Evaluation of Spatial Bias Correction for Impact Assessment of Climate Change on Floods using the d4PDF in the Chao Phraya River Basin, Thailand.

Abstract

The spatial bias heterogeneity of streamflow between the upstream main dams and the basin's outlet was investigated in a large ensemble climate simulation (with> 5000-year time slice simulation) database d4PDF, a commonly used dataset for impact assessment. After that, their combined correction. (two-step bias correction) was investigated and compared to a simpler, but more practical bias correction performed solely at the basin's outlet. The two-step bias correction clarified that, for d4PDF, the upstream bias impact was assumed insignificant because of the huge influence of downstream bias in the basin, hence the two methodologies resulted in similar future forecasts in the CPRB. However,the two-step bias correction of the global ensemble data was applied to the river discharge for impact estimations on floods with decreased uncertainty, revealing that a + 4K warmer climate will boost 100-year floods by 1.1-1.6 times in the CPRB than the prior climate (1961-2010). Future floods are projected to start in September, which is typically the start month for the present climate, as well as equally high in August, with floods lasting 10 to 50 days.