DIMINISHED SNOWMELT RUNOFF IN WESTERN RIVERS

Temperature has risen and snowpack has diminished over the past half-century across western North America, trends that are likely to continue and perhaps accelerate in coming decades. Predicting the effect of these trends on snowmelt runoff in major rivers has huge implications for water management across the semi-arid West. Many recent studies, including our group's analysis of changes in climate and streamflow in the Rio Grande headwaters, illustrate how predictability of surface water supplies has been reduced as the correlation between snowpack and subsequent surface runoff diminishes in a warmer climate. We find that spring season precipitation, which is less predictable than winter snowpack, accounts for an increasing fraction of runoff as snowpack declines. Furthermore, although diminished snowpack is generally expected to lead to reduced streamflow, quantifying long-term trends in 21st Century streamflow is extremely difficult. We find that much of the spread in climate model-based projections of Rio Grande streamflow in coming decades is related to projections of spring precipitation, which vary considerably between model simulations. Improvement in prediction skill for streamflow in major western rivers, both for interannual fluctuations and for long-term projections, should be realizable and would be of tremendous societal importance. Such improvements will require additional interdisciplinary research on climate prediction, characterization of snowpack and its surface radiative properties, surface-groundwater interaction in mountainous terrain, and better incorporation of complex topography and changing vegetation (including fire-disturbed forests) into streamflow prediction models.