基于温度植被干旱指数的土库曼斯坦典型绿洲干旱遥感监测

Remote sensing monitoring of drought in Turkmenistan oasis based on temperature /vegetation drought index

针对目前中亚地区广泛存在的农业干旱问题,选取土库曼斯坦典型绿洲为研究区,利用2幅MODIS影像和同期过境的Landsat影像,获取归一化植被指数(NDVI)和地表温度(Ts),构建NDVI-Ts特征空间,计算得到2001和2011年2种温度植被干旱指数TVDITM和TVDIMODIS。利用较高分辨率的LandsatTM数据对MODIS数据反演结果进行验证。结果表明:2001年、2011年TVDITM和TVDIMODIS的绝对误差分别为0.0178和0.0228;均方根误差分别为0.0226和0.0279;相关系数分别为0.949和0.922,说明TVDI可有效反映区域土壤干旱情况。研究区内干旱区域占总面积的60%以上,旱情由绿洲中心向外围逐步严重。近10年来,研究区内土壤湿度低值区呈现扩大趋势,湿润和正常所占比例下降了5.32%,干旱和重旱的比例上升了14%,旱情情况总体加剧。因此,利用MODIS影像的温度植被旱情指数可有效地监测中亚地区干旱情况。

Drought stress is an important issue in the agricultural production in central Asia. Selecting the typical MurghabTejen oasis in Turkmenistan as the study area, and by using two MODIS images and the two TM images in the same time, the land surface temperature （LST）/normalized difference vegetation index （NDVI） characteristic space was established, and two kinds of temperaturevegetation drought index （TVDI）, i.e., TVDI-TM and TVDI-MODIS, were calculated from the LST/NDVI of different sensor images in 2001 and 2011. The Landsat TM data with higher resolution were used to verify the results of the MODIS inversion data. The absolute error of the TVDITM and TVDI-MODIS was 0.0178 and 0.0228, the RMSE was 0.0226 and 0.0279, and the correlation coefficient was 0.949 and 0.922, respectively, indicating that TVDI could effectively reflect regional soil drought conditions. In the oasis, arid region occupied more than 60% of the total, and drought was becoming more serious from the center to the outer oasis. In the past ten years, the area of low soil moisture in the oasis presented an expanding trend, the mist and normal moisture area decreased by 5.32%, and the drought and severe drought area increased by 14%. Overall, the drought situation was aggravated. Therefore, using the TVDI of MODIS image could effectively monitor the droughts in central Asia.