Study On Soil Moisture Inversion Method Based On Remote Sensing

Soil moistureis an important indicator for agriculture drought and flood disaster. In aridand semi-arid area, drought occurs frequently, which results in serious losses to agriculturalproduction. Soil moisture retrieval based on remote sensing still faces a lot of difficultiesbecause of the influences of many factors such as geography, climate change and vegetationcoverage. Regarding to the influences of the vegetation coverage on soil moisture retrieving,with the MODIS remote sensing data used and pretreated by ENVI software, soil moistureinversion methods based on apparent thermal inertia, VSWI and TVDI are studied in this thesis.Soil moisture retrieving is carried out in different vegetation coverage situations inGuanzhong area in spring. The main research contents and conclusions are as follows:(1)With MODIS composite products data adopted instead of daily remote senseing data,apparent thermal inertia method is studied for soil moisture retrieving in low vegetationcoverage area. Soil moisture retrieving is then carried out according to the built apparentthermal inertia model. The results show that apparent thermal inertia has significant linearcorrelation with measured soil moisture and the RMSE and MRE of the apparent thermalinertia model built is5.7%and6.19%respectively, thus the apparent thermal inertia model isfeasible in low vegetation coverage situation.(2)In high vegetation coverage area, the VSWI method is used for soil moistureretrieval using daily MODIS L1B data. Three vegetation indexs of NDVI, MSAVI and EVI arecompared in this method for their applicability. MVSWI is put forward according to theinfluence of NDVI difference to VSWI, and it is compared with traditional VSWI. The resultsshow that the RMSE and MRE of the MVSWI model are4.79%and6.08%respectively,compared with7.03%and10.38%of the traditional VSWI model, which shows that theretrieval effects of MVSWI is better than the traditional VSWI.(3)The applicability of the TVDI method for soil moisture retrieval is researched withthe same data with VSWI. The results show that TVDI has no significant correlation withmeasured soil moisture data, and it is not suitable for soil moisture retrieval in April in studyarea.