Projections of Increasingly Earlier Snowmelt Runoff in the Western United States under a "Business as Usual" Climate Change Scenario

The timing of the snowmelt runoff pulse, the most important contribution to the annual flow for many gauges in western North America, is critical for water supply management and represents a seasonally integrated signal of climate change. A trend in this snowmelt runoff pulse towards earlier in the water year has already been observed for historic runoff records in the 1948-2000 period. This study examined the relationship between the temperature and precipitation changes predicted under a "Business as Usual" (BAU) climate change scenario for the 1995-2099 period and the projected future timing of the spring runoff pulse for a network of snowmelt-dominated gauges across western North America. Changes in streamflow timing, as compared to a 1951-80 baseline period, were projected with a simple and multiple regression analysis using temperature (TI) and precipitation (PI) indices. The predicted TI and PI were calculated from a downscaled PCM BAU ensemble member. The changes streamflow timing predicted by the regression model were compared with output from physically-based models for selected basin. The changes in temperature and precipitation obtained from BAU ACPI simulations show a continued warming trend in TI through the year 2099. The increasingly positive TI anomalies are mainly responsible for a continued shift towards an earlier timing of the snowmelt runoff. This hydroclimatic shift most strongly affects the Pacific Northwest, Sierra Nevada, and Rocky Mountain regions, where it reaches 20-40 days for many gauges. By contrast, the only modest changes predicted in PI do not cause any changes in CT. The TIPI regressions are completely dominated by TI. Whether the response in CT to the temperature trend will level out or increase at a certain threshold cannot be gleaned from a simple regression model, however, changes predicted by the regression model agree well with those produced by physically-based hydrologic simulations of selected rivers in the Pacific Northwest and the Sierra Nevada (California).