A comparison of photograph-interpreted and IfSAR-derived maps of polar bear denning habitat for the 1002 Area of the Arctic National Wildlife Refuge, Alaska

Polar bears (Ursus maritimus) in Alaska use the Arctic National Wildlife Refuge (ANWR) for maternal denning. Pregnant bears den in snow banks for more than 3 months in winter during which they give birth to and nurture young. Denning is one of the most vulnerable times in polar bear life history as the family group cannot simply walk away from a disturbance without jeopardizing survival of newly born cubs. The ANWR includes the “1002 Area”, a region recently opened for oil and gas exploration by the U.S. Department of the Interior (DOI). As a part of its mission, the DOI “… protects and manages the Nation's natural resources …” and is therefore responsible for conserving polar bears and encouraging development of energy potential. Because future industrial activities could overlap habitats used by denning polar bears, identifying these habitats can inform the decisions of resource managers tasked to develop resources and protect polar bears. To help inform these efforts, we qualitatively compared the distribution of denning habitat identified by two different methods: previously published habitat from manual interpretation of aerial photographs, and habitat derived by computer interrogation of interferometric synthetic aperture radar (IfSAR) digital terrain models (DTM). Because photograph-interpreted methods depicted denning habitat as a line and IfSAR-derived methods depicted habitat as a polygon, we assessed agreement between the two methods with distance measurements. We found that 77.5 percent of IfSAR-derived denning habitat (79.6 km2 ; 1.2 percent of the 6,837.0 km2 1002 Area) was within 600 m of photograph-interpreted habitat (3,026.9 km), including 53.9 percent within 200 m. This distribution differed from that of randomly distributed points, as only 49.4 percent of these occurred within 600 m of photograph-interpreted habitat, including 18.3 percent within 200 m. Both methods appear to identify the major physiographic features that polar bears might select for denning. IfSAR-derived methods identified habitat at greater frequency beyond major landscape features such as coastal bluffs, river banks and lakeshores, were more likely to identify isolated pockets of putative denning habitat, and were easier to implement than deriving habitat from photograph-interpretive efforts. However, previous research suggests that photograph-interpretation methods may identify denning habitat more correctly than computer interrogation of IfSAR DTMs. Future work should quantify the distribution of IfSAR-derived denning habitat relative to actual landscape features and polar bear maternal dens in the 1002 Area, and investigate the feasibility of habitat identification from finer grained DTMs.