Sim Update 7 Live Weather regression - a broken mess

That’s my suspicion, or hope. Like I said, it was completely off yesterday, today it was much better, not unusual that live weather is pretty off for days and then starts working again, not only after FS updates. Apparently it’s pretty different for different people too, so that may be server side issues indeed.

Except, it isn’t everywhere. And that was the whole point of my post, the weird urge to exaggerate every issue everyone has (<- see what I did there?) and spice it up with superlatives currently, instead of being factual, putting these facts together and heading to Zendesk to file a bug report.

We are all hitting these threads because we have more or less the same issues, and if we don’t that has some information value too. All the noisy drama OTOH has no information value and it makes the threads annoying to read for everyone. As if we wouldn’t know that it’s not great having to deal with this stuff.

Yeah this really is difficult, however they are not finished yet trying to blend that together they said. But generally, it will only reach so far and never fully meet the expectations of everyone looking out of the window. I should now, my local airfield is at the coast and the next METAR station is 50km away in the country, guess how often that METAR (or any other kind of weather repor for that mattert) has any resemblance with the actual (RW or not) weather at the airfield. So I guess I’d vote for “realistic” too if I’d have to.

Interesting, I was leaving EDDH roughly around that time too, but to the south and hit the cloud front south of the Elbe river just where I expected it, so I was at least sure it wasn’t the same country-wide generic weather as yesterday. But it sure isn’t quite what it’s supposed to be. I guess we’ll know more in a few days.

2 Likes

How does Active Sky

I don’t get your point. Vatsim uses the correct actual METARS. And other sims can simulate these correct actual METARS as well (like with Active Sky etc.). So it’s a bug coming from MSFS that MSFS has to sort out. Not Vatsim.

1 Like

The thing i mean with METAR and active sky and such things that use METAR to inject weather is that it will only use static local reports and make that into global weather. Maybe they could use many nearby METARs and then make an avarage of those in the future. I don’t know how to make it realistic and dynamic at the same time be good to use in VATSIM.

It would be cool if the sim itself could generate METAR based of the weather that happening in the sim instead. We could then use that in VATSIM instead of real METAR. Then the METAR will always match the weather in the sim and it could be dynamic at the same time.

And this is a new system meant to simulate the weather. We can’t use real METAR if it’s simulated in the sim itself. It should be close to the weather we have.

Also do you think they simulate it correct? I think it has big problems. It’s actually the same problems MSFS facing now when they integrated METAR in the sim. Thats why i thought this meteoblue sytem was much better.

1 Like

Hm…again: i don’t think. I KNOW that they simulate it correctly, except some very complex or special metar-reports. The metars at the departure or arrival airports are simulated acurately in P3D or XP. We get the pressure, winds, visibility and cloud coverage according to the newest metar. Outside of the airports area it will interpolate to the GFS-datas.
You are right that in proximity of the airport it can vary locally a lot within a few kilometers. But i don’t think we need this level of accuracy, because that would be extremely complex.

1 Like

No what we want are METAR’s that correctly represent the conditions in the simulator. People have it backwards. They want the simulator weather to honor the METAR when it should be the other way around.
That is how you get a dynamic and realistic weather simulation and at the same time be usable for flight planning that uses METAR and TAF to get an idea of the weather.

The simulator weather can never 100% match the real weather, because no computer power on the planet CAN simulate the weather 100% corrrectly, with all the known and unknown variables.
As I mentioned in another post I think the best solution for online network and MSFS is to utilize 24 or 48 hour old real-world (it still is real-world) weather data so the MSFS weather engine can better anticipate what will happen next and thus be able to depict weather uninterrupted and fully dynamic.
If the entire VATSIM community is flying on 19/11/2021 weather on the 21/11/2021 what difference does that make. Everyone will be flying within the same weather conditions based on real-world weather. The only “inaccuracy” here is timing, but seeing MD80’s and DC6’s on VATSIM I find it hard to believe that would be a problem for many. After all they value more the functionality and accuracy of the MD80, rather than current real-world flightplan that doesn’t read MD80. All that matters is that the VATSIM flightplan reads MD80 and then you have a correct combination of MD80 on the VATSIM flightplan and the MD80 actually flying in the simulator. At this point it doesn’t matter that the real world flightplan reads A320. Because the action on VATSIM and the simulator are taking place on VATSIM and the simulator and not the real world. The VATSIM ATC will also prefer to honor it as an MD80, because it needs to account for the aircraft specific performance.
I want a functioning weather simulation that best resembles what is happening outside in it’s functionality not it’s timeline.
I don’t want to go back to a flawed system with many unwanted side-effect just for the sake of being synchronous with the real-world weather on the timeline. Timeline and functionality are not the same.

that was a good one :slight_smile: :slight_smile: I think you are not the only one !!!

BTW: maybe you should merge these threads and others weather releated too, otherwise we will write everything twice everywhere !

hmm, may I ask why you then don´t use tools like unreal weather mod (is completely free and easy to install) or REX MSFS2020 - does exactly what you need and not that bad.
Yes the visibility layer is a “problem” but it could have been solved over time without completely ruining the enjoyable weather experience of many users (simmers) because of a “few people who want to fly on VATSIM”.
Do you use the default weather engine for VATSIM in the other SIMS ? (I don’t think so).
For something like that (even as a hardcore simmer like me) I have no understanding.
“Only those who fly on VATSIM with METAR are real SIMMERS” (shouldn’t mean that you agree with this), how often have I heard that (read between the lines).
Why do something evolutionary (since FS2004) and finally progressive (I don’t mean in terms of accuracy) but visual - change so much when there are already ways to do what you want to do (even with this tools or other “realistic” SIMS like P3D or XP11) !
I am also in favor of a very precise and above all, smoothly hybrid system (also with a visibility layer) - but certainly not in that state. (but haze and visiblity option is certainly something we “need”, it gives the SIM the rest of the touch)
But let’s wait and see, I think the girls and guys from Asobo and MS are certainly working hard to fix these bugs, for me and I speak only for myself, the SIM is no longer usable in this state (at the moment), besides really improved things like performance, night lighting, etc.,

I’m up for everyone to be happy and MSFS should of course not only be for “XBOX players” (hate the expression, please forgive me, don’t mean it) but this change should have been an explicit option (to choose it) -
and no offense at all !

That “METAR bubble” may also be due to a little twist: I know that from XP - the METARs from EDDH and EDHI are often quite different - when e.g. EDDH reports " 220750Z 32003KT 9999 2500NW BCFG FEW030 00/M01 Q1024 NOSIG" like this morning, EDHI reports " 220750Z VRB01KT CAVOK 00/M00 Q1024" and in XP this would cause a harsh change of weather when getting closer to Finkenwerder. If you depart E or N, the EDDH METAR would prevail, if you depart W or S, EDHI would be the last METAR the simulator gets. If EDHI reports “CAVOK” (and they often do) the effect would be exactly what you described.

Now this morning I checked the FS WX in EDDH and the weather wasn’t entirely off, there were few clouds with bottoms between 0 and maybe 1500’ but no patches of fog as reported (I also checked before sunrise with a similar METAR) and visibility wasn’t reduced at all in all directions (so nothing like “2500NW”).

Flying to EDHK from there was a surprise: The METAR said " 20850Z 29005KT CAVOK 03/01 Q1025" and this is what the webcam at the Holtenau lock was showing:

I checked the webcam because what FS showed wasn’t “CAVOK”, but it didn’t really match the screenshot either, that was more like flat puffy SCT. However, it was much closer to reality than the METAR.

Things like the issue with multiple stations in short distance or the quality of METARs from smaller airports (this is not the first time I observe things like this) make me think that adding METAR-based weather wasn’t a great idea.

In this example, both issues coincide quite badly. People want to include METAR reports because they think they are more accurate and current. But the reality seems to be that METARs may be more current and accurate at major airports, and smaller airports with equipment fallen in disrepair, only measuring actual wind data and putting placeholders in the report to make it a valid METAR are corrupting this advantage big time. A simulation-based dynamic weather may not be as current and not always 100% spot-on but overall much closer to reality, and it doesn’t force Asobo to hack in hair-rising kludges to blend together things that don’t belong together.

2 Likes

You don’t even have to compare it with METAR or the real weather you see outside. Just having the lowest cloud bases all on ground level tells you, that there is a bug in the system, regardless of where they get the information.

2 Likes

Yes, I saw the same in MSFS today. I also checked LOWS, and it wasn’t that bad either - aside from a missing low overcast layer which I have yet to see. It’s trending in the right direction, but I fear we have a long journey ahead when it comes to the accuracy of cloud coverage and visbility,.

Great example for the difficulties of a METAR based weather rendering: CAVOK means no significant cloud coverage BELOW 5000ft AGL, and your webcam picture illustrates that where many simmers would assume blue skies there can very well be several cloud layers, even overcast.

1 Like

I’m not sure if anyone else has the same experience, but the weather seems to be looking better today.

For example, I can now see overcast again:

1 Like

Are clouds rendering at the correct altitude now?

They definitely seem to be more accurate in terms of their overall altitude and coverage, haven’t tested thoroughly enough though to say it’s definitely fixed. Seems to be improved though compared to what I was seeing on SU7 release day.

3 Likes


Just done a departure out of Manchester and I have to say it looked very nice, and pretty accurate to the current conditions around this area.

2 Likes

Any issues with haze/fog pop-ins?

None so far, but that could be that I’ve not flown past any areas with such weather. It’s pretty clear across England today, I will keep an eye out though :slight_smile:

The question is of course if the whole weather pattern simply consists of these cumulus clouds of different degrees of coverage and at different heights and same extents - then we have METAR based engine but with corrected cloud heights (AGL) ! I will start a test myself and report back !!!
Thanks anyway!

1 Like

well the “METAR” test:
METAR: LOWW 221620Z 32009KT 9999 -RA FEW012 BKN050 06/05 Q1020 NOSIG

so now this is a great example WHY METAR is not suitable (alone) as an “ACCURATE” weather engine (apart from the same values you need for e.g. VATSIM flying), it doesn’t get any better than this example !
In reality we have a closed cloud cover, even if METAR says - BKN5000 and 1200FEW - it is mostly in nature (except for exceptions) a closed cloud cover !!! (this is also the case in reality because cloud expansion is never correctly represented by any values. Up to LOXT (I drove the route in real on the way home) - it was a closed cloud cover and pretty miserable visibility!

this is what it looks like in the sim, should be light rain:

It’s the same weather, I just adjusted the time manually so that you can see the clouds!

and this is exactly what I have been trying to explain to people since FSX and various weather engines came out, do it again here (because FS2004 and AS6.5 did it pretty well)
The advantage of MSFS is of course that the clouds are really “3D” and not bitmap based.
The weather is displayed correctly according to the METAR, but does not look “real” (or accurate) because the BKN5000 and the FEW1200 are probably closely meshed and it is not the same as in nature. (in this case, presumably a closed cloud cover). in addition, the low visibility is due to the light rain, which is not visible here at all, and nothing is “correct” any more - except that the images look quite “dramatic”.
In any case, rain with closed cloud cover looks different!
First of all, it was raining at the airport (in real) and I didn’t have rain because there was no cloud exactly above the runway (BKN5000) !
BUT something has definitely been changed, it has definitely become a little better !

It is just a fact that such a weather (if it was accurate) looked more “realistic” before SU7 and that´s it!

3 Likes

Yes I would on the whole probable agree here, it’s still not great, but something seems to have changed indeed

…I really hope you are joking, right?

I cannot believe that. LOL. I hope they won’t be that stupid. That’s really not gonna happen, for sure. Live weather means: 20-30 min delay is ok and will be tolerated because it’s what people are used in the actual sims. I think the best for you would be to stop flying on Vatsim if you want historical weather. I cannot imagine the number sof users lost if Vatsim would use a 2 day-old-weather. I guess it would be the end of that network.

1 Like