Trends and Carrying Capacity of Sea Otters in Southeast Alaska

Sea otter populations in Southeast Alaska, USA, have increased dramatically from just over 400 translocated animals in the late 1960s to >8,000 by 2003. The recovery of sea otters to ecosystems from which they had been absent has affected coastal food webs, including commercially important fisheries, and thus information on expected growth and equilibrium abundances can help inform resource management. We compile available survey data for Southeast Alaska and fit a Bayesian state-space model to estimate past trends and current abundance. Our model improves upon previous analyses by partitioning and quantifying sources of estimation error, accounting for over-dispersion of aerial count data, and providing realistic measurements of uncertainty around point estimates of abundance at multiple spatial scales. We also provide estimates of carrying capacity (K) for Southeast Alaska, at regional and sub-regional scales, and analyze growth rates, current population status and expected future trends. At the regional scale, the population increased from 13,221 otters in 2003 to 25,584 otters in 2011. The average annual growth rate in southern Southeast Alaska (7.8%) was higher than northern Southeast Alaska (2.7%); however, growth varied at the sub-regional scale and there was a negative relationship between growth rates and the number of years sea otters were present in an area. Local populations vary in terms of current densities and expected future growth; the mean estimated density at K was 4.2 ± 1.58 sea otters/km² of habitat (i.e., the sub-tidal benthos between 0 m and 40 m depth) and current densities correspond on average to 50% of projected equilibrium values (range = 1–97%) with the earliest-colonized sub-regions tending to be closer to K. Assuming a similar range of equilibrium densities for currently un-occupied habitats, the projected value of K for all of Southeast Alaska is 74,650 sea otters. Future analyses can improve upon the precision of K estimates by employing more frequent surveys at index sites and incorporating environmental covariates into the process model to generate more accurate, location-specific estimates of equilibrium density. © 2019 The Authors. The Journal of Wildlife Management Published by Wiley Periodicals, Inc.