Oh, The Weather Outside Is…

By on October 7, 2004

A post over on Signal Vs. Noise got me thinking about weather forecasting.

At least here in the Midwest (where weather occasionally kills you), the weather forecast is usually a very long affair consisting of discussions on front movements, dew points, barometric pressures, etc. Over time, we’ve all become accustomed to the weather man showing us all these fancy visualizations of the weather data. I know of at least one station here in Sioux Falls where the local news consists of the ‘First Look Weather’, the Top Story, the Current Conditions, another news story, the Weather Forecast, a little more news, and a Weather Wrap-Up. If they’re short on news, they throw in a ‘Weather Almanac’ piece to boot.

Why are the TV weather men trying to turn us into meteorologists? Basic knowledge of meteorology is probably the subject from grade school earth sciences that we use the most in our day to day lives, even though there’s a professional there to sort it all out for us. But why do we watch the weather? Just a few basic reasons, mostly:

  • How should I dress myself and the kids today?
  • Is it going to be nice enough to do something outside?
  • If I start something outside now (mowing, hiking), will it turn crappy before I get done?
  • If there’s something I need to do outside (yard work), should I do it now or another day this week?

So why doesn’t the weather forecast consist of this?

“Wear coats to work tomorrow. Don’t bother bringing sunglasses. If you need to mow, do it tonight, because it’s going to rain tomorrow morning.”

With all of the usability wonks running around on the Internet, you’d think that online forecasts would be better. But most of them are just trying to offer an online version of what’s on TV.

Weatherpixie.com at least takes a different stab at the weather, if only for the current conditions. You can get a little cartoon person standing in what appears to be your current conditions. It seems to put form over function, though. It’s fairly lousy weather here right now, but my little cartoon person has a t-shirt on. If she was my kid, I’d at least tell her to wear sleeves.

Instead of a little app that sits in the system tray and shows a sun or clouds with the current temp, I’d like to see one that looks out over the remainder of the day, makes some decisions, and shows a picture of a sweater, or an umbrella, or shorts. It could have little overlay icons for things like expected precipitation within the next few hours, or if precipitation is expected tomorrow. Or, even simpler, something that indicates when in the next few days is the best time to be outside. That’s all I really need to know. How should I dress, and when should I do my outside stuff?

So, am I just off on a rant here? Am I missing something? Are there other situations where our day-to-day data can be simplified? Let me know.



  1. Have you ever seen the panorama at Dunstan Orchard’s blog (http://www.1976design.com/blog/)? He takes the current weather and uses it to update what the header looks like. He has a detailed description of what he did and how he did it. See the “panorama information” link at the bottom-right of the header.

    The same process could be turned into a predictive tool like you mentioned, I would think.

  2. There is a station here in Des Moines that has a similar news format. Why? Because it works. It informs us, not of what to do, but of the raw data. I’m fully capable of assimilating raw data and making decisions on it., with one caveat: it’s almost always 24 hours ahead of reality.

    What I mean is this. All stations and the local National Weather Service branch make storm or rain predictions, but the storm actually arrives about 24 hours later than predicted. It has been this way since last Summer. I don’t know what is happening on the inside to cause this. Most TV stations have thier own equipment and trained (but apparently not tested) meteorologist. It’s not like they are all getting the same broadcast and adding their own flavor. Nevertheless, they all make the same mistake, over and over agian.

    If you’re ever down this way, look me up.

    1. dunstan orchad’s blog header is cool, but is it necessarily the actual weather at that local spot?

    2. deane in des moines: the major problem with weather predictions stems from mathematical physics, i.e., the initialization values in the equations. the very first initial values we calculate (or a computer model calculates) from the equations of motion are then approximated (taylor’s series)…that’s how we make a forecast, make a diagnosis of atmospheric conditions.

    the national weather service just went through an extreme makeover of data assimilation. it’s not any better, quite frankly, in large part b/c the nws is run like a business & not as a science institution. the europeans are beating us hands down! (and it’s because of the inital values, besides the fact that all the european member states got together the top mathematicians and meteorologists to work on this initialization problem.) you yourself may be capable of taking the weather data & assimilating it — but if your inital values aren’t very good, your forecast won’t be, either. just what are we working with? to give you an idea, the percentage of error in calculating just the horizontal components of wind (x and y) add up to 200%.

    also, the tv forecasters are not always the best meteorologists because they aren’t always trained to look at all terms in some fairly inegral equations for meteorology. the ams (american meteorological society) recently issued its first brand of policing by making it mandatory, after 2008, for all tv weather folk to have at least a bachelor’s of sci degree in atmospheric science. right now, don’t trust any forecast that goes beyond 4-5 days. even 5 days is pushing it.

Comments are closed. If you have something you really want to say, tweet @gadgetopia.