New Tool Provides Improved Seasonal Precipitation and Snowpack Forecasting for California Reservoirs It's February 1, your reservoir is at the top of conservation storage and snow water in the upper watershed is only 30% of normal. Do you start aggressively managing flood storage to hedge against a dry late winter and spring or continue operating as usual at the risk of summer water supply? Historic data aren't much help; in many Sierra locations, a low February snowpack leads just as often to normal April conditions as to critically dry ones. It sure would be nice if you could look out across the Pacific and see what might be in store for the next couple of months…and as it turns out, you can.

Working for the NASA Ames Research Center and the Los Angeles Department of Water and Power (LADWP), NHC and TetraTech scientists identified a mathematical relationship between California precipitation and the atmospheric circulation feature known as the North Pacific High (NPH). We have now developed a statistical model that successfully predicts April 1 snow water content, based on relative January strength of the NPH and February 1 snow water content, in the Eastern Sierra.

The model was calibrated and tested for four Owens Valley snow stations used by LADWP for reservoir operations planning. Over 66 years of hindcasting (1951-2016) and two years of forecasting (2017-2018), the model successfully predicted the relative April snow water level (i.e. dry, normal, or wet) from the January NPH location and February 1 snow surveys 75% of the time—with zero "opposite” forecasts (i.e. predicted wet/observed dry and vice versa)—for the principal station at Mammoth Pass. In addition to predicting values April 1 for snow water content, the model also calculates a probability distribution and confidence intervals, to provide a sense of the level of uncertainty of the prediction.

The observed April 1 snow water content (SWC) is plotted against the model forecasts at Mammoth Pass in the eastern Sierra Nevada. Each gray point represents a year in the hindcast period (1951-2016) used to develop and test the model. The red points represent the 2017 and 2018 forecasts, demonstrating successful predictions.

Based on applications to date, this statistical forecasting model offers significant potential to provide vitally needed improvements in seasonal forecasting to help water managers make operational decisions. NHC is currently seeking water managers or other interested partners who would like to develop precipitation forecasting models for other California watersheds.

For further information, contact Brady McDaniel or Patty Dillon.

Study methodology and results for the NASA Ames and LADWP applications are documented in two peer-reviewed journal articles:

Rath JS, Costa-Cabral M (2018, in press): A snowpack forecasting model for the eastern Sierra Nevada based on cointegration with the North Pacific High Sea-level pressure anomaly. Frontiers. Manuscript in press. Link to article.

Costa-Cabral M, Rath JS, Mills WB, Roy SB, Bromirski PD, Milesi C (2016): Projecting and forecasting winter precipitation extremes and drought in California using the North Pacific High sea-level pressure anomaly, J Climate. DOI: 10.1175/JCLI-D-15-0525. Link to article.