Analysis of extreme precipitation in Bangladesh based on reanalysis data and climate models

Understanding extreme precipitation is essential for mitigating the associated risk. This study analysed eight Coupled Model Inter-comparison Project phase 6 (CMIP6) models by capturing daily, seasonal, yearly, and extreme precipitation over Bangladesh during the period 1961-2014, treating ECMWF Reanalysis v5 (ERA5) rainfall reanalysis data as an observational reference. Sixteen extreme precipitation definitions were found in the literature, implemented, and used in a comparative analysis in Bangladesh, one of the most vulnerable countries to climate change. Definitions used in the literature, duration definitions, such as the Consecutive Dry Days (CDD) and the Consecutive Wet Days (CWD), frequency definitions, such as the days when precipitation is at least 10 mm, 20 mm, 30 mm, at least 65 mm but less than 100 mm, at least 100 mm but less than 115 mm, at least 115 mm but less than 205 mm, and at least 205 mm, and intensity definitions, such as SDII, RX1day, RX3day, RX5day, PRCPTOT, R95pTOT, and R99pTOT. Temporal trends in extreme events identified by 16 different methods were assessed using both the modified Mann-Kendall (MK) method and Sen's Slope Estimator (SSE).

Analysis of the ERA5 reanalysis product indicated a 0.33-day-per-year increase in consecutive dry days (CDD) during 1961-2014 over Bangladesh, revealing that Bangladesh may be drying out at a faster rate than previously anticipated. For average daily wet-day precipitation intensity (SDII), ERA5 and MIROC6 showed a decrease at annual rates of -0.07 and -0.06 mm per day, respectively. If these trends continue over the next few years, Bangladesh may encounter water scarcity. In addition, a decrease at a rate of -0.09 days per year in R65mm was found in ERA5, while other frequency definitions do not have statistically significant trends. MIROC-ES2L has the closest agreement to ERA5 in terms of Percentage BIAS (PBIAS), R-Squared (R2), Root Mean Squared Error (RMSE), and Mean Squared Error (MSE) metrics, followed by EC-Earth3-Veg-LR, IPSL-CM6A-LR, and MIROC6. Several recommendations for future studies and improvements are also included in this study. Results presented in this study, specifically a faster drying rate, fewer days with 65-100 mm of daily precipitation, and possibly less monsoon seasonal precipitation are important for the development of adaptation strategies in Bangladesh.