THE ANTARCTIC CLIMATE SYSTEM EXHIBITS CONSIDERABLE VARIABILITY ON MULTIPLE TIMESCALES AND THESE CLIMATE VARIATIONS CAN HAVE DRAMATIC IMPLICATIONS FOR SPECIES THAT LIVE IN THE REGION PARTICULARLY THOSE RELIANT ON THE TIMING OR CONCENTRATION OF SEA ICE. ANTARCTIC PACK-ICE SEALS ARE MAJOR PREDATORS OF ANTARCTIC KRILL AND ARE WHOLLY DEPENDENT ON SEA ICE FOR HABITAT AND FORAGING AND REPRODUCTION. LIKE OTHER MARINE MAMMALS PACK-ICE SEALS ARE EXTREMELY CHALLENGING TO SURVEY USING TRADITIONAL MEANS BECAUSE THEY ARE HIGHLY MOBILE WITH LARGE RANGES AND ONLY VISIBLE WHEN HAULED UP ON SEA ICE OR LAND. FOR THIS REASON WE HAVE HAD (UNTIL RECENTLY) NO COST-EFFECTIVE MEANS OF SURVEYING PACK-ICE SEALS AND VIRTUALLY NO PRECISE ESTIMATES OF GLOBAL ABUNDANCE OR POPULATION TRENDS. WHILE WE NOW HAVE THE CAPACITY TO FIND SEALS IN WORLDVIEW-3 IMAGERY OUR ABILITY TO IDENTIFY SEALS TO SPECIES REMAINS IN ITS INFANCY AND RELIES AT PRESENT ON RELATIVELY CRUDE DISTINCTIONS BETWEEN EACH SPECIES PREFERRED HABITAT. THIS ISSUE IS PARTICULARLY ACUTE ON THE ANTARCTIC PENINSULA WHERE WEDDELL CRABEATER AND LEOPARD SEALS OCCUPY SIMILAR AND OFTEN OVERLAPPING HABITAT. THE OBJECTIVES OF OUR PROJECT ARE TO BUILD A STATISTICAL MODEL OF PACK-ICE SEAL SPECIES IDENTIFICATION THAT WOULD ALLOW US TO INFER FOR EACH SEAL DETECTED IN COMMERCIAL SATELLITE IMAGERY THE PROBABILITY OF BELONGING TO EACH SPECIES IN THE REGION. THIS MODEL WOULD ALLOW US TO REFINE OUR POPULATION ESTIMATES OF PACK-ICE SEALS IN ANTARCTICA IMPROVE ALGORITHMS FOR PACK-ICE SEAL DETECTION USING COMMERCIAL SATELLITE IMAGERY AND WILL CONTRIBUTE TO ONGOING WORK TO FORECAST SEAL POPULATIONS UNDER ENVIRONMENTAL CHANGE.