Fusion of imaging spectroscopy and LiDAR for spatially explicit urban forest inventory

EPA Grant Number: F13F31235
Title: Fusion of imaging spectroscopy and LiDAR for spatially explicit urban forest inventory
Investigators: Alonzo, Michael Gerard
Institution: University of California - Santa Barbara
EPA Project Officer: Lee, Sonja
Project Period: September 28, 2014 through September 28, 2016
Project Amount: $84,000
RFA: STAR Graduate Fellowships (2013) RFA Text |  Recipients Lists
Research Category: Academic Fellowships , Fellowship - Geography

Objective:

Urban trees help reduce air pollution, lower temperatures and control stormwater runoff in cities. This study’s goal is to map the species and structure of the urban forest in order to understand the magnitude and spatial distribution of these ecosystem services. This research leverages high-resolution, remotely sensed imagery to delineate urban tree crowns, identify each crown’s species and measure the crown’s 3-D structure, including volume, leaf area and biomass. The primary deliverable is a novel analytical toolkit for enhanced mapping of urban forest structure and function.

Approach:

Three projects sequentially lead to fine-resolution maps of urban forest structure and function. First, the species of all canopy-dominant trees in an urban study are classified, using fused spectral and structural information extracted from hyperspectral imagery and LiDAR data, respectively. Second, key aspects of each tree’s structure, including leaf area index, are estimated from the 3-D LiDAR point cloud. Finally, the maps of species and leaf area are used, along with existing models connecting structure to function, to generate spatially explicit estimates of air pollution mitigation, urban cooling and stormwater runoff reduction.

Expected Results:

This study focuses on improving the methods employed to measure and analyze the urban ecosystem. Early results suggest that classification of tree species in a biodiverse setting is feasible with high-quality remotesensing products. Estimation of such structural parameters as leaf area index also has been shown to produce acceptable values compared to field measurements. It is possible that citywide maps of ecosystem function generated with remote sensing methods will be more accurate than existing field methods due to the elimination of sampling error.

Potential to Further Environmental/Human Health Protection

Globally, three-quarters of all humans are projected to live in cities by 2050. Air and water pollution are at critical levels in dense, urban areas where millions of inhabitants suffer the attendant health and lifestyle consequences. A robust and well-planned urban forest can significantly mitigate health risks associated with pollution while bolstering overall city livability.

Supplemental Keywords:

air pollution, urban climate, sensors

Progress and Final Reports:

  • 2015
  • Final