Reconciling Projections of Future Colorado River Stream Flow

Recent studies suggesting substantial changes in runoff may occur over the next century in the Colorado River Basin are of great concern to the region's water managers. The range of estimated impacts on Colorado River discharge have been from slight increases to a 45% decrease by mid century. While this range of projections and associated hypotheses may be of intellectual interest and stimulate scientific debate, to users and decision makers at the federal level, in the seven basin states, and internationally, providing conflicting information on future conditions is a serious impediment to drought and climate change planning.

Given the wide range of projected flows, the four NOAA Regional Integrated Sciences and Assessments in the western US (RISAs: Western Water Assessment, Climate Assessment for the Southwest, Climate Impacts Group of the Pacific Northwest, California Applications Program), Bureau of Reclamation and NOAA engaged in a coping with drought study to reconcile the range of estimates for future Colorado River flows. A secondary goal was to inform the National Integrated Drought Information System (NIDIS) on the process needed to convey to policy and decision makers the nature of the uncertainties associated with projections of future climate impacts.

An intercomparison was initiated to assess how the methodological approaches and models used to generate estimates of Colorado River flow in 2050 reflect actual uncertainty with associated risks, and how much uncertainty is due to differences in the methodological approaches and model biases. We first examined how the various approaches simulated the 20th century record of Colorado River stream flow due to shifts in the average and seasonality of temperature and precipitation, snow pack development and decline, and antecedent soil moisture conditions. Precipitation elasticities and temperature sensitivities of these approaches to simulate stream flow were evaluated by imposing a 10 percent reduction of precipitation and 1°C and 2 °C increases in temperature. We also examined the elevation dependency of runoff contributing to stream flow among the simulations and in comparison to the observed record.

An initial result of the intercomparison is a narrowing of the range of projected Colorado River flows at 2050 to decreases between -6% and -20%. Precipitation elasticity across the approaches for annual flows at Lees Ferry is on the order 2, whereas temperature sensitivity for annual flows at Lees Ferry ranges from -2%/°C to -9%/°C. In addition, differences in gridded time series of precipitation and temperature (Oregon State University PRISM approach versus University of Washington Maurer approach) were found to impact the simulated 20th century flows and sensitivity of future flows to climate change at 2050.

As part of our stakeholder engagement, we are also documenting how these research findings have and are informing policy and decision making in the basin, as well as institutional, social, behavioral issues such as mitigation and adaptation strategies. The results of our intercomparison will be readily applicable to address the needs of water planners elsewhere, especially across parts of the western U.S. and other middle latitude continental settings where water resources are heavily dependent on snowmelt runoff from mountainous headwater areas.