Remote Sensing of Aboveground Vegetation Structure, Biomass, and Water Content Across Spatial and Temporal Scales

Abstract

Vegetation plays a critical role in the interaction of terrestrial carbon, water, energy, and nutrient cycles at the Earth's surface, influencing global biospheric-atmospheric exchanges of carbon and water and regulating the climate system. However, natural and human-induced disturbances are increasingly affecting ecosystems, leading to reduced carbon storage by vegetation. Satellite remote sensing is used for spatiotemporal estimation of key vegetation variables such as structure, aboveground biomass (AGB), and vegetation water content (VWC), but coarse pixel resolution poses challenges for accurate calibration and validation of these estimates. In this dissertation, I explored novel remote sensing technologies and techniques to improve structure, AGB, and VWC estimation across multiple spatial and temporal scales. I incorporated environmental disturbance factors to benchmark the temporal sensitivity of these estimates to large-scale biomass change and to evaluate their accuracy in quantifying disturbance-driven biomass loss. Appendix A focused on enriching annual AGB estimates in North American arctic-boreal ecosystems using integrated of microwave and optical-multispectral satellite observations. This approach enhanced spatial AGB across the region and improved detection of biomass loss driven by large-scale wildfires. Appendix B explored the application of close-range photogrammetry and derived ultra-high spatial resolution 3D models for extracting structural plant information, improving biomass quantification, and assessing the impacts of physical disturbance for three morphologically distinct dryland shrub species. We found that the model integrating canopy area and mean shrub height yielded the most accurate species-agnostic AGB estimate, and adequately captured biomass loss driven by physical disturbance. Appendix C evaluated tower-level microwave reflectance and its relationship with eddy covariance flux measurements, vegetation greenness, soil moisture, and satellite microwave observations in a semi-arid grassland ecosystem during the summer 2021 growing season. While correlations were generally strong during the greening phase of the growing season, they were considerably weaker during the browning phase. We concluded this was driven by the temporal lag of tower-level microwave reflectance and attributed this to its high sensitivity to late-season residual vegetation moisture. Collectively, these findings advance understanding of remote sensing applications and techniques to improve capabilities for spatial estimation and temporal monitoring of vegetation properties which are central to global carbon and water cycles.