Hobbies And Interests
Input
NOAA̵7;s weather forecasts begin with an extremely large amount of weather data. Weather satellites covering the globe provide real-time information about cloud cover, temperature, and precipitation data, while local radar stations give a more finely grained picture of atmospheric conditions. A network of weather stations on the ground provide the most detailed information, including wind speeds, pressure, and other variables that can provide information on pressure systems and frontal boundaries. Combining all of this information provides a snapshot of current conditions, which provides the basis for computer modeling of the weather.
Computer Modeling
Weather forecasters feed this atmospheric data into powerful computers that use complex mathematical equations to simulate the behavior of air masses in the atmosphere. If the current data shows a high pressure system next to an area of low pressure, for instance, past observances and known physical laws suggest that the frontal boundary will move in a specific direction at a specific speed. The computer returns a forecast of what the atmospheric conditions will be like a few hours from now, and since that forecast comes from known data with relatively few unknown variables, it usually proves accurate. Then the meteorologists run the simulations again, this time using the predicted data, to generate a forecast a few more hours into the future. Every new simulation relies more and more on predicted data rather than observed data, and therefore becomes more open to error.
Long Range Forecasting
Generally, once these simulations reach the seven- to 10-day mark, the probability of accuracy is low enough that predicting anything specific becomes impossible. Beyond this limit, forecasters combine the last, best guesses of the computers with historical data and expected climate effects from large, persistent systems like El Nino or La Nina to provide the general conditions expected on a given date, usually described as a percentage chances of "above normal" or "below normal" temperature and precipitation. NOAA may occasionally issue longer-term forecasts about things like hurricane activity or out-of-the-ordinary precipitation levels in a region, but these predictions are generally unreliable at best.
The Future
As computers become more powerful, and meteorologists develop new models to explain and predict the behavior of the atmosphere, more accurate long range weather prediction may become possible. One technique that offers some advantages is the use of multiple mathematical models, running several simulations at once and aggregating the results into a composite forecast. However, accurate prediction of weather events beyond 10 days into the future is likely to remain a matter of luck for some time to come.