Wildlife species may respond to industrial development with changes in distribution. However, discerning a response to development from differences in habitat selection is challenging. Since the early 1990s, migratory tundra Bathurst caribou Rangifer tarandus groenlandicus in the Canadian Arctic have been exposed to the construction and operation of two adjacent open-pit mines within the herd's summer range. We developed a statistical approach to directly estimate the zone of influence (area of reduced caribou occupancy) of the mines during mid-July-mid-October. We used caribou presence recorded during aerial surveys and locations of satellite-collared cow caribou as inputs to a model to account for patterns in habitat selection as well as mine activities. We then constrained the zone of influence curve to asymptote, such that the average distance from the mine complex where caribou habitat selection was not affected by the mine could be estimated. During the operation period for the two open-pit mines, we detected a 14-km zone of influence from the aerial survey data, and a weaker 11-km zone from the satellite-collar locations. Caribou were about four times more likely to select habitat at distances greater than the zone of influence compared to the two-mine complex, with a gradation of increasing selection up to the estimated zone of influence. Caribou are responding to industrial developments at greater distances than shown in other areas, possibly related to fine dust deposition from mine activities in open, tundra habitats. The methodology we developed provides a standardized approach to estimate the spatial impact of stressors on caribou or other wildlife species.