HR: 0800h
AN: B51D-0237    [Abstracts]
TI: Landscape scale estimation of forest structure and composition using lidar, radar and high resolution imagery
AU: * Drake, J B
EM: jbdrake@mail.ucf.edu
AF: University of Central Florida, Department of Biology, Orlando, FL 32816-2368 United States
AU: Listopad, C
EM: claudia.listopad@us.bureauveritas.com
AF: University of Central Florida, Department of Biology, Orlando, FL 32816-2368 United States
AU: Noble, J
EM: jnoble@ttrs.org
AF: Tall Timbers Research Station, 13093 Henry Beadel Drive, Tallahassee, FL 32312 United States
AU: Masters, R
EM: rmasters@ttrs.org
AF: Tall Timbers Research Station, 13093 Henry Beadel Drive, Tallahassee, FL 32312 United States
AB: The spatial arrangement of forest canopy components such as leaves and branches directly influences a number of key ecosystem characteristics. Forest canopy structure affects many physical, micrometeorological characteristics (e.g., light, temperature, humidity) as well as numerous ecological processes such as tree growth and recruitment. Forest canopy structural and compositional information are critical for land management activities ranging from estimating fuel loads to mapping species' habitats. A persistent challenge has been to actually quantify the physical distribution of forest canopy structure and the variability of structure in space and time. Field studies are typically limited to small plot sizes (< 0.5 ha) and are both expensive and time-consuming. Remote sensing techniques are therefore necessary for estimating forest structure and composition at broad spatial scales. In this study we explore the use of three complimentary and widely used remote sensing techniques (lidar, radar and high resolution imagery) for estimating the structure and composition of pine and hardwood forests at Tall Timbers Research Station in the Red Hills of northwestern Florida. Small footprint airborne lidar data were used to create very high resolution (1 m) models of bare earth topography and canopy height. Additional canopy structural metrics such as canopy cover were also derived from the airborne lidar data. High resolution satellite imagery was used to estimate forest composition (e.g., pine vs. hardwood). Spectral and textural metrics were also derived from the imagery to estimate forest structural properties. Finally 30 m interferometric synthetic aperture radar data from the Shuttle Radar Topography Mission (SRTM) were used in conjunction with National Elevation Dataset (NED) to map the height of the scattering phase center (Kellndorfer et al. 2004). We then compared all remotely sensed estimates of forest canopy structure with field measurements of canopy structure (from approximately 1100 field plots) to evaluate the accuracy of each technique and possible error effects. We found that the lidar estimates of canopy structure were highly accurate and therefore we used the lidar canopy structural estimates to evaluate the imagery and radar structural estimates at the landscape scale. We also combined lidar structural estimates with estimates of forest composition from high resolution imagery to evaluate the effects of canopy height, canopy cover and forest composition on SRTM height estimates. We found that at the stand level SRTM data provides accurate forest height information which could be very helpful for landscape to regional scale applications This study illustrates the relative strengths and weaknesses of different remote sensing techniques for estimating forest structure and provides a foundation with which to explore how these techniques may be combined in a cost- and time-effective manner for future research and management efforts.
DE: 0439 Ecosystems, structure and dynamics (4815)
DE: 0476 Plant ecology (1851)
DE: 0480 Remote sensing
SC: Biogeosciences [B]
MN: Fall Meeting 2005