No miracle with SU11 
(that was expected since nothing in release notes about that), I confirm, clouds depiction in Live weather has not been improved, same unrealistic ones as always…
Above spain:
Asobo doesn’t add METAR cloud circles. Meteoblue does with their post processing. I dislike that post processing. Makes the weather behave strange. I have showed that earlier in this thread.
It’s unlikely that Meteoblue would process METAR every 10 minutes and serve the data to Azure/MSFS. The required infrastructure and the service itself is too heavy, unless they run all their models on Azure already. For the second, no meteorologist would design and release a product which so obviously degrades their original MSFS weather by a method without a perspective to ever work well (METAR blending).
Read through this and i think you will understand. We had the 2 model runs at release updated every 12 hours. After su7 it updates every hour with post processing.
Remember the hour of data that was missing after 01:00z after su7 were released?
My conclusion is that we have a different type of weather model now.
Pre su7= Weather data based on simulations are updated by “model runs”.
post su7= Weather data based on measurements are updated when measurement data become available.
I would like to have pre su7 weather model as an optional feature. That would make us all happy again 
So essentially rather than being based off of a model and “forecasted”, they’re using measured which implies “live” and we all know the only way to get live in the sim is through the METAR model. Hence the METAR issues. The only areas that would need to use a model of forecast data would be those that don’t have coverage through measured data, would explain the difference between spotting METAR influenced areas and those which look more natural.
To me both models are live. we all get same weather in the sim at the same time. But the simulated model behaves as weather should behave.
Neither of those models is the “real” weather in real time. Thats impossible to get into the sim.
Thats why we see different opinions here because we preffer different things.
Some preffer it to match METAR 100% and doesn’t care if it feels more static
Some preffer it to be dynamic 100% and doesn’t care if it doesn’t match METAR 100% (i preffer this and this i had at release 2020)
Now we are all forced to use the first.
METAR is live weather because it’s data that is being taken live from current conditions in 30-60 minute variables. A forecast model is not live, the only part live about it would be the moment it’s created because it’s taking the now conditions and guessing how they will develop into the future. That’s what creates the dynamics, when you take current conditions, it’s no longer dynamic, you’re forcing the weather to look a certain way. Only dynamic thing about METARS is the clouds moving.
But what system do you preffer?
Dynamic or fixed to METAR?
Both are weather. None of those are wrong. But the later make me not enjoy the sim weather. Because it doesn’t behaves like weather in my opinion. When i know what i had it only makes me dissapointed when weather suddenly switches.
This page describes their nowcast data, which is the data you see on their website, not the model data they send to Asobo.
Asobo doesn’t receive hourly updates for the entire world, or at least they have never said this. They repeatedly say that it’s model data only and that it takes quite a while for them to process all the weather data for the entire world and for the cloud servers to update. They now receive 6 hours worth of model data according to them, which would make sense because Meteoblue models are generally 6 and 12 hours depending on the area.
It’s important to separate Meteoblue’s nowcast data that they have on their website (Model + Live data) with the data they send Asobo (model data only). If Meteoblue was sending nowcast data to Asobo, Asobo would have no reason to do any METAR blending at all on their end, because it would already be done by Meteoblue (and much better). The clouds would also match the Nowcast map 100% of the time, which it doesn’t. It doesn’t match the Nowcast map 100% of the time and it doesn’t match the METAR string 100% of the time (at least not in my sim), which is the behavior we would expect with model data.
The model data has live weather data at the beginning of the data set and entirely forecast data onward to the end of the dataset. And model data is not random, so it could match live observations towards the end of the dataset as well. In other words, Model data can mostly match METAR sometimes, but if it doesn’t match it all the time, how can we conclude it’s coming from METAR? Just like if it doesn’t match the nowcast data, how can we conclude it’s coming from the nowcast data?
I know some people are using Rex, even though it’s METAR based it’s capable of a lot more clouds variety and to reproduce overcast skies with the right clouds, why is it that the METAR in MFS can’t do either?? I would have been much more forgiving with the METAR bobbles if they at least behaved like REX…but instead the whole weather in in shamble…
It does only mention nowcast as a postprocessing method in that link.
Asobo uses METAR as a post-processing source. They have said that too. Meteoblue injects METAR data into the weather model.
That’s true. Asobo openly talks about the METAR data they blend with Meteoblue and even gave us a METAR window so we can see the data they blend.
That’s how weather models work. They start with live data and predict where that live data will go spatially (outside of the live data coverage areas) and temporally (over time - up to 12 hours). So like I said (and Asobo also says), the beginning of the model data is live data in the areas where there is live data coverage from METAR and other live data sources. Which is part of the reason the hard transitions between datasets can be pretty rough. Some areas are transitioning from 6-12 hour old model data to live observations, because you’re going from the end of one dataset to the beginning of another.
Edit: In response to that video from 2021. This one is from a few months ago where they say they receive model data every 6 hours and interpolate between the hours and blend some METAR data every 30 minutes (winds etc). They specifically say they blend METAR data because of the age of the forecast data they get from Meteoblue. If it was hourly updated nowcast data, they wouldn’t feel the need to blend in METAR as well because it would already be live data in the coverage areas..
‘Initial conditions’ description by Meteoblue:
But Asobo says they not inject clouds from METAR. Then it needs to be Meteoblue.
This timelapse i did shows that either Asobo or Meteoblue changes the clouds from few into broken. In the beginning of the clip both insim METAR and the METAR on Meteoblue page says FEW and then both changes to broken and then a couple of minutes later the clouds changes.
Yes, Meteblue uses METAR. They are open about it, Asobo is open about it, nobody denies it, and you can read in great detail on their website (and other websites) about how weather models start with as much live data as possible to create the initial conditions. But that is only at the beginning of the 6+ hours of weather data since obviously they can’t pull live data from the future when generating a 12 hour weather model…
It shows that the METAR string changed from “few” to “broken” and then quite a while afterwards there was some shift in the clouds.. You simply called it going from “few” to “broken” in your post to coincide with the METAR window, yet If that video was the timelapse with no other windows open, I would not have described that as transitioning from “few” to “broken”. I would have described it as the clouds moving towards the camera, the entire cloudscape appeared changed a bit, and then the clouds continued moving.. That could be caused by so many different things, including the potentially world-wide hard transition that is supposed to happen multiple times a day with model data only. Or it could have been a relatively small change in cloud coverage that wasn’t smoothed out properly, or it could just be a bug. The point is, to conclude that the clouds are coming from METAR, they should match METAR consistently.
The fact that the clouds quite often don’t match METAR is a pretty big deal. And it’s not like it’s just off a bit, but quite dramatically. I’ve seen “clear skies” in the METAR string yet overcast as far as the eyes can see. I’ve seen “overcast” in the METAR string yet not a single cloud in the sky. You can’t just ignore that fact, or the fact that many of the people who claim the clouds = METAR also complain about the clouds not matching METAR. That’s a much bigger piece of evidence then seeing clouds act weird every once in a while, especially since the clouds can also act weird in Meteoblue-only areas, such as the middle of the ocean where you see all the behavior you guys use as evidence of METAR (circular storms, sudden shifts, low density cumulus etc).
Those changes/swtiches in clouds appeard in su7. And that specific update implemented METAR.
I hope we can have the pre su7 weather back as an option. There i know only winds, temp and pressure disturbed. Nothing else. I would most preffer release weather because if we want METAR disturb then the post su7 weather are still there to choose.
Jorg Neumann: I don’t know if you noticed, it works really well in the simulator, but we couldn’t get it to work as well in the world map, so we left the world map the same, but the goal is to get that working as well in the coming months. We tried, and tried, and tried and it doesn’t yet look right. It doesn’t really matter, but it’s one of these details that we want to get right.
Agree, METAR bubbles will never look right in the world map. Still we are forced to accept them in the simulator.
If he thinks it works well in the simulator he needs his head looking at. Does he actually do point to point flights using live weather? I doubt it.
If they can have worldmap same then it means they still have the old weather system active? Why can’t we get that weather model as an optional feature then in the simulator as well? Works so much better. In the video you did it shows how static it is with METAR.
Weather don’t behave like that. Looks like clouds goes in a loop and pulsating and stuff and stays fixed in place.
And in that time frame you recorded it had several hard transitions that can’t be the 6 hour updates. It updates much more frequent than every 6 hours.
I want my fluid weather system back again
The forecast for KUTS on Sat morning. Sunday night will be overcast.
Checking again on Sunday and still predicting overcast overnight.
Night time arrives and the forecast holds.
The cloud map shows > 95 % coverage between 18:00 and 21:00
But the sim has other ideas.
Top down view contains your regular patchwork of METAR cloud circles.
39 hours, 6 weather updates all containing overcast for that time and the sim refuses to show it properly.
Oh great and now there’s this new product, FSi Panel, which finds an airport anywhere in the World that exactly matches some weather criteria, based on METAR. If MS ever goes back to not forcing weather to METAR we’ll have another bunch of screamers. It would be best if the METARs in MSFS were made based on the MSFS weather, and not from the RW reports, and that would be the source of data for FSi Panel. But what are the odds that both developers are visionary enough, when even MS shut down theirs.