Estimation of Surface Soil Moisture Using Vegetation and Thermal Indices from MODIS Images (Case study: Gonbad-e Qavus)

Estimating soil moisture along a vast area is mostly expensive and sometimes undoable. One of the newly available technique is remote sensing, which allows the possible surface soil moisture estimates in a large scale. Therefore, the objective of this research was to investigate the possibility of estimating surface soil moisture using MODIS satellite images in compare to soil moisture measured data. To conduct this research, vegetation and temperature indices from MODIS images were extracted from October 2013 to June 2014 period for Gonbad-e Qavus region, then measurements were calibrated with field measurement data. The results indicated that significant but weak correlations are available between NDVI and NMDI indices with measured surface soil moistures (R= 0.62), while there were highly significant, but negative correlations (p ≤ 0.01; R = -0.89) for LST index and thermal bands of 31 and 32 with surface soil moistures. The TVDI index also had highly significant, but negative correlation (p ≤ 0.01; R = -0.90) with surface soil moisture. There were also highly significant linear and exponential correlations between SWI index and surface soil moisture. In conclusion, the results indicate promising relations for thermal and combined models to estimate surface soil moisture relative to vegetation models.