The Impacts of Soil Moisture on Grassland and Oak Savanna

Although soil moisture is a key element in land surface processes and terrestrial ecosystems, estimating soil moisture at large scale is limited by the need for multiple model inputs and lack of appropriate data sources. To explore the potential of estimating soil moisture at large scale using optical satellite sensors, this dissertation investigated the responses of Californian grassland and oak savanna to the change of soil moisture, in terms of carbon assimilation, remotely sensed measures of chlorophyll content and canopy water content, at different spatial, temporal, and spectral scales. I evaluated the impacts of soil moisture on gross primary production (GPP) at different stages of growth of grassland and oak savanna. I found GPP in the late growing season was controlled by declining soil moisture, but there was a 10-20 day lag in the response of GPP to soil moisture. To assess the impacts of soil moisture on chlorophyll content and canopy water content and to explore the potential of estimating soil moisture at large scale using optical remote sensing data, the relationships between soil moisture and greenness based vegetation indices (VIs) and canopy water based VIs derived from AVIRIS and MODIS data were analyzed. Greenness based VIs had stronger relationship with soil moisture in the grassland and oak savanna than canopy water based VIs, particularly during the dry-down season. However, relationships between soil moisture and VIs were modified by vegetation types, soil properties, and viewing geometry. The differences in sensors were less evident for the grassland than on oak savanna. Furthermore, since actual evapotranspiration (AET) plays a critical role in ecosystem water balance and ecosystem responses to water stress, the correlation between AET and signals from webcam imagery was analyzed to complement with the studies of the relationships between soil moisture and VIs from satellite imagery. Significant relationships between AET and relative green brightness and VIs derived from webcam images were observed in Californian annual grassland. The results obtained in this study provide evidence to estimate soil moisture at large scale in a water-limited environment using optical remote sensors.