OOS 23-8 - Integrating LiDAR and Advanced Land Imager (ALI) data in landcover mapping of the U.S. Virgin Islands (USVI) and using the U.S. National Vegetation Classification to integrate Puerto Rico and the USVI with national GAP landcover datasets

Wednesday, August 10, 2011: 10:30 AM
12A, Austin Convention Center
William A. Gould, International Institute of Tropical Forestry, San Juan, PR and Gary S. Potts, International Institute of Tropical Forestry, USDA Forest Service, Río Piedras, PR
Background/Question/Methods

The land cover of Puerto Rico and the US Virgin Islands (USVI) have been recently mapped as part of the National Gap Analysis Program. Traditionally, GAP projects have relied on satellite imagery from the Landsat program to derive land cover habitat maps at 30m spatial resolution. The Puerto Rico GAP incorporated the Landsat 7 ETM+ 15m panchromatic band to enhance the spatial resolution of the multispectral bands in order to improve the delineation of habitats. Current Landsat 7 ETM+ imagery and scene acquisition is limited by the scan line correction error and the use of criteria that includes cloud-cover forecasts to guide Landsat image collection. These limitations make the collection and use of images difficult for tropical humid regions.  As a consequence, for the USVI GAP, we used pan-sharpened 10m spatial resolution data from the Advanced Land Imager (ALI) onboard the Earth Observation 1 (EO-1) satellite. Additionally, various products were derived from discrete raw LiDAR data, i.e., bare earth surface elevations, canopy cover, and canopy height, to refine the land cover classification.

The map products for Puerto Rico and the USVI are used in modeling animal species-habitat relationships and mapping vertebrate species distributions as part of the National Gap Analysis Program.  They have not been integrated with the National Gap landcover map pending the development of a compatible vegetation classification. We have modified the original classification to best fit with the U.S. National Vegetation Classification (NVC) system in order to accomplish this, and discuss progress and problems in this process.

Results/Conclusions

Seventy land cover classes have been described for Puerto Rico and mapped at a 15m spatial resolution, fifty-one of these occur in the USVI and have been mapped at a 10m spatial resolution. Initial classification is based on the hierarchical controls of climate, substrate, and topography on vegetation and includes forested, shrubland, and grassland classes of natural vegetation divided in dry, mesic, wet, and flooded categories on differing substrates (e.g., calcareous or non calcareous). Problems in integrating the mapped classes with the NVC are related gaps in vegetation description of tropical forests types specific to the islands of the U.S. Caribbean, and inconsistencies in the level of detail of description of tropical vegetation — where the number of plant species is high relative to temperate regions and variation in species composition related to disturbance, land use history and biogeography complicates classifications based on species composition.

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