NOAA Advises California Water Managers on Improving QPF for Folsom Dam Operation

Contact: Marty Ralph

On June 6, 2003, NOAA/OAR's Environmental Technology Laboratory (ETL) and NOAA/NWS contributed to the multi-disciplinary Lower American River Science Conference held in Sacramento, California. The California Weather Symposium, which was conducted as part of the conference, consisted of invited speakers who directed their presentations toward the theme of "Forecasting Extreme Precipitation in the Sierra Nevada and Implications for the American River Watershed." The presentations were followed by a panel discussion on the topic. A total of roughly 75 participants, including water managers, hydrologists, flood control managers, meteorologists, and media, discussed many facets of the flood threat to Sacramento. The need for improved forecasts, and the scientific and practical approaches that could lead to improvements, were described by speakers from several organizations, including two speakers from NOAA. The meeting highlights the critical role NOAA can play in addressing one of the greatest flood risks in the nation, and the value of working closely with those responsible for acting upon forecasts of extreme events. A key topic of discussion was the potential impact of a test bed approach, which would bring together the diverse set of people and efforts related to this problem in a way that generates breakthroughs that might not otherwise occur.

The invited speakers and the titles of their presentations were as follows:

1. Tim Washburn, General Counsel, Sacramento Area Flood Control Agency: Connecting the Dots: Fish, Groundwater and Weather.
2. Maurice Roos, Chief Hydrologist, California Dept. of Water Resources: Extreme Precipitation in the American River Basin
3. Dr. Beth Faber, Research Hydraulic Engineer, Hydrologic Engineering Center (HEC), Army Corps of Engineers: The Potential for Adaptive Reservoir Operations Provided by Forecast Information
4. Paul Pugner, Chief, Water Management, Sacramento District, Army Corps of Engineers: Spring Forecast Based Operations, Folsom Dam, California.
5. Dr. Dan Cayan: Climate Research Division, SCRIPPS: Precipitation Structure in the Sierra Nevada of California in Winter.
6. Wes Junker*, Senior Forecaster NOAA/NWS/Hydrometeorological Prediction Center (HPC) What We Know About Northern California Extreme Rainfall Events: An HPC Perspective.
7. Theresa Carpenter, Research Hydrologist, Hydrologic Research Center, San Diego, CA: Characterizing Precipitation Forecast Errors for the Folsom Lake Watershed
8. Dr. Marty Ralph*, Chief, NOAA/ETL/Regional Weather and Climate Applications Division Extreme Precipitation on the American River Watershed: Strategies for Improving Forecasts

*Panelists, with Owen Rhea (formerly of the NWS/California Nevada River Forecast Center)

The meeting covered the origins and problems with current flood control and water supply capabilities on the Lower American River, including Folsom Dam with its 977,000 acre-foot (AF) storage capacity (of which between 400,000 and 670,000 AF are kept empty during the winter flood season). Experts described the need for, and challenges of, developing a "Forecast-based operations" plan for Folsom Dam. The maximum 3-day estimated unimpaired runoff from the 1.2 M acres (1861 mi2) above the dam was nearly 1,000,000 AF in both the February 1986 (976,000 AF) and January 1997 (988,000 AF) floods. Maximum mean 3-day inflow rates of 166,000 ft3 s-1 (CFS) were observed in the 1986 flood, and 164,300 CFS in the 1997 flood. These are well above the dam's maximum controlled discharge rate of 115,000 CFS (= 9,504 AF/hour), which would pose a serious problem if the reservoir had already filled. Since the dam was completed in 1955, these two floods plus two others have exceeded the dam's original design flood volume of 978,000 AF over 6 days. The 1986 and 1997 floods exceeded the 6-day volume in just 3 days. These recent larger floods raise doubts about the ability of the dam to control even greater floods, thus potentially requiring uncontrolled releases that could cause downstream levees to fail, which would lead to a multi-billion dollar disaster. A formal ACE report in 2002 estimated that a 200-year flood would cause $10 B in damage to Sacramento. Due to the many competing demands on operating a multipurpose dam (e.g., power generation, water supply, flood control, fisheries health), it has not been possible in the past to make decisions on flood releases based on weather forecasts. Observed stream flow and rain gauge data are the currently acceptable bases for making releases in floods. However, it has been recognized that improved quantitative precipitation forecasts (QPF) of extreme rainfall events and ensuing runoff, including estimates of forecast uncertainty, provide the opportunity to change this.

A proposal from the various agencies involved has been made that would allow as much as 30,000 additional acre feet of water to be stored during spring months (near the end of the flood season) for water supply, fisheries preservation and lake recreationist needs, as long as 24-48 h QPF could be used to release this water before a storm strikes. This water would give dam operators more flexibility in meeting the competing demands on the reservoir. In particular, it could improve flow and water temperature conditions in the lower river so as to increase the numbers of migratory fall-run salmon and steelhead occupying the river. Steelhead are listed as a threatened species. Fall-run salmon are being considered for listing. Increasing the numbers of these fish in Central Valley rivers and streams is essential to continuing water supply and hydropower operations.

Wes Junker, Senior Forecaster of NWS/NCEP/HPC, summarized the QPF skill to date (threat score of 0.2 for 5 inch 24 h QPF in CA; scores are higher for 1 or 2 inch totals in strong events), which is still poor, but is also much better than elsewhere in the country because the predictability of cool-season orographic rainfall is greater than most non-tropical extreme summer convective rainfall. Although it appears that current skill is still too low to be effective in dam operations, he pointed out that the potential for improvements are great in this area due to the orography and due to possible use of 1-2 inch thresholds in events with 5 inch QPF. Dr. Marty Ralph, of NOAA/OAR/ETL, then summarized four candidate methods for improving 0-48 h QPF in the region based on NOAA's experimental work in the area over the last few years. The approaches included assessing and improving current operational orographic rainfall algorithms and forecast techniques, improved 24-48 h numerical weather prediction through mesoscale targeted dropsondes, and better nowcasting through monitoring of snow level and better radar-derived rainfall estimates.