Spatial-temporal variation characteristics and attribution analysis of potential evapotranspiration in Beijing-Tianjin-Hebei region

This paper aims to investigate the spatial and temporal variation characteristics of the potential evapotranspiration (ET0) and its response to climate change using the collected data from the daily weather observation of 23 meteorological stations in the Beijing-Tianjin-Hebei region for 57 years from 1960-2016. The specific parameters of dataset included the daily average air temperature, the maximum and minimum air temperature, relative humidity, wind speed and sunshine hours. We chose the Penman-Monteith formula to calculate the daily ET0 at each site, and then analyzed the spatial and temporal variation characteristics of the calculated ET0. A sensitivity coefficient was used to quantitatively study the sensitivity of ET0 to various meteorological elements and its spatial and temporal distribution. The contribution rate was set to evaluate the influence of meteorological elements changes on the variation of ET0 during the calculation. The results show that: 1) The spatial distribution of ET0 shows a decreasing trend from south to north across the Beijing-Tianjin-Hebei region, except a high ET0 in the central region, such as Tanggu, Huangye and Baoding stations. There was a linear decline rate of -0.92mm /a. e in the whole distribution of ET0. The spatial distribution of ET0 MK (Penman-Monteith) decreased from the northwest to southeast. In temporal characteristics, spring has the most obvious decrease, followed by autumn, summer, and winter. 2) The sensitivity coefficient of ET0 to the average air temperature, sunshine hours, relative humidity and wind speed are 0.26, 0.28, 0.31 and -044, respectively, indicating the most sensitive to the changes of relative humidity in the Beijing-Tianjin-Hebei region, followed by the wind speed and sunshine hours. The sensitivity coefficient of ET0 to meteorological elements presents in a fluctuating trend from1960-2016, the sensitivity coefficient of ET0 to the wind speed and relative humidity in a downward trend, while the sensitivity coefficient of ET0 to average temperature and sunshine hours in an upward trend. The spatial distribution of the sensitivity coefficient from northwest to southeast: the wind speed and average temperature gradually increased, while the sunshine hours and relative humidity gradually decreased. The spatial distribution characteristics of the sensitivity coefficient of ET0 to relative humidity and average temperature were the opposite situation, where the high sensitivity zone of average temperature belongs to the low sensitive zone of relative humidity. The opposite phenomenon also occurred in the spatial distribution characteristics of the sensitivity coefficient of ET0 to sunshine hours and wind speed sensitivity, i.e., the high sensitivity zone of sunshine hours belongs to the low sensitive zone of wind speed. 3) The contribution rate of meteorological elements to ET0 depends on the sensitivity coefficient and the relative change of meteorological elements. ET0 has the most sensitivity to the changes of relative humidity, but the change rate of the relative humidity was small. ET0 has the lower sensitive to the changes of wind speed, but the change rate of the wind speed was large, resulting the most contribution to ET0 in the whole region was the wind speed. However, the stations number of the wind speed as the dominant factor decreased with the measure time, whereas the stations number of the average temperature as the dominant factor increased with the measure time, especially in the period of 2010—2014 to 2010s, 12 stations with the average temperature as the dominant factor. The stations number of the sunshine hours as the dominant factor also increased with the measured time. It infers that the average temperature and sunshine hours have much significant influences on ET0 in recent years. From 2013 to 2016, the correlation coefficients of the winter average air temperature, sunshine hours, relative humidity, wind speed and PM2.5 concentration were -0.44, -0.53, 0.42 and -0.35, respectively, while the correlation coefficients of ET0 and PM2.5 concentration were -0.37. Therefore, it is possible that the haze, especially winter haze, has a certain impact on meteorological elements, and thereby on ET0 in the Beijing-Tianjin-Hebei region.