ECMWF GRIB data

[franksingleton]

Like the UK Met Office, ECMWF has always charged for its GRIB data. PredictWind did some form of deal, I have no idea about details, and now provides both ECMWF and UK GRIB data products to its paying customers. As a loss leader, I understand that they make limited amounts of data freely available. Some while ago, I cannot give the date, Windy,com began providing ECMWF GRIB information on their app. As they also have a commercial service, I guess that they also pay ECMWF and that the rest of us freeloaders benefit.

Yesterday, I noticed that Ventusky.com also had ECMWF GRIBs on their browser version but not (yet?) on their app. Then, when using the WeatherTrack app, I saw that they, too, had ECMWF data as one of their many options.

Experimenting with the app, I saw that they provide the data only on a 0.4 degree resolution, not on the full 0.1 degree as PredictWind say that they do. As there is smoothing in all NWP models, the effective resolution of any model is around 5 grid lengths. Consequently, there should be no loss of real information using WeatherTrack as opposed to paying PW, at least for a limited range of elements, wind, pressure, temperature. I surmise that both Windy and Ventusky are also using the 0.4 degree data in their apps. While these are two good presentations, I always wonder and have never checked the download costs of either. The advantage of WeatherTrack is that you can control download size , useful when roaming and more so over a satphone data connection.

I have to wonder when the Met Office will also be providing its GRIB data as freely as other Met services. A letter that I wrote to the Royal Met Society has brought no apparent change oh heart yet.

Does anyone have any more information on these issues? Any corrections or additional information?

 

[franksingleton]

Perhaps I did not search well enough, quite likely, I am still not sure. But, after contacts with Ventusky, WeatherTrack and OpenGribs, I have now seen ECMWF Open Data - Real Time - User Documentation - ECMWF Confluence Wiki. I have not followed every link here, but it is now clear that ECMWF is quite content for their data to be used with due acknowledgement. Further, a limited amount of data are available at no charge. These are Wind, Temperature and Wave height.



The question now is when and whether the Met Office will follow suit. They are, effectively, joint owners of ECMWF. Continued refusal to provide GRIB data looks odd.

 

[Laser310]

Didn't we just have a thread about this?

UK Met grib is available in expedition (0.18 and 0.09 deg resolutuon) at no cost - they are encrypted and can only be viewed in Expedition.

While Expedition costs money, you can download the program for free and operate it in Demo mode for free. I believe you can get and view the UK Met grib in the free Demo Mode. What you can't do is any kind of instrument connectivity or optimal routing.

to get the UK met, which is on Expedition's grib server, you need an expedition account - but these are available for no cost on the expedition website.

ECMWF is also available off the expedition server - but not at the highest resolution (8km) that is available from PW

 

[franksingleton]

This thread concerns ECMWF and their release of (some) data free of charge. This is welcome.

Whatever anyone may or may not get from Expedition I, personally, do not know. What I do know is that the Met Office charges for big datasets. If you can get data free from Expedition, then they are paying for it. I do not know their business model so cannot comment.

 

[Praxinoscope]

Slight off thread comment, in that I haven’t been able to access ECMWF medium range forecasts since Friday last week, site either freezes or message appears can not find this package.

 

[Mataji]

Perhaps this is a good place to ask, as I have little clue to what this is all about, what’s wrong with PocketGrib.

 

[franksingleton]

Perhaps this is a good place to ask, as I have little clue to what this is all about, what’s wrong with PocketGrib.

Nothing.

GRIB data are made available by National Weather Services and by ECMWF. Pocketgrib uses the US NOAA GFS and the US Navy COAMPS. Other apps provide data from other models. Some sailors like to see more than one model. There are differences. ECMWF is a world leader and it’s forecasts are likely to be the best, particularly beyond 5 days. The UK UM is a leading model with a slight edge on other countries.

Until recently, ECMWF only supplied data on prepayment. They now make a limited amount of data free to anyone wishing to make them available to us end users. Windy.com pays ECMWF and kindly makes the data freely available. Ventusky, I understand, does not pay but now makes wind and temperature data freely available. WeatherTrack GRIB app also does not pay but can now provide wind, temperature and wave forecasts from ECMWF.

For me, the GFS is good enough although I do like to see how consistent the various models are. Were it available as easily, I would prefer to use the UK UM but do not think that it would be so greatly more useful to warrant paying. Others like to place more reliance on other forecast models. That is why the release of ECMWF data is interesting and why some would like the UK GRIB data to be as freely available as the GFS and the German ICON.

Apologies for a long answer to a one sentence question.

 

[Roberto]

Ecmwf gribs at 0.4° grid can be downloaded with Saildocs queries. From Saildocs site:


ECMWF "Open Data":
===================
The European ECMWF model data is highly regarded but the full catalog is not available for distribution without prohibitive fees. Previously, only very limited data was available as "WMO Essential" data to satisfy international obligations. However, ECMWF has now extended their "Open Data" project to grib data. From March 8 2022 Saildocs now provides this data which can be freely distributed.
Available parameters are WIND, MSLP, HGT500, and TEMP (2-meter air temp).
The code is "ECMWF", default parameters are surface pressure and wind. The grid is 0.4 x 0.4 deg, forecasts are available from 00 to 144 hours (6 days) at 3-hour increments and 150 to 240 hours (ten days) at 6-hour increments. The data is updated twice a day for the 00z and 12z forecast, and generally available by HH+8, or 08:00 and 20:00 utc.
A typical request might be:
send ECMWF:35N,19N,123W,102W|0.4,0.4|0,3..72|MSLP,WIND

 

[franksingleton]

Thank you, Roberto. I had not thought about Saildocs. How did you pick that up? Their update should be about 0700/1900. Windy.com can often be late in issuing the latest forecast, up to 2+ hours later than the scheduled release time.
Using the Windy compare facility, there is usually little real difference between ECMWF, GFS, ICON for the first 5 days. Each is usually within 5kn of the others, ie within the likely ensemble range. Beyond 5 days, ECMWF should be best.

 

[franksingleton]

One thing leads to another. In my searching I looked at Saildocs to find that they now provide ICON and ECMWF GRIBs aswell as GFS and others including ocean currents. I have put a SailDocs Models - Franks-Weather - The Weather Window page on my site. This is taken from a rather long, unwieldy Saildocs write up

For many, Saildocs may seem for the nerds. However, for anyone with limited bandwidth, using email over satphone, HF/SSB radio or just with poor connections, it is the most efficient and flexible of all GRIB services.

 

[Laser310]

For many, Saildocs may seem for the nerds

one thing about saildocs, that is not known to many of the users, is the availability - by return email - of compressed versions of many of NOAA's graphics.

so; you don't have to navigate to a web page, which is impossible with iridium handsets/go.., and you get smaller files. They are often only 10 or 20kb. Obviously, these are more useful for those in US waters..., but they are also useful in the caribbean, and the atlantic generally.

for example:
an email sent to query@saildocs.com.., with this text in the body: send PYEA86.tif

will return a compressed (15kb) version of the 0Z tropical surface analysis covering much of the north atlantic. As most of you know, this is a human product. I download surface analyses and forecasts every day at sea. Gribs are not enough...

note that while gif versions of the images are available on the noaa website, you need request the tif version from saildocs. you will need a tif viewer - i use irfanview, which is free. several PC nav programs let you import and georeference tif files.

because tif files can not be uploaded here, i have uploaded a png version.

 

[franksingleton]

If you use the MailASail facility, http://weather.mailasail.com/w/uploads/Franks-Weather/Mailasailcharts.pdf, it is even easier because Ed Wildgoose has produced a list that obviates you searching NOAA’s site. It is on my Marine Weather Forecasts, Charts and Actuals on the Internet - Franks-Weather - The Weather Window page.