4. Discussion
This study aims to explore the spatiotemporal variation of ETref and other climatic variables that influence ETref. We focused to identify the seasonal and annual trend patterns of ETref of the study area. It was observed that the MT had an increasing trend in most locations in the country. Many studies also reported that global temperature was increasing substantially over the last few decades in an alarming rate which resulting not only altering the global water budget but also the more frequent drought and flood (IPCC 2007; Feng et al. 2014; Islam et al., 2019). Recently, a study by Islam et al. (2017) also found frequently and intense droughts phenomena occurring in Bangladesh that also support our result of an increasing trend of MT which might result from the severe change of climatic parameters throughout the country.
Wind speed and sunshine hours were reducing considerably in most of the regions and it can be alarming under climate change scenario. Syeda and Nasser (2012) reported that a significant decreasing trend of the sunshine hour had been observed over Bangladesh that might be the result of the increasing air pollution in our atmosphere and also the soared anthropogenic activities for the past few decades. Lian et. Al. (2015) studied that characteristics of changing climate ought to change the water cycle which can be resulted in frequent occurring of drought and flood also indicate the same changes occurring like our study. Most areas in Bangladesh experiencing a considerable declining of SD coupling with the rising temperature. Many scientists i.e. Jhajharia et al. (2012) and Chu et al. (2017), considered that the rising of aerosols and clouds resulting from air pollution was the main reason behind the lowering SD.
A significantly decreasing trend of ETref also found in the study done by Rahman et al. (2019) which also supports our findings. A slightly increasing trend of ETref was found northeast to the southwest part of Bangladesh that contradicted the studies done by Yang et al. (2013). But Mojid et al. (2015) found an increasing trend of ETref in Bogra and Rangpur stations similar to our studies. Some published works like Fan and Thomas (2013) and Tao et al. (2015), it was evident that SD played a pivotal role in quantifying ETref variations in several locations of Bangladesh. According to Tao et al. (2015), it is evident that SD and WS are the most prominent parameters than MT and RH which act behind the disparity of ETref.
Nonetheless, various scholars in their works (Chu et al. 2017; Li et al. 2018) confirmed that SD and WS act as a key influencer about ETref which is analogous to our study. On the other hand, The PCC analysis revealed that SD, MT, and RH were the main drivers of ETref and WS had little effect. The declining amount of RH in the air also influencing the reduced ETref as like as the other climatic variables. Though an overall decreasing trend of ETref was found in the study area coupled with increasing temperature which should be taken under consideration by the policymaker to manage water for future agricultural water practice and drought risk reduction phases.
The decrease in SD and WS was found in the study area which should resultant a decline ETref but a significant increase in MT offset the impact and followed the “evapotranspiration paradox” in Bangladesh which is supported the study conducted by Herath et al. (2017), Feng et al. (2017). Studies by Roderick and Farquhar (2002) establish the fact that a decline in SD resulted in the lessening ETref which were similar to our study.
The  studies done by Rahman et al. (2019) and Mojid et al. (2015) showed that MT, SD, and WS had a significant influence on the ETref occurring that also support our study. The WS and SD were declining notably in some parts of Bangladesh. In the meantime, in Bangladesh strange fluctuations were noticed in MT and SD that can be triggered by global warming. Changes in hydrological processes resulted from changing climate ought to enhance the frequency of drought, flood and reallocating the water resources in Bangladesh (Islam et al., 2019).
5. Conclusion
We found decreased in ETref with time across Bangladesh which is similar to the “evapotranspiration paradox” in different portions of the globe. Though a significant rise in MT, but SD, R, and WS was found as the key influential factors of decreasing annual and seasonal ETref in Bangladesh. This study indicates that the declining SD and WS would gradually decrease ETref in Bangladesh. Spatially, the southern part of Bangladesh reflected the higher ETref while the northwestern region exhibited the lower ETref on an annual and seasonal scale. In a seasonal scale, the ETref in the western part tends to be elevated in the pre-monsoon compared to other seasons. Moreover, it is interesting to note that decline in ETref across Bangladesh in the last four decades has been prominent during the all seasons except for monsoon. The declining ETref might have implications in atmospheric evaporation balance, irrigation, and reservoir management required in different regions of Bangladesh. A decrease in ETref and variations in water shortage can influence the hydrologic cycle in the study region which has created an environmental crisis. The outcomes of the research can be useful for designing adaptation measures of climate change and also mitigating drought effects on agricultural water resources and ecosystems. The coupling of satellite-derived ETref and climatic datasets can be utilized for better evaluating spatial variations in ETref and climatic variables in the future study. Also, research should be carried out to appraise the variations in soil moisture water balance because of a decline in ETref for evaluating the effects of decreasing ETref on irrigation practices, hydrological system, and drought adaptation in Bangladesh, especially in the northwest region.