Bias Correction of CMIP6 Global Circulation Model Rainfall Data Using Scale Distribution Mapping Technique with Spatial Rainfall Data over Thailand

Abstract

Spatial bias correction of precipitation data is crucial for climate change studies, as rainfall data from Global Climate Models (GCMs) exhibit biases when compared to observational data. This study aims to develop an approach for correcting spatial biases in precipitation data from CMIP6 Global Climate Models, utilizing scale-adjustment techniques combined with spatial distribution corrections. The study employed data from 15 GCMs, rainfall observations from the Thai Meteorological Department, and blended satellite infrared radiation and station observation data. Results from the blended precipitation data correction show that correlation coefficients (R) increased from 0.65-0.67 to 0.68, Root Mean Square Error decreased from 122-126 mm to 84-85 mm, and percentage bias was reduced from -25% to -15% to approximately -5% to 4%. The bias correction of GCM precipitation data demonstrated similar performance in both calibration and validation periods, with all statistical metrics showing improvement. The overall performance evaluation revealed that models BCC-CSM2-MR, CanESM5, and CESM2 provided the best correction results in terms of bias reduction. This study significantly contributes to climate change adaptation planning in Thailand, as the bias-corrected data can be used more effectively for future climate predictions.