3 minutes ago, VFXSimmer said:

My understanding is that the INTENT is for Meteoblue to incorporate Metars as additional sample points into their worldwide prediction model.  That is NOT yet what we have in SU7.  The last Q&A said this is still scheduled for next year.

MS/Asobo has been pretty clear that the path towards the ultimate weather model is a multi-step process and this is only part 1.  Until the point that the metar data is helping to drive the entire prediction model there will always be a inconsistency between the two. Once the METAR data is plugged directly into the global weather prediction, the pops and inconsistencies will go away.  It’s really the only way to do so.

 

There’s no reason to implement anything until it’s working as intended.  Or, go the DX12 route and have it as an opt-in beta.  That way they can test part 1,2,3 etc, until it’s at a point where it can be pushed as a mandatory update.  Cloud/haze pop-ins, msl instead of asl metar cloud injections, and no overcast are pretty big regressions from SU6 weather.

13 hours ago, Flic1 said:

Why does it matter if it exactly fits the real-world?

Because of flying online it's needed. And many would actually like at the most 1 hour old data, not 12 hour old forecast data. Not having a flightsim that incorporates METAR weather which governs airfield operations is massively ridiculous. Just because ASOBO needs to refine the interpolation between the forecast model and the METAR data, you just want to knee jerk it right out of the Sim. Give it a rest and let them work on it. 

Edited November 20, 20214 yr by B777ER

33 minutes ago, VFXSimmer said:

That is NOT yet what we have in SU7.  The last Q&A said this is still scheduled for next year.

No!

https://www.twitch.tv/videos/1208240572?t=01h07m49s

33 minutes ago, VFXSimmer said:

I think it would be best if we avoid the knee-jerk “sky is falling” reaction every step along the way (pardon the pun).

METAR implementation is in SU7 (Twitch link), late 2022 is something vague that "will blow us away" (YT embedded link). I think it would be best if we stay realistic.

 

 

Edited November 20, 20214 yr by Tom_L

9 hours ago, Dominique_K said:

The ultimate goal is certainly not to have what we had in P3D. The Asobo weather was lacking but it was still way better than that. 

 

P3D has its faults for sure but with HiFi weather engine and the interpolation they had with it, it was much more accurate than meteoblue wx in MSFS. So is the "ultimate goal" as you put it to have just pretty clouds and westher that can be up to 12hr old? 

I just did a flight out of Bob Hope  (KBUR) because I had not had any issues like some of you are having .

Here is the current metar :

KBUR	201853Z VRB05KT 4SM HZ CLR 16/09 A3009 RMK AO2 SLP180 T01610089

This what I observed :

I thought it was pretty close to the observed weather .

John

 

19 minutes ago, Tom_L said:

that "will blow us away"

Standard Neumanish. We already have been blown away 13 times, maybe time to let us glue the pieces together. 

4 hours ago, sightseer said:

What's wrong with a simulated METAR?  Why does it need to match real world 100%?  .

Wow, I think I can speak for many when I say I hope Asobo laugh at this and then move on. Good grief.... 

3 hours ago, Dominique_K said:

I suppose you are right. This hot and interesting gentlemen's discussion reflects two visions :

- one which gives priority to the replication of a flight in the exact weather conditions and  demands METAR

- another which is satisfied by flying in convincing weather conditions, which was until now the heart of the Asobo engine, because it gives things like moving fronts that we had not seen before and smooth transitions.

The trigger was the quite awful implementation of the weather in SU7 under the pretence of injecting METAR. And, in the background, lies the inability so far of Asobo to integrate a decent planning tool in their sim

So maybe the solution is giving some kind of options but ASobo cannot continue to ruin their earlier work. Or rather they can but I hope they don't.  

It can be boiled down to those that want real weather and those that want fake fantasy land weather that don't fly online or could care less about simulating real world conditions. 

3 hours ago, regis9 said:

I see you're a fellow Nova Scotian so likely very much aware of how much our actual weather can differ from modelled weather 😉

From reading this thread clearly some are happy with a non-metar weather solution but frankly I don't get how that's anything but a major step backwards in realism...I want the real world conditions in the sim.  That makes it possible to use real world weather resources to plan your flight.  As noted here, so many times I've planned a VFR flight using Skyvector only to load the sim and find rain, low clouds etc.  Even for IFR flights it's annoying to plan for a certain approach based on real world data only to get to the airport and find the winds completely different from what you had planned and from what the current METAR is showing.

I do enjoy flying in the conditions I can see out my window, especially in the winter when I do most of my simming and the weather around here is pretty bad at times and challenging to fly in.

Indeed, I do. In our area, not even the observed weather matches what we actually experience because there are mountains between us and the weather station. If we have rain coming our of low-hanging clouds, the weather radar still shows clear skies 😄

There is no question that METARs aren't the perfect source for weather information. I currently fly a lot in Northern Manitoba, where many METAR stations are automated. They give you correct temperature and air pressure, but they do not observe the conditions (cloud coverage, precipitation). Until SU6, I usually had the choice between Asobo's omnipresent thunderstorm, or always flying with clear skies using REX's METAR-based weather 🙂

I think Asobo is on the right path, but they still need to put more work into the weather engine. They need to enable us to do flight planning with real-world weather information, which usually consists of METARs and/or something like Foreflight or Skyvector (in the US). I am quite optimistic that this will eventually be the case.

Peter

58 minutes ago, Fiorentoni said:

It's the wrong way trying to push real live data INTO the sim, this won't work since there is no adequate real life data; METARs are not "live" either since they are a mere description of something that could look different from location to location (e.g. scattered or broken clouds) AND most importantly METARs can be almost an hour old. That's not "live" weather in any way.

So instead they should focus on their prediction model which creates are similar weather to real life (though not identical) and then create in-game METARs from the in-game weather, which can be pushed OUT of the sim and then be used by e.g. Simbrief or whatever. This is the only way to have correct METARs worldwide AND have no transitions and other restrictions of pushing real life METAR data into the sim.

Every 6 minutes a speci can come out.  That's very up to date.  If there is not a significant change in conditions (related to an aircraft flying a visual vs instrument approach, convection, VFR/IFR etc), then no speci is needed and the hourly metar is acceptable.  SCT040 to SCT080 is not significant - and in the sim it won't matter either.  FEW008 to OVC008 is a change from VFR to IFR, and will generate a speci - and msfs should reflect this important change in conditions.

14 minutes ago, johnbow72 said:

I just did a flight out of Bob Hope  (KBUR) because I had not had any issues like some of you are having .

Here is the current metar :

KBUR	201853Z VRB05KT 4SM HZ CLR 16/09 A3009 RMK AO2 SLP180 T01610089

This what I observed :

I thought it was pretty close to the observed weather .

John

 

You thought 4 miles and clear skies looks like your screens?  CLR doesn't look like cumulus clouds directly over the sensor.

16 minutes ago, johnbow72 said:

I just did a flight out of Bob Hope  (KBUR) because I had not had any issues like some of you are having .

Here is the current metar :

KBUR	201853Z VRB05KT 4SM HZ CLR 16/09 A3009 RMK AO2 SLP180 T01610089

This what I observed :

I thought it was pretty close to the observed weather .

John

 

That is not clear skies!  Haze is not formed clouds!

Here is KDAL Dallas Love Field today when its metar was Clear Skies, few at 25,000, 10 miles visibility.  This captured at 1.5 miles off the end of the runway on approach.  The runway and tower both just beginning to come into view.   



Temperature was 20°C, Dewpoint 11°C, = Relative Humidity of 56.2%.
 

Edited November 20, 20214 yr by fppilot

I could be wrong, but I assume that the METAR implementation as of now isn't the intended final form of it. I don't have MSFS yet, but I've been following and I agree that from the reports the current implementation seems pretty bad and reminiscent of FSX era live weather and EA in P3D right now with the sudden transitions. That's by no means a desirable situation, especially since the weather is supposed to be one of the main selling points of MSFS.

However, Hifi have developed a very good weather engine that uses METARS and forecast/interpolation. Perhaps I'm not looking closely enough, but I can't remember ever seeing a sudden transition. If it's even happening, it's subtle enough for me to completely miss it as long as I'm concentrating on flying rather than staring at the weather to wait for it to change. Given that Hifi have accomplished this, I have no doubt that Asobo can pull this off too and deliver a live weather implementation that blends METARs into the predictive model to fine tune it/make it more precise with much better transitions than SU7 introduced.

I do think that much of the frustration stems from the fact that MS/Asobo seem to have made another 'half-baked' update (at least with regards to weather) live instead of properly finishing it first to get into a state in which it doesn't feel like a downgrade or like it broke a lot things that were working before. I'm no technical expert, but, again, I don't believe this is what it will be from now on. It seems like a rather raw implementation along with errors that, frankly, shouldn't have happened (using MSL instead of AGL, locking the 10SM report to always mean exactly 10SM) with proper quality control/testing of the update first. I'm sure this will be improved upon to get to a state in which METAR reports and the predictive model work together much better.

12 minutes ago, fppilot said:

That is not clear skies!  Haze is not formed clouds!

Here is KDAL Dallas Love Field today when its metar was Clear Skies, few at 25,000, 10 miles visibility.  This captured at 1.5 miles off the end of the runway on approach.  The runway and tower both just beginning to come into view.   



Temperature was 20°C, Dewpoint 11°C, = Relative Humidity of 56.2%.
 

I understand now the issue , thanks for the feedback . Here is a current image from KBUR :

37 minutes ago, johnbow72 said:

I understand now the issue , thanks for the feedback . Here is a current image from KBUR :

Looks exactly like how it does in SIM. Except for the clouds. Visibility/haze looks spot on.