Mechanistic modeling of environmental drivers of woolly mammoth carrying capacity declines on St. Paul Island

• Published in: Ecology
• Main Authors: Wang, Yue, Porter, Warren, Mathewson, Paul D., Miller, Paul A., Graham, Russell W., Williams, John W.
• Format: Article in Journal/Newspaper
• Language: English
• Published: Ecological Society of America 2018
• Subjects: Ecology; Beringia; carrying capacity; dynamic vegetation model; extinction; Holocene; island biogeography; LPJ-GUESS; mechanistic niche model; megafauna; Niche Mapper; woolly mammoth; Bering Land Bridge
• Online Access: https://lup.lub.lu.se/record/7e031d50-021c-4ab1-a674-25c6bfb45598; https://doi.org/10.1002/ecy.2524

On St. Paul Island, a remnant of the Bering Land Bridge, woolly mammoths persisted until 5,600 yr BP with no known predators or competitors, providing a natural system for studying hypothesized environmental drivers of extinction. These include overheating due to rising temperatures, starvation, and drought. Here, we test these hypotheses using Niche Mapper and LPJ-GUESS to mechanistically estimate mammoth metabolic rates and dietary and freshwater requirements and, from these, estimate variations in island carrying capacity on St. Paul for the last 17,000 yr. Population carrying capacity may have been several hundred individuals at the time of initial isolation from the mainland. Adult mammoths could have fasted for two to three months, indicating a necessary ability to access snow-buried forage. During the Holocene, vegetation net primary productivity increased, but shrinking island area overrode increased net primary productivity (NPP), lowering carrying capacity to ~100 individuals. NPP and freshwater availability alternated as critical limiting factors for this island population during the environmental changes of the late Pleistocene and Holocene. Only two or three individuals could have been sustained by the freshwater surplus in crater lakes (up to 18 individuals under the most optimistic parameter sensitivity experiments), suggesting that the St. Paul mammoth population was highly dependent on coastal freshwater sources. The simulations are consistent with the available proxy data, while highlighting the need to retrieve new paleohydrological proxy records from the coastal lagoons to test model predictions. More broadly, these findings reinforce the vulnerability of island megaherbivore populations to resource limitation and extinction.