WRAL WeatherCenter Blog

I have heard that meteorologists are able to predict what the weather will be for an entire month, but only provide a weeks worth for viewers. Is this true?

Posted August 5, 2007 12:53 p.m. EDT

MIKE MOSS SAYS:       Emily,      It depends a bit on how you define "predict." Computer models of atmospheric behavior could certainly be run out for a month in advance, although typically about the longest operational models are currently run is out to 16 days or so, with statistical models that refine the raw output from those in terms of high and low temperatures and precipitation probabilities running out to about eight days. The question that arises is at what point do the predictions become so unreliable as to be no better than (a) assuming the weather stays the same as it is, (b) assuming the weather will match the long term climatalogical average for the period in question, or (c) guessing. There isn't a cut and dried answer to that question, but on average somewhere around the 5-7 day time frame the forecasts become so subject to changing from one day to the next that it probably isn't of much value to use them for specific planning, especially with regard to cloud cover and precipitation. Some studies of "predictability" have suggested that even with a greatly enhanced observation system that samples, with high accuracy, the atmosphere/ocean/earth system at a much higher spatial and temporal resolution than we can achieve now, and with essentially unlimited computational power, about two weeks may be the limit for any kind of specific weather forecast that would consistently outperform a-c above.