HR: 1340h
AN: H33D-1408    [Abstracts]
TI: Limitations of Snow-Water Equivalent and Precipitation Data in Seasonal Flow Forecasting in the Upper Klamath River Basin of Oregon and California
AU: * Risley, J C
EM: jrisley@usgs.gov
AF: U.S. Geological Survey, Oregon Water Science Center 10615 SE Cherry Blossom Drive, Portland, OR 97216 United States
AU: Roehl, E A
EM: ed.roehl@advdatamining.com
AF: Advanced Data Mining Services, LLC, 2123 Old Spartanburg Road PMB 104, Greer, SC 29650-2704 United States
AB: Water managers in the upper Klamath Basin, located in south-central Oregon and northeastern California, rely on accurate forecasts of spring and summer streamflow to optimally allocate increasingly limited water supplies for various demands that include irrigation for agriculture, habitat for endangered fishes, and hydropower production. Federal agencies make forecasts on the 1st of each month, from January through May, of the total volume of water expected to pass a stream gage or flow into a reservoir during an entire summer irrigation season. Often the forecasts are based on output from flow forecast models that use real-time snow-water equivalent (SWE) and precipitation data as input. Because the January and February forecasts are significantly less accurate than those made in the spring, we were interested in quantifying the limitations of real-time SWE and precipitation data in forecasting future flows. Using over 20 years of daily SWE, precipitation, and flow time-series from five sites in the upper Klamath Basin, we first decomposed the flow records into annual periodic, long-term climatic, and chaotic traces. Being the component of the original flow records with their seasonality and long-term trends removed, the chaotic traces were then lag correlated with SWE and precipitation records. After a 120 day lag (approximately 4 months), all of the correlation coefficients between the chaotic flow traces and the SWE and precipitation records were less than 0.4. From this, we could infer that SWE and precipitation data older than 120 days provide forecast models only with information regarding the annual seasonal and long-term climate patterns and do not provide information that is unique and specific for the upcoming irrigation season. These results also support the need to find new climate variables, such as mid-oceanic indicators, to improve forecast model accuracy rather than using just real-time SWE and precipitation conditions.
DE: 1816 Estimation and forecasting
DE: 1833 Hydroclimatology
DE: 1860 Streamflow
DE: 1872 Time series analysis (3270, 4277, 4475)
SC: Hydrology [H]
MN: Fall Meeting 2005