提出三温模型结合MODIS数据反演区域蒸散发的方法,在内蒙古草原开展案例研究,以2008年植被生长季(7—10月)的波文比系统观测数据为标准,对该方法进行检验.结果表明：三温模型反演的蒸散发量,平均值、最大、最小值分别为4.58 mm/d、9.03 mm/d、1.28 mm/d;蒸散发反演结果在空间上分布较均匀,与草原的均一性相吻合,在时间上蒸散发的数值先逐渐增大,8月后逐渐减小,与观测结果相一致;三温模型反演的蒸散发量与观测值之间的最小、最大绝对误差分别为0.11 mm/d、1.64 mm/d,平均绝对误差为0.58 mm/d、平均相对误差为17.10%.三温模型在1 km空间尺度的反演精度较理想.

Estimation of evapotranspiration (ET) is an obstacle in researches of both water and energy balance. In this paper, three-temperature (3T) model was revised for remote sensing application. A case study was performed in a semi-arid grassland from July to October in 2008 in Taibus Banner of Inner Mongolia, China. Eleven cloudless MODIS images at 1 km scale were used to estimate ET and data observed from two Bowen ratio systems in the corresponding period was adopted to validate the performance of the 3T model. Results showed that: (1) the modeled ET varied from 1.28 mm/d to 9.03 mm/d in the growing season, with a mean value of 4.58 mm/d; (2) the modeled daily ET distributed homogeneously in space, and temporally, ET value increased and then decreased gradually as the growing season changed from beginning to the end; (3) the mean absolute error between the observed and modeled ET was 0.58 mm/d, with maximum and minimum absolute errors of 1.64 mm/d and 0.11 mm/d, respectively, and the mean absolute percent error was 17.10%. All the performances indicated that the 3T model has a good potential in ET estimation.