Lapse rate adjusted bias correction for GCMs precipitation data: An Application to the Monsoon Asia Region

Abstract

Abstract Bias correcting General Circulation Models (GCM’s) data is necessary before it is used in the climate change impact assessment studies at regional scales. Most of the bias correction methods consider raw GCM’s and locally observed data for correcting the systematic bias in the GCM’s data. However, highly varying topographic conditions and associated lapse rate properties perhaps impact on the bias correction process. Therefore, we have introduced a novel bias correction method where raw GCM data was first adjusted for local lapse rates and later was bias corrected with a simple linear regression coefficient. Monsoon Asia region was used as the study region to evaluate the proposed lapse rate regression (LR-Reg) based bias correction along with linear scaling (LS) and quantile mapping (QMap) bias correction methods. Asian Precipitation Highly Resolved Observational Data Integration Towards Evaluation of Water Resources (APHRODITE) precipitation data was used as reference data to evaluate LS, QMap and LR-Reg bias correction methods. The comparison results show that the LR-Reg bias correction method was more promising and reduced significant bias from GCM’s precipitation data. The relative reduction in the Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) values by LR-Reg over LS method was upto 30% while the relative reduction in MAE and RMSE values by LR-Reg over QMap was upto 50%. Future projected precipitation under shared socio-economic pathways (SSP245 and SSP585) scenarios showed that the increase in precipitation was upto 50% mostly in the northern and central parts of China and in the Himalayan belts.