Forecasts are produced by assuming normal weather (precipitation, temperature) in the
future. Forecast runoff is modeled using the current antecedent conditions and
future scenarios. The further the departure from normal weather, the larger the change.
Monthly updates usually result in more accurate forecasts as the year progresses since the period
of known data increases and the period of unknown (forecasted) data decreases.
In dealing with future scenarios, assumptions are made to how/when weather patterns occur.
If the weather does not follow its "normal" pattern, forecasts can change dramatically.
For example, a major storm will create much greater runoff than normal.
Similarily, a major hot spell will create much-above-average runoff from snowmelt.
The forecast based on normal weather is the probable forecast. Other
forecasts are produced using lower quartile, upper quartile, and lower decile (10% chance
of precipitation, being lower) precipitation. These forecasts provide a reasonable
range of (lower to upper quartile) volume and a reasonable minimum (lower decile).
|