Selection of Predictors for Downscaling Rainfall in the East Coast of Peninsular Malaysia

Statistical downscaling techniques are widely used for downscaling coarse resolution GCM projections at local or regional scales. One of the major challenges in statistical downscaling is to select GCM variables that have strong relation with local climate. The objective of the present study is to compare various statistical approaches, such as canonical correlation analysis, principle components analysis, simple correlation, and multiple step regression for the selection of appropriate set of GCM variables for downscaling rainfall in the east coast of peninsular Malaysia, which is considered as the most vulnerable zone to climate change in Malaysia. Twenty-six large-scale atmospheric variables reanalysed as a proxy for current observation of GCM variables by National Center for Environmental Prediction (NCEP) at forty-two grid points surrounding the study area were used for this purpose. The results reveal that the NCEP variables at grid points mostly located in the northeast of the study area have higher influence on the rainfall in the study area. The study also indicates that three NCEP variables, namely, relative humidity at 850 hPa at northwest grid point, vorticity at southeast grid point, and relative humidity at 850 hPa at southeast grid point have high correlation with rainfall at all stations under study. It is expected that the finding of the study can be useful for downscaling and future projection of rainfall in the east coast of peninsular Malaysia.

General circulation model; Statistical downscaling; Rainfall; East coast of peninsular Malaysia