As wildfire activity increases in the western US due to warmer and dryer conditions, managers are increasingly concerned about fire-facilitated transitions from forest to non-forest. Concerns are heightened in California and the Southwest, regions that are currently being affected by increased fire activity, and where fire exclusion and historic forest management activities have led to unnaturally dense vegetation and abundant ladder fuels in some forest types. As such, this study was intended to quantify risk of fire-facilitated transition from forest to non-forest in California, Arizona, and New Mexico. To do so, we compartmentalized our study into three modules reflecting the incremental nature of our approach. The first is the 'climate module', in which we developed and implemented an approach to assess potential climate-induced changes in vegetation and fire regimes over broad spatial scales. The climate module, however, assumes that changes in climate, whether gradual or abrupt, result in immediate changes to vegetation and fire regimes; this assumption is strained given that climate-mediated changes in vegetation are often delayed and require the catalyst of disturbance. Consequently, the ‘disturbance module’ was developed to characterize the drivers of and map the potential for stand-replacing fire as a disturbance agent. The 'climate and disturbance module' applied the methods from the climate module and incorporated stand-replacing fire from the disturbance module to evaluate the potential for fire-facilitated conversion to non-forest. Results from the climate module indicate that vegetation patterns and fire regime characteristics (i.e. fire frequency and severity) and will likely shift in response to climate change. We revealed a potential climatic tipping point in which small changes in climate may result in conversion from forest to non-forest. In terms of fire regime characteristics, we show that universal increases in fire frequency and severity should not be expected; both increases and decreases in fire frequency and severity are projected and depend on the bioclimatic context. However, results from the climate module should only be interpreted as a longer range projection (at least 50-100 years) since disturbance as a catalyzing agent was not considered. The disturbance module reveals that fuel and fire weather are the most important factors driving stand-replacing fire, though fuel was on average 1.7 times more influential than weather across our study area. Geospatial layers depicting the probability of stand-replacing fire were produced in the disturbance module and are available for download. In turn, these maps were used in the climate and disturbance module to characterize the potential for fire-facilitated conversion to non-forest. The climate and disturbance module shows that 4.3% of forest in the southwestern US is at risk of fire-facilitated conversion to non-forest when assuming fire burns under average weather conditions. When assuming that fire burns under extreme weather conditions, 30% of forest is at risk of conversion to non-forest. In conducting this study, we delivered numerous in-person and web-based presentations to land managers and other interested parties (e.g. non-governmental organizations) in the southwestern US and elsewhere. We also delivered several presentations at professional conferences and published our results in peer-reviewed scientific journals. The mapped fire severity predictions (and metadata) we produced in the disturbance module are available online for download. See Appendix B for a full accounting of our science delivery.