To maintain diverse habitat and migration corridors, land managers need to incorporate climatic variation and competitive interactions with exotic species into long-term conservation plans. We combined field studies with spatial analyses of native cottonwood trees, (Populus fremontii, P. deltoides, P. angustifolia), and the invasive exotic tamarisk tree (Tamarix spp.). We found four major patterns. 1. A common tree, broadleaf cottonwood, currently has a greater potential niche than other cottonwoods (narrowleaf (P. angustifolia) and hybrids), and tolerates the greatest environmental variation, but can become rare under extended extreme drought conditions. 2. Sixty-two percent of the variation in the mortality of broadleaf cottonwoods on the Colorado Plateau (P. fremontii) and sixty-nine percent of the variation in mortality in Central Arizona was associated with the density of tamarisk. 3. Tamarisk will increase cottonwood forest fragmentation. 4. Ongoing drought is causing a dominance shift from native cottonwoods to exotic tamarisk in mainstem river channels across the southwest. We used the Desktop GARP software to produce a spatial model that accurately predicted the distribution of upper- and lower-elevation cottonwood species, and their overlap was a significant predictor of hybrid tree locations as verified by three independent validations. These finding have 2 major implications. 1. Our method identifies riparian areas in the southwest U.S. that are most drought sensitive and most resilient, which provides a basis for prioritizing management. 2. An exotic species is impacting native riparian forests in a manner similar to climate change. We argue that proactive management, coordinated across a large region, should maintain climate refugia in the most resilient areas and mitigate compounding pressures in the most vulnerable areas to preserve biodiversity through future climate fluctuations.