The National Weather Service's new wind-chill scale
November 26, 2001 6:25 AM   Subscribe

The National Weather Service's new wind-chill scale has ruffled some feathers. The new formula is slightly more complicated than the one we've been using since 1945. It's also different from the Real-Feel Temperature Index, the one proposed in 1999 by Accu-Weather, the World's Weather Authority[TM]. Accu-Weather is cranky: they say the new formula is confusing, that change is confusing, and that private businesses are better able to make these decisions.

Who takes it? Whose chill-factor formula reigns supreme?
posted by gleuschk (30 comments total)
 
How about mine? Open your window. If it feels cold, wear a coat. If it feels warm, don't. If the "real-temperature" is going to drop below 32, take a coat. Real hard.
posted by benjh at 6:57 AM on November 26, 2001


From the Real-Feel link: "The RealFeel Temperature takes into account the effects of multiple parameters, including ambient temperature, wind speed, solar intensity, humidity, precipitation intensity/type, elevation and atmospheric pressure. Wind Chill only takes into account two variables - temperature and wind speed..." Sounds like we have a winner here.
posted by alumshubby at 7:02 AM on November 26, 2001


Here in North Dakota, I've read comments by schoolboard members commenting about how the revised windchill scale will result in fewer "snow days" (which are often used when it's dangerously cold out) and more outside recess time in the winter for students. Of course they apparently don't realize that the NWS hasn't simply waved a magic wand and said,"Let North Dakota be habitable in the winter"---all that's been done is change the scale, not the temperature. It will still get ungodly cold here and any adjustments to past policy should only be converting the old numbers in the policy to match up with their revised equivalents.
posted by nathan_teske at 7:05 AM on November 26, 2001


Oh yeah, I care about this issue.
posted by fleener at 7:14 AM on November 26, 2001


From the Real-Feel link: "The RealFeel Temperature takes into account the effects of multiple parameters, including ambient temperature, wind speed, solar intensity, humidity, precipitation intensity/type, elevation and atmospheric pressure. Wind Chill only takes into account two variables - temperature and wind speed..." Sounds like we have a winner here.

If the two variables in the equation account for enough of the variance in the outcome variable, then adding extra variables really doesn't get you much. So the sheer number of variables in the equation doesn't always tell you how well the equation works. To compare the two equations, we would have to test there predictions against actual body heat loss under a variety of conditions, and see which equation is a better fit.

Accuweather is well known for hyping their predictions and waving their hands vigorously at the error involved. (The prediction for the 15th day of their 15 day forecast has a huge amount of error associated with it, for instance, something which is constantly ignored in their [Accuweather's] press releases).

fleener: Oh yeah, I care about this issue.

*waves magic wand to inprove the fleener chill*
posted by iceberg273 at 7:16 AM on November 26, 2001


Thanks, icy.

By the way, what means "well-known"? This was the first time I'd heard of Accu-Weather.
posted by gleuschk at 7:30 AM on November 26, 2001


By the way, what means "well-known"? This was the first time I'd heard of Accu-Weather.

"Well-known" to media outlets, perhaps. In the U.S. at least, the firm provides a huge percentage of weather forecasts to newspapers and TV and radio stations. Sometimes a station will actually say, "Your WTVN Accu-Weather forecast" (which makes it sound like Accu-Weather is their "trademark" for weather forecasts) but just as often, media outlets don't credit the source. Chances are, if you pay attention to weather forecasts at all, you hear Accu-Weather's forecasts every day.
posted by kindall at 7:38 AM on November 26, 2001


By the way, what means "well-known"? This was the first time I'd heard of Accu-Weather.

And thus is my weather obsession laid bare.

Accuweather has been around for a while, and was the first company to offer 15 day forecasts. Weather.com offered 10 days forecasts at the same time, and their was some controversy in the weather forecasting world about whether or not the 15 day forecasts were particularly valid. I read an article on this not to long ago, where the Accuweather folk were waxing eloquent about their 15 day forecasts, but I didn't blog it and so it is now lost to me forever.
posted by iceberg273 at 7:44 AM on November 26, 2001


I have trouble taking Accu-Weather seriously when they provide a forecast like 'A good deal of cloudiness with a shower in spots.' It sounds like someone had an English public school education and is *really* proud of it.
posted by darukaru at 7:48 AM on November 26, 2001


Further reading:
Why calculating windchill is important.
A study on the accuracy of weather forecasting (a side issue to this thread).
Interpreting wind chill as a loss of energy.
A short biography of Paul A. Siple who coined the term "wind chill".
posted by iceberg273 at 7:51 AM on November 26, 2001


"Your WTVN Accu-Weather forecast"

You're right, now that I think of it - I hear this all the time. I guess I figured there just weren't that many good catchy names for weather forecasts, so lots of people jumped on "Accu-Weather".

I really am curious about whether this is an issue that's better served by government departments or private companies. The information I can find all seems to focus on the US/Canada. I tried (not terribly hard) to find information for other countries. Anybody know what they do in, say, China? or Greenland?
posted by gleuschk at 8:10 AM on November 26, 2001


Nice link!
posted by rushmc at 8:17 AM on November 26, 2001


I really am curious about whether this is an issue that's better served by government departments or private companies.

The public versus private forecast issue has been covered some by WeatherZine.

Anybody know what they do in, say, China?

In at least one case they work together.
posted by iceberg273 at 8:28 AM on November 26, 2001


Links to different national meteorological services around the world.
posted by liam at 8:29 AM on November 26, 2001


That China/IBM link is stunning - thanks!
posted by gleuschk at 8:39 AM on November 26, 2001


Interesting links, thanks gleuschk, liam and iceberg.

I also have a weather fixation, but it coincides with Atlantic hurricane season. (Only 4 days left, whew...)

In the comparison chart showing wind chill at 5 degrees, there's as much as a 40-45 degree difference between the old and new systems. That's a rather big jump. I can see how this could have a big impact when determining if it's too cold for work/school/etc. Political conspiracy at work?
posted by groundhog at 8:43 AM on November 26, 2001


why would anyone complain that a formula is confusing? plug it into a computer and the computation is e-z -- unless the formula you're given is incomplete.
posted by moz at 8:49 AM on November 26, 2001


Nice Iron Chef Reference!
posted by aj100 at 8:59 AM on November 26, 2001


If the two variables in the equation account for enough of the variance in the outcome variable, then adding extra variables really doesn't get you much

Maybe I'm not being rigorously scientific here, but in an area with dramatically variable fall and winter humidity, I can tell ya I'd like to see the humidity figured into how I'm going to feel at the bus stop. Dry air doesn't feel nearly as chilly as damp air at relatively mild temperatures.
posted by alumshubby at 10:43 AM on November 26, 2001


By the way, what means "well-known"? This was the first time I'd heard of Accu-Weather.

More clever PR: a company which would suffer in no apparent way, cries in protest simply to get it's name in the media. Try that your your business!
posted by ParisParamus at 10:48 AM on November 26, 2001


Heh, anyone else notice that after 60mph the 'old' scale actual gets warmer as the windspeed goes up?
posted by delmoi at 11:37 AM on November 26, 2001


Maybe I'm not being rigorously scientific here, but in an area with dramatically variable fall and winter humidity, I can tell ya I'd like to see the humidity figured into how I'm going to feel at the bus stop. Dry air doesn't feel nearly as chilly as damp air at relatively mild temperatures.

Right. This is true at mild temperatures (due to differences in rate of evaporation from the skin, not due to the ability of moist air to conduct heat better), but wind chill calculations are most useful for cold temperatures.

At low temperatures, humidity actually has very little effect on wind chill. From a paper presented at the Internet Workshop on Windchill by Robert Steadman:

The humidity effect . . . may be appreciable (-5 to +6 K) in hot weather but is negligible (-1 to 0 K) when ZT [dry bulb temperature] < 0°c./i>

The heat index calculation uses humidity because the ability of the human body to cool itself by evaporation is reduced in high humidity, which leads to overheating at high temperatures (and the myriad of problems that then occur). Wind chill, on the other hand, is a measure of how fast your body is cooling. At low temperatures, humidity doesn't affect rates of cooling much one way or another (at least in comparison to the cooling due to the wind). So adding humidity to the mix really doesn't buy you much for a cold weather calculation (which wind chill is). In fact, if anything, low humidity should increase cooling (remember, evaporation from your skin surface cools your body; at low humidty there is a higher rate of evaporation because of a lower water vapor pressure), but fortunately, the effect is tiny.

posted by iceberg273 at 11:48 AM on November 26, 2001


I find it hard to take seriously any formula that doesn't evaluate to the actual temperature when the wind speed is zero. Both the old and the new formulas fail this test. AccuWeather isn't even in the game, since they don't seem to offer a formula at all.
posted by anewc2 at 11:48 AM on November 26, 2001


Humidity is a major, MAJOR defect if it isn't included. Remember: hypothermia (sp?) kicks in at temperatures way above freezing. To a smaller degree (pi), wet air makes cold feel ALOT colder.

By the way, what ever happened to Bob Harris?
posted by ParisParamus at 12:01 PM on November 26, 2001


Cranky and obnoxious, I'd say. They've "known all along" that the old formula was imperfect--but didn't do anything to improve on it for most of that "all along," and then trademarked their change so nobody else can use it.

The wind chill formula may not be perfect, but I can go to the National Weather Service, or Weather Underground (does anyone else find that an odd name for a business?), or anywhere, and they're all using the same definitions. Meaning that I know what they're talking about.

I'd like to see what their weather forecasts look like without government input. They'd have to pay for their own weather satellites, fund the development of their own computer models for hurricanes, and monitor the weather at thousands of locations without using any of the government-funded weather stations. Or they can admit that publicly funded, shared research and data are useful, and stop kvetching that they don't have a lock on the most recent way of warning people about really cold days.
posted by rosvicl at 12:02 PM on November 26, 2001


And what happened to the font? And formatting?
posted by ParisParamus at 12:19 PM on November 26, 2001




Unclosed i tag. fixed now :)
posted by nathan_teske at 12:35 PM on November 26, 2001


or not.. damn
posted by nathan_teske at 12:35 PM on November 26, 2001


Unclosed i tag. fixed now

It is all my fault, and I cannot see the damage I have caused, because in my browser the italics end at the end of my post. Alas.
posted by iceberg273 at 12:41 PM on November 26, 2001


AccuWeather isn't even in the game, since they don't seem to offer a formula at all.

They say (somewhere - can't find it at the moment) that it's patent-pending, and they'll make the formula public when the patent is approved. They do give the following slightly surreal example, couched as a criticism of the old gummint scale:

"The Wind Chill is dramatically inaccurate when winds are nearly calm. For example, if the temperature is 10 degrees Fahrenheit with calm winds, the Wind Chill would be 5, while The RealFeel Temperature is a more accurate 18 under average conditions for other parameters. "

I never did very well with subtraction (good thing it's not needed for a PhD), but isn't that hogwash?
posted by gleuschk at 12:42 PM on November 26, 2001


« Older From Dismal Science To Joyful Art   |   The next supposed Japanese video game craze is... Newer »


This thread has been archived and is closed to new comments