Habitat fragmentation and habitat loss diminish population connectivity, reducing genetic diversity and increasing extinction risk over time. Improving connectivity is widely recommended to preserve the long-term viability of populations, but this requires accurate knowledge of how landscapes influence connectivity. Detectability of landscape effects on gene flow is highly dependent on landscape context, and drawing conclusions from single landscape studies may lead to ineffective management strategies. We present a novel approach to elucidate regional variation in the relative importance of landscape variable effects on gene flow. We demonstrate this approach by evaluating gene flow between isolated, genetically impoverished mountain goat (Oreamnos americanus) populations in Washington and much larger, genetically robust populations in southern British Columbia. We used GENELAND to identify steep genetic gradients and then employed individual-based landscape genetics in a causal modeling framework to independently evaluate landscape variables that may be generating each of these genetic gradients. Our results support previous findings that freeways, highways, water, agriculture and urban landcover limit gene flow in this species. Additionally, we found that a previously unsupported landscape variable, distance to escape terrain, also limits gene flow in some contexts. By integrating GENELAND and individual-based methods we effectively identified regional limiting factors that have landscape-level implications for population viability.