Accidental introductions of rodents present one of the greatest threats to indigenous island biota. On uninhabited remote islands, such introductions are most likely to come from shipwrecks. Here we use a comprehensive database of shipwrecks in Western Alaska to model the frequency of shipwrecks per Aleutian and Bering Sea island, taken as a proxy for the likelihood of rodent introductions, using physical variables, and the intensity of nearby fishing traffic and activity as predictors. Using data spanning from 1950 to 20114, we found that shipwrecks were particularly common in the 1980s to early 2000s, with a major peak in wrecks during the late 1980s. Amount of fishing activity within 5 km of an island was the strongest predictor of shipwrecks, followed by the strength of tidal currents and density of large-vessel traffic. Islands with the highest frequency of shipwrecks are all in the eastern Aleutians, including Unimak, Unalaska, and Akun Islands. By contrast, the largest seabird colonies are in the western Aleutian and Pribilof Islands, including Buldir, Kiska, and Saint George islands. Multiplying the frequency of a shipwreck by the number of seabirds breeding per island provides a measure of risk. The risk of rodent introductions from shipwrecks to seabirds was then greatest for Saint George (Bering Sea), Buldir (Western Aleutians) and Saint Matthew islands (Bering Sea). Keeping these high-risk islands rodent free should be a conservation priority. Most islands with a high predicted frequency of shipwrecks already have established rodent populations and therefore few remaining seabirds. Of those islands with established rodent populations, Attu and Kiska Islands would make suitable targets for eradication, given their relatively low expected frequency of shipwrecks for their size.
