Integrated modelling of climate change impacts on water resources and quality in a lowland catchment: River Kennet, UK

Summary

An integrated approach to climate change impact assessment is explored by linking established models of regional climate (SDSM), water resources (CATCHMOD) and water quality (INCA) within a single framework. A case study of the River Kennet illustrates how the system can be used to investigate aspects of climate change uncertainty, deployable water resources, and water quality dynamics in upper and lower reaches of the drainage network. The results confirm the large uncertainty in climate change scenarios and freshwater impacts due to the choice of general circulation model (GCM). This uncertainty is shown to be greatest during summer months as evidenced by large variations between GCM-derived projections of future low river flows, deployable yield from groundwater, severity of nutrient flushing episodes, and long-term trends in surface water quality. Other impacts arising from agricultural land-use reform or delivery of EU Water Framework Directive objectives under climate change could be evaluated using the same framework.

Introduction

Climate change could have far reaching consequences for water resources (Arnell, 2003a, Arnell, 2004, Leavesley, 1994, Pilling and Jones, 1999, Wilby et al., 1994), the physiochemistry (Hejzlar et al., 2003, Webb et al., 2003, Wilby et al., 1997) and ecology of freshwater environments (Beaugrand and Reid, 2003, EA, 2005, Hiscock et al., 2004, Moss et al., 2003, Sommer et al., 2004). Interpretations of past and future climatic impacts are confounded by other environmental drivers such as changes in agriculture or land management (Dils and Heathwaite, 2000, Whitehead et al., 2002a). Consequently, an integrated approach to impact assessment is required in which high resolution climate change scenarios drive process-based models of freshwater systems to quantify the likely hydrological, water quality and ecological impacts (Worrall et al., 2003).

Climate change impact assessment involves recognising three key aspects of uncertainty. First, there are uncertainties linked to general circulation models (GCMs), in particular: (i) future emissions of greenhouse gases, (ii) their conversion into atmospheric concentrations and (iii) subsequent radiative forcing (Allen et al., 2001, Jenkins and Lowe, 2003, New and Hulme, 2000, Webster et al., 2003). Second, there are uncertainties in the representation of climatology at regional scales, including differences between dynamical and statistical downscaling methods (Leung et al., 2003, Prudhomme et al., 2002). Third, there are parameter and structural uncertainties in the hydrochemical and ecological models used for impact assessment (Bathurst et al., 2004, Jakeman et al., 1993, Wilby, 2005).

This paper focuses on the first source of uncertainty (i.e., future climate impacts projected by different emissions scenarios/GCMs) and has two main aims. First, to develop an integrated approach to climate change impact assessment at the river catchment scale. Second, to explore the range of uncertainty in future river flow and quality indicators arising from the choice of GCM driven by a limited range of emission scenarios. The first aim will be addressed by linking water resource (CATCHMOD) and water quality (INCA) models to GCM output using statistical downscaling techniques. The second aim will be achieved by driving the CATCHMOD and INCA models with downscaled daily precipitation and evaporation series arising from three GCMs. The River Kennet provides an ideal case study because the water resources of south-east England are expected to face growing pressure from urbanisation and projected reductions in summer rainfall (Limbrick et al., 2000).