Surface Soil Moisture Model with NDVI (Case Study: Rangeland of Khorasan Razavi Province

Soil moisture is one of the critical parameters in climatic, ecological and hydrological modeling and vegetation growth. Soil moisture is important for environmental studies but due to ground-based point measurement of soil moisture is expensive and impractical for watershed measurement, this parameter is not be used widely to simulate and predict models. Overall and uniform view on different part of land, repetitive photography from wide and inaccessible areas along with data with regular periods and decrement effects of plans on environment are important criteria's of remote sensing. Rangeland growth is heavily functioning of the water availability therefore the vegetation index from satellites may respond to the change of soil moisture. We used MODIS sensor images with the consideration of economical aspects, availability of images and high radiometric, temporal and spatial resolution to estimate the surface soil moisture. MODIS images were provided by Iranian Space Agency for the day's availability of soil moisture field data matched by satellite images. Radiometric and geometric correction was conducted by orbital parameters on all images and the images were converted to BIL format and shifted in relation to accurate boundary of Caspian sea. Using bathymetry map of Caspian Sea, atmospheric correction was conducted by dark pixel method. This case study analyzes the correlation the surface soil moisture measured gravimetric sampling and MODIS NDVI (250m by 250m), compares the collected coincided and lagged soil moisture and MODIS NDVI. Also estimates the soil moisture using NDVI by linear regression models during growing season (April-Aug) from 2003 through 2005 in arid and semi arid rangeland in Khorasan Razavi Province. NDVI shows lag to soil moisture because of vegetations response lag it. Results show that surface soil moisture in arid and semi arid rangelands has moderate correlation with simultaneous and lagged NDVI and can be estimated using NDVI during growing season. Stronger relations can be obtained with surface soil moisture data that are lagged by 2 weeks with respect to the vegetation index.