Selection of CMIP5 multi-model ensemble for the projection of spatial and temporal variability of rainfall in peninsular Malaysia

Abstract

This study uses a multi-model ensemble (MME) for the assessment of the spatial and temporal variations of rainfall in peninsular Malaysia under climate change scenarios. The past performance approach was used for the selection of GCM ensemble from a pool of Coupled Model Intercomparison Project Phase 5 (CMIP5) GCMs. The performances of four bias correction methods, namely, scaling, gamma quantile mapping, generalized quantile mapping, and power transformation were assessed to select the most suitable method for the downscaling of daily rainfall of selected GCMs based on APHRODITE rainfall at a spatial resolution of 0.25° × 0.25°. The downscaling model was used for the projections of daily rainfall for the period 2010–2099 for four representative concentration pathways (RCP) scenarios, namely, RCP2.6, RCP4.5, RCP6.0, and RCP8.5. Random forest regression algorithm was used to develop the multi-model ensemble (MME) mean of GCM-projected rainfall for different RCPs in order to show the changes in rainfall for three future periods, 2010–2039, 2040–2069, and 2070–2099. The results revealed four GCMs, BCC-CSM1.1(M), CCSM4, CSIRO-Mk3.6.0, and HadGEM2-ES as the most suitable GCMs for the projection of daily rainfall of peninsular Malaysia. The power transformation was found as the most suitable method for the correction of biases in GCM daily rainfall. The MME mean of projected rainfall showed the increase in rainfall in peninsular Malaysia for all the scenarios and future periods. The maximum increase in annual rainfall was projected by 15.72% during 2070–2099 for RCP8.5. The variability of future rainfall was also found to increase along with mean rainfall. The increase in rainfall variability was projected by 26.15% for RCP8.5 during 2070–2099. The spatial pattern of rainfall changes revealed more variability in future rainfall in the northeast where frequency of hydro-climatic disasters is higher compared to other regions. The results indicate the possible increase in hydro-climatic disaster in peninsular Malaysia due to climate change.