那曲东部土壤水分MODIS遥感反演研究

Study on Soil Moisture by Remote Sensing Retrieval based on MODIS Data in Eastern Naqu

利用那曲东部2014年8月至2015年7月土壤水分观测资料与同期MODIS数据建立了研究区土壤水分遥感监测模型。表观热惯量法(ATI)反演土壤水分结果不理想,基于MODIS 8天合成数据以及晴空数据拟合结果的决定系数分别为0.4503和0.3753,晴空条件下ATI方法监测效果较差。基于单窗方法建立的四种模型中,三次多项式模型拟合效果较好,决定系数为0.5475,分析认为排除冬季数据后建模效果更好。结论:基于单窗方法的三次多项式模型能较好的反演研究区土壤水分,不足之处为对天气要求较高,若无晴空遥感数据,将影响土壤水分监测工作的开展。

The paper developed the remote sensing monitoring models for soil moisture based on observation data and MODIS data from August 2014 to July 2015 in the East of Naqu. The ATI models of soil moisture were not satisfactory for the R2 values of 0.4503 and 0.3753, respectively, based on 8-day and clear sky MODIS data. And the monitoring effect was even worse in clear atmosphere. The cubic polynomial model with R2 value of 0.5475 gave the best result in four models based on the reflectance of MODIS at the seventh band. And the analysis indicated that the model worked better excluding winter data. It concluded that the retrieved results of the cubic polynomial model based on Mono-Window Algorithm was in good agreement with the observation data in the study area, and its disadvantage was that soil moisture monitoring cannot be carried out without remote sensing data of clear sky.