Inversion Study Of Wheat Drought Based On HJ Satellite Data

Drought is one of the most serious natural disasters on agricultural production. Remote sensing technology has achieved many results in the field of crop drought monitoring which has the advantage of fast, large area and non-destructive in monitoring. This paper extracted temperature vegetation drought index (Temperature Vegetation Dryness Index, TVDI) for winter wheat drought monitoring by using HJ satellite data, took surface temperature retrieval and atmospheric correction of remote sensing data as case study which is important for the drought monitoring process. The main work and results are as follows:(1) Typical spectral characteristics before and after correction of the residents, vegetation and water showed that atmospheric correction can effectively eliminate the "increase" effect in visible band caused by Rayleigh scattering and the water vapor absorption in near infrared band. Atmospheric correction result and MODIS surface reflectance product with high accuracy had high homogeneity in the reflectivity of vegetation, water and residents, among them the average error of vegetation was12.8%.Comparison of the characteristics of NDVI changes before and after correction, it could be found that atmospheric correction had corrected NDVI value of mixed pixel, made it more reasonable.(2) TM and HJ-1B/IRS inversion results had high homogeneity in the spatial distribution trend and statistical results. Relevance of NDVI and RSR showed that vegetation index increases with the decrease of surface temperature, the drop rate of temperature decreases with the vegetation coverage increases. It means that when the vegetation coverage is high, canopy structure has high homogeneity, which resulted in small space difference of canopy temperature, then reduces the spatial resolution requirements of RS images.(3) Due to the low indicative of NDVI to vegetation information when the vegetation coverage is low, the maximum surface temperature does not decreases with the increase of NDVI in all the interval as the general theoretical model. Therefore, only considered the interval of NDVT>0.2when fitted the dry and wet edge, it will lead to the lower fitting precision of no-wheat class, but improve the fitting precision of winter wheat which is the research object in this paper. The TVDI index and the measured soil moisture of different depth had different degrees in correlation, especially the coefficient in the depth of10-20cm reaches about-0.65～-0.8level, which showed that TVDI has high indicative significance for soil moisture particularly the shallow soil, and could be used for drought inversion. According to the linear relationship between TVDI and the measured soil relative humidity, we obtained the distribution map of soil relative humidity in the depth of10-20cm, and divided the drought grade, the rainfall data proved that the inversion results are reasonable.