How forecasts are made

Most of the annual streamflow in Arizona originates as snowfall that has accumulated in the mountains during the winter and early spring. As the snowpack accumulates, hydrologists estimate the runoff that will occur when it melts.

Measurements of snow water equivalent at selected manual snow courses and automated Snow Telemetry (SNOTEL) sites, along with precipitation and streamflow values, are used in statistical and simulation models to prepare runoff forecasts. These forecasts are coordinated between hydrologists in the Natural Resources Conservation Service (NRCS) the National Weather Service, and the Salt River Project.

Forecasts of any kind are not perfect. Streamflow forecast uncertainty arises from three primary sources: (1) uncertainty of future weather conditions, (2) uncertainty in the forecasting procedure, and (3) errors in the data. The forecast, therefore, must be interpreted not as a single value but rather as a range of values with specific probabilities of occurrence. The middle of the range is expressed by the 50% exceedance probability forecast, for which there is a 50% chance that the actual flow will be above, and a 50% chance that the actual flow will be below, this value. To describe the expected range around this 50% value, four other forecasts are provided, two smaller values (90% and 70% exceedance probability) and two larger values (30%, and 10% exceedance probability). For example, there is a 90% chance that the actual flow will be more than the 90% exceedance probability forecast.

The wider the spread among these values, the more uncertain the forecast. As the season progresses, forecasts become more accurate, primarily because a greater portion of the future weather conditions become known. This is reflected by a narrowing of the range around the 50% exceedance probability forecast. Users should take this uncertainty into consideration when making operational decisions by selecting forecasts corresponding to the level of risk they are willing to assume about the amount of water to be expected. If users anticipate receiving a lesser supply of water, or are concerned about having an adequate water supply, they may want to base their decisions on the 90% or 70% exceedance probability forecasts. On the other hand, if users anticipate receiving too much water, or are concerned about the threat of flooding, they may want to base their decisions on the 30% or 10% exceedance probability forecasts. Regardless of the forecast value users choose, they should be prepared to deal with either more or less water.

SUMMARY
As of February 15, snowpack levels are below normal to normal throughout the major basins of the state. Precipitation for the first half of February was well below normal to above normal in the major river basins. The Salt and Verde River reservoir system stands at 76 percent of capacity, while San Carlos Reservoir is at 10 percent of capacity. The mid-month forecast calls for well below normal to normal runoff in all basins for the spring runoff period.

SNOWPACK
Snow water equivalent levels in the state’s major river basins are below normal to normal, ranging from 91 percent of median in the Gila River Basin, to 76 percent of median in the Verde River Basin. The statewide snowpack is below normal at 78 percent of median.

SALT RIVER BASIN as of February 15, 2020
Below normal streamflow levels are forecast for the basin. In the Salt River, near Roosevelt, the forecast calls for 85% of median streamflow through May, while at Tonto Creek, the forecast calls for 93% of median streamflow through May. Snow survey measurements show the Salt snowpack to be at 81% of median.   back...