Evaluating climate models to analyze drought conditions in the western region of Bangladesh

Abstract

Being susceptible to natural disasters, Bangladesh is one of the most disaster-prone countries. Among the various natural calamities, droughts are a frequent occurrence in the western region of the country. Hence, this study first compared the efficacy of five bias-corrected Regional Climate Models (RCMs) - MIROC, NOAA, MPI, IPSL, and CCCma - for the Western region of Bangladesh with observed monthly precipitation and subsequent SPI values. Various evaluation methods- RMSE, Taylor Diagram, Mann–Whitney U Test, and t-test, were applied to precipitation and SPI3 values for the historical base period. Through these analyses, the MIROC model exhibited the highest level of accuracy. Accordingly, future projections for short-term droughts (SPI-3) and their characteristics were conducted using the ensemble of top three climate model under the RCM 8.5 W/m2 scenario. Short-term droughts are anticipated to become less frequent and severe in the 2060s and 2080s compared to the 2020s. Nonetheless, North-West region is projected to be more drought-prone than South-West until 2100. This research shows the importance of evaluating the better-fitting RCMs for assessing historical droughts and making reliable projections for the future. The methodology and findings can be employed in evidence-based decision-making and applied in other drought-prone areas to understand future drought risks.

Keywords

• RCM
• SPI
• Drought prediction
• Taylor diagram
• Mann–Whitney U test
• RCP 8.5