Future Influence of Tropical Indian Ocean Sea Surface Temperature on Extreme Wet Days Rainfall During March–May in Tanzania Using <scp>CMIP6</scp> Models

Abstract

ABSTRACT This study assesses the performance of CMIP6 models in simulating monthly sea surface temperature (SST) over the Tropical Indian Ocean (TIO) and extreme wet days (EWDs) in Tanzania using daily precipitation data (1981–2014) during March–May. Observational datasets from daily CHIRPSv2.0 and monthly ERSST were employed for validation. The study also examines future climatological mean and trend of SST and EWD during 2015–2048 and 2049–2082. Furthermore, the future influence of the TIO SST and Indian Ocean Dipole (IOD) on EWDs during 2049–2082 was also assessed. Results from historical models display varying accuracy, with most models underestimating EWDs, and provide relatively accurate simulated SST. ACCESS‐CM2, CMCC‐CM2‐SR5 and CMCC‐ESM2 models demonstrate high variance in EWDs (1.0–2.0), while simulated SST showed greater variance, peaking at 7.16 for CMCC‐CM2‐SR5, with reduced biases. ACCESS‐CM2, CMCC‐CM2‐SR5, MIROC6, NESM3, and EC‐Earth3‐Veg‐LR models exhibit moderate to high Taylor skill score values &gt; 0.5 and statistically significant correlations at the 99% confidence level, indicating a strong ability to replicate observed spatial and temporal patterns. These models also provide valuable insights into future climate projections during 2015–2048 and 2049–2082. Projections under SSP2‐4.5 and SSP5‐8.5 indicate increasing trends in EWDs, especially under SSP5‐8.5, during 2049–2082, with values reaching +0.8 days per decade. Southwestern and coastal regions exhibit the strongest positive trends at the 90% confidence level, while some western areas show slight decreases. Ensemble results of SST indicate consistent and robust warming signals, particularly under SSP5‐8.5 during 2049–2082. Singular value decomposition reveals substantial connections between TIO SST and EWDs during 2049–2082 with dominant positive anomalies across models and significant correlations ranging from R = 0.58 to R = 0.78. Furthermore, IOD in January is highly expected to influence MAM EWDs during 2049–2082, with notable significant spatio‐temporal positive correlations at the 90% confidence level for EC‐Earth3‐Veg‐LR, CMCC‐CM2‐SR5 and the ensemble mean. These results provide critical insights for future planning in different sectors.