Analysis of climate change impact on precipitation in Danjiangkou reservoir basin by using statistical downscaling method

General circulation models (GCMs) are often used to assess the impact of climate change. As they operate on coarse scales, the simulation results obtained from GCMs are not particularly in accordance with comparatively smaller river basin scales. A statistical downscaling method based on a least square support vector machine (LS-SVM) is proposed for downscaling daily total precipitation series from GCMs outputs for the Danjiangkou reservoir basin in China. The downscaling method is developed and validated using large-scale predictor variables derived from the National Center for Environmental Prediction - National Center for Atmospheric Research (NCEP/NCAR) reanalysis data set. The performance of the LS-SVM downscaling method is also compared to a well-known statistical downscaling model (SDSM). The downscaling results suggest that the LS-SVM is an efficient method for downscaling daily precipitation series in the study region. The statistical relationship obtained is used to project future precipitation from CGCM2 for IPCC SRES A2 and B2 scenarios. The downscaling results, corresponding to both scenarios, show that there is an increase in the average annual precipitation downscaled from CGCM2 by about 277.51 mm for the IPCC SRES A2 scenario and by about 157.65 mm for the IPCC SRES B2 scenario, by the 2080s. Copyright 2009 IAHS Press.

Analysis of climate change impact on precipitation in Danjiangkou reservoir basin by using statistical downscaling method Conference Paper