Changes and causes of rainfall extremes over northeast Bangladesh under the warming world

Abstract

Northeast Bangladesh is located in the Meghna basin. The local community of northeast Bangladesh is very much concerned about the flash flood as it damages the Boro rice during the pre-monsoon season. Hence, long-term or seasonal flood prediction is essential for many management decisions in agriculture and food security, water and disaster risk reduction.El Niño–Southern Oscillation (ENSO) can be significant concerning seasonal flood prediction.They are also apprehensive about the impact of climate change on the long-term rainfall pattern, which could influence the hydro-climatic extremes like floods, droughts, and other extreme events. Therefore, this study focused on determining the changes of future rainfall extremes under the warming world and to find out if there exist any teleconnectionsbetween the pre-monsoon rainfall and large scale process e.g., ENSO.

The trends of extreme rainfall indices over northeast Bangladesh during the pre-monsoon and monsoon seasons were analyzed for the period 1984-2016. With access to the highest number of available rainfall stations in northeast Bangladesh, the trends of extreme rainfall events were investigated using the Mann-Kendall trend test and Sen’s slope estimator. The Standard Normal Homogeneity and the Pettitt tests were used in appraising the quality of the data. Among seven stations, the rainfall of Sunamganj station was found inhomogeneous and was not considered for trend analysis. Most of the rainfall extremes indices showed a decreasing trend during the pre-monsoon as well as monsoon season, with the most significant reduction during the monsoon season. The total seasonal rainfall and consecutive wet days showed a decreasing trend in both seasons. The consecutive dry days (CDD) showed an increasing trend in the monsoon season only. Moreover, a decreasing trend was observed in one-day maximum rainfall (RX1), five-day maximum rainfall (RX5), the intensity of the daily rainfall over 25 mm (R25) during the pre-monsoon and 50 mm (R50) during the monsoon.

The future trend and changes in rainfall extremes for northeast Bangladesh were examinedfor the periods of 2041-2070 and 2071-2099. Six regional climate models (RCMs) over the coordinated regional downscaling experiment (CORDEX) South Asia domain considering two representative concentration pathways (RCPs), namely RCP4.5 and RCP8.5, were used for this purpose. The multi-model ensemble mean of the extreme rainfall indices was generated using the Bayesianmodel averaging (BMA) approach. The BMA mean is a weighted average related to each RCM’s predictive skill during the training period.

Most of the extreme indices showed an increasing trend during the pre-monsoon season for all futuretime slices except 2071-2099 for RCP4.5, while they showed a decreasing trend for the baseline period (1976-2005) for the same season. Most of the extreme indices showed a decreasing trend during the monsoon season for all future time slices, which is similar to the baseline period.

The seasonal rainfall, together with other extreme indices, is expected to increase in the future relative to the baseline period, except for a decrease of CDD during both pre-monsoon andthe monsoon season. The average pre-monsoon rainfall of the study area is projected to increase by 12.93% and 18.42% under RCP4.5 for the period 2041-2070 and 2071-2099, respectively. The increase of the pre-monsoon rainfall for those periods will be 18.18%, and 23.85%, respectively under RCP8.5. The average monsoon rainfall of the study area is projected to increase by 4.96% and 2.27% under the RCP4.5 for the period 2041-2070 and 2071-2099, respectively. These increases in monsoon rainfall for that period will be 6.56% and 6.40%, respectively for RCP8.5. It was also noted that all the extreme indices except consecutive wet days (CWD) are expected to increase significantly at the 95% confidence level during the pre-monsoon season. Therefore, the study area is expected to experience more frequent floods in the future in both the pre-monsoon and monsoon seasons as a consequence of climate change. In particular, the intensity and the magnitude of the flash flood in the pre-monsoon are expected to increase in the future as the extreme indices are likely to increase significantly in the pre-monsoon season.

The present study also examined the relationship between El Niño Southern Oscillation (ENSO) and pre-monsoon rainfall, particularly in April over the Meghna basin and its response under the warming world. Firstly, the relationship between April rainfall over the Meghna basin and the heat low over central India during April was determined. The heatwave creates a low-pressure system (Cyclonic) in central India and the high-pressure system (Anticyclonic) in the Bay of Bengal. These two-systems trigger the south-westerly moisture flow from the Bay of Bengal towards the Meghalaya Mountain region and cause heavy rainfall over the Meghna basin. The result showed that there is a high inverse correlation (ρ> 0.55) between April rainfall over the Meghna basin and heat low over central India during April. Considering these findings, the relationship of different ENSO indices (e.g., ESOI, SOI, ONI, MEI) for several months (e.g., January, February, and March) with April rainfall was determined. It was found that the Oceanic Niño Index (ONI) during January has the highest correlation value (ρ=0.52) and the maximum spatial coverage of the correlation value for which it is statistically significant (ρ=0.32 at the95% confidence level)with April rainfall. It was also found that in most of the cases, floods in April occurred either during El Niño events or even neutral eventsbut not during the La-Nina events during January. Finally, the relationship between ONI index during January and the heat low over central India during April was determined and found that there is a high inverse correlation of heatwave over central India in April with ONI index during January (ρ=-0.55). This infers that if the ENSO index during January is positive, there is a possibility of heat low over central India during April. On the other hand, it was shown that the heat low over central India has an impact over heavy rainfall during April over the Meghna basin. Therefore, it can be stated that the El Niño during January is related to heavy rain during April over the Meghna basin.

As ENSO would impact flash floods over the Meghna basin, how ENSO would be influenced under the warming world was also studied. The result showed that the intensity of El Niño event increases with global warming under extreme scenario while it is opposite in the case of a La Niña event. However, there is no significant change in ENSO amplitude under the warming climate. Hence, Northeast Bangladesh would experiencemore frequent flooding in April as the El Niño event is expected to increase remarkably in the future.