TVDI模型的农业旱情时空变化遥感应用研究

Abstract

温度植被干旱指数（TVDI）是一种通过反演土壤湿度来反映农业干旱状况的重要方法，其中能量平衡和植被指数的变化是影响TVDI模型精度的主要因子。在研究比值植被指数（RVI）、归一化植被指数（NDVI）、增强型植被指数（EVI）和修正土壤调节植被指数（MSAVI）和下垫面温度（<em>T</em>_s）的基础上，引入DEM数据并对<em>T</em>_s做地形校正，减少了地形起伏对能量平衡的影响，建立不同植被指数的<em>T</em>_s-Ⅵ特征空间，选择与土壤湿度相关性最好的温度植被干旱指数（TVDI），获取研究区2005年作物生长季（5-9月）的干旱状况，并用同步的气象信息对干旱状况进行验证。结果表明：根据不同时期采用不同植被指数的TVDI模型，经过地形校正后能够更好地反映研究区的农业干旱状况。;Temperature Vegetation Dryness Index (TVDI) is an important tool that reflects agriculture dry situation by inverting soil moisture. The changes of energy balance and vegetation index are two main factors to influence the precision of the TVDI. The MODIS (Moderate….) data products, as RVI(Ratio Vegetation Index), NDVI(Normalized Difference Vegetation Index), EVI(Enhanced Vegetation Index), MSAVI(Modified Soil Adjusted Vegetation Index), and <em>T</em>_s (Land Surface Temperatures), are applied and the DEM (ASTER-GDEM) data are used to correct the <em>T</em>_s data for the reduction of the topographic influences by topographic relief. The TVDI is then employed by comparison of different vegetation index, where the TVDI is more sensitive to soil moisture. Thus the dry situation in the study area is analyzed during the plant growth time and compared by the synchronous meteorology data. The results indicate that: (1) terrain correction can effectively prevent the decrease of TVDI value from a lower surface temperature for a higher pixel. The correlation between <em>T</em>_s-NDVI index and measured values on May is compared, <em>R</em>< sup>2</sup> will increase from 0.4634 to 0.5859 by terrain correction. It shows that the terrain corrected TVDI can improve effectively the estimation of soil moisture. (2) By comparing the correlation between <em>T</em>_s-NDVI, <em>T</em>_s-EVI, <em>T</em>_s-RVI, <em>T</em>_s-MSAVI and soil moisture,all the TVDIs present the negative correlations with soil moisture. The best correlations between the soil moisture and TVDIs can be always found, such as <em>T</em>_s-MSAVI in June, July and September 2005, <em>T</em>_s-EVI in May, and <em>T</em>_s-NDVI in August. Thus a TVDI feature space for different periods by these vegetation indexes are built for inversion of drought conditions. By comparison with agricultural meteorology, the results are acceptable. (3) Large area of the study area was humid from May to September 2005, drought occurred in the West on August, and humid was located in East on June. Therefore, compared with the measured data, the terrain corrected TVDI model is robust to eliminate the terrain and land cover influences to land surface temperature for inversion of soil moisture in the study area. And it is faithful to predict the agricultural drought condition in the study area during 2005 crop growth season.