Urban trees are strategically important in a city's effort to mitigate their carbon footprint, heat island effects, air pollution, and stormwater runoff. Currently, the most common method for quantifying urban forest structure and ecosystem function is through field plot sampling. However, taking intensive structural measurements on private properties throughout a city is difficult, and the outputs from sample inventories are not spatially explicit. The overarching goal of this dissertation is to develop methods for mapping urban forest structure and function using fused hyperspectral imagery and waveform lidar data at the individual tree crown scale.

Urban forest ecosystem services estimated using the USDA Forest Service's i-Tree Eco (formerly UFORE) model are based largely on tree species and leaf area index (LAI). Accordingly, tree species were mapped in my Santa Barbara, California study area for 29 species comprising >80% of canopy. Crown-scale discriminant analysis methods were introduced for fusing Airborne Visible Infrared Imaging Spectrometry (AVIRIS) data with a suite of lidar structural metrics (e.g., tree height, crown porosity) to maximize classification accuracy in a complex environment. AVIRIS imagery was critical to achieving an overall species-level accuracy of 83.4% while lidar data was most useful for improving the discrimination of small and morphologically unique species. LAI was estimated at both the field-plot scale using laser penetration metrics and at the crown scale using allometry. Agreement of the former with photographic estimates of gap fraction and the latter with allometric estimates based on field measurements was examined. Results indicate that lidar may be used reasonably to measure LAI in an urban environment lacking in continuous canopy and characterized by high species diversity. Finally, urban ecosystem services such as carbon storage and building energy-use modification were analyzed through combination of aforementioned methods and the i-Tree Eco modeling framework. The remote sensing methods developed in this dissertation will allow researchers to more precisely constrain urban ecosystem spatial analyses and equip cities to better manage their urban forest resource.