Variation of evapotranspiration and its response to vegetation productivity in the North China Plain

The North China Plain is a main grain production area in China, where the shortage of water resources is the main factor restricting regional grain production and socioeconomic development. Clarifying the temporal and spatial variation of evapotranspiration (ET) and analyzing the main driving factors are critical for exploring regional water resource evolution and optimizing water resource management. Based on the PML_V2 (Penman-Monteith-Leuning Evapotranspiration V2) remote sensing ET product released in 2019 with a spatial resolution of 500 m and temporal resolution of 8-day, Theil-Sen Median slope estimation and Mann-Kendall trend analysis were used to evaluate the changing trend of ET; and the correlation coefficient method was used to analyze the relationship between ET and vegetation productivity. To evaluate the ET variation at the pixel scale, the significance of variation, and driving factors in four agricultural areas representing three agricultural area types were selected: Shijiazhuang and Baoding, Hengshui, and Dezhou, which represented the piedmont plain of Taihang Mountains, central low plain, and Yellow River irrigation area, respectively. The results showed that the annual average ET was 588.1 mm from 2001 to 2019 in the whole North China Plain, the interannual variability was characterized by a low-high-low dynamic trend, and the maximum (665.4 mm) and minimum (542.2 mm) ET occurred in 2015 and 2001, respectively. The ET trends during different crop growth seasons were significantly different. During the wheat growth season, the overall ET trend was declining, possibly resulting from the policies, such as wheat conversion to fallow, and limitation of groundwater pumping, which are being implemented to alleviate the groundwater funnel in North China. The overall ET trend was significantly upward in the corn growth season. Additionally, there were significant differences among the annual average ET for different land use types. The ET in 85.5% of the agricultural land areas showed an upward trend, of which 42.3% increased significantly and was mainly distributed in the Yellow River irrigation area. For the annual average ET in urban land, the areas with decreasing and increasing trends were 50.9% and 49.1%, respectively. Urbanization resulted in a significant decline in ET in the expanding areas of large cities, whereas an increasing trend was observed in the downtown regions of large cities, such as Beijing and Tianjin. Correlation analysis showed that areas with a positive correlation between ET and NDVI (normalized difference vegetation index) accounted for 76.54% of the North China Plain, and areas with a positive correlation between ET and GPP (gross primary production) accounted for 87.6% of the entire region. The stronger correlation between ET and GPP indicated the influence of higher crop productivity on ET in major grain-producing areas, which was also proven by the correlation between ET and vegetation productivity in the four typical agricultural areas. There were significant correlations between ET and GPP/NDVI in the Yellow River irrigation area represented by Dezhou. The only significant correlation between ET and GPP was observed for the central low plain, represented by Hengshui. Non-significant correlations between ET and GPP/NDVI were seen in the piedmont plain represented by Shijiazhuang and Baoding, possibly resulting from multiple ET driving factors, including vegetation productivity.