Evaluation of Simple Statistical Downscaling Methods for Monthly Regional Climate Model Simulations with Respect to the Estimated Changes in Runoff in the Czech Republic

• Published in: Water resources management 2013-12, Vol.27 (15), p.5261-5279
• Main Authors: Hanel, Martin, Mrkvičková, Magdalena, Máca, Petr, Vizina, Adam, Pech, Pavel
• Format: Article
• Language: English
• Subjects: Atmospheric Sciences; Bias; Civil Engineering; Climate change; Climate models; Computer simulation; Drought; Droughts; Earth and Environmental Science; Earth Sciences; Earth, ocean, space; Environment; Environmental impact; Estimates; Exact sciences and technology; Geotechnical Engineering & Applied Earth Sciences; Hydrogeology; Hydrology; Hydrology. Hydrogeology; Hydrology/Water Resources; Planning; Precipitation; RCM; Runoff; Sampling; Simulation; Statistical methods; Studies; Water resources; Water resources management
• ISSN: 0920-4741; 1573-1650
• DOI: 10.1007/s11269-013-0466-1

An ensemble of fifteen regional climate model (RCM) simulations has been used to estimate the climate change impacts on runoff and several drought characteristics for 250 basins in the Czech Republic. The scenario series of precipitation and temperature have been derived with four simple statistical downscaling methods (SDMs): the delta change and bias correction method, both in two alternatives considering the changes/biases in the mean, and in the mean and variance. Bootstrap resampling has been used to assess the effect of sampling variability and the differences in the estimated changes in runoff obtained by different SDMs were evaluated. Further simplification of the SDMs (spatial-average changes/biases and ensemble-average changes in precipitation and temperature) have been considered. It was shown, that given the spread between the projections of individual RCM simulations and the sampling variability, the differences in the estimated changes in mean runoff between the SDMs are not very large. The same partly holds also for the effect of spatial averaging. In general, the SDMs accounting for variability have led to smaller decrease (or larger increase) in runoff and the decrease was also smaller for bias correction methods than in the case of delta change methods. In contrast to changes in mean runoff, significant differences between the estimates based on different SDMs were found for the drought characteristics. In addition, the averaging of the changes in precipitation and temperature over the RCMs resulted in much stronger decrease in runoff than indicated by ensemble average changes in runoff.