• •

### 基于MODIS数据的博斯腾湖流域地表蒸散时空变化

1. 1. 南京信息工程大学遥感学院, 江苏 南京 210044
2. 新疆巴音郭楞蒙古自治州博斯腾湖研究所, 新疆 库尔勒 841000
• 收稿日期:2011-12-25 修回日期:2012-02-13 出版日期:2012-11-20 发布日期:2012-11-20
• 作者简介:

作者简介：徐永明（1980- ）,男,江苏连云港人,讲师,主要从事资源环境遥感研究。E-mail：xym30@263.net

• 基金资助:
国家环保行业科研专项(200909048)、国家重大专项水体污染控制与治理项目(2009ZX07106-004)资助

### Spatio-temporal Variations of Land Surface Evapotranspiration of Bosten Lake Basin Based on MODIS Data

Yong-ming XU1(), Qiao-hua ZHAO1, Ya-er BA2, Shu-ying BAI1, De-yong SUN1

1. 1. School of Remote Sensing, Nanjing University of Information Science and Technology, Nanjing,Jiangsu 210044, China
2. Bosten Lake Science Research Institute of Bayangol Mongol Autonomous Prefecture, Kuerle, Xinjiang 841000, China)
• Received:2011-12-25 Revised:2012-02-13 Online:2012-11-20 Published:2012-11-20

Abstract:

Land surface evapotranspiration is an important process in the energy balance and water cycle between the atmosphere and the land surface. Detailed understanding of the spatial and temporal variations of land surface evapotranspiration is critical for hydrologic cycle study and water resource management, especially in arid and semi-arid area. Satellite remote sensing provides a straightforward and consistent way to observe evapotranspiration over large scales with more spatially detailed information than traditional in-situ observation. In this paper, remote sensing data from the Moderate Resolution Imaging Spectroradiometer (MODIS) on the EOS/AQUA satellite and ground meteorological data from July 2002 to September 2009 were employed to estimate the land surface evapotranspiration in Bosten Lake Basin, Xinjiang Uygur Autonomous Region. Four remotely-sensed variables (land cover, land surface temperature, normalized difference vegetation index and surface albedo) and three ground measured variables (air temperature, precipitation and wind velocity) were used. First the net radiation, ground heat flux and sensible heat flux were estimated and therefore the latent heat flux could be calculated based on the surface energy balance approach. Then the accumulated daily evapotranspiration was derived by up-scaling the instantaneous latent heat flux under the assumption of constant daytime evaporative fraction. The validation of the estimated evapotranspiration gives a satisfactory accuracy with the mean absolute error of 12.39 mm and the mean relative error of 14.15%. The annual mean land surface evapotranspiration of Bosten Lake Basin exhibits a distinguishable spatial pattern, high in the northwest (higher than 700 mm) and low in the southeast (lower than 150 mm). The Bosten Lake and the small lake wetland exhibit the highest evapotranspiration while the desert around Bosten Lake exhibits obviously lower evapotranspiration than other areas, indicating that the spatial distribution of evaporatranspiration is highly influenced by land cover. The seasonal variation of mean evapotranspiration shows a unimodal pattern, with the evapotranspiration in summer accounts for 48.10% of the whole year. Evapotranspiration is significantly positively correlated with precipitation and air temperature, suggesting that the evapotranspiration in Bosten Lake Basin increases with precipitation and air temperature. Furthermore, the two climate factors show different contributions to evapotranspiration in different seasons. In spring, the precipitation has more impact on evapotranspiration than air temperature; in summer, both precipitation and air temperature have obvious high correlation coefficients with evapotranspiration; in autumn and winter, air temperature has greater effect on evapotranspiration than air temperature. This study shows that satellite remote sensing is an effective approach for estimating spatial land surface evapotranspiration at regional scale with the aid of meteorological data. The spatial and temporal variations of evapotranspiration provides important reference for the water resource management and ecological environment research of the Bosten Lake Basin.

• TP79