Fusion of LiDAR and Imagery for Estimating Canopy Fuel Metrics in Eastern Washington ForestsEPA Grant Number: FP916945
Title: Fusion of LiDAR and Imagery for Estimating Canopy Fuel Metrics in Eastern Washington Forests
Investigators: Erdody, Todd
Institution: University of Washington
EPA Project Officer: Just, Theodore J.
Project Period: September 1, 2008 through September 1, 2009
RFA: STAR Graduate Fellowships (2008) RFA Text | Recipients Lists
Research Category: Academic Fellowships
Fuel regimes of the American west have been altered as a result of fire suppression and timber management. This has led to increased fuel loading, and as a result land managers are in need of precise information about the fuels they manage, especially around the wildland urban interface. Canopy fuel metrics such as canopy height, canopy base height and canopy bulk density are specific inputs for wildfire models such as FARSITE and FlamMap. These models are used to predict wildfire growth and crown fire initiation and propagation based on a combination of fuels, weather and topographic information. Currently, the raster layers for these metrics in these models are homogenous at coarse spatial resolutions. If finer resolution data were used, accurate quantification of the forest with more spatial heterogeneity can be accomplished. Light Detection and Ranging (LiDAR) and color near-infrared imagery are systems that have been utilized for measuring various forest structure characteristics at high resolution. LiDAR can provide spatial, 3-dimensional information while high-resolution imagery can provide spectral information in the visible and near-infrared bands. This project will assess how well these two different high-resolution remote sensing technologies can estimate canopy fuels in the forests of eastern Washington State.
Regression models derived from LiDAR metrics, imagery metrics and a fusion of the two will be developed at the Ahtanum State Forest for Ponderosa pine (Pinus ponderosa) stands representative of eastern Washington State. Models will be validated using a cross-validation procedure as well as with data from a similar forest type in the Colville National Forest. Raster layers can then be produced for use in wildfire models and subsequently in smoke models.
Through a fusion of LiDAR and high-resolution imagery, and thus a combination of spatial and spectral information, estimations of canopy fuels will be improved. By improving the ability to estimate canopy fuels at higher resolutions, spatially explicit fuels layers can be created and used in wildfire and smoke models, like FOFEM, leading to more accurate estimations of crown fire risk and particulate emissions. Air quality districts and fire managers will be able to make improved assessments of smoke impacts and will be able to plan as needed by notifying the public about air quality and associated health risks.