Why is Wind speed in hurricanes so wrong?
@Siff Where do you get the real data from for everywhere on the planet? It's easy to say you have real data in somewhere like the USA, which is covered by hundreds of observation points, along with upper air measurements taken twice daily. What about everywhere else?
How do you verify a forecast for a location which there is no credible observation data? Remote sensing can help, but it only goes so far.
stevecrye | Premium last edited by
@TZ 100% correct TZ. Others in this thread clearly don't understand weather modeling. Windy pulls the model data - NAM 3k, ECMWF 9k or GFS 22k - from NOAA and other agencies and then displays it on a great GUI - all 100% free. Windy does not generate model data. Prior to Windy I was forced to look at the raw data - ugh. I'm hoping soon that Windy can also show other "popular" models such as HRRR, OP4-, BAK40, etc. Also, a skew-t at a click would be wonderful! Keep up the good work. Windy is an invaluable tool for pilots and sailors, who use it every day. As for the haters - take the time to learn how to use Windy - for example, I just now selected Wind Gusts and see 155 MPH near the eye of Irma - very accurate. And study meteorology , it might help you get a clue ... if all that fails, go use some other app/website, most are not free.
You are absolutely right!
Windy is the best "tool" for visualization of weather models and the only one that provides, for free, data from the ECMWF. Unfortunately some people cannot distinguish model data (forecasted values) from real measurments (observations). And some others even expect to have forecast radar images !!! ignoring that radar "can see" only what happens now ... forecast models "visualize" the future (with all the inaccuracies that include).
P.S. sorry for my poor english
If someone wants to compare actual wind measurements with model's outputs:
Here can find ocean wind measurements from scatterometer (satellite instrument)
@Siff , it is mentioned in the about section. It says the site uses three weather models. i don't see why you feel they need to spell out the things they do not use.
@georgewells , news agencies are most likely getting their data from the National Hurricane Center, who are in turn getting their wind speed data by flying an airplane into the storm. This website relies on a few standard models to present it's graphics, not actual observations.
Siff mentioned that he had been comparing weather station data to Windy's reported surface winds and they agreed. I too have been doing this following Irma (I live in Orlando). I went to weather channel or other local weather data for a location and compared it to Windy and they did seem to be very close (for wind speed). Weather stations measure the actual wind speed from instruments that are 10m off the ground (http://www.wral.com/weather/blogpost/1283652/). I would assume that the reported hurricane wind speeds are measured the same way, but I agree that there’s a big disparity between the reported speed and Windy’s speed. Another very similar site, www.ventusky.com, has the same disparity.
I don't care about the politics for sure. I also think this is a great graphic to indicate what's happening and going to happen. Whether there is some error in the system of 20 or 40 %, it's a great way to show laymen some technical information that isn't a bunch of stuff that only engineers, meteorologists, or geniuses can comprehend. GREAT JOB
AnemoiVenti last edited by
Windy needs to get the Winds right... kind of a poetic necessity. :)
I agree with many that that the wind data sets (wind & wind gusts) are are not easy to nail at all - let alone consistently. But I truly believe it needs to be mastered over time... especially in the angry weather world we have brought ourselves.
I think Ventusky is making a good effort here (though not sure on data accuracy audits re wind).
When Cat5 Matthew rolled thru Florida last year - the wind data projections were wildly off the further out from the storm centers.. useless actually.
When I looked that the micro-local Dark Sky wind speed projections... they were spot on.
I think that (in the end) there needs to be a artful balance between macro and micro data. That said: when you are in a storm - most want to know what is going to happen to me where I am... and those we love.
Master the winds... Windy!
Stay safe everyone....
Windy states very clearly that winds are arrived at using all relative, applicable and resourceful data and they are defined. The arrows represent wind velocity, (relative speed and direction), within a certain volume suspended above the surface. There may be a billion directions and speeds in that volume, but the one chosen represents the speed and direction expressing the central or typical value in a set of data, in particular the mode, median, or, (most commonly), the mean.****
IMO windy is a great tool-very discerning engineering.
Mike Boyle last edited by
Steve, this is great and I thank you (I live in Florida)
A better work around is take the measurement at 100 meters or 500 meters.
This is a compromise of sorts.
This is a worthwhile project and may save many lives here.
I am reposting on Facebook with some editing. all the tool bars and are can be distracting. I wish you could do false colour white green and blue.
Mike Boyle last edited by
ralphthemagician last edited by
@Siff I get what you're saying, but you can't really compare this to the general media. Modern entertainment journalism just makes stuff up. They don't care. Their information comes from the aether. They just want you to keep watching through the commercials to the next segment. They'd say the winds are made of explosive gas and are gusting at Mach 2.5 if it was believable.
Please at least post a notice of wind info not being related to actual real time experience. I used this to track storm as my daughter and crewof S/Y Ocean Star are stranded in Tortola. I was am so confused and it absolutely added to all the stress!!! Did not know what info to trust.
Finally realized the tracking portion is the only real time data that is helpful . All that other effort to post data is USELESS if it is not actual experience in real time.
I am amazed at the accuracy of Windy hurricane 'forecasts'. OK the max wind speed is off, but you can tell the severity of the storm by its size and color bands. I kept some screen grabs captured last week using the european model and noticed Windy 'forecast' for Irma was spot on (going up west coast of Florida starting Sunday 10 Sept). I now am looking at forecast of Jose for Sept 18 ....wow a monster off the USA east coast and yet another hurricane forming at 20N 40W (while the GFS model does not show this at all). FABULOUS TOOL, CONGRATULATIONS!
@Siff The explanation makes perfect sense from a data science perspective. The grid is aggregating data...it is a window function. The simplest example is an average (rectangular window function). Take the average of 10, 10, 20, & 80. The average is 30. Clearly, 80 was included, but the output (30). When the OP says that it "smoothes" away extremal data, they literally mean smooth:
That said, there are better methods that can preserve extremal measurements. These are called "High Resolution Schemes" or "Total Variation Diminishing" methods:
These let you recover data like this:
The challenge with using these for weather is scale and computing power.
Stay safe out there!
Ricky_Lightning last edited by
@Siff The sources that are providing the data sets for the likes of windy.com, has been proven to be selective in what data they distribute. For example the reporting of Earthquakes - Dutchsinse (A YouTuber who operates a channel warning various parts of the world of any imminent earthquake related threats) regularly demonstrates examples of earthquakes being reported on other nations websites, (other meaning non-US) but not by the USGS. This is a regular occurrence. Ergo if the source is unreliable why bother using its data?
@Ricky_Lightning See my post, this is not a "data withholding" problem. It's math. There are only so many ways to aggregate a bunch of measurements into one while preserving the character of the data.
Ricky_Lightning last edited by
@meteo-GR what would be the point in comparing the two as you've suggested? The source of the data in both examples, is from the same source.