Paper No. 9-8
Presentation Time: 10:10 AM
HUNTING FOR NATURE’S REFRIGERATORS AMONG COLLUVIAL AND KARST LANDFORMS OF THE APPALACHIAN MOUNTAINS
Cold-air traps throughout the Central Appalachians support unusual local ecosystems that commonly represent refugia for plant and animal species normally found in more northerly climates or at higher elevations. Historically coined “nature’s refrigerators”, several well-known examples lie in as parks or designated as natural areas, although some have been destroyed or degraded though anthropogenic activity such as highway construction and timbering. An unknown number of potentially-at-risk sites remain undocumented in the scientific literature and virtually unmapped. Several approaches have contributed to an ongoing effort to better understand the factors that control the distribution of these unique ecosystems. Word-of-mouth networking and review of “nature” literature has proven to be a successful, but time-consuming, approach to enlarging the inventory of well-documented sites, particularly in areas oft-visited by highly literate outdoor enthusiasts. Hand-held thermal imaging, such as FLIR camera imagery, is very revealing on warm afternoons during spring seasons when minimal leaf cover and sharp thermal contrasts allow identification of cold air outflows at a sub-meter scale. Leaf cover during the growing season reduces the range of detection for cold-air traps, although close-range FLIR images do allow identification and delineation of cold air outflows under tree canopies. Unfortunately, the scale and resolution of widely available satellite-based thermal imagery is typically too coarse to allow delineation of most known ecologically significant cold air traps. LiDAR-derived DEMs supplemented by high-quality aerial photography provide an opportunity to characterize the characteristics of landforms that support cold-air traps, but our preliminary findings indicate that nature’s refrigerators form in a variety of settings that may be difficult to accurately predict using simple attribute-based models.