Compare maybe have NEMS offset
I am looking at the "compare" for ECMWF, GFS, NEMS and I think NEMS may be offset. The location I am looking is 50N 145W (Ocean Station PAPA). I compare with MeteoBlue and the 3hrly winds match up, but the meteoblue times are UTC, while the windy/compare ones say they are UTC-10hr, so the NEMS time series appear to be offset 10hr.
I am not sure to have clearly understood the trouble.
Meteoblue express the time in UTC when Windy do it in local time, which is UTC-10h for your position. NEMS forecasts displayed by Meteoblue according to UTC time must be the same as the ones of Windy according local time (UTC-10h), e.g. Meteoblue forecast at 12:00 must be the same as Windy forecast at 02:00 local time.
Also keep in mind that windy provides data from the "old" meteoblue model.
@idefix37 Thanks for responding. MeteoBlue reports in UTC. Windy is local. We agree. However if you look at the actual numbers, windspeed. the NEMS matches meteoBlue at the same time 00utc == 00 local so it is offset.
example below sunday 00utc is 28kts , look at the compare figure at the bottom NEMS 28 kts at 00local
I think this is a problem, after I posted for a short time the wind time series was fixed, but the daylight was then out of whack. I did a screen grab - here it is. It was rolled back though, apparently.
see how time and daylight got offset after?
You are right. There is a problem with NEMS at your position 50ºN 145ºW
Here is the ECWMF forecast.
Local time at your location is now 9:00
And with NEMS
Local time at the same moment is 19:00 !
Then I get a black label « Not in sync ». Do you have it too?
Obviously NEMS display is wrong at your location.
Hi @oliveRidley @idefix37 @Gkikas-LGPZ , thanks for your reports and comments. The NEMS model really seems to be offset in ocean locations due to a problem with time zones. We will investigate this issue and find a solution.
It seems that offset occurs only in the seas/oceans locations it is easily visible when selecting a location in distant time zone +/- 10. Recently, we have changed timezones API and I think this something related to it. @Ivo