Evaluation of gridded dataset in estimating extreme precipitations indices in Pakistan

Anthropogenic-induced climate change triggered increased extreme precipitation events, posing a significant threat to vulnerable regions like Pakistan. However, the lack of reliable long-term in-situ records hampers the monitoring of climatic extremes in the country. This study assessed the skill of four daily gridded precipitation datasets (APHRODITE, CHIRPS, CPC, and PGF) in tracking changes in precipitation extremes from 1985 to 2016 at 42 meteorological stations across Pakistan. This study employed Sen slope estimator in determining the change and the Mann–Kendall (MK) and its modified version (mMK) to test the significance of the changes. Spatial analysis of trends based on in-situ and gridded datasets revealed substantial increases in most precipitation extremes. However, there were large variations in the skill of gridded products in estimating precipitation extremes. The APHRODITE and PGF overestimated or underestimated trend significance, respectively. The CPC outperformed the others, exhibiting an RMSE of 3.96, Pbias of 18.5, NSE of 0.23, md 0.53, and a correlation coefficient of 0.6 in estimating changes at a 95% confidence level. Moreover, CPC also demonstrated superior performance in estimating trends at a 99% confidence level. It also performed best in estimating mMK trend significance at a 99% confidence level. Therefore, the study recommends CPC for tracking extreme precipitation trends in the region.