Effects of ENSO on California Precipitation
Maurice RoosCalifornia Department of Water Resources
Sacramento, CA. USA
mroos@water.ca.gov
Background
First some background on what we do. My group's primary activity is water supply forecasting, including snowmelt runoff, and (jointly with the NWS California-Nevada River Forecast Center) flood forecasting on the major rivers of northern California. The former is a longer range forecast about six months into the future (or to the end of the water year on September 30) which could employ longer ranger weather forecasts if sufficient skill was demonstrated. The precipitation projection used for flood forecasting is much shorter (in hours up to a couple of days) and probably not a good target for eastern tropical Pacific warm or cool event forecasts. However, there may be some definitive statements on the change in probabilities of flood events during a season which can be developed from El Niño/La Niña conditions which could be useful in emergency planning.
Water supply forecasts are used by water project operators to schedule reservoir operations and water system deliveries. In general, reservoir management is very conservative so as to avoid shortage, that is, to avoid reductions in planned deliveries after the delivery season has started. Changes in delivery to State Water Project water contractors will occur only if the 99 percent probable amount surpasses the 90 percent probability value projected in December. For example, if the conventional December 1, 90 percent probability amount was 11 million acre feet of Sacramento River system runoff, the predicted runoff as of January 1 (or any subsequent date) must exceed this amount at the 99 percent probability to change the allocation.
Goals
My major objective in attending is to learn what information is available on relationships between La Niña and worldwide weather patterns, especially those which might affect California during the winter wet season. There are indications from past such events of dry winters in southern California and Arizona, but a much more mixed outlook for northern California. In fact, some of our worst floods have come in La Niña years. We wonder whether the major river flood threat is sufficiently high to warrant special preparations over and above those made for every winter. Would such a forecast be convincing enough to ask our Legislators and Governor for some predeployment funds for flood-fighting materials and activities as was done last year, based on the El Niño related forecasts?
From a water-operation study point, there are two thresholds of long-range forecasting which would be useful. The first is a reliable forecast of a wet or dry season early in the water year. (The water year is the 12-month period which begins October 1 and extends through September 30). Generally, not much precipitation occurs after mid-April, so the seasonal forecast in November or on December 1 only needs to extend about 5 months. Such a forecast would still be quite valuable on January 1, at about the 35 percent mark in the accumulation season. By early February many of the major operational planting decisions are made, with mostly small adjustments being made as a result of later weather events. At that point, the major benefit would shift to a reliable forecast of the next wet season (a 15-month forecast) which would affect the amount of reservoir carryover to save for a possible dry winter next season. This is especially true if the current year is dry and users need to weigh whether to use most of their stored water in the current season or to take some losses to ensure better supply during the following year.