Comparison of hydrological simulations of climate change using perturbation of observations and distribution-based scaling

Projected climate change effects on groundwater and stream discharges were investi-gated through simulations with a distributed, physically based, surface water-groundwater model. Input to the hydrological model includes precipitation, reference evapotranspira-tion, and temperature data of the HIRHAM4 regional climate model (RCM). The aim of this study was to determine whether the choice of bias-correction method, applied to the RCM data, affected the projected hydrological changes. One method consisted of perturbation of observed data (POD) using climate change signals derived from the RCM output, while the other consisted of distribution-based scaling (DBS) of the RCM output. Distribution-based scaling resulted in RCM control period data closely approaching the observed climate data and thereby considerably improved the simulation of recharge and stream discharges. When comparing the simulations using both methods, only small differences between the projected changes in hydrological variables for the scenario period were found. Mean annual recharge increased by 15% for the DBS method and 12% for POD, and drain flow increased by 24 and 19%, respectively, while the increases in base flow were similar (7%). For both methods, daily stream discharges up to and including the median showed little change, while increases occurred for the higher quantile values. This study showed that the choice of bias-correction method did not have a significant influence on the projected changes of mean hydrological responses in this catchment, although further analysis is necessary to determine whether extremes are affected. Furthermore, the characteristics of the hydrological system likely reduced the sensitivity of the projected changes to the choice of method.