Assessment of Water Availability in a Central- European River Basin (Elbe) Under Climate Change

Transcription

1 Impacts and Adaptation Article ID: (28) Suppl Assessment of Water Availability in a Central- European River Basin (Elbe) Under Climate Change Fred F. Hattermann 1, Joachim Post 2, Valentina Krysanova 1, Tobias Conradt 1, Frank Wechsung 1 1 Potsdam Institute for Climate Impact Research, Telegrafenberg A62, Potsdam D-14473, Germany 2 German Aerospace Center (DLR); German Remote Sensing Data Center (DFD), Wessling D-82234, Germany Abstract: The Elbe region is representative of humid to semi-humid landscapes in Central Europe, where water availability during the summer season is the limiting factor for plant growth and crop yields, especially in the loess areas with high crop productivity having annual precipitation lower than mm. This paper summarizes the results of the first phase of the GLOWA (GLObal WAter)-Elbe project and tries to assess the reliability of water supply in the German part of the Elbe river basin for the next years, a time scale relevant for the implementation of water and land use management plans. One focus of the study was developing scenarios which are consistent with climate and land use changes considering possible uncertainties. The concluding result of the study is that nature and communities in parts of Central Europe will have to deal with considerably lower water resources under scenario conditions. Key words: water resources; land use change; climate change; vegetation Introduction There have been numerous studies of the potential impacts of climate change on water resources and agriculture over the last decade [1J7]. A review on climate change, climate models and water resources management can be found in Varis et al. [8]. However, in most studies, the climate change impact was investigated for the hydrological regime, land use and agriculture separately, applying different and incompatible tools, and without addressing the possible feedbacks and inherent uncertainty of the results [9]. Regional characteristics are taken into account in the study by using climate forcing data which were produced by a statistical downscaling method [1]. The land use scenario applied in this study has been derived assuming that changes will mainly affect agriculture. The anticipated trend in global agricultural economics and the trend in crop yields under climate change have been integrated in development of the land use scenario, whereby impacts of climate change on vegetation growth have been taken into account to assure consistency in the climate and land use scenario. The eco-hydrological model used to analyze the impacts of changes in climate and land use on hydrology and crop yields is the model SWIM (Soil and Water Integrating Model) [11]. SWIM is applied to propagate the inherent uncertainty in water supply under climate change caused by the uncertainty in the climate input data, whereby SWIM transforms the change in climate into altered hydrological quantities. This study, carried out in the framework of the first phase of the German GLOWA-Elbe project, tries therefore (1) to develop and apply a transient climate and land use change scenario which is consistent in terms of its climate and socio-economic boundary conditions, and takes feedbacks of the water cycle and vegetation into account, (2) to investigate and quantify the climate and land use change impacts on water resources in a region where the per capita water supply is very low, and where the water availability is the main constraint for vegetation growth and crop yields, and (3) to quantify the uncertainty of water Received: October 29, 27 Corresponding author: F. F. Hattermann, hattermann@pik-potsdam.de 28 Editorial Office of 42 Adv. Clim. Change Res., 28, 4 (Suppl.): 42J

2 availability. 1 Material and methods 1.1 The modelling strategy The GCM (Global Circulation Model) simulation run selected to apply as boundary condition in the study was produced by the coupled atmospherejocean model ECHAM4-OPYC3 [12], which was driven by the IPCC SRES (International Panel on Climate Change Special Report Emissions Scenario) A1 [13]. The regional climate change scenario was derived using the statistical downscaling model STAR (STAtistisches Regionalmodell) [14] (== 1 in Fig. 1) based on the assumption that the most reliable result of GCM simulations is the trend in temperature. One hundred long-term transient time series of the possible future climate (2J2) were calculated in a way that they reflect the temperature changes calculated by the GCM in the given scenario as well as the observed regional climate pattern. The IPCC temperature scenario selected gives a rather moderate temperature increase of approximately 1.4 until 2. The anticipated scenario increase in temperature can be, due to the robustness of the temperature signal with only small differences of temperature increase in the different IPCC scenarios until 2, regarded as representative for the model region and different scenarios [1]. The stochastic character of the possible climate change in the region was taken into account in the scenario derivation by incorporating a conditioned Monte Carlo approach in the downscaling process and producing 1 realizations. They cover the possible range of climate change in the German Elbe basin under various scenario WATSIM Global crop market development conditions. For more information about the derivation of the 1 realizations and the generation of climate change uncertainty [1]. The next step was to derive a land use change scenario, consistent with the socio-economic story line of the regional climate change scenario. This was done based on the assumption that land use changes in the Elbe basin will mainly affect arable land. Following this assumption, the eco-hydrological model SWIM was used to calculate potential crop yields under climate change for the nine main crops cultivated in the study area ( == 2 in Fig. 1). The climate change scenario taken was the most probable variant of the 1 scenario realizations produced by STAR. The results represented one input for the regional agroeconomic model RAUMIS (Regionales Agrar-und Umweltinformationssystem fur Deutschland) [16]. They were used by RAUMIS to optimize the potential operational income of farmers in the study area under climate change and to calculate corresponding county specific crop distributions (land use scenario variant A) (== 3 in Fig. 1). The second input was global crop market conditions as defined by the underlining IPCC storyline and produced by the global agro-economic model WATSIM [17]. These were the input for RAUMIS to optimize the potential operational income of farmers under socio-economic change in a second land use scenario (land use scenario variant B). The spatial aggregation level of the simulations is the administrative level of counties. The county specific crop distributions derived by RAUMIS had to be transformed into crop rotations for each spatial element considered in the eco-hydrological model SWIM (== 4 in Fig. 1), which are much smaller than the average county area, and which are not defined by ECHAM4 Global climate change scenario Landscape Regional Global RAUMIS 1) Optimized operational income of farmers 3 2) Crop distributions under global change 4 2 SWIM 1) Potential crop yields unter climate change 2) Regional water balance under global change STAR Regional climate change 1 Fig. 1 Flow chart of the modelling procedure Adv. Clim. Change Res., 28, 4 (Suppl.): 42J 43

3 Fred F. Hattermann et al.: Assessment of Water Availability in a Central-European River Basin (Elbe) Under Climate Change administrative boundaries but built based on information about the environmental characteristics of the basin (the so called hydrotopes or Hydrological Response Units, HRUs). A crop generator was developed and implemented in SWIM to disaggregate the county specific crop distributions and to calculate crop rotations for each hydrotope under the constraint that the simulated crop distributions in the catchment agree each year with the county statistics. This was done considering seven soil fertility classes, whereby commercially more valuable crops with mostly higher nutrient and water demand were allocated in areas having fertile soils and vice versa, as it is recorded in the local crop statistics. The final step was to transform the regional climate and land use changes into altered hydrological quantities (==in Fig. 1). This was done by running SWIM 1 times, each time forced by another realization of the climate change scenarios, and with the land use scenarios as spatial boundary condition. 1.2 The basin under study The river Elbe is the most easterly located large river flowing into the North Sea. The total Elbe basin has a catchment area of 148,268 km 2. The German part of the Elbe, where the model was applied, covers 8,26 km 2 from the Czech border to Neu Darchau, the downstream gauge station not influenced by the tide of the North Sea (see Fig. 2), and additional 16,148 km 2 in the inter-tidal zone or drained by rivers influenced by the tide. The total length of the Elbe is 192 km, 728 km of that in Germany. About 2% of the catchment areas are covered by arable land, 29% by forest, 6% by settlements and industry and % by lakes, mining pits and other land use forms. The Elbe and its estuaries are regulated by 273 dams for flood protection and freshwater supply. Despite of flood protection measures, several extreme floods occurred during the last decades in the region, culminating in the disastrous August 22 flood in the Elbe basin. The flood was caused by a low pressure system called Vb ( five b ), a circulation pattern that is known to produce heavy and intensive rainfall in central Europe [19]. Climatically, the Elbe basin is one of the driest regions in Germany, with mean annual precipitation below mm in the lee of the Harz Mountains (western part of the basin), where the loess plains with high agricultural productivity Germany Magdeburg N 1 km Hamburg Gadow Blankenstein Berlin Wechselburg Czech Republic Fig. 2 The German Elbe basin [18] Dresden Poland are located. The long-term mean annual precipitation over the whole German part of the basin is 69 mm in the period 191J2 (corrected), and the long-term mean discharge at the estuary is 877 m 3 /s with an average inflow from the Czech Republic of 31 m 3 /s (ATV-DVWK 2) [2]. The per capita water supply is only 68 m 3 /a, being one of the lowest in Europe. The average per capita water supply in Germany is ca. 22 m 3 /a, with approximately 137 m 3 /a in the Rhine basin and 43 m 3 /a in the German Danube basin [21]. The per capita water demand in Germany is 49 m 3 /a, about 14% of it being for public freshwater supply. Over the last two decades, decreasing water levels in rivers and groundwater have been observed in the lowland parts of the basin. Groundwater recharge, especially, is extremely sensitive to changing conditions of climate and land use, since it represents the residual of the water balance [22]. 2 Results and discussion 2.1 Reference period Climate downscaling Observed climate data of the period 191J2 were chosen to validate the regional climate model STAR [1]. The validation criterion was that the stability of the main statistical characteristics of the regional climate (variability, frequency distribution, annual cycle and persistence) had to be maintained. A set of 1 realizations of the climate for this period has been generated. By comparison of the climate statistics of the individual realizations with the observed trend in =Wechsung F, Hattermann F, Post J, et al. Algorithm for the generation of spatial crop cover patterns on arable land and its application in a modelling framework for accessing hydrological impacts of land use and climate change in the Elbe river basin, 28 (in preparation) 44 Adv. Clim. Change Res., 28, 4 (Suppl.): 42J

4 climate, it was found that the median trend in precipitation of the 1 variants has the highest correlation with the observed trend for the integral of the total basin, indicating an equally balanced distribution around the most likely realization Hydrology The eco-hydrological model has been at first adjusted to the hydrological processes in the Elbe basin using a multiobjective, multi-site and multi-scale validation [23]. The model was tested in 12 subbasins located in different regions of the German Elbe basin with catchment areas from 28 to 23,69 km 2 (multi-site calibration). The calibration was done on a daily time step using records of observed river discharge and groundwater table dynamics for comparison (multiobjective calibration). A non-generic automatic calibration was performed using a Latin Hypercube method [24] to make sure that all physically meaningful parameter combinations are considered in the modelling procedure. The same method was applied to analyse the model sensitivity and to quantify the uncertainty of the simulation results. Afterwards, fine-tuning of the model was done by hand. The simulated river discharge was compared with the measured discharge for a 12-year period (calibration period 1981J1986, validation period 1987J1992). 2.2 Scenario period Climate The trend of climate change under scenario conditions and the range of the 1 variants are shown in Table 1. The observed trend of precipitation in the Elbe River basin differs from the global trend of higher precipitation due to the regional orography (the Harz Mountains in the west of the basin) and an increase in west wind circulation pattern. It is assumed that this trend will last into the future and as a result also a small negative trend in precipitation. Due to Q / (m 3 /s) Observed Simulated Year Fig. 3 Comparison of the simulated and observed river discharge of the Elbe basin (Neu-Darchau, catchment area 86, km 2 ) [23] the uncertainties in the regional climate change, the possible range of change in annual precipitation is between approximately J7 mm and 8 mm (Table 1). A trend analysis reveals that the precipitation trend is only significant for the two driest realizations (with a 9% confidence), one of them is realization Land use The eco-hydrological model estimated a basin wide average decrease in crop yields under climate change of approximately 11%J1% for different C3 cereals and no significant change in the productivity of C4 plants like maize. On the basis of these results, the agro-economic model RAUMIS calculated new crop distributions for the counties in the German Elbe basin, whereby the local farmers will optimize their operational income under climate change. The main land use changes are in areas, where cost-effective crop cultivation will be impossible under altered climate conditions because of the low fertility and water holding capacity of the local soils and/or the decrease in precipitation. Considering only climate change as driver (land use scenario variant A) would induce only a small shift from agricultural land to fallow land (decrease in cropland area Table 1 The scenario trend in temperature and precipitation for the entire German Elbe basin (difference of the average 1961J199 and 21J2) Temperature/ Precipitation/mm Summer Winter Average Summer Winter Sum Lowland (J43.3; 72.)* J8. (J22.;.9)* J1.7 (J6.8; 73.4)* Loess region (J39.6; 12.) J16. (J34.7; 12.)* J.1 (J74.3; 9.1)* Mountains (J42.6; 19.)* J18.9 (J27.7; J.9)* J16.6 (J7.3; 13.1)* Total Elbe basin (J42.; 84.2)* J11. (J2.7; J2.8)* J.6 (J68.2; 81.)* *Range of change between the driest and the wettest climate scenario variant Adv. Clim. Change Res., 28, 4 (Suppl.): 42J 4

5 Fred F. Hattermann et al.: Assessment of Water Availability in a Central-European River Basin (Elbe) Under Climate Change of approximately.9%), and also the shares of the different crops cultivated in the area are relatively stable. The changes are more pronounced when considering the liberalization of global crop markets in the second land use scenario (land use scenario variant B): The share of fallow land of the total crop area becomes approximately 29.8% (beforehand lower than 7.%), the share of winter wheat drops from 26.4% to 21.7%, and the one of maize drops from 7.% to 2.%. The scenario taken to produce the following results is variant A Hydrology Annual results The analysis of the impacts of climate change on the hydrology in the German Elbe basin has been performed by comparing the hydrological processes in the period 1961J199 against the period 21J2 (Fig. 4). The hydrological decade 1961J199 is referred to as the reference period, the period 21J2 as the climate change or scenario period. The climate change period is considerably shorter than the reference period because the annual precipitation during the scenario period has a relatively steady decrease. Averaging over a longer time period would conceal the problems for water users due to water shortage at the end of the planning period (and beyond). The relevance of the length of the scenario time period which has been selected to compare with is illustrated in Table 2, where also a longer scenario period has been chosen to calculate the possible changes in hydrological quantities. The uncertainty of the results is expressed by confidence intervals indicating the range of possible changes. The uncertainty in hydrology is the impact of the uncertainty in the climate input data shown in Table 1. The annual change in hydrology is significant only for the driest realizations comparing the means of the reference and the scenario period for each of the 1 realizations. While 2 climate realizations show a significant trend in precipitation, 14 realizations have a significant trend in evapotranspiration, also 14 realizations a significant trend Table 2 The range of possible changes in the hydrological cycle under climate change. Shown are the changes for the driest, median and wettest climate change variant (difference in % of the average 1961J199 (reference) and 21J2), and the number of scenario variants which have a decrease in the corresponding hydrological component Variant Reference Dry/% * Median/% * Wet/% * Number of dryer variants Precipitation 66.7 J1.3 (J12.) J1. (J.) 14.8 (8.8) 9 Evapotranspiration (3.) 6.8 (6.4) 14.9 (11.8) 1 Direct runoff J44.2 (J31.7) J11.2 (J6.2) 32.7 (9.1) 69 Ground-water recharge 78.8 J7. (J69.8) J28. (J22.).2 (J1.3) 87 * In brackets are the differences of the reference period 1961J199 and the longer period 246J2 8% included % included Average 1961J199 Average reallzation 32 Number of realizations (a) (b) (c) (d) mm mm Fig. 4 The distribution of the -year average during 21J2 of the 1 variants, their median and 1th, 2th, 7th and 9th percentiles, and the median of the reference period 1961J199 [18] (a) precipitation, (b) evapotranspiration, (c) groundwater recharge, (d) direct runoff 46 Adv. Clim. Change Res., 28, 4 (Suppl.): 42J

6 in groundwater recharge and 11 realizations a significant trend in river runoff. The effects are not necessarily combined: realization 32, for example, has a strong and significant trend in precipitation, but not in evapotranspiration, where the possible increase in evapotranspiration due to the increase in temperature is limited by the water (precipitation) available, and thus reduces the trend signal Daily dynamics Figure shows the daily river discharge of the 1 realizations at gauge Neu-Darchau, averaged for the reference period 1961J199 and the scenario period 21J 2. One main result is that the average discharge under scenario conditions is in 8 % of the variants lower for the scenario period than during the reference period. In addition, the annual flow regime under climate change is different from the one under reference conditions: due to higher temperatures, the vegetation season expands into the late autumn, and as a result the river water level rises more slowly at the end of the summer term. More and more pronounced low water situations can be observed in late summer and early fall. Floods are generated in early spring in the reference period stimulated by snow melt. But due to higher temperatures, the runoff retention of snow will be lower under climate change, and the period 21J2 shows the highest flood events during January. The consequence for water resources management is that longer periods with low discharge as well as severe flood events in early winter have to be taken into account in the planning process. Discharge/(cm 3 /s) 1 Average 1961J199 9 Average realization 32 8 % included 7 1% included Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov.Dec. Fig. Five year average daily water discharge at gauge station Neu-Darchau for the reference period and under scenario conditions (21J2) [18] Sub-regional impacts Figures 6 and 7 illustrate the regional changes in precipitation and groundwater recharge (comparison of reference period against scenario realization 32), while the Figs. 6d and 7d show the uncertainty of the results expressed by the coefficient of variability. The overall pattern is that rainfall will decrease in the northern part, while the southern part of the basin shows a more complex pattern of areas with increasing and areas with decreasing annual precipitation. This latter result is supported by Menzel et al. [4], where a different statistical downscaling method is used to generate a climate scenario for the Mulde catchment located in the southern part of the German Elbe basin. The trend in groundwater recharge follows this general pattern with large decreases in areas with decreased annual precipitation. Noticeable is as shown in Fig. 4 that the relatively small change in precipitation has a relatively high impact on groundwater recharge (as already discussed for the annual changes in precipitation and groundwater recharge). The areas, where the local annual groundwater recharge is negative due to plant water uptake from groundwater (red areas in Fig. 7 (a, b)) become larger under scenario conditions. The greatest relative changes occur in the lee region of the Harz Mountains, where also the change in precipitation is the highest. These are the areas where, due to the high water holding capacity of the loess soils, groundwater recharge is very low also during the reference period. The uncertainties in precipitation expressed by the coefficient of variability of the 1 variants are the highest in the lee of the Harz Mountains, and the same is visible for groundwater recharge, with a spatial focus in the loess plains (see Fig. 2). The highest uncertainties in groundwater recharge occur in areas where the total annual amount of recharge during the reference period has been small anyway. Small changes in precipitation have a great impact on recharge in these areas, and a coefficient of variability of maximum 14% for precipitation in these areas result in a coefficient of more than % for groundwater recharge. 3 Summary and conclusions The model system presented in this study allows assessing the uncertainty of water supply under global change considering the most important drivers, land use and climate, whereby impacts of the trend in climate on land use can be taken into account. The overall result is that nature and society in the catchments of the German Elbe basin are confronted with severe changes in water availability and river flow regimes, although the results are associated with high uncertainty. The trend in temperature of 1.4 in the period 21J2 will stimulate plant growth and lead to a prolongation of the vegetation period. As a result, plant transpiration will intensify in early spring Adv. Clim. Change Res., 28, 4 (Suppl.): 42J 47

7 Fred F. Hattermann et al.: Assessment of Water Availability in a Central-European River Basin (Elbe) Under Climate Change (a) (b) Preciopitation/mm 4J4 4J J J6 6J6 6J7 7J7 7J8 8J8 8J9 9J9 9 Preciopitation/mm 4J4 4J J J6 6J6 6J7 7J7 7J8 8J8 8J9 9J9 9 (c) (d) Prec./mm change J2 to J1 J1 to J1 J1 to J J to J J1 1J1 1J3 2 Variability/% lower 3 3J4 4J J6 6J7 7J8 8J9 9J1 1J11 11J12 12J13 13J14 Fig. 6 The mean annual precipitation of 1961J199 (a) and 21J2 (b), the change in precipitation (c) and the coefficient of variability (d) in the Elbe region [18] and continue into late autumn. This will influence the river flow regime, where the rise of the river water level will be later in early winter and the recession of the level earlier in spring. More and longer low water intervals can be observed in late summer and early fall. The increase in evapotranspiration will have the highest impacts on groundwater recharge, the residual of the local water balance, which is very sensitive to changes in water supply. The study shows that a relatively small and insignificant change in precipitation in combination with a more significant increase in temperatures can result in severe and significant changes in hydrological variables, in particular in river flow and groundwater recharge. The uncertainty of the results induced by the uncertainty in the development of the regional climate change and propagated by the ecohydrological model is much higher than the uncertainty generated by the performance of the eco-hydrological model itself [23], indicating how sensitive the water cycle is against change in climate. However, a very important result is that within all the uncertainties about the possible future development of water resources in the area, some very robust trends can be detected: there is less water availability in summer over nearly all of the 1 realizations (even when precipitation shows a slight increase), and the high flow period is earlier in the year (due to the earlier snow melt). It seems to be that the relative robust and certain increase in temperature outbalances the relative uncertain trend in precipitation, because temperature stimulates evapotranspiration, having nearly the order of magnitude as precipitation. This very 48 Adv. Clim. Change Res., 28, 4 (Suppl.): 42J

8 (a) (b) Recharge/mm J J1 1J1 1J2 2J2 2J3 3J3 3J4 4J4 4 Recharge/mm J J1 1J1 1J2 2J2 2J3 3J3 3J4 4J4 4 (c) (d) Recharge change/mm J27 to J2 J2 to J22 J22 to J2 J2 to J17 J17 to J1 J1 to J12 J12 to J1 J1 to J7 J7 to J J to J2 J2 to J2 J7 7J1 1J12 12J1 1J17 Variability/% J1 1J1 1J2 2J2 2J3 3J3 3J4 4J4 4J J Fig. 7 The mean annual groundwater recharge of 1961J199 (a) and 21J2 (b), the change in recharge (c) and the coefficient of variability (d) in the Elbe region [18] robust result leads also the conclusion that downscaling of GCM results to the regional scale and application of hydrological models is a possible and valuable procedure, when quantifying also the propagation of uncertainties. The impacts of the anticipated land use change on the water balance in the Elbe basin are not significant when only climate change is the driver for the land use change (land use scenario variant A). Additional studies show that the impacts become more severe, when global socioeconomic developments are also taken as drivers for the land use scenario (land use scenario variant B, saying that socio-economic drivers have a higher impact on land use in the target area than climate changes) [2]. Here, the management of land use becomes very important [26], for example, the types of land cover which will replace agriculture (e.g. grassland, forests or settlements). Acknowledgements The authors would like to thank all their colleagues at the Potsdam Institute for Climate Impact Research (PIK) who contributed to this paper with technical help. Part of this work was supported by the German BMBF programme GLOWA Elbe. References [1] [2] Eisenreich S J, Climate Change and the European Water Dimension. Ispra, Italy: European Commission-Joint Research Centre, 2 Hattermann F F, Krysanova V, Post J, et al. Understanding consequences of climate change for water resources and water-related secors in Europe. In: The Adaptiveness of Adv. Clim. Change Res., 28, 4 (Suppl.): 42J 49

9 Fred F. Hattermann et al.: Assessment of Water Availability in a Central-European River Basin (Elbe) Under Climate Change IWRM, an Analysis of European IWRM Research. IWA Publisher, 28: 99J112 [3] Bronstert A, Niehoff D, Burger G. Effects of climate and land-use change on storm runoff generation: present knowledge and modelling capabilities. Hydrological Processes, 22, 16 (2): 9J29 [4] Menzel L, Burger G. Climate change scenarios and runoff response in the Mulde catchment (Southern Elbe, Germany). Journal of Hydrology, 22, 267 (1/2): 3J64 [] Chiew F H S, McMahon T A. Modelling the impacts of climate change on Australian streamflow. Hydrological Processes, 22, 16 (6): 123J124 [6] Kundzewicz W Z, Budhakooncharoen S, Bronstert A, et al. Coping with variability and change: floods and droughts. Natural Resourses Forum, 22, 26: 263J274 [7] Eckhardt K, Ulbrich U. Potential impacts of climate change on groundwater recharge and streamflow in a central European low mountain range. Journal of Hydrology, 23, 284: 244J22 [8] Varis O, Kajander T, Lemmela R. Climate change and water: from climate models to water resources management and vice versa. Climatic change, 24, 66: 321J344 [9] Bronstert A. Rainfall-runoff modelling for assessing impacts of climate and land-use change, Hydrological processes, 24, 18 (3): 67J7 [1] Gerstengarbe F W, Werner P C. Simulation results of the regional climate model STAR. In: Wechsung F, Becker A, Graefe P. Integrated Analysis of the Impacts of Global Change on Water, Environment and Society in the Elbe Basin, PIK-Report 9, 2: 1436J179 ( in German) [11] Krysanova V, Muller-Wohlfeil D- I, Becker A, et al. Development and test of a spatially distributed hydrological/ water quality model for mesoscale watersheds. Ecological Modelling, 1998, 16 (2/3): 261J289 [12] Roeckner E, Bengtsson L, Feichter J, et al. Transient climate change simulations with a coupled atmosphere-ocean GCM including the tropospheric sulfur cycle. Journal of Climate, 1999, 12 (1): 34J332 [13] IPCC. Climate Change 21: The Scientific Basis. Contributions of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge University Press, 21 [14] Werner P C, Gerstengarbe F W. Proposal for the development of climate scenarios. Climate Research, 1997, 8: 171J182 [1] Orlowsky B, Gerstengarbe F W, Werner P C. A resampling scheme for regional climate simulations and its performance compared to a dynamical RCM. Theoretical and Applied Climatolog, 28, 92 (3/4): 29J223 [16] Henrichsmeyer W C C, Lohe W, Meudt M, et al. Entwicklung eines gesamtdeutschen Agrarsektormodells RAUMIS96. Endbericht zum Kooperationsprojekt. Forschungsbericht fur das BML (94 HS 21). Bonn/Braunschweig, 1996 [17] Von Lampe M. A Modelling Concept for the Long-Term Projection and Simulation of Agricultural World Market Developments-World Agricultural Trade Simulation Model WATSIM. Dissertation, University of Bonn, 1999 [18] Hattermann F F. Integrated Modelling of Global Change in the German Elbe River Basin. Dissertation, University of Potsdam, 2 [19] Becker A, Grunewald U. Flood risk in Central Europe. Science, 23, 3: 199 [2] ATV-DVWK German Society for Water Management, A.u.A.e.V. The Elbe River and the main tributaries: pressures and trends, characterization and perspectives. Hennef, 2 (in German) [21] Stanners D, Bourdeau P. Europe s Environment: the Dobrls Assessment. Luxembourg: European Environment Agency, Office for Official Publications of the European Communities, 199 [22] Hattermann F, Krysanova V, Wechsung F, et al. Integrating groundwater dynamics in regional hydrological modelling. Environmental Modelling & Software, 24, 19 (11): 139J 11 [23] Hattermann F F, Wattenbach M, Krysanova V, et al. Runoff simulations on the macroscale with the ecohydrological model SWIM in the Elbe catchment-validation and uncertainty analysis. Hydrological Processes, 2, 19: 693J 714 [24] Tarantola S. SimLab 1.1, Reference Manual, 2 [2] Hattermann F F, Gomann H, Conradt T, et al. Impacts of global change on water-related sectors and society in a transboundary central European river basinjpart 1: project framework and impacts on agriculture. Adv. Geosci., 27, 11: 8J92 [26] Niehoff D, Fritsch U, Bronstert A. Land-use impacts on storm-runoff generation: scenarios of land-use change and simulation of hydrological response in a meso-scale catchment in SW-Germany. Journal of Hydrology, 22, 267 (1/2): 8J93 Adv. Clim. Change Res., 28, 4 (Suppl.): 42J