OOS 10-7 - Remote sensing of habitat characteristics for at-risk species on military installations in the Southeast

Tuesday, August 4, 2009: 10:10 AM
Mesilla, Albuquerque Convention Center
H. Alexis Londo1, David L. Evans2, Scott A. Tweddale3, Joelle M. Carney1, Scott Roberts1 and Peter V. Campbell4, (1)Department of Forestry, Mississippi State University, Mississippi State, MS, (2)Hawk Ridge Bird Observatory, Duluth, MN, (3)Construction Engineering Research Laboratory, US Army Corps of Engineers, Champaign, IL, (4)Sandhills Sub-Office, U.S. Fish and Wildlife Service, Southern Pines, NC
Background/Question/Methods

Military installations in the southeastern United States include areas inhabited by several federally endangered or at-risk species, including the red-cockaded woodpecker (RCW Picoides borealis) and the gopher tortoise (Gopherus polyphemus). Field assessment of pine forest habitat conditions and suitability for these species is costly and time consuming, requiring measurements to describe vegetation size, density, structure, and age, often over large landscapes. Remote sensing techniques can be useful in attaining many of these measurements. The location, size, and distribution of individual trees can be derived from LiDAR (Light Detection and Ranging), and these measurements expanded to the stand level for further analyses. The vertical distribution of LiDAR points can be used to describe the vertical structure of the vegetation in a stand – often an important element of habitat suitability. Classification of multi-spectral imagery can yield information on tree species or types (pine vs. hardwood) which can provide predictions of forest covertype. This remotely sensed information  can be used to determine current habitat suitability. Areas potentially available or suitable for management to create new habitat, or enhance current habitat values can be identified, as can areas not suitable for management.

Results/Conclusions

A remote sensing based spatial analysis to assess RCW habitat was developed on Ft. Bragg and surrounding areas in North Carolina. Object-oriented image classification yielded covertype and distinguished loblolly (Pinus taeda) and longleaf (Pinus palustris) pine with an accuracy of 81% when combining both pine species into one class, and 74% when classifying pine species separately. LiDAR-based estimates of mean basal area per hectare for canopy trees was not significantly different from field measurements of basal area for all trees on 69 plots distributed across three sample areas (α = 0.05). Density of midstory/understory hardwoods derived from LIDAR in 4 height strata were positively correlated with field measurements of total cover. The techniques developed in the RCW study are being investigated for suitability in the assessment of gopher tortoise habitat at Camp Shelby in Mississippi.

Copyright © . All rights reserved.
Banner photo by Flickr user greg westfall.