Climatically sensitive species are naturally expected to experience disproportionate impacts from climate change. Although many vertebrates select microhabitats that allow them to persist toward the edge of their physiological limits, climate-driven models are often conducted at spatial scales too coarse to capture these fine-scale behaviors. One such climatically limited species is the monotypic Mountain Beaver (Aplodontia rufa), a species endemic to western North America. Mountain Beavers have physiological constraints that limit their distribution to cool, moist habitats. Although they have persisted through past periods of climate change, increasing temperature since the last glacial maximum is believed to have had a strong influence in reducing their range. This is particularly true for the federally endangered Point Arena subspecies of Mountain Beaver (Aplodontia rufa nigra; hereafter PAMB), found toward the southwestern edge of their range in Mendocino County on the coast of California. Previous efforts to model PAMB habitat suitability were conducted at the landscape scale. However, as a climatically sensitive taxon with limited mobility, it is important to understand the influence of microclimates on habitat suitability, particularly in the context of global climate change. We created downscaled climate data that we combined with known locations of PAMB to create a fine-scale species distribution model. Mean temperature was the strongest predictor of PAMB presence, with an increased probability of presence at lower mean temperatures. Although PAMB may be living at the edge of their physiological limits, they appear to be selecting habitat at a microclimate scale similar to elsewhere in their range, which allows them to persist at this margin. Conservation efforts for PAMB should prioritize habitat restoration, as well as maintaining and increasing connectivity between microrefugia, to provide cooler areas for PAMB persistence.