Remote Sensing Procedures to Update Forested Geospatial Datasets after a Landscape Altering Event

No Thumbnail Available



Journal Title

Series/Report No.

Journal ISSN

Volume Title



The creation of accurate geospatial datasets like vegetation and fire fuel loads is a time consuming effort and these datasets are routinely used by resource managers. Therefore the accuracy of these datasets is vital. Vegetation and fire fuel load datasets often represent a dynamic landscape and landscape altering events such as a wildland fire or a hurricane can drastically change that landscape. The goal of this research is to investigate the use of automated change detection techniques that can not only indicate areas of change but also quantify the magnitude of change that occurred as well. Hurricane Isabel did extensive damage to the forest landscapes of central Virginia in September of 2003, specifically Petersburg National Battlefield. The Rocky Top Fire occurred in July of 2002 in Shenandoah National Park, resulting in a mosaic pattern of burns, covering roughly 1500 acres. The objective of this research was to test the use of remote sensing procedures to update vegetation and fire fuel load spatial datasets. First, using digital orthorectified photomosaics, the automated feature extraction technique Visual Learning System's Feature Analyst, was employed to delineate forest damage following Hurricane Isabel. Second, the satellite based remote sensing technique Normalized Burn Ratio, was utilized to delineate and quantify burn severity on vegetation after the Rocky Top Fire. A third objective was to estimate fire behavior differences between the existing pre-event and the remotely sensed post-event fuel load datasets using the FARSITE model, thereby cataloging the potential need for vegetation and fuel load updates. The results of this research show that, 1) VLS Feature Analyst is an excellent indicator of downed woody debris, 2) the Normalized Burn Ratio is the best technique available for indicating and quantifying the effects of a wildland fire on the landscape, 3) changes in assigned Fuel Models, especially in the Logging Slash group, affect FARSITE outcomes, and 4) Fuel Models should be assigned based on expected fire behavior, not on the total fuel loading.



Remote Sensing, Natural Disturbance, Automated Feature Extraction, Wildland Fire, Change Detection, Burn Severity, National Park Service, GIS, Fuel Mapping





Natural Resources