top border

Please note, not all links may be active. This site is a snapshot of an earlier time.

For Today, Partial Clearing a Good Bet
Seven Days From Now, All Bets are Off

Andrew Freedman @ 7:40 AM

A coworker of mine whom I consider to be a typical weather information consumer has an unhealthy amount of pent-up anger at Chicago weatherman Tom Skilling. He says Skilling, who is the chief meteorologist for WGN TV and widely regarded as one of the best TV meteorologists in the business, takes far too long to get to the seven-day forecast.

Dan Stillman's Forecast

Forecast Confidence: Medium-HighToday: The front that brought overnight rain will slowly make its way through the area. Thus, despite morning sunshine, we may see the return of off-and-on clouds and an occasional shower (30 percent chance) through early afternoon. The chance of rain decreases significantly after that. High temperatures will be in the upper 60s. (Skinscast and Natcast can be seen at the bottom of this post.) Tonight: Clearing skies and cool with lows in the low 50s in town, mid-to-upper 40s in the burbs. Tomorrow: High pressure takes control, which means a beautiful, mostly sunny day with high temps climbing to the low-to-mid 70s.

Check back tomorrow for Jason's week-ahead forecast.

"[Skilling] talks for 40 minutes about weather systems in Nebraska before giving me the seven-day," my coworker said. "You know, a lot of people feel the same way I do, about him," he stated over inter-office instant messenger, as if he was planning to run against him in the next weathercaster of the year election. (There is no such election, but if there is, my money's on Skilling).

Usually when confronted with such anti-weatherman rhetoric I'll launch into a passionate defense of the necessities of showing the 500 mb chart or the national dewpoint analysis, only to fail to bridge the chasm that exists between the weather junkie TV watcher and the everyday viewing public. The same thing happened to me in D.C. when I debated the merits of each weathercaster there, discussing, say, a Tony Pann or a Bob Ryan. This debate brings up a larger issue, however. It highlights the goal of most viewers when they watch the TV weathercast and begs the question: Should a TV weathercast really be geared towards giving the seven-day forecast? Why is that the ultimate goal?

Clearly the seven-day is what most people are watching for, otherwise they would put it at the beginning of the weathercast. But why is this? And how accurate are these forecasts? Shouldn't it be more important to present an overall picture of the weather and its immediate effects on people? Where has this fervent drive toward an extended forecast come from? I doubt the pressure comes from within the weather room (interesting side note, many TV weather department offices don't have windows) from TV mets eager to show off extended forecasting skills. I believe many TV meteorologists would rather not have it at all, since busting forecasts seven days out ends up reducing their credibility in the long run.

Maybe I'm old fashioned, but 10-day forecasts make me uncomfortable. They're useful for trendspotting, rather than as accurate pinpoint measurements, yet I worry that viewers take them at face value, and that's the weathercaster's fault.

I used to watch a local TV meteorologist in Boston, whom I've long regarded as an excellent forecaster, curse at the task of projecting temperatures five days in advance. He'd complain about how unreliable such outlooks are, and how viewers don't realize how poor a forecast they're getting. Granted, this was a few years ago, but I don't believe extended-range forecasting has advanced leaps and bounds since then (Josh Larson, please correct me if I'm wrong on this).

According to ForecastAdvisor, a web site started by Eric Floehr, a computer consultant in Marysville, Ohio, that makes comparisons of major weather forecasting firms' accuracy available to the public, a nationwide survey of one- to nine-day forecasts in 2005 showed that high temperature forecasts were about 39 percent accurate at 10 days out. Low temperature forecasts were about 37 percent accurate for the same time period. This is better than purely predicting the weather based on climatology alone (ForecastAdvisor says using climatology yields an accuracy of 33 percent), but it is not a dramatic improvement.

I think each TV station should consider putting a disclaimer on their extended range forecasts similar to what WZZM 13 does in Grand Rapids Michigan. Their extended forecast carries the following qualifier: "WZZM 13 On-Target meteorologists... pride themselves on their accuracy. They work hard at researching information to provide you with the best forecast possible. The 7 day forecast is provided to you as a "planner" and of course may change as quickly as the weather itself. Please revisit our website often for updates and be sure to watch WZZM 13 News at 5:00am, Noon, Take Five, 5:30pm, 6pm and 11pm for the latest 13 On Target Weather information."

That's a lot better than the typical "Here's your seven-day, now over to sports" that many stations routinely resort to. And maybe there lies the difference. Perhaps the people in Grand Rapids are being better served than the people in Washington or New York or Boston or even Chicago, despite Skilling's popularity. The key to presenting long-range forecasts is to also include some idea as to the confidence in those predictions and the accuracy, emphasizing the fluid nature of the atmosphere and the computer models that help predict the fluctuations of the atmosphere. Simply stating "here's the seven-day" and leaving it at that doesn't cut it in my book, because it leads people to angrily denounce the weatherman five to six days later when it rains on their golf game. Or whatever. Not everyone plays golf. Perhaps a different analogy is better. And yes, maybe pacifying the viewership is necessary, and the seven-day should be shown at the top of the forecast, and the rest of the broadcast can be solving the mystery of how that forecast was arrived at.

Maybe then my friend at work will stop whining.

Mets at Nationals
October 1, 2006 -- 7:05pm -- RFK Stadium
First Pitch9th InningWeatherChance of Rain
A shower or two is possible, but probably nothing that could cancel the game. Even a delay is unlikely.

Natcast bids adieu for the offseason. See ya next April!

Comments are closed for this archived entry | Link | email post Email this post