Evaluating Vegetation Response to Water Stress Using Close-Range and Satellite Remote Sensing

Drought is a weather related natural disaster that occurs in virtually all climatic zones of the world. In the last century, almost all parts of the contiguous United States have experienced several prolonged drought events with considerable impacts on the agricultural economy and environment. With changing climates, the droughts are expected to be more severe, longer, and widespread in many parts of the world including sections of the United States. Understanding the response of vegetation to water stress using remote sensing technologies will enhance our ability to detect and monitor drought. This research evaluates the response of vegetation to drought-related water stress at the leaf, canopy, and landscape scales using remotely sensed reflectance and/or thermal data. At the leaf level, a crop water stress index model was developed using high spatial resolution thermal imagery to estimate Relative Water Content (RWC) in soybean leaves. The model showed a higher accuracy in RWC determination (85%) compared to the raw temperature based RWC determination (69%). At the canopy level, multi-year close-range reflectance based vegetation indices (VIs) were correlated with soil moisture measured at four depths of maize and soybean root zone. Results indicated that maize VIs were significantly related to soil moisture at deeper depths and kept the soil moisture memory up to previous 45-days. Soybeans VIs were significantly related to soil moisture at shallower depths and kept a relatively shorter (5-days) memory of soil moisture compared to maize. At the landscape scale, Terra-MODIS Land Surface Temperature (LST) and Normalized Difference Vegetation Index (NDVI) products were used to detect drought-induced stress in vegetation including corn, soybeans, and three grassland cover types across the state of Nebraska. Results indicate that the majority of the land cover pixels experienced significantly higher daytime and nighttime LSTs and lower NDVI during the drought-year growing season compared to the non-drought year. The findings of this dissertation research will contribute toward the development of more robust tools for monitoring drought stress in vegetation. Advisors: Sunil Narumalani and Brian D. Wardlo