Spatio-temporal characterization of rainfall in Bangladesh: an innovative trend and discrete wavelet transformation approaches

Abstract

The analysis of detailed characteristics of rainfall is of utmost importance for understanding water resources, irrigation, and agriculture. This study investigated detailed characteristics of rainfall of Bangladesh from 1966 to 2019. Rainfall features like the precipitation concentration index (PCI) and seasonality index (SI) were estimated to characterize spatial patterns of rainfall regimes, and innovative trend analysis (ITA) and percent bias (PBIAS) were used to detect the trend, and its reliability was tested by using the Mann–Kendall (MK) or modified Mann–Kendall (mMK) test. The magnitude of changes was computed by using Sen’s slope estimator (Q), and discrete wavelet transform (DWT) was employed to find out the dominant periodicity of the trend of annual series. SI and PCI revealed that rainfall is mainly seasonal and markedly with the long dry season, and the distribution of rainfall is irregular for the major portion of Bangladesh. The result of ITA and PBIAS showed a similar trend for all the stations and time scales. ITA showed that 23 stations experienced significant (α = < 0.05) increasing (39.13%) and decreasing (39.13%) trends in annual rainfall. For the seasonal scale, post-monsoon rainfall is dominated by significant increasing trend (60.87%), whereas winter rainfall is dominated by significant decreasing trend (65.28%); besides, both decreasing and increasing trends were found in pre-monsoon and monsoon in Bangladesh. The analysis of DWT confirmed that 4- to 16-year periodic cycle was the dominant component for annual rainfall. The results of detailed analysis of rainfall will be helpful for policymakers and scientists to focus on regional-scale planning about flood and drought situation of the country that will be ultimately helpful for agricultural development and environmental planning.