Drivers of Landscape-scale Nutrient Export Detected from Satellite Remote Sensing and Imaging Spectroscopy

This dissertation characterizes factors influencing nutrient export from forested landscapes using satellite multispectral remote sensing and imaging spectroscopy. The chief objective is to identify drivers of water quality by investigating relationships among landscape-scale physiographic, climatologic and biophysical variables. Chapter 1 develops regional-scale indicators of water quality using hypertemporal remote sensing data (MODIS) for the state of Wisconsin. The approach generates an algorithm to predict water quality across years using MODIS-derived inputs, with the ability to predict future years as data become available. Chapter 2 utilizes MODIS data to predict nitrate-nitrogen loads from forested components of catchments in the Chesapeake Bay watershed on a near-continuous-time basis. This study makes novel use of functional concurrent linear models to integrate temporal variation in environmental factors into the analyses and provide insights into both the timing and magnitude of potential drivers of changes in water quality. Chapter 3 focuses on the development of generalizable algorithms to map canopy foliar traits across space and time using imaging spectroscopy data obtained from NASA's AVIRIS sensor. In this study, I develop calibrations for the determination of leaf chemical composition (nitrogen, carbon, and fiber constituents) and morphology (leaf mass per area) of temperate and boreal tree species using imaging spectroscopy. This chapter also demonstrates techniques to explicitly propagate uncertainties from the leaf to the image scale. Chapter 4 employs a structural equation modeling approach to assess the relative influences of foliar biochemistry, derived from imaging spectroscopy, watershed physiography and human land use patterns on water quality in watersheds across the Upper Midwestern United States. Overall, the approaches developed in this research may help spatially target development and restoration policies towards building more resilient landscapes, especially with respect to water quality. This study demonstrates that recent advances in satellite and airborne imaging technologies enhance our ability to understand landscape-level processes associated with water quality, and that we can develop more standardized and accessible data sets for use by a broader audience of managers and stakeholders.