Content area
Full Text
A witty, oxymoronic proverb from statistician George Box reads as follows: "All models are wrong; some models are useful." When I first heard it, the quoter-my husband-used it to refer to his work as a geneticist. The very concept of races among humans, he was explaining to me, was exasperatingly, annoyingly imperfect-and yet grudgingly utilitarian.
Surely, every branch of science would like to claim the Box tautology as most fitting to its own field. So without apology, I'll not resist the urge to steal it from genetics (as my husband stole it from statistics) and declare it most fitting to forecasting, however unoriginal my thievery may seem. The very word models so aptly calls to mind those models forecasters use, half use, and often discard, in the preparation of a forecast. Indeed, such is the mark of a skilled forecaster: the admission that all the models are wrong.
Such is also true of meteorologists themselves-those who forecast, those who broadcast, those who teach, and those who write papers intelligible to but a select few. That is, like the Eta or the Global Forecast System (GFS), the models for people are wrong. And, like the Eta or the GFS, they are nonetheless useful. Anyone who has ever taken the Myers-Briggs personality test knows the catharsis of learning he is an "INTJ" or an "ESFP." If he is wise, he also knows where he strays from his model and how not to let this sliver of truth become a self-fulfilling prophecy. Still, there is much to be gained from shrewd pattern recognition, in the science and in the business of weather. (A forecaster once told me that when he blurred his eyes intentionally, he could discern hook echoes where his fellows often missed them.) Patterns, like models, work when we allow the edges to soften.
Perhaps I've been in a unique position to have worked at two companies where broadcast meteorologists and operational meteorologists labor together daily and in close proximity. Perhaps the schism between the two is more glaring when we're all stuck in the same building. I've seen the latter accuse the former of a whole host of meteorological transgressions, such as relying too heavily on quantitative precipitation forecasts and National Weather...