ECMWF

WX OM

The ECMWF model hours are now 360 hours in the new 49r1 upgrade. Any plans to add it?

https://www.ecmwf.int/en/newsletter/181/earth-system-science/ifs-upgrade-improves-near-surface-wind-and-temperature

zekun_ni

This. In my experience the AIFS forecasts after +240 is subject to a certain amount of model bias and is nearly useless for a number of locations including PNW where I'm living. It'd be a decent improvement for Windy to replace those forecasts with the updated ECMWF IFS.

idefix37

@WX-OM
@zekun_ni

It is the Ensemble model which has been extended up to 15 days. The HRES model is still providing 10-day forecast.
The medium range charts from ECMWF are still over 10 days as said here.

See more information about the new cycle 49r1.
And specially

zekun_ni

@idefix37

I believe you are misinterpreting the documentation provided by ECMWF. In fact, the text in the red box highlighted by you has said exactly that there is no distinct "HRES" (which formerly provided 10-day forecast) any more, and "HRES" is now the same as the ensemble control, which goes up to 15 days.

idefix37

@zekun_ni
Yes, but explain why HRES is always displayed as a 10-day forecast?
To get the 15 days, you need to replace HRES with the Ensemble model Control. It is not the same data source and probably requires an access agreement.
See these related posts
This will be surely unified in the future.

zekun_ni

@idefix37
At least in Open Data the operational IFS forecasts have been directly extended to 360 hours without any change on the data location. Example: https://data.ecmwf.int/forecasts/20241117/00z/ifs/0p25/oper/

Technically it could be working out differently in paid real-time subscriptions, which I have no knowledge about. But even if you can't manage to get that data, you can use Open Data as a fallback option. Of course it's only a fallback since the resolution is lower, but compared to AIFS (which is also in 0.25-deg resolution) it won't be worse.

مراقب مناخي

For the next IFS upgrade, Cycle 49r1, planned for
October 2024, the HRES and the ENS control are being
made bit identical. Both will run to 15 days for 00 and
12 UTC

From IFS Cycle 50r1, planned for October 2025, we
will stop producing the current ENS control, and the
data stream that is currently called the HRES will be
called 'ENS control' instead. This single, unperturbed
forecast will continue to be made available earlier
than the ensemble perturbed members, in line with
the agreed dissemination schedule

idefix37

@zekun_ni
So you would accept a lower resolution just to have 5 days more in the forecast, with low accuracy?

Filip_K

Hi all, these are official ECMWF release notes regarding 49r1.

HRES is indeed forecasting 15 days now:

We are working on acquiring and using the new data as soon as possible. I can't give any precise day yet though.

WCT

@idefix37 yes, that doesn’t matter to me

WCT

@WorldwideCycloneTracking I don’t think that they are now low accuracy, in fact, they are as quite as or more accurate than the GFS model...

zekun_ni

@Filip_K

Thanks for your attention on this matter. Waiting for further updates from you and Windy!

WCT

@Filip_K Thank you so much. It means a lot to me. But it makes me wonder will it be available in map form?

Filip_K

Hi @WCT, yes, there is no need not to show it in all its glory (:

WCT

@Filip_K Thank you so much !!!

zekun_ni

@Filip_K As far as I can see the update is now online. Thank you very much for your effort!

przemo1986

@zekun_ni aktualizacja nie zawiera map na 16 dni a chyba model miał zawierać wizualizacje

przemo1986

@Filip_K said in ECMWF:

Hi @WCT, yes, there is no need not to show it in all its glory (:

Aktualizacja danych jest ale widzę że nie wprowadziliście wizualizacji danych modelu na 16 dni czy będzie taka aktualizacja ? Czy tylko na 10 dni ?

Filip_K

Hi @przemo1986,
the forecast is available in detail (lower panel) for full 15-16 days already (basic forecast, meteogram and airgram), where it was previously extended by AIFS. Full functionality for extended forecast, like map visualization, will be available in next version, which will be available in following weeks.

Thank You for Your patience.

illili

@Filip_K I see on desktop V44.0.1 at least (iOS 18 with Windy app V43.0.4 does not have it) that ECMWF extends out to 15 days now.

Is that leveraging the AIFS that was previously only in the forecast detail, or is that just the "standard" ECMWF model that @WX-OM mentioned a few weeks back?

Is there any plan to have both versions of the ECMWF available?

Thank you for the hard work!