Estimation of soil moisture status in the Texas High Plains using remote sensing

Knowledge of both soil moisture content and latent heat are of great importance to many environmental applications, including the monitoring of plant water requirements, plant growth and productivity, as well as for irrigation and cultivation management procedures. Soil moisture is however difficult to measure at the spatial scale of a catchment. The conventional point measurement methods such as neutron probe or gravimetric method are not appropriate for understanding of the spatial and temporal behavior of soil moisture. By taking the advantage of the strengths of multi-spectral satellite imagery, new approaches have suggested for estimating soil moisture content and latent heat to achieve higher accuracy and spatial resolution. Considering that the most current approaches for soil moisture estimation are based on converting surface thermal emittance to surface temperature (T_s), a modified trapezoid method, Thermal Ground cover Moisture Index (TGMI), using combination of raw data from multiple red, near infrared (NIR), and thermal infrared (TIR) channels is suggested. This approach is based on correlating ground cover (GC) and TIR data from satellite imagery to ground measured soil moisture. Also, another index, the Perpendicular Soil Moisture Index (PSMI), is proposed for monitoring soil moisture status from space, based on the sensitivity analyses of the impact of soil moisture on the surface reflectance and emittance in TIR-GC space. Soil moisture maps were produced for each image acquisition date for the study area using these two indices. Statistical analysis of estimated and field-measured volumetric water content from a large number of fields indicates that the procedure for estimating soil moisture from remote sensing imagery using TGMI is accurate enough that, on average, estimates of soil moisture determined using this index should be within 4 percent of their true values. Furthermore, PSMI is highly correlated to measure volumetric water content with correlation coefficient equal to 0.7. In addition, a new approach that estimates a stress factor, F_s, which reduces latent heat as the stomata closes in response to reduced soil moisture using TIR-GC space has been proposed. Considering the effect of F_s in spectral crop coefficient approach (K_sc), crop water use (CWU) can be estimated more accurately for monitoring actual evapotranspiration. Comparing hourly values of CWU estimated using the combination of F_s and K_sc method with actual measurements of evapotranspiration made using the eddy covariance method showed that the proposed method was accurate with correlation coefficient equal to 0.84.

This dissertation won 1st Place in the Texas Tech University Outstanding Thesis and Dissertation Award, Biological Life Sciences, 2015.