Print PageEphemeral wetlands are ecologically important ecosystems throughout North America and occur frequently in the Atlantic coastal plain ecoregion. They are known to support 86 species recognized by the US Fish and Wildlife Service as threatened or endangered, and numerous others that are state protected. Despite the growing consensus of their importance and imperilment, ephemeral wetlands have yet to become a national conservation priority. They are often cryptic on the landscape and methods to detect them remotely have been ineffective at the landscape scales necessary for resource management. This study seeks to fill wetland detection gaps, by implementing and expanding new remote sensing methods. We subjected high resolution LiDAR elevation data to various relief models designed to elucidate fine-scale geomorphology, specifically small localized changes in concavity, as a location predictor of pools. Relief models were then processed with local indicators of spatial association (LISA) in order to improve their predictive success.
Results/Conclusions
We also conducted field verification of 114 wetlands selected using a random stratified design proportional to land cover, to measure model commission (α) and omission (β) error rates. Commission error was approximately 8% (n=70) and tentative omission error was 11% (n=44). Omission error rates may be inflated due to model buffering system and/or edge effects and is undergoing further investigation. These results suggest our local relief models captured small geomorphologic changes that successfully predict ephemeral wetland boundaries in the extremely low-relief Atlantic Coastal Plain. Many small wetlands are centers of biodiversity in forested landscapes and such analyses will aid in their detection and provide information integral for landscape-scale management.
