Study On Inversion Of Soil Moisture In Arid Area Based On VIIRS-TVDI

Soil moisture plays a significant role in agricultural monitoring and yield estimation. Remote sensing monitoring over soil moisture features such merits as fast effectiveness, strong dynamic contrast and large monitoring areas. Meanwhile, TVDI integrates the reaction of both vegetation growth state and land surface temperature to soil moisture, thus satisfying the demand of this study. The VIIRS sensor carried by Suomi-NPP, as the successor of MODIS/AVHRR, has not been used in soil moisture monitoring in domestic yet. Since TVDI method was put forward, the vegetation index is used in vegetation characterization, but dying vegetation in vegetation index has no contribution to the latent heat of transformation, which can cause the error of the TVDI model. To solve above problems, the study carried out the following work and drew conclusions:1) VIIRS sensor data EDR daily surface reflectance product(GIGTO-VI1~5BO) is studied in this paper and the normalized differential vegetation index(NDVI) of the 125 th Regiment of the 7th Agricultural Shi in Xinjiang from June 29 to July 10, 2015 is calculated. In addition, temperature vegetation dryness index(TVDI) is optimized and dry and wet edges are extracted to form characteristic space based on the EDR daily land surface temperature product(GMTCO-VLSTO) so as to calculate TVDI result; The measured ground soil moisture content data in the same period are adopted for verification and this model is promoted to the economic belt regions in the whole north slope of Tianshan Mountain. And in the same way, the soil moisture is also inverted by the OLI/TIRS sensor data of 29 th June.The correlation index R~2 between TVDI result and measured ground data is 0.85; the uncorrelated two-tailed test value between them is smaller than 0.001. The soil moisture model built has sound reaction to water content in unsaturated farmland soil. The large breadth and fine-time resolution ratio of VIIRS data further improves timeliness of data processing in large area monitoring; only with 1~4 images can all areas in Xinjiang be completely covered for daily monitoring;2) The R~2 between OLI/TIRS sensor inversion results and the moisture content of data measured ground correlation was 0.85, which can meet the demand of modeling. The final result is in line with the field distribution of soil moisture content, and present the consistency with the VIIRS inversion results. In the application of small scale, OLI/TIRS's advantage of high resolution takes full advantage, which can remedy the VIIRS inversion results with the insufficient in respect of spatial resolution;3) Using the trichotomy pixel model, the study optimize the TVDI model by replace the characteristic NDVI with PV, as the NPV has no contribution to the latent heat transition. The accuracy of final result is higher than that of the original model, which means the new model can express the actual situation of soil moisture content better.