They’re working on allowing the API to be publicly available. I think we already have a wishlist about this a while back.
3 Likes
What is your source…
I only know the “Under Investigation” Status of the last Feedback Snapshot and Jörg’s respond in the Q&A sessions, which were never actually an answer, but rather shirked an answer… Saying “they are working on it” would be a mile-long step forward
Working on it can mean a million things though. Under investigation to look for a possible viability can still be considered “working on it”. English is my third language so it might be a poor choice of words from my end.
I consider anything that they have been aware of but hasn’t made any announcement as “no” is the same as “working on it”.
they say we’ve been working on it for 3 years, it’s no longer possible to take it seriously 3 years it would have already been done, today we are given games or simulations in Kit, MSFS 2024 won’t have a Weather Radar API because they’ll still have that ■■■■ blue weather
2 Likes
Asobo, can you guys give us an update on weather radar progress? As you’ll know, it’s by far the most-requested feature. You guys have an amazing platform, but as a real-world pilot who uses a weather radar every flight, the lack of it in MSFS stands out as the platform’s only major deficiency. Many thanks!
6 Likes
Come on vote VOTE…
I want to see at least 99999999999999999999 votes till Friday to finally get some nice weather radar for my adventurous thrilling Alaskan Fenix bushflights 
When’s the last time Asobo acknowledged the existence of this issue? This thread has tens of thousands of views and thousands of votes 
9 Likes
Please also integrate the weather for a third provider
1 Like
The Microsoft game that was improperly called Flight Simulator in its first version in 1982 is taking on the authentic and legitimate meaning of its name. Finally we start talking about “simulator”. It seems clear that in the flight simulator there are subsystems that represent distinct simulations of dynamic physical events, starting from the geoid with everything that it includes, i.e. nature, which with the achievements of the human being also influence what covers planet earth : the atmosphere. This is an envelope of gas retained both by the force of gravity and by the magnetic field (which counteracts the solar wind which would otherwise sweep it away) and participates mostly in its rotation. The gas that makes up the atmosphere has a chemical composition that varies based on altitude and location and has a rather complex structure. Based on this premise, it is understood that to obtain a flight simulator the “environment” must be created. To return to the topic, simulating the weather actually means building a weather simulator with these characteristics https://www.youtube.com/watch?v=mY2ozPHn0w4 and connecting it inextricably to the geoid (another simulator). With this premise, we begin to lay the foundations of infinite subsystems to start working on a flight simulator. MSFS2024 represents the starting point of this concept. Most likely we will see the 3D weather in the new simulator, where the philosophy is to simulate and not play which are two different concepts.
5 Likes
I’d love that! The problem is lack of granular data in the various micro and mesoscale weather systems and climates around the world. The soundings, observations, and radar data in the lower atmosphere are few and far between in enough places that a full dynamic, global weather simulation (and all the chaotic interactions on those scales that are required to make it accurate across the world) isn’t really possible at this time.
We can model the synoptic scale (and sort of down to the mesoscale in very short-range modeling, and only in some places) at this time, but accurately getting it smaller than that is a long way off.
This is literally the root of the issue that is dividing the camps that are clamoring for better weather.
3 Likes
Yes, we either have already known data or future predicted data. We have nothing inbetween. The inbetween data is missing to make the 100% accurate weather simulation in flight simulators we all want. I wonder when we will get that data?
I think we all need to accept plausible weather in flight simulators.
Then what type of data that should be used we have different opinions of.
The only thing i know for sure is that predicted and observed weather is completely different data. Observed is fixed data and will always remain the same over time and predicted is simulated that can change over time.
Me personally preffer an atmosphere that is simulated but the weather thats happening in the sim may be different than the real weather. That makes it flow without any kind of transitions. And then the simulated weather thats happening in the sim should be observed and reported like the real weather.
Data from observations is a big part of what goes into predictive modeling. This is incredibly important because what’s being observed is part of the big picture of what causes subsequent weather. Things change, they diverge from the model, it goes on to change what’s predicted in the next model run, and the cycle continues to go that way, in many cases further diverging from the baseline. That’s the crux of the misunderstanding I spoke of.
The spatiotemporal gaps in data means weather modeled at any level lower than synoptic will not be plausible. And it will need to be re-injected anyway, when things suddenly don’t match up. So much for the notion of “flow.” This just furthers my point.
Agree, but the predictive model is still a completely different thing than an observation.
Observation data is known data and predictive data are unknown data.
Known data is fixed values and can’t be dynamic. Unknown data are unfixed data that can be dynamic. The real weather is dynamic and are able to change.
Irl observations are made from the always dynamic weather. The real weather are not instantly changed because an observation report different things. The observation needs to be updated instead.
This is a straw argument because all of the data you mention are subject to the same process, whether short-term or long-term, static or dynamic. Nobody who correctly understands the issue is asking for “static” weather based on only one input, and the sim is not currently doing that, anyway, for clouds. I do not know why this keeps on being trotted out as if it is the case. However, the difference in application in the sim is the interval in which it is updated, the divergence that occurs from reality in between updates (model runs), and the degree to which corrections must be made from that divergence.
I understand you’re advocating for smoother, realistic aesthetics, but the argument is framed incorrectly and misses some major inconsistencies that would be (actually were in the distant past) revealed using that method. The problem is that it’s not taking into account that model runs are few and far between, relatively, and all sorts of divergence happens (the observations from which go into the next model refresh) in the ensuing time. You can just “let it go” and build to what the model thinks it was going to do at the time the model was run, however 3, 6, and 12 hours from now, it’s going to diverge increasingly, to the point that, without a refresh, it’s unusable and unrealistic. Storms and fronts, for instance, will be hundreds of miles off, or exist/not exist entirely, the resultant winds will be off, and coverage and ceilings will be completely off. The only way to make aeronautical decisions would be to dig back for a 12-24 hour old forecast and use that. Forget tying it in to any reasonably current satellite, radar, or surface observations, just fly based on a very old forecast, which is a dangerously laughable prospect in aviation.
Contrast that to the fact that some areas of the world have what’s known as rapid refresh that will spit out updated models at a fairly small interval (even hourly), and despite not necessarily being as comprehensive or detailed as the larger runs, they are useable. Asobo could could use those and decrease the update interval to more closely match what’s happening in those parts of the world.
That said, advocating for smoother animation is absolutely something I can get on board with, but I will absolutely oppose the disposal of more granular reality in the quest for that, especially when it’s based on falsities and repeated straw arguments. We could, instead, focus our efforts on asking for better processing and animation between the shorter model runs/refreshes, which would still undergo the same conventions as the longer runs, except the output would be closer to what’s really happening in the 12 hours between runs. Yes, you’d have at least one (relatively minor) “hitch” per hour in which something might suddenly appear or dissipate as the model is refreshed (versus a major hitch in the longer interval), but again, that’s fairly close to how it works in the real world, and could be overcome with proper animation/rendering, whatever you want to call it.
1 Like
My own thoughts is that the more accurate to real world weather we needs it to be the more static it feels because the only accurate data we have is fixed data. The weather needs to stay as those fixed values, if not it becomes inaccurate to those fixed values. I know the weather in the sim is dynamic but it feels more static/generic now compared to what i bought 2020. I’m not expecting them to revert the sim weather. I hope they show some real progress of weather soon though. But i’m not expecting that either to happen before MSFS 2024 is released.
1 Like
Hey let me intervene shortly and calibrate this discussion because it is extremely interesting but can’t converge.
There’s an opinion which Perry has been representing for a long time and many subscribe to, that it is OK to have a weather simulation in MSFS that is perhaps not entirely matching the observation at all time and locations, but is on the other hand able to benefit from the consistency between the atmospheric variables which only the weather models with full set of physics representing equations can ensure. In this world we only desire Microsoft to improve the fidelity of the weather representation in the simulator by utilizing all the available data from the weather models, for example visibility.
The other opinion which CharlieFox00 represents and defends with exquisite arguments and logic is that MSFS should faithfully render the weather at the observation locations, and somehow ensure the transitions in time and space by some kind of blending. While theoretically possible, this approach would require many more data assimilation cycles in the weather model than even Microsoft can afford and than it is practical. Without it, the mismatch of the real observation the weather picture shall be tied to, and the surrounding modelled data, is causing unnatural artifacts. Pilots who value accuracy at the airport over continuity can be totally fine with this, and this is also OK.
The bottom line is, there is no one perfect solution, and at least two imperfect ones that meet many users’ expectations.
3 Likes
I think it’s possible and practical - they’re already doing it much better than it was when the sim first came out!
But you’re correct that it’s still not accurate to the micro/mesoscale, and I don’t know if it ever will be, unless they institute an extremely rapid refresh based on radar and satellite data (the most accurate and rapidly-refreshed combination of observations), kind of outside and maybe as a togglable addition to the model runs they seem to be relying on currently.
My theory to that end is that you could take the current radar data and blend them to when the next radar refresh occurs. There will always be a minimum lag of whatever the cycle refresh is (6-10 minutes, at minimum), plus the time it took for the radar data to be scanned and sent, so another 6-10 minutes, plus processing time. This could theoretically get down to 15-25 minutes from reality, which might be the closest we can ever expect.
Speed, intensity, and location of echoes are knowns, then you just have to dedicate processing power to blending and rendering between the last update and the next. Allow 3 or so minutes to blend and I think it can be done. It would be marginally better than it is now, temporally, but would immensely improve spatial accuracy. Just make it a toggle to please all sides.
As far as the “old way” (two years ago), I don’t think there’s enough data (and certainly we can’t produce new data to this effect) about what reality it was based on at that time. All we could see is that the end result clearly did not match real-time observations in the majority of examples. I mean it was wildly off in many cases. The first fix for this was kludgy, but like I said above, it’s improved since then.
But my point remains - no matter what, the 12 hour modeling or 1 hour modeling are both subject to the same limitations in terms of output and depiction. They both must eventually refresh, and it’s really that and processing power to accomplish it that are in question. We don’t know exactly where the bounds were and what kind of looping/blending was going on to get from T+0 to T+12 because to my knowledge, nearly nobody sat and followed a storm through the 12-hour or 6-hour refresh cycle and recorded whether it matched observed reality (and subsequent divergence from the model) anywhere in the area it affected. Instead, it mostly manifested in the end-result of complaints from users checking results against observations at the aerodrome. The middle, in between airport reporting stations, got lost in the shuffle.
So to be fair, the output at the airport is only a small part of my view on this - weather should be regarded much more holistically, as each interface and interaction goes on to spur new ones and are part and parcel of the entire flying experience, the enroute portion definitely included. So much so, that it gets frustrating when the folks on the other side of the discussion (opposite Perrry’s stance) only focus on the airport, which contributes to the breakdown of the discussion and sends us running to our camps.
I’m advocating for the same thing, better rendering of layers, smoother, realistic transitions, but also increased accuracy throughout, and the only way I can see to do that is to use how weather actually behaves and throw it into the sim with a small transition delay, refreshing it every hour instead of every 3, 6, 9, 12, etc.
Tl;Dr - I want better accuracy, I think it’s possible, a toggle is fine, and please don’t categorize me into the camp that only pays attention to airports.
1 Like
Why do we want those old transition systems? I would preffer a weather system that evolves into different conditions completely seamless as they advertised before MSFS were released.
2 Likes
You’re going by the advertising on a sim that was basically released as a beta and has seen, what, 13 updates and improvements since?
Here’s the deal: the rendering changed for the worse. The accuracy changed for the better. Whether the two are directly related is unknown, but theoretically it doesn’t have to be that way, it just takes more work to resolve.
No matter what, it has not, is not, and never will be “seamless.” You just didn’t find the seams. I did, constantly. And they were bad. The seams now aren’t as bad as some people are making them out to be - it’s not static, it’s still dynamic, it still morphs and changes, it’s just missing several types of clouds that should mask and blend that.
Yes, i do.
I think you do that as well ”live-weather” even if both you and i know thats impossible to get 100% all over the world all the time.
For you it may have been improvements for your flight experience, for me it has been the opposite.