Performance Evaluation of Bias Correction Methods and Projection of Future Precipitation Changes Using Regional Climate Model over Thanjavur, Tamil Nadu, India

The use of regional climate models in climate change risk assessments is uncertain because of the threat of biases. Prior to using simulations, some corrections must be made. The aim of this study is to evaluate the effectiveness of simple and complex correction approaches on CORDEX (Coordinated Regional Climate Downscaling Experiment) precipitation datasets over Thanjavur in Tamil Nadu, India. The various bias correction approaches were assessed using the two-sample cross-validation technique, and the best method was chosen using statistical metrics such as the correlation coefficient, root mean square error, and percent bias for correcting the future scenario datasets. According to the findings, all methods significantly improved raw simulated estimations after correction. The Delta Change method was used to adjust the RCP2.6 and RCP8.5 scenarios datasets for this study region because it produced good agreement (0.98) and low wet bias (1%) against observed precipitation. In addition, this study found that the annual mean precipitation projections showed an increase in precipitation intensity of between 4.2 and 6.8% under the RCP2.6 scenario and between 14 and 20.1% under the RCP8.5 scenario, compared to observed data (1976-2005). Furthermore, the analysis showed that most years in the twenty-first century will experience heavy precipitation, resulting in floods. This research has offered reliable knowledge about future precipitation to environmentalists, urban administrators, and policy-makers, enable them to make plausible strategies to mitigate the detrimental effects of climate change.