Spatiotemporal trends in reference evapotranspiration and its driving factors in Bangladesh

Abstract

This research investigates spatiotemporal variations in ETo and the controlling factor of those variations using the modified Mann-Kendall test, empirical Bayesian kriging model, Morlet wavelet analysis (MWA), and cross-wavelet transform (XWT) model relying on daily climate data sets obtained from 18 meteorological stations for the period 1980–2017. Additionally, the stepwise linear regression analysis and partial correlation coefficient (PCC) were employed to determine the variables driving the changes in ETo. The investigation exhibited a decline in annual for −1.19 mm year−1 and seasonal (−0.40 mm decade−1 during pre-monsoon, −0.47 mm decade−1 during post-monsoon, −0.50 mm decade−1 during winter) ETo, which indicates the existence of “evapotranspiration paradox” in Bangladesh, similar to many regions across the globe. The trend test depicted that despite the increase in mean temperature (MT), a noteworthy decrease in sunshine duration (SD), and wind speed (WS) are the main reasons for the reduction in ETo. Spatial analysis of ETo revealed the highest annual values in the southwest while the lowest in the northwest. Two cycles, 1–3 and 3–5 years were found significant in the annual and seasonal ETo. The outcomes revealed coherence among ETo with meteorological factors at different time-frequency bands, which is noteworthy. Stepwise regression and PCC showed that the impact of meteorological factors on ETo varies on the annual and seasonal scales where MT, RH, and SD are the major factors responsible for the variations of ETo in both annual and seasonal scales. These outcomes of the research can be advantageous for designing irrigation and management of sustainable water resources to mitigate climate change impacts as well as controlling anthropogenic activities.