Development of Streamflow Projections under Changing Climate Conditions over Colorado River Basin Headwaters

By W. Paul Miller, Thomas Piechota, Subhrendu Gangopadhyay, Tommy Pruitt, Hydrology and Earth System Sciences, published 2011
 

Abstract: The current drought over the Colorado River Basin has raised concerns that the US Department of the Interior, Bureau of Reclamation (Reclamation) may impose water shortages over the lower portion of the basin for the first time in history. The guidelines that determine levels of shortage are affected by relatively short-term (3 to 7 month) forecasts determined by the Colorado Basin River Forecast Center (CBRFC) using the National Weather Service (NWS) River Forecasting System (RFS) hydrologic model. While these forecasts by the CBRFC are useful, water managers within the basin are interested in long-term projections of streamflow, particularly under changing climate conditions. In this study, a bias-corrected, statistically downscaled dataset of projected climate is used to force the NWS RFS utilized by the CBRFC to derive projections of streamflow over the Green, Gunnison, and San Juan River headwater basins located within the Colorado River Basin. This study evaluates the impact of changing climate to evapotranspiration rates and contributes to a better understanding of how hydrologic processes change under varying climate conditions. The impact to evapotranspiration rates is taken into consideration and incorporated into the development of streamflow projections over Colorado River headwater basins in this study. Additionally, the NWS RFS is modified to account for impacts to evapotranspiration due to changing temperature over the basin. Adjusting evapotranspiration demands resulted in a 6% to 13% average decrease in runoff over the Gunnison River Basin when compared to static evapotranspiration rates. Streamflow projections derived using projections of future climate and the NWS RFS provided by the CBRFC resulted in decreased runoff in 2 of the 3 basins considered. Over the Gunnison and San Juan River basins, a 10% to 15% average decrease in basin runoff is projected through the year 2099. However, over the Green River basin, a 5% to 8% increase in basin runoff is projected through 2099. Evidence of nonstationary behavior is apparent over the Gunnison and San Juan River basins.

A Non-Parametric Approach for Paleo Reconstruction of Annual Streamflow Ensembles

By Subhrendu Gangopadhyay, Benjamin L. Harding, Balaji Rajagopalan, Terrance J. Fulp, Water Resource Research, submitted for review, 2008
 

Abstract: A non parametric method for paleo reconstruction of streamflows using tree ring chronologies is developed. In this method, first a Principal Component Analysis (PCA) is performed on the chronologies of the overlap period (i.e., the contemporary period for which both the tree ring and streamflow data are available) and the leading Principal Components (PCs) that capture up to 90% of the variance are retained. For any year’s chronology K-nearest neighbors (K-nn) are identified from the observed period in the PC space, which results in K observed years. The corresponding streamflow in the K years are potential ensemble members. This set of K streamflows are resampled using a weight function that gives most weight to the nearest neighbor and least to the farthest, thus providing an ensemble. Also, a weighted mean of the K streamflows can be computed to provide a mean value. This method is applied to the naturalized streamflow (1906-2005) at the important Lees Ferry gauge on the Colorado River and the tree ring chronologies (1400-1905). The cross-validated streamflow reconstructions for the contemporary period compares very well with the observed flows and also with other published streamflow reconstructions. The proposed non-parametric method provides an ensemble of streamflows for each year in the past and consequently a realistic confidence interval. In addition, the method can be used to reconstruct structured and even non-numerical data for use in water resources modeling.

Regional Analysis of Trend and Step Changes Observed in Hydroclimatic Variables around the Colorado River Basin

By W. Paul Miller and Thomas Piechota, Journal of Hydrometeorology, 9(5), 2008
 

Abstract: Recent research has suggested that changes in temperature and precipitation events due to climate change have had a significant impact on the availability and timing of streamflow. In this study, monthly temperature and precipitation data collected over 29 climate divisions covering the entire Colorado River basin and monthly natural flow data from 29 U.S. Geological Survey (USGS) gauge locations along the Colorado River are investigated for trend or step changes using parametric and nonparametric statistical tests. Temperature increases are persistent (at least 10 climate divisions over 6 months in trend analysis) throughout the year over the Colorado River basin, whereas precipitation only notably increased over 17 climate divisions (during trend analysis) during February and remained relatively unchanged otherwise. These results correspond with changes in naturalized streamflow throughout the year. Streamflow increases are recorded between November and February but exhibit a decreasing trend over the traditional peak runoff season (April through July). Under trend analysis, 18 flow stations exhibited increasing trends in January and 19 flow stations exhibited decreasing trends in June. It is likely that increasing temperature trends have affected the character of precipitation in the Colorado River basin, causing a change in the timing of runoff events.

For more information relating to these publications, please contact CRBclimateresearch@usbr.gov.