Water managers are increasingly interested in better understanding and planning for projected resource impacts from climate change. In this management-guided study, we use a very large suite of synthetic climate scenarios in a statistical modeling framework to simultaneously evaluate how (1) average temperature and precipitation changes, (2) initial basin conditions, and (3) temporal characteristics of the input climate data influence water-year flow in the Upper Colorado River. The results here suggest that existing studies may underestimate the degree of uncertainty in future streamflow, particularly under moderate temperature and precipitation changes. However, we also find that the relative severity of future flow projections within a given climate scenario can be estimated with simple metrics that characterize the input climate data and basin conditions. These results suggest that simple testing, like the analyses presented in this paper, may be helpful in understanding differences between existing studies or in identifying specific conditions for physically based mechanistic modeling. Both options could reduce overall cost and improve the efficiency of conducting climate change impacts studies.