Comparing aerial detection and photo interpretation for conducting forest health surveys

Abstract

Surveys to detect damage on forested lands occur annually across the United States. Historically, much of these data have been collected by observers in aircraft documenting visible damage using a technique known as aerial detection survey (ADS) or sketch mapping. ADS is an efficient method but produces coarse scale data. We used one meter natural color imagery to conduct a photo interpretation survey (PIS) to compare the resulting data with the annual ADS data. The study area consists of a portion of the Black Hills National Forest that is currently experiencing mountain pine beetle-induced conifer mortality. Ground plots were used to compare how accurately each survey method delineated forest damage and results indicate a statistically significant difference (alpha 0.05) between ADS and PIS and a trend toward no significant difference when a buffer is considered. Substantial differences in the average size of polygons, total number of polygons, and total area mapped were observed. Our results indicate that current estimates of mortality intensity in our study area are low and remedies should be investigated. Photo surveys over the Black Hills National Forest are less efficient but yield damage polygons with higher accuracy and with fewer errors of commission than traditional ADS. Copyright © 2013 by the Society of American Foresters.