilbrook,B., Johannessen, T., Olsen, A., Bellerby, R., Wong, C.S., Delille, B., Bates, N.R., deBaar, H.J.W., 2009. Climatological mean and decadal change in surface oceanpCO2, and net sea–air CO2 flux over the global oceans. Deep Sea Research Part II56 (8–10), 554–577.

Tomas, C.R., 1997. Identifying Marine Phytoplankton. Academic Press, New York.Utermöhl, H., 1958. Zur Vervollkommnung der quantitativen Phytoplankton-

Methodik. Mitteilungen—Internationale Vereinigung Für Theoretische undAngewandte Limnologie 9, 1–38.

Volk, T., Hoffert, M.I., 1985. Ocean carbon pumps: analysis of relative strengths andefficiencies in ocean-driven atmospheric CO2 changes. In: Sundquist, E.T.,Broecker, W.S. (Eds.), The Carbon Cycle and Atmospheric CO2: Natural Varia-tions Archean to Present. AGU, Washington, DC, pp. 99–110.

Waite, A., Fisher, A., Thompson, P.A., Harrison, P.J., 1997. Sinking rate versus cellvolume relationships illuminate sinking rate control mechanisms in marinediatoms. Marine Ecology Progress Series 157, 97–108.

Wolters, M., 2002. Determination of silicate in brackish or seawater by flowinjection analysis: QuickChemMethod 31-114-27-1-D, Methods Manual. LachatInstruments, Loveland, CO p. 12.

Zar, J.H., 1996. Biostatistical Analysis. Prentice Hall, Inc., Englewood Cliffs, NewJersey.



Appendix

Appendix: Conference contributions

List of abstracts

Abstract 1 (English):

PRIMARY PRODUCTION IN THE BALTIC SEA TRANSITION ZONE – SEASONAL PATTERNS AND VERTICAL DISTRIBUTION TODAY AND WITH A CHANGING CLIMATEpresented orally at the Nordic Marine Science Conference in Strömstad, Sweden the 13th-16th of September 2010.

Abstract 2 (Danish):

PRIMÆRPRODUKTION I DE INDRE DANSKE FARVANDE: BETYD-NINGEN AF DYB PRIMÆRPRODUKTION I RELATION TIL FORÅR-SOPBLOMSTRING OG ÅRLIG PRIMÆRPRODUKTIONOrally presented at the Danish Marine Science meeting the 18th-20th of January 2011 in Ebeltoft, Denmark.


VERTICAL, HORIZONTAL AND TEMPORAL DISTRIBUTION PAT-TERNS IN PRIMARY PRODUCTIONPresented orally at the Marine Strategy conference the 14th-16th of May 2012.


SUB-PYCNOCLINE PRIMARY PRODUCTION ENHANCES SEDIMEN-TATION: A STUDY ON A TEMPERATE STRATIFIED MARINE SYSTEMOral presentation at the NAACOS seminar on the biological pump and its (bio-logical) control. Copenhagen, Denmark the 5th of December 2012.


PRIMÆRPRODUKTIONEN RESPONDERER PÅ DANMARKS REDU-CEREDE NÆRINGSTILFØRSLER FRA LAND TIL HAVOrally presented at the Danish Marine Science meeting the 21th-23th of January 2013 in Roskilde, Denmark.


THE VERTICAL DISTRIBUTION OF PRIMARY PRODUCTION MAY CHANGE IN RESPONSE TO REDUCED NUTRIENT LOADINGPresented orally at the ASLO conference in New Orleans, USA the 17th-22th of February 2013.


Appendix

Abstracts in collaboration with other researchers


PRIMARY PRODUCTION PROVIDED BY A 3D ECOSYSTEM MODEL COVERING THE NORTH SEA – BALTIC SEA AS SUPPORT TO THE MSFDPresented by Marie Maar orally at the Marine Strategy conference the 14th-16th of May 2012.


EUTROPHICATION AND FISH PRODUCTION: A CHALLENGE FOR SCIENCE AND MANGEMENT IN COASTAL SYSTEMSPresented by Stiig Markager orally at the BSSC conference the 26h-29th of August 2013.

Abstracts


PRIMARY PRODUCTION IN THE BALTIC SEA TRANSITION ZONE – SEASONAL PATTERNS AND VERTICAL DISTRIBUTION TODAY AND WITH A CHANGING CLIMATE

Maren Moltke Lyngsgaard, Stiig Markager and Katherine Richardson

Climate change is predicted to increase temperatures and precipitation in the Bal-tic Sea area. A higher freshwater outfl ow from the Baltic Sea will change the hy-drological characteristics in the area including, strength, frequency and depth of the pycnocline. The position of the pycnocline relative to the depth of the photic zone could be the key factor determining the effects of climate change on primary production (PP) in this stratifi ed area.

This study investigates the seasonal patterns and vertical distribution of the PP from 1998 to 2008 in the Baltic Sea transition zone. Pycnocline depth and area primary production were calculated for several stations based on data from the Danish monitoring program including CTD-profi les, water chemistry and PP characteristics from the 14C-technique. Area PP shows highest monthly mean val-ues in spring and summer with inter annual variation highest in spring months (February and March). Furthermore the study showed that 10-15 % of the annual area PP was produced below the mixed layer with the largest contribution (up to 60 %) coming from the summer months (May-July). This emphasize the impor-tance of PP associated with the pycnocline and hence its position relative to the photic zone. A sensitivity study showed that small changes in the interpolation of PP characteristics (alpha, Pmax ) signifi cantly altered the value for area PP. These results stress the importance of having consensus in the algorithms when analys-ing PP data in relation to climate change. The results furthermore indicate a great loss of PP if the pycnocline depth increases.


Appendix


PRIMÆRPRODUKTION I DE INDRE DANSKE FARVANDE:BETYDNINGEN AF DYB PRIMÆRPRODUKTION I RELATION TIL FORÅRSOPBLOMSTRING OG ÅRLIG PRIMÆRPRODUKTION


Klimaændringer vil forårsage højere temperaturer og øget nedbør i Østersøen. En højere ferskvandsudstrømning fra Østersøen vil ændre de hydrologiske forhold i de indre danske farvande inklusiv styrke, frekvens og dybde på lagdelingen. Placeringen af lagdelingen i forhold til dybden på den eufotiske zone kan være en afgørende faktor, når effekten af klimaændringer på fremtidig primærproduktion (PP) skal bestemmes. Dette studie undersøger årstidsvariation og vertikal forde-ling af PP fra 1998 til 2009 i de indre danske farvande. Dybde på lagdelingen er defi neret og areal PP beregnet for fl ere stationer ud fra CTD profi ler, vandkemi og PP parametre (14C metode) målt som et led i den Danske overvågning. Areal PP viser høje månedlige middelværdier i forår og sommer, med en høj variation mel-lem årene. Variationen mellem år er særlig høj under foråret (februar og marts). Middelværdier for alle stationer viser at 15 % af den årlige PP bliver produceret under den opblandede zone, hvoraf størstedelen kommer fra de lagdelte som-mermåneder (april – august). Dette understreger betydningen af PP i- og omkring lagdelingen og derfor dennes position i forhold til den eufotiske zone. Forårsop-blomstringens tilskud til den årlige areal PP udgør 10-15 %, hvilket svarer til det samme eller mindre end tilskuddet fra den dybe PP. Det store tilskud fra de dybere lag indikerer en betydelig reduktion i den årlige PP hvis 1) lagdelingens dybde øges i sommermåneder og/eller 2) lys dæmpningen øges som et resultat af en øget mængde opløst organisk materiale i den eufotiske zone.


VERTICAL, HORIZONTAL AND TEMPORAL DISTRIBUTIONPATTERNS IN PRIMARY PRODUCTION


Primary production informs about the organic input and oxygen production in a water body. Both parameters are important in determining the quality of marine waters or the potential for a rich fi sheries industry. The concentration of chloro-phyll a in the water is often used as an estimate for primary production, however, this measurement can only tell us the ‘end product’ from a long and complicated process, where a large part of the plant material is recycled. It is, therefore, more accurate and informative to measure the primary production rate in the process of estimating the amount of organic material and oxygen produced.

In this study we analysed 1182 coastal water primary production measurements from the time period from 1998 to 2009. We expected primary production to be greatest during the spring bloom period. However, we found a general seasonal var-iation with the highest production during summer months. Production at this time was highly supported by a deeper primary production from below the pycnocline. This sub-pycnocline primary production accounts for app. 7-29 % of the annual pri-mary production depending on the hydrodynamics in the area. For the specifi c study area (Baltic Sea transition zone), it accounts for almost half of the oxygen consump-tion in the waters below the pycnocline. This study shows that when including the sub-pycnocline primary production, not only the annual estimate increases, but also the seasonal variation changes. We therefore strongly suggest that no survey pro-grams are based only on surface primary production measurements.


Appendix


SUB-PYCNOCLINE PRIMARY PRODUCTION ENHANCES SEDIMEN-TATION: A STUDY ON A TEMPERATE STRATIFIED MARINE SYSTEM

Maren Moltke Lyngsgaard, Katherine Richardson, Stiig Markager and Michael Olesen

Analysis of the vertical distribution of primary production (1254 14C measure-ments) in relation to density profi les, was carried out for the time period from 1998-2011 at six different stations within the Baltic Sea transition zone. The re-sults showed that primary production below the pycnocline (sub-pycnocline pri-mary production, SPP) contributes approximately 20 % to the annual primary pro-duction in the region. Furthermore, the effect of nitrogen load (tonnes of nitrogen year-1) on the vertical distribution of primary production was analysed. Results showed that reduced N-loadings not only decrease the primary production occur-ring above the pycnocline, but it also relocates it towards the sub-pycnocline lay-ers. The sub-pycnocline primary production was negatively correlated to nitrogen loading (R2 = 0.74, p < 0.005). The relocation means that more oxygen will be produced in sub-pycnocline waters. The vertical distribution of sedimentation in relation to primary production and size-fractionation of phytoplankton was ex-amined at one of the stations from the analysis above in year 2010. The results showed that phytoplankton size- and sedimentation was signifi cantly higher for the sub-pycnocline layer compared to the mixed layer. Consequently, the export of organic material may remain the same despite the signifi cant decrease in N-loadings, seen over the past decade for the region, if more production is allocated towards sub-pycnocline layers.


PRIMÆRPRODUKTIONEN RESPONDERER PÅ DANMARKS REDU-CEREDE NÆRINGSTILFØRSLER FRA LAND TIL HAV


Danmark opnåede i 2011 at have reduceret kvælstof tilførslen, fra land til hav, med 46 % sammenlignet med perioden før 1994.

Vi kan i dette studie vise, at der er en statistisk signifi kant positiv sammenhæng mellem kvælstoftilførsel og størrelsen på den årlige primærproduktion (R2=0.33, p<0.05). Den tydeligste sammenhæng fås når kvælstoftilførslen beregnes som summen af tilførsler fra september året før til marts måned. Desuden viser studiet, at maksimum primærproduktion i de indre danske farvande fi nder sted midt på sommeren.

Den vertikale fordeling af primærproduktion responderer også på ændringen i kvælstoftilførsel. Ved at differentiere mellem primærproduktion over- og under springlaget viser resultaterne af vores analyse en signifi kant negativ sammenhæng mellem den årlige produktion under springlaget og kvælstof tilførslen (R2=0.57, p<0.05). Dette betyder at en reducering i kvælstof tilførsel introducerer en om-fordeling af primærproduktionen, fra at fi nde sted i det opblandede lag til at fi nde sted under springlaget.

Et nærstudie af sedimentationen i Aarhus bugt (sommeren 2010) viser at en langt større andel af produktionen synker ud når produktionen sker i og under spring-laget set i forhold til en produktion oppe i den opblandede del af vandsøjlen (en faktor 2-3). Derfor giver en allokering af primærproduktion fra den opblandede del af vandsøjlen til laget i-og under springlaget anledning til en større sedimente-


Appendix

ring samlet set. Og det kan derfor lade sig gøre at opretholde den samme fl uks af organisk materiale til bunden trods en kraftig nedgang i tilførslen af næringssalte og den samlede primærproduktion. Omfordelingen betyder samtidig en væsentlig iltproduktion unders springlaget.

Den statistiske analyse er baseret på 1254 primærproduktions målinger fra 6 sta-tioner (Aarhus bugt, Lillebælt, Storebælt (2 stationer), Øresund og Ålborg bugt) i perioden fra 1998-2011 fra MADS databasen, Bioscience, Aarhus Universitet.


THE VERTICAL DISTRIBUTION OF PRIMARY PRODUCTION MAY CHANGE IN RESPONSE TO REDUCED NUTRIENT LOADING


Danish coastal waters have experienced serious hypoxia during recent decades in response to eutrophication. From the mid 80s to 2007, management initiatives have succeeded in reducing nutrient loadings into the stratifi ed Danish coastal waters by 49 % for nitrogen and 87% for phosphorous. This reduction in nutrient loading does not, however, seem to have resulted in lower primary production. This study demonstrates that, while depth-integrated primary production has not changed, there has been a signifi cant increase in the percentage of the primary production occurring below the pycnocline during the period 1998 to 2011. 1254 14C primary production measurements show that sub-pycnocline primary produc-tion constitutes approximately 20 % of the annual primary production in these coastal waters. The magnitude of the sub-pycnocline production affects oxygen concentration in the bottom water, sedimentation and food quality for benthic organisms.


PRIMARY PRODUCTION PROVIDED BY A 3D ECOSYSTEM MODEL COVERING THE NORTH SEA – BALTIC SEA AS SUPPORT TO THE MSFD

Marie Maar, Stiig Markager, Eva Friis Møller, Maren M. Lyngsgaard, Peter Henriksen

The use of models is essential for the MSFD as a tool that can support monitoring activities, the setup of indicators and make scenarios for reaching the ecological goals in MSFD. However, models are never better than the data for their calibra-tion. In this respect, it is essential to calibrate and validate the models to rate measurements and not just standing stocks. In Danish waters the primary produc-tion is the only rate measured regularly. We have therefore optimized the biogeo-chemical ERGOM model for primary production using the thousands of primary production measurements available for the Kattegat/Belt Sea area and experimen-tal data. ERGOM was also modifi ed to be valid for both the North Sea – Baltic Sea by adding silicate, modifying zooplankton grazing impact and adjusting light parameterizations and is one of the few models that are validated for both areas. ERGOM is coupled to the 3d circulation model HBM (Hiromb-BOOS model) with a fi ne two-way nested grid in the Kattegat to Arkona Basin.


Appendix


EUTROPHICATION AND FISH PRODUCTION: A CHALLENGE FOR SCIENCE AND MANGEMENT IN COASTAL SYSTEMS

Stiig Markager, Maren Moltke Lyngsgaard, J. Rasmus Nielsen and Katherine Richardson

Increasing inputs of nutrients tends to enhance phytoplankton primary produc-tion in coastal areas starting a cascade of negative effects in the ecosystem. Often it is assumed that these also include a decrease in fi sh production with negative effects on commercial catches. However, this seems like a paradox as the sim-plest assumption would be that a higher production at the bottom of the food web also should enhance production at the top. In this paper, we quantify the positive relationship between nutrient loadings to the Kattegat/Belt Sea area and phyto-plankton primary production. Moreover, we show that higher loadings introduce a shift in the vertical distribution of phytoplankton production, so that more of the primary production takes place above the pycnocline rather than below. We pro-pose that this shift in vertical distribution has implications for sedimentation, food availability for the benthic fauna, fi sh production and oxygen in the bottom water. Together this causes a cascade mechanism that potentially can explain a negative relationship between nutrient inputs and fi sh production in coastal ecosystems. Such coupled mechanisms are crucial to understand when we aim at an integrated coastal zone management or an ecosystem approach to management as is the goal for EU’s Marine Framework Strategy Directive

VERTICAL DISTRIBUTION OF PELAGIC PHOTOSYNTHESIS– implications for marine ecosystem dynamics

Because the tiny plants in the ocean (phytoplankton) need light to carry out photosynthesis, it has generally been assumed that it is phytoplankton distribution and activity near the surface that determines ecological structure and function in marine environments. This PhD focusses on the importance of understanding the vertical distribution of phytoplankton and their activity in controlling the dynamics of marine systems. In particular, this PhD provides important new understan-ding of how nutrient enrichment of coastal ecosystems through human activities infl uences the function of marine ecosystems – understanding that can be useful in assessing the success of efforts to reduce impacts of nutrient enrichment. In addition, the PhD provides examples of cases where the vertical structure of phy-toplankton activity can be important in terms of ocean carbon storage and fi sheries.

Vertical distribution of pelagic photosynthesis – implications for m

arine ecosystem dynam

icsM

aren Moltke Lyngsgaard • P

hD thesis • 2013

PhD thesis 2013Maren Moltke Lyngsgaard


ISBN: xxxxxxxxxxxxxxx

Documents

VERTICAL DISTRIBUTION OF PELAGIC PHOTOSYNTHESIS Moltke Lyngsgaard.pdf · 2014. 1. 20. · PhD thesis by Maren Moltke Lyngsgaard 5 Acknowledgements First of all, I would like to thank