Estimation of regional evapotranspiration over the Hufu Plain using STME and MODIS data

The Hufu Plain is a vital grain production base, with sufficient solar radiation, energy, fertile soils as well as intensive agricultural management. In the region, irrigation has ensured stable and high crop yields over the years. The continuous extraction of groundwater and surface water has induced severe water shortage in the Hufu Plain region. Therefore reasonable and efficient use of the limited water resources was necessary for sustainable agricultural production. Here in this study, we developed a remote sensing evapotranspiration (ET) model, STME (a single-source trapezoid model for evapotranspiration), and took 115 MODIS (moderate resolution imaging spectrometer) images (local surface temperature and reflectance) for October 1, 2011 to September 30, 2012 to estimate regional surface water deficit index (WDI) and evapotranspiration in the Hufu Plain. Two typical farmland ecosystems were selected as the investigated objectives, one was cropland ecosystem in Luancheng, the other one was orchard ecosystem in Zhaoxian. We compared the estimated ET by STME with observations by the eddy covariance system. Results suggested that the STME model well estimated daily ET. The average observed net radiation (Rn) was 4.10 mm and the average estimated Rn was 4.69 mm, with a root mean square difference (RMSD) of 0.80 mm for Zhaoxian orchard ecosystem. The average observed daily ET was 2.86 mm and the average estimated daily ET was 3.01 mm, with RMSD = 0.95 mm for Zhaoxian orchard ecosystem. Also the average observed daily ET was 2.67 mm and the average estimated daily ET was 2.44 mm, with RMSD = 0.87 mm for Luancheng cropland ecosystem. To interpret the temporal and spatial variations in regional ET, we used the STME model to estimate regional ET in the Hufu Plain study area. Furthermore, the water deficit index (WDI) reflected the severity of drought in the region. ET for orchard ecosystems in October was more than that for cropland ecosystems. The regional ET was less than 1 mm in November. In April, ET for cropland ecosystems was more than that for orchard ecosystems. Then in May, ET for cropland and orchard ecosystems was somehow similar as there was vigorous vegetation growth. In June, ET for cropland ecosystems was less than that for orchard ecosystems because wheat of cropland was harvested and maize just planted. In July, ET for the whole region was the maximum. This suggested that ET was not only related to vegetation growth, but also associated with soil moisture. In the months of August and September, ET decreased as crops matured and were harvested. WDI varied with time and place and it was practicable in guiding irrigation. STME, a mathematically-based model for calculating the vertex of trapezoid framework and water deficit index, simplified and clarified ET estimation process.