Incorrect wind speed
-
They are reporting Dorian wind speeds of 175 mph, but Windy says the highest putter band winds are 83 mph.
John M
John Miller
-
Please see many other threads regarding this subject, including this recent one:
https://community.windy.com/topic/9456/dorian-wind-speed-and-pressure/2
And the detailed discussion here:
https://community.windy.com/topic/4415/why-is-wind-speed-in-hurricanes-so-wrong
Additionally, Windy does not create any forecast data but instead only visualises forecast and actual data received from various third party providers.
There are many useful but also complex and expert level tools and data available via Windy, therefore the user must always take care:
- To understand what information they are displaying and viewing
- To understand from where and when the information was sourced
- To understand what this information actually means.
Comparing Forecast data with Actual data can be a common mistake.
Around 1300 UTC the Windy Hurricane Tracker was reporting wind speeds of 142kn / 164mph / 264kph.
Hope this helps :)
-
In addition a further recent topic here with a screenshot explanation of how comparisons can easily be invalid depending on the data used:
-
@skypics234
I see highest (forecasted) wind speed 139 mph (2 Sep. / 15 Z)
-
... or even higher (3 Sep / 00 Z)
-
Yes, unfortunately all of these new threads reporting Windy to be 'wrong' are unfortunately a situation of different data being compared with each other.
-
The forecasted wind GUSTS do not fall in line with forecasted GUSTS of other sources at all. We see it all the time in NY. Your predicted GUSTS are multiple times more than other sources. We use professional custom prepared forecasts as well as public ones. Your gust predictions are always much higher than any by far. Sustained wind info seems good. Compare your forecasted gusts head to head with others and you'll see. Maybe you are using some kind of bad math to convert speeds to different units or something...?
-
@hope4411 I usually see good comparisons with wind speed (sustained) but not accurate with GUSTS. just to be clear... and I am not even talking Hurricanes....just everyday winds and forecasts...thanks...
-
@hope4411 said in Incorrect wind speed:
The forecasted wind GUSTS do not fall in line with forecasted GUSTS of other sources at all. We see it all the time in NY. Your predicted GUSTS are multiple times more than other sources. We use professional custom prepared forecasts as well as public ones. Your gust predictions are always much higher than any by far. Sustained wind info seems good. Compare your forecasted gusts head to head with others and you'll see. Maybe you are using some kind of bad math to convert speeds to different units or something...?
What exact data for wind gusts are you comparing to what other exact data? And are you referencing the same forecast issue times?
As I mentioned earlier, Windy does not create any forecast data but instead only visualises forecast and actual data received from various third party providers.
These third party providers ARE the reputable and official (and normally governmental) global weather data providers, such as ECMWF, GFS, NEMS, ICON-EU, AROME, NAM, etc.
Many other Weather forecast providers use this same data too (especially the free GFS model), and some add their own manipulations as well.
If you have already done comparisons please share actual screenshot examples if there appears to be a consistent error.
Here is a screenshot of the forecast for New York (but this is just an example, a specific location can yield different results).
On the second line you can see the Wind Gusts from each available model, in this case GFS, ECMWF, METEOBLUE, and NAM, for the next days.
Even between the four models displayed here you can easily see a 50% or 100% difference between the Wind Gust forecast. Each model interprets it's own data differently and gives a different result.
https://www.windy.com/40.730/-73.990/wind?gfs,gust,40.183,-73.755,7,i:pressure
-
Also, have you considered using Windy's "Observation vs Forecast" Feature?
How To Use The "Observation vs Forecast" Feature:
https://community.windy.com/topic/9425/how-to-use-the-observation-vs-forecast-feature -
Example. Binghamton, NY. Wednesday 9/4 at 2pm you show wind predicted at 12mph with gusts to 42 mph for ECMWF. GFS is 9 mph and 22 mph. Meteoblue is 2 mph and 11 mph. NAM is 11 mph and 21 mph.
Our custom forecast is 11 mph winds with no gusts. NOAA is about the same prediction.
-
NWS and NOAA sources we look at. Again sustained winds are pretty close to yours (such as the GFS), but it is the Windy GUST info seems off sometimes by 10-20 mph and almost always much higher. We just looked at Brewster NY and similar results. Maybe you can check NWS or NOAA to see theirs and compare head to head. We have seen times where 30-40 mph gust difference.
-
@hope4411 said in Incorrect wind speed:
Example. Binghamton, NY. Wednesday 9/4 at 2pm you show wind predicted at 12mph with gusts to 42 mph for ECMWF. GFS is 9 mph and 22 mph. Meteoblue is 2 mph and 11 mph. NAM is 11 mph and 21 mph.
Our custom forecast is 11 mph winds with no gusts. NOAA is about the same prediction.
Where is the "error" from Windy in relation to this?
Each Model is giving it's own interpretation and it's own forecast result, and as you can see, each one is different, including your custom forcast too.
If you think the Models are wrong, then that may be so. But the Model accuracy is not within the control of Windy.
And all of this data is Forecast data, so there is no telling who is "right" yet.
-
Again the sustained winds are usually pretty close, so we are confused why the GUST data would only be off...puzzling.
I understand you feel like I am bashing etc. I am not. I love your site. I am just reporting that the models or converstion or something that you use are very, very high for gust data compared to NWS and NOAA that we use. The sustained match ours well, so we know something is coordinating well. That is what I am asking you to review...is there an issue with your gust data...please compare as I mentioned. You will see big differences in GUSTS at a given time and day to NWS/NOAA, while the sustained match up pretty well.
-
@hope4411 said in Incorrect wind speed:
NWS and NOAA sources we look at. Again sustained winds are pretty close to yours (such as the GFS), but it is the Windy GUST info seems off sometimes by 10-20 mph and almost always much higher. We just looked at Brewster NY and similar results. Maybe you can check NWS or NOAA to see theirs and compare head to head. We have seen times where 30-40 mph gust difference.
Where do you think NWS and NOAA are getting their data from?
(Here is an interesting link with available Datasets at the left side: https://www.ncdc.noaa.gov/data-access/model-data/model-datasets/global-forcast-system-gfs).You stated in another thread that you "know a thing or two about weather and forecasts" but you keep repeating "yours" in reference to Windy.
I'll say again. Windy does not create any forecast data. Windy does not have a meteorologist on staff. Windy does not interpret the data. Windy ONLY visualises the data.
And in the case of New York Windy is visualising four different sets of Model data for comparison.
-
@hope4411 said in Incorrect wind speed:
Again the sustained winds are usually pretty close, so we are confused why the GUST data would only be off...puzzling.
I understand you feel like I am bashing etc. I am not. I love your site. I am just reporting that the models or converstion or something that you use are very, very high for gust data compared to NWS and NOAA that we use. The sustained match ours well, so we know something is coordinating well. That is what I am asking you to review...is there an issue with your gust data...please compare as I mentioned. You will see big differences in GUSTS at a given time and day to NWS/NOAA, while the sustained match up pretty well.
No, I don't think that you are bashing, I think that you do not fully understand how Windy works (see my above post).
Also you need to provide more concrete data for comparison, otherwise we are not comparing apples with apples. Please add screenshots with ALL references, time points, sources, etc.
Eg: Raw GFS data from your side vs the GFS data displayed on Windy. NB: other 'Forecasts' from other sources, even if GFS based, can also still be different if this source is adding some of their own manipulations as well.
And finally of course, all of this will still only be Forecasts, so it will still not be clear who is "right".
I will mention it to the Windy team and see if they are aware of any specific issues, but I suspect not.
I suspect that this is just simply the result of different Models (and even individual Meteorologists / Forecasters such as your custom forecast) giving different results.
Hope this helps :)
-
Brewster, NY Wednesday 9/4 12pm (in MPH)
(Windy)
Source Sustain Gust
ECMWF 12 28
GFS 12 22
METEO 1 16
NAM 8 30(other)
NWS 11 22Brewster, NY Wednesday 9/4 3pm (in MPH)
(Windy)
Source Sustain Gust
ECMWF 12 30
GFS 13 23
METEO 0 17
NAM 11 22(other)
NWS 10 20
Mechanicville, NY Wednesday 9/4 12pm (in MPH)
(Windy)
Source Sustain Gust
ECMWF 12 35
GFS 13 22
METEO 4 16
NAM 16 26(other)
NWS 14 24Mechanicville, NY Wednesday 9/4 3pm (in MPH)
(Windy)
Source Sustain Gust
ECMWF 7 33
GFS 10 22
METEO 3 14
NAM 10 19(other)
NWS 13 21 -
I put examples of some quick comparisons I did in the prior message. I believe what I am finding is that our NWS is more closer to your GFS or NAM. Just to note...It is far off from the METEO data. Most importantly, I had been using the default ECMWF (until now!) but see that is usually the highest gust speed by a lot and likely why I felt it was so far "off". That one seems to always be too high in our opinion. Maybe switching to the other GFS or NAM will be better for us. Does that make sense to you based on my findings?
-
Yes it makes sense.
Looking very quickly I was going to say in general that I don't really see a problem with the data that you posted (overall, in general).
ECMWF does seem in general higher, but comparing all 5 models by eye to me I see the same general average in most cases.
Are a couple of models higher or lower in some cases? Yes. But each Model uses different methodology, that is the point.
It is also why multiple Models are used by Windy when available, and why Windy makes various Comparisons possible for the user.
-
OK, thanks. I believe in our observations over the last couple years, the predicted Gusts seem too high compared to our other sources of weather station info, as well as the media reports. We normally leave the default on 'ECMWF' in Windy, but now we will try to use 'GFS' (or 'NAM') and see if we are running in "more parallel" between our info and Windy, specifically for predicted Gusts. Again, thanks for the help and good work on this. As an electric utility company, we depend on great information to help us with storms, power outages, etc...so you see that the wind speed, gusts, etc is crucial data to us.
