Water managers in the Colorado River Basin are increasingly challenged to balance conflicting multi-sectoral water demands. The state of Colorado’s West Slope basins play a critical role in delivering water to the Lower Colorado River Basin and supporting the state of Colorado’s local economy and natural environment. Internal variability – irreducible uncertainty stemming from interactions across non-linear processes within the hydroclimate system – complicates future vulnerability assessment. The historical streamflow record in the West Slope represents a single realization of an inherently stochastic process, which does not capture the full extent of internal variability. Climate change further complicates drought vulnerability assessment – downscaled climate modeling suggests the potential for significant declines in West Slope streamflows by mid-century. In this work, we contribute a novel multi-site Hidden Markov Model (HMM)-based synthetic streamflow generator that captures the stationary internal variability of streamflows across the six West Slope Basins. Our HMM preserves the spatial and temporal correlation structures within and between basins by modeling two “hidden” climate states, representing wet and dry hydrologic conditions. To capture spatially compounding drought events, our HMM uses a common state for all basins in any given year and a single transition matrix that dictates the likelihood of persisting and transitioning between a wet and dry state. We explore the impact of a middle-of-the-road climate change scenario by perturbing the HMM parameters to generate a climate-adjusted ensemble of streamflow. Our results reveal that drought events that emerge from the system’s stationary internal variability can exceed historically observed droughts in magnitude, duration, and spatial extent. Streamflow records from the climate-adjusted ensemble reveal a regime shift, where previously rare droughts become routine events. We illustrate the increased multi-sectoral drought vulnerability stemming from these drought events by routing the streamflow ensembles through StateMod, the state of Colorado’s water allocation model. Our methodology can be expanded to other snow-dominated river basins, and our results can be used to inform future Colorado River planning efforts.