Analysis of Trend of Evaporation Changes and Determining the Role of Factors Affecting it Using Quantile Regression and Bayesian Quantile regression(Case Study: Hashem-Abad Station, Gorgan)

Climate change and global warming are some of the issues and concerns of human beings today that have important effects on rain, evaporation, runoff, and finally water supply and causes the severity and weakness of these parameters, increasing the occurrence of severe weather events and lack of available water, which causes irreparable damage. Evapotranspiration is an important component in the hydrological cycle that is affected by various factors such as air temperature, relative humidity, wind speed and number of sunshine hours, and these factors are also affected by climate change. Due to the occurrence of climate change in the country in recent decades, it is important to study the changes in the trend of climatic parameters and their role in evapotranspiration in order to apply management methods to reduce evaporation in water resources. The Mann-Kendall trend test is an often-used method to examine changes in the data time series, but this test only expresses changes in the center of the data series, so Quantile and Bayesian multiple regression methods are used to study the trend of changes in different parts of the data time series and also to investigate the role of various parameters on a specific parameter. Therefore, the purpose of the present research is to study and compare the trend of changes in evaporation and climatic parameters affecting it and determining the role of these factors on evaporation using Quantile and Bayesian multiple regression methods at Hashemabad Gorgan station located in Golestan province.

In the first step, the meteorological data time series of evaporation, and the factors affecting it including average temperature, relative humidity, sunshine hour and wind speed were prepared for Hashemabad Gorgan synoptic station with a statistical period of more than 30 years (1984-2018) and the seasonal data series of these data were formed. The non-parametric Mann-Kendall test was performed to investigate the trend of changes in evapotranspiration and the factors affecting it and then Quantile and Bayesian regression was used to investigate the changes in various quantiles of data series and also to determine the role of changes in different values of each of these meteorological factors on evaporation.

investigation of the trend in the evaporation data series and the factors affecting it based on the Man-Kendall test shows that the aforementioned trend test shows a decreasing trend in evaporation only for winter, while the lowest changes in trend in different quantiles are in this season, and in Summer the increasing trend in evaporation has the highest slope in the range of 0.66-1.14% for lower and upper quantiles, respectively. In spring, the increasing slope of 0.75 in the upper quantiles gradually changes to a decreasing slope of -0.72. In autumn, the lower quantiles have a higher slope up to -0.66 and in winter, the quantiles have the lowest slope between -0.14 to -0.08. The results of investigating the factors affecting evaporation also show that changes in temperature have a greater share in the changes of daily evaporation and the bigger the quantile, the increasing slope increases. The relationship between relative humidity and evaporation follows a decreasing slope and ranks second after temperature. Finally, it can be concluded that in Hashemabad station of Gorgan, the Mann-Kendall trend test could not detect significant changes that have occurred in the process of the evaporation trend, but these changes have been revealed the best by the Quantile regression method. In summer, which is the driest and hottest season of the year, the trend of increase in evaporation is more intense, and due to summer cultivation in this region, water consumption in agriculture has increased in recent years and with this trend will increase further in the future.

Given the fact that climate change may not have occurred only in the average value of a data series, but extreme events might have occurred in some parts of the series so it is necessary to study different parts of the data series using methods such as Quantile and Bayesian multiple regression and the results of this research emphasizes this matter. Also, due to the necessity of studying evapotranspiration in the application of management methods for water resources, useful and practical results can be brought out by studying the changes in different ranges of evapotranspiration, especially it’s extreme and high values as well as the effects of changes in different ranges of climatic parameters on evaporation.