In my opinion as well. But the opinions varies as much as the weather
I totally agree with you.
I think the reason why people (including myself) moan about the weather not matching METAR is the fact that when they added METAR into the depiction model, it negatively affected the global weather model, with bubbles of weather, clouds looking wrong, visibility suddenly decreasing, etc, etc.
It is, therefore, galling that the after all these sacrifices, the weather doesn’t match the METAR anyway, rendering it essentially useless in terms of it being used to generate this weather in-sim. If we have to have the negatives of using it we might as well actually have the weather matching it.
In my example yesterday at Gibraltar, the METAR indicated overcast at 700ft, and various forecasts, including meteoblue, which, along with METAR, is supposed to be used to generate the weather in-sim, also indicated low overcast and poor visibility, whilst the sim showed essential unlimited vis and clear skies.
Why was this the weather in sim? What data was being used to generate these clear skies, it was not the METAR nor meteoblue.
We want to know what the data source is so we can use it for flight planning.
But if the model is going to be negatively affected by METAR, I want the weather to at least correspond with that, at the moment it is a mess, and the integration of METARs into the weather model, especially for cloud and visibility depiction, serves no purpose. It ruins the look of the weather and the weather doesn’t even match the METAR anyway.
I agree…I really think they can fix this, but for some reason it has been a big secret on why nothing has happened. Or,if it is happening, how about an update other than “bug logged” it has said that for almost two years. I have no doubt they have the talent and the brain power to figure this out. I just think we need an update, and not just a “oh we are working on it” I mean a real update.
Yes you are correct the metar doesn’t match the weather anyway, and in the real world, guess what, it doesn’t either except for that exact moment in time when the reading was taken. I understand why it’s being used, and it’s a good baseline, but it is just one of many pieces when looking at the weather and what it’s really doing. It has become predictable in the sim on what the weather will be…if you want a real challenge, especially on a crumy day, go with a forecast based model. You never know what you might get. I have seen forecasts calling for murder and mahem with sigmets out for severe turbulence and embedded thunderstorms only to experience mostly smooth conditions. The point is, you never really know what the weather is going to do…that’s the challenge and what goes into making good decisions about whether or not you are going to make your chosen flight. If the simmer wants a real life experience, the forecast model that is unpredictable is the only way to go.
Oh and VATSIM…all you really need is a correct altimeter setting, visibility, wind direction and speed accuracy for proper runway assignement in the terminal area. Beyond that, the clouds, rain, and other visual immersion items are a bonus. Obviously the more accurate they are the better your experience, but let’s be honest…none of us really have the most accurate weather depiction in MSFS. I have seen many metars in the sim that don’t produce the correct visibility, precip, or clouds and everyone still flies on VATSIM. The bottom line in my opinion is Asobo and MS need to improve the clouds and the associated weather that comes with them. If they do that, I can’t see many people complaining anymore.
100% in agreement with everything here.
I’ve been here since launch checking the forum. But correct me if i’m totally wrong here but we as users has been confused where to find the actual source in use since release? That were my issue also at release when MSFS were released 2020. None has actually said what source were in use more than a multiforecast model from MB i think it is NEMS 30 model. But if it’s a multimodel in use it can’t be only NEMS 30 right? Because a Multimodel uses many different models to find the most accurate model for that specific location. And using a Multimodel makes it impossible to find the source because we as users don’t know what model that are in use at that specific moment/location or whatever. That made the community to beg for METAR or other sources to be able to find a match? Because we as users needs to find a match because we are used to find a match of weather in flight simulators.
The problem i see is this:
We as users are good at telling how a feature should be but we are bad at explaining how to make that happen.
I have a perfect example:
Users says turbulence is overdone but do those users also explain in detail why it’s overdone and how to make it better? And what does “overdone” really mean? Is that really good feedback?
I was with you until here. Show me the -12 hr and -6 hr model runs and tell me how it’s going to reconcile with what’s actually happening without producing major discontinuities or overly-large areas of denial with poor weather. To be fair, some weather can be depicted with models - like winds and temps aloft, but anything less than synoptic scale is a no-go, especially when dealing with areas of convection.
The problem is everybody is focusing on METAR - that’s where these conversations inevitably turn, however it’s not evident that METAR has been driving anything other than surface wind and visibility at airports for a while now. METAR is not sufficient for large-scale weather generation, but neither are forecast models.
We need to go even more granular, using satellite and radar, where available, to drive weather generation and animation.
It’s not that hard to understand because Asobo said they introduced METAR for visuals of weather in SU7. Ever since su7 the weather has been generic,
“Not hard to understand?” Prove to me that METAR are still being used for cloud generation. I’ve provided many examples that indicate otherwise.
I know there is other posts that have shown that already. But my point were that Asobo mentioned they introduced the METAR for visuals in su7 and since then the weather has been generic. It may be other causes for that but we can’t report things that we don’t know anything about. Thats why we report that the METAR implementation caused the weather to get worse. It’s Asobo that needs to find the reason for that if it’s not the METAR that causes the issue.
Live-weather is a secret. We as users can only speculate.
Ok, so we’ve agreed that METAR is not currently the problem. That was left behind many months ago. The accuracy since then is demonstrably better.
Why do we keep falling back to that? Because the graphical representation also changed? Why can’t we focus on fixing that instead of derailing the thread every time back to a problem that no longer exists?
Is it because METAR is the single most common weather word that all simmers, regardless of experience level seem to know? A common ground with which to pile all grievances? The only thing visible to us in the sim because no other forecast or observational tools exist?
I think using real time nexrad would be a good idea. Not sure if they could do that, but even a radar summary chart/mosaic or something that depicted areas of precipitation would be great to integrate into the cloud generation. With radar and satellite imagery I can’t understand why they couldn’t draw clouds based on the layers the radar and satellite indicated. I know from that they can obtain at least the precipitation tops, and with satellite imagery I am sure they could estimate those layers as well. Maybe that’s way to complicated.
Absolutely agree. We can get it down to about a 10 minute resolution and distribution in the US, but I’m not sure about the rest of the world - speculating that this exact thing may be part of the complexity - integrating this in one region and that in another.
Or maybe Asobo just needs to explain everything pertaining to “what are they thinking and doing regarding live weather”…that might solve all the speculation.
Yes, absolutely agree. So much speculation and trial and error. I do think it’s important, however, that we end-users get together and establish a common understanding of the language around weather and the operational ramifications of any weather changes. The old “you don’t know what you don’t know” issue plagues us all to varying degrees - which could be alleviated by some deeper insight from the Asobo team.
There has been a MEGATON of effort to find out reasons for the shortcomings of Live Weather.
As the inner mechanism of Live weather generation is mostly unknown that has led to many assumptions, not always helpful in the course of the discussion and in finding a common goal.
My conclusion is that it makes little sense to try to explain to developers, who are probably knees deep in code every single day, how their system works.
I hope that we don’t need to get into those topics again once MSFS2024 is released, but should it be necessary, my hope would be that we users combine efforts in reporting and documenting shortcomings with comparisons to reality (a picture is worth a thousand words) where the weather simulation is not
- accurate or at least reasonable
- visually appealing and
- meteorologically and physically plausible
and leave it to the developers to figure out how to resolve the respective issue.
In the end it comes down to how many users can be convinced to vote, and it has become my honest belief that most users won’t bother with lengthy, repetitive discussions of probable causes or technical issues.
I think the issue is the new mb model we have since su7 that tries to be more accurate but gets less varied.
But as i’ve said, nobody knows for sure except those people that codes live-weather for MSFS.
I’m seeing the word “integrate” this and that into the MSFS weather system thrown around a lot in this thread. While sounding smart and doable, it’s in fact the fundamental problem in producing the weather analyses.
Combining various types of observations with the pre-existing approximation (forecast) from a previous analysis time, consistently with the governing equations, is being done at only a few leading weather centers. A complete weather model and a lot of computing power is needed, and only the observation data that the model can “understand” can be used. Certainly not SCT or BKN from an airport report.
Forcing the weather system in MSFS to a particular observation at one moment in time and a micro location can only produce a chimera of a weather, and it is understandable that Meteoblue can’t and in fact even won’t go there.
If “integration” is left to software engineers and game designers to satisfy user feedback, we’ll stay in this limbo forever.
Thus, there are quite literally two perfect options, each for its kind of users. A model-based weather system, similar to the original one and incrementally improving in tact with the overall progress of the numerical weather prediction technology, or a static weather source based system with always perfect match to the selected static source.
Nothing is “static,” that’s a misnomer that needs to be quashed right out of the gate. It’s not “static” right now, and if they figured out how to correctly render the various cloud types, y’all wouldn’t even notice. We keep going back to this notion that it’s not dynamic and it completely is. It just doesn’t have enough stratus and cirrus, and the CBs are weak.
But there will always be inputs They will either be:
Farther apart, having more leisure to let the weather do whatever the forecast guesses it’s going to do between updates, but leading to more inaccuracy and large discontinuities the farther one strays from the model update. This will almost certainly negate any external tools used to make decisions.
Or closer together, having weather do whatever it actually does, but forcing the renderer to be more on top of blending, animating, and filling in gaps to make it aesthetically pleasing. This will allow the use of most external tools.
Don’t just take my word for it. You have to look at a model run right as it’s published, then look at the actual observed weather 6 or 12 hours later (depending on which model you use). The difference between the model and observed is the error at which planning and in-flight decision-making tools will be unusable and the discontinuity that will have to be rectified when the new model is injected. Until folks do this and explain how to make it work from a flight planning or aeronautical decision making perspective, and stop handwaving the discontinuity it will create, their stance doesn’t hold water. Unless, of course, you’re okay with removing a huge part of what aviators do.
Simple fact is that the weather looks much more generic and simple now compared to what i bought 2020.
I hope they can restore that. If not, i will not buy 2024 when i know that version also can get worse over time.