Seasonal patterns of biogeochemical conditions of the

Seasonal patterns of biogeochemical conditions of the water column and sediment - water interface near the submarine outfall in the Santa Marta Bay, Colombian

Caribbean

Diana Marcela Arroyave Gómez

Universidad Nacional de Colombia Facultas de Minas, Departamento de Geociencias y Medio Ambiente

Medellín, Colombia

2020

CONDICIONES BIOGEOQUÍMICAS EN LA COLUMNA DE AGUA Y EN LA INTERFACE SEDIMENTO - AGUA EN LAS CERCANÍAS DEL EMISARIO SUBMARINO DE

LA BAHÍA DE SANTA MARTA (CARIBE - COLOMBIANO) PRODUCIDAS POR PATRONES

CLIMÁTICOS ESTACIONALES.


Universidad Nacional de Colombia Facultas de Minas, Departamento de Geociencias y Medio Ambiente

Medellín, Colombia

2020

Seasonal patterns of biogeochemical conditions of the water column and sediment - water interface near the submarine outfall in the Santa Marta Bay, Colombian

Caribbean


This thesis is presented for the degree of Doctor en Ciencias del Mar

Adviser: Ph.D. Francisco Mauricio Toro Botero

Universidad Nacional de Colombia

Co-adviser (a):

Ph.D. Marco Bartoli

Parma University, Italy

Universidad Nacional de Colombia Facultad de Minas, Departamento de Geociencias y Medio Ambiente

Medellín, Colombia

2020

I

Resumen

Las condiciones biogeoquímicas en la interface sedimento-agua y en la columna de agua cerca de

la descarga del emisario de aguas residuales de Santa Marta (ESM) fueron estudiadas en el periodo

de surgencia (S) y no surgencia (NS) mediante mediciones de propiedades sedimentarias y flujos

de nutrientes bentónicos, así como, con la implementación del modelo acoplado hidrodinámico-

ecológico AEM3D. Las propiedades sedimentarias (materia orgánica, contenido de C, N, P, δ13C

y δ15N, y potencial redox) y el metabolismo béntico (respiración aerobia, desnitrificación, nitrato

amonificación y reciclaje de nutrientes) fueron analizadas en cuatro estaciones ubicadas en la

proximidad, a 100 m, 750 m y 1800 m de distancia de la descarga del efluente de aguas residuales

no tratadas en ambos periodos climáticos en el Área Costera de Santa Marta (ACSM). En cada

sitio, se muestrearon núcleos de sedimento a profundidades entre 20 m y 30 m. Luego, los flujos

de nutrientes fueron medidos en el laboratorio vía incubaciones oscuras; en paralelo a los flujos, se

midieron la desnitrificación y la reducción desasimilatoria del nitrato a amonio vía IPT (por sus

siglas en inglés, Isotope Pairing Tecnnique). Los resultados indican que los sedimentos permitieron

trazar el impacto del emisario (a 750 m y 1800 m con una contribución del carbono orgánico

terrestre del ∼ 40 y ~20 %, respectivamente). Los resultados sugieren altas demandas de oxígeno

de los sedimentos en la proximidad del emisario, así como una supresión de la desnitrificación e

incrementos en la libración de nitrógeno amoniacal por medio reducción desasimilatoria de nitrato

a amonio (por sus siglas en inglés, DNRA), el cual se vió incrementado durante el período de

surgencia.

Por otra parte, el modelo AEM3D fue aplicado para analizar las variaciones estacionales de los

parámetros fisicoquímicos y biológicos en la columna de agua del ACSM bajo dos cargas

diferentes de nutrientes y materia orgánica provenientes del efluente del agua residual del emisario

(caudal de 1.0 m3 s-1 y 2.5 m3 s-1) durante los periodos de NS y S. El modelo fue configurado,

calibrado y validado con base en mediciones de metabolismo béntico obtenida dentro del periodo

de simulación, imágenes satelitales de temperatura superficial del mar (TSM) y clorofila-a (Chl-

a), información de base datos de HYCOM, campañas de campo y literatura. El modelo fue capaz

de reproducir la magnitud y la compleja dinámica de las rápidas transiciones de temperatura,

nutrientes y fitoplancton, incluido el tiempo y la duración de los períodos de estratificación y

II

mezcla durante las temporadas NS y S. El modelo también pudo capturar el efecto de la fertilización

de la surgencia y del emisario. El campo de viento fue el principal forzante de la hidrodinámica

costera y la dispersión de la pluma. Los tiempos de residencia promedio más bajos de la pluma (3.7

± 0.4 días) correspondieron al período de mayor intensidad de la surgencia. La temperatura, la luz

y los nutrientes fueron los factores que limitaron el crecimiento del fitoplancton. Las

concentraciones carbono orgánico total (COT), fósforo total (TP) y fosfato (PO43-) aumentaron

levemente en los dos escenarios de carga de aguas residuales. El crecimiento del fitoplancton fue

limitado tanto en el periodo de NS como en S debido a los grandes cambios de temperatura y

advección y mezcla en el área costera, lo que resultó en una gran dilución de las cargas de

nutrientes. Los grandes y rápidos cambios en la temperatura y el ambiente altamente energético

desacoplaron el crecimiento del fitoplancton con el suministro de nutrientes en los compartimentos

bentónico y pelágico. El modelo demostró ser una herramienta de investigación y de gestión

razonablemente confiable para predecir la dinámica de nutrientes y fitoplancton, y para analizar el

papel individual de diferentes entradas de nutrientes durante los periodos de NS y S. El principal

resultado del modelo sugiere un impacto limitado de los nutrientes del emisario y de la surgencia

sobre la calidad química y biológica del agua en el ACMS. Sin embargo, los análisis de sedimentos

revelaron la ocurrencia de un impacto orgánico pronunciado, alterando la dinámica biogeoquímica

de los sedimentos y sugieren mantener este sistema continuamente monitoreado y estudiado, a

través de una combinación de actividades experimentales, monitoreo basado en imágenes de

satélites y enfoques de modelación. Esto parece particularmente importante debido a las crecientes

presiones antropogénicas en las áreas costeras y en las cuencas hidrográficas, y a los cambios

globales en curso que afectan el clima, la intensidad del viento, la temperatura del agua y las tasas

de mezcla.

Palabras claves: emisario de aguas residuales, metabolismo béntico, surgencia, tiempo de

residencia, AEM3D, desnitrificación y nitrato amonificación.

III

Abstract

The biogeochemical conditions at the sediment-water interface and along the water column near

the discharge of the Santa Marta sewage outfall (SMSO) were studied during the non upwelling

(NUPW) and upwelling (UPW) seasons by sedimentary properties and benthic metabolism

measurements, as well as, by the implementation of a coupled 3D hydrodynamic-ecological model

(AEM3D). Sediment properties (organic matter quantity, C, N and P pools and δ13C, δ15N and

redox potential) and benthic metabolism (aerobic respiration, denitrification, nitrate

ammonification and nutrient recycling) were analyzed in four stations located in the proximity and

100, 750 and 1800 m far from the untreated wastewater effluent discharge in both seasons in the

Santa Marta Coastal Area (SMCA). From each site, sediment cores were collected between 20 and

30 m depth. Then, the nutrient fluxes were measured in the laboratory via dark incubations;

sequentially to fluxes denitrification and dissimilative nitrate reduction to ammonium were

measured via the r-IPT (Isotope Pairing Tecnnique). The results indicate that the sediments trace

the impact of the outfall (at 750 m and 1800 m with a contribution of terrestrial organic carbon of

~ 40 and ~ 20%, respectively). The results suggest significantly higher sediment oxygen demands

(SOD) in the outfall vicinity, as well as a suppression of denitrification and increments in the

ammonia nitrogen release through disassimilatory reduction of nitrate to ammonium (DNRA),

which was increased during the UPW season.

On the other hand, AEM3D model was applied to analyze the seasonal variations of water physico-

chemical and biological parameters in SMAC under two different nutrient and organic matter loads

from wastewater outfall (flow-rate of 1.0 m3 s-1 and 2.5 m3 s-1) and along the NUPW and UPW

season. The model was set up, calibrated and validated based on benthic metabolic measurements

carried out within the simulation period, satellite–derived chlorophyll-a (Chl-a) and sea surface

temperature (SST) maps, HYCOM database and field and literature water quality data. The model

was able to reproduce the magnitude and timing of complex dynamics and fast transitions of

temperature, nutrients, and phytoplankton, including the time and duration of stratification and

mixing periods during the NUPW and UPW seasons. The model was also able to capture the effect

of fertilization from upwelling and from the outfall plume. The wind field was the main driver of

nearshore hydrodynamics and the outfall plume dispersion. The shortest average residence times

IV

of the outfall plume (3.7 ± 0.4 days) corresponded to the period of highest upwelling intensity.

Temperature, light intensity and nutrients were the factors that limited phytoplankton growth. The

plume concentrations of TOC, TP and PO43- increased slightly under two scenarios of different

wastewater loading. The phytoplankton growth was limited in both NUPW and UPW seasons due

to large changes in temperature and advection and mixing in the coastal area, resulting in large

dilution of nutrient loads. Wide and fast changes in the temperatures and the highly energetic

environment uncoupled phytoplankton growth and nutrient supply in the benthic and pelagic

compartments. The model proved to be a reasonably reliable research and management tool to

predict nutrient and phytoplankton dynamics, and to analyze the individual role of different inputs

during NUPW and UPW seasons. The main model outputs suggest limited impact of the nutrients

from the outfall and from upwelling to the chemical and biological quality of the water in the

SMCA. However, sediment analyses revealed the occurrence of a pronounced organic impact,

altering sediment biogeochemical dynamics and suggest maintaining this system continuously

monitored and studied, via combination of experimental activities, satellite-based monitoring and

modeling approaches. This seems particularly important due to increasing anthropogenic pressures

on the coastal area and on watersheds and to ongoing global changes affecting climate, wind

intensity, water temperature and mixing rates.

Keywords: sewage outfall, benthic metabolism, upwelling, residence time, AEM3D,

denitrification and nitrate ammonification.

V

Statement of the candidate’s contribution

This thesis was completed during my candidature for the degree of Doctor en Ciencias del Mar at

The Universidad Nacional de Colombia, Medellín campus, and during a research stay in the Parma

University, Italy. The main content of this thesis (Chapter 2 and 3) is a compilation of two

manuscripts prepared for publication as standalone journal articles. The content of this thesis,

including field and laboratory work, statistical analysis, modelling, interpretations and writing, is

based on my own ideas and work, undertaken under the supervision of Prof. Francisco Mauricio

Toro Botero, PhD, of Universidad Nacional de Colombia and Prof. Marco Bartoli, PhD, of Parma

University. Co-authors of the papers listed below have contributed with discussions and revisions

to the manuscripts. Here I provide a declaration of my contribution to each of those publications.

Chapter 2 has been published as “Arroyave Gómez D.M., Gallego Suárez D., Bartoli M. and Toro-

Botero M. 2020; Spatial and seasonal variability of sedimentary features and nitrogen benthic

metabolism in a tropical coastal area (Taganga Bay, Colombia Caribbean) impacted by a sewage

Outfall; Biogeochemistry. 150:85–107”. I conducted the planning and carrying out the field

campaigns, laboratory work, data processing and statistical analysis under close supervision of my

advisers and Prof. Gallego. Prof. Bartoli participated in the conducted the fieldwork and laboratory

works for data collection of sediment cores. The initial manuscript was written by myself and

carefully revised and edited by Prof. Marco Bartoli. During the development of field campaigns, a

system for the incubation of large sediments cores was also developed (see Final Appendices)

together with Ph.D. Fabio Alexander Suarez of the company “Faro Tecnológico” and tested in field

campaigns. Finally, the data of these cores were not included in the final results due to the lack of

representation (few cores). Nevertheless, this was a good research exercise for the future work.

Chapter 3 has been submitted to the journal Marine Pollution Bulletin as: “Arroyave Gómez D.M.,

Bartoli M., Bresciani M., Luciani G. and Toro-Botero M. “Biogeochemical modeling of a tropical

coastal area undergoing seasonal upwelling and impacted by untreated submarine outfall”. I

conducted the field campaigns to obtain water quality data, as well as, data processing, statistical

analysis and set up, calibration and validation of 3D-coupled hydrodynamic-ecological model for

the Santa Marta Coastal Area (SMCA) under close supervision of my co-advisers. The initial

VI

manuscript was written by myself and carefully edited by my advisers and co-authors. The satellite

images were provided by Dr. Mariano Bresciani of the Istituto per il Rilevamento Elettromagnetico

dell'Ambiente (Milan, Italy) and Dr. Giulia Luciani of the Politecnico di Milano - Lecco Campus,

Department of Civil and Environmental Engineering. I processed the satellite images to include in

the model set up and analyzed the confidence of the simulations.

Thanks to Prof. Marco, during my research stay in his aquatic ecology laboratory at the University

of Parma, I had the opportunity of to work on a parallel project and a paper in which is proposed a

method to measure the nitrification in bioturbated sediments by coupling anoxic sediment slurries

and membrane inlet mass spectrometry (MIMS) to quantify 15N-NO3- and 14N-NO3

- in water

samples. My participation in the paper was as co-author. The paper was published as Paula

Carpintero Moraes, Diana Marcela Arroyave Gómez, Fabio Vincenzi, Giuseppe Castaldelli, Elisa

Anna Fano, Marco Bartoli and Sara Benelli. 2019. Analysis of 15NO3- via anoxic slurries coupled

to MIMS analysis: an application to estimate nitrification by burrowing macrofauna. Water, 11,

2310; doi:10.3390/w11112310 (see Final Appendices).

VII

Acknowledgements

I would like to thank to my advisers, Associate Prof. Mauricio Toro Botero (Universidad Nacional

de Colombia, campus Medellín) and Associate Prof. Marco Bartoli (Parma University, Italy), as

well as Associate Prof. Darío Gallego Suárez (Universidad Nacional de Colombia, campus

Medellín) for their guidance, support and enthusiastic motivation throughout my PhD study. In

particular, I would like to thank Mauricio, for encouraging me to pursue the PhD in marine science,

for teaching me mathematical modeling and supporting my ideas. I am very grateful to Marco, who

in spite of his busy schedule accepted Mauricio's invitation to be my co-advisor on the subject of

biogeochemistry and benthic metabolism and he came to the Santa Marta city to teach me about

sampling and benthic measurements in sediment cores. The November field trip at Santa Marta

city and my research stay at the Parma University under his guidance was a great research

experience. I am also grateful to Darío, for teaching me about water quality and treatment and for

always encouraging me pursue MSc and PhD studies.

Throughout this study, I was supported financially by a Ph.D. scholarship of Doctoral Training

Program from Colombia (COLCIENCIAS – 2014). My thesis was funded by COLCIENCIAS in

the Call No. 714 - 2015 – Research projects, technological development and innovation in

environment, oceans and biodiversity (Contract – FP44842-020-2016). Prof. Marco Bartoli’s trip

to Colombia to carry out field campaigns was supported by the call Fellows Colombia Program

(ICETEX - 2017) – Inter-institutional Ph.D Program – Marine Sciences and Universidad Nacional

de Colombia, campus Medellín (Researchers Mobility Call – 2017). Prof. Bartoli is thanked by

supporting chemistry analyses in Parma University (Italy) and Ferrara University (Italy).

I acknowledge all the people who kindly collaborated with developing of my thesis. Prof. Andres

Franco and Alfonso Gamero of the Universidad Jorge Tadeo Lozano, campus Santa Marta are

acknowledged for providing laboratory facilities and logistic support for this study. Prof. Paola

Iacumin and Dott. Antonietta Di Matteo from Geology Department of Parma University and Irma

Lubiene from Klaipeda University are acknowledged for the technical assistance in isotopic

composition and nutrients analyses of benthic measurements. Prof. Darío Gallego Suarez, Sandra

Milena Franco Rúa and Alexander Hincapie Silva from Sanitary Engineering Laboratory at

VIII

Universidad Nacional de Colombia campus Medellín are acknowledged for the technical assistance

in nutrient analysis of water samples collected for the water quality modeling. The chlorophyll-a

analyzes were carried out by Instituto de Investigaciones Marinas y Costeras José Benito Vives de

Andréis (INVEMAR). I am thankful with Johann Camil Delgado, Maximiliano Arredondo, Lina

Maria Ramírez and Tayrona dive center at Taganga (Jose, Freddy and Wile) for their help in core

collection and sampling. I thank Lina Marcela Jaramillo for administrative assistance for field trips.

I am also grateful to Lina Maria Ramírez and Mateo Novoa Parra who contributed to the

hydrodynamic model configuration with their undergraduate thesis. Fabio Suarez of Faro

Tecnológico company is acknowledged by support in the field campaigns logistic. Dr. Mariano

Bresciani of Istituto per il Rilevamento Elettromagnetico dell'Ambiente (Milan, Italy) and Dr.

Giulia Luciani of the Politecnico di Milano - Lecco Campus, Department of Civil and

Environmental Engineering (Italy) are grateful for providing temperature and chlorophyll-a

remote-sensing data who processed and supplied MODIS and Sentinel -3A data to end products.

To all my fellow students and officemates, thank you for your support and friendship during my

study – Adriana, Mauricio, Juan, Victor, David, Eliana, Maria Angelica, Lina, Angelica, Carlos,

Manuel C., Sebastian G., Alejandro, Juan David, Natalia, Eileen, Luna, Alexander, Ricardo, Mateo

B., Steffani, Camila, Oscar, Manuel C. (and many more). Also, thanks to Sara B. and Erika R. at

Parma University.

My deepest gratitude belongs to my family. I want to thank my parents Nelly and Leonardo, for

their immense love, prayers, caring and sacrifices for educating and preparing me for my future. I

also express my thanks to my brother Gustavo, sister-in-law Nora and niece Sofia for their valuable

support and love.

IX

TABLE OF CONTENTS

CHAPTER 1. .......................................................................................... 22

GENERAL INTRODUCTION ............................................................ 22

1.1. MOTIVATION ............................................................................................................................................... 22 1.2. OBJECTIVES ................................................................................................................................................. 23

1.2.1 General objective ....................................................................................................................................... 23 1.2.2. Specific objectives ..................................................................................................................................... 23

1.3. THESIS OVERVIEW ....................................................................................................................................... 23 1.4. REFERENCES ............................................................................................................................................... 25

CHAPTER 2. .......................................................................................... 26

SPATIAL AND SEASONAL VARIABILITY OF SEDIMENTARY FEATURES AND NITROGEN BENTHIC METABOLISM IN A TROPICAL COASTAL AREA (TAGANGA BAY, COLOMBIA CARIBBEAN) IMPACTED BY A SEWAGE OUTFALL ............... 26

2.1. ABSTRACT ......................................................................................................................................................... 26 2.2. INTRODUCTION .................................................................................................................................................. 27 2.3. MATERIALS AND METHODS ................................................................................................................................ 29

2.3.1. Study area .................................................................................................................................................. 29 2.3.2. Sediment cores collection .......................................................................................................................... 30 2.3.3. Net flux measurement at the sediment-water interface .............................................................................. 31 2.3.4. Denitrification and DNRA measurements ................................................................................................. 33 2.3.5. Sedimentary features ................................................................................................................................. 33 2.3.6. Contribution of terrestrial organic matter to Taganga Bay sediments ..................................................... 34 2.3.7. Statistical Analysis .................................................................................................................................... 35

2.4. RESULTS ............................................................................................................................................................ 36 2.4.1. Water and Sedimentary features................................................................................................................ 36 2.4.2. Organic carbon and nitrogen isotopic composition in the Taganga Bay sediments ................................. 38 2.4.3. Aerobic and anaerobic metabolism ........................................................................................................... 39 2.4.4. Net N2 fluxes and rates of denitrification and nitrate ammonification ...................................................... 41 2.4.5. Nutrient fluxes at the sediment-water interface ......................................................................................... 42 2.4.6. Multivariate analysis ................................................................................................................................. 46

2.5. DISCUSSION ....................................................................................................................................................... 46 2.5.1. Sedimentary features ................................................................................................................................. 46

2.5.1.1. Carbon isotopic signature along the transect as tracer of anthropic impact .............................................................. 46 2.5.1.2. The isotopic signature of N in sediments suggests multiple co-occurring processes ............................................... 49 2.5.1.3. The stoichiometry of organic C and N ........................................................................................................................ 50

2.5.2. Metabolic rates .......................................................................................................................................... 51 2.5.2.1. Oxygen and nutrient fluxes .......................................................................................................................................... 51 2.5.2.2. Microbial N transformations ........................................................................................................................................ 53

2.6. CONCLUSIONS ................................................................................................................................................... 55 2.7. APPENDICES CHAPTER 2 .................................................................................................................................... 57

2.7.1. Materials and methods .............................................................................................................................. 57 2.7.2. Results ....................................................................................................................................................... 59

X

2.8. REFERENCES ...................................................................................................................................................... 63

CHAPTER 3 ........................................................................................... 73

BIOGEOCHEMICAL MODELING OF A TROPICAL COASTAL AREA UNDERGOING SEASONAL UPWELLING AND IMPACTED BY UNTREATED SUBMARINE OUTFALL ............ 73

3.1. ABSTRACT ......................................................................................................................................................... 73 3.2. INTRODUCTION .................................................................................................................................................. 74 3.3. MATERIALS AND METHODS ................................................................................................................................ 77

3.3.1. Study area .................................................................................................................................................. 77 3.3.2. Model description ...................................................................................................................................... 79 3.3.3. Water quality field monitoring for SMCA ................................................................................................. 80 3.3.4. The AEM3D Hydrodynamic model set up for SMCA. ............................................................................... 81 3.3.5. AEM3D ecological model set up for SMCA. ............................................................................................. 83 3.3.6. Calibration and Validation of AEM3D model for SMCA .......................................................................... 85 3.3.7. Statistical treatment and assessment of AEM3D model performance ....................................................... 87 3.3.8. Analyzed Scenarios using the Model AEM3D ........................................................................................... 87

3.4. RESULTS ............................................................................................................................................................ 88 3.4.1. Water quality monitoring for SMCA ......................................................................................................... 88 3.4.2. Calibration and validation of AEM3D model for SMCA .......................................................................... 88 3.4.3. Wind field and hydrodynamics of SMCA during upwelling and non-upwelling seasons .......................... 98 3.4.4. Nutrients and phytoplankton dynamics at different loads of the outfall along NUPW and UPW seasons.

........................................................................................................................................................................... 102 3.5. DISCUSSION ..................................................................................................................................................... 105

3.5.1. Circulation patterns during non-upwelling and upwelling seasons in the SMCA ................................... 106 3.5.2. Effect of increased nutrient load of the outfall on phytoplankton and nutrients dynamics during NUPW

and UPW seasons in SMCA .............................................................................................................................. 107 3.5.3. Implications for Benthic – pelagic coupling of SMAC ............................................................................ 110

3.6. CONCLUSIONS ................................................................................................................................................. 112 3.7. APPENDICES CHAPTER 3 .................................................................................................................................. 113

3.7.1. Materials and methods ............................................................................................................................ 113 3.7.2. Results ..................................................................................................................................................... 114

3.8. REFERENCES .................................................................................................................................................... 124

CHAPTER 4 ......................................................................................... 134

CONCLUSIONS AND RECOMMENDATIONS ............................ 134

4.1. RESEARCH SUMMARY ...................................................................................................................................... 134 4.2. RECOMMENDATIONS FOR FUTURE WORK ......................................................................................................... 136

5. FINAL APPENDICES .................................................................... 137

5.1. INCUBATION AND STIRRING SYSTEM FOR LARGE SEDIMENTS CORES ................................................................ 137 5.2. COMPLEMENTARY WORK ................................................................................................................................. 138

XI

List of figures

Fig. 2.1 Location of the Santa Marta sewage outfall (SMSO) and of the sediment sampling

stations in Taganga Bay (a) and of the study area in Caribbean Colombia (b). Sampling

stations coordinates are reported in decimal degrees. ................................................... 31

Fig. 2.2 Signatures of δ13C (V-PDB) and δ15N(AIR) in surface sediments (0-3 cm) of 4

stations (S1 to S4) located along increasing distances from a sewage outfall. Station S1 is

the closest to the outfall. Measurements were performed in November 2017 and in

January – February 2018, during non upwelling (NUPW) and upwelling season (UPW),

respectively. .................................................................................................................. 40

Fig. 2.3 Relative contribution of terrestrial versus marine organic carbon in surface

sediments (0-3 cm depth) calculated for 4 stations along increasing distance from a

sewage outfall in Taganga Bay – Colombian Caribbean. Measurements were carried out

during the non upwelling (NUPW) and upwelling (UPW) season. Averages ± standard

deviations (n=4) are reported. ....................................................................................... 41

Fig. 2.4 Dark oxygen fluxes measured at the 4 stations during the non upwelling (NUPW) and

upwelling (UPW) season. Averages (n=5) and standard deviations are reported. ......... 43

Fig. 2.5 Dark net N2 fluxes measured as N2:Ar ratios at the 4 stations during the non

upwelling (NUPW) and upwelling (UPW) season. Averages (n=5) and standard

deviations are reported. ................................................................................................ 44

Fig. 2.6 Rates of denitrification of water column nitrate (Dw) and of nitrate produced via

nitrification (Dn) measured in the dark at the 4 stations during the non upwelling

(NUPW) and upwelling (UPW) season. Averages (n=12) and standard deviations are

reported. ....................................................................................................................... 44

XII

Fig. 2.7 Dark rates of nitrate ammonification (DNRA) measured at the 4 stations during the

non upwelling (NUPW) and upwelling (UPW) season. Averages (n=4) and standard

deviations are reported. ................................................................................................ 47

Fig. 2.8 Dark fluxes of inorganic nutrient measured at the 4 stations during the non upwelling

(NUPW) and upwelling (UPW) season. Averages (n=5) and standard deviations are

reported. ....................................................................................................................... 49

Fig. 2.9 RDA plot showing the angles vectors representing sedimentary properties and those

nutrient fluxes reflect their (linear) correlation (scaling 1). .......................................... 52

Fig. 2S.1 The images show the overnight preincubation of intact sediment cores collected

from S1, submersed and with the stirring system on (a) and how surface sediments from

S1 (b) and S4 (c) looked. In sediments from S1 a well-developed mat of Beggiatoa is

evident. ......................................................................................................................... 57

Fig. 2S.2 RDA plot showing the angles between TOC, δ13C and δ15N that reflect linear

correlation (scaling 2). .................................................................................................. 60

Fig. 3.1 Domain location and sampling stations. Panel (a): The star and the dots indicate the

location of the SMSO and of the water column sampling stations in SMCA. HYM1 and

HYM2 are the Hybrid Coordinate Ocean Model (HYCOM) stations used to set

temperature and salinity boundary conditions for the AEM3D model. Panel (b): The

study area in the Colombian Caribbean. Panel (c):Transect of 1800m from the SMSO

both towards Taganga Bay and Santa Marta Bay in which it was analyzed the increase

of wastewater flow-rate with the model. ....................................................................... 79

Fig. 3.2 Water quality monitoring field data for SMCA – Colombian Caribbean at 10 stations

during NUPW (26th November – 2017) and UPW (21th August 2017, 26th January 2018

XIII

and 4th February) seasons. Average ± standard deviations are reported. Panels on the

left report pooled data from all stations and depths in the different seasons while panels

on the right report data relative to the different stations from all seasons and depths. 89

Fig. 3.3 Comparison between simulated and measured temperature profiles during the

calibration (November 2017) and validation (January and February 2018) periods. Field

data correspond to CTD measurements during: a) 26th November 2017, within the

NUPW season; b) 26thJanuary 2018, within the UPW season; c) 4th February 2018 within

the UPW season. AEM3D model simulation results include the range of all the simulated

profiles generated from the day before to the measurement day. CTD data were not

measured at P2 and P4 in January. .............................................................................. 91

Fig. 3.4 Sea surface temperature (SST) satellite images from MODIS during NUPW and

UPW seasons for (a) 26th November 2017 (NUPW), (c) 14th December 2017 (UPW) and

(e) 15th January 2018 (UPW) and modelled SST at noon for SMCA on (b) 26th November

2017 (NUPW), (d) 14th December 2017 (UPW) and (f) 15th January 2018 (UPW). NMAE

between simulated results and satellite images is also reported. ................................... 92

Fig. 3.5 Temperature – Salinity plots of stations located in Taganga (P2 and P4) and Santa

Marta (P7) bays in SMCA during 26th November 2017 (NUPW) and 4th February 2018

(UPW) for (a) sampling data and (b) AEM3D model results. ....................................... 93

Fig. 3.6 Modelled temperature vertical profiles at different times in SMAC: (a) Calibration

period (November 2017 - NUPW) and (b) Validation period (December 2017 – February

2018 - UPW) at station P4; (c) Calibration period (November 2017 –NUPW) and (d)

Validation period (December 2017 –February 2018 - UPW) at station P9. ................... 93

Fig. 3.7 Chl-a satellite images from Sentinel-3A during NUPW and UPW seasons for: (a)

November 26th 2017 (NUPW), (c) January 3rd 2018 (UPW) and (e) January 30th 2018

(UPW) and modelled Chl-a at noon for SMCA (b) November 26th 2017, (d) January 3rd

XIV

2018 and (f) January 30th 2018. NMAE between simulated results and satellite images

are reported. ................................................................................................................. 97

Fig. 3.8 Wind field measurements at Taganga Bay for the simulation period (November 2017

– February 2018): (a) wind speed and (b) wind direction (measured clockwise from the

north). ......................................................................................................................... 100

Fig. 3.9 Average residence time (“water age”) series: (a) November 2017 - NUPW and (b)

December 2017 – February 2018 - UPW season. Average surface flow velocity series (Vx

and Vy): (c) NUPW and (d) UPW season. Flow velocity series (Vx and Vy) at a depth of

40 m: (e) NUPW and (f) UPW season. Series of average surface temperature of at 40 m

(g) NUPW and (h) UPW season near to SMSO during the simulation. ....................... 101

Fig. 3.10 Surface-averaged nutrient concentrations in NUPW and UPW season close to

SMSO during the simulation: (a) NH4+ (NUPW) and (b) NH4

+ (UPW), (c) NOx (NUPW)

and (d) NOx (UPW), (e) PO43-(NUPW) and (f) PO4

3-(UPW), (g) TP (NUPW) and (h) TP

(UPW), (i) DO (NUPW) and (j) DO (NUPW), (k) Chl-a (NUPW) and (l) Chl-a (UPW).

.................................................................................................................................... 104

Fig. 3S. 1. Comparison of model simulation profiles (grey color) against field monitoring

data (red circles) for NH4+, NOx, PO4

3-, TP, DO and Chl-a during the calibration period

in November 2017 (NUPW) at 6 stations located in Taganga (P2, P3, P4 and P5), Santa,

Santa Marta (P7) and Gaira (P9) bays. Field data correspond to measurements carried

out on Novemver 26th 2017. ................................................................................................ 115

Fig. 3S. 2. Comparison of model simulation profiles (grey color) against field monitoring data

(red circles) for NH4+, NOx, PO4

3-, TP, DO and Chl-a during the validation period in

December 2017 – February 2018 (UPW) at stations located in Taganga (P2, P3, P4 and

P5), Santa Marta (P7) and Gaira (P9) bays. Field data correspond to measurements

XV

carried out on January 26th 2018. Field data were not measured at P2 and P4 in January.

NH4+ field data were below the detection limit at all stations. ........................................ 116



3-, TP, DO and Chl-a during the validation period in

December 2017 – February 2018 (UPW) at stations located in Taganga (P2, P3, P4 and

P5), Santa Marta (P7) and Gaira (P9) bays. Field data correspond to measurements

carried out on February 4th 2018. ...................................................................................... 117

Fig. 3S. 4. Simulated-surface residence time and flow velocity of the SMSO plume in

November 2017 - NUPW and December – February 2018 – UPW seasons in SMCA: (a)

southwesterly (SW) winds gust, (b) low wind speed in the NUPW, (c) changes in wind

direction from WNW – NNE in afternoon in the NUPW, (d) NNE winds but slightly in

the north direction, (e) NNE winds, (f) changes in the direction from NNE – WS in the

morning in the UPW, (g) low intensity southwest winds in the UPW season, (h) changes

in the wind directions E-SW in the morning, (i) and (k) NNE winds in the UPW, (j)

decrease the intensity of NNE in the mornings and (k) WS winds in the UPW season.

............................................................................................................................................... 118

Fig. 3S. 5. Average residence time (“water age”) series (a) November 2017 - NUPW season

and (b) December 2017 – February 2018 – UPW season. Average surface flow velocity

series (Vx and Vy) in (c) NUPW and (d) UPW season. Flow velocity (Vx and Vy) at a

depth of 40 m in (e) NUPW and (f) UPW season. Series of water mean flow – rate of 2.5

m3·s-1 during the simulation. .............................................................................................. 119

Fig. 3S. 6. Modelled PAR extinction profiles at different times in SMAC: (a) Calibration

period (November 2017 - NUPW) and (b) Validation period (December 2017 – February

2018 - UPW). ........................................................................................................................ 120

Fig. 3S. 7. TOC quantification along a 1800m transect from the SMSO towards Taganga and

Santa Marta bays for flow – rates of 1.0 m3·s-1, 2.5 m3·s-1 and without outfall. Results

XVI

based on modelled average surface concentrations in: (a) NUPW and (b) UPW season.

............................................................................................................................................... 121

Fig. 3S. 8. Chl-a concentration quantification through a 1800m transect from the SMSO

towards Taganga and Santa Marta bays for flow – rates of 1.0 m3·s-1and without outfall.

Results based on modelled average surface concentrations of Chl-a between February

14th -15th 2018 UPW season. ............................................................................................... 122

Fig. 3S. 9. Chl-a satellite images from Sentinel-3A during UPW season for (a) February 14th

2018 (c) February 15th 2018 and modelled Chl-a at noon for SMCA (b) February 14th

2018, (d) February 15th 2018. NMAE between simulated results and satellite images are

reported. ............................................................................................................................... 123

XVII

List of tables

Table 2.1 Sedimentary features measured in the upper 3 cm layer of stations S1 to S4.

Sampling stations are located along a gradient of impact, with S1 closest to the sewage

outfall. Sediments were collected in November 2017, NUPW and in January-February

2018, UPW in Taganga Bay – Colombian Caribbean. ...................................................... 37

Table 2.2 Results of two-way ANOVA to test the effects of the factors sampling season and stations on sedimentary features of Taganga Bay. ............................................................ 38

Table 2.3 Results of the two-way ANOVA on the effects of the factors sampling season, sampling stations and of their interaction on benthic fluxes, denitrification and nitrate ammonification rates. See the text for more details. .......................................................... 45

Table 2.4 Results of two-way PERMANOVA to test the effects of the factors sampling season and stations on nutrient fluxes, Dn, Dw and DNRA. ............................................. 48

Table 2.5 Fluxes of O2 and nutrients and rates of denitrification, N2 Fixation and DNRA from different estuarine zones (µmol·m-2·h-1). ................................................................... 56

Table 2S.1 Summary of sedimentary features analysis and of process rate measurements

carried out at each station. ................................................................................................... 58

Table 2S.2 Pearson correlation coefficients (r) between sedimentary features and N-related

processes ................................................................................................................................. 59

Table 2S.3 Sedimentary properties significantly correlated with the axes in RDA analysis 61

Table 2S.4 Significance test of the relationship between the sedimentary properties and

nutrient fluxes on RDA1 and RDA2. ................................................................................... 61

Table 2S.5 Pairwise test of interactions in two ways PERMANOVA on nutrient fluxes, Dn,

Dw and DNRA ....................................................................................................................... 62

XVIII

Table 3.1 Results of Kruskal-Wallis rank sum test to analyze the effect of seasons, stations

and depths on nutrients, Chl-a and DO in the water column of SMCA. ......................... 90

Table 3.2 Parameter description and values used for the simulations of AEM3D ecological

model for SMCA. .................................................................................................................. 95

Table 3S.1 Wastewater effluent quality of SMSO for simulated scenarios in the AEM3D

model .................................................................................................................................... 113

Table 3S.2 Statistical comparison between model simulations and field data of surface (0

m), 20m and 40 m deep in waters of Santa Marta Coastal Area (SMCA) for several

variables of AEM3D model. ............................................................................................... 114

22

Chapter 1.

General Introduction

1.1. Motivation

The potential impact of land-use alterations, nutrient and organic enrichment on the

biogeochemistry of tropical estuaries and coastal areas is not well understood (Downing et al.,

1999; Smith et al., 2012; Boynton et al., 2018). Tropical estuaries are among the most

biogeochemical active regions in the biosphere (e.g. corals, fishing, etc.). These systems are

subjected to complex processes due to factors such as highly variable hydrodynamics, large organic

matter inputs of land sources (e.g. wastewater outfall, rivers, harbors, etc.), as well as alterations

in physico-chemical composition in the water masses by seasonal upwelling (Corredor et al., 1999;

Capone & Hutchins, 2013; Cloern et al., 2014) . A higher natural and anthropogenic nutrient loads

increase primary production (either pelagic or benthic) resulting in dissolved oxygen depletion

(hypoxia and anoxia), water quality decline, changes in ecological structure including loss of

biodiversity, increments in algal blooms and changes in sediment biogeochemistry (Diaz, 2001;

Grall & Chauvaud, 2002; Smith et al., 2012; Davidson et al., 2014). These alterations will probably

have a greater impact on tropical systems than those already observed in temperate areas, probably

due to the much higher water temperatures heterotrophic activity, and to the lower oxygen

solubility and availability (Downing et al., 1999). In this thesis, benthic metabolism measurements

and the implementation of a 3D-coupled hydrodynamic – ecological model was used to analyze

the effects of the wastewater effluent discharge along with a seasonal upwelling the physical

variables, the nutrient dynamics and phytoplankton in the benthic and pelagic compartment of a

tropical estuary located in the Santa Marta Coastal Area (SMCA) in the Colombian Caribbean. The

understanding of these processes is essential for developing effective management strategies in

tropical coastal areas.

Chapter 1. Introduction

23

1.2. Objectives

1.2.1 General objective

To analyze how the biogeochemical conditions (nitrogen and phosphorus) in the water column and

in the sediment - water interface near the Santa Marta sewage outfall (SMSO) are modified by

changes in the local water circulation generated by seasonal climatic patterns and currents and

future variations of the wastewater effluent discharge.

1.2.2. Specific objectives

• To evaluate the nutrient fluxes at the sediment-water interface and their effect on the water

column in the area near the Santa Marta sewage outfall (SMSO) by sediment cores

incubations in the laboratory.

• To evaluate the effect of local water circulation patterns generated by seasonal climatic

patterns on the biogeochemical conditions in the area near the Santa Marta sewage outfall

(SMSO) by using a coupled hydrodynamic-ecological model (AEM3D).

• To determine, from the results of the AEM3D model and the sediment cores incubations,

the possible implications of climatic variability and future increases in the wastewater

effluent discharge on the oxygen regime and the biogeochemical cycles in the Santa Marta

Bay.

1.3. Thesis overview

This thesis is based on two main research chapters (Chapters 2-3) which have been prepared for,

or published in, peer-reviewed scientific journals. The Chapter 1 presents a brief overview of the

study and its goals and Chapter 4 contains a brief summary of the main conclusions derived from

Chapters 2-3.


24

In the chapter 2, sediment properties (organic matter quantity, C, N and P pools and δ13C and δ15N)

and benthic metabolism (aerobic respiration, denitrification, nitrate ammonification and nutrient

recycling) were analyzed at increasing distances from the marine outfall and during the non

upwelling (NUPW) and upwelling (UPW) seasons in Santa Marta Coastal Area (SMAC). Carbon

and N stable isotopes were used to determine the organic source in sediments (marine and

terrestrial) and the extent of the organic pollution. Samplings were collected in November 2017

(NUPW) and in January-February 2018 (UPW) at 4 stations (S1-S4) located in the proximity and

at 100, 750 and 1800 m far from the outfall. From each site, plexiglass liners were collected at

20<x<30 m depth for sediment characterization (C, N and P content and δ13C and δ15N, pore water

Eh and S2-) and metabolic measurements (O2, N2, CH4, Fe2+, Mn2+, NH4+, NO2

-, NO3-, PO4

3- and

SiO2). Fluxes were measured in the laboratory via dark incubations; sequentially to fluxes

denitrification and dissimilative nitrate reduction to ammonium were measured via the r-IPT.

Chapter 3 presents the methodology followed to calibrate and validate the mathematical model.

Water quality samples were collected during four field campaigns in August 2017 (minor

upwelling) and November 2017 (major non upwelling) and January – February 2018 (major

upwelling). This dataset and satellite images of temperature and chlorophyll-a for Santa Marta

Coastal Area (SMCA) were used to set the open boundary conditions for nutrients, as well as, to

calibrate and validate of model. In addition, benthic measurements derived from experiments

within the simulation period provided critical data to refine the calibration parameters. The 3D-

coupled hydrodynamic-ecological model (AEM3D) based on the ELCOM – CAEDYM model was

used to predict circulation patterns, residence time, nutrient and chlorophyll-a dynamics under

different organic loads of the Santa Marta sewage outfall (SMSO), including seasonal non-

upwelling (NUPW) and upwelling (UPW), allowing the understanding of their implications on the

benthic-pelagic coupling.


25

1.4. References

Boynton, W. R., Ceballos, M. A. C., Bailey, E. M., Hodgkins, C. L. S., Humphrey, J. L., & Testa,

J. M. (2018). Oxygen and Nutrient Exchanges at the Sediment-Water Interface: a Global

Synthesis and Critique of Estuarine and Coastal Data. In Estuaries and Coasts (Vol. 41, Issue

2). Estuaries and Coasts. https://doi.org/10.1007/s12237-017-0275-5

Capone, D. G., & Hutchins, D. A. (2013). Microbial biogeochemistry of coastal upwelling regimes

in a changing ocean. Nature Geoscience, 6(9), 711–717. https://doi.org/10.1038/ngeo1916

Cloern, J. E., Foster, S. Q., & Kleckner, A. E. (2014). Phytoplankton primary production in the

world’s estuarine-coastal ecosystems. Biogeosciences, 11(9), 2477–2501.

https://doi.org/10.5194/bg-11-2477-2014

Corredor, J. E., Howarth, R. W., Twilley, R. R., & Morell, J. M. (1999). Nitrogen cycling and

anthropogenic impact in the tropical interamerican seas. Biogeochemistry, 46(1–3), 163–178.

https://doi.org/10.1007/BF01007578

Davidson, K., Gowen, R. J., Harrison, P. J., Fleming, L. E., Hoagland, P., & Moschonas, G. (2014).

Anthropogenic nutrients and harmful algae in coastal waters. Journal of Environmental

Management, 146, 206–216. https://doi.org/10.1016/j.jenvman.2014.07.002

Diaz, R. J. (2001). Overview of hypoxia around the world. Journal of Environmental Quality,

30(2), 275–281. https://doi.org/10.2134/jeq2001.302275x

Downing, J. A., Mcclain, M., Twilley, R., Melack, J. M., Elser, J., Rabalais, N. N., Lewis, W. M.,

Turner, R. E., Corredor, J., Soto, D., Kopaska, J. A., & Howarth, R. W. (1999). The Impact

of Accelerating Land-Use Change on the N-Cycle of Tropical Aquatic Ecosystems : Current

Conditions and Projected Changes Source : Biogeochemistry , Vol . 46 , No . 1 / 3 , New

Perspectives on Nitrogen Recycling in the Temperate and Tropical Ame. Biogeochemistry,

46, 109–148. https://doi.org/10.1002/mds.25809

Grall, J., & Chauvaud, L. (2002). Marine eutrophication and benthos: The need for new approaches

and concepts. Global Change Biology, 8(9), 813–830. https://doi.org/10.1046/j.1365-

2486.2002.00519.x

Smith, J., Burford, M. A., Revill, A. T., Haese, R. R., & Fortune, J. (2012). Effect of nutrient

loading on biogeochemical processes in tropical tidal creeks. Biogeochemistry, 108(1–3),

359–380. https://doi.org/10.1007/s10533-011-9605-z

26

Chapter 2.

Spatial and seasonal variability of sedimentary features and

nitrogen benthic metabolism in a tropical coastal area

(Taganga Bay, Colombia Caribbean) impacted by a sewage

outfall

2.1. Abstract

The effects of anthropogenic pressures in coastal areas are extensively studied in temperate but not

in tropical zones, where their impact might be amplified by high water temperatures and upwelling

phenomena. Sedimentary features and benthic metabolism were studied during the non upwelling

(NUPW) and upwelling (UPW) seasons in Taganga Bay (Colombia). The bay is impacted by a

submarine outfall of virtually untreated, organic and nutrient-rich wastewater. Samplings were

performed in November 2017 (NUPW) and in January-February 2018 (UPW) at 4 stations located

in the proximity and 100, 750 and 1800 m far from the outfall, respectively, at depths between 22

and 28 m. Aerobic respiration, denitrification, dissimilative nitrate reduction to ammonium

(DNRA) and nutrient fluxes were measured.

The influence of the outfall was detectable 750 and 1800 m away from the point pollution source,

where δ13C data suggested that ∼ 40 and ~20 % of organic inputs were terrigenous, respectively.

In the proximity of the outfall benthic oxygen demand peaked and the presence of Beggiatoa mats

suggested reoxidation of sulphides, that were abundant in pore water. Under sulfidic conditions,

DNRA was the major driver of nitrate demand, whereas at stations far from the outfall,

denitrification dominated nitrate consumption. Organic matter and nitrate inputs to the bay during

the UPW season enhanced the effects of the outfall by increasing aerobic respiration and DNRA.

Higher N availability during the UPW season reversed fluxes of molecular nitrogen and turned the

sediments of 3 out of 4 sites from net sinks to net N2 sources. Results from this study suggest that

Chapter 2. Sedimentary features and benthic metabolism

27

the analysis of sediments allows tracing the impact of the outfall and that such impact is enhanced

during the upwelling season. In tropical areas, marine outfalls and upwelling may act in synergy

and contribute to ecosystem deterioration due to high temperatures, increase of microbial

respiration, sulphide toxicity and benthic biodiversity loss.

2.2. Introduction

Organic matter mineralization in coastal sediments provides an important ecosystem service such

as benthic-pelagic coupling; nutrient regeneration from sediments may support 15–32% of nitrogen

(N) and 17–100% of phosphorus (P) demand by phytoplankton (Aller 1988; Bonaglia et al. 2014;

Carstensen et al. 2014; Mermillod-Blondin et al. 2004; Boynton et al. 2018). Sediments may also

act as nutrient traps, however the supply of organic matter to the benthic compartment and the

availability of oxygen are considered primary factors in controlling the direction and magnitude of

solute fluxes (Banta et al. 1995; Hopkinson et al. 2001; Smith et al. 2012). Excess inputs of organic

matter to sediments may uncouple microbial respiration and oxygen availability leading to anoxia,

buildup of free sulphides (S2-+HS-+H2S), simplification of benthic macrofauna communities and

loss of biogeochemical services as denitrification or P-retention (Rosenberg and Loo 1988; Grall

and Chauvaud 2002; Howarth et al. 2011). Oxygen shortage produces profound changes in aquatic

communities, affecting energy transfer across trophic levels and biogeochemical cycles (Diaz

2001; Testa and Kemp, 2011). Under hypoxic/anoxic conditions, the accumulation of toxic,

reduced anaerobic metabolites within sediments and/or water depresses the activity of microbes,

plants and benthic fauna (Testa and Kemp 2011). In turn, this creates a positive feedback, as

bioturbation and bioirrigation by infauna has a key role in organic matter degradation and solute

transport (e.g. O2, NO3-) in sediments (Aller 1994; Bonaglia et al. 2014; Carstensen et al. 2014).

Strictly anoxic and chemically reduced sediments favor P mobility and its availability in water

column and reduce P burial via precipitation or co-precipitation with iron (Fe) and manganese (Mn)

oxides and hydroxides and organic matter (Testa and Kemp 2011; Zilius et al. 2014). Under

sulphidic conditions, the activity of denitrifers is inhibited, whereas that of nitrate ammonifiers is

stimulated, and the role of sediments switches form net N sink to net N source (Carstensen et al.

2014). Such switch is increasingly reported and has deep implications for benthic N cycling (An

and Gardner 2002; De Brabandere et al. 2015; Hall et al. 2017).


28

Superimposed to or interacting with the effects of organic matter enrichment are the seasonal

variations of physical variables in water column such as temperature, vertical stratification and

circulation patterns. These factors, that include upwelling phenomena, are important as they control

the vertical distribution and the supply of dissolved oxygen, nutrients and oligoelements in the

water mass (Testa and Kemp 2011). There is a general scarcity of studies targeting benthic

metabolism measurements and of the potential impact of land-use alteration, nutrient and organic

enrichment on sediments biogeochemistry in low latitude tropical coastal zones (Downing et al.

1999; Boynton et al. 2018). Downing et al. (1999) suggested that alterations in the N-cycle will

probably have a greater impact on tropical systems than those observed in temperate areas, due to

higher water temperatures and heterotrophic activity, and to lower oxygen solubility. Upwelling

phenomena represent another potentially important, understudied factor regulating these dynamics,

as they may alter locally the temperature of the water mass and its physico-chemical composition

(e.g. chlorophyll-a, suspended solids, oxygen, macro- and micronutrient content). Upwelling may

either smooth the effects of organic enrichment or enhance them, as it represents in many areas a

natural eutrophication phenomenon.

In this study, the combined effect of anthropogenic organic input associated to untreated sewage

outfall and of natural upwelling on benthic biogeochemistry was analyzed in Taganga Bay, a

tropical system located in Santa Marta Coastal Area (SMCA), Caribbean Colombia. There is a

strong interest in understanding the effects of multiple stressors, including wastewater discharge,

and seasonal upwelling on the functioning of SMCA, and in particular on the health of coral

meadows, as the area is considered a hotspot of biodiversity in the Colombian Caribbean

(Bayraktarov et al. 2013; Bayraktarov and Wild 2014; Bayraktarov et al. 2014). Despite the

importance of preserving ecosystem services in the area, no biogeochemical studies targeting the

impact of wastewater outfall, alone or in combination with the seasonal upwelling phenomena were

carried out to date. In particular, the sediment compartment was never studied in Taganga Bay,

despite its memory for past and ongoing pollution and its importance in regulating dissolved

oxygen and nutrient content in the water column.

Sediment properties (organic matter quantity, C, N and P pools and δ13C and δ15N) and benthic

metabolism (aerobic respiration, denitrification, nitrate ammonification and nutrient recycling)

were analyzed at increasing distances from a marine outfall and along the non upwelling and

upwelling seasons in SMCA. Carbon and N stable isotopes were used to determine the organic


29

source in sediments (marine and terrestrial) and the extent of the organic pollution. The study had

a threefold aim: to trace the spatial impact of the outfall on the sediments, to quantify its effects on

benthic respiration and nutrient recycling and to analyze whether such effects are enhanced by the

natural upwelling. I hypothesized a large stimulation of heterotrophic activity due to the organic

inputs from the outfall and a suppression of biogeochemical sediment services as N removal via

denitrification, sulphide precipitation and nutrient retention. I also hypothesized that organic inputs

from marine upwelling act in synergy with those from the outfall, enhancing the risk of sediment

anoxia and nutrient regeneration.

2.3. Materials and methods

2.3.1. Study area

Taganga Bay has an open conformation, subject to oceanic and continental influences, and is

impacted by multiple pressures including a submarine wastewater outfall, the harbor activity, the

Manzanares and Gaira river discharge and by seasonal surface runoff (Escobar 1988; Ramos-

Ortega et al. 2008; Mancera-Pineda et al. 2013)(Fig. 2.1). River discharge and surface runoff are

low or null in the dry seasons (December- April and July-August) and peak in the wet seasons

(May-June and September-November) (Arévalo-Martínez and Franco-Herrera 2008).

The Santa Marta sewage outfall (SMSO, Fig. 2.1) discharges on average 1 m3s-1 of untreated

wastewater derived from ~500,000 inhabitants (García 2013). The city nearly doubles its resident

population during high tourist season, peaking in January and resulting in increased wastewater

discharge (Díaz-Rocca and Causado-Rodríguez 2007). The outlet is located between Santa Marta

and Taganga bays (11.26 Lat and -74.22 Lon) (García 2013). The sewage outfall consists of 1m

diameter tubing and extends 428 m along the coastline (García et al. 2012). Sewage only receives

a preliminary treatment for large solids removal (Díaz-Rocca and Causado-Rodríguez 2007).

During 2006, the loadings of total N, P and total suspended solids (TSS) released to the bay from

the outfall were estimated in ~1,100 T N·yr-1, ~500 T P·yr-1 and ~12,300 T TSS·yr-1, respectively

(García 2013). The maximum extension of the area affected by discharge from SMSO was

estimated at 600 m from the outfall from water column fecal coliforms concentration, whereas

sediments were not analyzed (García 2013).


30

Taganga Bay is highly influenced by the Southern Caribbean upwelling system (Rueda-Roa &

Muller-Karger, 2013). It is subjected to strong seasonality caused by the Caribbean Low – level Jet

of Northeast (NE) Trade Winds and the oscillations of the Intertropical Convergence Zone

(Andrade and Barton 2005). Major (December- April) and minor (July-August) upwelling events

occur during the dry season while minor (May-June) and major (September-November) no-

upwelling periods occur during the rainy season (Fajardo 1979; Andrade and Barton 2005;

Arévalo-Martínez and Franco-Herrera 2008). Seasonal upwelling leads to changes in

physicochemical variables such as temperature decrease (from 30 °C to 21 °C), salinity increase

(from 33 to 38), oxygen subsaturation conditions (< 91% ) (Ramírez 1981; Salzwedel and Müller

1983; Bayraktarov et al. 2014), and increase of nitrate and chlorophyll-a concentration, turning the

system from oligotrophic to mesotrophic (Arévalo-Martínez and Franco-Herrera 2008; García-

Hoyos et al. 2010; Paramo et al. 2011).

Water and sediment sampling was carried out at four stations (S1 to S4, located in the proximity

and at 100, 750 and 1800 m from the outfall, respectively) (Fig. 2.1). Specifically, S1 was located

near the outfall diffusers whereas S4 was close to Punta Venado, in a sandy area with low coral

covering (Martínez and Acosta 2005; Vega-Sequeda et al. 2008). On every sampling occasion the

physico-chemical features of the water overlying sediments (temperature, salinity and dissolved

oxygen) were analyzed with a YSI 560 multiple probe.

2.3.2. Sediment cores collection

Sediments were sampled with minimum disturbance by scuba divers using transparent plexiglass

liners (internal diameter = 4 cm, length = 20 cm), in order to have approximately equal heights (8-

9 cm) of sediment and water column. Sampling was conducted during major non upwelling rainy

season –NUPW- (S1: 27 November, S2: 28 November, S3: 30 November and S4: 2 December

2017) and during major upwelling dry season –UPW- (S1: 26 January, S2: 28 January, S3: 31

January and S4: 4 February 2018). Twenty sediment cores were collected at each station. The cores

were intentionally scattered along a 22-28 m depth profile, due to the variable topography of the

bottom and in order to have overlapping depths among sites. Once collected, the cores were bottom

and top capped and brought to the boat where the top lid was removed and the cores were

submersed in a tank containing in-situ water cooled with ice packs (Fig. 2.S1, supplementary

material). Different tanks were used for the different stations. Bottom water (∼40 L per site) was


31

collected through four large PVC pipes equipped with valves at both ends. Pipes were submersed

and the valves were closed near the bottom by divers. Cores were stored vertically in the dark, and

immediately transferred to the laboratory where they were equipped with magnetic stirrers (more

details provided in next paragraphs). During the overnight preincubation the cores were maintained

submersed with the top open and the stirring on; the water in the tank was maintained at in situ

temperature and at 100% oxygen saturation.

Fig. 2.1 Location of the Santa Marta sewage outfall (SMSO) and of the sediment sampling stations

in Taganga Bay (a) and of the study area in Caribbean Colombia (b). Sampling stations coordinates

are reported in decimal degrees.

2.3.3. Net flux measurement at the sediment-water interface

The incubations were carried out at the chemistry laboratory of the Jorge Tadeo Lozano University,

campus Santa Marta, in a temperature-controlled room maintained at 29 and 23 °C during the

NUPW and UPW season, respectively. The day after sampling, dark incubations began by sealing

each core with gas-tight lids. During the preincubation and incubation periods the temperature of

the incubation tank varied by less than 0.3 °C. The water inside the cores was gently stirred by


32

Teflon- coated magnetic bars without suspending the sediment (Fig. 2.S1, supplementary material).

The magnetic bars were driven by an external magnet rotated by a motor at 40 rpm. Five cores per

station were used to measure the fluxes of dissolved gas (O2:Ar, N2:Ar and CH4), metals (Fe2+ and

Mn2+) and inorganic nutrients (NH4+, NO2

-, NO3-, PO4

3- and SiO2) (Table S1, supplementary

material) ). Solutes concentrations were measured at the beginning and at the end of the incubation,

assuming that their uptake or release rate during the incubation was linear. At time zero, four water

samples (ca. 60 ml) were collected from the incubation tank before cores were capped. Sediment

incubation time was maximum 6 h to keep oxygen within 20-30% of the initial value. Incubation

time was set with pilot incubations carried out before starting the experiments. At the end of the

incubation, lids were removed and water samples were collected from each core. Subsamples of

12 ml were transferred into exetainers, poisoned with 200 µl of ZnCl2 solution (7 M) and later

analyzed for O2:Ar and N2:Ar ratios and for CH4 concentrations with a Membrane Inlet Mass

Spectrometer (MIMS) equipped with a furnace with a copper reduction column (Bay instruments).

Subsamples of 15 ml were transferred to glass vials, acidified with 50 µl of concentrated HNO3

and analyzed for Fe2+ and Mn2+ with a Varian 240FS AA. Subsamples of 20 ml were filtered

(Whatman, GF/F), refrigerated and stored at -20 °C for later nutrient analysis. Nutrients were

measured with a continuous-flow analyzer (San++, Skalar) according to Grasshoff et al. (1983).

Detection limits were 0.1 µM for NH4+ and NO3

-, 0.05 µM for NO2- and PO4

3- and 0.3 µM for

SiO2. The Dissolved Inorganic Nitrogen (DIN) was calculated as DIN = NH4+ + NO2

- + NO3-.

Dissolved oxygen (DO) was also measured in the incubation tank and in the cores with a calibrated

oxygen sensor (Orion Star Water TM A326). Solute fluxes across the sediment - water interface

were calculated according to the equation (2.1) (Dalsgaard et al. 2000):

� = ��∙�∙� (2.1)

where F is flux of measured solutes (μmol m-2 h-1), C0 is concentration at time zero (μmol l-1), Cf

is concentration at the end of incubation (μmol l-1), V is the volume of water in the core (L), A is

area of sediment surface in the core (m2) and t is the incubation time (h). Fluxes directed from the

sediment to water column were considered as positive.


33

2.3.4. Denitrification and DNRA measurements

Sequential to net fluxes, the isotope pairing technique ITP (Nielsen, 1992) was used to measure the

dark rates of denitrification and nitrate ammonification (DNRA). The IPT allows partioning NO3-

source for denitrification in the contribution of nitrate diffusing to the anoxic sediment from the

water column (Dw) and denitrification of nitrate produced within the sediment due to nitrification

(Dn). Besides denitrification (and DNRA) the concentration series approach I adopted allows to

test for the occurrence of the anaerobic oxidation of NH4+ to N2 (ANAMMOX) (Dalsgaard et al.

2005; Trimmer et al. 2006; Burgin and Hamilton 2007; Koop-Jakobsen and Giblin 2009). Different

amounts of 15NO3- from a 30 mM Na15NO3 solution (>98% Sigma Aldrich) were added to the water

column of each of the 12 replicate cores to perform a concentration series experiment (Dalsgaard

et al. 2000). Three levels of 15NO3- concentration (20, 40 and 60 µM, each level with 4 replicates)

were created adding increasing amounts of the stock 15NO3- solution to the cores water phase (Table

S1, supplementary material). The incubation length was similar to that for fluxes, in order to keep

oxygen variations within 20-30% of the initial value. At the end of the incubation the sediment and

water were gently mixed. An aliquot of the slurry was transferred to a 12 ml Exetainer and poisoned

with 200 µl of ZnCl2 solution (7 M); 14N15N and 15N15N abundance in N2 were analyzed by MIMS.

As calculated 28N2 production was independent from the level of 15NO3- addition (no detectable

ANAMMOX), denitrification rates were calculated according to the equations and assumptions of

Nielsen (1992). Another aliquot of the slurry (20 ml) was transferred in 50 ml falcon tubes, treated

with KCl (1 M), shaken for 1 hour in the dark, centrifuged, filtered and frozen for later analysis of

the 15NH4+ fraction in the ammonium pool. 15NH4

+ was determined via chemical oxidation to 29N2

and 30N2 by alkaline hypobromite addition as described in Warembourg (1993). DNRA was then

calculated as described in Bonaglia et al. (2014). All cores were finally sieved (mesh size 500 µm)

to check for the occurrence of large macrofauna.

2.3.5. Sedimentary features

After incubations, the analysis of sedimentary features was conducted on subsamples from the

upper sediment layer (0-3 cm) in four cores per site (Table 2.S1, supplementary material). Such

sediment horizon was decided on the basis of the homogeneous color of sediment (Fig. 2.S1,

supplementary material) and in order to get information about the most reactive, upper layer. Each

sediment slice was homogenized and a subsample was taken using a cut-off 5 ml syringe. Water


34

content and porosity were determined after drying at 60 °C until constant weight. Then each sample

was homogenized and ground with a porcelain mortar and pestle prior analysis of total organic

carbon (TOC), total nitrogen (TN), isotopic measurements (δ13C and δ15N), total phosphorus (TP),

total inorganic phosphorus (TIP) and organic material content (OM). TOC, TN, δ13C and δ15N were

measured with a Thermo Finnigan EA/NA – 1100 CHN elemental analyzer coupled with an isotope

ratio mass spectrometer after removing carbonates with HCl (25% v/v) and subsequent rinsing with

deionized water to neutralize and dry (30 °C). The samples were run in duplicate and the analytical

precision was ±0.2% for TOC and ±0.01% for TN. TOC and TN percent measured on sediment

was corrected by the weight change due to the procedure (Tucker and Giblin 2010). For δ13C, the

reference was Vienna Pee Dee Belemnite (VPDB), and for δ15N, it was atmospheric nitrogen.

Isotopic ratios were expressed in the usual δ –notation (part per mill, ‰). Samples replicate

analyses were within ± 0.2% for δ13C and ± 0.3% for δ15N. Organic matter content (%) was

measured by loss on ignition (LOI) of 0.2 g sediment at 350°C in a muffle furnace. TP was

extracted from ashes with 37% HCl whereas TIP was extracted after treating 0.2 g of dry sediment

with 1 M HCl. Both TP and TIP were determined spectrophotometrically (Aspila et al. 1976). Total

Organic phosphorus (TOP) was estimated as difference between TP and TIP. Redox Potential (Eh)

was measured with a Thermo Fisher ORP sensor Orion Star TM A326, inserted within the gently

homogenized sediment slices. The sensor response was checked against a standard redox solution

(200 mV). As the sediment slice was exposed to air the measured Eh value can be slightly

overestimated. The porewater from the same slice was extracted from the sediment via

centrifugation (4000 rpm, 5 minutes) in glass vials filled with N2 and analyzed for sulphide

concentration by spectrophotometry (Cline 1969).

2.3.6. Contribution of terrestrial organic matter to Taganga Bay sediments

The contribution of terrestrial and marine organic matter to sediment organic carbon was estimated

using a two-end-member mixing model based on the equation derived by Calder and Parker (1968)

taken from Schlünz et al. (1999). This approach has been widely applied in organic matter pollution

studies of marine sediments (Li et al. 2016; Zhou et al. 2006). The terrestrial organic carbon

contribution was calculated using the equation (2.2):

Terrestial TOC %�=�δ13Cmarine-δ13COrg�

�δ13Cmarine-δ13Cterrestrial� x 100% (2.2)


35

The contribution of marine organic matter to the TOC was then estimated from equation (2.3):

Marine TOC %�=100-Terrestial OC(%) 2.3�

where δ13Cmarine is the δ13C of marine end member, δ13Cterrestrial is the δ13C of terrestrial end member

and δ13Corg is the measured value in sediment samples. A δ13Cterrestrial value of -27.0 ‰ was used as

end member based on the low δ13C values of sediment samples in the study area (Li et al. 2016).

For the marine end member (δ13Cmarine) a value of -20.3 ‰ was used as this is a typical carbon

isotopic composition for diatoms in temperate marine ecosystems (Gearing et al. 1984).

Furthermore, Gearing et al. (1984) indicated that planktonic isotope ratios varied little with water

temperature (0 to 20 °C). A prevalence of diatoms and dinoflagellates has been found in the Santa

Marta and Taganga bays (Ramírez-Barón et al. 2010; Garcés-Ordóñez et al. 2016).

2.3.7. Statistical Analysis

A two-way analysis of variance (2-way ANOVA) was carried to test the significance of the factors

season (NUPW versus UPW period), sampling site and of their interactive effects on sedimentary

properties, stable carbon and nitrogen isotopes (δ13C and δ15N), sedimentary fluxes and microbial

N transformations. Rcmdr package of the R- Project for Statistical Computing (R version 3.5.1)

was used to fit different models to the data (Fox, 2005). Kolmogorov-Smirnov’s test and Levene’s

tests were employed to assess the data normality (p>0.05) and variance homogeneity (p>0.05),

respectively. In the case of heteroscedasticity, data were 1/x, √x, ln(x), log(x2) or log(x)

transformed. A goodness of fit was obtained after transformations. For significant factors, post hoc

pairwise comparisons were performed using the Tukey HSD test. Significance was accepted for p

<0.05.

Correlations between sediment properties and nutrient fluxes were analyzed using simple and

multiple lineal regression models, Pearson – test analysis (significance level p<0.0001) and

redundancy analysis (RDA), respectively. RDA was performed in the R environment using the

vegan package. Before RDA analysis, a data exploratory multivariate analysis through Principal

Component Analysis (PCA) (results not shown) was made on sedimentary properties (explanatory

variables) and nutrient fluxes to detect and remove redundant variables lead to multicollinearity

and affect data interpretation. Then, in the RDA analysis, one replicate of nutrient fluxes was


36

randomly deleted to have a balanced arrangement along with sedimentary properties (4 replicates).

This process was repeated three times to verify the validity of RDA results. The significance of

sedimentary properties was tested against 999 permutation of ANOVA function included in vegan

package (p<0.05). RDA results were interpreted based on the scaling used. The angles vectors

representing sedimentary properties and nutrient fluxes reflect their linear correlation (scaling 1).

Angles between all vectors reflect linear correlation (scaling 2) (Legendre and Legendre 1998).

Two ways PERMANOVA was applied to analyze the multivariate variance to the raw variables of

nutrient fluxes, denitrification and DNRA. The adonis function within vegan package was used to

perform PERMANOVA test against 999 permutation. Post hoc pairwise comparisons were made

using the function pairwise adonis in R. Significance was accepted for p <0.05.

2.4. Results

2.4.1. Water and Sedimentary features

The upwelling altered the bottom water physico-chemical conditions in Taganga Bay. It resulted

in a nearly 7 °C decrease of average water temperatures, from 29.16±0.25 to 22.58 ± 0.21 °C, in a

decrease of DO concentration and saturation rates, from 235 ± 10 µM (123% saturation) to 208 ±

15 µM (106% saturation) and in an increase of salinity, from 35.98 ± 0.19 to 37.05 ± 0.02.

Sediments from the four stations were sandy, with comparable density, porosity and water content

but with marked differences in terms of OM, C, N and P content, and redox status (Table 2.1).

There was a clear gradient of impact from the sediments closest to the outfall (S1), that were more

organic, sulphidic and covered by Beggiatoa mats, to the outer station (S4) that had less organic

loaded and more oxidized sediments. The two-way ANOVA adopted for the analysis of sediment

features showed a good fit to experimental data (R2 > 75%, Table 2.2). The season, station and

interaction (season x station) terms were statistically significant for all parameters except for TOP.

The effect of station was higher for TOC and TN than that of season while both effects were strong

for OM, TP and TIP (Table 2.2). Sediment C, N, P, and OM content were in general highly

correlated across the stations (Table 2.S2, supplementary material).


37

Table 2.1 Sedimentary features measured in the upper 3 cm layer of stations S1 to S4. Sampling stations are located along a gradient of

impact, with S1 closest to the sewage outfall. Sediments were collected in November 2017, NUPW and in January-February 2018, UPW

in Taganga Bay – Colombian Caribbean.

November 2017, NUPW January-February 2018, UPW

Station S1 S2 S3 S4 S1 S2 S3 S4

Density (g cm-3) 1.73 ± 0.04 1.80 ± 0.11 1.89 ± 0.04 1.89 ± 0.08 1.68 ± 0.10 1.71 ± 0.05 1.96 ± 0.02 1.82 ± 0.06

Porosity 0.53 ± 0.02 0.47 ± 0.05 0.46 ± 0.01 0.43 ± 0.01 0.54 ± 0.03 0.44 ± 0.02 0.35 ± 0.03 0.35 ± 0.01

Water (%) 30.66 ± 1.92 26.10 ± 4.55 24.26 ± 0.43 22.98 ± 1.05 31.97 ± 2.00 25.63 ± 1.56 18.03 ± 1.52 19.45 ± 0.59

OM (%) 3.78 ± 0,99 2.59 ± 0,47 2.70 ± 0,35 2.27 ± 0,18 6.68 ± 0.64 3.32 ± 0.47 2.23 ± 0.13 2.11 ± 0.38

TOC (%) 0.34 ± 0.15 0.36 ± 0.22 0.18 ± 0.03 0.20 ± 0.02 1.32 ± 0.35 0.55 ± 0.25 0.16 ± 0.07 0.14 ± 0.01

δ13C(V-PDB) (‰) -24.69± 0.28 -24.50±0.75 -23.62± 0.24 -21.77± 0.42 -26.12 ± 0.56 -25.53± 0.20 -23.07± 0.54 -21.36 ± 0.23

TN (%) 0.044 ± 0.009 0.022 ± 0.004 0.026 ± 0.003 0.012 ± 0.001 0.134 ± 0.012 0.038 ± 0.002 0.029 ± 0.010 0.012 ± 0.001

δ15N(AIR) (‰) 3.26 ± 0.39 3.35 ± 0.40 6.23 ± 0.37 4.45 ± 0,88 4.92±0.98 13.40±2.75 7.77±3.40 4.00±2.09

TP (%)) 0.063 ± 0.016 0.051 ±0.008 0.058 ± 0.004 0.039 ± 0.005 0.123 ± 0.017 0.074 ± 0.008 0.054 ± 0.006 0.044 ± 0.002

TIP (%) 0.051 ± 0.007 0.049 ± 0.014 0.051 ± 0.004 0.035 ± 0.005 0.105± 0.014 0.063 ± 0.013 0.043 ± 0.008 0.033 ± 0.011

TOP (%) 0.019 ± 0.009 0.008 ± 0.009 0.007 ± 0.004 0.004 ± 0.001 0.029 ± 0.014 0.010 ± 0.008 0.012 ± 0.004 0.017 ± 0.008

TOC/TN (molar) 9 ± 3 13 ± 2 8 ± 2 20 ± 3 12 ± 4 14 ± 7 7 ±4 13 ± 1

Eh (mV) - 432 ± 21 -320 ± 63 -101 ± 48 49 ± 102 NM NM NM NM

H2S+HS-+S2- (µM) 870 ± 130 170 ± 80 0 0 NM NM NM NM


38

Table 2.2 Results of two-way ANOVA to test the effects of the factors sampling season and

stations on sedimentary features of Taganga Bay.

Dependent variable

Independent variable

F P Ƞ2 Adjusted R2

Tukey HSD test

OM Season 60.66 <0.0001 0.48 0.98 Station 30.06 <0.0001 0.35 S1-S2, S1-S3, S1-S4, S2-

S4 Interaction 6.33 0.002 0.07

TOC Season 10.25 0.0006 0.16 0.93

Station 22.40 <0.0001 0.52 S1-S3, S1-S4, S2-S3, S2-S4

Interaction 5.70 <0.05 0.13

TN Season 35.62 <0.0001 0.12 0.98 NUPW-UPW Station 165.81 <0.0001 0.82 S1-S2, S1-S3, S1-S4, S2-

S4, S3-S4 Interaction 4.35 <0.05 0.02

TP Season 106.38 <0.0001 0.62 0.98 NUPW-UPW

Station 29.26 <0.0001 0.26 S1-S2, S1-S3, S1-S4, S2-S4, S3-S4


TIP Season 27.78 <0.0001 0.44 0.94 Station 11.94 <0.0001 0.28 S1-S4, S2-S4, S3-S4


TOP Season 52.56 <0.0001 0.78 0.76

δ13C Season 13089.30 <0.0001 0.98 0.98 NUPW-UPW Station 124.90 <0.0001 0.01 S1-S3, S1-S4, S2-S3,

S2-S4, S3-S4 Interaction 10.18 <0.0001 0.001

δ15N Season 88.46 <0.0001 0.71 0.94 NUPW-UPW

Station 6.97 0.001 0.08 S1-S3, S3-S4 Interaction 9.51 0.0002 0.09

OM, TOC, TN, TP, TIP, δ15N data were 1/x transformed while TOP data was √x. δ13C data were not transformed.

Significant differences in Tukey HSD test (P <0.05).

Ƞ2=0.01 represents a small effect, Ƞ2=0.10 a medium effect, Ƞ2=0.25 a large effect (Keith, 2015).

2.4.2. Organic carbon and nitrogen isotopic composition in the Taganga Bay sediments

The sedimentary C and N isotopic composition displayed significant spatial and seasonal variations

(Fig. 2.2). The most and last depleted average δ13C values were measured in S1 and in S4,

respectively, both during the UPW season (Table 2.1). The seasonal effect was considerably


39

stronger than the spatial variation and the interaction (Table 2.2). The pairwise comparisons

between S1 and S2 did not show a clear distinction, but differences between these stations and

stations S3 and S4 were significant. Most and least depleted values of δ13C were negatively

correlated with TOC, OM, TN, TP, and TIP but not with TOP, or δ15N (Table 2.S2, supplementary

material).

Values of δ15N also displayed high spatial and temporal variability with the more depleted values

measured in S1 and S2 during the NUPW season (Table 2.1). Average δ15N values increased from

the NUPW to UPW season, in particular at S2, and the effect of season was greater than the effect

of station and of the interaction. Pairwise comparisons suggested significant differences between

δ15N values measured at S1-S3 and S3–S4 (Table 2.2). Outcomes of pairwise and Pearson’s

coefficient indicate that spatial distribution of δ15N did not correspond to that of TN. Furthermore,

there were no significant correlations with other sedimentary properties.

The molar ratio TOC:TN calculated at the different sites and over the two sampling seasons varied

by an order of magnitude, from 3 to 24 (Table 2.1). In general, TOC:TN molar ratios were lower

at S1 and S3 in both seasons; the highest ratio was found at S4 during the NUPW season.

The relative contribution of terrestrial organic carbon ranged from 13.0 to 94.9 %, with the highest

average contribution calculated for S1 during the UPW season (86.8 ± 8.4 %) and declining from

S1 to S4 (Fig. 2.3). At S3, 750 m far from the outfall, the average terrestrial contribution of organic

C nearly halved as compared to S1 (45.5 ± 7.3 %, n=8) while at S4, 1.8 km far from the outfall, it

was minor (18.9 ± 5.7 %, n=8) and marine sources prevailed. There was an increase of terrestrial

TOC average values from the NUPW to UPW season for S1 and S2, whereas the opposite occurred

at S3 and S4.

2.4.3. Aerobic and anaerobic metabolism

Oxygen uptake rates displayed clear spatial and seasonal trends, with decreasing respiration along

the transect S1-S4 and with higher rates measured at S1 and S2 during the UPW as compared to

the NUPW season (Table 2.3 and Fig. 2.4). The effects of season and sampling station were highly

significant but there was an interaction between the two factors (Table 2.3). Rates of oxygen

consumption ranged from -1018.4 to -4519.0 µmol O2·m-2·h-1 (NUPW season) and from -1269.6

to -6311.0 µmol O2·m-2·h-1 (UPW season). In both investigated seasons the average O2 uptake rate

measured at S1 was nearly 3.6 times higher than that at S4. During the UPW season oxygen demand


40

increased significantly (S1 and S2, by 24 and 62%, respectively) or tended to increase (S3 and S4,

by 4 and 11%, respectively) as compared to the NUPW season. Variation coefficients were

generally low (<15%) at all stations and sampling seasons. Sediment sieving at the end of the

incubations revealed the occasional presence of a few, small oligochaetes.

Fluxes of ferrous iron, manganous manganese and methane were below the limit of detection of

our methods (<1 µmol Fe2+ or Mn2+ m-2h-1 and <0.1 µmol CH4 m-2h-1) and are not reported.

Fig. 2.2 Signatures of δ13C (V-PDB) and δ15N(AIR) in surface sediments (0-3 cm) of 4 stations

(S1 to S4) located along increasing distances from a sewage outfall. Station S1 is the closest to the

outfall. Measurements were performed in November 2017 and in January – February 2018, during

non upwelling (NUPW) and upwelling season (UPW), respectively.


41

Fig. 2.3 Relative contribution of terrestrial versus marine organic carbon in surface sediments (0-

3 cm depth) calculated for 4 stations along increasing distance from a sewage outfall in Taganga

Bay – Colombian Caribbean. Measurements were carried out during the non upwelling (NUPW)

and upwelling (UPW) season. Averages ± standard deviations (n=4) are reported.

2.4.4. Net N2 fluxes and rates of denitrification and nitrate ammonification

The net fluxes of N2 were characterized by small scale variability, as suggested by high variation

coefficients, and by marked seasonal differences, depending upon the interaction with the factor

site (Fig. 2.5 and Table 2.3). Rates were both positive, suggesting the dominance of denitrification

over N2 fixation, and negative, suggesting the dominance of N2 fixation. In the transition from the

NUPW to the UPW season negative N2 fluxes measured in S1, S2 and S3 became positive whereas

at station S4 net fluxes were negative in both seasons. Net N2 fluxes ranged from -109.6 to 67.6

µmol N m–2 h–1 during the NUPW season and from 50.0 to 186.3 µmol N m–2 h–1 during the UPW

season.

S a m p l i n g s i t e

S 1 S 2 S 3 S 4

Ter

rest

rial T

OC

(%

)

0

2 0

4 0

6 0

8 0

1 0 0

NUPW

UPW


42

Calculated rates of 28N2 production from the IPT concentration series were always independent

from the level of 15NO3- added to the core water phase, suggesting that ANAMMOX was not

contributing to N2 production at the sampling stations. Rates of total denitrification were very low

at three out of four stations (2<x<4 µmol N m-2h-1) with S3 as only exception (8<x<11 µmol N m-

2h-1). At all sampling sites denitrification was mostly sustained (90%) by production of nitrate

within sediments by nitrification (Fig. 2.6). Total denitrification (Dtot) showed high significant

spatial and temporal variations (Table 2.3). Lowest average values of Dtot were found at S1 during

the UPW season while highest values were observed at S3 during the NUPW season. Dtot

decreased at S1 (by ∼ 35 %) and S3 (by ∼ 25 %) whereas rates remained virtually unchanged at S2

and S4 between the NUPW and UPW season. During the same time lag, average rates of Dw

increased in all stations. At S1, S2 and S3 Dw increased by a factor of ∼2.4 whereas at S4 rates

increased by a factor of ∼ 3.

Rates of DNRA were consistently higher during the UPW season, except at S4, and during both

seasons DNRA rates decreased from S1 to S4 (Table 2.3 and Fig. 2.7). More in detail, S1 and S2

displayed similar rates, higher than those in S3 and S4, during the NUPW season, whereas during

the upwelling period DNRA rates decreased from S1 to S4. Significant positive correlations were

found between DNRA and TOC and Dw whereas negative correlations were observed between

DNRA and Dn. (Table 2.S2, supplementary material). The comparison of the share of nitrate

consumption between Dtot and DNRA indicated that the latter process was responsible for a major,

similar fraction of nitrate consumption (60 to 65%) at stations S1 and S2 in both UPW and NUPW

seasons. The contribution of DNRA to nitrate uptake was smaller and consistent in the two seasons

at station S3 (15-17%) and S4 (29-31%).

2.4.5. Nutrient fluxes at the sediment-water interface

Dissolved inorganic N, P and Si fluxes were low as compared to what expected from oxygen

demand and the elemental composition of sediments and did not follow clear patterns (Fig. 8).

Fluxes of NO3- and NH4

+ were significantly different between sampling periods and among stations

(Table 2.3). The effect of season was stronger for NO3- while the effect of the interaction was

important for NH4+. NO3

- flux dominated DIN fluxes at S1 in both seasons, and at S2 and S4 only

in the UPW season. Significant positive correlations were found between NO3- and O2 fluxes (Table

2.S2, supplementary material). Ammonium fluxes dominated inorganic N exchanges at S3 in both


43

seasons, while at S2 and S4 they were relevant only in November. Fluxes of NO2- were negligible

and are not reported.

Fig. 2.4 Dark oxygen fluxes measured at the 4 stations during the non upwelling (NUPW) and

upwelling (UPW) season. Averages (n=5) and standard deviations are reported.

DIP fluxes were generally low and ranged from -0.56 to 0.17 µmol P·m-2·h-1. They displayed a

significant temporal variation whereas there was no significant difference among stations (Table

2.3). Significant spatial and seasonal effects were determined also for SiO2 fluxes, that ranged from

-3.19 µmol Si m–2 h–1 at S3 during the UPW season to 7.04 µmol Si m–2 h-1 measured at S3 during

the NUPW season. Sediments were sources of SiO2 at S1 and S2 in both seasons and at S3 and S4

during the NUPW season, whereas SiO2 was taken up at S3 and S4 during the UPW season.

Sampling site

S1 S2 S3 S4

O2 f

lux (

µm

ol m

-2h

-1)

-6000

-5000

-4000

-3000

-2000

-1000

0

NUPW

UPW


44

Fig. 2.5 Dark net N2 fluxes measured as N2:Ar ratios at the 4 stations during the non upwelling

(NUPW) and upwelling (UPW) season. Averages (n=5) and standard deviations are reported.

Fig. 2.6 Rates of denitrification of water column nitrate (Dw) and of nitrate produced via

nitrification (Dn) measured in the dark at the 4 stations during the non upwelling (NUPW) and

upwelling (UPW) season. Averages (n=12) and standard deviations are reported.

Sampling site

S1 S2 S3 S4

N2 flu

x (

µm

ol N

m-2

h-1

)

-100

-50

0

50

100

NUPW

UPW

NUPW

Sampling site

S1 S2 S3 S4

De

nitri

fica

tion

ra

te (

µm

ol N

m-2

h-1

)

0,0

0,1

0,2

0,3

0,4

0,5

0,6

3,0

6,0

9,0

12,0 UPW

Sampling site

S1 S2 S3 S4

Den

itrificatio

n r

ate

(µ

mol N

m-2

h-1

)

0,0

0,1

0,2

0,3

0,4

0,5

0,6

3,0

6,0

9,0

12,0DnDw

DnDw


45

Table 2.3 Results of the two-way ANOVA on the effects of the factors sampling season, sampling

stations and of their interaction on benthic fluxes, denitrification and nitrate ammonification rates.

See the text for more details.

Dependent variable


F P Ƞ2 Adjusted

R2 Tukey HSD test

O2 flux

Season 986.19 <0.0001 0.73

0.99

NUPW-UPW

Station 210.12 <0.0001 0.23 S1-S2, S1-S3, S1-S4, S2-S3, S2-S4,S3-S4

Interaction 25.36 <0.0001 0.03

NO3- flux

Season 561.88 < 0.0001 0.78

0.98

NUPW-UPW

Station 91.11 <0.0001 0.19 S1-S2, S1-S3, S1-S4,

S2-S3, S2-S4 Interaction 5.15 0.005 0.01

NH4+ flux

Season 9.44 0.0006 0.11 0.81

Station 9.31 0.0001 0.16 S1-S4,S2-S4

Interaction 33.14 <0.0001 0.56

DIN

Season 104.30 < 0.0001 0.38

0.89

NUPW-UPW

Station 92.45 <0.0001 0.51 S1-S2, S1-S3,S1-S4, S2-S3, S2-S4, S3-S4

Interaction 8.64 0.0002 0.05

DIP flux Season 161.73 <0.0001 0.89 0.89 NUPW-UPW

DSi flux Season 19.96 <0.0001 0.25

0.97

Station 12.95 <0.0001 0.25 S1-S2,S2-S3,S2-S4 Interaction 15.41 <0.001 0.29

N2 flux Season 13.79 <0.001 0.37

0.53 NUPW-UPW

Interaction 2.49 < 0.05 0.2

Dw Season 220.71 <0.0001 0.78

0.94 NUPW-UPW

Station 10.93 <0.0001 0.06 S1-S2,S1-S3, S2-S4,

S3-S4

Dn Season 162.22 <0.0001 0.61

0.90 NUPW-UPW

Station 38.78 <0.0001 0.22 S1-S3, S1-S4, S2-S3,

S2-S4, S3-S4

Dt Season 168.86 <0.0001 0.62

0.90 NUPW-UPW

Station 39.03 <0.0001 0.21 S1-S3, S1-S4, S2-S3,

S3-S4

DNRA

Season 655.21 <0.0001 0.66

0.98 NUPW-UPW

Station 217.09 <0.0001 0.33 S1-S3, S1-S4, S2-S3,

S2-S4, S3-S4 Interaction 3.30 < 0.05 0.005

Squared root reflected transformation was applying to O2 while ln(x) reflected transformation was used to NO3-. NH4

+, SiO2, N2 and DIN data were

not transformed. DIP data was log(x2) transformed while Dw and DNRA data were log(x). Dn and Dt data were 1/x transformed.

Significant differences in Tukey HSD test (P <0.05).

Ƞ2=0.01 represents a small effect, Ƞ2=0.10 a medium effect, Ƞ2=0.25 a large effect (Keith, 2015).


46

2.4.6. Multivariate analysis

The PCA analysis allowed to identify a subset of sedimentary properties (TOC, δ13C and δ15N) and

of nutrient fluxes (O2, N2, NH4+, NO3

-, DIP, SiO2, Dw, Dn and DNRA) with essential information

for the RDA analysis. TOC, δ13C and δ15N were the most significant variables explaining 41.90 %

of changes in the nutrient fluxes in the two axes of RDA that were statistically significant (p<0.001,

Table 2.S3 and Table 2.S4 in supplementary material). Fluxes of NH4+, SiO2 and DNRA were

directly correlated with the increase of TOC while Dn and NO3- were inversely correlated (Fig. 9

and Fig S2 in supplementary material). TOC and δ13C were not correlated with δ15N, whereas there

was a strong negative correlation between TOC and δ13C. δ15N showed slightly higher correlation

with N2, DIP and Dw than with other fluxes. Results from the RDA align with what described in

previous paragraphs and confirm that δ15N is a poor predictor of the effects of the outfall on the

benthic compartment. Two-way PERMANOVA showed highly significant differences over the

stations, seasons and interaction on nutrient fluxes, Dw, Dn and DNRA. Pairwise test pointed out

that stations and interaction were significantly different from each other (Table 2.4 and Table 2.S5

in Supplementary material).

2.5. Discussion

2.5.1. Sedimentary features

2.5.1.1. Carbon isotopic signature along the transect as tracer of anthropic impact

Carbon isotopic signature along the analyzed transect traced the spatial variation of anthropogenic

TOC, TP and TIP in sediments, as reported elsewhere by Church et al. (2006); Gao et al. (2008)

and Dang et al. (2018). In Taganga Bay the most depleted average δ13C values were found in the

sediments with the highest oxygen average uptake, chemically reduced conditions (very negative

redox potential and buildup of free sulphides in the pore water, and presence of Beggiatoa mats),

suggesting a large, measurable impact of the organic matter delivered from the outfall. The average

δ13C values at S1 and S2 were similar to or more depleted than values reported in marine sediments

impacted by untreated and primary treated sewage effluents in Portugal, Canada and China (-24.0

to -25.4‰, Sampaio et al. 2010; Burd et al. 2013 Li et al. 2016). In contrast, the less impacted

sediments of station S4 showed the most enriched average δ13C values, lowest oxygen average


47

uptake, oxidized sediments without free sulphides in the upper horizon and higher relative

percentage of marine material.

Fig. 2.7 Dark rates of nitrate ammonification (DNRA) measured at the 4 stations during the non

upwelling (NUPW) and upwelling (UPW) season. Averages (n=4) and standard deviations are

reported.

The sedimentary gradients determined along the transect of anthropogenic organic matter input are

consistent with those indicated by Hargrave et al. (2008). These authors provide a nomogram for

benthic organic enrichment zonation based on Eh and total free sulphides, conforming to our

findings, with sediments at S1 and S2 chemically reduced (Eh < -150 mv), grossly polluted (∼75%

of terrigenous OC) and with Beggiatoa mats clearly visible on the sediment surface. The

appearance of white Beggiatoa mats on sediments is a consequence of the increase (S2- above 500

– 1000 µM) and vertical migration of free sulfides in pore water (Hargrave et al. 2008) which may

explain the production of NH4+, due to DNRA (Preisler et al. 2007). Sediments from station S3

possibly showed a transitory organic enrichment, oxic type B (∼ -50 < Eh < 100) with a lower

relative proportion of terrigenous OC. Sediments from station S4 had positive redox potential and

there were no indications of anaerobic metabolism end-products accumulation in the pore water,

suggesting low impact.

Sampling site

S1 S2 S3 S4

Nitra

te a

mm

onific

atio

n r

ate

s

(µm

ol N

m-2

h-1

)

0

1

2

3

4

5

6

NUPWUPW


48

The influence of sewage effluent on sedimentary features, as evidenced by carbon isotopes

signature, was clearly detectable at 750 m (S3) and 1800 m (S4) far from the outfall (∼45 % and

∼20% contribution of terrestrial organic carbon, respectively). These results suggest that the

analysis of sediments reveals a much more spatially extended impact of the outfall in the study area

than previously evaluated, at least by a factor of 3. In fact, according to García (2013), the

maximum extension of an area affected by SMSO discharge is 600 m, if the concentration of fecal

coliforms in the water column is considered. Terrigenous OC at S4 may possibly reflect a high

background value in Taganga Bay, due to the influence and/or mixing of wastewater outfall with

other anthropogenic sources such as riverine inputs, surface runoff and harbor activities due to

currents and transport process.

The significant interactions between the effects produced by the submarine outfall and the

upwelling might be partly due to the overlapping of the UPW and the high touristic season, when

the organic and nutrient loads generated by the town of Santa Marta peak. It can be expected that

the further increase of the loads generated by the outfall may act in synergy with the effects of the

upwelling on particulate matter, nutrient and oligoelements mobilization. The two-end-member

mixing model I employed supports this statement as it suggests during the UPW season an increase

of terrestrial OC at S1 and S2 and an increase of marine OC at S3 and S4. Our interpretation for

what observed close to the outfall is that the increase of the sewage loads generated during the

highest tourist season of Santa Marta city (December-January) produces a much large, local effect

than that generated by the UPW. On the contrary, the marine OC increment calculated for the outer

stations is the result of the dominant net effect of the UPW-associated particulate transport.

Table 2.4 Results of two-way PERMANOVA to test the effects of the factors sampling season

and stations on nutrient fluxes, Dn, Dw and DNRA.


F Adjusted R2 P Pairwise

Season 65.889 0.12 <0.001 Station 134.295 0.72 <0.001 S1-S3, S1-S4, S2-S3, S2-

S4 S3-S4 Season:Station 22.210 0.12 <0.001 Table S5. Supplementary

Material


49

Fig. 2.8 Dark fluxes of inorganic nutrient measured at the 4 stations during the non upwelling

(NUPW) and upwelling (UPW) season. Averages (n=5) and standard deviations are reported.

2.5.1.2. The isotopic signature of N in sediments suggests multiple co-occurring processes

The nonlinear variation of δ15N along the analyzed transect likely reflects the complexity of the

biogeochemical N-cycle (Bedard-Haughn et al. 2003; Meyers 1997). Such complexity might be

exacerbated in impacted ecosystems, where the level of fractionation can be affected by short-term

temporal or spatial variability of dominant microbial N transformations (Bedard-Haughn et al.

2003). The average δ15N values at S1 and S2 were similar or more enriched than values reported

in other impacted areas in Portugal, Canada and China (2.1 to 5.2‰, Sampaio et al. 2010; Burd et

al. 2013; Li et al. 2016). δ15N-enriched values were also observed at S3, where total denitrification

rates were nearly 3-4 times higher than at the other stations. This is in agreement with the

preferential consumption of 14N by denitrifiers (Mariotti et al. 1981; Bedard-Haughn et al. 2003;

Sampling site

S1 S2 S3 S4

NO

3

- flu

xe

s (

µm

ol N

m-2

h-1

)

-25

-20

-15

-10

-5

0

5

NUPW

UPW

Sampling site

S1 S2 S3 S4

NH

4

+ flu

xe

s (

µm

ol N

m-2

h-1

)

-4

-2

0

2

4

6

8

10

12

NUPW

UPW

Sampling site

S1 S2 S3 S4

PO

4

3- f

luxe

s (

µm

ol P

m-2

h-1

)

-0,3

-0,2

-0,1

0,0

0,1

0,2

NUPW

UPW

Sampling site

S1 S2 S3 S4

SiO

2 flu

xe

s (

µm

ol S

i m

-2h

-1)

-4

-2

0

2

4

6

NUPW

UPW


50

Alkhatib et al. 2012; Rooze and Meile 2016). Slightly depleted δ15N values at S4 could be due to

N2 fixation rates (Bedard-Haughn et al. 2003).

What emerges from our seasonal results is that processes as denitrification, leading to higher δ15N

signatures, and N2-fixation, leading to lower δ15N signatures, may vary over relatively short time

and spatial scales. Net N2 fluxes in fact were mostly negative during the NUPW season, when the

benthic system need to import reactive nitrogen for microbial growth, to sustain heterotrophic

processes in organic-enriched sediments. During the UPW season, only a few months later, nitrate

concentrations in the water column nearly doubled (from 1.10 to 1.89 µM) and organic inputs from

terrestrial and marine sources increased. Under these circumstances the net fluxes of N2 were

reversed in 3 out of 4 stations, suggesting that denitrification exceeded-fixation.

2.5.1.3. The stoichiometry of organic C and N

The sedimentary ratios of TOC and TN have been widely used as proxies to elucidate the source

and fate of organic matter in marine sediments (Meyers 1997; Li et al. 2016; Wang et al. 2018).

Generally, TOC:TN ratios higher than 15 characterize terrigenous material (Meyers 1997; Wang

et al. 2018) whereas values close to 6-8 are generally associated to fresh phytoplankton detritus

(e.g. from upwelling waters, as determined in the study area by Bayraktarov & Wild (2014)). Ratios

at S1, S2 and S3 indicate a mixture of terrestrial and marine sources, whereas at S4 the high

TOC:TN ratio is unexpected. S4 is far from the outfall and has the lowest relative proportion of

terrigenous organic carbon; its high TOC:TN ratio can be explained by settling of micro- and

macroalgae growing on reef substrate due to anthropogenic nutrient input and upwelling, as

reported by Vega-Sequeda et al. (2008) at Punta Venado, near station S4. High TOC:TN ratios in

algal organic matter can be due to N-limitation, supported by negative N2 fluxes, enhanced

synthesis of lipid-rich organic matter or enhanced preservation of carbon-rich components and

degradation of nitrogen-rich components (Meyers 1997). Similarly to δ15N, TOC:TN ratios have

limitations as tracers of organic pollution and do not allow to discriminate organic matter sources

to marine sediments due to the complexity of N-cycle biogeochemical processes (Bedard-Haughn

et al. 2003; Meyers 1997).


51

2.5.2. Metabolic rates

2.5.2.1. Oxygen and nutrient fluxes

Regardless the origin of the organic matter to sediments and despite marked decrease of bottom

water temperatures during the UPW season, the OC input from the marine side stimulated or

maintained elevated oxygen uptake in all stations. Bayraktarov and Wild (2014) found that from

75 to 79 % of the annual organic matter input to the sediments of Santa Marta coastal area was

supplied during UPW, with higher inputs at exposed than at sheltered sites. They also reported

higher oxygen uptake at sheltered sites due to lower water currents promoting particles

sedimentation, which is consistent with oxygen uptake rates at S4 that were 26 and 23 % higher

than at S3 in the NUPW and UPW seasons, respectively.

Sedimentary oxygen uptake peaked at S1 with rates comparable or even higher than those reported

for estuaries and marine ecosystems that receive high loadings of organic matter (Table 5). At

station S1 rates of oxygen and nitrate consumption were nearly 4 and 12 times higher than those

measured far from the outfall, respectively. As sediment sieving revealed the absence of significant

contribution of macrofauna to benthic processes, I speculate that most oxygen consumption was

microbial or chemical and that most small-scale variability was due to the distribution of Beggiatoa

colonies, patchily colonizing surface sediments at S1 and S2.

Inorganic nutrient fluxes (ammonium, reactive Si and P) were generally low when compared to

fluxes measured in different impacted coastal areas (Table 2.5) and did not show a clear spatial

gradient as it was observed for sedimentary features, oxygen uptake rate and nitrate fluxes. The

absence of a consistent trend of inorganic nutrient regeneration between stations proximal to and

distant from the outfall, suggests mechanisms that are uncoupled to oxygen uptake and not redox-

dependent. Alternatively, decomposition might be incomplete and lead to the release of dissolved

organic forms of N and P, which were not measured in the present study. Limited efflux of

inorganic N, P and Si might be a consequence of poor bioturbation by macrofauna in the Taganga

Bay sediments, in turn due to organic-impacted and sulphidic sediments. Measured inorganic N,

Si and P regeneration might be low as compared to expected rates (i.e. from oxygen uptake and

sedimentary C:N or C:P ratios) due to the refractory macromolecular quality of the terrestrial

organic carbon and to inefficient anaerobic mineralization rates. The presence of Beggiatoa mats

at S1 and S2 and the absence of measurable Fe2+, Mn2+ and CH4 fluxes suggests that most of the


52

organic matter is mineralized via suphate reduction and that large fraction of the oxygen uptake is

due to reoxidation processes.

Fig. 2.9 RDA plot showing the angles vectors representing sedimentary properties and those

nutrient fluxes reflect their (linear) correlation (scaling 1).

As the sewage outfall has an installed capacity of 2.5 m3s-1 (García 2013) the loads of organic

matter and nutrients from the outfall are expected to increase. Under this scenario, sedimentary

features and fluxes should be continuously monitored as saturation of processes retaining nutrients

or replacement of denitrification by DNRA may turn sediments permanent sources of N and P to

the water column. I do not think that present analytical techniques allow to trace significant increase

of nutrient concentration due to sediment regeneration in the water mass, as the dilution is very

large. Attempts, made in the proximities of floating fish cages, revealed no differences in nutrient


53

concentrations along increasing distances from the cages. However, bioassays based on

phytoplankton revealed much higher primary production (eutrophication, sensu Nixon, 1995) in

the proximity of cages. This ultimately means that sediments provide much evident and measurable

impact of the outfall than the water column dissolved nutrients, that sedimentary nutrient fluxes

need to me monitored and that phytoplankton may respond to very small increase of water column

nutrients (Dalsgaard et al., 2006).

The recent increase of phytoplankton blooms in the Santa Marta coastal area, including potentially

toxigenic microalgae, such as Pseudo-nitzschia, Anabaenopsis, Gonyaulax, Gyrodinium,

Gymnodinium, Prorocentrum, Scripsiella y Cochlodinium represents a warning signal about the

effects of anthropogenic nutrient inputs to this system (Garcés-Ordóñez et al. 2016). Such blooms

suggest subtle changes in water (and sediment) chemistry that likely affect nutrient concentrations

and stoichiometry and that should be carefully considered.

2.5.2.2. Microbial N transformations

Discrepancies between measurements based on N2/Ar and IPT have been widely discussed; N2/Ar

might overestimate true rates of denitrification due to reactions between O2 and N2 in the mass

spectrometer ion source. As during dark incubations O2 decreases, such reactions may lead to an

apparent increase of N2 (Eyre et al. 2002, 2013; Lunstrum Aoki 2016). As the MIMS employed in

our study was equipped with a copper reduction column maintained at 600°C I think that most

oxygen interferences were removed. Also the IPT presents some limitations depending on site-

specific sediment features. At S1and S2 for example visible mats of Beggiatoa might have affected

the results of my denitrification measurements (Song et al. 2013). Beggiatoa have large

intracellular pools of nitrate which prevent correct evaluation of 15NO3- to 14NO3

- ratio in the

denitrification zone, possibly leading to a slight underestimation of the nitrate reduction rates. I

acknowledge that future application of the IPT in the Taganga Bay sediments with Beggiatoa will

require to determine the vertical distribution and intracellular NO3- content of these microbes,

together with additional experiments specifically targeting rates in individual cells (see the review

by Robertson et al. (2019) on how to tackle IPT limitations).

Results from N2/Ar ratio measurements suggest a seasonal shift between the dominance of N-

fixation and denitrification in 3 out of 4 sites, that may be related to increased availability of water

column nitrate or ammonium from mineralization processes during the UPW season. A similar


54

dynamic, regulated by large seasonal availability of reactive N, is reported by Newell et al. (2016)

in the Waquoit Bay and by Zilius et al. (2018) in a coastal lagoon.

Results from the IPT suggest the co-occurrence of DNRA and denitrification in sediments, as also

reported elsewhere (Burgin and Hamilton 2007; Gardner and McCarthy 2009; Newell et al. 2016;

Bonaglia et al. 2014). The seasonal increase of water column nitrate was addressed to the upwelling

(Arévalo-Martínez and Franco - Herrera 2008; Ramírez-Barón et al. 2010) and not to the outfall,

due the huge dilution factor. Such increase was likely responsible for higher nitrate demand by

organic sediments via DNRA or it may have favored the growth of primary producers competing

with bacteria and potentially impacting corals (Christensen et al., 2000; Diaz-Pulido and Garzón-

Ferreira 2002; Dalsgaard 2003;Franco-Herrera et al. 2006). In Taganga Bay DNRA increased from

the NUPW to UPW season in all stations.

DNRA increases with salinity and eutrophication likely due to higher sulphate reduction rates and

sulphide accumulation, as shown in the present study and by others (Gardner et al. 2006; Dong et

al. 2011; Bernard et al. 2015). The predominance of DNRA over denitrification in tropical estuaries

may be due to both an energetic advantage of nitrate ammonifiers over denitrifiers when competing

for limited nitrate, and to higher affinity for nitrate by the nitrate ammonifiers at tropical

temperatures permanently > 24 °C. These explanations can be realistic for my case study, with

rates of DNRA having similar patterns as those of oxygen demand and sulphide concentration in

pore water.

Station S3 displayed the highest rates of denitrification in the two campaigns, likely due to

moderate carbon enrichment, lower microbial oxygen consumption, absence of free sulphides in

porewater and a molar ratio TOC:TN similar to Redfield ratio (8 ± 3). Ammonium availability and

oxygen penetration likely drove nitrification in surface sediments, leading to high rates of coupled

denitrification. At S4 denitrification rates were low likely due to the scarce availability and high

competition for reactive nitrogen, resulting in net import and minimum loss.

Expected increase of organic load to the Taganga Bay sediments from the outfall and during the

UPW/touristic season will turn the benthic system more and more sulphidic. The suppression of

denitrification and the increase of DNRA may ultimately lead to lower net N losses, higher NH4+

mobility and stimulation of algal growth, as reporter in Bernard et al (2015). Results from this

study support the evidence of strong and complex relationships between variations in

oceanographic conditions and biological process in the near coastal zone and of positive feedback


55

between upwelling and impacts associated to the outfall and the touristic season (Paramo et al.,

2011).

2.6. Conclusions

Sediments seem to track past and ongoing impacts from a submarine outfall better than the water

column. Taganga Bay sediments allow to trace the organic pollution from a sewage outfall and

reveal that the spatial extent of the impact is much larger, by a factor of 3, than what hypothesized

from water column analysis of fecal coliforms. Seasonal analysis of sediment features and

metabolic rates suggest also an interaction between the effects of submarine outfall and those of

upwelling, due to additional nutrient and organic matter inputs from the open sea to the coastal

area. The cascading effects of organic inputs, increasing respiration, sulphide build up, suppression

of denitrification and increase of ammonification may dramatically impact the macrofauna

community, leading to species loss over increasing areas, and stimulate the growth of algae, with

negative implications for corals, water transparency and tourism. This is a real concern, as the loads

released by the outfall are expected to increase. Tropical areas as Taganga Bay seem intrinsically

more vulnerable to these impacts than temperate areas due to the different temperature regimes,

which may easily uncouple microbial respiration and oxygen availability.


56

Table 2.5 Fluxes of O2 and nutrients and rates of denitrification, N2 Fixation and DNRA from different estuarine zones (µmol·m-2·h-1). Site Zone O2 NO3

- NH4+ DIP SiO2 DIN Dt N-fixation DNRA Methode

Reference

Boston Harbor, Massachusettsa

Temperate -291.67 - 6250

- - -16.67 – 325.00

- -1.25 – 883.33

167 - - Stoichiometric Giblin et al. (1997)

Whites Point Los Angeles County sewage outfallb

Temperate -875.00 - -958.33

- 20.83 – - 41.67 c

87.50 – 137.50

11.25 – 42.92

75.00 – 166.67

- - - - - Berelson et al. (2002)

Laguna Madre and Baffin Bay, Texas

Sub-tropical

- 3.8 ± 3.0 110±17 - - - 32 ± 5 58 ± 5 41 ± 13 ICCF/MIMS N2:Ar 15NO3

−addition

Gardner et al. (2006)

Sabine Lake, Texas

Sub-tropical

- 1.29 ± 5.07

34.7 ± 23.1

- - - 30.1 ± 10.1

49.7 ± 14.4

0.91 ± 0.81

East Matagorda Bay, Texas

Sub-tropical

- 11.1 ± 3.2 23.2 ± 4.8

- - - 22.5 4.4

2.8 ± 2.1

3.6 ± 0.6

Nueces Estuary, Texas

Sub- tropical

- 0.7 ± 1.1 38.7 ±16.8

- - - 40.0 ± 6.6 19.1 ± 7.4 10.3 ± 4.0

Florida Bay, USA

Sub -tropical

- -2.0 – 2.0

7 – 540 0.1 – 13 - - 0 - 350 - 10 -240 MIMS Gardner & McCarthy

(2009) Mae Klong, Thailand

Tropical - 3800 ± 300

- - - - - 0–7.4 - 0.3–22.8

Acetylene Block

Dong et al.

(2011)

Cisadane, Indonesia

Tropical - 5800 ± 400

- - - - - 0–103 - 0–1137

Vunidawa-Rewa, Fiji

Tropical - 3900 ± 400

- - - - - 0–2.6 - 0.2–10.2

Moreton Bay Australia

Sub –tropical

-257 ± 32 - -988 ± 35

0 -35 -10 – 60 - - - 6.5 ± 2 - 32 ± 5

40–67 - MIMS N2:Ar Ferguson & Eyre (2012)

Little Lagoon, Alabama, USA

Sub -tropical

-342 - - 439

-93 - -121

45–110 -1.0 – 1.0 - - 13 - 55 1.6 - 2.6 1-168 MIMS Bernard et al. (2015)

Golfo de Mexico, hypoxic zone

Sub – tropical

-408 - - 1800

−26.59± 11.49

77.4±14.6 3.29 ± 1.66

- 75.81± 16.13

2–492 0–147 ± 39 1.12±0.44 - 31.4±10.8

ICCF/MIMS N2:Ar, 15NO3

− addition

(McCarthy et al. 2015)

Waquoit Bay, Massachusetts.

Temperate -997 ± 97d ND 540± 51d 34 ± 8d - - 0–28 ± 14 49 ± 8 – 103 ± 4

5.7 ± 4.8 ICCF /MIMS N2:Ar 15NO3

− and 30N2 additions

Newell et al. (2016)

Taganga Bay – Caribbean Colombia

Tropical -1018.4 - -6311.0

-21.36 - 2.83

-3.36 - 13.43

-0.56 - 0.17 -3.19 - 7.04 - 12.3 -14.8 1.27 – 14.87

1.00 – 5.40 ICSI/MIMS N2:Ar, 15NO3

− addition

This study

a Nutrient fluxes values at the site where sewage sludge was directly pumped before implementing wastewater treatment in Boston Harbor, Massachusetts. The average annual oxygen consumption was 5541.67 µmolO2·m-2·h-1 b Nutrient fluxes were measured by free-vehicle benthic chamber devices (landers). The wastewater discharge is located at 60 m. Receiving medium temperature between 11 and 14 °C c Nitrate flux. dIncubation 5 h. eICCF intact core continuous-flow incubations, ICSI intact core static incubations, MIMS membrane inlet mass spectrometry.


57

2.7. Appendices Chapter 2

2.7.1. Materials and methods

Additional information about sediment sampling and features and summary of sedimentary

and process analyses (Fig. 2.S1 and Table 2.S1).

Fig. 2S.1 The images show the overnight preincubation of intact sediment cores collected

from S1, submersed and with the stirring system on (a) and how surface sediments from S1

(b) and S4 (c) looked. In sediments from S1 a well-developed mat of Beggiatoa is evident.


58

Table 2S.1 Summary of sedimentary features analysis and of process rate measurements

carried out at each station.

Measurement Method Replicates (Layer) Parameters

Sedi

men

t-w

ater

fl

uxes

Gas fluxes Membrane inlet

mass spectrometry 5 O2:Ar, N2:Ar, CH4,

Nutrient fluxes

Spectrophotometry 5 NH4+, NO2

-, NO3-, PO4

3-, SiO2

Metal fluxes Atomic absorption 5 Fe2+, Mn2+

Mic

robi

al N

pr

oces

ses

Denitrification (IPT,

concentrations series)

Membrane inlet mass spectrometry

12 (4 x 3 15NO3- level) 29N2, 30N2

DNRA (15NO3

- addition)

Membrane inlet mass spectrometry

4 15NH4+

Sedi

men

tary

fea

ture

s

Total C and N and isotopic composition

Mass spectrometry 4 (0-3 cm) TOC, TN, δ13C, δ15N

Total P, total inorganic P

Spectrophotometry 4 (0-3cm) TP, TIP (TOP=TP-TIP)

Organic content

Loss on ignition 4 (0-3 cm) OM

Redox potential

Potentiometry 5 (0-3 cm) Eh

Pore water sulphides

Spectrophotometry 5 (0-3 cm) H2S+HS-+S2-


59

2.7.2. Results

Additional information about statistical analysis results: Person’s correlations (Table S2),

RDA analysis (Fig. 2.S2, Table 2.S3 and Table 2.S4) and pairwise test of PERMANOVA

(Table 2.S5).

Table 2S.2 Pearson correlation coefficients (r) between sedimentary features and N-related

processes

Dn DNRA Dw NO3- O2 OM TIP TN TOC

TOC:TN TOP TP δ13C

δ15N

Dn 1.00

DNRA -0.48b 1.00

Dw -0.19 0.45b 1.00

NO3- 0.18 -0.49b -0.28 1.00

O2 0.34 -0.87a -0.42b 0.65a 1.00

OM -0.21 0.70a 0.30 -0.87a -0.78a 1.00

TIP -0.22 0.73a 0.40b -0.73a -0.79a 0.87a 1.00

TN -0.23 0.69a 0.40b -0.88a -0.78a 0.94a 0.88a 1.00

TOC -0.41b 0.72a 0.39 -0.75a -0.75a 0.88a 0.91a 0.86a 1.00

TOC:TN -0.48 0.06 -0.11 0.23 0.09 -0.10 0.04 -0.19 0.24 1.00

TOP -0.07 0.28 0.33 -0.51b -0.26 0.54b 0.17 0.45b 0.32 -0.31

1.00

TP -0.20 0.72a 0.49b -0.81a -0.76a 0.94a 0.89a 0.92a 0.88a -0.09

0.58b 1.00

δ13C 0.17 -0.90a -0.37 0.49b 0.82a -0.70a -

0.76a -0.69a -0.70a -0.03

-0.26 -

0.75a 1.00

δ15N -0.01 0.27 0.54b 0.24 -0.29 -0.04 0.13 0.01 0.08 0.07

-0.03 0.13 -0.29 1.00

aSignificant at P < 0.0001 bSignificant at P <0.05


60

Fig. 2S.2 RDA plot showing the angles between TOC, δ13C and δ15N that reflect linear

correlation (scaling 2).


61

Table 2S.3 Sedimentary properties significantly correlated with the axes in RDA analysis

RDA1 RDA2 Adjusted R2 P δ13C 0.997 0.074 0.77 0.001 δ15N -0.334 0.943 0.56 0.001 TOC -0.993 -0.122 0.74 0.001

Significance level P<0.05

999 permutations

Table 2S.4 Significance test of the relationship between the sedimentary properties and

nutrient fluxes on RDA1 and RDA2.

Df Variance F P Model 3 4.12 7.89 0.001 Residual 28 4.88

Significance level P<0.05

999 permutations


62

Table 2S.5 Pairwise test of interactions in two ways PERMANOVA on nutrient fluxes, Dn,

Dw and DNRA

Interactions (Season:Station) F Adjusted R2 P NUPW-S1 vs NUPW-S2 139.53 0.96 0.029 NUPW-S1 vs NUPW-S3 140.64 0.96 0.024 NUPW-S1 vs NUPW-S4 214.07 0.97 0.036 NUPW-S1 vs UPW-S1 7.34 0.55 0.059 NUPW-S1 vs UPW-S2 8.72 0.59 0.055 NUPW-S1 vs UPW-S3 151.53 0.96 0.024 NUPW-S1 vs UPW-S4 219.09 0.97 0.038

NUPW-S2 vs NUPW-S3 3.16 0.34 0.119 NUPW-S2 vs NUPW-S4 59.81 0.91 0.021 NUPW-S2 vs UPW-S1 54.27 0.90 0.033 NUPW-S2 vs UPW-S2 562.65 0.99 0.035 NUPW-S2 vs UPW-S3 4.92 0.45 0.038 NUPW-S2 vs UPW-S4 105.65 0.95 0.031

NUPW-S3 vs NUPW-S4 20.49 0.77 0.031 NUPW-S3 vs UPW-S1 58.13 0.91 0.024 NUPW-S3 vs UPW-S2 463.46 0.99 0.031 NUPW-S3 vs UPW-S3 1.38 0.19 0.249 NUPW-S3 vs UPW-S4 15.51 0.72 0.031 NUPW-S4 vs UPW-S1 75.20 0.93 0.034 NUPW-S4 vs UPW-S2 676.68 0.99 0.031 NUPW-S4 vs UPW-S3 38.29 0.86 0.019 NUPW-S4 vs UPW-S4 4.52 0.43 0.058 UPW-S1 vs UPW-S2 2.50 0.29 0.182 UPW-S1 vs UPW-S3 57.52 0.91 0.033 UPW-S1 vs UPW-S4 70.90 0.92 0.030 UPW-S2 vs UPW-S3 578.74 0.99 0.029 UPW-S2 vs UPW-S4 861.43 0.99 0.031 UPW-S3 vs UPW-S4 50.76 0.89 0.029

Significance level p<0.05


63

2.8. References

Alkhatib M, Lehmann MF and Del Giorgio PA (2012) The nitrogen isotope effect of benthic

remineralization-nitrification- denitrification coupling in an estuarine environment.

Biogeosciences, 9(5), 1633–1646. https://doi.org/10.5194/bg-9-1633-2012

Aller RC (1988) Benthic fauna and biogeochemical processes in marine sediments: the role

of burrow structures BT - Nitrogen Cycling in Coastal Marine Environments. In T. H.

Blackburn & J. Sørensen (Eds.), Nitrogen Cycling in Coastal Marine Environments

(JOHN WILEY, pp. 301–338).

Aller RC (1994) Bioturbation and remineralization of sedimentary organic matter: effects of

redox oscillation. Chemical Geology, 114(3–4), 331–345. https://doi.org/10.1016/0009-

2541(94)90062-0

An S and Gardner WS (2002) Dissimilatory nitrate reduction to ammonium (DNRA) as a

nitrogen link, versus denitrification as a sink in a shallow estuary (Laguna Madre/Baffin

Bay, Texas). Marine Ecology Progress Series, 237, 41–50.

https://doi.org/10.3354/meps237041

Andrade CA and Barton ED (2005) The Guajira upwelling system. Continental Shelf

Research, 25(9), 1003–1022. https://doi.org/10.1016/j.csr.2004.12.012

Arévalo-Martínez DL and Franco - Herrera A (2008) Características oceanográficas de la

surgencia frente a la ensenada de Gaira, Departamento de Magdalena, época seca menor

de 2006. Boletín de Investigaciones Marinas y Costeras, 37(2), 131–162.

Aspila K, Agemian Hand Chau ASY(1976) A Semi-automated Method for the

Determination of Inorganic , Organic and Total Phosphate in Sediments. Analyst, 101,

187–197.

Banta GT, Giblin AE, Hobbie JE et al. (1995) Benthic respiration and nitrogen release in

Buzzards Bay , Massachusetts. Jornal of Marine Research, 53, 107–135.

Bayraktarov E, Bastidas-Salamanca ML and Wild C (2014) The physical environment in

coral reefs of the Tayrona National Natural Park (Colombian Caribbean) in response to

seasonal upwelling. Boletín de Investigaciones Marinas y Costeras, 43(1), 137–157.


64

Bayraktarov E, Pizarro V, Eidens C et al. (2013) Bleaching susceptibility and recovery of

Colombian Caribbean corals in response to water current exposure and seasonal

upwelling. PLoS ONE, 8(11), 1–11. https://doi.org/10.1371/journal.pone.0080536

Bayraktarov E and Wild C (2014) Spatiotemporal variability of sedimentary organic matter

supply and recycling processes in coral reefs of Tayrona National Natural Park,

Colombian Caribbean. Biogeosciences, 11(11), 2977–2990. https://doi.org/10.5194/bg-

11-2977-2014

Bedard-Haughn A, Van Groenigen JW and Van Kessel C (2003) Tracing15N through

landscapes: Potential uses and precautions. Journal of Hydrology, 272(1–4), 175–190.

https://doi.org/10.1016/S0022-1694(02)00263-9

Berelson WM, Johnson K, Coale K et al. (2002) Organic matter diagenesis in the sediments

of the San Pedro Shelf along a transect affected by sewage effluent. Continental Shelf

Research, 22, 1101–1115.

Bernard RJ, Mortazavi B and Kleinhuizen AA (2015) Dissimilatory nitrate reduction to

ammonium (DNRA) seasonally dominates NO3− reduction pathways in an

anthropogenically impacted sub-tropical coastal lagoon. Biogeochemistry, 125(1), 47–

64. https://doi.org/10.1007/s10533-015-0111-6

Bonaglia S, Nascimento FJA, Bartoli M et al. (2014) Meiofauna increases bacterial

denitrification in marine sediments. Nature Communications, 5(5133), 1005–1011.

https://doi.org/10.1038/ncomms6133

Boynton WR, Ceballos MAC, Bailey EM et al. (2018) Oxygen and nutrient exchanges at the

sediment-water interface: a global synthesis and critique of estuarine and coastal data.

Estuaries and Coasts, 41(2), 301–333. https://doi.org/10.1007/s12237-017-0275-5

Burd B, Macdonald T and Bertold S (2013) The effects of wastewater effluent and river

discharge on benthic heterotrophic production, organic biomass and respiration in

marine coastal sediments. Marine Pollution Bulletin, 74(1), 351–363.

https://doi.org/10.1016/j.marpolbul.2013.06.029

Burgin AJ and Hamilton SK (2007) Have we overemphasized the role of denitrification in


65

aquatic ecosystems? A review of nitrate removal pathways. Frontiers in Ecology and

the Environment, 5(2), 89–96. https://doi.org/10.1890/1540-

9295(2007)5[89:HWOTRO]2.0.CO;2

Carstensen J, Conley DJ, Bonsdorff E, Gustafsson BG et al. (2014) Hypoxia in the Baltic

Sea: Biogeochemical cycles, benthic fauna, and management. Ambio, 43(1), 26–36.

https://doi.org/10.1007/s13280-013-0474-7

Church TM, Sommerfield CK, Velinsky DJ et al. (2006) Marsh sediments as records of

sedimentation, eutrophication and metal pollution in the urban Delaware Estuary.

Marine Chemistry, 102(1–2), 72–95. https://doi.org/10.1016/j.marchem.2005.10.026

Cline JD (1969) Spectrophometric determination of hydrogen sulfide in natural waters.

Limnology and Oceanography, 14(3), 454–458.

https://doi.org/10.4319/lo.1969.14.3.0454

Dalsgaard T (2003) Benthic primary production and nutrient cycling in sediments with

benthic microalgae and transient accumulation of macroalgae. Limnology and

Oceanography, 48(6), 2138–2150. https://doi.org/10.4319/lo.2003.48.6.2138

Dalsgaard T, Nielsen LP, Brotas V et al. (2000) Protocol Handbook for Nitrogen Cycling In

Estuaries. National Environmental Research Institute, Silkeborg, Denmark. T.

Dalsgaard (ed.). ISBN: 87-7772-535-2

Dalsgaard T, Thamdrup B and Canfield DE (2005) Anaerobic ammonium oxidation

(anammox) in the marine environment. Research in Microbiology, 156(4), 457–464.

https://doi.org/10.1016/j.resmic.2005.01.011

Dalsgaard, T., & Krause-Jensen, D. (2006). Monitoring nutrient release from fish farms with

macroalgal and phytoplankton bioassays. Aquaculture, 256(1-4), 302-310.

Dang DH, Evans RD, Durrieu G et al. (2018) Quantitative model of carbon and nitrogen

isotope composition to highlight phosphorus cycling and sources in coastal sediments

(Toulon Bay, France). Chemosphere, 195, 683–692.

https://doi.org/10.1016/j.chemosphere.2017.12.109


66

De Brabandere L, Bonaglia S, Kononets MY et al (2015). Oxygenation of an anoxic fjord

basin strongly stimulates benthic denitrification and DNRA. Biogeochemistry, 126(1–

2), 131–152. https://doi.org/10.1007/s10533-015-0148-6

Diaz-Pulido G, and Garzón-Ferreira J (2002) Seasonality in algal assemblages on upwelling-

influenced coral reefs in the Colombian Caribbean. Botanica Marina, 45(3), 284–292.

https://doi.org/10.1515/BOT.2002.028

Díaz-Rocca LH and Causado-Rodríguez E (2007) La insostenibilidad del desarrollo urbano:

El caso de Santa Marta–Colombia. Clío América, 1(1), 64–100.

Diaz RJ (2001) Overview of hypoxia around the world. Journal of Environmental Quality,


Dong LF, Sobey MN, Smith CJ et al. (2011) Dissimilatory reduction of nitrate to ammonium,

not denitrification or anammox, dominates benthic nitrate reduction in tropical estuaries.

Limnology and Oceanography, 56(1), 279–291.

https://doi.org/10.4319/lo.2011.56.1.0279

Downing JA, Mcclain M, Twilley R et al. (1999) The Impact of Accelerating Land-Use

Change on the N-Cycle of Tropical Aquatic Ecosystems : Current Conditions and

Projected Changes., Biogeochemistry, 46, 109–148.

Escobar A (1988) Estudio de algunos aspectos ecologicos y de la contaminacion bacteriana

en la Bahia de Santa Marta, Caribe Colombiano. Boletin de Investigaciones Marinas y

Costeras, 18, 39–57.

Eyre BD, Maher DT and Squire P (2013) Quantity and quality of organic matter (detritus)

drives N2 effluxes (net denitrification) across seasons, benthic habitats, and estuaries.

Global Biogeochemical Cycles, 27(4), 1083–1095.

https://doi.org/10.1002/2013GB004631

Eyre BD, Rysgaard S, Dalsgaard T and Christensen PB (2002) Comparison of isotope pairing

and N2:Ar methods for measuring sediment denitrification - Assumptions,

modifications, and implications. Estuaries, 25(6 A), 1077–1087.

https://doi.org/10.1007/bf02692205


67

Fajardo G (1979) Surgencia costera en las proximidades de la península colombiana de La

Guajira. Boletin Cientifico CIOH, 2, 7–19.

Ferguson A and Eyre B (2012) Interaction of benthic microalgae and macrofauna in the

control of benthic metabolism, nutrient fluxes and denitrification in a shallow sub-

tropical coastal embayment (western Moreton Bay, Australia). Biogeochemistry, 112(1–

3), 423–440. https://doi.org/10.1007/s10533-012-9736-x

Fox J (2005) The R Commander: A Basic-Statistics Graphical User Interface to R. Journal

of Statistical Software, 14(9), 1–42.

Franco-Herrera A, Castro L and Tigreros P (2006) Plankton dynamics in the south-central

Caribbean Sea: Strong seasonal changes in a coastal tropical system. Caribbean Journal

of Science, 42(1), 24–38.

Gao J, Wang Y, Pan S et al. (2008) Spatial distributions of organic carbon and nitrogen and

their isotopic compositions in sediments of the Changjiang Estuary and its adjacent sea

area. Journal of Geographical Sciences, 18(1), 46–58. https://doi.org/10.1007/s11442-

008-0046-0

Garcés-Ordóñez O, Arteaga E, Obando P et al. (2016) Atención a eventuales emergencias

ambientales en la zona marino-costera del departamento del Magdalena. Convenio

CORPAMAG-INVEMAR; código: PRY-CAM-011-14. Informe técnico final. (Issue 14).

García-Hoyos LM, Franco-Herrera A, Ramire-Barón JS et al. (2010) Dinámica océano-

atmósfera y su influencia en la biomasa fitoplanctónica en la zona costera del

epartamento del Magdalena. Boletín de Investigaciones Marinas y Costeras, 39(2),

307–335.

García F (2013) Modelación de los efectos del emisario submarino de santa marta sobre la

calidad del agua. PhD Dissertation, Universidad de Antioquia, Facultad de Ingeniería.

García F, Palacio C and Garcia U (2012) Water quality at Santa Marta Coastal Area (

Colombia ). Dyna, 79(173), 85–94.

Gardner WS and McCarthy MJ (2009) Nitrogen dynamics at the sediment-water interface in

shallow, sub-tropical Florida Bay: Why denitrification efficiency may decrease with


68

increased eutrophication. Biogeochemistry, 95(2), 185–198.

https://doi.org/10.1007/s10533-009-9329-5

Gardner WS, McCarthy MJ, An S et al. (2006) Nitrogen fixation and dissimilatory nitrate

reduction to ammonium (DNRA) support nitrogen dynamics in Texas estuaries.

Limnology and Oceanography, 51(1 II), 558–568.

Gearing JN, Gearing PJ, Rudnick DT et al. (1984) Isotopic variability of organic carbon in a

phytoplankton-based, temperate estuary. Geochimica et Cosmochimica Acta, 48(5),

1089–1098. https://doi.org/10.1016/0016-7037(84)90199-6

Giblin AE, Hopkinson CS and Tucker J (1997) Benthic metabolism and nutrient cycling in

Boston Harbor, Massachusetts. Estuaries, 20(2), 346–364.

https://doi.org/10.1007/BF02690378

Grall J and Chauvaud L (2002) Marine eutrophication and benthos: the need for new

approaches and concepts. Global Change Biology, 8(9), 813–830.

Grasshoff KM, Ehrhardt and K Kremling (1983) Methods of Seawater Analysis, 2nd ed.

Berlin: Verlag Chemie.

Hall POJ, Almroth E, Bonaglia S et al. (2017) Influence of Natural Oxygenation of Baltic

Proper Deep Water on Benthic Recycling and Removal of Phosphorus, Nitrogen,

Silicon and Carbon. Frontiers in Marine Science, 4(February), 1–14.

https://doi.org/10.3389/fmars.2017.00027

Hargrave BT, Holmer M and Newcombe CP (2008) Towards a classification of organic

enrichment in marine sediments based on biogeochemical indicators. Marine Pollution

Bulletin, 56(5), 810–824. https://doi.org/10.1016/j.marpolbul.2008.02.006

Hopkinson CS, Giblin AE and Tucker J (2001) Benthic metabolism and nutrient regeneration

on the continental shelf of Eastern Massachusetts, USA. Marine Ecology Progress

Series, 224, 1–19. https://doi.org/10.3354/meps224001

Howarth R, Chan F, Conley DJ et al. (2011) Coupled biogeochemical cycles: Eutrophication

and hypoxia in temperate estuaries and coastal marine ecosystems. Frontiers in Ecology

and the Environment, 9(1), 18–26. https://doi.org/10.1890/100008


69

Koop-Jakobsen K and Giblin AE (2009) Anammox in tidal marsh sediments: The role of

salinity, nitrogen loading, and marsh vegetation. Estuaries and Coasts, 32(2), 238–245.

https://doi.org/10.1007/s12237-008-9131-y

Legendre P, Legendre L. Numerical Ecology. 2nd ed. Amsterdam: Elsevier, 1998. ISBN 978-

0444892508.

Li Y, Zhang H, Tu C et al. (2016) Sources and fate of organic carbon and nitrogen from land

to ocean: Identified by coupling stable isotopes with C/N ratio. Estuarine, Coastal and

Shelf Science, 181, 114–122. https://doi.org/10.1016/j.ecss.2016.08.024

Lunstrum A and Aoki LR (2016) Oxygen interference with membrane inlet mass

spectrometry may overestimate denitrification rates calculated with the isotope pairing

technique. Limnology and Oceanography: Methods, 14(7), 425–431.

https://doi.org/10.1002/lom3.10101

Mancera-Pineda J, Pinto G and Vilardy S (2013) Patrones de distribución espacial de masas

de agua en la bahía de Santa Marta, Caribe Colombiano: Importancia relativa del

upwelling y outwelling. Boletín de Investigaciones Marinas y Costeras, 42(2), 329–360.

Mariotti A, Germon JC, Hubert P et al. (1981) Experimental determination of nitrogen kinetic

isotope fractionation: Some principles; illustration for the denitrification and

nitrification processes. Plant and Soil, 62(3), 413–430.

https://doi.org/10.1007/BF02374138

Martínez S and Acosta A (2005) Cambio temporal en la Estructura de la comunidad coralina

del área de Santa Marta - Parque Nacional Natural Tayrona (Caribe Colombiano).

Boletin de Investigaciones Marinas y Costeras, 34, 161–192.

https://doi.org/10.1213/ane.0b013e31816e5128

McCarthy MJ, Newell SE, Carini SA et al. (2015) Denitrification Dominates Sediment

Nitrogen Removal and Is Enhanced by Bottom-Water Hypoxia in the Northern Gulf of

Mexico. Estuaries and Coasts, 38(6), 2279–2294. https://doi.org/10.1007/s12237-015-

9964-0

Mermillod-Blondin F, Rosenberg R, Norling K et al. (2004) Influence of bioturbation by


70

three benthic infaunal species on microbial communities and biogeochemical processes

in marine sediment. Aquatic Microbial Ecology, 36, 271–284.

https://doi.org/10.3354/ame036271

Meyers PA (1997) Organic geochemical proxies of paleoceanographic,paleoimnologic,and

paleoclimatic processes. Organic Geochemistry, 27(5–6), 213–250.

https://doi.org/doi:10.1016/S0146-6380(97)00049-1

Newell SE, McCarthy MJ, Gardner WS et al. (2016) Sediment nitrogen fixation: a call for

re-evaluating coastal N budgets. Estuaries and Coasts, 39(6), 1626–1638.

https://doi.org/10.1007/s12237-016-0116-y

Nielsen LP (1992) Denitrificaction in sediment determined from nitrogen isotope pairing.

FEMS Microbiology Ecology, 86, 357–362.

Nixon SW (1995) Coastal marine eutrophication: A definition, social causes, and future

concerns. Ophelia, 41(1), 199–219. https://doi.org/10.1080/00785236.1995.10422044

Paramo J, Correa M and Núñez S (2011) Evidencias de desacople físico-biológico en el

sistema de surgencia en la Guajira, caribe Colombiano. Revista de Biologia Marina y

Oceanografia, 46(3), 421–430. https://doi.org/10.4067/S0718-19572011000300011

Preisler A, De Beer D, Lichtschlag A et al. (2007) Biological and chemical sulfide oxidation

in a Beggiatoa inhabited marine sediment. ISME Journal, 1(4), 341–353.

https://doi.org/10.1038/ismej.2007.50

Ramírez-Barón JS, Franco-Herrera A, García-Hoyos LM and López-Cerón DA (2010) La

comunidad fitoplanctónica durante eventos de surgencia y no surgencia, en la Zona

Costera del Departamento del Magdalena, Caribe colombiano. Boletín de

Investigaciones Marinas y Costeras, 39(2), 233–263.

Ramírez G (1981) Características Fisico-Químicas de la Bahía de Santa Marta (Agosto 1980-

Julio 1981). Boletín de Investigaciones Marinas y Costeras, 13, 111–121.

Ramos-Ortega LM, Vidal LA, Vilardy S et al. (2008) Análisis de la contaminación

microbiológica (coliformes totales y fecales) en la bahía de Santa Marta, caribe

colombiano. Acta Biológica Colombiana, 13(3), 87–98.


71

Robertson EK, Bartoli M, Brüchert V (2019) Application of the isotope pairing technique in

sediments: Use, challenges, and new directions. Limnology and Oceanography:

Methods, 17(2), 112–136. https://doi.org/10.1002/lom3.10303

Rooze J and Meile C (2016) The effect of redox conditions and bioirrigation on nitrogen

isotope fractionation in marine sediments. Geochimica et Cosmochimica Acta, 184,

227–239. https://doi.org/10.1016/j.gca.2016.04.040

Rosenberg R and Loo LO (1988) Marine eutrophication induced oxygen deficiency: Effects

on soft bottom fauna, western Sweden. Ophelia, 29(3), 213–225.

https://doi.org/10.1080/00785326.1988.10430830

Rueda-Roa DT and Muller-Karger FE (2013) The southern Caribbean upwelling system :

Sea surface temperature , wind forcing and chlorophyll concentration patterns. Deep-

Sea Research Part I, 78, 102–114. https://doi.org/10.1016/j.dsr.2013.04.008

Salzwedel H and Müller K (1983) A summary of Meteorological and hydrological data from

the Bay of Santa Marta, Colombian Caribbean. Boletin de Investigaciones Marinas y


Sampaio L, Freitas R, Máguas C et al. (2010) Coastal sediments under the influence of

multiple organic enrichment sources: An evaluation using carbon and nitrogen stable

isotopes. Marine Pollution Bulletin, 60(2), 272–282.


Schlunz B, Schneider RR, Muller PJ et al. (1999) Terrestrial organic carbon accumulation

on the Amazon deep sea fan during the last glacial sea level low stand. Chemical

Geology, 159(1–4), 263–281.

Smith J, Burford MA, Revill AT et al. (2012) Effect of nutrient loading on biogeochemical

processes in tropical tidal creeks. Biogeochemistry, 108(1–3), 359–380.

https://doi.org/10.1007/s10533-011-9605-z

Song GD, Liu SM, Marchant H, Kuypers MMM Lavik G (2013) Anammox, denitrification

and dissimilatory nitrate reduction to ammonium in the East China Sea sediment.

Biogeosciences, 10(11), 6851–6864. https://doi.org/10.5194/bg-10-6851-2013


72

Testa JM and Kemp WM (2011) Oxygen – Dynamics and Biogeochemical Consequences.

In Treatise on Estuarine and Coastal Science (Vol. 5). Elsevier Inc.

https://doi.org/10.1016/B978-0-12-374711-2.00505-2

Trimmer M, Risgaard-Petersen N, Nicholls JC et al. (2006) Direct measurement of anaerobic

ammonium oxidation (anammox) and denitrification in intact sediment cores Mark.

Marine Ecology Progress Series, 326, 37–47. https://doi.org/10.3354/meps326037

Tucker J and Giblin A (2010) Quality Assurance Project Plan ( QAPP ) for Benthic Nutrient

Flux Studies : 2010. Boston: Massachusetts Water Resources Authority, Report 201, 65

pp.

Vega-Sequeda J, Rodríguez-Ramírez A, Reyes-Nivia MC et al. (2008) Formaciones

Coralinas Del Área De Santa Marta: Estado Y Patrones De Distribución Espacial De La

Comunidad Bentonica. Boletin de Investigaciones Marinas y Costeras, 37(2), 87–105.

Warembourg FR (1993) Nitrogen fixation in soil and plant systems. Nitrogen isotope

techniques, pp.127-156.

Zhou J, Wu Y, Zhang J et al. (2006) Carbon and nitrogen composition and stable isotope as

potential indicators of source and fate of organic matter in the salt marsh of the

Changjiang Estuary, China. Chemosphere, 65(2), 310–317.

https://doi.org/10.1016/j.chemosphere.2006.02.026

Zilius M, Bartoli M, Bresciani M, Katarzyte M et al. (2014) Feedback mechanisms between

cyanobacterial blooms, transient hypoxia, and benthic phosphorus regeneration in

shallow coastal environments. Estuaries and Coasts, 37(3), 680–694.

https://doi.org/10.1007/s12237-013-9717-x

Zilius M, Vybernaite-Lubiene I, Vaiciute D et al. (2018) The influence of cyanobacteria

blooms on the attenuation of nitrogen throughputs in a Baltic coastal lagoon.

Biogeochemistry, 141(2), 143–165. https://doi.org/10.1007/s10533-018-0508-0

Chapter 3. Water column biogeochemistry modeling

73

Chapter 3

Biogeochemical modeling of a tropical coastal area

undergoing seasonal upwelling and impacted by untreated

submarine outfall

3.1. Abstract

Coastal tropical regions with seasonal upwelling display chemical, biological and hydrodynamic

features highly variable in time and space. A coupled 3D hydrodynamic-ecological model was

applied to the Santa Marta Coastal Area (SMCA, Colombian Caribbean) to provide insights into

the role of external stressors (e.g. wastewater outfall and upwelling) on the water quality and

benthic – pelagic coupling. The model was set up, calibrated and validated based on benthic

metabolic measurements carried out within the simulation period, satellite–derived chlorophyll-a

(Chl-a) and sea surface temperature (SST) maps, HYCOM database and field and literature water

quality data.

The model was able to reproduce the magnitude and timing of complex dynamics and fast

transitions of temperature, nutrients, and phytoplankton, including the time and duration of

stratification and mixing periods during the non upwelling (NUPW) and upwelling (UPW) seasons.

The model was also able to capture the effect of fertilization from upwelling and from the outfall

plume. The upwelling triggered a decrease in water temperature and coincided with increases in

wind intensity and the shortest average residence time of the outfall plume. Temperature, light and

nutrients were the factors that limited phytoplankton growth. The plume concentrations of total

organic carbon (TOC), total phosphorus (TP) and phosphate (PO43-) increased slightly under two

scenarios of different wastewater loading. Wide and fast changes in the temperatures and the highly

flushed environment uncoupled phytoplankton growth and nutrient supply in the benthic and

pelagic compartments. The model proved to be a reasonably reliable research and management tool

to predict nutrient and phytoplankton dynamics, and to analyze the individual role of different


74

inputs during NUPW and UPW seasons. The model outputs suggest limited impact of the nutrients

from outfall and from upwelling to the chemical and biological quality of the water in the SMCA

due to the large dilution of nutrient load. However, sediment analyses revealed the occurrence of a

pronounced organic impact, altering sediment biogeochemical dynamics and suggest keeping this

system continuously monitored and studied, via combination of experimental activities, satellite-

based observations and modeling approaches. This seems particularly important due to increasing

anthropogenic pressures on the coastal area and on watersheds and to ongoing global changes

affecting climate, wind intensity, water temperature and mixing rates.

Key words: Sewage outfall, upwelling, AEM3D, modeling, remote sensing, residence time

3.2. Introduction

The benthic-pelagic coupling has a crucial role in the dynamics of nutrient cycling, energy transfer

in food webs and ecosystem services in marine systems (Griffiths et al., 2017). Anthropogenic and

natural factors regulate benthic – pelagic coupling directly and indirectly through their effects on

physical, biogeochemical and biological components of coastal and estuarine ecosystems (Capone

& Hutchins, 2013; Griffiths et al., 2017). Anthropogenic nutrient loads from agricultural or densely

populated watersheds increase coastal primary production (either pelagic or benthic), sometimes

resulting in water quality decline, increase of entity and duration of algal blooms, loss of

biodiversity, changes in sediment organic load and biogeochemistry (e.g. suppression of

denitrification and P-retention and increase of dissimilative nitrate reduction to ammonium

(DNRA) and of P-mobility) (Diaz, 2001; Grall & Chauvaud, 2002; Reopanichkul et al., 2009;

Smith et al., 2012; Davidson et al., 2014; Arroyave Gómez et al., 2020). All these cascade

consequences of eutrophication may lead to dissolved oxygen (DO) depletion in the water column

and in sediments and in the buildup of sulphides in pore water, determining a positive feedback,

further increasing water chemical and biological quality deterioration (Diaz & Rosenberg, 2008).

Additionally, interacting with the effects of anthropogenic nutrient inputs are upwelling

phenomena (Arroyave-Gomez et al., 2020). Coastal upwelling regions display chemical and

biological features highly variable in time and space, experiencing large and sudden changes of

nutrients, Chl-a, temperature, salinity and DO (Capone & Hutchins, 2013; Howard et al., 2017).


75

These regions exhibit highly variable hydrodynamics, faster flushing and lower retention time than

other coastal and estuarine systems (Howard et al., 2017).

There is a limited understanding of how coastal and estuarine ecosystems respond to both

anthropogenic nutrient inputs and seasonal upwelling through their interactive effects on the

benthic–pelagic coupling and on water chemistry (Capone & Hutchins, 2013; Griffiths et al., 2017).

Paradoxically, such knowledge gaps are more pronounced in tropical areas where metabolic

activities are stimulated by elevated water temperature regimes but where nutrient processing and

the consequences of eutrophication are understudied (Downing et al., 1999; Miguel, 2018).

Quantification and prediction of material exchange between the benthic and pelagic systems and

their sensitivity to different drivers and characteristics (e.g. bathymetry, local hydrodynamics,

regional climatic fluctuations, seasonal upwelling intensity, external nutrient inputs, global climate

change, etc.) require the enforcement of empirical, experimental and modelling approaches

(Capone & Hutchins, 2013; Griffiths et al., 2017).

Coupled hydrodynamic-ecological models allow the exploration of complex feedback loops

between physical processes and biogeochemical cycles (Robson & Hamilton, 2004; Spillman et

al., 2007; Burger et al., 2008). However, applications of these models have frequently included

coarse representation of mineralization and regeneration of nutrients associated to bottom

sediments, as well as limited validation of nutrient fluxes at the sediment–water interface (Soetaert

et al., 2000; Paraska et al., 2014). Benthic–pelagic flux representations in ecological models are

logistically challenging due to scarce datasets related to the high intrinsic spatial and temporal

variability of benthic-pelagic coupling (Wilson et al., 2013; Capet et al., 2016). Despite these

limitations, the approach chosen to represent the remineralization from sediments (e.g. ranging

from simple parametrizations to vertically resolved dynamic sediment diagenesis models according

to Soetaert et al., 2000) should be defined based on the importance of the benthic- pelagic coupling

in the studied system (Wilson et al., 2013).

The Santa Marta Coastal Area (SMCA), located in the Colombian Caribbean Sea, is highly

influenced by seasonal upwelling (Andrade & Barton, 2005; Rueda-Roa & Muller-Karger, 2013)

and it is affected by several anthropogenic pressures (e.g. a sewage outfall, harbor activities, rivers

discharge). Seasonal coastal upwelling causes a high temporal variability in water temperature,

nutrient and Chl-a concentrations and water circulation patterns (Arévalo-Martínez & Franco -

Herrera, 2008; Ramírez-Barón et al., 2010). These factors, along with anthropogenic nutrient inputs

are the major drivers controlling biological diversity and key biogeochemical and ecological


76

services in this area (e.g. the presence of corals and fishes as well as the denitrification, P-retention

or sulphide buffer capacity of sediments, etc.) (Bayraktarov et al., 2013; Bayraktarov & Wild,

2014; Arroyave Gómez et al., 2020). A recent experimental study targeting sediment

biogeochemistry along a transect from impacted to not impacted areas showed the complexity and

local variability of benthic nitrogen cycling. In the same study, it was demonstrated that sediments

trace, via their organic matter and δ13C signature, the extended impact the sewage outfall produces

interacting with the seasonal increase of nutrients from upwelling (Arroyave Gómez et al., 2020).

The SMCA thereby provides an excellent, complex and challenging scenario to analyze the

interactive effects of seasonal upwelling and anthropogenic nutrient inputs on benthic–pelagic

coupling and coastal biogeochemistry. With this respect modelling tools, if properly calibrated,

may be able to capture the net effects of so many variables on water chemical and biological quality

(Spillman et al., 2007; Machado & Imberger, 2012). Using measurements of benthic metabolism

in a coupled hydrodynamic–ecological mathematical model may provide important insights into

biogeochemical cycles and management of a tropical coastal area affected by anthropogenic

nutrient inputs and seasonal upwelling. Despite the highly variable nutrient dynamics and transport

processes in SMCA, and the ecological and recreational values of this coastal area, no water

column biogeochemical modeling study has been carried out to date.

In this study the coupled 3D hydrodynamic-ecological model AEM3D was applied to analyze the

seasonal variations of water physicochemical and biological parameters in SMAC and to provide

insights into the benthic-pelagic coupling, under two different nutrient and organic matter loads

from untreated wastewater outfall (average and maximum flow-rate) and along the NUPW and

UPW seasons. The model was set up, calibrated and validated by (i) benthic metabolic

measurements dataset; (ii) satellite–derived Chl-a and SST maps; (iii) temperature and salinity data

from the HYCOM database and, (iv) field and published literature meteorological and water quality

data. I hypothesized a limited effect of sewage outfall on the water chemical and biological quality

in the SMCA due to its high flushing rates and short water renewal time, resulting in a large dilution

of nutrients and organic matter and limited increase of their concentrations along the water column.

However, phytoplankton growth may show transient increase resulting in Chl-a peaks, due to the

elevated capacity of algal cells to assimilate nutrients in a diluted medium and to respond to an

even small increase of their availability. I remark, for example, that experimental activities

targeting the determination of the impact of the outfall on the water microbiological quality

revealed a very small area of impact, much smaller than that evidenced by sediment analyses. Even


77

small changes in the concentration or ecological stoichiometry of nutrients may increase pelagic

primary production or change the community composition of phytoplankton, with both changes

very difficult to highlight in the open sea via traditional sampling and measurement activities

(Dalsgaard & Krause-Jensen, 2006). I also remark that biogeochemical modeling of the water

column may allow testing interactive effects of increased loads (e.g. outfall, rivers, etc.) and

upwelling events along ongoing global changes, potentially affecting water temperatures,

stratification and mixing and with them oxygen concentrations, determining strong impacts in

SMCA.

3.3. Materials and methods

3.3.1. Study area

The Santa Marta Coastal Area, located in the Colombian Caribbean, includes Gaira, Santa Marta

and Taganga bays and the Tayrona National Natural Park (TNNP) (Diaz et al., 2000; Vega-Sequeda

et al., 2008). The SMCA has an open conformation, subject to oceanic and continental influences,

impacted by multiple stressors including a submarine wastewater outfall, harbor activities, the

Manzanares and Gaira river discharges and by seasonal surface runoff (Escobar, 1988; Mancera-

Pineda et al., 2013; Ramos-Ortega et al., 2008) (Fig. 3.1). River discharge and surface runoff are

low or null during the dry seasons (December- April and July-August) and peak during the rainy

seasons (May-June and September-November) (Arévalo-Martínez & Franco - Herrera, 2008).

The Santa Marta sewage outfall (SMSO, Fig. 3.1) discharges on average 1 m3s-1 of untreated

wastewater derived from ~500,000 inhabitants (García, 2013). However, as the touristic pressure

in the areas is expected to increase, the SMSO has an installed capacity of 2.5 m3s-1 (García, 2013)

and will deliver to the bay larger amounts of organic matter and nutrients. At present the Santa

Marta city nearly doubles its resident population during the high tourist season, peaking in January

and resulting in increased wastewater discharge (Díaz-Rocca & Causado-Rodríguez, 2007;

Moscarella et al., 2011). The SMSO outlet is located between Santa Marta and Taganga bays (11.26

Lat and -74.22 Lon) (García, 2013). The sewage outfall has a 1 m diameter tubing extending 428

m from the coastline. The wastewater is released through 31 alternating diffusers installed both

sides in the last 120 m of the tubing, at 56 m depth (Moscarella et al., 2011). The sewage only

receives a preliminary treatment for the removal (90%) of large solids (Díaz-Rocca & Causado-

Rodríguez, 2007). During 2006, the loadings of total nitrogen (N), phosphorus (P) and suspended


78

solids (TSS) released to the bay from the outfall were estimated in ~1,100 T N·yr-1, ~500 T P·yr-1

and ~12,300 T TSS·yr-1, respectively (García, 2013).

The regional climate and oceanographic dynamics of SMCA are strongly determined by the

strength of the Caribbean Low – level Jet of the Northeast (NE) Trade Winds and North (N) –

South (S) oscillations of the Intertropical Convergence Zone (ITCZ) (Andrade & Barton, 2005).

Parallel NE winds to the coastline trigger an upwelling of subsurface waters due to the Southern

Caribbean upwelling system (Andrade & Barton, 2005; Arévalo-Martínez & Franco - Herrera,

2008; Paramo et al., 2011; Rueda-Roa & Muller-Karger, 2013). Major (December- April) and

minor (July-August) UPW events occur during the dry season with strong NE trade winds (mean

3.5 m·s-1) while minor (May-June) and major (September-November) NUPW periods occur during

the rainy season with low wind incidence (mean 1.5 m·s-1) (Fajardo, 1979; Andrade & Barton,

2005; Arévalo-Martínez & Franco - Herrera, 2008). Seasonal upwelling leads to changes in

physicochemical variables such as temperature decrease (from 30 °C to 21 °C), salinity increase

(from 33 to 38), oxygen under-saturation conditions (< 91% ) (Ramírez, 1981; Salzwedel & Müller,

1983; Bayraktarov et al., 2014) and increase of nitrate and Chl-a concentrations, turning the system

from oligotrophic to mesotrophic (Arévalo-Martínez & Franco - Herrera, 2008; García-Hoyos et

al., 2010; Paramo et al., 2011). Surface currents also modify circulation patterns contributing to

high spatial and temporal variability of nutrient concentrations and phytoplankton abundance. The

Caribbean Current, flowing west –northwest and strengthened by the NE trade winds, influences

the UPW during the dry season while the Panama–Colombia Countercurrent (SE - NE), reinforced

by the SW winds during the rainy season, causes inputs of fresh and brackish waters from Ciénaga

Grande de Santa Marta (CGSM) which in turn receives mixed water inputs from the Magdalena

River, the Caribbean Ocean and from rivers draining the western margin of Sierra Nevada de Santa

Marta (SNSM) (Ramírez, 1981; Andrade, 2003; Franco-Herrera et al., 2006; García-Hoyos et al.,

2010; Mancera-Pineda et al., 2013).


79

Fig. 3.1 Domain location and sampling stations. Panel (a): The star and the dots indicate the

location of the SMSO and of the water column sampling stations in SMCA. HYM1 and HYM2 are

the Hybrid Coordinate Ocean Model (HYCOM) stations used to set temperature and salinity

boundary conditions for the AEM3D model. Panel (b): The study area in the Colombian Caribbean.

Panel (c):Transect of 1800m from the SMSO both towards Taganga Bay and Santa Marta Bay in

which it was analyzed the increase of wastewater flow-rate with the model.

3.3.2. Model description

The Aquatic Ecosystem 3D Model (AEM3D) used in this study was developed by HydroNumerics

(Hodges & Dallimore, 2016) and is based on the ELCOM – CAEDYM model developed earlier

by the Center for Water Research (CWR), the University of Western Australia (Hodges, 2000;

Hipsey et al., 2012). ELCOM, a high-resolution 3D hydrodynamic model, uses a fixed grid

structure to solve the unsteady Reynolds-averaged Navier-Stokes equations, subject to boundary


80

forcings and hydrostatic approximations (Hodges & Dallimore, 2016). CAEDYM consists of a

series of process-based partial differential equations that dynamically simulate concentrations of

biogeochemical variables (e.g. phytoplankton, particulate and dissolved organic matter, dissolved

inorganic matter, etc.) accounting for primary and secondary production, nutrient cycling (C, N, P

and Si), oxygen dynamics, and sediment-water interactions (Spillman et al., 2007). Fluxes of

dissolved inorganic and organic nutrients from the sediments are dependent on temperature and

DO concentration of the water layer immediately above the sediment surface (Burger et al., 2008;

Özkundakci et al., 2011). ELCOM – CAEDYM has been successfully applied to study transport

processes, biogeochemical cycles and environmental management in several lakes, estuaries and

coastal oceans (Robson & Hamilton, 2004; Romero et al., 2004; Spillman et al., 2007; Machado &

Imberger, 2012;). This model has been previously described in detail by Robson & Hamilton

(2004) and Spillman et al. (2007).

3.3.3. Water quality field monitoring for SMCA

The water column sampling was carried out at 10 stations for physicochemical analysis (Fig. 3.1a),

with stations P1 and P10 located 5.2 Km northwest and 9.5 Km southwest from the SMSO diffuser,

and representing offshore and CGSM water physicochemical features, respectively. Stations P2,

P3, P4 and P5 were located in Taganga Bay, with station P3 located in the proximity of the diffusers

and stations P2 and P4 located 2 Km north and 1 Km northeast from diffusers, respectively. Station

P5 was located 0.6 Km southwest from the diffusers, station P6 was located near the harbor, station

P7 represented the central part of the Santa Marta bay, station P8 was close to the Manzanares

River discharge and station P9 was located between Gaira and Santa Marta bays.

Water quality samples were collected during four field campaigns on August 21th 2017 (minor

UPW) and November 26th, 2017 (major NUPW) and on January 26th and February 4th, 2018 (major

UPW). The first campaign was affected by the tropical storm Harvey which crossed the northern

Colombian Caribbean Sea the day before, increasing precipitation events in the area. Due to

heterogeneous topography, water samples were taken with a 5 L Niskin bottle at the surface, at 20

m and at 40 m depth in nearshore stations. At station S1, located offshore, water samples were

collected at the surface, 20 m, 60 m and 120 m depth. DO was measured in situ using a multi-probe

sensor Orion Star TM A326 whereas water temperature and salinity profiles were measured by a

CTD (Conductivity – Temperature – Depth). Unfiltered subsamples were used for analysis of total


81

phosphorous (TP) whereas filtered subsamples (Whatman, GF/F) were analyzed for ammonium

(NH4+), NOx

- (nitrate + nitrite), phosphate (PO43-) and total dissolved phosphorus (TDP).

Subsamples for Chl-a were filtered (Whatman GF/C glass fiber filters) and filters were stored

together with nutrient samples at -20 °C for later analysis. Ammonium was determined by

indophenol method, NO2- by sulfanilamide, NO3

- by cadmium reduction, PO43- by ascorbic acid

method, TP and TDP by ascorbic acid method after digestion with perchloric acid according to

APHA (2012). The Chl-a was extracted with 90% acetone (Strickland & Parsons, 1972). This

dataset was used to set the open boundary conditions for nutrients, and also for the model

calibration and validation.

3.3.4. The AEM3D Hydrodynamic model set up for SMCA.

The simulation domain extended roughly 19 Km along the coast (N-S) and 15 km cross-shore in

the N and 5 Km in the S, and from the coastline to near the 450 m isobath (Fig. 3.1a). This wide

physical domain was determined based on the approach proposed by Hillmer & Imberger (2007)

to capture the variability of dominant ecological processes from the advective flow through open

border in a coastal ecosystem. The domain was defined as a compromise in terms of running time

(computational effort) and the limited area for a coastal region due to the steep bottom gradient

which descends rapidly to depths of more than 500 m at relatively short distances from the coastline

(Fig. 3.1a) (García-Hoyos et al., 2010).

The bathymetry was obtained from nautical chart digitization from the Oceanographic and

Hydrographic Research Center (CIOH acronym in Spanish) from Colombia (www.cioh.org.co).

Horizontal and vertical grid configurations were set based on a representation of the SMSO

wastewater discharge as a rising buoyant plume (García, 2013), condition achieved by forcing the

flow-rate as a mean volume flux time series calculated through the plane plume equation (Fischer

et al., 1979) for the diffuser characteristics (length, depth and flow-rate) as detailed by Machado &

Imberger (2012). A mean volume flux of 230 m3 s-1 was estimated from the mean flow – rate (1.0

± 0.16 m3·s-1) for 2006 (García, 2013) and mean depth of 43 m (the diffusers are located between

30 and 56 m depth). This mean volume flux is equivalent to a mean initial dilution at the water

column surface by a factor of 230 which is in agreement to the range of dilutions reported by García

(2013). In order to estimate the potential impact of the nutrient load increase due to the raise in

wastewater discharge up to 2.5 m3 s-1, a volume flux of 315 m3 s-1 was calculated. After testing


82

several configurations for horizontal and vertical grid sizes, the outfall discharge was simulated as

an inflow through the bottom face of 3 cells of 60 x 60 x 1.5 m located in E-W direction and geo-

referenced for the diffuser, allowing to reproduce the buoyant plume and correspondent

gravitational flow (Machado & Imberger, 2012). The conformation of diffuser cells (volume 16200

m3) and a 45 s computational time step allowed to vary the volume fluxes from 230 m3 s-1 (10350

m3 per time step) to 315 m3 s-1 (14175 m3 per time step), satisfying both, the Courant – Friedrichs

- Lewy (CFL) stability condition and the fact that the volume of injected effluent in each step size

will not exceed the volume of receiver’s cells. From the configuration of the diffuser cells, a non-

uniform horizontal grid was adjusted to increase both the spatial resolution (near to outfall and

shoreline) and computational efficiency (Spillman et al., 2007; Machado & Imberger, 2012). The

diffusers were located in cells of 60 m x 60 m, which increased in size steadily at 7% until reaching

490 m in the western, 600 m in the northern and 642 m in the southern borders. The vertical grid

was set to 51 variable thickness layers ranging from 0.5 m near the surface and progressively

increasing up to 1.5 m for diffuser cells (14 cells between 30 and 56 m depths) and up to a maximum

of 41 m at the deepest point in the domain (around 447 m depth). Thin vertical top layers provide

high resolution to capture the buoyant plume spreading on the surface and seasonal stratification

of the water column (Machado & Imberger, 2012). The resulting grid consisted of 81 x 69 x 51

cells with 115240 maximum wet cells. The north, south and west offshore boundaries of the domain

were modeled as open boundaries. These boundaries were forced by data of Sea Surface Height

(SSH) and temperature and salinity vertical profiles from the GOFS 3.1, 41-layer

HYCOM+NCODA Global 1/12° re-analysis. The northern and western boundaries were forced

with data from HYM1 while the southern boundary was forced with data from HYM2 (Fig. 3.1a).

The time and space resolutions of GOFS 3.1 are 3 hours and 0.08°, respectively. The HYCOM re-

analysis data have been intensively used to provide boundary conditions to regional and coastal

models (Chassignet et al., 2007). Within the domain, temperature and salinity vertical profiles were

initialized with station P1 averaged CTD data. Meteorological forcing was applied uniformly on

the entire domain free surface. Wind direction and speed, solar radiation, air temperature, relative

humidity, atmospheric pressure and precipitation were obtained, in hourly intervals, from a weather

station located approximately 3 Km east-northeast of outfall at Taganga Bay (Fig 3.1a).

Water age simulation was carried out to estimate the fluid residence time to understand the

exchange and fate of nutrients and Chl-a with open waters (Machado & Imberger, 2012).


83

3.3.5. AEM3D ecological model set up for SMCA.

The ecological model was configured to simulate the C, N, P, DO and Si biogeochemical

transformations with diatoms as the dominant phytoplankton group in the SMCA (>90%, e.g.

Chaetoceros ssp. and Skeletonema ssp.) ( Ramírez-Barón et al., 2010; Garcés-Ordóñez et al., 2016).

Three forms of dissolved inorganic nutrients (NOx-, NH4+, PO4

-), dissolved organic matter

(dissolved organic carbon (DOC), dissolved organic nitrogen (DON), dissolved organic

phosphorus (DOP)) and particulate organic matter (particulate organic carbon (POC), particulate

organic nitrogen (PON), particulate organic phosphorus (POP)) were modeled. The August and

November average profiles were adopted as initial conditions for the nutrients. Mean vertical

concentration profiles of NOx-, NH4+, PO4

-, DOP and DO from P1 were used as initial conditions

and open boundaries conditions along the northern and western boundaries while mean vertical

nutrient profiles from P10 were applied as open boundary conditions in the south. The inputs of

Manzanares and Gaira rivers were not directly simulated for the model calibration and validation

because these discharges vary seasonally and the mean precipitation from November 2017 to

February 2018 was very low (~18 mm). Mean vertical nutrient profiles in P8 were used as initial

condition profiles as background chemical conditions of the Manzanares River whereas data from

P6 were used to set the background chemical conditions of the harbor and the surface runoff from

main streets of the Santa Marta city. POP was assumed as the difference between TP and PO4-3.

POC and PON were set to a single domain – wide value of 8 µM C and 1.2 µM N, respectively,

for the upwelling region based on literature reports (Martiny et al., 2013) while vertical profiles

were used as initial conditions for DOC, DON and Si were fixed from mean values of southeastern

Caribbean Sea and SMCA (Lorenzoni et al., 2013; Mancera-Pineda et al., 2013). The average

nutrient concentrations in open borders were interpolated over time based on field data and

previous studies in the area (García-Hoyos et al., 2010; Mancera-Pineda et al., 2013). Due to the

high spatial and seasonal variability of Chl-a in the SMAC (Franco-Herrera et al., 2006; Mancera-

Pineda et al., 2013), MODIS Chl-a remote-sensing maps (three maps per month for the simulated

period from November 2017 to February 2018) were used to set the horizontal variation through

an average initial horizontal file of Chl-a. Every MODIS Chl-a map was adjusted to the domain

size and interpolated to the same non-uniform horizontal grid of the hydrodynamic model. These

grids were averaged to obtain the initial horizontal Chl-a 2D file. This file integrated the

background of all pollution sources. Then, monthly Chl-a averages over the domain’s three


84

boundaries were interpolated in time to force the model. Effluent quality of SMSO was estimated

from average measurements for 2006 (García, 2013) and implemented with literature values for

untreated wastewater (Table S.1) (Metcalf & Eddy, 2003; Henze & Comeau, 2008). The effluent

quality and volume fluxes were assumed constant during the simulations. The discharge input from

bottom cells depicting the diffuser was imposed diluted by the volume flux.

In this study diatoms, as the dominant algal group, were modelled with a focus on the N cycle

because it was found that temperature and nitrate concentration significantly influence the

abundance of the phytoplankton community in UPW and NUPW seasons in SMCA (Arévalo-

Martínez & Franco - Herrera, 2008; Ramírez-Barón et al., 2010). Nutrient dynamics were

simulated along with phytoplankton basic processes as growth (with potential limitation by light,

nutrients and temperature), settling, resuspension and a loss lumped term (accounting for

respiration, excretion, mortality and grazing) (Hillmer & Imberger, 2007; Machado & Imberger,

2012). The N cycle included processes of nitrification and denitrification. Heterotrophic bacteria

activity was not modeled explicitly due to scarcity of data and lack of parameter data. However,

the nutrient pathways catalyzed by bacteria were included in the mineralization of the particulate

organic pools (POC, POP and PON) (Bruce et al., 2006). Particulate detritus settling (according to

the Stoke’s law), resuspension and light attenuation was simulated (Machado & Imberger, 2012).

Dissolved nutrient exchanges at the sediment-water interface were simulated through their changes

as a function of temperature and DO ( Robson & Hamilton, 2004; Spillman et al., 2007; Burger et

al., 2008). Sediment PO4-, NH4, NO3

- and Si release rates, sediment oxygen demand (SOD) and

sediment composition, inputs to CAEDYM, were based on a study of spatial and temporal

variability of sedimentary properties and benthic metabolism along a transect from the SMSO to

outer, non-polluted sediments within the modeling period (Arroyave Gómez et al., 2020). The

measurements also included the quantification of N2 Fixation, denitrification (Dw, Dn and Dt) and

DNRA. The range of benthic nutrient fluxes considered the increase of organic matter and the ratio

between marine and terrestrial C inputs, measured with δ13C due to both seasonal upwelling and

the increase in wastewater discharge produced by the tourist peak season (December 2017 –

January 2018). Maximum PO4-, NH4, NO3

- and Si sediment release rates were specified in the

model. SOD rates near to SMSO were highly variable, both in space and time, ranging from 0.78

to 4.85 g DO m-2d-1 (Arroyave Gómez et al., 2020). Similar to the Chl-a, the SOD horizontal

variation was included in the model through a 2D-file. Average SOD values were geo-referenced

for 4 stations located in the proximity (2.95 ± 0.36 g DO m-2d-1), 100 m (1.32 ± 0.07 g DO m-2d-1),


85

750 m (1.26 ± 0.10 g DO m-2d-1) and 1800 m (0.89 ± 0.09 g DO m-2d-1) far from the outfall in the

same hydrodynamic model non-uniform horizontal grid. An average value from the three stations

away from the outfall was used for depths less than 50 m (1.16 g DO m-2d-1), depths between 50

and 100 m (0.58 g DO m-2d-1) and for depths higher than 100 m (0.29 g DO m-2d-1). This space

differentiation of SOD allows representing the variation of a higher SOD in sediments nearshore

due to a higher organic carbon availability. The geochemical module of the model was not enabled

for the simulations because the benthic fluxes were uncoupled to oxygen uptake and no redox –

dependence in near and far stations from the outfall. Furthermore, the absence of measurable Fe2+

and Mn2+ fluxes was observed (Arroyave Gómez et al., 2020).

3.3.6. Calibration and Validation of AEM3D model for SMCA

The period November 2017 – February 2018 was chosen for the simulations, based on the

availability of the several data sets required for the models (e.g. meteorological, water quality, and

benthic metabolism data). The model sensitivity analysis and calibration were carried out for the

major NUPW season (November 2017) while model validation was carried out for the major UPW

season (December 2017 – February 2018). AEM3D model was calibrated against field data for

November 26th for CTD temperature and salinity vertical profiles at stations from P2 to P10 while

DO, NOx-, NH4+, PO4

-, TP and Chl-a were calibrated against field vertical profiles at P2, P3, P4,

P5, P7 and P9. Profiles at these stations were not used to force the model. The calibration was

carried out by comparing field campaign profiles with the model outputs in a time interval that

includes all model profiles generated from the day before to the measurement day. The surface

wind drag coefficient and horizontal diffusion coefficient were calibrated for the hydrodynamic

model while algal respiration mortality, excretion and grazing (Kr), nitrification rate (Knit),

mineralization maximum rates of DON (μMINDON) and DOC (μMINDOC) and DOP (μMINPON) were

calibrated for the ecological model. Other internal parameter values of the AEM3D model were

derived from several sources: field data and experimental analysis (benthic metabolism), site-

specific parameters and literature values.

The model was validated for field vertical profiles of January 26th and February 4th without

changing hydrodynamic and ecological parameters determined during the calibration. For January

26th the hydrodynamic model was validated with profiles of P3 and from P4 to P10 while the


86

ecological model was validated with profiles P3, P5, P7 and P9. The model was validated on

February 4th in the same stations considered for the calibration.

Remote sensing maps of SST and Chl-a were used for the model validation. SST maps were

obtained from the Level-2 products of the Moderate Resolution Imaging Spectroradiometer

(MODIS) sensor, on-board the AQUA satellite of National Aeronautics and Space Administration

(NASA) (1-km nominal resolution for the thermal bands). This satellite has sun-synchronous and

near-polar orbits and acquires at daytime, providing complete global coverage of the Earth every

1 to 2 days, since its launch in 2002 (NASA, 2020). The Level-2 SST data products, in Celsius

degrees (°C), are generated from Level-1 radiance data measured at two infrared bands of the

MODIS sensor, centered around 11 μm and 12 μm, by applying sensor calibration and atmospheric

corrections (Minnett et al., 2002). Level-2 data products have been used in several marine studies

(Delgado et al., 2014; Kozlov et al., 2014; Ghanea et al., 2016; García, 2020). All Level-2 AQUA

SST products available for the analysis period were downloaded from the NASA ocean color group

website (http://oceancolor.gsfc.nasa.gov/). All the images were subsetted, masked with clouds and

quality flags distributed within the Level-2 products, and reprojected in the UTM cartographic

projection.

The remote sensed Chl-a maps were obtained from Level-1 EFR images acquired by the Ocean

and Land Colour Instrument (OLCI) sensor (300 m spatial resolution), on-board Sentinel-3A

satellite of the European Space Agency (ESA). The images were processed with the Case 2

Regional Coast Colour processor (C2RCC), a neural net developed for optically complex (Case 2)

waters (Brockmann et al., 2016), widely used by the scientific community of water remote sensing

(Toming et al., 2017; Ruescas et al., 2018; Ogashawara, 2019). Regarding the C2RCC

parametrization and due to the width of the processing area it was implemented the default option

to adopt the European Centre for Medium-Range Weather Forecasts (EMCWF) data as the source

of meteorological data of air temperature and atmospheric pressure. EMCWF data are distributed

within the OLCI L1-EFR products. Water salinity was set at default values of 35 PSU for all

images. Also, the inherent optical water properties were left by default. The Chl-a maps obtained

by C2RCC were masked for clouds and perturbations using the masks produced by C2RCC:

Cloud_risk, Rtosa_OOS, Rtosa_OOR, Rhow_OOS and Rhow_OOR, the latest indicating the input

spectrum to the atmospheric correction neural net or the input spectrum to derive the WQPs were

not within the training range of the respective neural nets or outside the range of the expected

results. All the processing was carried out with free and open source software: the European Space


87

Agency’s SNAP (Zuhlke et al., 2015) toolbox, the free software for satellite images processing and

visualization distributed by ESA, and QGIS. Finally, Chl-a satellite maps used during calibration

were different from those adopted for the model validation.

3.3.7. Statistical treatment and assessment of AEM3D model performance

Kruskal – Wallis non-parametric test was used to assess significant differences for the water quality

variables monitored in the field campaigns for the factors season, station and depth. For significant

factors, post hoc pairwise comparisons were performed using Dunn’s test for Kruskal– Wallis. The

statistical test was carried out by the Rcmdr package of the R- Project for Statistical Computing (R

version 3.5.1) (Fox, 2005).

Model performance over the calibration and validation period was assessed for each output variable

based on plot visual inspections of simulated against observed data along with the calculation of

the normalized mean absolute error (NMAE), as defined by (Alewell & Manderscheid, 1998):

�� = ∑ |��|��

!∙�" (3.1)

where Pi is the simulated variable, Oi is the observed variable, Ō is the average observed value and

n is the number of observations. A NMAE value of zero indicates a perfect agreement between

predicted and measured data (Alewell & Manderscheid, 1998).

3.3.8. Analyzed Scenarios using the Model AEM3D

The calibration and validation baseline scenario was compared with a scenario in which the

wastewater flow-rate increased to 2.5 m3 s-1 (corresponding to a volume flux of 315 m3 s-1), keeping

the outfall average composition constant (Table 3S.1). Simulations without including the outfall

were also carried out to set a reference case.

In order to estimate the increase of signature of total organic carbon (TOC) with the incremented

wastewater loading, the modelled average surface concentrations of TOC were quantified through

an 1800 m transect from the outfall both towards Taganga Bay and Santa Marta Bay (Fig 3.1c).


88

This transect overlapped that where benthic metabolism measurements were carried out in the

Taganga Bay within the simulation period.

3.4. Results

3.4.1. Water quality monitoring for SMCA

Nutrient and Chl-a concentrations displayed large variability among stations, seasons and depths

(Fig. 3.2). Significant differences between NUPW and UPW seasons were found for NOx, NH4+,

TP, Chl-a and DO (Kruskal-Wallis test, p<0.05). The pairwise comparisons between seasons

during the simulation period (November 2017 (NUPW) – February 2018 (UPW)) also showed a

significant difference for these variables (Dunn test, p < 0.05) (Table 3.1). NH4+, PO4

3-, TP, Chl-a

and DO concentrations decreased while NOx increased between NUPW and UPW seasons. NH4+

and PO4- concentrations below detection limits were around 46 and 38 %, respectively, and NH4

+

concentrations were not detectable in the monitored stations during the field campaign of January

26th (Fig 3.2). The Chl-a had a wide range of variation (0.07 – 7.2 µg Chl-a·L-1). Significant

differences between stations were only found for TP between stations P5-P7, P5-P9 and P4-P7

(Kruskal-Wallis test, p<0.05 and Dunn test, p < 0.05). Stations P4 and P5 were close to the area

influenced by the discharge and transport of outfall effluents. Chl-a and DO had significant

differences along the water depths (Table 3.1).

3.4.2. Calibration and validation of AEM3D model for SMCA

Comparisons of modelled temperature, vertical and surface profiles, during calibration and

validation periods against the field data and SST satellite images are reported in Fig. 3.3 and Fig.

3.4, respectively. They show a high correspondence with NMAE values in a range of 0.01 - 0.02

(Fig. 3.4 and Table 3.S2), indicating that the model reproduced reliable temperature values fitting

spatially and temporally variable water temperatures during NUPW and UPW seasons. Spatial

gradients of surface temperature computed by the model were slightly underestimated in both

seasons, more in the southern than in the northern area (Fig. 3.4). A temperature gradient between

the southern-northern open boundaries was evident due to the inputs of warmer waters from CGSM

in the southern boundary. However, this gradient was less pronounced in the UPW season as


89

compared to the NUPW season due to the water inflow in northeast - southeast direction along the

northern border by the NNE winds.

Fig. 3.2 Water quality monitoring field data for SMCA – Colombian Caribbean at 10 stations

during NUPW (26th November – 2017) and UPW (21th August 2017, 26th January 2018 and 4th

February) seasons. Average ± standard deviations are reported. Panels on the left report pooled

data from all stations and depths in the different seasons while panels on the right report data

relative to the different stations from all seasons and depths.

.


90

Table 3.1 Results of Kruskal-Wallis rank sum test to analyze the effect of seasons, stations and

depths on nutrients, Chl-a and DO in the water column of SMCA.

Parameter

Season Station Depth χ2 Df P Dunn

Test χ2 Df P Dunn

Test χ2 Df P Dunn

Test NOx (mg N·l-1)

(n = 82) 68.94 3 <0.001 Aug. - Feb.

Aug. - Jan. Aug. - Nov. Nov. - Jan. Nov. - Feb.

12.24 9 0.20 2.77 4 0.60 -

NH4 (mg N· l-1) (n = 53)

20.36 2 <0.001 Aug. - Feb. Nov. - Feb.

9.24 9 0.42 8.26 4 0.08 -

PO4 (mg P· l-1) (n = 34)

2.78 3 0.43 - 13.38 9 0.15 1.82 2 0.40 -

TP (mg P·l-1) (n = 63)

10.19 3 0.02 Aug. - Feb. Nov. - Jan. Nov. - Feb.

18.43 9 0.03 P1-P7, P4-P7, P5-P7, P1-P9, P5-P9

7.32 4 0.12 -

Chl-a ( µg/L) (n = 83)

14.49 3 0.002 Aug. - Feb. Aug. - Jan.

8.56 9 0.48 30.84 4 <0.001 0-20, 0-40

DO (mg O2· l-1) (n= 127)

78.69 3 <0.001 Aug. - Feb. Aug. - Jan. Nov.- Feb. Nov.- Jan.

8.98 9 0.44 42.14 7 <0.001 0-10, 0-20, 0-30, 0-40,

0-50

χ2: Chi-squared. Kruskal-Wallis test significance level (p<0.05). Significant differences in Dunn test (P <0.05).

The model captured the presence of two water masses with different ranges of temperature and

salinity based on temperature – salinity plots from field and simulated data (Fig 3.5 a and Fig 3.5

b). High temperatures and low salinities of the water column were observed on November 26 while

low temperatures with slightly higher salinities were found on February 4th. Modeled temperatures

and salinities on November 26th (NUPW) varied between 28 - 30 °C and 34 - 36, respectively,

whereas temperatures and salinities on February 4th (UPW) ranged between 20.5 – 23.5 °C and

36.2 –36.7, respectively.

The development of water column stratification and mixing periods were evident both in the

measured and simulated temperature profiles for the NUPW and UPW seasons, respectively (Fig.

3.3 and Fig. 3.6). The formation of thermoclines during the NUPW season were observed in the

range of 2 to 15 m in the simulated profiles at stations P4 and P9 (Fig. 3.6 a and Fig. 3.6 c), and

were in agreement with the field data (Fig. 3.3 a), indicating their presence at depths between 3 and

12 m. Both field (Fig. 3.3 b and Fig. 3.3 c) and modeled data (Fig. 3.6 b and Fig. 3.6 d) show how

temperature gradients decreased and the thermocline was absent in the UPW season.


91

Fig. 3.3 Comparison between simulated and measured temperature profiles during the calibration

(November 2017) and validation (January and February 2018) periods. Field data correspond to

CTD measurements during: a) 26th November 2017, within the NUPW season; b) 26thJanuary

2018, within the UPW season; c) 4th February 2018 within the UPW season. AEM3D model

simulation results include the range of all the simulated profiles generated from the day before to

the measurement day. CTD data were not measured at P2 and P4 in January.


92

NMAE=0.01

(a) (b)

NMAE=0.01

(c) (d)

NMAE=0.02

(e) (f)

Fig. 3.4 Sea surface temperature (SST) satellite images from MODIS during NUPW and UPW

seasons for (a) 26th November 2017 (NUPW), (c) 14th December 2017 (UPW) and (e) 15th January

2018 (UPW) and modelled SST at noon for SMCA on (b) 26th November 2017 (NUPW), (d) 14th

December 2017 (UPW) and (f) 15th January 2018 (UPW). NMAE between simulated results and

satellite images is also reported.


93

Fig. 3.5 Temperature – Salinity plots of stations located in Taganga (P2 and P4) and Santa Marta

(P7) bays in SMCA during 26th November 2017 (NUPW) and 4th February 2018 (UPW) for (a)

sampling data and (b) AEM3D model results.

(a) (b)

(c) (d) Fig. 3.6 Modelled temperature vertical profiles at different times in SMAC: (a) Calibration period

(November 2017 - NUPW) and (b) Validation period (December 2017 – February 2018 - UPW) at

station P4; (c) Calibration period (November 2017 –NUPW) and (d) Validation period (December

2017 –February 2018 - UPW) at station P9.


94

The parameters adjusted for AEM3D ecological model during the calibration period are shown in

Table 3.2. The model was generally able to reproduce the magnitude and dynamics of field

measurements of nutrients and Chl-a during the calibration (NUPW) and validation (UPW) periods

based on adjusted vertical profiles (Fig. 3S.1, Fig. 3S.2 and Fig. 3S.3), Chl-a remote sensing data

(Fig. 3.7) and NMAE (Table 3S.2). Both field data and satellite images indicate the high spatial

and temporal variability of Chl-a (Fig. 3.2 and Fig. 3.7) as well as a gradient between the southern-

northern open border. Like temperature, this gradient was progressively reduced during the

upwelling season. High Chl-a concentrations were also observed along the offshore border (Fig.

3.7 e).

The satellite images showed several Chl-a patches in the SMCA likely due to the CGMS water

inputs, wastewater outfall, rivers, surface runoff, seasonal upwelling and plankton productivity in

the reef community (Fig. 3.7). The influence of CGSM waters on the SMCA was evident in satellite

images retrieved in November 2017 and January 2018, showing an increase of Chl-a up to 9.6 µg

Chl-a L-1 on November 26th over the southern open border and an increment of Chl-a up to 1.5 µg

Chl-a L-1 on January 30th along the western open border (Fig. 3.7 a and Fig. 3.7 e). Chl-a in January

likely increased due to the combined effects of both tourism and upwelling season. The Chl-a

increase nearshore on January 3rd in areas with nearby rivers, harbor, surface runoff and outfall was

possibly due to the increase of tourists in SMCA (Fig. 3.7 c). The increase of Chl-a along the north

open border observed on January 30th was likely due to seasonal upwelling caused by nutrient-rich

water inputs in direction NE-SW (Fig 3.7 e). A Chl-a patch with concentrations between 2.0 and

2.4 µg Chl-a L-1 near the outfall discharge area was observed without the apparent influence of

other pollution sources or the upwelling on November 26th (Fig. 3.7 a). This effect extended around

1.5 km towards Taganga Bay and 1.0 km towards Santa Marta Bay and to the north. A Chl-a patch

at a distance of 1.5 km north from the outfall was also evident. Transport and dispersion processes

of the outfall plume and other pollution sources might have stimulated phytoplankton growth and

caused the patch formation. A Chl-a patch with high concentrations, up to 7.2 µg Chl-a L-1, was

also evident on November 26th in the north, in TNNP where reef communities are present (Fig. 3.7

a). Despite the high variability of Chl-a, the model better captured the magnitude and timing of

Chl-a in the northern (Taganga Bay and TNNP) than in the southern sector of the analyzed area.

Furthermore, modeled Chl-a had a slightly better fit to field data and satellite images in the

upwelling period than in the non-upwelling period as shown to the lower NMAE values (Fig. 3.7

and Table 3.S2).


95

The simulated and field concentrations of NOx and DO had a good fit both for the calibration and

validation periods, with NMAE values less than 0.10. TP had a better fit for depths of 20 and 40 m

than for surface concentrations (Table 3.S2). Despite the limited data on PO43- for the calibration,

this parameter had a good fit with the field data on February 4th (Fig. 3.S3). The simulated NH4+

had a lower vertical adjustment with the field data than the other nutrients. The high NH4+

concentrations on November 26th (NUPW) coincided with high Chl-a concentrations observed

from the satellite image for that day on the southern border, close to the outfall and river discharges

(Fig. 3.7 a).

Table 3.2 Parameter description and values used for the simulations of AEM3D ecological model

for SMCA.

Parameter description Parameter Value Literature Values/remarks

Photosynthetically Active Radiation (PAR) fraction of incident solar radiation fPAR 0.45 0.45a,b Extinction coefficient for PAR (background) (m-1) KePAR 0.25 0.285a, 0.20b, 0.16-

0.48 at Gaira Bayc Bulk transfer coefficient for heat at air-water interface - 0.0013 0.0013d Bulk transfer coefficient for momentum at air-water interface - 0.0013 0.0013d Wind surface drag coefficient - 0.0013 0.0013 b,d Drag coefficient on bottom - 0.005 0.005 b,d Horizontal diffusion coefficient of momentum (m2 s-1) - 5.0 1.0b, 5.0e Oxygen

Maximum DO consumption by sediments (g DO m-2 d−1)

SSOD 1.13 0.2 -1.0a, 0.78 – 4.85 at Taganga Bay f, 0.39 – 5.52g

Dissolved Oxygen (DO) half-saturation constant for SOD (mg DO L-1) KSOD 0.5 0.5,a,b,h 1.5i Photosynthetic stoichiometry ratio of DO to C (g DO g C-1) YO:C 2.67 2.67a,b,h,i Nitrification stoichiometry ratio of N to C (g DO g N-1) YO:N 3.43 3.43 a,b,h,i Phosphorus

Maximum rate of microbial decomposition of POP to DOP at 20 °C (d−1) μDECPOP 0.06 0.06,a 0.01h, 0.05i Maximum rate of DOP mineralisation to PO4 at 20 °C (d−1) μMINPON 0.001 1.0a, 0.002-0.8g, 0.05i

DO half saturation constant mediating sediment FRP (g O m−3) KsFRP 3.0 3.0a, 1.0i Maximum FRP sediment flux (g P m−2 d−1)

SFRP

0.0001 0.0020a, -0.0004 – 0.0001 at Taganga Bayf, 0.00002-0.049g, 0.0005h, 0.004i

Maximum DOP sediment flux (g P m−2 d−1) SDOP 0.0001 0.0005a Nitrogen

Maximum rate of microbial decomposition of PON to DON at 20 °C (d−1) μDECPON 0.01 0.05a, 0.01i Maximum rate of DON mineralisation at 20 °C (d−1) μMINDON 0.001 1.0,a 0.024 - 0.3g Maximum NH4 sediment flux (g N m−2 d−1)

SNH4 0.001

0.007a, 0.010h, -0.0011 – 0.0045 at Taganga Bayf

Maximum NO3 sediment flux (g N m−2 d−1 SNO3 -0.0031

-0.005a, -0.0072 – 0.0010 at Taganga Bayf, 0.02i

Maximum DON sediment flux (g N m−2 d−1) SDON 0.001 0.001a DO half saturation constant mediating dissolved nitrogen sediment fluxes (g O m−3) KsNH4, KsNO3,

KsDON 1.5 1.5a , 0.5h, 1.0i

Nitrification rate at 20 °C (d−1) Knit 0.001 0.020a,h, 0.20i DO 1/2 saturation constant for nitrification(mg DO L-1) Knit 2.0 2.0a,h, 4.0i Denitrification rate (d−1). Kden 0.002 0.01a,h, 1.3 – 16.6

µmol N·m-2·h-1 at sediments – Taganga Bayf, 0.4i, 0.1j


96

2 – 50 nmol NL-1d-1k DO 1/2 saturation constant for denitrification(mg DO L−1) Kden 2.0 2.0a, 0.5h, Temperature multiplier for nitrification/denitrification (–) ϑnit; ϑden 1.08 1.08a,i Carbon

Maximum rate of microbial decomposition of POC to DOC at 20 °C (d−1) μDECPOC 0.01 0.01a, 0.07t Maximum rate of DOC mineralisation at 20 °C (d−1) μMINDOC 0.001 0.001a, 0.07h Maximum DOC sediment flux (g C m−2 d−1) SDOC 0.001 0.001a, 0.005t DO half saturation constant mediating dissolved organic carbon sediment flux (g O m−3) KSDOC 0.5 0.5a, 0.5t Specific light attenuation coefficient due to the action of POC (mg POC L-1 m-1) KePOC 0.02 0.02b, 0.05t Specific light attenuation coefficient due to the action of DOC (mg DOC L-1 m-1) KeDOC 0.01 0.004a, 0.01b, 0.001t Silice

Maximum Si sediment flux (g Si m−2 d−1) SSi 0.0013 0.12a , -0.002 – 0.005 at Taganga Bayf,

DO half saturation constant mediating dissolved silica sediment flux (g DO m−3) KsSi 1.5 1.5a Diameter of POM particles (m) dPOM 5.0x10-6 5.0×10−6 a,t Density of POM particles (Kg m-3) ρPOM 1070 1070a Settling velocity of particulate detritus (POM), used for POC, PON and POP VSPOM Calculated from

Stoke’s law Phytoplankton

Internal stoichiometry ratio of C to Chl-a (g C (g chl-a)−1) YC:Chl-a 42 60a, 42b,i, 47l Maximum potential growth rate (d-1) µmax 0.75 2.5a, 1.0-3.9g, 1.6i, 1.0l,

0.12-0.75 at Taganga Bay at 25 °C and 35 % salinityg

Internal stoichiometry ratio of P to Chl-a (g P (g Chl-a)-1) YP:Chl-a 0.9 0.3a, 0.9b Half-saturation constant for PO4

- uptake (mg P L-1) KP 0.003 0.001a, 0.001-0.3g, 0.003b, 0.0037l

Internal Stoichiometry ratio of N to Chl-a (g N (g Chl-a)-1) YN:Chl-a 12 12a , 9b Half-saturation constant for NO3 + NH4 uptake (mgL-1) KN

0.01 0.03a, 0.07b, 0 –0.5g,

0.004 -0.011l Internal Stoichiometry ratio of Si to Chl-a (g Si (g Chl-a)-1) YSi:Chl-a 12 12a, Half-saturation constant for SiO2 uptake (mgL-1) KSi 0.04 0.05a, 0.028b, 0.022-

0.126n Temperature multiplier for growth (-) uG 1.07 1.02a, 1.06b, 1.02-

1.14g, 1.07i Standard Temperature for growth (°C) TSTD 20 15a, 20b, 19o, 20 at

Taganga Bayp Optimum Temperature for growth (°C) TOPT 25 20a, 27b, 26o, 25 at

Taganga Bayp Maximum Temperature for growth (°C) TMAX 35 37a, 34b, 32o Algal respiration mortality, excretion and grazing (d-1) Kr 0.001 0.27a, 0.085b, 0.001-

0.17g, 0.15i, Temperature multiplier for metabolic losses (–) ur 1.06 1.12a, 1.06b, 1.07i,

1.02-1.14q Fraction of metabolic loss that goes to CO2 as respiration (-) Fres 0.8 0.8a, 0.7b,i Fraction of mortality and excretion that is DOM (remainder is POM) Fdom 0.2 0.2a Fraction of phytoplankton DO lost to photorespiration fDOpres 0.014 0.014b,h, 0.14i Initial light intensity for photosynthesis (µE m-2 s-1) Ik 190 50a, 120b, 40-100q. 146

± 47 non-upwelling period and 230 ± 58 upwelling period at TNNPr

Light Saturation for maximum production (µE m-2 s-1) Is 400 400b Specific light attenuation coefficient for phytoplankton (m2 (mg chl-a)−1) KextP 0.02 0.02a,b, 0.020 0.01–

0.03g Settling velocity (m s-1)

Vs 3.5x10-7 3.5x10-7a, 3.7x10-6b, 9.26x10-7 - 3.7x10-6s

Critical shear stress for phytoplankton resuspension (N m-2) tCP 0,001 0,001b,i Resuspension rate constant (mg Chl-a m-2 s-1) α, αf 0.008 0.008b,h Half-saturation constant of available phytoplankton mass for resuspension (mg Chl-a m-2) Kmass 0.01 0.01b,h, 1.0i

Sources: a Spillman et al. ( 2007); b Machado & Imberger (2012); c Barragán et al. (2009); d Woodward et al. (2017); e Hillmer & Imberger (2007); f Arroyave Gómez et al. (2020); g Hamilton & Schladow (1997); h Romero et al. (2004); i Robson and Hamilton (2004); j Jørgensen & Bendoricchio (2001); k Glud et al. (2015); l Ganguly et al. (2013); m Carrera et al. (2018); n Martin-Jézéquel et al. (2000); o Robson & Hamilton (2003); p Barreto Hernández & Velasco (2014); q Griffin et al., 2001), r Eidens et al. (2014); s Seip & Reynolds (1995); t Gal et al.(2009).


97

NMAE=0.52

(a) (b)

NMAE=0.24

(c) (d)

NMAE=0.51

(e) (f)

Fig. 3.7 Chl-a satellite images from Sentinel-3A during NUPW and UPW seasons for: (a)

November 26th 2017 (NUPW), (c) January 3rd 2018 (UPW) and (e) January 30th 2018 (UPW) and

modelled Chl-a at noon for SMCA (b) November 26th 2017, (d) January 3rd 2018 and (f) January

30th 2018. NMAE between simulated results and satellite images are reported.


98

3.4.3. Wind field and hydrodynamics of SMCA during upwelling and non-upwelling seasons

The wind field was the main driver of SMCA nearshore hydrodynamics and the outfall plume

dispersion during both November 2018 (NUPW) and December 2017 – February 2018 (UPW)

seasons (Fig. 3.8, Fig. 3.9 and Fig 3S.4). The water residence time was highly variable for both

seasons (Fig. 3.9). A week of simulation was considered as model spin-up period for both seasons

because it was the time needed to stabilize the residence time in a range around 3.7 – 5.7 days

between November 8th - 21th in NUPW and around 3.7- 7.3 days between December 8th - January

7th (Fig. 3.9 a and Fig. 3.9 b). The average residence times between November 8th – 30th (NUPW)

and between December 8th – February 27th (UPW) were 5.6 ± 1.1 d and 4.2 ± 1.1 days, respectively.

East-southeast (ESE) winds with average direction 122 ± 100° (clockwise from north) and low

speed (1.6 ± 1.4 m·s-1) dominated during the NUWP season while north-northeast (NNE) winds

with average direction 46 ± 69° and high speed (5.0 ± 2.2 m·s-1) prevailed during the UPW season

(Fig. 3.8 a and Fig. 3.8 b). The high variability of the water residence time during the last days of

November and December – mid January was possibly due to the alternation of variable winds

periods during the transition from NUPW to UPW seasons (Fig. 3.9 a and Fig. 3.9 b). The

November 22th – 30th easterly winds average direction (86 ± 97°) and speed (2.4 ± 1.7 m·s-1), the

December 8th – 18th north-northeast NNE winds average direction (69 ± 71°) and speed (2.9 ± 2.0

m·s-1) and the December 30th – January 14th east-southeast winds average direction (104 ± 105°)

and speed (2.6 ± 2.0 m·s-1) were variable winds periods with a high variation in both direction and

speed (Fig. 3.8 a and Fig 3.8 b). The daily wind pattern showed a trend with southwesterly (SW)

winds blowing in the mornings and early afternoon while easterly (E) winds, slightly to the

southeast, prevailed at night and early morning in the period of the greatest intensity of NUPW

between November 8th- 21th. This daily pattern had a slight increase in the westerly component in

the southwest winds whereas the eastern component of winds began to turn towards the north –

northeast (NNE) in variable winds periods. The NNE winds with high average speed (greater than

5m·s-1) prevailed all day during upwelling periods in December 19th -29th and January 15th –

February 27th (Fig. 3.8). The period of greatest intensity and duration of upwelling had an average

residence time of 3.7 ± 0.4 days.

The plume and water flow transport along the open borders depended on the wind field in both

seasons (Fig. 3S.4). Average wind speeds equal to or greater than 2.0 m·s-1 in southwest direction

in the afternoon generated water flow input along north (N) and northeast (NE) directions through


99

the southern and western open borders, forcing a plume displacement towards the north (Taganga

Bay) as occurred in November 8th -9th (Fig. 3S.4 a). The offshore plume transport towards the

northern and western open borders was caused by low wind speed during all day in NUPW season

(from November 12th to 24th) (Fig. 3S.4 b). This was also the case for the offshore advected plume

during the mornings and early afternoons when there was a dominance of low intensity southwest

winds in the UPW season such as December 29th-30th, January 1st and 8th, February 25th-27th (Fig.

3S.4 g and 3S.4 l). Increments in the plume residence time during periods of time such as November

24th-26th, December 8th-12th and January 1st-3rd were observed due to sharp changes in both the

wind speed and direction such as NWN - NNE in the afternoon of November 26th (Fig. 3S.4 c),

NNE - SW (Fig. 3S.4 f) in the morning of December 11th, E-SW in the morning of January 3rd (Fig.

3S.4 h). These changes led to water input through three open borders or to the reduction in the

water outflow. On the other hand, high average NNE wind speeds (equal to or greater than 4 m·s-

1) but strongly in northern direction caused both reductions in the residence time (November 27th-

30th and January 5th-6th) (Fig. 3S.4 d) and the advection of the plume towards the south (Santa

Marta Bay) (Fig. 3S.4 e) due to water flow from the north to the south border and in SW direction.

The whole day NNE high speed winds produced a water flow input through the north open border

(in direction S and SW) and its exit through the south and west open borders in periods of high

intensity UPW season (Fig. 3S.4 i and Fig. 3S.4 k). The decrease in the intensity of NNE winds

and SW winds of low speed caused a decrease in the SW component of water currents of the UPW

season. This produced a slight inflow through the southern border (Fig. 3S.4 j and Fig. 3S.4 l).

The flow velocity pattern (X and Y components) for both seasons was similar for the surface and

at 40m depth during the simulation period. The X and Y components of the averaged flow velocity

for surface were -0.45 ± 0.11 m s-1 (range -0.56 – 0.41 m·s-1) and -0.13 ± 0.03 m s-1 (range -0.38 –

0.01 m·s-1), respectively, with a dominant north-northeast (NNE) flow direction. The average flow

velocity of these components for 40 m depth were 0.05 ± 0.03 m s-1 (range -0.10 – 0.14 m s-1) and

0.04 ± 0.05 m s-1 (range -0.10 – 0.17 m s-1), respectively, with a dominant northeast (NE) flow

direction. The Y velocity component was higher than the X component in both seasons for surface

water flow (Fig. 3.9 c and Fig. 3.9 d). This pattern had a shift from noon November 28th where the

flow velocity components were reversed: the X component reached its maximum and positive

value (currents from west to east) while the component Y reached its minimum and negative value

(currents from north to south). This was caused by an abrupt change from low-speed west-

southwest winds to high speed north-northeast (NNE) but strongly to north which coincided with


100

a sharp decline of residence time of the plume from 7.0 to 4.8 days in a few hours (the afternoon

and night of that day) due to water inputs along north border in NS direction (Fig. 3S.4 d). The

entering water flow had a higher surface temperature (~30.5 °C) than the plume (~29.0°C), which

caused an appreciable difference between the average surface temperature and 40 meters up to

November 30th (Fig. 3.9 g). The seasonal change in temperature caused by the upwelling was

evident both in the outfall vicinity (Fig. 3.9 h) and in the ACSM (Fig. 3.6 b and Fig. 3.6 d) in the

simulation period. There was a progressive decrease in temperature until reaching average and

minimum temperatures (23.0 ± 1.8 ° C, a minimum value of 20.6 ° C) close to the outfall in the

period of greatest intensity of the upwelling (Fig. 3.9 h).

The increased wastewater flow-rate to 2.5 m3·s-1, corresponding to a volume flux of 315 m3·s-1,

caused an increment in the plume horizontal turbulent dispersion, which reduced the influence of

currents directions and thermal changes observed in the base scenario on November 28th (Fig. 3S.5

g and Fig. 3.9 g).

Fig. 3.8 Wind field measurements at Taganga Bay for the simulation period (November 2017 –

February 2018): (a) wind speed and (b) wind direction (measured clockwise from the north).


101

Fig. 3.9 Average residence time (“water age”) series: (a) November 2017 - NUPW and (b)

December 2017 – February 2018 - UPW season. Average surface flow velocity series (Vx and Vy):

(c) NUPW and (d) UPW season. Flow velocity series (Vx and Vy) at a depth of 40 m: (e) NUPW

and (f) UPW season. Series of average surface temperature of at 40 m (g) NUPW and (h) UPW

season near to SMSO during the simulation.


102

3.4.4. Nutrients and phytoplankton dynamics at different loads of the outfall along NUPW

and UPW seasons.

The one-week model spin-up period was also evident in the Chl-a dynamics but was less apparent

for nutrients in both seasons (Fig. 3.10). The modeled average surface concentrations of nutrients

and Chl-a in the outfall vicinity during the simulation period captured the seasonal variability of

field data. NH4+, PO4

3- and DO decreased on average by 460 %, 3 % and 7 %, respectively, while

NOx and TP increased by 25 % and 14 % between the two seasons, respectively (Fig. 3.10).

Modelled NH4+ had a drastic decrease from December until the end of the simulation period, which

agrees with both the SMCA water quality background and the field data during this period (Fig.

2).

The modelled average surface Chl-a close to the outfall plume during the NUPW season was 1.29

± 0.12 µg Chl-a L-1, range = 1.11 -1.72 µg Chl-a L-1 whereas during the UPW season it averaged

0.81 ± 0.35, range = 0.10 -1.48 µg Chl-a L-1. These values corresponded to an average decrease of

around 59% between the two seasons. Phytoplankton growth was strongly related to temperature

as shown by the similarities between the surface average time series of temperature (Fig. 3.9 g and

Fig. 3.9 h) and Chl-a (Fig. 3.10 k and Fig. 3.10 l). The warmer water inputs through the northern

open border from November 28th to 30th (NUPW) (Fig. 3.9 g) stimulated the phytoplankton growth

and nutrients uptake due to the slight decrease in NH4+, NOx

-, PO43-, TP and DO (Fig 3.10). The

lowest temperatures and Chl-a concentrations were reached during the period of highest intensity

of the upwelling (January 15th - February 27th). Despite the lowest temperatures, shortest residence

times and a most advective environment, several Chl-a peaks were also found (Fig. 3.10 l). Chl-a

peaks close to the outfall coincided with a high nutrient supply, a deeper mixing layer (Fig. 3.6 b

and Fig 3.6 d), a slightly increased incident radiation, lower extinction of photosynthetically

available radiation (PAR) (Fig. 3.S6) and a greater Chl-a inputs through the northern border during

the last days of February. In general, higher Chl-a concentrations were observed in the outfall

vicinity than in the rest of the domain without the influence of the upwelling between February

13th-16th (Fig. 3.10 l). The upwelling caused by water inputs with Chl-a concentrations in the

northern border, which were higher on February 26th -27th (Fig. 3.10 l), showing an evident effect

on the Chl-a close to the outfall. The simulated surface Chl-a were in the range of field data and

satellite images (0.51 - 2.69 µg Chl-a L-1).


103

The increase of outfall loading due to the increment of flow-rate to 2.5 m3·s-1 produced an average

increase of nutrients in the plume surface layer for NH4+ (~ 6%), NOx

- (~ 5%), PO43- (103%) and

TP (~ 60%) while a slight decrease in Chl-a of around (~ 2 %) and unchanged DO concentrations

for NUPW season. The decrease of Chl-a between November 28th and 30th was produced by the

outfall wastewater increment which reduced the effect of warmer waters inputs on the plume,

limiting the growth of phytoplankton and the nutrients uptake in comparison with the base scenario

(Fig. 3.10 and Fig. 3.S5). The increased discharge from the outfall also produced an increase in

nutrient concentrations for NH4+ (~ 28%), PO4

3-(~ 103 %), TP (~ 51%) and Chl-a (~2%) and a

decrease for NOx- (~1.8 %) and DO (~4%) during the UPW season in comparison with the base

scenario.

The simulated surface concentrations of TP, PO43- and TOC were slightly higher in the outfall

vicinity for both seasons, in contrast with NH4+ and NOx

-. Although TP, PO43- and TOC were

rapidly diluted, some variability was observed in their spatial distribution. The CGSM waters input

at the southern border during the NUPW season dominated the NOx- dynamics, whereas its

increment during the UPW season was due to the seasonal upwelling. The increased outfall load

produced an increment in the TOC signature on the transect of around 20%, 15-20% and 14-16%

for distances of 600m, 1200m and 1800m in both seasons, respectively (Fig. 3.S7).

On the other hand, in the same transect, the average surface Chl-a distribution between February

14th-16th was evaluated for the base scenario where a Chl-a peak was observed. The average Chl-a

decreased around 12-19%, 45-51% and 54-60% at distances of 600 m, 1200 m and 1800 m,

respectively (Fig. 3.S8).

The simulations without the outfall showed a higher Chl-a than in the two scenarios evaluated with

the wastewater discharge during the NUPW season and part of the UPW season (Fig 3.10 k and

Fig 3.10 l). This is due to the fact that Chl-a concentrations in the open borders (western and

southern) are higher or of the same magnitude as near the outfall in November and December.

However, this effect is progressively reduced in January and February. The Chl-a is higher for both

scenarios with the outfall between around February 10th -20th. On the other hand, higher

temperature and lower advection favor the increment in Chl-a without the outfall. The water

temperature was higher on average by nearly 1.7 ° C and 1.1 ° C without outfall during NUPW and

UPW seasons, respectively.


104

Fig. 3.10 Surface-averaged nutrient concentrations in NUPW and UPW season close to SMSO

during the simulation: (a) NH4+ (NUPW) and (b) NH4

+ (UPW), (c) NOx (NUPW) and (d) NOx

(UPW), (e) PO43-(NUPW) and (f) PO4

3-(UPW), (g) TP (NUPW) and (h) TP (UPW), (i) DO

(NUPW) and (j) DO (NUPW), (k) Chl-a (NUPW) and (l) Chl-a (UPW).


105

3.5. Discussion

Carbon, nitrogen and phosphorus biogeochemistry in coastal tropical areas is understudied and

research efforts need to be intensified in order to understand how interactive ongoing global

changes and local pressures affect coastal ecosystem functioning. Coastal ecosystems may

apparently buffer multiple pressures but in reality, they may progressively erode their capacity to

be resilient, reach tipping points and suddenly display nonlinear responses, changing their status.

In many areas of the world stable coral reefs coastal systems were replaced by macroalgae-

dominated stable systems, due to combination of excess nutrient loads and other threats as excess

fishing pressure (Norström et al., 2009; Reopanichkul et al., 2009; Fung et al., 2011). Such

transitions among stable states can be progressive or sudden and unexpected, and can be very

difficult to reverse, with important economic implication for local populations.

This work aimed at modelling different physicochemical parameters in a complex tropical system

where multiple actors and stressors interact. Apart from the hydrodynamic aspects, the complexity

in SMCA arises from the interaction between two potential sources of eutrophication (increase of

organic matter fixation rates, sensu Nixon, 1995): the untreated wastewater outfall from nearly

500,000 people and nutrient inputs during UPW periods. Other nutrient sources are also potentially

important (e.g. river-associated loads) but they were not directly addressed in this study by seasonal

effect and scarce data. This modelling effort follows an experimental work recently published,

targeting benthic biogeochemistry along a transect from the sediments nearly the outfall to

increasingly distant stations (Arroyave-Gomez et al., 2020). Such experimental work was carried

out during NUPW and UPW seasons and produced data that were used as input data to the model

applied in this study. Interestingly, detailed analysis of sediment composition and metabolic rates

highlighted a much larger than expected area impact of the outfall in the bay and an interaction

between the outfall and the upwelling, affecting process rates. These results should represent an

early warning signal for this touristic area which includes a national park and valuable coral reef

ecosystems.

Sediments have longer memory of impacts than the water column (Abessa et al., 2005; Burd et al.,

2013), as advective processes, mixing and dilution can produce steep changes in water chemistry

and biology (e.g. phytoplankton blooms and algal community composition). Sediments are

therefore easier targets than the pelagic system to address ongoing changes, and pelagic systems

need modeling tools to reproduce and understand the biogechemical processes, as traditional water


106

sampling is not enough. Besides experimental benthic data, this works took advantage of satellite

remote sensing of Chl-a and SST, and of water sampling along a grid of stations.

The simulations were generally able to reproduce the magnitude and timing of complex dynamics

and fast transitions of temperature, nutrients, and phytoplankton, including the time and duration

of stratification and mixing periods during the NUPW and UPW seasons. The NMAE statistical

values were comparable with other modelling studies (Gal et al., 2009; Machado & Imberger,

2012) which supports the goodness of the simulations performed. I believe that the model could be

further implemented and used to predict the physical and biochemical variables under a wider range

of climate conditions (e.g. wind direction and speed), of organic matter and nutrient inputs, and

under NUPW and UPW season in SMCA.

3.5.1. Circulation patterns during non-upwelling and upwelling seasons in the SMCA

The model was able to reproduce the influence of winds as main driver of water currents,

circulation patterns, change of physical variables (e.g. the temperature from 30°C to 21 °C and the

salinity from 33 to 38) and seasonal upwelling intensity in SMCA as described by Arévalo-

Martínez & Franco - Herrera (2008), Bayraktarov et al. (2014) and García (2013). Southwest winds

drive currents from the SW to NE in NUPW season carrying warmer water inputs from the CGSM

in the south while the strong NE winds drive currents from to NE to SW during UPW season

leading to colder water inputs in the north (Bayraktarov et al., 2014). The model also reproduced

the presence of two different water masses due to these changes in circulation. High temperature

and low salinity waters correspond to the effect of the Panama-Colombia countercurrent that

produces the CGSM inputs in NUPW season in SMCA with high load of organic material and Chl-

a, which considerably reduces the water transparency and restricts the photic zone depth (Franco-

Herrera et al., 2006; Mancera-Pineda et al., 2013). On the other hand, low temperature and high

salinity waters during the UPW season correspond to Subtropical Subsurface Water located

between 100 and 200 m deep in the Caribbean Sea (Arévalo-Martínez & Franco - Herrera, 2008;

García et al., 2012; Paramo et al., 2011). The simulations captured the stratification and mixed

patterns of the water column coinciding with the NUPW and UPW seasons, respectively (García

et al., 2012; García, 2013). The thermoclines predicted by the model nearshore fall within the

thermocline thicknesses estimated (up to 12 m) during the stratification periods in the SMAC

(García et al., 2012; García, 2013). Predicted dominant circulation at 40 m depth in the outfall


107

vicinity was to the northeast (NE), which is closely consistent with the currents field measurements

east northeast (ENE) (García, 2013).

The model captured sharp changes, in the scale of days and hours, of currents and residence time,

which led to fast temperature changes (November 28th to 30th). This was possibly due to a good

resolution of the boundary conditions (e.g. meteorology, temperature and salinity) and bathymetry.

The HYCOM data reproduced with a good fit temperature and salinity field profiles for the SMCA

(García, 2013). However, the underestimation at the southern border was likely due to the HYCOM

station location that was 2.8 km north of the border. Such location likely limited the possibility to

capture the highly variable dynamics of temperature along this border.

3.5.2. Effect of increased nutrient load of the outfall on phytoplankton and nutrients

dynamics during NUPW and UPW seasons in SMCA

The main limitation factor of phytoplankton growth in SMAC was the water temperature, which

had a high variability on the scale of days and hours. The changes of this variable revealed that the

phytoplankton growth showed a slightly higher coupling with the nutrients supply in the outfall

plume during the NUPW season than during the UPW season. Despite the increase in the nitrate

concentration produced by the upwelling, a decoupling between nitrate supply and Chl-a growth

was evident. The decoupling between phytoplankton and nutrient supply due to the increase of

advection and water mass mixing during UPW season in SMCA has been described by Arévalo-

Martínez & Franco - Herrera (2008) and Paramo et al. (2011), which was also supported in this

study by the estimates of residence time in both seasons. On the other hand, the simulations indicate

that during the NUPW and transition seasons, the impact of the outfall was less compared to the

influence of CGSM, whereas in the higher intensity period of the upwelling these were

progressively reduced. The nutrient inputs from the outfall caused a higher stimulation of

phytoplankton growth picks than the rest of the domain between February 13th -16th (Fig. 3.10 l),

and its effect was intensified by the upwelling in agreement with Mancera-Pineda et al. (2013). In

the simulations, this influence was more evident during the last days of February when high Chl-a

concentrations entered through the northern border and Chl-a peaked with 1.48 µg Chl-a L-1 in the

plume. Higher Chl-a concentrations by seasonal upwelling have been observed between February

and April in the SMCA, with values between 0.007 and 1.61 µg Chl-a L-1 (Ramírez-Barón et al.,

2010).


108

Water temperature and nitrate concentrations have been found as significant variables for

phytoplankton growth in the SMCA (Ramírez-Barón et al., 2010). In addition, the light intensity

also limited phytoplankton growth because of the increase of the Chl-a from February 13th,

corresponding to higher PAR extinction. Phytoplankton growth rate in nutrient-rich estuaries is

determined in large part by light availability, which varies with incident solar irradiance, turbidity,

and depth of the mixed layer (Cloern et al., 2014). The increase of the mixed layer increases the

photic layer leading to a higher availability of light for photosynthesis, increasing the incorporation

of nutrients from the phytoplankton (Cloern et al., 2014). Therefore, phytoplankton appeared to

respond moderately to high nutrient concentrations from the outfall during the period of greatest

intensity of the UPW season in agreement with lower water temperatures for optimal growth,

higher advection and mixing (e.g. lower residence times) and a higher availability of light.

In general, the model showed a better adjustment of temperature and Chl-a with the field data

during the UPW season than during NUPW season due to the reduction of horizontal gradients

caused by the influence of the CGSM waters. The model did not capture the high Chl-a

concentrations in the southern border in NUPW as a result of higher underestimation of the

modelled temperature over this border, the advective flows across border did not allow to reproduce

phytoplankton growth rate (Hillmer & Imberger, 2007) and some numerical dispersion due to its

coarse resolution (Spillman et al., 2007), since the main focus of this study was the region in the

vicinity of the outfall.

The nutrients showed a high variability attributed to the plume advection in the outfall vicinity.

Nevertheless, slightly higher concentrations of TP, PO43- and TOC in the outfall vicinity were

possibly due to their high concentrations in the wastewater effluent (Table 3S.1) and low

concentrations along the open border. The simulated nutrient and Chl-a had the same order of

magnitude near the outfall and in the SMCA as those reported by Ramírez-Barón et al. (2010),

Mancera-Pineda et al. (2013) and Garcés-Ordóñez et al. (2016). Low NH4+ concentrations during

the UPW season are likely due to both progressive reduction of NH4+ inputs by CGSM and rivers

(Mancera-Pineda et al., 2013), and the uptake of this nutrient by phytoplankton.

From the nutrients perspective it was not possible to delimit the area affected by the outfall because

of multiple factors like high dilution, uptake, low concentrations, resulting in unclear trend along

the considered transect. However, the combination the several approaches (e.g. experimental,

modeling and via remote sensing) allowed determining the environmental impact of SMSO in this

coastal marine ecosystem, a highly variable and complex hydrodynamics. Chl-a satellite images


109

traced an affected area of about 1500 m on a day characterized by high residence time during the

simulation period which is in agreement with the estimated spatial impact on sediment around 1800

m using stable carbon isotopes (Arroyave Gómez et al., 2020). In addition, the stimulation of

phytoplankton growth in the outfall plume during short time periods was estimated by the model

during the NUPW and UPW seasons.

It is intrinsically difficult to catch the effects of a point pollution source in the sea. Analyzing the

impact of cage fish farms in coastal areas, (Dalsgaard & Krause-Jensen, 2006 and references

therein) highlighted that studies focusing on nutrient and phytoplankton concentration analysis

along transects from the cages to increasingly distant stations, repeated in different moments of the

day, were not able to catch significant impact of fish metabolic activity, sediments impacts and fish

feed loss on the water column. However, the use of micro and macroalgal bioassays (the

measurement of algal growth in dialysis bags suspended at increasing distances from the cages)

allowed to catch significant differences in primary production (higher closer to the cages). The

work of (Dalsgaard & Krause-Jensen, 2006) allows to infer that present analytical capacity is not

enough to catch very small concentration differences in a highly diluted and mixed medium as the

open sea, but that organisms with nutrient extraction capacity in a diluted medium as phytoplankton

allow to reveal local impacts. These important outcomes should be taken into account when

discussing the results of this study. In the SMAC the outfall may apparently not determine a

measurable (analytically) or visible (via satellite remote sensing) impact on the physicochemical

or biological features of the study area, but this does not mean that there is no impact. What is

likely true is that locally the outfall increases pelagic primary production as well as grazing by

zooplankton or organic matter settling on sediments, with no apparent negative consequences on

the pelagic ecosystem. This hypothesis is supported by sediments analysis reported in Arroyave-

Gomez et al. (2020). Moreover, the simulations carried out under a scenario of increasing nutrient

input from the outfall reveal a small (2%) but measurable increase of phytoplankton in the area.

The increment of wastewater loading due to the increment of flow-rate to 2.5 m3 s-1 produced a

slightly higher decoupling between phytoplankton growth and nutrient supply for the NUPW

season than UPW season (e.g. slight increase in Chl-a corresponded to a slight decrease in NOx-).

This shows a synergistic effect between nutrient supply, temperature and light and advection due

to both the increase of discharge as well as mixing of upwelling season. In highly mixing and

dispersive environments, as the outfall vicinity, the phytoplankton growth is not tightly coupled

with nutrient source (Leon et al., 2012).


110

3.5.3. Implications for Benthic – pelagic coupling of SMAC

The maximum sediment nutrient release rates (NH4+, NO3

-, PO43+ and Si) and SOD specified by

the model were based on measurements derived from experiments conducted in Taganga Bay

within the simulation period and provided critical data to refine the accuracy of the simulations

(Arroyave Gómez et al., 2020). Low mineralization rates of carbon, nitrogen and phosphorus in

the water column obtained during the calibration process coincided with the low regeneration rates

of inorganic nutrients from the sediment to the water column found in the Taganga Bay (Table 3.2)

(Arroyave Gómez et al., 2020), indicating a refractory macromolecular quality of organic matter.

This organic matter recalcitrant nature could possibly be attributed to bound nitrogen in form of

humic acid dissolved from Magdalena River and CGSM, as well as, organochlorine pesticide

residues absorbed in suspended organic matter (Corredor et al., 1999; Espinosa et al., 1995). The

low nitrification rate obtained during calibration may be due to the inhibition of nitrifying bacteria

by low turbidity and high intensity of light in tropical estuaries (Corredor et al., 1999). The model

also considered the sediment nitrate sink and the high variability of SOD accounted high sediment

respirations nearshore (Arroyave Gómez et al., 2020). The model proved to be quite reliable in

reproducing the dynamic and timing of nutrients and phytoplankton in NUPW and UPW seasons

in the SMAC in a wind field wide range. This was because the model parameters were intensively

calibrated and complemented with site-specific parameters and experimental measurements.

The increment of wastewater loading increased the TOC signature (Fig. 3S.7) in the water column

which could cause a raise in the NH4+ release via DNRA from sediment considering the high

fraction of particulate organic material (POM) of SMSO effluent (Table S1) and the significate

correlation found on transect towards Taganga bay between TOC and DNRA in the sediment

within the simulation period (Arroyave Gómez et al., 2020). The high NH4+ variability of the field

data coincided with the seasonal influence of CGSM waters, river discharges and runoff as

described by Mancera-Pineda et al. (2013). Nevertheless, the NH4+ contribution via DNRA could

possibly be higher in the SMCA than that found in the outfall vicinity (Arroyave Gómez, et al.,

2020). The Chl-a, particulate and dissolved organic material from the CGSM waters, can be

considerably higher than that of the outfall. The seasonal variability of the Magdalena River

sediment plume extends until the SMAC (Ramirez, 2017) whereas CGSM is a hypertrophic

shallow coastal lagoon with Chl-a concentrations fluctuating between 5 and 181 µg Chl-a L-1

(Hernández & Gocke, 1990). CGSM was classified as a highly productive system with the highest


111

annual phytoplankton primary production of the 131 estuarine and coastal ecosystems analyzed by

Cloern et al. (2014). The predominance of DNRA over denitrification has been found in tropical

estuaries with large organic matter inputs (e.g. anthropogenic and natural) (Dong et al., 2011;

Bernard et al., 2015; Arroyave Gómez et al., 2020). Even though DNRA and heterotrophic N

fixation has possibly a higher potential importance in coastal ecosystem to N budget and

productivity at regional and global scales (Bernard et al., 2015; McCarthy et al., 2015; Newell et

al., 2016), these processes have not been included or are underrepresented by the ecological model

(Gardner et al., 2006). Therefore, more research in tropical systems must be carried out to

accurately model the coastal marine N budget.

The model represented the sediment nutrients release by the static model which regulates its release

according to the changes in temperature and DO. This does not consider the effect of the external

nutrient load on the sediments. The inclusion of feedback between water column and sediment and

the nitrogen remineralization via DNRA requires the implementation of a diagenetic model that

considered this process and/or co-occurrence between DNRA and denitrification. However, its

implementation requires combination of extensive benthic measurements database under different

spatiotemporal conditions and testing several approaches in numerical models (e.g. depth-

integrated diagenesis models, etc). The approach combinations would allow understanding major

shifts and budget N in SMCA considering the recent increase of phytoplankton blooms in the area

(Garcés-Ordóñez et al., 2016) which represents a warning signal due to overlapping several nutrient

sources (e.g. anthropogenic and upwelling). The implementation of a vertically resolved dynamic

sediment diagenetic model in a biogeochemical model as the one applied in this study, it is not

realistic due to the high computational cost considering the wide modeled area, the steep bottom

gradient, and the plume transport and dispersion simulation (Wilson et al., 2013). Nevertheless, the

biogeochemical model configuration could be considered a viable option to represent sediments

due to lower residence time, high dilution capacity and slight decoupling between phytoplankton

and nutrients in pelagic systems and in the coupling benthic-pelagic. The results of the simulations

suggest that the model is useful for predicting some important variables of water quality in the area

nearshore.


112

3.6. Conclusions

The implemented model provides useful predictions of nutrients and Chl-a dynamics in SMCA

over a wide range of temperature and hydrodynamic conditions during the NUPW and UPW season

due to the use of a good resolution boundary conditions data (e.g. HYCOM and meteorological

data and Chl-a distribution from satellite images), as well as, an intensive calibration process of the

ecological parameters. The model predicted the seasonal sequence of phytoplankton in the outfall

vicinity due to changes produced by seasonal upwelling (e.g. fertilization) and the factors that most

limited it growth rates (temperature, light and nutrients). Growth limiting factors and

hydrodynamics were changing rapidly, at the scale of hours and days. In addition, the model

captured short periods in which there was a moderate phytoplankton growth stimulation by

nutrients supply near the outfall diffuser in both seasons and seasonal upwelling, but this showed

a slight uncoupling in both seasons by changes of water temperature and advection and mixing.

The model proved to be a reasonably reliable management tool to predict nutrient and

phytoplankton dynamics, and to analyze the individual role of different inputs during NUPW and

UPW seasons. This work is not conclusive in the sense that more experimental data (e.g. primary

production measurements, water column metabolic rates) should be collected in order to further

improve the data input. Monitoring activities should include water physicochemical parameters

and process rates, sedimentary features and benthic metabolism and the diversity and abundance

of benthic and pelagic organisms, in order to catch ongoing changes and to contrast them. Sediment

analyses and preliminary outcomes from the present work suggest that the effluents from the

sewage outfall should be treated in order to reduce the loads of particulate organic matter and

nutrients. This is urgent as the load is expected to more than double in the near future, in order to

maintain the ecological status of this area which supports the tourism industry.


113

3.7. Appendices Chapter 3

3.7.1. Materials and methods

Table 3S.1 Wastewater effluent quality of SMSO for simulated scenarios in the AEM3D model

Variable (units) Wastewater effluenta NH4-N (mg N/L) 21.37 NOx-N (mg N/L) 2.00 PO4-P (mg P/L) 10.27 bDOC (mg C/L) 115.70 cDON (mg N/L) 9.55 cDOP (mg P/L) 3.75 bPOC (mg C/L) 173.54 cPON (mg N/L) 1.97 cPOP (mg P/L) 1.39

a Untreated effluent quality data measured at SMSO for 2006 García (2013) b Calculated from soluble BOD/total BOD and BOD/TOC ratios of 0.4 and 1.0 for untreated wastewater, respectively (Metcalf & Eddy, 2003; Henze and Comeau, 2008) c Calculated from soluble N/ TN and soluble P/TP ratio 0.83 and 0.73 for untreated wateswater, respectivaly (Henze and Comeau, 2008)


114

3.7.2. Results

Table 3S.2 Statistical comparison between model simulations and field data for surface (0 m), 20m

and 40 m depths of Santa Marta Coastal Area (SMCA) for several variables of the AEM3D model.

Variables Depth (m) Calibration

(26 Nov. 2017 -NUPW) NMAE

Validation (26 Jan. 2018 -UPW)

NMAE

Validation (4 Feb. 2018 - UPW)

NMAE

Temperature (°C)

0 0.02 0.01 0.02

20 0.02 0.02 0.02

40 0.00 - 0.01

Salinity (ups)

0 0.04 0.04 0.03

20 0.03 0.02 0.01

40 0.03 - 0.01

NH4+

(mg N·L-1)

0 0.49 - 0.21

20 0.34 - 0.30

40 0.46 - 0.30

NOx (mg N·L-1)

0 0.05 0.22 0.12

20 0.07 0.27 0.18

40 0.10 - 0.23

PO43-

(mg P·L-1)

0 - - 0.86

20 0.14 - 0.39

40 - - 0.50

TP (mg P·L-1)

0 0.87 - 1.00

20 0.23 0.74 0.50

40 - - 0.30

DO (mg O2·L-1)

0 0.08 0.06 0.08

20 0.01 0.03 0.05

40 0.01 - 0.08

Chl-a (µg Chl-a·L-1)

0 0.46 0.44 0.36

20 3.16 2.31 1.64

40 5.78 - 0.44


115



3-, TP, DO and Chl-a during the calibration period in November

2017 (NUPW) at 6 stations located in Taganga (P2, P3, P4 and P5), Santa, Santa Marta (P7) and

Gaira (P9) bays. Field data correspond to measurements carried out on Novemver 26th 2017.


116

Fig. 3S. 2. Comparison of model simulation profiles (grey color) against field monitoring data (red circles) for NH4

+, NOx, PO43-, TP, DO and Chl-a during the validation period in December 2017 –

February 2018 (UPW) at stations located in Taganga (P2, P3, P4 and P5), Santa Marta (P7) and Gaira (P9) bays. Field data correspond to measurements carried out on January 26th 2018. Field data were not measured at P2 and P4 in January. NH4

+ field data were below the detection limit at all stations.


117

Fig. 3S. 3. Comparison of model simulation profiles (grey color) against field monitoring data (red

circles) for NH4+, NOx, PO4

3-, TP, DO and Chl-a during the validation period in December 2017 –

February 2018 (UPW) at stations located in Taganga (P2, P3, P4 and P5), Santa Marta (P7) and

Gaira (P9) bays. Field data correspond to measurements carried out on February 4th 2018.


118

(a) Nov. 8th in the afternoon (b) Nov. 19th in the early morning

(c) Nov. 26th in the afternoon (d) Nov. 28th in the afternoon and night

(e) Nov. 30th in the afternoon (f) Dec. 11th in the morning (g) Dec. 29th in the morning (h) Jan. 3rd in the morning

(i) Jan. 26th in the morning (j) Feb. 3th in the morning (k) Feb. 4th in night (l) Feb. 26th at noon

Fig. 3S. 4. Simulated-surface residence time and flow velocity of the SMSO plume in November 2017 - NUPW and December – February 2018 – UPW seasons in SMCA: (a) southwesterly (SW) winds gust, (b) low wind speed in the NUPW, (c) changes in wind direction from WNW – NNE in afternoon in the NUPW, (d) NNE winds but slightly in the north direction, (e) NNE winds, (f) changes in the direction from NNE – WS in the morning in the UPW, (g) low intensity southwest winds in the UPW season, (h) changes in the wind directions E-SW in the morning, (i) and (k) NNE winds in the UPW, (j) decrease the intensity of NNE in the mornings and (k) WS winds in the UPW season.


119

Fig. 3S. 5. Average residence time (“water age”) series (a) November 2017 - NUPW season and

(b) December 2017 – February 2018 – UPW season. Average surface flow velocity series (Vx and

Vy) in (c) NUPW and (d) UPW season. Flow velocity (Vx and Vy) at a depth of 40 m in (e) NUPW

and (f) UPW season. Series of water mean flow – rate of 2.5 m3·s-1 during the simulation.


120

(a)

(b) Fig. 3S. 6. Modelled PAR extinction profiles at different times in SMAC: (a) Calibration period

(November 2017 - NUPW) and (b) Validation period (December 2017 – February 2018 - UPW).


121

Fig. 3S. 7. TOC quantification along a 1800m transect from the SMSO towards Taganga and Santa

Marta bays for flow – rates of 1.0 m3·s-1, 2.5 m3·s-1 and without outfall. Results based on modelled

average surface concentrations in: (a) NUPW and (b) UPW season.


122

Fig. 3S. 8. Chl-a concentration quantification through a 1800m transect from the SMSO towards

Taganga and Santa Marta bays for flow – rates of 1.0 m3·s-1and without outfall. Results based on

modelled average surface concentrations of Chl-a between February 14th -15th 2018 UPW season.


123

NAME=0.48

(a) (b)

NAME=0.66

(c) (d)

Fig. 3S. 9. Chl-a satellite images from Sentinel-3A during UPW season for (a) February 14th 2018

(c) February 15th 2018 and modelled Chl-a at noon for SMCA (b) February 14th 2018, (d) February

15th 2018. NMAE between simulated results and satellite images are reported.


124

3.8. References

APHA, 2012. Standard methods for the examination of water and wastewater, 22nd edition edited

by E. W. Rice, R. B. Baird, A. D. Eaton and L. S. Clesceri. American Public Health

Association (APHA), American Water Works Association (AWWA) and Water

Environment Federation (WEF), Washington, D.C., USA.

Abessa, D. M. S., Carr, R. S., Rachid, B. R. F., Sousa, E. C. P. M., Hortelani, M. A., & Sarkis, J.

E. (2005). Influence of a Brazilian sewage outfall on the toxicity and contamination of

adjacent sediments. Marine Pollution Bulletin, 50(8), 875–885.


Alewell, C., & Manderscheid, B. (1998). Use of objective criteria for the assessment of

biogeochemical ecosystem models. Ecological Modelling, 107(2–3), 213–224.

https://doi.org/10.1016/S0304-3800(97)00218-4

Andrade, C. A., & Barton, E. D. (2005). The Guajira upwelling system. Continental Shelf

Research, 25(9), 1003–1022. https://doi.org/10.1016/j.csr.2004.12.012

Andrade, Carlos Alberto. (2003). Evidence for an eastward flow along the Central and South

American Caribbean Coast. Journal of Geophysical Research, 108(C6), 3185.

https://doi.org/10.1029/2002JC001549

Arévalo-Martínez, D. L., & Franco - Herrera, A. (2008). Características oceanográficas de la

surgencia frente a la ensenada de Gaira, Departamento de Magdalena, época seca menor de

2006. Boletín de Investigaciones Marinas y Costeras, 37(2), 131–162.

Arroyave Gómez, D. M., Gallego Suárez, D., Bartoli, M., & Toro-Botero, M. (2020). Spatial and

seasonal variability of sedimentary features and nitrogen benthic metabolism in a tropical

coastal area (Taganga Bay, Colombia Caribbean) impacted by a sewage outfall.

Biogeochemistry, 150(1), 85–107. https://doi.org/10.1007/s10533-020-00689-0

Barragán, G. R. G., Canosa, A., & Niño, J. P. (2009). Bacterioplancton en bahía Gaira, Mar

Caribe (Colombia): Comparación de la variabilidad en abundancia y biomasa bacteriana

durante diferentes períodos. Boletin de Investigaciones Marinas y Costeras, 38(2), 75–90.

https://doi.org/10.25268/bimc.invemar.2009.38.2.172

Barreto Hernández, A., & Velasco, L. A. (2014). Aislamiento y cultivo de microalgas bentónicas

del Caribe colombiano bajo diferentes condiciones de temperatura. Intropica, 9(May 2014),

23. https://doi.org/10.21676/23897864.1422

Bayraktarov, E., Bastidas-Salamanca, M., & Wild, C. (2014). The physical environment in coral


125

reefs of the Tayrona National Natural Park (Colombian Caribbean) in response to seasonal

upwelling. Boletín de Investigaciones Marinas y Costeras, 43(1), 137–157.

Bayraktarov, E., Pizarro, V., Eidens, C., & et al. (2013). Bleaching susceptibility and recovery of

Colombian Caribbean corals in response to water current exposure and seasonal upwelling.

PLoS ONE, 8(11), 1–11. https://doi.org/10.1371/journal.pone.0080536

Bayraktarov, E., & Wild, C. (2014). Spatiotemporal variability of sedimentary organic matter

supply and recycling processes in coral reefs of Tayrona National Natural Park, Colombian

Caribbean. Biogeosciences, 11(11), 2977–2990. https://doi.org/10.5194/bg-11-2977-2014

Bernard, R. J., Mortazavi, B., & Kleinhuizen, A. A. (2015). Dissimilatory nitrate reduction to

ammonium (DNRA) seasonally dominates NO3− reduction pathways in an

anthropogenically impacted sub-tropical coastal lagoon. Biogeochemistry, 125(1), 47–64.

https://doi.org/10.1007/s10533-015-0111-6

Brockmann, C., Doerffer, R., Peters, M., Stelzer, K., Embacher, S., & Ruescas, A. (2016).

Evolution of the C2RCC neural network for Sentinel 2 and 3 for the retrieval of ocean

colour products in normal and extreme optically complex waters. Living Planet Symposium,

740(54).

Bruce, L. C., Hamilton, D., Imberger, J., Gal, G., Gophen, M., Zohary, T., & Hambright, K. D.

(2006). A numerical simulation of the role of zooplankton in C, N and P cycling in Lake

Kinneret, Israel. Ecological Modelling, 193(3–4), 412–436.

https://doi.org/10.1016/j.ecolmodel.2005.09.008

Burd, B., Macdonald, T., & Bertold, S. (2013). The effects of wastewater effluent and river

discharge on benthic heterotrophic production, organic biomass and respiration in marine

coastal sediments. Marine Pollution Bulletin, 74(1), 351–363.


Burger, D. F., Hamilton, D. P., & Pilditch, C. A. (2008). Modelling the relative importance of

internal and external nutrient loads on water column nutrient concentrations and

phytoplankton biomass in a shallow polymictic lake. Ecological Modelling, 211(3–4), 411–

423. https://doi.org/10.1016/j.ecolmodel.2007.09.028

Capet, A., Meysman, F. J. R., Akoumianaki, I., Soetaert, K., & Grégoire, M. (2016). Integrating

sediment biogeochemistry into 3D oceanic models: A study of benthic-pelagic coupling in

the Black Sea. Ocean Modelling, 101, 83–100.

https://doi.org/10.1016/j.ocemod.2016.03.006


126

Capone, D. G., & Hutchins, D. A. (2013). Microbial biogeochemistry of coastal upwelling

regimes in a changing ocean. Nature Geoscience, 6(9), 711–717.

https://doi.org/10.1038/ngeo1916

Carrera, S., Velasco, L. A., & Barreto-Hernández, A. (2018). Potencial de microalgas bentónicas

del Mar Caribe como alimento en maricultura Potential of benthic microalgae of the

Caribbean Sea as food in mariculture. Revista de Biología Marina y Oceanografía, 53(3),

321–333.

Chassignet, E. P., Hurlburt, H. E., Smedstad, O. M., Halliwell, G. R., Hogan, P. J., Wallcraft, A.

J., Baraille, R., & Bleck, R. (2007). The HYCOM (HYbrid Coordinate Ocean Model) data

assimilative system. Journal of Marine Systems, 65(1-4 SPEC. ISS.), 60–83.

https://doi.org/10.1016/j.jmarsys.2005.09.016

Cloern, J. E., Foster, S. Q., & Kleckner, A. E. (2014). Phytoplankton primary production in the

world’s estuarine-coastal ecosystems. Biogeosciences, 11(9), 2477–2501.

https://doi.org/10.5194/bg-11-2477-2014

Corredor, J. E., Howarth, R. W., Twilley, R. R., & Morell, J. M. (1999). Nitrogen cycling and

anthropogenic impact in the tropical interamerican seas. Biogeochemistry, 46(1–3), 163–

178. https://doi.org/10.1007/BF01007578

Dalsgaard, T., & Krause-Jensen, D. (2006). Monitoring nutrient release from fish farms with

macroalgal and phytoplankton bioassays. Aquaculture, 256(1–4), 302–310.

https://doi.org/10.1016/j.aquaculture.2006.02.047

Davidson, K., Gowen, R. J., Harrison, P. J., Fleming, L. E., Hoagland, P., & Moschonas, G.

(2014). Anthropogenic nutrients and harmful algae in coastal waters. Journal of

Environmental Management, 146, 206–216. https://doi.org/10.1016/j.jenvman.2014.07.002

Delgado, A. L., Jamet, C., Loisel, H., Vantrepotte, V., Perillo, G. M. E., & Piccolo, M. C. (2014).

Evaluation of the MODIS-Aqua Sea-Surface Temperature product in the inner and mid-

shelves of southwest Buenos Aires Province, Argentina. International Journal of Remote

Sensing, 35(1), 306–320. https://doi.org/10.1080/01431161.2013.870680

Díaz-Rocca, L. H., & Causado-Rodríguez, E. (2007). La insostenibilidad del desarrollo urbano:

El caso de Santa Marta–Colombia. Clío América, 1(1), 64–100.

Diaz, J. M., Barrios, L. M., Cendales, M. A., & et al. (2000). Areas coralinas de Colombia: Vol.

publicacio (INVEMAR (ed.); Issue November).

Diaz, R J. (2001). Overview of hypoxia around the world. Journal of Environmental Quality,


127


Diaz, Robert J, & Rosenberg, R. (2008). Spreading dead zones and consequences for marine

ecosystems. Science (New York, N.Y.), 321(5891), 926–929.

https://doi.org/10.1126/science.1156401

Dong, L. F., Sobey, M. N., Smith, C. J., & et al. (2011). Dissimilatory reduction of nitrate to

ammonium, not denitrification or anammox, dominates benthic nitrate reduction in tropical

estuaries. Limnology and Oceanography, 56(1), 279–291.

https://doi.org/10.4319/lo.2011.56.1.0279

Downing, J. A., Mcclain, M., Twilley, R., Melack, J. M., Elser, J., Rabalais, N. N., Lewis, W. M.,

Turner, R. E., Corredor, J., Soto, D., Kopaska, J. A., & Howarth, R. W. (1999). The Impact

of Accelerating Land-Use Change on the N-Cycle of Tropical Aquatic Ecosystems : Current

Conditions and Projected Changes Source : Biogeochemistry , Vol . 46 , No . 1 / 3 , New

Perspectives on Nitrogen Recycling in the Temperate and Tropical Ame. Biogeochemistry,

46, 109–148. https://doi.org/10.1002/mds.25809

Eidens, C., Bayraktarov, E., Hauffe, T., Pizarro, V., Wilke, T., & Wild, C. (2014). Benthic

primary production in an upwelling-influenced coral reef, Colombian Caribbean. PeerJ,

2014(1), 1–22. https://doi.org/10.7717/peerj.554

Escobar, A. (1988). Estudio de algunos aspectos ecologicos y de la contaminacion bacteriana en

la Bahia de Santa Marta, Caribe Colombiano. Boletin de Investigaciones Marinas y


Espinosa, L., Ramírez, G., & Campos, N. (1995). Análisis de residuos organoclorados en los

sedimentos de manglar en la Ciénaga Grande de Santa Marta y la Bahía de Chengue, Caribe

colombiano. An. Inst. Invest. Mar. Punta Betín, 24, 79–94.

Fajardo, G. (1979). Surgencia costera en las proximidades de la península colombiana de La

Guajira. Boletin Cientifico CIOH, 2, 7–19.

Fox, J. (2005). The R Commander: A Basic-Statistics Graphical User Interface to R.

JournalofStatisticalSoftware, 14(9), 1–42.

Franco-Herrera, A., Castro, L., & Tigreros, P. (2006). Plankton dynamics in the south-central

Caribbean Sea: Strong seasonal changes in a coastal tropical system. Caribbean Journal of

Science, 42(1), 24–38.

Fung, T., Seymour, R. M., & Johnson, C. R. (2011). Alternative stable states and phase shifts in

coral reefs under anthropogenic stress. Ecology, 92(4), 967–982. https://doi.org/10.1890/10-


128

0378.1

Gal, G., Hipsey, M. R., Parparov, A., Wagner, U., Makler, V., & Zohary, T. (2009).

Implementation of ecological modeling as an effective management and investigation tool:

Lake Kinneret as a case study. Ecological Modelling, 220(13–14), 1697–1718.

https://doi.org/10.1016/j.ecolmodel.2009.04.010

Ganguly, D., Robin, R. S., Vardhan, K. V., Muduli, P. R., Abhilash, K. R., Patra, S., &

Subramanian, B. R. (2013). Variable response of two tropical phytoplankton species at

different salinity and nutrient condition. Journal of Experimental Marine Biology and

Ecology, 440, 244–249. https://doi.org/10.1016/j.jembe.2013.01.008

Garcés-Ordóñez, O., Arteaga, E., Obando, P., & et al. (2016). Atención a eventuales emergencias

ambientales en la zona marino-costera del departamento del Magdalena. Convenio

CORPAMAG-INVEMAR; código: PRY-CAM-011-14. Informe técnico final. (Issue 14).

García-Hoyos, L. M., Franco-Herrera, A., Ramire-Barón, J. S., & et al. (2010). Dinámica océano-

atmósfera y su influencia en la biomasa fitoplanctónica en la zona costera del epartamento

del Magdalena. Boletín de Investigaciones Marinas y Costeras, 39(2), 307–335.

García, F. (2013). Modelación de los efectos del emisario submarino de santa marta sobre la

calidad del agua. PhD Dissertation, Universidad de Antioquia, Facultad de Ingeniería.

García, F., PALACIO, C., & García, U. (2012). Distribución Vertical De Temperatura Y

Salinidad En El Área Costera De Santa Marta ( Colombia ) Vertical Distribution of

Temperature and Salinity At Santa Marta Coastal Area ( Colombia ). Dyna, 79(171), 232–

238.

García, M. J. L. (2020). SST comparison of AVHRR and MODIS time series in the Western

Mediterranean Sea. Remote Sensing, 12(14), 1–12. https://doi.org/10.3390/rs12142241

Gardner, W. S., McCarthy, M. J., An, S., & et al. (2006). Nitrogen fixation and dissimilatory

nitrate reduction to ammonium (DNRA) support nitrogen dynamics in Texas estuaries.

Limnology and Oceanography, 51(1 II), 558–568.

Ghanea, M., Moradi, M., Kabiri, K., & Mehdinia, A. (2016). Investigation and validation of

MODIS SST in the northern Persian Gulf. Advances in Space Research, 57(1), 127–136.

https://doi.org/10.1016/j.asr.2015.10.040

Glud, R. N., Grossart, H. P., Larsen, M., Tang, K. W., Arendt, K. E., Rysgaard, S., Thamdrup, B.,

& Nielsen, T. G. (2015). Copepod carcasses as microbial hot spots for pelagic

denitrification. Limnology and Oceanography, 60(6), 2026–2036.


129

https://doi.org/10.1002/lno.10149

Grall, J., & Chauvaud, L. (2002). Marine eutrophication and benthos: the need for new

approaches and concepts. Global Change Biology, 8(9), 813–830.

Griffin, S. L., Herzfeld, M., & Hamilton, D. P. (2001). Modelling the impact of zooplankton

grazing on phytoplankton biomass during a dinoflagellate bloom in the Swan River Estuary,

Western Australia. Ecological Engineering, 16(3), 373–394. https://doi.org/10.1016/S0925-

8574(00)00122-1

Griffiths, J. R., Kadin, M., Nascimento, F. J. A., Tamelander, T., Törnroos, A., Bonaglia, S.,

Bonsdorff, E., Brüchert, V., Gårdmark, A., Järnström, M., Kotta, J., Lindegren, M.,

Nordström, M. C., Norkko, A., Olsson, J., Weigel, B., Žydelis, R., Blenckner, T., Niiranen,

S., & Winder, M. (2017). The importance of benthic–pelagic coupling for marine ecosystem

functioning in a changing world. Global Change Biology, 23(6), 2179–2196.

https://doi.org/10.1111/gcb.13642

Hamilton, D. P., & Schladow, S. G. (1997). Prediction of water quality in lakes and reservoirs.

Part I - Model description. Ecological Modelling, 96(1–3), 91–110.

https://doi.org/10.1016/S0304-3800(96)00062-2

Henze, M., & Comeau, Y. (2008). Biological Wastewater Treatment: Principles Modelling and

Design. In M. C. M. van M. Henze & G. A. E. and D. B. Loosdrecht (Eds.), Biological

Wastewater Treatment: Principles Modelling and Design. IWA Publishing, London, UK.

Hernández, C. A., & Gocke, K. (1990). Productividad primaria en la Ciénaga Grande de Santa

Marta, Colombia. An. Inst. Invest. Mar. Punta Betín, 19(20), 101–119.

Hillmer, I., & Imberger, J. (2007). Influence of advection on scales of ecological studies in a

coastal equilibrium flow. Continental Shelf Research, 27(1), 134–153.

https://doi.org/10.1016/j.csr.2006.09.004

Hipsey, M., Romero, J., Antenucci, J., & Hamilton, D. (2012). Computational Aquatic Ecosystem

Dynamics Model: CAEDYM v3 v3.2 Science Manual (pp. 1–119). Centre for Water

Research University of Western Australia.

Hodges, B. (2000). Numerical Techniques in CWR-ELCOM (code release v.1). Centre for Water

Research, Univ. Western Australia.

Hodges, B., & Dallimore, C. (2016). Aquatic Ecosystem Model: AEM3D v1.0 User Manual.

Howard, M. D. A., Kudela, R. M., & McLaughlin, K. (2017). New insights into impacts of

anthropogenic nutrients on urban ecosystem processes on the Southern California coastal


130

shelf: Introduction and synthesis. Estuarine, Coastal and Shelf Science, 186, 163–170.

https://doi.org/10.1016/j.ecss.2016.06.028

Jørgensen, S. E., & Bendoricchio, G. (2001). Fundamentals of Ecological Modelling (Elsevier

(ed.)). http://library1.nida.ac.th/termpaper6/sd/2554/19755.pdf

Kozlov, I., Dailidiene, I., Korosov, A., Klemas, V., & Mingelaite, T. (2014). MODIS-based sea

surface temperature of the Baltic Sea Curonian Lagoon. Journal of Marine Systems,

129(October 2018), 157–165. https://doi.org/10.1016/j.jmarsys.2012.05.011

Leon, L. F., Smith, R. E. H., Malkin, S. Y., Depew, D., Hipsey, M. R., Antenucci, J. P., Higgins,

S. N., Hecky, R. E., & Rao, R. Y. (2012). Nested 3D modeling of the spatial dynamics of

nutrients and phytoplankton in a Lake Ontario nearshore zone. Journal of Great Lakes

Research, 38(SUPPL.4), 171–183. https://doi.org/10.1016/j.jglr.2012.02.006

Lorenzoni, L., Taylor, G. T., Benitez-Nelson, C., Hansell, D. A., Montes, E., Masserini, R.,

Fanning, K., Varela, R., Astor, Y., Guzmán, L., & Muller-Karger, F. E. (2013). Spatial and

seasonal variability of dissolved organic matter in the Cariaco Basin. Journal of

Geophysical Research: Biogeosciences, 118(2), 951–962.

https://doi.org/10.1002/jgrg.20075

Machado, D. a., & Imberger, J. (2012). Managing wastewater effluent to enhance aquatic

receiving ecosystem productivity: A coastal lagoon in Western Australia. Journal of

Environmental Management, 99, 52–60. https://doi.org/10.1016/j.jenvman.2011.12.020

Mancera-Pineda, J., Pinto, G., & Vilardy, S. (2013). Patrones de distribución estacional de masas

de agua en la Bahía de Santa Marta, Caribe Colombiano: Importancia relativa del upwelling

y outwelling. Boletín de Investigaciones Marinas y Costeras (Invemar), 42(2), 329–360.

Martin-Jézéquel, V., Hildebrand, M., & Brzezinski, M. A. (2000). Silicon metabolism in diatoms:

Implications for growth. Journal of Phycology, 36(5), 821–840.

https://doi.org/10.1046/j.1529-8817.2000.00019.x

Martiny, A. C., Vrugt, J. A., Primeau, F. W., & Lomas, M. W. (2013). Regional variation in the

particulate organic carbon to nitrogen ratio in the surface ocean. Global Biogeochemical

Cycles, 27(3), 723–731. https://doi.org/10.1002/gbc.20061

McCarthy, M. J., Newell, S. E., Carini, S. A., & et al. (2015). Denitrification Dominates

Sediment Nitrogen Removal and Is Enhanced by Bottom-Water Hypoxia in the Northern

Gulf of Mexico. Estuaries and Coasts, 38(6), 2279–2294. https://doi.org/10.1007/s12237-

015-9964-0


131

Miguel, L. L. A. J. (2018). Study on seasonal hydrology and biogeochemical variability in a

tropical estuarine system, Central Mozambique Coast, Africa. Marine Pollution Bulletin,

131(December 2017), 674–692. https://doi.org/10.1016/j.marpolbul.2018.04.071

Minnett, P. J., Evans, R. H., Kearns, E. J., & Brown, O. B. (2002). Sea-surface temperature

measured by the Moderate Resolution Imaging Spectroradiometer (MODIS). International

Geoscience and Remote Sensing Symposium (IGARSS), 2(C), 1177–1179.

https://doi.org/10.1109/igarss.2002.1025872

Moscarella, M. V., García, F., & Palacio, C. (2011). Microbiological water quality of Santa

Marta Bay , Colombia. Dyna, 167, 132–141.

Newell, S. E., McCarthy, M. J., Gardner, W. S., & et al. (2016). Sediment nitrogen fixation: a

call for re-evaluating coastal N budgets. Estuaries and Coasts, 39(6), 1626–1638.

https://doi.org/10.1007/s12237-016-0116-y

Nixon, S. W. (1995). Coastal marine eutrophication: A definition, social causes, and future

concerns. Ophelia, 41(1), 199–219. https://doi.org/10.1080/00785236.1995.10422044

Norström, A. V., Nyström, M., Lokrantz, J., & Folke, C. (2009). Alternative states on coral reefs:

Beyond coral-macroalgal phase shifts. Marine Ecology Progress Series, 376(Hatcher 1984),

293–306. https://doi.org/10.3354/meps07815

Ogashawara, I. (2019). The use of sentinel-3 imagery to monitor cyanobacterial blooms.

Environments - MDPI, 6(6). https://doi.org/10.3390/environments6060060

Özkundakci, D., Hamilton, D. P., & Trolle, D. (2011). Modelling the response of a highly

eutrophic lake to reductions in external and internal nutrient loading. New Zealand Journal

of Marine and Freshwater Research, 45(2), 165–185.

https://doi.org/10.1080/00288330.2010.548072

Paramo, J., Correa, M., & Núñez, S. (2011). Evidencias de desacople físico-biológico en el

sistema de surgencia en la Guajira, caribe Colombiano. Revista de Biologia Marina y

Oceanografia, 46(3), 421–430. https://doi.org/10.4067/S0718-19572011000300011

Paraska, D. W., Hipsey, M. R., & Salmon, S. U. (2014). Sediment diagenesis models: Review of

approaches, challenges and opportunities. Environmental Modelling & Software, 61, 297–

325. https://doi.org/10.1016/j.envsoft.2014.05.011

Ramírez-Barón, J. S., Franco-Herrera, A., García-Hoyos, L. M., & López-Cerón, D. A. (2010).

La comunidad fitoplanctónica durante eventos de surgencia y no surgencia, en la Zona

Costera del Departamento del Magdalena, Caribe colombiano. Boletín de Investigaciones


132

Marinas y Costeras, 39(2), 233–263.

Ramírez, G. (1981). Características Fisico-Químicas de la Bahía de Santa Marta (Agosto 1980-

Jilio 1981). Boletín de Investigaciones Marinas y Costeras, 13, 111–121.

Ramirez, M. A. (2017). Maximiliano Arredondo Ramirez.

Ramos-Ortega, L. M., Vidal, L. A., Vilardy, S., & et al. (2008). Análisis de la contaminación

microbiológica (coliformes totales y fecales) en la bahía de Santa Marta, caribe colombiano.

Acta Biológica Colombiana, 13(3), 87–98.

Reopanichkul, P., Schlacher, T. A., Carter, R. W., & Worachananant, S. (2009). Sewage impacts

coral reefs at multiple levels of ecological organization. Marine Pollution Bulletin, 58(9),

1356–1362. https://doi.org/10.1016/j.marpolbul.2009.04.024

Robson, B. J., & Hamilton, D. P. (2004). Three-dimensional modelling of a Microcystis bloom

event in the Swan River estuary, Western Australia. Ecological Modelling, 174(1–2), 203–

222. https://doi.org/10.1016/j.ecolmodel.2004.01.006

Robson, Barbara J., & Hamilton, D. P. (2003). Summer flow event induces a cyanobacterial

bloom in a seasonal Western Australian estuary. Marine and Freshwater Research, 54(2),

139–151. https://doi.org/10.1071/MF02090

Romero, J. R., Antenucci, J. P., & Imberger, J. (2004). One- and three-dimensional

biogeochemical simulations of two differing reservoirs. Ecological Modelling, 174(1–2),

143–160. https://doi.org/10.1016/j.ecolmodel.2004.01.005

Rueda-Roa, D. T., & Muller-Karger, F. E. (2013). The southern Caribbean upwelling system :

Sea surface temperature , wind forcing and chlorophyll concentration patterns. Deep-Sea

Research Part I, 78, 102–114. https://doi.org/10.1016/j.dsr.2013.04.008

Rueda-Roa, Digna T., & Muller-Karger, F. E. (2013). The southern Caribbean upwelling system:

Sea surface temperature, wind forcing and chlorophyll concentration patterns. Deep-Sea

Research Part I: Oceanographic Research Papers. https://doi.org/10.1016/j.dsr.2013.04.008

Ruescas, A. B., Mateo-García, G., Camps-Valls, G., & Hieronymi, M. (2018). Retrieval of case 2

water quality parameters with machine learning. International Geoscience and Remote

Sensing Symposium (IGARSS), 2018-July, 124–127.

https://doi.org/10.1109/IGARSS.2018.8518810

Salzwedel, H. and Müller, K. (1983). A summary of Meteorological and hydrological data from

the Bay of Santa Marta, Colombian Caribbean. Boletin de Investigaciones Marinas y



133

Seip, K. L., & Reynolds, C. S. (1995). Phytoplankton functional attributes along trophic gradient

and season. Limnology and Oceanography, 40(3), 589–597.

https://doi.org/10.4319/lo.1995.40.3.0589

Smith, J., Burford, M. A., Revill, A. T., Haese, R. R., & Fortune, J. (2012). Effect of nutrient

loading on biogeochemical processes in tropical tidal creeks. Biogeochemistry, 108(1–3),

359–380. https://doi.org/10.1007/s10533-011-9605-z

Soetaert, K., Middelburg, J. J., Herman, P. M. J., & Buis, K. (2000). On the coupling of benthic

and pelagic biogeochemical models. Earth Science Reviews, 51(1–4), 173–201.

https://doi.org/10.1016/S0012-8252(00)00004-0

Spillman, C. M., Imberger, J., Hamilton, D. P., Hipsey, M. R., & Romero, J. R. (2007).

Modelling the effects of Po River discharge, internal nutrient cycling and hydrodynamics on

biogeochemistry of the Northern Adriatic Sea. Journal of Marine Systems, 68(1–2), 167–

200. https://doi.org/10.1016/j.jmarsys.2006.11.006

Toming, K., Kutser, T., Uiboupin, R., Arikas, A., Vahter, K., & Paavel, B. (2017). Mapping

water quality parameters with Sentinel-3 Ocean and Land Colour Instrument imagery in the

Baltic Sea. Remote Sensing, 9(10). https://doi.org/10.3390/rs9101070

Vega-Sequeda, J., Rodríguez-Ramírez, A., Reyes-Nivia, M. C., & et al. (2008). Formaciones

Coralinas Del Área De Santa Marta: Estado Y Patrones De Distribución Espacial De La

Comunidad Bentonica. Boletin de Investigaciones Marinas y Costeras, 37(2), 87–105.

Wilson, R. F., Fennel, K., & Paul Mattern, J. (2013). Simulating sediment-water exchange of

nutrients and oxygen: A comparative assessment of models against mesocosm observations.

Continental Shelf Research, 63, 69–84. https://doi.org/10.1016/j.csr.2013.05.003

Woodward, B. L., Marti, C. L., Imberger, J., Hipsey, M. R., & Oldham, C. E. (2017). Wind and

buoyancy driven horizontal exchange in shallow embayments of a tropical reservoir: Lake

Argyle, Western Australia. Limnology and Oceanography, 62(4), 1636–1657.

https://doi.org/10.1002/lno.10522

Zuhlke, M., Fomferra, N., Brockmann, C., Peters, M., Veci, L., Malik, J., & Regner, P. (2015).

SNAP (sentinel application platform) and the ESA sentinel 3 toolbox. Sentinel-3 for Science

Workshop, 734. http://library1.nida.ac.th/termpaper6/sd/2554/19755.pdf

134

Chapter 4

Conclusions and recommendations

4.1. Research summary

This thesis had as main aim the understanding of biogeochemical processes in both benthic and

pelagic compartments in the Santa Marta Coastal Area (SMCA), a tropical coastal ecosystem

affected by a seasonal upwelling and an untreated wastewater effluent from a submarine outfall.

The spatial and temporal variability of biogeochemical conditions was analyzed through high-

standard experimental measurements of benthic metabolism and sedimentary features (e.g.

denitrification, dissimilative nitrate reduction to ammonium (DNRA), C and N stable isotopes) and

the calibration and validation of a coupled 3D hydrodynamic - ecological model using temperature

and chlorophyll-a satellite remote sensing in two periods during non-upwelling (NUPW) and

upwelling season (UPW). The results indicated:

The interaction of effects between the sewage outfall and seasonal upwelling were

determined by metabolic rates and sedimentary features in the benthic environment, while

the ecological model captured the effect of fertilization from both sources in the outfall

plume due to the increase in chlorophyll-a. Similarly, the model reproduced the increase of

nitrate and salinity, the decrease of temperature and a slightly decline of dissolved oxygen

during seasonal upwelling. The high inputs of dissolved and particulate organic matter have

a high potential to impact the benthic more than the pelagic system near the outfall. In fact,

high sedimentation rates of organic matter decouple carbon input and mineralization rates

and result in the accumulation of sulfides in pore water and the development of Beggiatoa

mats on the sediment surface. Nutrient accumulation and increased primary production are

more difficult to catch in a highly dispersive and mixed coastal system.

Chapter 4. Conclusions and recommendations

135

The high level of δ15N fractionation in sediment stressed the high complexity of nitrogen

cycling, which is mediated by different microbial processes, each with a specific regulation

and differentially impacted by anthropogenic pressures. Experimental results suggest

however that sulphides depress denitrification and enhance nitrate ammonification and in

general N recycling from sediments.

The combination of several approaches (e.g. experimental, modeling and remote sensing)

allowed determining the environmental impact of a sewage outfall in a coastal marine

ecosystem with a highly variable and complex hydrodynamics. In this study, a spatial

impact on sediments around 1800 m was estimated using stable carbon isotopes whereas

Chl-a satellite images traced an affected area of about 1500 m on a day characterized by

high residence time during the simulation period. In addition, stimulation of phytoplankton

growth in the outfall plume during short time periods was estimated by the model during

the UPW season.

The results of the model suggest a synergy between mixing – driven changes of

temperature, light and nutrients on the phytoplankton growth. The increase of the euphotic

zone, that was an indirect effect of higher water mixing during the seasonal coastal

upwelling was accompanied by short-term stimulation of phytoplankton growth, also

sustained from the outfall nutrients supply and by lower temperatures, closer to the optimal

growth temperatures.

Tropical coastal ecosystems with multiple organic matter inputs such as SMCA have a

lower buffering capacity with increasing organic matter loads due to the predominance of

DNRA over denitrification. This is a fundamental concern about the key biogeochemical

and ecological services in this area (e.g. the presence of corals and fishes as well as the

denitrification, P-retention or sulphide buffer capacity of sediments, etc.).

The approaches and findings contribute to a better understanding of biogeochemical

processes under the influences of anthropogenic nutrient inputs and seasonal upwelling in

the study area and in other understudied tropical coastal regions.

Chapter 4. Conclusions and recommendations

136

4.2. Recommendations for future work

The following is a list of some recommendations for further work arising from this thesis.

As the biogeochemical and transport processes in SMCA are highly influenced by the local

and regional meteorological conditions, research into the effects of these changes could

yield interesting results and insights. Changes at three levels could be analyzed: (i) Regional

climatic fluctuations caused by El Niño-Southern Oscillation (ENSO) cycle. La Niña years

refer to the cold phase of ENSO while El Niño years refer to the warm phase of ENSO. (ii)

The frequency and intensity of extreme events like tropical storms. (iii) A long-term scale

due to climate change.

The nutrients input in the SMCA from multiple sources includes those derived by sewage

outfall, rivers, harbor, seasonal runoff and upwelling and Ciénaga Grande de Santa Marta

(CGSM) and Río Magdalena. Therefore, the understanding of major shifts and budget of

nutrients, as well as of the cumulative effect of sediment transport, requires the application

of different approaches such as field monitoring, benthic metabolism measurements

(denitrification, DNRA, N2 fixation, etc), micro and macroalgae bioassays to analyze

primary production rates and numerical modeling (e.g. depth-integrated diagenetic models,

diagenetic model, etc).

As the DNRA and heterotrophic N2 fixation have possibly a higher potential importance in

coastal ecosystem to N budget and productivity at regional and global scales, research of

these processes should be oriented to improve their inclusion in ecological models.

The analysis of stable isotopes in dissolved nitrate (δ15NNO3 and δ18ONO3) and suspended

particulate matter (δ15NPN and δ13CPN) can be used to trace dissolved nitrate, ammonium

and particulate organic matter, as well as to determine rates of nitrification of ammonium

and the dominant N forms in the sewage effluent from outfall and other anthropogenic

nutrient sources. This could contribute to improve understanding the fate and cycling of

anthropogenic nitrogen in the water column in NUPW and UPW seasons in SMCA.

137

5. Final Appendices

5.1. Incubation and stirring system for large sediments cores

In the project that funded my thesis, I worked with PhD Fabio Alexander Suarez of company “Faro

Tecnologico” in the design and setting up of an incubation and stirring system for large sediments

cores (Fig 5S.1). This system was field tested. Finally, data of these sediments cores were not

included in the final results due to the lack of representation (few cores). However, this was a good

research exercise for the future work.

(a) Incubation and stirring system for transparent plexiglass liners (internal diameter = 15 cm and length = 60 cm)

(b) Silicone – coated magnetic stirrers located on the top covers

(c) System of sediment cores sampling and fixing the bottom cover

(d) Bottom water samplers

Fig. 5S.1. Incubation and stirring system for large sediments cores

External magnet rotated by a motor at 30 rpm

138

5.2. Complementary work

Co-author in the paper:

Paula Carpintero Moraes, Diana Marcela Arroyave Gómez, Fabio Vincenzi, Giuseppe Castaldelli,

Elisa Anna Fano, Marco Bartoli and Sara Benelli. 2019. Analysis of 15NO3- via anoxic slurries

coupled to MIMS analysis: an application to estimate nitrification by burrowing macrofauna.

Water, 11, 2310; doi:10.3390/w11112310

Paper available in the following website: https://www.mdpi.com/2073-4441/11/11/2310

Documents

Seasonal patterns of biogeochemical conditions of the