A methodology to assess the combined effect of climate change and groundwater overexploitation over the Upper Guadiana basin, Spain

There is a growing concern about the combined effect of climate change and groundwater overexploitation on the availability of water resources in the Upper Guadiana basin (UppGb) in central Spain. General Circulation Models (GCMs) are used to evaluate the possible impact of climate change based on future scenarios of greenhouse gas emissions. However, the output of these models cannot be applied directly to hydrological models because their spatial resolution is coarse and because their simulated precipitation is highly biased. A stochastic downscaling method for generating daily spatial rainfall fields was developed. The model termed Stochastic Rainfall Generating Process (SRGP) incorporates two major non-stationarities -- changes in the frequencies of different precipitation generating mechanisms (frontal and convective), and spatial non-stationarities caused by the interactions of meso-scale atmospheric circulation patterns (ACP) with topography (orographic effects). SRGP was developed to incorporate good climate outputs simulated by GCMs (i.e. ACP), and actual observations. These capabilities enabled us to (1) use SRGP as a downscaling method for climate change impact study, and (2) generate stochastic rainfall fields conditioning to the information of rain gauges. The latter capability was used to investigate the effect of rainfall spatial variability (RSV) on the hydrological response in the UppGb. RSV exerted a major influence on the response of the system especially on the groundwater recharge and the aquifer related responses.GCMs considered in the fifth assessment report of the Intergovernmental Panel on Climate Change were used to evaluate the impact of climate change. The RCP8.5 future emission scenario (GCM-RCP8.5) and the GCM historical control (GCMH) were selected. The climate change was assumed to be the accumulated effects of increases in Temperature, changes in annual and climatological ACP frequency, and changes in probability and volume of rain. Transformations were applied to correct the bias in the temperature, probability and volume of rain, whereas the ACP sequences were used directly. The SRGP method was employed as a rainfall downscaling method for the GCMs. GCMH was used to evaluate the hydrological response obtained with GCMs as driving climate variables, introducing the concept of stochastic equivalence. This evaluation was based on the comparison of the hydrological response obtained with actual observations and transformed (b