Climate Change Impacts on the Catchment Contribution to ...uu.diva-portal.org/smash/get/diva2:170783/FULLTEXT01.pdfFigure 1. Lake Mälaren, Sweden. A synthesis paper (Paper IV) features

ACTA

UNIVERSITATIS

UPSALIENSIS

UPPSALA

2007

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 347

Climate Change Impacts on theCatchment Contribution to LakeWater Quantity and Quality

KAREN MOORE

ISSN 1651-6214ISBN 978-91-554-6980-1urn:nbn:se:uu:diva-8236

�� !"� !##$ �� #%##&� �� & �� & '��( �� )�� (

��

*�� +( !##$( ,�� ,�� -�� ,�� ,�� .�/� 0��1�� 1��( 2�� ( �� 34$( 3" ��( ��(-56� 7$�87�8""4897�#8�(

2 /�� :�� ;�� ) )�� :�� < �� =0.>� � �� & �� )�� :�� :��(?�� )�� &� �� )�� )�� ( >�

� �� ) @�/� 52 �� )�� &�)� �� &� � �� &�� ( >� �� =�� & .�/� *�� 5)�� /�� )�� &� �� &�� ( '��/ ��&& �� )�� (2 �� & �� &� �� A�-�B �� &� &��

�� )�� & ��/ �-� �� &�� &�� ( - �� )�� )��&�) �� ) ��/ �� &&( - )�� -� �� )�� (�� &� � )�� &�� )�� &� �� =�� &

.�/� *�� =0.> �� /� ��/� �� /� ��(2 �� &�� & �� &�� ( �� ;�� /� �� &�� /� ��(.�� & ��

A,� *B &�� &� �� A� ,B�� C�� &� �� & �� /��(

� �� ,�� )�� :�� .�/� *�� =0.>� ��

�� !�� "�#�� $%� �� &'($)* ��

D +�� *�� !##$

-55� �9"�89!�4-56� 7$�87�8""4897�#8��%�%��%��%��8�!39 A��%EE��(/�(��E��<��F��%�%��%��%��8�!39B

To my family, with love

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals: I Moore, K., Rosenzweig, C., Goldberg, R., Pierson, D.C., Pettersson, K., Schneiderman, E.M., Zion, M.S. and Lounsbury, D.G. (2006). Impacts of projected climate change on phosphorus and sediment loadings for a New York City water supply reservoir. Verhandlungen

Internationale Vereinigung für theoretische und angewandte Limnologie 29 (4): 1829-1832. II Moore, K., Pierson, D., Pettersson, K., Schneiderman, E., and

Samuelsson, P. Effects of warmer world scenarios on hydro-logic inputs to Lake Mälaren, Sweden and implications for nutrient loads. (Accepted by Hydrobiologia)

III Markensten, H., Moore, K., and Persson, I. Lake phytoplankton response to a future climate scenario – a model approach. (Submitted) IV Moore, K., Jennings, E., May, L., Järvinen, Arvola, L.,

Tamm, T., Järvet, A., Nõges, T., Pierson, D., and Schneider-man, E. Modelling the Effects of Climate Change on Inorganic Nitrogen Transport from Catchments to Lakes. (manuscript)

V Moore, K., Pierson, D., Pettersson, K., and Weyhenmeyer, G. Temporal variation in colored dissolved organic matter (CDOM) fluorescence in the Swedish river Hedströmmen. (manuscript) Paper I is reproduced with the kind permission from the publisher.

Contents

Introduction.....................................................................................................9

Aims of the thesis..........................................................................................10

Methods ........................................................................................................11 Study sites............................................................................................11 Catchment model selection..................................................................13 Model overview...................................................................................13 Main model features ............................................................................14 Data requirements................................................................................15 Catchment model calibration and validation .......................................16 Climate model forcings........................................................................16 Model coupling for lake response........................................................18 Methods development for DOC monitoring and model calibration ....18

Results and Discussion .................................................................................21 Catchment responses to a warmer climate...........................................21 Evaluating the catchment contribution to lake productivity for a warmer scenario...................................................................................24 Future work..........................................................................................24

Conclusions and Perspectives .......................................................................26

Effekter av klimatförändringar på vattenflöde till sjöar och deras vattenkvalitet (Swedish summary)................................................................28

Acknowledgements.......................................................................................30

References.....................................................................................................32

Abbreviations

CDOM Chromophoric (or colored) dissolvedorganic matter

DIN Dissolved inorganic nitrogen DIP Dissolved inorganic phosphorus DOC Dissolved organic carbon GCM General circulation model GIS Geographic Information System GWLF Generalized Watershed Loading

Functions model IPCC Intergovernmental Panel on Climate

Change N Nitrogen NYCDEP New York City Department of En-

vironmental Conservation P Phosphorus PROBE PROgram for Boundary layers in the

Environment, physical lake model PROTBAS PROTech Based Algal Simulations,

phytoplankton model RCAO Rossby Centre Atmosphere-Ocean

coupled regional climate model RCM Regional Climate Model Si Silica SMHI Swedish Meteorological and Hydro-

logical Institute SRES IPCC Special Report on Emissions

Scenarios VENSIM Ventana Simulation Environment

modelling software

9

Introduction

Increases in greenhouse gas concentrations since about 1750 have perturbed the earth-atmosphere system, leading to surface warming and other changes in climate (IPCC, 1995). During the 20th century, changes in precipitation patterns and a warming trend have been observed at what appears to be an unprecedented rate (Stott et al., 2000). On a regional scale, annual precipita-tion in northern Europe increased by 10 to 40% and decreased about 20% in southern Europe, while average annual air temperature increased by about 0.8 C in Europe overall during the 20th century (IPCC, 2001). A key ques-tion in light of climate change projections is how will aquatic systems re-spond to changes in driving variables such as temperature and precipitation?

Eutrophication has been a perennial concern for water managers (Janus and Vollenweider, 1981). Apart from cultural influences on eutrophication, there is some indication that warming periods and climate oscillations are linked to increased occurrences of harmful algal blooms (HABs) (Edwards et al., 2006). Another concern is that carbon flux will increase with climate warming. Increases in the release of humic substances from catchment soils may affect water color and lake metabolism. Aside from looking at patterns in historical data, an alternative approach is to model lake and river re-sponses to climate drivers and test the sensitivity of the model system to changes in driving variables such as precipitation and temperature.

Simulation models are useful tools for exploring impacts of climate change on water quantity and quality. While it is impossible to predict the future, it is reasonable to make projections using current knowledge. Climate change projections are based on plausible assumptions of economic and po-pulation growth to estimate human impacts on greenhouse gas emissions and their effects on the climate system. This thesis evaluates the results of using warmer world scenarios to drive GWLF, a simple catchment-scale model that simulates water quantity on a daily time step coupled with nutrient load-ing functions to simulate monthly and annual nitrogen, phosphorus, and silica fluxes. Additionally, a method for obtaining high-frequency data that can be related to carbon flux is described.

10

Aims of the thesis

The focus of this thesis is to investigate climate influences on catchment hydrology and nutrient transport and the possible consequences for lake pro-ductivity. The aims of this thesis were to: 1. Examine the magnitude and direction of hydrological and nutrient load-

ing changes with projections of climate warming at the catchment scale. 2. Evaluate the catchment contribution to lake phytoplankton growth and

succession for a warmer world scenario. 3. Compare regional changes in dissolved inorganic nitrogen export for

multiple catchments and climate scenarios. 4. Develop an approach to monitoring of organic matter inputs that charac-

terizes its dynamic nature and provides appropriate data for calibrating a dissolved organic carbon model.

11

Methods

Study sites

The first site for water quantity and quality modelling (Paper I) was the West Branch of the Delaware River (catchment area 912 km2), the principal inflow to Cannonsville Reservoir (42º N 30º W) in upstate New York. This site was selected because it is part of the New York City Water Supply sys-tem and implications of climate change are of keen interest for drinking wa-ter concerns and maintenance of river flows for aquatic habitat conservation.

The remaining work in this thesis (Papers II, III, IV, and V) includes rivers that flow into the Galten basin of Lake Mälaren, Sweden (Figure 1). Lake Mälaren is the third largest lake in Sweden (Willén, 2001), and is the drinking water supply for Stockholm. Lake Mälaren has a complex hydro-graphy; Galten is one of five lake sub-basins and and has four principal river inflows (Arbogaån, Hedströmmen, Köpingsån, and Kolbäcksån, all featured in Paper II) located at the western end of the Norrström drainage. The Gal-ten catchment represents about 38% of the total catchment area for the Norr-ström drainage basin of Lake Mälaren (SMHI, 1993). The lake area of Gal-ten is 61.2 km2, with a large catchment (8508 km2). Given the large catch-ment to lake area ratio (139:1), catchment influences are likely to have a strong impact on lake function (Paper III). The principal land cover is forest dominated by coniferous species (GIS data source: Lantmäteriet, Sweden). Galten is a eutrophic shallow basin with a short water retention time (Willén, 2001). Galten was selected as test case in part because of its history of algal blooms. Phosphorus reductions in the 1960s and 1970s resulted in water quality improvements (Wallin, 2000), but Galten basin remains eutrophic, with an average total phosphorus concentration of 50 μg L-1 (Willén, 2001). The strong influence of the catchment on lake water quality makes it a good candidate for studying the connection of climate impacts on water quantity and quality.

12

Figure 1. Lake Mälaren, Sweden.

A synthesis paper (Paper IV) features dissolved inorganic nitrogen model-ing results from the Arbogaån river inflow to Galten (Figure 1) along with other sites in northern and western Europe (Table 1). Methods development for DOC monitoring in preparation for future modeling (Paper V) was done on the river Hedströmmen near its inflow to Galten (Figure 1).

Table 1 Sites for dissolved inorganic nitrogen (DIN) modelling results (Paper IV).

Country Lake Basin Latitude

Longitude Sub-

catchment name

Sub-catchment

area (km2)

Modeler

United Kingdom

(UK)

Esthwaite Water

54.4º N 2.9º W Esthwaite 15.6 Linda May

Ireland (IE)

Lough Leane

52.03º N 9.55º W Flesk 332 Eleanor Jennings

Finland (FI) Pääjärvi 61.04º N

25.08º E Mustajoki 76.8 Marko Järvinen

Sweden (SE)

Mälaren, Galten basin

59.45º N 16.19º E Arbogaån 3808 Karen Moore

Estonia (EE) Võrtsjarv 58.08º N

25.92º E Tarvastu 96.5 Toomas Tamm

13

Catchment model selection

There are several models that are employed for hydrological and water qual-ity studies that would have been suitable for this examination of potential climate change impacts. The HBV hydrological model (Bergström, 1995) and related nutrient models such as HBV-N (Arheimer and Brandt, 1998) are among the obvious choices. A comprehensive evaluation of water quality models assembled for the EUROHARP project gives a description of nine model choices that are in common use (Schoumans and Silgram, 2003; Sil-gram and Schoumans, 2004). The GWLF model presented here is in current use in North America and Europe in different platforms under different names with modifications (BasinSim: Dai et al., 2000; GWLFXL: Hong and Swaney, 2004; CSIM: Mörth et al., 2007). It might be considered a “fast and frugal” model (Carpenter, 2003) for getting a first approximation of climate impacts on heterogenous catchments spanning a range of sizes.

GWLF (the Generalized Watershed Loading Functions) model (Haith et al., 1992) was selected because it provides a compromise between simple empirical export coefficients and complex chemical simulations models. The underlying assumption is that water and sediment movement give an ap-proximation of mechanisms for nutrient transport at the catchment scale. The original use of the model was as a screening tool to evaluate non-point sour-ce loadings from mixed agricultural watersheds. GWLF has also been used for setting regulatory guidelines and projecting potential change under dif-ferent management and land use scenarios. Model application can be ex-tended to evaluate climate change when land use and management are held constant.

Model overview

GWLF was developed by Haith and Tubbs (1981) and validated by Haith and Shoemaker (1987) to simulate monthly dissolved and total phosphorus and nitrogen loads in streamflow. The original model was written in Quick-BASIC v. 4.5 for use with the MS DOS operating system. It is available at no cost from the principal author (Dr. Douglas Haith, Cornell University, Ithaca, New York, USA). The GWLF User’s Manual provides model up-dates in nutrient loads from septic systems and modifications to the urban runoff and groundwater portions of the model (Haith et al., 1992). Paper I uses the model version described by Schneiderman et al., 2002. Several model modifications were made during the course of 10 years of model ap-plication. Model revisions included: addition of unsaturated leakage between unsaturated and saturated subsurface reservoirs; revised timing of sediment

14

export; inclusion of urban sediments and dissolved nutrients; tracking of particulate nutrients from point sources; and revised timing of septic system loads (Schneiderman et al., 2002). Papers II, III, and IV use model ver-sions also created by the Water Quality Modeling group of the New York City Department of Environmental Protection (NYCDEP) for the European Union CLIME project (Climate and Lake Impacts in Europe). In all cases, GWLF was written in Vensim, a visual modeling software package (Ventana Systems, Inc.). One advantage of using Vensim software is that it makes model conceptualization, modification, and testing easy to perform and communicate using a graphical interface. This allowed the model to be read-ily adapted and applied to the European watersheds in the CLIME project. The Vensim interface also facilitated the calibration of the model in order to obtain an optimal correspondence between the simulated and measured nu-trient export rates.

Main model features

The primary features of the GWLF model that originated with Haith and colleagues at Cornell University (Haith and Tubbs, 1981; Haith and Shoe-maker, 1987) remain unchanged. The model is driven by daily temperature and precipitation data and water balances are calculated on a daily interval. Streamflow consists of runoff and groundwater discharge components. Mod-elling of diffuse nutrient sources is a function of land use and runoff. Dis-solved nutrient loads are derived by multiplying runoff by a land use-specific nutrient concentration, and agricultural practices such as manure spreading can also be taken into account in the model. Runoff is calculated using the SCS curve number method (Ogrosky and Mockus, 1964). These curve num-bers are based on soil hydrologic groups that are classified on the basis of their characteristic infiltration rates. Runoff curve numbers for different rural and urban land use categories have been estimated as a function of land use and soil hydrologic group for the USA. (SCS, 1986), and these curve num-bers are further adjusted as a function of antecedent moisture (Haith, 1985). Sediment and particulate nutrient loss are based on USLE, the Universal Soil Loss Equation (Wischmeier and Smith, 1978), and estimates of soil nutrient concentration. The contribution of nutrients from septic systems is based on population density and estimates of system performance. Point source load-ings (i.e. treated sewage) are also accounted for in the overall estimates of nutrient flux. Nutrient loads can be summed by month, season, or year. An overview of the main elements of the model is shown in Fig. 2. For this the-sis, dissolved nutrients are the focus, with the exception of Paper I, where particulate phosphorus is included. Additionally, the Swedish sites (Papers II, III, IV) include inputs from septic systems as a point source, using litera-ture values from the TRK study (Brandt and Ejhed, 2002), rather than using

15

estimates of population and septic system performance. This was done to simplify data requirements by relying on a trusted data source and reducing assumptions about the functional status of septic systems.

Figure 2. Conceptual overview of GWLF model components (source: E. Schneiderman).

Data requirements

Data requirements include spatial data for land use and soils, ideally in GIS format. Spatial data for Paper I were assembled by NYCDEP (see Schnei-derman et al., 2002). Land use data for Sweden were obtained from the Na-tional Land Survey, Lantmäteriet. Data were reclassified to match the CORINE level 3 classification system (European Commission CooRdinate Information on the Environment) used at the sites discussed in Paper IV. Soils data for Paper I were obtained from NYCDEP; the source was the Natural Resource Conservation Service (NRCS). Swedish soils data for Pa-pers II, II, and IV were obtained from the Geological Survey of Sweden, Sveriges Geologiska Undersökning (SGU). Data for other sites discussed in

16

Paper IV were obtained by local modeling experts (Table 1) and sources are reported in Järvet et al., 2006.

Meteorological data (daily precipitation and air temperature) for Paper I was obtained from NYCDEP. Swedish weather and river discharge data for Papers II, III, and IV were obtained from the Swedish Meteorological and Hydrological Institute (SMHI). Data for other sites discussed in Paper IV were obtained by local modeling experts (Table 1) and sources are reported in Järvet et al., 2006.

Land use-specific nutrient concentrations, point sources, and values for septic system inputs for Paper I were collated by NYCDEP. Swedish nutri-ent data for Papers II, III, and IV were derived from published results of the TRK study (Transport, Retention och Källfördelning, Brandt and Ejhed, 2002); nutrient concentrations for agricultural land were from the JRK study (Typområden på jordbruksmark, Kyllmar and Johnsson, 1998). Nutrient concentrations for baseflow were estimated from observations of river water quality averaged for flows at or below the 20th percentile for long-term river discharge records from SMHI corresponding to water quality measurements from the Swedish Agricultural University (Sveriges lantbruksuniversitet, SLU) for Papers I, II, III, and IV. Water quality data for Sweden were primarily from SLU (http://info1.ma.slu.se), with the exception of Paper V that used data from Uppsala University and Norrvatten.

Catchment model calibration and validation

The catchment model was calibrated and validated using long-term historical data for river discharge and nutrients (N, P, Si) (Papers II and III). To evaluate model performance for Papers II, III, and IV, the Nash-Sutcliffe statistic was used to test for differences between GWLF model output and measured values (Nash and Sutcliffe, 1970). A value of one indicates a per-fect correspondence between measured and modeled values. In Paper III, the root mean square error (RMSE) was used to validate the performance of the PROBE model (Mayer and Butler, 1993).

Climate model forcings

Model output from the Goddard Institute for Space Studies (GISS) general circulation model (GCM) (Russell et al., 1995) was used directly to drive the GWLF model in Paper I. For all sites discussed in Papers II, III, and IV, data were provided by the Rossby Centre of SMHI, Norrköping. Climate forcings for these papers were taken from the Rossby Centre Atmosphere-Ocean (RCAO) regional climate model and the Hadley Centre regional cli-mate model (HadRM3p), both of which were driven using boundary condi-

17

tions from two general circulation models: the Max Planck Institute ECHAM4/OPYC3 and the Hadley Centre HadAM3H for two IPCC emis-sions scenarios (A2, B2) (Samuelsson, 2004; Table 2). A 30-year “time slice” was used for a control period (1961-1990) and future period (2071-2100) for Papers II and IV. Paper III used a 21-year time slice (1970-1990 for the observed period). For both A2 and B2 scenarios, increases in popula-tion growth are projected, but the rate of increase is greater for scenario A2 (IPCC, 2001). The use of multiple models allows for some consideration of the range of uncertainty in model representation of the climate system.

One method used to downscale climate forcings from the regional to the local catchment scale is based on monthly average changes in precipitation and temperature, known as the “delta change” approach. The approacch was used in its simplest form as described by Hay et al., (2000) in Papers II, III, and IV (SE, FI). For other sites in Paper IV (UK, IE, EE), a weather gene-rator (Jones and Salmon, 1995) was used in a stochastic model to generate sets of daily weather variables for future scenarios (work performed by the Climate Research Unit, East Anglia, UK; Watts et al., 2004).

Single 30-year time series for the historical period (1961-1990) and future scenarios (2071-2100) for cases in which the weather generator was not ap-plied were then resampled by month to create multiple (100) recombinations of daily weather timeseries. The result was 100 30-year synthetic weather data sets that were used to drive GWLF. All summary statistics for Papers II and IV are based on these multiple iterations that were intended to more fully explore the distributional properties of model outputs.

Table 2. Climate models and scenarios used for future projections of climate change.

Abbreviation General Circulation Model (GCM)

Regional Climate Model (RCM)

SRES Scenario

(IPCC, 2001)

E A2 ECHAM4/OPYC3 RCAO A2

E B2 ECHAM4/OPYC3 RCAO B2

H A2 HadAM3H RCAO A2

H B2 HadAM3H RCAO B2

Had A2 HadAM3P Had RM3p A2

Had B2 HadAM3P Had RM3p B2

18

Model coupling for lake response

The PROTBAS (PROTech Based Algal Simulations, Markensten and Pier-son, 2007) was used to simulate the response of lake phytoplankton growth and succession. Physical lake parameters for thermal structure were from the PROBE (PROgram for Boundary layers in the Environment) model (Svens-son, 1998). A schematic diagram of model coupling is shown in Figure 3. Paper III employs output from one RCM (Regional Climate Model) to force the catchment model (GWLF), the physical lake model (PROBE), and the biological response model (PROTBAS). Paper V provides the data for mak-ing addition of a dissolved organic carbon (DOC) model feasible in the fu-ture.

Figure 3. Conceptual overview of model coupling (source: D. Pierson).

Methods development for DOC monitoring and model calibration

Paper V describes a field monitoring station that uses a fixed-wavelength fluorometer (WetLabs WetStar ex/em 370/460 nm) as the main tool for col-lecting high-frequency data on colored or chromophoric dissolved organic matter (CDOM). Hourly measurements of CDOM, river stage height, and

19

river temperature are recorded with a Campbell CR10X data logger (Camp-bell Scientific Instruments) at the Dömsta gauge (SMHI 61-2219) on river Hedströmmen. A GSM modem allows for remote collection of data. Grab samples analyzed for DOC, turbidity, and filtered absorbance (420 nm), and water color (Pt-Co Hazen scale) were also collected to examine the relation-ships between CDOM, dissolved organic carbon, and water color.

20

21

Results and Discussion

Catchment responses to a warmer climate

For four Swedish rivers and six future climate scenarios, the greatest change in streamflow resulted from increased winter precipitation in the form of rain rather than snow, reduced snow accumulation, and decreased spring snow melt (Paper II). This translated to a change in the timing of peak river dis-charge, with increased winter flows and a reduction in the spring melt peak. A similar shift is seen in nutrient loadings, with decreased loading during the period from spring to autumn and increased loading in the winter months. Annual dissolved inorganic nitrogen (DIN) and phosphorus (DIP) loads in-creased for four out of six future scenarios, with the exception of the HadRM3p A2 and B2 scenarios, which decreased due to a decrease in an-nual river discharge.

0

20

40

60

jan feb mar apr may jun jul aug sep oct nov dec

Mon

thly

str

eam

flow

, mm

control H A2 H B2 E A2

E B2 Had A2 Had B2

Figure 4. Monthly streamflow based on 100 30-year simulations obtained by resampling (points show median of 100 mean monthly values).

22

Although the test case for New York (Paper I) used direct output from a

different climate model than was used in Papers II, III, and IV, and some high-flow events were missed because precipitation was averaged over a vast area (4º x 5º grid), a change in timing and delivery of phosphorus was seen. The New York catchment showed a decrease in dissolved phosphorus load and streamflow for the future scenario used (A2). Paper I provides the only case in which particulate phosphorus was included in this thesis, and there was an increase in the cumulative particulate phosphorus loss for the future scenario. Since a shift was also seen in the timing of peak flows to the winter, the results suggest that erosion increased during the dormant season.

Paper IV focused on dissolved inorganic nitrogen loads for sites in five European countries in two climate regions. According to the Köppen classi-fication system, Sweden, Finland, and Estonia have a “humid continental” climate that is moist with severe winters and Ireland and the UK have a “ma-rine” climate that is moist with mild winters (Ahrens, 1994). For all sites, a shift in seasonal export patterns of DIN were observed for multiple future climate projections. As with the New York case (Paper I), the northern Eu-ropean sites showed a similar pattern of increased fluxes in winter that were explained by precipitation changes for future climate projections. Changes in the timing and magnitude of peak DIN load seen in the northern European sites (Sweden, Finland, Estonia) occurred due to increased winter rainfall and reduced snow cover and spring melt runoff. The sites in western Europe (Ireland, UK) experienced DIN load increases in winter or early spring (January through April in Ireland and November through March in the UK). A summary of the annual percentage change in dissolved organic nitrogen loads is shown in Table 3. Overall, the timing of nutrient delivery for all six future climate scenarios is strikingly different from historical conditions (Figure 5).

Table 3 Annual percentage change in median daily dissolved inorganic nitrogen loads.

Climate Model/Scenario Site

E A2 H A2 Had A2 E B2 H B2 Had B2

Flesk (IE) -4 -2 -7 +10 -4 +2

Esthwaite (UK) +43 +17 +4 +57 +4 +6

Mustajoki (FI) +25 +3 -10 +20 +3 +10

Arbogaån (SE) +19 +2 -12 +12 +2 -11

Tarvastu (EE) +50 +2 – +34 -4 –

23

Figure 5. Simulated DIN load (median of daily values aggregated by month for multiple realisations of a 30-year time series (black line: control period, 1960-1991; gray envelope:

shows highest and lowest median values for six climate change projections. a. Esthwaite Water (UK); b. Flesk, L. Leane (Ireland); c. Mustajoki, L. Pääjärvi (Finland); d. Arbogaån,

L. Mälaren (Sweden). Note scale changes on y-axis.

A cursory comparison of hydrological and nutrient loading results in this thesis (Papers II and IV) was made with other Swedish studies (Andréasson et al., 2004; Arheimer et al., 2005). Hydrological modelling results for the HBV model applied to a comprehensive study for the whole of Sweden showed a change in runoff ranging from +2 to -35% for the Norrström basin (Andréasson et al., 2004). Our results for the same scenarios range from +3 to +22%. Andréasson et al. (2004) presented results on a regional scale and their approach differed in how the climate signal was transferred to the hy-drological model. However, the seasonal pattern in runoff for an example from southern Sweden (Arheimer et al., 2005) showed a similar pattern to the results presented in Papers II and IV for four scenarios. Arheimer et al. (2005) also reported increases in annual nitrogen load due to higher winter flows, as did Papers II and IV for the E A2, E B2, H A2, and H B2 scenar-ios. This is notable because different models with different structures and assumptions converged with a similar outcome. GWLF omits biogeochemi-cal processes, whereas Arheimer et al. (2005) take into account nitrogen retention and transformation, in addition to transport.

24

Evaluating the catchment contribution to lake productivity for a warmer scenario

The impact of changes in water quantity, quality, and the timing of delivery from the catchment on receiving waters (lakes and reservoirs) is a critical question for drinking water supply, recreation, and ecosystem function. The main message in Paper III is that climate affects both the inputs from the catchment and in-lake processes, and the warming projections may have significant impacts on lake trophic status. The impact of the catchment con-tribution on phytoplankton growth for a warmer scenario was to promote the growth of cyanobacteria that was accompanied by a decline in diatom bio-mass. Additionally, an increase in lake temperature favored an increase in total phytoplankton biomass. Simulations using coupled watershed–phytoplankton models suggest that lake water quality may deteriorate if warming projections are borne out.

Future work

A logical extension of the work presented in this thesis is to model the or-ganic carbon inputs from the catchment. Future work includes use of the GWLF hydrological model coupled with a process-based model for estimat-ing dissolved organic carbon (DOC) flux (Jennings et al., 2005). The intent is to simulate DOC flux for future scenarios since increased fluxes of or-ganic carbon and colored humic compounds are another important water quality concern. It is essential that observations used for model calibration reflect the dynamics of the model (Jørgensen and Bendoricchio, 2001). To obtain the appropriate data for model calibration, a river monitoring station was established in 2003 on river Hedströmmen and initial results of monitor-ing are reported in Paper V.

High-frequency monitoring of colored organic matter (CDOM) fluores-cence, a proxy for DOC, successfully captured the dynamic behavior of or-ganic matter inputs (Paper V). Results from two contrasting water years show CDOM concentration and flux are highly variable. Changes in concen-tration were often of short duration (a few hours to a few days), but could change by as much as 50% in less than 24 hours. The catchment response to rainfall and runoff events varied with season, antecedent conditions, and the magnitude and duration of storms, making high-frequency CDOM fluores-cence measurements a relatively inexpensive option that allows a more de-tailed examination of organic matter contributions from streams and rivers.

The relationship between CDOM and dissolved organic carbon (R2 = 0.75) make it possible to convert CDOM to DOC concentration for model-ling purposes. Continuation of this work is planned, using the high-

25

frequency CDOM fluorescence data for model calibration and simulation of DOC flux.

26

Conclusions and Perspectives

The main results of the papers presented in this thesis can be summa-rized as follows:

� Changes in streamflow and dissolved nutrients (N, P, and Si) for six fu-

ture climate projections gave a consistent picture of change in their sea-sonal patterns. For four Swedish rivers, changes in the timing of peak river discharge and nutrient delivery were attributed to increases in winter precipitation, reduced snowpack, and earlier snow melt. The direction and magnitude of change in annual loads depended on the climate scenario used and river drainage area. Four out of six future climate projections showed an increase in streamflow and associated DIN and DIP loads. Two projections (HadRM3p A2 and B2) were the exception, and declined due to lower annual precipitation. The spring melt peak observed histori-cally was reduced for all six future scenarios examined, and peak runoff and dissolved phosphorus and nitrogen load maxima occurred in winter rather than early spring.

� The change in timing of nutrient delivery to one Swedish lake for one moderate future warming scenario had a negative impact on lake water quality. Cyanobacteria biomass increased and diatom biomass decreased as a consequence. Total phytoplankton biomass increased in response to increasing lake water temperature. Measures taken to ameliorate eutro-phication problems improved the water quality for Lake Mälaren in the past (Willén, 2001). Our results from simulations with coupled catch-ment and lake models suggest that climate warming may promote in-creases in algal blooms and require futher nutrient reductions to maintain water quality improvements.

� On a superregional scale, model results for DIN flux reflect the effects of dynamic changes in hydrology, including changes in the proportion of surface runoff and subsurface flow to total streamflow with their differing signature concentrations of nitrogen. For the sites in northern Europe (Sweden, Finland, and Estonia), peak DIN loads occurred in winter when projected air temperature increases are highest. Loss of snow cover, which impacts soil freezing and thawing, and increases in winter precipi-tation can be expected to increase winter DIN loads. The sites in western Europe (Ireland and the UK) experienced DIN load increases in winter or

27

early spring (January through April in Ireland and November through March in the UK). Warmer temperatures extend the growing season and this could conceivably offset some of the increases in DIN loading.

� A monitoring approach for capturing the dynamic behavior of organic matter inputs through the use of automated measurements of colored dis-solved organic matter (CDOM) fluorescence provides appropriate data for tracking pulses of organic matter and for calibrating a dissolved or-ganic carbon model.

28

Effekter av klimatförändringar på vattenflöde till sjöar och deras vattenkvalitet (Swedish summary)

En nyckelfråga när det gäller förutsägelser om klimatförändringars effekter är hur akvatiska system reagerar när drivvariabler som temperatur och ne-derbörd förändras.

Simuleringsmodeller är användbara instrument för att undersöka effekter av klimatförändringar på vattenflöde och vattenkvalitet. Den här avhandling-en utvärderar resultaten vid användandet av scenarier för ett varmare klimat för att driva en enkel avrinningsområdesmodell, GWLF, som simulerar vat-tenflöde med daglig upplösning kopplat till funktioner för näringsämnesbe-lastning för att beskriva kväve-, fosfor- och kiseltransporter.

Ett första test av GWLF med en grov modell för atmosfärisk cirkulation applicerad på ett avrinningsområde I New York indikerade att förändringar i den årliga belastningen av löst fosfor minskade med det årliga vattenflödet samt att den partikulära fosforbelastningen ökade i ett framtida klimatscena-rio. Vattenflödet ökade i januari, februari och april, men minskade under alla övriga månader I framtidsscenariet. När GWLF användes för Mälarens del-bassäng Galtens avrinningsområde tillsammans med en regional klimatmo-dell, som skalats ned till avrinningsområdet, kunde en förändring i årligt och säsongsflöde tillsammans med belastningen av löst fosfor och kväve påvisas i alla fyra huvudinflöden. Riktningen och storleken på förändringarna var dock beroende av vilket klimatscenario som användes och karaktären på de olika åarnas avrinningsområden. Den historiska vårfloden minskade i alla sex klimatscenarier som testades och maximala flöden och näringsämneshal-ter förekom under vintern i stället för under tidig vår.

Det biologiska svaret på ett varmare klimat modellerades för Galtenbas-sängen i Mälaren. Här användes en växtplanktonmodell (PROTBAS). Vat-ten- och näringsämnesinflöde modellerades med avrinningsområdesmodel-len GWLF och användes tillsammans med en fysikalisk sjömodell (PROBE) för att driva växtplanktonmodellen för en historisk period och ett framtids-scenario. I en varmare framtid ökade cyanobakteriernas biomassa beroende på tidsförändringar för näringsämnesinflödet (löst kväve, fosfor och kisel)

29

och mängden kiselalger minskade. Den förutspådda temperaturhöjningen orsakade en ökning i total växtplanktonbiomassa.

Belastningen av löst kväve på sjöar modellerades för fem europeiska av-rinningsområden, vilket möjliggjorde en jämförelse av modellresultat för Galtens avrinningsområde i Sverige med avrinningsområden i Finland, Est-land, England och Irland. Modellresultaten visar effekterna av dynamiska förändringar i hydrologi inkluderande förändringar i andelen ytavrinning i förhållande till det totala flödet. Skiften i säsongsfördelningen av det lösta kvävets flöden konstaterades för alla avrinningsområden vid ett flertal fram-tidsscenarier. I norra Europas avrinningsområden skedde förändringarna i tidpunkt och storlek för maximala flöden beroende på högre vinterneder-börd, minskad snömängd och därmed lägre vårflod. Avrinningsområdena i västra Europa (Irland och England) hade högre vinterflöden av löst kväve (januari-april på Irland och november-mars i England).

Slutligen redovisas en automatisk högfrekvensmätning som ett första steg för att andvända avrinningsområdesmodellen GWLF kopplad till en pro-cessbaserad flödesmodell för löst organiskt kol för att karaktärisera dagens inflöde via åar samt framtida uppskattningar. Timvisa mätningar av fluore-scens från färgat organiskt material i Hedströmmen (inflöde till Galten) un-der en tvåårsperiod visade att den var mycket varierande. Utan komplemen-tära högfrekvensmätningar skulle episoder med höga koncentrationer eller utspädning missas vid reguljär manuell provtagning. Relationen mellan fluo-rescens och koncentrationen av löst organiskt kol visar att den kan användas för att uppskatta halterna av löst organiskt kol. Metoden tycks kunna ge data för att kunna kalibrera flödesmodellen för löst organiskt kol och kan även användas för övervakning av inflödet av löst organiskt kol till sjöar.

30

Acknowledgements

I would like thank all who supported me during my studies. First, I want to thank my supervisors, Don Pierson and Kurt Pettersson for giving me such a fantastic opportunity to work in Sweden. Also, a big thank you to your fami-lies for their hospitality and kindness.

Many thanks to Hampus Markensten for help throughout my project work, including programs and valuable advice, and to Irina Persson for your collaboration and friendship. To Rahmat Naddafi, Thorsten and Stephanie Blenckner, Gesa Weyhenmeyer, and Emil Rydin, thanks for your encour-agement and friendship. It was a pleasure to be at Norr Malma field station together.

The staff at Erken Laboratory have been so helpful over the years. I would have been stuck many times without the help of Björn Mattsson who came to the rescue many times and made things work again. Thanks to He-lena Enderskog, Olga Matzèn, Erika, and Monika for the sample analysis and help in all seasons. Maria Kahlert, I will always remember the research school of 2004. Karin Pettersson as leader of the 2005 research school, your support was essential for its success. Thanks to you both and to the students for that great experience! My memory of working with the modelling class from Belarus in 2006 is also a fond one. I’m sure I learned as much as the students through the experience.

Although my visits to the limnology department in Uppsala have been sporadic, I appreciate everyone there for making me feel welcome. Thanks to Lars Tranvik for allowing me to come as an external student. It has been a great privilege. Thanks to Peter Eklöv for the annual reviews and for recog-nition of the amount of work involved. A big thanks to Jan Johansson for DOC analysis. You were always so accomodating with sudden requests for sample analysis.

There are many people in the CLIME project that made my work a joy. Thanks to Glen George for the inspiration and guidance. I am grateful to Eleanor Jennings and Pam Naden for help and mentoring during the project. Thanks to Linda May and her daughter Nicola for the inspiring picture of Nicola that was priceless. Patrick Samuelsson provided essential climate data and patient answers to my many questions. There were some special meetings connected to the project that will always be a happy memory. Spe-cial thanks to Uli Nickus and Hansjörg Thies in Austria; Judit Padisák in Hungary; Tiina and Peeter Nõges and Anu Reinart in Estonia; Lauri Arvola

31

and Marko Järvinen in Finland; Norman Allott and Eleanor Jennings in Ire-land for your hospitality. There are many others in the CLIME project whose kindness is much appreciated.

There are several to thank at the New York City Department of Environ-mental Protection. I must first say that David Lounsbury saved the day with his frequent advice on GIS problems. Thanks David for your help with both my Master’s and Ph.D. projects. Lorraine Janus has been a mentor and en-courager throughout, and I am grateful for her support. Thanks to Elliot Schneiderman and Mark Zion for the lessons in modeling and your insights.

I appreciate the collaboration with Cynthia Rosenzweig and Rich Gold-berg at the Goddard Institute of Space Studies in Manhattan. Your help al-lowed me to get some preliminary results at an early stage of my Ph.D. stud-ies.

My connection to the Institute of Ecosystem Studies in Millbrook, New York has remained strong over the years. I appreciate the use of the library, and a big thanks to Amy Schuler for her assistance. Thanks to all the staff who made me feel welcome. I often made my visits when I was getting stuck in place and needed fresh inspiration to move forward.

A very special thanks to my family for all their support. My husband Larry showed tremendous patience throughout and did not complain when I could not tend to the household or cook (although maybe the cooking part was a blessing in disguise). To my children, a huge thanks for understanding when I was absorbed in my studies. Thanks to Ian for the technical support, such as building my main modeling computer. Thanks to Kyle for keeping up the pace in school and weathering my frequent trips away from home. Thanks to my parents and entire family for encouragement and for under-standing this preoccupation I have with science. They have always been supportive of my studies, even in my early years at Vassar College. A spe-cial remembrance of the contribution made by my mother-in-law, Marie Hines, who helped with the children during my absences. Marie died of lung cancer ten days after my return from Sweden in May 2005. It was heart-breaking and we miss her dearly.

Thanks to my friends at Mitchell Hollow Mission Church for your sup-port. In particular, Johanna Overton and her husband Pastor David Overton have been wonderful friends and encouragers since our arrival in Windham. Thanks to Pastor Todd and Susan Luce, Rose and Kai Schleupen, Liz and Tim Buckner and others for your prayerful support.

I am sure I have left someone important out, but if you read this and missed your name, it is because my brain was tired and any omissions were unintentional. Just know that kindness matters, and I appreciate all the help I have gotten along the way. Tack så mycket!

Funding support from the European Commission Environment and Sus-tainable Development Programme (contract EVK1-CT-2002-00121, CLIME) and from Malméns stiftelse are is gratefully acknowledged.

32

References

Ahrens, C.D. (1994) Meteorology Today: An Introduction to Weather, Climate, and the Environment, 5th ed. West Publishing Company, Minneapolis/St. Paul.

Andréasson, J., Bergström, S., Carlsson, B., Graham, L.P. and Lindström, G. (2004) Hydrological change–Climate change impact simulations for Sweden. Ambio 33: 228–234.

Arheimer, B., Andréasson, J., Fogelberg, S., Johnsson, H., Pers, C.B. and Persson, K. (2005) Climate change impact on water quality: Model results from southern Sweden. Ambio 34: 559–566.

Arheimer, B. and Brandt, M. (1998) Modelling nitrogen transport and reten tion in the catchments of southern Sweden. Ambio 27: 471–480.

Bergström, S. (1995) The HBV model. Pages 443-476 in V.P. Singh (ed.) Computer Models of Watershed Hydrology. Water Resources Publications, Littleton, Colorado.

Brandt, M. and Ejhed, H. (2002) TRK Transport – Retention – Källfördelning. Belastning på havet. Naturvårdsverket Rapport 5247, Stockholm.

Carpenter, S.R. (2003) The need for fast and frugal models. Pages 455–460 in C.D. Canham, J.J. Cole and W.K. Lauenroth (eds.), Models in Ecosystem Science. Princeton University Press, Princeton, N.J.

Dai, T., Wetzel, R.L., Tyler, R.L. and Lewis, E.A. (2000) BasinSim 1.0: A Windows Based Watershed Modeling Package. Virginia Institute of Marine Science, College of William and Mary. (http://www.vims.edu/bio/models/basinsim.html)

Edwards, M., Johns, D.G., Leterme, S.C., Svendsen, E. and Richardson, A.J. (2006) Regional climate change and harmful algal blooms in the northeast Atlantic. Limnology and Oceanography 51: 820–829.

Haith, D.A. (1985) An event-based procedure for estimating monthly sedi

ment yields. Transactions of the American Society of Agricultural Engineers 28: 1916-1920.

33

Haith, D.A., Mandel, R. and Wu, R.S. (1992) GWLF, Generalized Watershed Loading Functions Version 2.0 User’s Manual. December 1992 (corrected and reprinted, January 1996). Dept. of Agricultural and Biological Engineering, Cornell University, Ithaca, New York.

Haith, D.A. and Shoemaker, L.L. (1987) Generalized watershed loading functions for stream flow nutrients. Water Resources Bulletin 23: 471–478.

Haith, D.A. and Tubbs, L.J. (1981) Watershed Loading Functions for Non point Sources. ASCE. Journal of Environmental Engineering 107(EE1):121-137.

Hay, L.E., Wilby, R.L. and Leavesley, G.H. (2000) A comparison of delta change and downscaled GCM scenarios for three mountainous basins in the United States. J. Am. Water Resources Assoc. 36: 387– 397.

Hong, B. and Swaney, D. (2004) GWLFXL Users’ Manual. Cornell University. (http://www.eeb.cornell.edu/biogeo/usgswri/GWLFXL/gwlfxl.doc)

IPCC (1995) Climate change 1995: The science of climate change. J.T. Houghton, L.G. Meira Filho, B.A. Callander, N. Harris, A. Kattenberg and K. Maskell (eds.). Cambridge University Press, Cambridge.

IPCC (2001) Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergov- ernmental Panel on Climate Change. J.T. Houghton, Y. Ding, D.J. Griggs, M. Noguer, P.J. van der Linden, X. Dai, K. Maskell and C.A. Johnson (eds.). Cambridge University Press, Cambridge.

Janus, L.L. and Vollenweider, R.A. (1981) The OECD Cooperative Programme on Eutrophication, Canadian Contribution. Scientific Series No. 131, National Water Research Institute, Inland Waters Directorate, Canada Centre for Inland Waters, Burlington, Ontario.

Jennings, E., Naden, P., Järvinen, M., Moore, K., Nic Aongusa, C., Tamm, T., and Järvet, A. (2005) Report on Testing of the DOC Model in “Warm World” Mode. Workpackage 5 deliverable, CLIME project EVK1-CT-2002-00121.

Jones, P.D. and Salmon, M. (1995) Development and integration of a

stochastic weather generator to a crop growth model for European agriculture. MARS Project, Final Report to Institute of Remote Sensing Applications, Agricultural Information Systems (ISPRA)., Contract No. 5631-93-12, ED ISP GB.

34

Järvet, A., Järvinen, M., Jennings, E., Koncsos, L., May, L., Moore, K., Nic Aongusa, C., Padisák, J., Pierson, D. and Tamm, T. (2006) Report on testing of delivery and internal dynamics of P and N Models in “Warm World” Mode. Workpackage 4 deliverable, CLIME project EVK1-CT-2002-00121.

Jørgensen, S.E. and Bendoricchio, G. (2001) Fundamentals of Ecological Modelling, 3rd ed. Elsevier, Amsterdam.

Kyllmar, K. and Johnsson, H. (1998) Nutrient losses from arable land within the period 1984–1995. Results from the water quality monitoring programme Typområden på jordbruksmark, Swedish University of Agricultural Sciences, Division of Water Quality Management, Report Ekohydrologi No. 44.

Markensten H. and Pierson, D.C. (2007) Weather driven influences on phytoplankton succession in a shallow lake during contrasting years. Application of PROTBAS. Ecological Modelling doi:10.1016/j.ecolmodel.2007.1004.1023.

Mayer, D.G. and Butler, D.G. (1993) Statistical validation. Ecological Modelling, 68: 21–32. Mörth, C-M., Humborg, C., Eriksson, H., Danielsson, Å., Medina, M.R., Löfgren, S., Swaney, D.P. and Rahm, L. (2007) Modelling riverine nutrient transport to the Baltic Sea: A large-scale approach. Ambio 36: 124–133.

Nash, J.E. and Sutcliffe, I.V. (1970) River flow forecasting through conceptual models. Part 1: A discussion of principles. Journal of Hydrology 10: 282–290.

Ogrosky, H.O. and Mockus, V. (1964) Hydrology of agricultural lands. In V.T. Chow (ed.) Handbook of Applied Hydrology, McGraw-Hill, New York.

Russell, G.L., Miller, J.R. and Rind, D. (1995) A coupled atmosphere-ocean

model for transient climate change studies. Atmosphere–Ocean 33:683-730.

Samuelsson, P. (2004) RCAO and HadRM3p simulation results with focus

on CLIME sites. CLIME Deliverable 5 (contract EVK1-CT-2002- 00121).

Schoumans, O.F. and Silgram, M. (eds.) (2003) Review and literature evaluation of nutrient quantification tools. EUROHARP report 1-2003, NIVA report SNO 4739-2003, ISBN 82-557-4411-5, Oslo, 120 pp. Schneiderman, E.M., Pierson, D.C., Lounsbury, D.G., and Zion, M.S. (2002) Modeling the hydrochemistry of the Cannonsville watershed with

35

GWLF. Journal of the American Water Works Association 38: 1323–1347. SCS (1986) U.S. Soil Conservation Service Technical Release 55: Urban Hydrology for Small Watersheds. USDA (U.S. Department of Agriculture). June 1986.

Silgram, M. and Schoumans, O.F. (eds.) (2004) Modelling approaches: Model parameterisation, calibration and performance assessment methods in the EUROHARP project. EUROHARP report 8-2004. NIVA report SNO 4740-2003, Oslo, 18 pp.

SMHI (1994) Svenskt Vattenarkiv: Avrinningsområden i Sverige Del 3. Vattendrag till Egentliga Östersjön och Öresund. Swedish Meteorological and Hydrological Institute, Report No. 50, Norrköping.

Stott, P.A., Tett, S.F.B., Jones, G.S., Allen, M.R., Mitchell, J.F.B. and Jenkins, G.J. (2000) External control of 20th century temperature by natural and anthropogenic forcings. Science 290: 2133–2137.

Svensson, U. (1998) PROBE Program for Boundary Layers in the Environment System description and manual. Swedish Meteorological and Hydrological Institute, Report No. 24, Norrköping. Wallin, M. (ed.) (2000) Mälaren–miljötillstånd och utveckling 1965–1998.

Mälarens vattenvårdsförbund, Uppsala.

Watts, M., Goodess, C. M., and Jones, P. D. (2004): The CRU Daily Weather Generator: BETWIXT Technical Briefing Notre 1. The Climate Research Unit, University of East Anglia, UK.

Willén, E. (2001) Four decades of research on the Swedish large lakes Mälaren, Hjälmaren, Vättern and Vänern: The significance of monitoring and remedial measures for a sustainable society. Ambio 30: 458–466.

Wischmeier, W.H. and Smith, D.D. (1978) Predicting rainfall erosion losses: A guide to conservation planning. Agricultural Handbook 537, U.S. Department of Agriculture, Washington D.C.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 347

Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science andTechnology, Uppsala University, is usually a summary of anumber of papers. A few copies of the complete dissertationare kept at major Swedish research libraries, while thesummary alone is distributed internationally through theseries Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Science and Technology.(Prior to January, 2005, the series was published under thetitle “Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-8236

ACTA

UNIVERSITATIS

UPSALIENSIS

UPPSALA

2007

Documents

Climate Change Impacts on the Catchment Contribution to ...uu.diva-portal.org/smash/get/diva2:170783/FULLTEXT01.pdfFigure 1. Lake Mälaren, Sweden. A synthesis paper (Paper IV) features