Performance assessment of bias correction for precipitation data from General Circulation Models using Linear Scaling and Quantile Mapping Method in the upper Chao Phraya River Basin

Abstract

The present research proposed to assess the performance of bias correction (BC) techniques for precipitation in the upper Greater Chao Phraya River Basin. The research used observational and General Circulation Model (GCM) data from three models in the Sixth Phase of the Coupled Model Intercomparison Project (CMIP6). The models are CanESM5, CNRM-CM6-1, and INM-CM-5.0. The linear scaling (LS) and quantile mapping (QM) methods were used to rectify the bias, while RMSE, NSE, and PBIAS were utilized to assess the efficacy of the approaches. The calibration period spanned from 1995 to 2004, whereas the validation phase extended from 2005 to 2014. During the calibration period, the LS method had RMSE values of 38.149 – 40.681 mm., NSE values of 0.752 – 0.782, and PBIAS values of 0.044 – 0.095%. During the validation period, they were 54.786 to 58.562 mm., 0.652 to 0.696, and 23.334 to 26.207%, respectively. On the other hand, the QM method showed RMSE, NSE, and PBIAS values between 58.452 and 66.151 mm., 0.343 and 0.487, and 4.495 and 9.101%, respectively, during the calibration period. The validation period was between 71.533 and 118.256 mm., -0.418 to 0.481, and 14.034 to - 35.022%, respectively. The performance evaluation of bias correction methods revealed that the LS method maintained stable seasonal trends, while QM adjusted extreme rainfall in certain periods but remained fluctuating during the dry season. This research recommends using LS for long-term studies and QM for heavy rainfall event analysis to support water resource management planning under climate change conditions.