Abstract

This article applies formal detection and attribution techniques to investigate the nature of observed shifts in the timing of streamflow in the western United States (US). Previous studies have shown that the snow hydrology of the western US has changed in the second half of the 20th century. Such changes manifest themselves in the form of more rain and less snow, in reductions in the snow water contents and in earlier snowmelt and associated advances in streamflow “center” timing (the day in the “water-year” on average when half the water-year flow at a point has passed). However, with one exception over a more limited domain, no other study has attempted to formally attribute these changes to anthropogenic increases of greenhouse gases in the atmosphere.

Using the observations together with a set of global climate model (GCM) simulations and a hydrologic model (applied to three major hydrological regions of the western US – the California region, the Upper Colorado River basin and the Columbia River basin), we find that the observed trends toward earlier “center” timing of snowmelt-driven streamflows in the western US since 1950 are detectably different from natural variability (significant at the p<0.05 level). Furthermore, the non-natural parts of these changes can be attributed confidently to climate changes induced by anthropogenic greenhouse gases, aerosols, ozone, and land-use. The signal from the Columbia dominates the analysis, and it is the only basin that showed detectable signal when the analysis was performed on individual basins. It should be noted that although climate change is an important signal, other climatic processes have also contributed to the hydrologic variability of large basins in the western US.