Management and Conservation Article Estimating Carrying Capacity for Sea Otters in British Columbia EDWARD J. GREGR, 1 SciTech Environmental Consulting, 2136 Napier Street, Vancouver, BC V5L 2N9, Canada LINDA M. NICHOL, Fisheries and Oceans Canada, Pacific Biological Station, 3190 Hammond Bay Road, Nanaimo, BC V9T 6N7, Canada JANE C. WATSON, Malaspina University-College, 900 Fifth Street, Nanaimo, BC V9R 5S5, Canada JOHN K. B. FORD, Fisheries and Oceans Canada, Pacific Biological Station, 3190 Hammond Bay Road, Nanaimo, BC V9T 6N7, Canada GRAEME M. ELLIS, Fisheries and Oceans Canada, Pacific Biological Station, 3190 Hammond Bay Road, Nanaimo, BC V9T 6N7, Canada ABSTRACT We estimated carrying capacity for sea otters (Enhydra lutris) in the coastal waters of British Columbia, Canada, by characterizing habitat according to the complexity of nearshore intertidal and sub-tidal contours. We modeled the total area of complex habitat on the west coast of Vancouver Island by first calculating the complexity of the Checleset Bay–Kyuquot Sound (CB–KS) region, where sea otters have been at equilibrium since the mid-1990s. We then identified similarly complex areas on the west coast of Vancouver Island (WCVI model), and adapted the model to identify areas of similar complexity along the entire British Columbia coast (BC model). Using survey data from the CB–KS region, we calculated otter densities for the habitat predicted by the 2 models. The density estimates for CB–KS were 3.93 otters/km 2 and 2.53 otters/km 2 for the WCVI and BC models, respectively, and the resulting 2 estimates of west coast of Vancouver Island complex habitat carrying capacity were not significantly different (WCVI model: 5,123, 95% CI ¼ 3,337–7,104; BC model: 4,883, 95% CI ¼ 3,223–6,832). The BC model identified the region presently occupied by otters on the central British Columbia coast, but the amount of coast-wide habitat it predicted (5,862 km 2 ) was relatively small, and the associated carrying capacity estimate (14,831, 95% CI ¼ 9,790–20,751) was low compared to historical accounts. We suggest that our model captured a type of high-quality or optimum habitat prevalent on the west coast of Vancouver Island, typified by the CB–KS region, and that suitable sea otter habitat elsewhere on the coast must include other habitat characteristics. We therefore calculated a linear, coast-wide carrying capacity of 52,459 sea otters (95% CI ¼ 34,264–73,489)—a more realistic upper limit to sea otters in British Columbia. Our carrying capacity estimates are helping set population recovery targets for sea otters in Canada, and our habitat predictions represent a first step in Critical Habitat identification. This habitat-based approach to estimating carrying capacity is likely suitable for other nonmigratory, density-dependent species. (JOURNAL OF WILDLIFE MANAGEMENT 72(2):382–388; 2008) DOI: 10.2193/2006-518 KEY WORDS British Columbia, carrying capacity, endangered species, Enhydra lutris, habitat model, optimum habitat, recovery, sea otter, Species at Risk Act (SARA), species–habitat relationship. Sea otters (Enhydra lutris) are currently listed as Threatened under the Canadian Species at Risk Act (SARA), though downlisting to Special Concern was recommended in April 2007 by the Committee on the Status of Endangered Wildlife in Canada (2007). The recovery planning process for species at risk under SARA requires recovery goals. Population recovery goals may be set relative to estimates of the regional carrying capacity (Laidre et al. 2001, 2002). We developed models to estimate habitat area and derived carrying capacity estimates to assist in formulating recovery goals for the British Columbia sea otter population. Sea otters are a nonmigratory, nearshore species that feed primarily on benthic invertebrates in depths ,40 m (Riedman and Estes 1990, Bodkin et al. 2004). Sea otter populations are density dependent with growth believed to be regulated by prey abundance. When sea otters expand into areas where prey are abundant, population growth is rapid until food becomes limiting and the population reaches equilibrium (i.e., births are offset by mortality and emigration; Estes 1990). This density dependence and nonmigratory behavior allows a habitat approach to be used for estimating carrying capacity (K). Using a habitat approach, K is estimated by multiplying density of sea otters calculated for a portion of their range where they are at equilibrium by the total area defined as suitable habitat (DeMaster et al. 1996). This approach has been used in California (DeMaster et al. 1996, Laidre et al. 2001), Washington (Laidre et al. 2002), and Alaska, USA (Burn et al. 2003). The approach requires 1) sufficient survey data of representative areas where the population is at equilibrium and 2) spatial definitions of sea otter habitat. Calculation of available habitat requires a spatial model supported by physical data. In each of the above studies, the modeling approach was adapted to available physical data to characterize sea otter habitat. For instance, comprehensive substrate data available for coastal California and Wash- ington allowed Laidre et al. (2001) to divide the Californian coast into 3 habitat classes: sandy, rocky, and mixed. Defining the seaward extent of the habitats as the 40-m isobath allowed Laidre et al. (2001) to calculate the area of each habitat class. Laidre et al. (2001) estimated maximum densities of otters (0.92–5.15 otters/km 2 ) for each habitat class based on survey data from representative sections of each habitat class where otters were believed to be at equilibrium. In Washington, where the 40-m isobath extends 10–15 km from shore, use of the 40-m isobath as a seaward habitat boundary was unreasonable, so a K based on coastline length was used because it was a better indicator of K in that area (Laidre et al. 2002). Burn et al. (2003) used a simpler habitat model for the Aleutian Islands, defining suitable otter habitat as the total area contained within 400 m from shore, the 40-m isobath, and bays and fjords ,6 km across. 1 E-mail: ed@scitechconsulting.com 382 The Journal of Wildlife Management  72(2)