This dataset holds streamflow sequences for each of three drought scenarios developed to characterize plausible future drought conditions in the Colorado River Basin. These sequences were produced using the methods described in Center for Colorado River Studies Future of the Colorado River Project white paper 4 entitled “The Future Hydrology of the Colorado River Basin” by Salehabadi, Tarboton et al. (2020) and paper Salehabadi, H., D. G. Tarboton, B. H. Udall, K. G. Wheeler and J. C. Schmidt, (2022), "An Assessment of Potential Severe Droughts in the Colorado River Basin," JAWRA Journal of the American Water Resources Association, https://doi.org/10.1111/1752-1688.13061. This study defined three drought scenarios: (1) Millennium Drought, (2) Mid 20th Century Drought and (3) Paleo Tree Ring Severe Drought. The first two were defined using the US Bureau of Reclamation Natural flows from 2000-2018 and 1953-1977 respectively. The last was defined using the years 1576-1600 from the Meko et al., 2017 tree ring reconstruction of streamflow at Lees Ferry. 100 streamflow traces, each 42 years long were produced for each scenario by resampling years with replacement. Resampling from identified past drought scenarios, provides test droughts based on past flows that are more severe, due to the variety in the sampling, than any past droughts that have actually occurred. They are nevertheless plausible, since they are derived from past records. We used a nonparametric resampling approach referred to as “Water Year Block Disaggregation” to split the simulated annual flow at Lees Ferry into monthly flow at each of the 29 Colorado River Simulation System (CRSS) natural inflow sites. For the first two scenarios where there are historic natural flows at the 29 CRSS sites, this selects the entire water year block of monthly flows across sites for the corresponding drought year. For the paleo scenario, where there are not flows at each of the sites, the historic natural flow year with annual flow at Lees Ferry closest to the paleo flow is selected, and then flows across the sites and months adjusted by the ratio of paleo flow to closest historic flow.