Quantification of uncertainties in projections of extreme daily precipitation simulated by CMIP6 GCMs over homogeneous regions of India.

Global climate model (GCM) projections are subject to significant uncertainties. Quantifying uncertainties in climate change projections improves credibility and makes climate data more reliable. This study aims to quantify the uncertainties in projected extreme precipitation during the 21st century over the homogeneous rainfall regions of India simulated by Coupled Model Intercomparison Project Phase 6 (CMIP6) GCMs. The percentile‐based square root error variance (SREV) method estimates model, scenario and ensemble uncertainties in projections of extreme precipitation. The uncertainty is investigated at four thresholds: 95th, 99th, 99.9th and 100th percentiles. The results show that the wet northeast region has a greater SREV, which is consistent with previous studies. At 99th and 99.9th percentiles, relative model SREV is dominant over the northeast (NE) region. However, at the 95th percentile high relative model SREV is found over the northwest (NW) region during southwest (June, July, August and September) and NE (October, November and December) monsoon seasons. Model uncertainty is the main source of uncertainty, followed by scenario and ensemble uncertainties. The study indicates that the arid NW region in India has a higher level of uncertainty than other regions with homogeneous rainfall. These findings will assist policymakers in planning infrastructure development in arid regions of India. [ABSTRACT FROM AUTHOR]