Atmosphere Academy: Weather and Math
More properly referred to as atmospheric numerical predictions, weather models are mathematical equations that try to represent the atmosphere.
The basic premise is simple enough. If you can figure out an equation to represent the changes in speed and movement of a developing storm system, you can get an accurate forecast by solving the equation. In practice, it’s a little more complicated.
The atmosphere is a fluid, which is to say that its always in motion, all over the place. Equations representing the atmosphere can quickly become rather convoluted, involving questions of mass, momentum, and thermal energy. Thank goodness we have computers to help us with the computations, and it’s a job even for them.
Rather than analyzing ever nook and cranny of air all over the globe (which would be impossible), computers instead divide the atmosphere into imaginary grids, where equations can be computed on a manageable scale. This gives us a good estimate of how storms are moving, but it’s far from the whole story. That’s why computer models give us wrong forecasts from time to time, which gets us back to the original question; how can we improve our reliability?
One way is to increase the number of grids, which gives us better resolution and hopefully a more accurate output. Of course, this increases the work load for the computer exponentially. A better, more common sense way to improve accuracy, is to make sure you properly initialize the model.
Before any model is run, it’s fed the latest weather information from weather balloons, dropsondes, and other instruments. After all, the computer has to know where storms are now, before it can figure out where the storms are headed. The more accurate our current observations, the better chance that the forecast will be correct.
And some models handle certain situations better than others. Some models are geared more toward tropical meteorology; others do a better job with land-based systems. So the number 1 best way to improve forecast accuracy is to know which model to use in a given situation, and that comes only from experience.
Story summary image
More: continued here