Development of a new crop water stress index for agricultural crops using the red, near-infared and thermal infrared remote sensing imagery

The application of remote sensing technology in evaluating the crop water stress index is hampered by soil background effects related to partial ground cover. This study investigates the explicit contribution of soil information as an added dimension to the Vegetation Index Temperature/Trapezoid method, as well as data fusion technique to increase visual interpretation of water stress to formulate a new index of water stress. The study was conducted in Lubbock and Hale counties which includes an irrigation study at USDA-ARS in 2006-2008, Texas Tech irrigation studies at N Quaker and New Deal in 2007, and the AgriLife extension irrigation study in Halfway, TX in 2007-2008. Remote sensing missions were carried out using the Texas Tech Airborne Multispectral Remote Sensing System. Digital counts in Red and Near-Infrared images were converted to reflectance. Thermal Infrared images were converted into absolute temperatures using the known reflectance of calibration tarps and thermal sensors on the ground. Estimation of ground cover and the calculation of soil brightness were carried out using observed reflectance in the Red and NIR. The normalization of soil brightness into a Soil Brightness Index was carried out using the temporal reflectance values of soil brightness. Differences between surface temperature and air temperature were calculated. The fusion of the differences between the surface minus air temperatures, ground cover and soil brightness index into the Hue and Saturation & Intensity domain (respectively) improved the identification of water stress levels and the non-living features. Degree of water stress seemed to follow the ground cover across locations. The non-stressed cotton was observed consistently to be above 60 percent ground cover. Below this value, cotton begins to experience some degree of stress which can be associated with the acclimation of cotton to plant-available-water. Low ground cover however, can be non-stressed, with soil brightness index approaching 0. The 3-dimenstional CWSI is applicable to cotton under various irrigation types. The ground cover estimates and normalized soil brightness improved the discrimination of water stress levels using the 3-dimensional crop water stress index.