Thursday, January 17, 2008

Windchill update

I've updated my website main page to include a graphic on windchill readings from my station (dial image showing latest windchill, plus hi/low stats) Click the dial image to get more windchill info. Windchill will be the main story Friday.

IMPORTANT: Wind readings from my station are much lower than exposed sites such as the airport, or rooftop locations. My anemometer is only about 15 feet off the ground, and is surrounded by trees and buildings. As a result, windchill readings from my site frequently underestimate the actual windchill for the area. This is why my site should never be used to make official decisions with respect to windchill! The station at PJ school has better exposure for winds (on the rooftop) for Charleswood area windchill values.


  1. I just read some wind chill forecasts for select areas of Manitoba. The morden area has a value of -46 predicted for the morning. This takes into account -30 temperatures and 30km/h winds. The Steinbach area says -43 conditions with -30 temperatures and 30km/h winds as well.

    Could anyone enlighten me as to why the exact same temperature and wind forecasts could yield different wind chill values?

  2. Sharp eye Scott :)

    The reason for the inconsistency is due to the calculation of the forecast windchills, which is done on a 3 hr basis based on the forecast temperature and wind speed for each location. This data is input via a forecast grid, and then the text forecast you see is an automated translation of that gridded data. The wind speed in the text forecast is a general average for the day, but it could vary from that average by 5 or 10 km/h at any time. Because of this, the windchill value could vary by several units between two sites that seem to have the same forecast.

    So it's likely that the Steinbach low of -30C coincided with a wind speed of 20 km/h that a windchill of -43. Meanwhile, the -30C at Emerson was occurring with a 30 km/h wind giving a windchill of -46.

    Someday, the gridded data will be available for all to see on Weatheroffice.. which will give more detailed and clearer weather information for each location every 3 hours, or every hour. Until that time, we'll have to put up with the limitations of the text forecast.

    Hope that clears things up a bit for you! And bundle up.. regardless if the windchill is -43 or -46!

  3. Thanks for the information rob,

    I was also wondering, since we are on the topic of wind chill, does environment canada use sustained speed or gusts for its forecast? I know that some days there is no "gusting wind" forecast and therefore there would no gust wind chill calculation. But on those days when the wind is gusting does EC use the highest gust for its calculation or the sustained?

  4. I doubt either Environment Canada or the US National Weather Service use the gusts. I use the gusts in the unofficial forecasts I do for my area when it looks like the highest wind gusts could coincide with the lowest temperatures, since if you're outside, whatever the highest gust is would relevalt, since that's what you'll feel. Of course, since my forecasts are unofficial, I'm not subject to specific rules. I just try to be accurate.

  5. That's correct. Official windchill readings are based on sustained wind speed.. gusts are not considered.

  6. Thanks, I thought I was right. Gee, I should have checked my post before submitting it. I had a bad typo. It should have said: "whatever the highest gust is would be what is relevant, since that's what you'll feel." I guess that's what happens when it's 13 degrees in the house and you have cold hands. That's the official excuse anyway.

  7. Tomorrow will be a cold one regardless of the wind.

    Thank-you again for responding to my questions it did seem unlikely that National Weather Services would use gusts for reading as they are too unpredictable.

    My prediction is -31 tomorrow morning for the south-east as it is still clear as of 11:00 p.m. and the clouds won't move in intime to stop significant cooling. Wind chill -43 to -46!

  8. I'm calling for wind chills possibly as low as -46C (-50F) in my area. Looks like we're on the same page.

  9. Friday Jan 18th - Not as bad out there as I thought it was going to be today, at least here in Winnipeg.. temperatures not quite as cold and winds not quite as strong as I expected this morning.. thus windchills "only" around -40. Still cold and worthy of windchill warnings.. but we've had worse.

    Forecast calls for winds to pick up this afternoon so we may see colder windchills this afternoon into tonight.

  10. The wind seems lighter here as well, but it's probably much worse in open areas.

  11. Wait for the forecast for Saturday night, though. Ugly. (Hint: it's REALLY close to -50.)


  12. Windchill forecast. that is.

  13. Rob,

    For my website I am looking for some advice on calculating accurate temperature and wind forecasts. I thought you would be an excellent person to ask in addition to any other individual in this forum who has knowledge on the subject.

    1) What are/is the best model to look at to determine temperature and wind for short-range forecasting (i.e. within 3 days).

    2) After looking at the models what are some ways to determine whether the low/high should be calibrated according to other features. I mean how do you figure what a difference in the temperature will be from the result of clear skies, temperature advection, 850mb temps, etc.

    3) For wind is it better to look at isobar maps or to look at an actual wind forecast map. And how do you determine a fairly accurate average wind speed for a certain time-frame.

    Thank-you in advance!

  14. Scott..

    First of all, it's important to realize that there is no "best" model. If there was, that's all we would use. They all have different physics packages and algorithms which perform well in some cases, poor in others. Generally, they're all very good at getting the basic idea right in the first 48-72 hours.. and then you have to use your expertise and meteorological skills to fine tune the details for your specific area. I have to tell you that Environment Canada's forecast temperature and wind algorithms from their regional model (called UMOS) is very good, and getting hard to beat. Generally though, the human forecaster can add the most value to a model forecast within the first 12-24 hours, but after that, it's getting tougher and tougher to improve upon a model forecast.

    In the short term (first 12-24 hours or so) nothing beats looking at good real-time data and making your projections from what's really happening. It's always best to make temperature and wind correlations with what's happening upstream to get a rough idea with what to expect in your area. Then you make adjustments for local effects such as cold air drainage, downsloping effects, expected cloud conditions, snow cover, etc, etc. Also, be aware of sharp temperature and wind boundaries (such as across a warm front or cold front) that can really throw off model forecasts if the frontal positions are not handled well.

    That's very general advice, but if you need further specifics, let me know your email address and I can send you more detailed information.

  15. my email address is,

    I am still a beginner so terms like cold air drainage and downsloping are new to me.

    I would appreciate more information if you have!

    Thank-you again for all the help.