A study of the impact of climate change on local precipitation using statistical downscaling

Abstract

This study investigates the potential of projecting regional precipitation over Yunlin County using a singular‐value‐decomposition (SVD) based statistical downscaling scheme to cope with climate change; the change rate of local precipitation and its connection with the variability of large‐scale circulation are also discussed. It is found that rainfall over Yunlin County is closely tied to the large‐scale circulation over the East Asian monsoon region, and that general circulation models (GCMs) perform reasonably in simulating the mean states of sea level pressure (SLP) and meridional wind field at 850 hPa (V850) over this region. Consequently, the two large‐scale variables, SLP and V850, both of which are taken from seven GCMs involving the 20C3M, A1B, and B1 scenarios, are used as predictors for downscaling. The result shows that the downscaling schemes based on these predictors show a skillful and stable performance in reproducing historical seasonal rainfall anomalies over Yunlin County, especially when the strategies of using two predictors and multiple model ensemble (MME) are considered. It also shows that comparing the mean states of 20C3M‐downscaled and scenario‐projected seasonal local rainfall in the coming 36 years with respect to the period of 1975–2000 reveals a suppressive pattern in the wet season and an increasing pattern in the dry season; the finding may be associated with the strength of large‐scale circulation under different scenarios. Overall, this study demonstrates that useful information concerning the impact of climate change on regional precipitation can be obtained by downscaling schemes using outputs of GCMs as predictors.