Spatial and Temporal Distributions of Evapotranspiration in Gannan Grassland Based on MODIS Data
基于MODIS数据的甘南草原区域蒸散发量时空格局分析

In the context of increasing pressures on water resources and degrading grassland ecosystems in Gannan, this study was aimed at deriving ET across the grassland in Gannan using the Surface Energy Balance System (SEBS) model capable of being applied to large areas in conjunction with Moderate-resolution Imaging Spectroradiometer (MODIS) data produces. Specifically, the daily ET in July of the grassland was estimated, with analyzing the temporal and spatial distributions and changes in ET during the period 2000-2009. Meanwhile, the authors employed the fractal cross-hatching of image method to describe the remotely sensed ET and its distribution characteristics over the study area so as to gain a greater understanding of the differences in spatial distribution and changing trends in ET. The results of the fractal calculation indicated a fractal number of cross-hatching of 1.237 in 2000 and 1.337 in 2009. Marked variations in the profile curve in 2009 indicated a large gradient of ET, whereas a smaller gradient in spatial variation in ET occurred in 2000. Essential factors affecting ET distribution are the patch size of various landscapes, and the number and the combination of space. In particular, the uniformity degree of the distribution in landscapes plays a crucial role in the fractal number of ET. Results also show that as a whole, the average daily ET showed a decreasing trend, with a relatively larger change found in the Bailongjiang to Wubaixiang in Zhouqu County, where the magnitude of decrease was approximately 50%. There was, however, only 7 mm ET occurring over areas which had been sufficient in water resources and expected to show large ET rates, including imudoutang and Sangke grassland scenic pond, peaceful deer farm near the west bank of the Yellow River tributaries, the Gahai Nature Reserve, Branch was bitter River, Bailongjiang, the Dam River, and the Taohe River. This might be related to decreased water resources over these areas. It was also found that the discrepancy between the simulations and the observations were about 0.30 mm, with showing the smallest discrepancy of only 0.07 mm and the largest discrepancy of 0.48 mm. This might be indicative of the reliability of the simulation performed in this study.