The combined impact of killer whale consumptive and non-consumptive effects on northern sea otter population viability in the Western Aleutians Archipelago, Alaska

Predators can alter the abundance, distribution, and behavior of prey populations through consumptive and non-consumptive effects. In the Aleutian Archipelago of Alaska, killer whales (Orcinus orca) are considered the most probable cause of northern sea otter (Enhydra lutris kenyoni) population declines in the southwestern Alaska stock, which led to their listing as a threatened distinct population segment under the Endangered Species Act. Much of the research attention in the Aleutian Archipelago region has focused on the consumptive effects of killer whales on sea otter population dynamics. Here, we explore non-consumptive effects by accounting for restricted sea otter habitat use within discrete predation refuges characterized by areas of shallow, complex reef habitats close to shore. We constructed Population Viability Analysis (PVA) models that incorporated sea otter count data collected by aerial and skiff-based methods over six decades (1959 to 2021) to inform uplisting (to endangered) and downlisting (delisting from ESA) criteria. Our models incorporated both density dependent effects and density-independent effects on sex and age structure, which we termed predation hazards. Prior to 1990, predation hazards were negligible, fluctuated at high values between 1990 and the early 2000s, and then declined as sea otter populations reached low densities. We estimated a current regional abundance of 2,405 sea otters (95%CI = 1,734 to 3,238) in the Western Aleutians Management Unit. Our base PVA model that considered only inter-island fragmentation indicated the risk of the regional sea otter population becoming endangered was <5% when there were at least 1,500 otters (95%CI = 1,200 to 2,100), and provided a delisting threshold of 2,100 sea otters. A PVA model that accounted for restricted habitat use of sea otters within discrete predation refuges (i.e. inter-island and intra-island fragmentation) indicated a less encouraging potential for sea otter recovery. The probability of the population becoming endangered increased to >10% and the delisting threshold increased to >10,000 sea otters (nearly 5x higher). Our results indicate sea otters within fragmented predation refuges could be more susceptible to the effects of stochastic processes with potentially limited ability for rescue effects. Overall, our research reveals the importance of evaluating both consumptive and non-consumptive effects when considering conservation and management plans for at-risk populations thought to be limited by a predator. ### Competing Interest Statement The authors have declared no competing interest.