Forecasting ecological and evolutionary strategies to global change: an example from habitat selection by lemmings

Ecologists and evolutionary biologists must develop theories that can predict the consequences of global warming and other impacts on Earth's biota. Theories of adaptive habitat selection are particularly promising because they link distribution and density with fitness. The evolutionarily stable strategy that emerges from adaptive habitat choice is given by the system's habitat isodar, the graph of densities in pairs of habitats such that the expectation of fitness is the same in each. We illustrate how isodars can be converted into adaptive landscapes of habitat selection that display the density- and frequency-dependent fitness of competing strategies of habitat use. The adaptive landscape varies with the abundance of habitats and can thus be used to predict future adaptive distributions of individuals under competing scenarios of habitat change. Application of the theory to three species of Arctic rodents living on Herschel Island in the Beaufort Sea predicts changes in selection gradients as xeric upland increases in frequency with global warming. Selection gradients will become more shallow for brown lemming (Lemmus trimucronatus) and tundra vole (Microtus oeconomus) strategies that preferentially exploit mesic habitat. Climate change will cause selection gradients for the alternative strategy of using mostly xeric habitat to become much steeper. Meanwhile, the adaptive landscape for collared lemmings (Dicrostonyx groenlandicus), which specialize on xeric Dryas-covered upland, will become increasingly convex. Changes in the adaptive landscapes thus predict expanding niches for Lemmus and Microtus, and a narrower niche for Dicrostonyx. The ability to draw adaptive landscapes from current patterns of distribution represents one of the few methods available to forecast the consequences of climate change on the future distribution and evolution of affected species.