10 December 2010

Why It Never Seems to Rain on Weather Forecasters...

From xkcd this past Wednesday, 8 December, 2010:

A great view on the phenomenon from a fellow geek. I can testify that this seems to happen all too often...especially when it's a person with any academic background in meteorology watching the weather radar, or wherever students and/or professionals in meteorological sciences gather in any number for classes, conferences, parties, whatever. At Colorado State, our cartoonist would have earned (dubious) honor as "Weather Weenie of the Week," which was rarely awarded outside of the North Atlantic hurricane season...

Unofficially, it's called a "weather hole," and a couple of friends from my time in the Atmospheric Sciences graduate program at Colorado State University spent some time studying the conventional wisdom.  Their article, published in 2005 in the Bulletin of the American Meteorological Society, was titled "Do Meteorologists Suppress Thunderstorms? Radar-Derived Statistics and the Behavior of Moist Convection." Their abstract:
Meteorologists and other weather enthusiasts sometimes lament that they live in weather holes—places that receive less exciting weather than do their surroundings. This belief seems to stem from countless hours spent gazing at thunderstorms on displays of radar reflectivity. To test objectively whether radar observations truly bear out this belief, the authors analyzed 6 yr of composite reflectivity from the Weather Surveillance Radar-1988 Doppler (WSR-88D) network. Statistics for 28 target cities, selected for their prominent meteorological communities, are compared with statistics for random points in the conterminous United States to see whether any of the targets is truly a weather hole or, perhaps, a hot spot—the counterpart to a hole. Holes and hot spots are defined by the frequency of convective echoes at a target relative to echoes in the surrounding region, and by the probability that convective echoes near a target were followed shortly by a convective echo at that target.

The data do, indeed, reveal mesoscale variability in occurrences of thunderstorms, as well as distinct signatures of storms' motion and the footprints of stormy regions at each target. However, although the data support the basic concept of convective weather holes and hot spots, only one of the meteorological targets fully met the authors' criteria for a hole and only one fully met their criteria for a hot spot. During the 6 yr studied, nearly all of the selected targets experienced convective storms about as often as their immediate surroundings did.

These results suggest that meteorologists are unnecessarily cranky about the frequency of storms in their hometowns. Meteorologists' belief that they live in weather holes may reveal the need to explore more deeply the statistical behavior of moist convection. The authors comment on some of the strengths and weaknesses of using composite reflectivity alone for that exploration and for determining weather holes and hot spots. Finally, the authors speculate that, with the proper quality control, statistics might serve in the near future as very powerful tools for probabilistic forecast guidance.
The full article (pdf) is available to everyone, even if you're not an AMS member or journal subscriber. Yes, we scientists who remain responsible keep working, following up on every idea and questioning every assumption, even if it means our cherished illusions end up lying on the floor, shattered to a zillion pieces, each offering its own little reflection of a sad, sad reality... [sniff]

No comments: