Yellow Rails (Coturnicops noveboracensis) are among the most secretive bird species in North America. They are poorly sampled by common survey protocols, and as a result their occurrence across much of their range is uncertain. We compiled occurrence records of the species and used resource selection functions to classify habitats as selected, neutral, or avoided using four different land cover maps in the oil sands region of northeastern Alberta. We assessed the accuracy of these maps using 279 previously unsurveyed locations and showed that a consensus-based ensemble classifier predicted occurrence more accurately than any single map. We combined the four maps into one map that rated habitat on a scale from 0 (consensus avoided) to 8 (consensus selected). Occupancy analysis showed increasing occupancy rates in areas with higher habitat suitability classes, with maximum occupancy rates of 0.18 (95% CI: 0.07-0.32) in class 8 habitat. We combined detections of 169 male Yellow Rails at surveyed locations with model predictions for unsurveyed locations to produce two population estimates for our study area, based on two estimates of the detection radius of the species. The estimate assuming a 150-m detection radius was 2747 males (95% CI: 588-5563), and the estimate assuming a 250-m detection radius was 1650 males (95% CI: 416-3266). Although estimates contained substantial uncertainty, our results suggest a larger number of Yellow Rails in the region than previously thought, which alters the current understanding of the distribution of this species. We estimated that about 17% of the population in our study area resides on oil sands leases that cover 14% of the study area, in habitats facing ongoing and future industrial development. The availability of a habitat map based on empirical evidence and detailed analyses for this species of conservation concern will improve targeted monitoring and promote mitigation of potential effects of development.