Location‐specific factors influence patterns and effects of subsistence sea otter harvest in Southeast Alaska

To better understand the spatial context of population dynamics of sea otters (Enhydra lutris) in Southeast Alaska (SEAK), we investigated the spatial and temporal patterns of subsistence sea otter harvest and assessed the effect of harvest on population growth. U.S. federal law permits subsistence harvest of sea otters and sale of clothing and handicrafts made by coastal Alaska Natives. Hunters are required to self‐report these harvests along with information on date, location, age class, and sex. Using harvest data collected from 1988 to 2015, we developed a spatially explicit, age‐structured, density‐dependent population simulation model to explore the potential impacts of harvest on sea otter population dynamics. We examined patterns of harvest and simulation model results at two spatial scales: the SEAK stock and three smaller subregions that vary in sea otter occupation time and carrying capacity: Sitka Sound, Keku Strait, and the Maurelle Islands. Annual sea otter harvest in SEAK increased from 55 animals in 1988 to a reported maximum of 1449 animals in 2013. Estimated mean annual harvest rate was 2.8% at the SEAK stock scale, but ranged from 0% to 39.3% across the three focal subregions described above. Across all subregions (n = 55), annual sea otter harvest rate was strongly influenced by time since recolonization, sea otter population density, and proximity to communities with sea otter hunters. The simulation model predicted population trends and per capita harvest rates similar to those estimated from aerial survey data, providing a reasonable approximation of population dynamics. Results of the simulation model suggested that current harvest levels can reduce population size at both the SEAK and subregional scales. Variation in harvest impacts was a function of subregion‐specific factors, including time since recolonization and population status with respect to carrying capacity. We found that subsistence harvest and its population effects were scale‐ and location‐dependent, indicating that higher spatial and temporal resolution of sea otter population and hunting data could help address emerging sea otter management and conservation concerns in this region. Variation in ecological and demographic processes across different scales can lead to spatial structure in populations (Turner 1989, Dunning et al. 1992). Therefore, effective management of populations requires information about population status and dynamics at spatial scales relevant to the species in question. For populations that are hunted for subsistence, harvest data can provide