I’ve recorded plenty of evidence to the contrary, presented way upthread. The weather portrayed by MB in the sim is often (not always) hours ahead or behind the hours-old forecast model from which the data were sent to MSFS. This creates a discrepancy between the METARS, which are anywhere from a few minutes to an hour-ish old.
The picture I posted comparing sim radar and real world radar in Northern California during an approaching front (the impetus for me being in this thread) clearly showed that the front as drawn in-sim was miles (or hours) ahead of where it was in real life, covering a far too large area. The result was an area of way too intense rain in an area that was reviving little to none at the time. The METARs in this case were “protecting” the various airports from this discrepancy by removing or attenuating the MB precip, which, as verified by real-time radar data, should have not yet arrived. Further testing confirms this to be the case often.
That said, I’m usually okay with the transition to more accurate weather, as provided by the METAR as I approach the airport. It just looks goofy if you ever get a birds-eye view of it (which is often the case with the in-game radar), something better blending could fix.
Kind of a personal, yet irrelevant question, but if we’re comparing measurements, it’s well over 30 years.
Except that it’s not. It’s a fairly accurate depiction of what the most recent real-world weather is in the vicinity of an airport. Where the discrepancy lies is that the MB data are either not granular enough, not accurate in the various layers, or the timing of those is worse than the lag in the METAR (or combinations of all three).
As I offered way upthread, the solution is to either blend them better (this solves the “overriding” issue), or time/granularize the injected MB data better. You’d think that they would build/update the model data around the METAR, use it as a foundational element (rather than a juxtaposition), since that’s giving real-world observation versus a model run.
Well yeah but that’s the point @Perrry and others are making.
Sacrificing believability for the sake of accuracy. I also remember weather not being accurate before SU7 but it also looked much, MUCH better than it looks now. Right now it looks like it’s fixed. Very unnatural, like the weather engine was stripped of its own life. Before SU7 the sim was able to render a somewhat believable Cb clouds for example. Right now I can’t see them anywhere or I see them absolutely everywhere.
There’s very little stratus clouds hanging in the valleys, only unnatural looking fog layers around airports with METAR reports.
I was the one who wanted the METAR implementation as well because since the sim’s beginning I haven’t encountered a single foggy day or night where I fly.
I couldn’t have known they’d bork the weather so much though.
Exactly, the accuracy is more important than the visuals or behaviour. As long as it’s accurate nobody in the flight sim community cares how it looks and behaves it seems.
Well, we have different opinions and i don’t want to change anyones opinion. I hope we can have the old system as an optional feature. But i think thats not possible. And i start to give up on that. It’s sad when i had that at release and suddenly removed. Well, thats how software development works these days, we do not know anymore what we will buy. It can completely change into something else next month.
“Pre su7 i were excited to see how the sky would look like everytime i was flying. Now i know how it looks like before i start it up. Thats the thing i really miss.“
We’re talking about a sim. The weather won’t be 100% 1:1. However, if the sims engine was modeled for one set of data what would be the point of forcing it to use data it’s not designed for. The engine was never meant for METARS to run the show since it was built around a forecasted model which METARS are not. I can’t tell you how many times I’ve seen my local airport reporting clear when in fact there were clouds or high level clouds blanketing the entire area. I know (not that I can test it) that if the only data being injected was MB that there’d be a proper depiction BECAUSE METARS have no reason to report anything roughly above 20,000-25,000 feet.
Sometimes when the weather is goofing up I’ll switch to using REX Skyforce which is only METAR based and while it does provide some better depictions in some cases, it entirely influences and affects the entire visual area making whatever the closest METAR is to my location the same for the entire 40NM mile visual area. I’ve checked this before when looking out my window and seeing clouds about 20 NM away and yet loading up at my local airport about 4 nm away saying clear, all of those clouds are gone. Switching from METARS back to the MSFS weather, those clouds will be there again. I know that MSFS made some improvements so that the ‘bubbles’ or influence of close airports are better honored. Take for example an area like LA. That is a weather depiction mess, in a METAR sense. You have four airports within 20 miles of each other and so the sim has to literally draw mini areas which honor each airports reporting station. This is highly unrealistic. As weather is essentially one thing, the way I look at it would be taking a cloth and then poking holes in it where each reporting station is. This would explain the whole METAR bubble mess. If you’re just using a forecasted source like MB without METARS, you’d end up with just one big dynamic blanket of weather, which I think is what everyone really wants because it’s how the world works outside and is quite apparent in the pre SU7 photos.
Another big headache or issue is that METARS update however often they feel they need to. It’s not 100% locked in to just updating once every hour. This again becomes an issue when you’re talking about immersion or ‘believability’ because as previous developers/METAR based products have had to fight with is how to smooth out the transitions to weather changing or fixing the ‘popping’ of weather. With forecasted data that is 6 hours or more you don’t have that issue. I can sit at an airport for 3 hours and watch the weather grow and build around me. If I had 4 stations around me updating independently from another, there’s going to be ‘popping’ and rough transitions. Developers have tried to minimize that by allowing you to choose the update time up to 60 minutes. But if we’re talking about ‘accuracy’ then you’re going to miss out on that as weather is very dynamic and so again in my opinion, forecasted data wins out on that one.
You said you’ve been simming for over 30 years. As someone relatively close to that number, I can say old habits die hard. Some people get stuck into thinking that the way it’s been done for years and years or in every single preceding sim is the only way and that something new just doesn’t work. Anyone who has been involved with FS not just as a pilot but someone who has tested many different programs trying to chase the most realistic options available out there would agree 10/10 that using METARS over forecasted data is antiquated. I see METARS as a ‘live’ slide show of the weather. I want a live movie of the weather, and that’s what using MB’s forecasted data offers. If I am flying and I get a sudden weather change or wind shift because MB injected their next set of data, I know I don’t have to deal with that for another 6 hours. With METARS I can expect that maximum every hour or however often a station around me updates. To me, that’s unacceptable and breaks immersion as well as having lived with that for over 20 years with all the attempts to fix it or minimize it being largely unsuccessful. A weather environment that is solid for 6 hours versus every 15-45 minutes, as a flier, which one would you want?
Live weather was atrocious after SU7 for about one year. Thankfully it is now back to being reasonable, but boring and artificial.
We had SIM updates, WORLD updates and AIRCRAFT updates, so I sincerely hope we will see a WEATHER AND ATMOSPHERE UPDATE some day that includes everything once promised in the infamous weather feature video without sacrificing accuracy.
Live weather in its current, castrated state has imho become the weakest part of the sim, and with the already announced and scheduled updates in all the other key areas of the sim the discrepancy will some day be evident to even those who don’t care today.
As I have indicated in this post above I am optimistic that Jorg Neumann, head of MSFS, isn’t too excited about the current state of the atmosphere either, but I also think it will take time, for sure far beyond SU12.
Agreed. It’s quite lacking or at least has become predictable, unvaried, and just plain. What’s the point of developing, releasing, and updating all these different aircraft and scenery areas if no one wants to get in the plane and go enjoy the sky. Could get my kicks off of Google Earth or something like that if I wanted to sight see. I’m starting to understand the phrase when people refer to the sim as being a sight seeing simulator now. I’ll still do my flying and enjoy what it offers. But I know it can definitely do a lot more and in my opinion the weather has regressed a lot. But I’ve already explained my thoughts and reasoning in my last post about it. Fingers crossed. We’ll get there. I can’t imagine having one of the head developers who seems passionate about the detail and accuracy of the weather letting the lack of attention go on for much longer, I hope. Seems to be that supporting developers that bring in $$ and bring in new simmers (game sales) is still their focus. They aren’t driving any sales specifically from weather but hopefully they can realize that if more attention gets paid to the core function of the sim that it will also drive and inspire new people to join or to buy different planes which still creates $$. There’s got to be a point where they’ve maxed out all of these areas and hopefully soon enough the only areas really left to work on or improve will be AI/ATC and weather. Let’s just hope that’s not at the tail end of their development agreement.
It’s obvious the engine wasn’t built around METARs, I said as much.
This is a fallacious argument. A “big dynamic blanket of weather” would be one that models everything, everywhere, all the time - you’d set it off and let it go. In reality, it has to be updated, injected with updated data from MB every so often (I believe 4 or 6 hours last time I was in this discussion). That ruins the notion of “dynamic,” because it’s still following updates.
A notional cloud deck gracefully sweeping across the landscape will either suddenly shift, disappear, or reappear, not because there was moisture that was elevated to the lifted condensation level, but because an updated dataset told it that it was wrong and needed to be moved. We had plenty of remorphs prior to su7, they just didn’t surround an airport in a certain radius or happen as frequently. Which brings me back to my point - blend better.
Also, I’ve taken close to 10,000 screenshots and have hundreds of videos of all my flights since pre su1. There is no discernible difference post su6 except the sudden changes in airport radii, which we agree can be improved.
Yes, I proved that, that’s how I first ended up in this thread months ago. The problem is it was the MB model that was unrealistic as proven by real-time radar and METAR.
There are criteria to this. It’s not “whenever they feel they need to.”
True. It can just be old, wrong information, which I’ve proven over and over.
Which can be completely erroneous.
There’s the problem
We want the same thing. The problem is the movie is wrong.
As I’ve said before, I want a weather environment that allows for accurate, useable forecasting and observational tools. Part of aeronautical decision making is using those tools to develop a flight plan and exercise good aeronautical decision making prior to and during the flight. This often comes down to precise timing around the margins of the interface between weather systems, which is where the MB forecast method often fails.
There are three ways to successfully accomplish this:
The sim uses modeled weather that is more accurate (it has been proven that the current model is often not), simply generating weather that matches observations without injecting them.
The sim can eschew observations but generates a complete set of forecast tools tantamount to what the FAA, etc, offers. This would simply make observations and forecasts based on what’s in the sim. I think you’d find the data generated by this would show more inaccuracy than would be considered acceptable.
Continue blending the forecast and observational models (but improve the blending). I’d actually like to see them integrate RW radar, but I’m guessing that’d drive people who believe so strongly in MB’s accuracy even more nuts.
One person’s one hour is the next person’s six hours.
I want one that’s accurate, especially in the terminal environment. I don’t care how long the period is. But all the data I use to execute a flight, based on not only forecasts, but real-world observation of METAR, PIREP, and radar (which all feed back into future forecasts), is more timely than a 6 hour old forecast that spawned a thing that has diverged from reality. Lots can change in the ensuing period. If I plan a VFR, non-FIKI flight that is meant to be near the margins of icy IFR weather, I am using a lot of tools enroute (including the mk 1 eyeball) to ensure that I remain out of that. If I run into unforecast weather and need to divert, I divert, but I’m still using real-time observations (on my own and/or with the help of ATC) to make the diversion decision. A stale MB model isn’t accurate enough or provide enough information to accomplish that.
Don’t get me started on the dynamics of thunderstorms, especially severe ones, which I will never expect the MB forecast model to accurately portray. Aircraft often have to pick their way around those and radar is the best tool for that, forget METAR. Some storms come and go in well less than and hour. Where’s that in the MB model?
That got a chuckle from me. Zero realistic interaction in flight, forget aircraft and systems. No chance of any weather. Haha. There is a lot of difference between Google Earth and the sim, weather aside.
“People are saying…” always love that. The plural of anecdote is not data.
There’s a part of me that thinks that “people” don’t know what real weather looks like and how it interacts on a scale farther than their eye can see or longer than their conceptions can predict. Apply that notion to the sim and we might generate another subset of people that might call that a “weather sight-seeing simulator.”
Well can we all agree that weather needs to be focused on though? Then if they need to blend it better or what they do with it as long as we all will be happy in the end. I really hope they can nail both the fluidness and accuracy at the same time. The fluidness and variety is the most important to me. Because weather is fluid and varied in the real world. And i want them to get the weather in the sim feel like real weather.
They said the weather should be seamless and the weather is not seamless anymore.
Since release they have only focused on the accuracy part of weather. And that feels like an endless task. Will never be 100% accurate all over the world at the same time. The progress of weather has stalled and moved backwards instead of forward.
There’s an option 4, the easiest to implement and can give the most pleasure: tolerating the inaccuracy and just enjoying the holistic weather simulation. With unreasonable expectations out of the picture the developers could also focus on the needed improvements.
You must not of watched Mattius’ from MB’s weather video. Yes in fact the entire global world is modeled when the weather is injected. That’s why I can fly four hours without seeing weather shift or pop. In his video he explains and shows how the modeling works. He drops multiple batteries and from the beginning point that is hour 0. As the batteries scatter and move away those are subsequent hours. Of course hour 0 will be most accurate but there’s a better chance than none that hour 5 will still be accurate for what is happening outside, that’s what model predictability is for. Now, take using METARS. You drop the batters down tracks that only go 45 minutes and then they keep resetting over and over. Sure it’s accurate but is it really that dynamic? It’s forcing an element with a mind of it’s own (weather) to follow specific paths. That’s not how weather works. Even meteorologists get it wrong quite often. That’s what makes it fun and immersive, you don’t know what you’re going to get just like outside. With METARS we’ve constricted the window of being dynamic from 6 hours down to below 60 minutes or less. Imagine going outside looking up into the sky and seeing the weather rapidly transition from clear skies into thunderstorms out of nowhere, you’d think it’d be the end of the world. In the real world we can watch that happen and build around us, METARS don’t allow that because they only show what the data says for that ‘snapshot’ in time. Aka the slideshow. METAR update, next slide. Update, next slide. If somehow METARS could be constantly updated, that would be a better deal but they aren’t.
Next time you’re on final with partly cloudy skies and then all of a sudden it’s 1/2 visibility with 30 knot gusts and rain, let me know how that landing works out. As someone coming from previous sims, that’s exactly what you and I both had to live and deal with.
Injected being the key phrase. It is not modeling all the data that go into weather in the sim, specifically lapse rates and lift (amongst others). At the time, it couldn’t have done orographic lift if it wanted to as those features have only been recently implemented.
Nevertheless, it is still being fed a “this is what might happen” forecast and allowed to run with that for several hours. A blob painted on the map and allowed to become what it’s forecast to become, when it is forecast to become.
What ends up happening and the timing thereof is the issue.
A) in my thousands of hours in the sim, I’ve never dealt with an unexpected shift of that magnitude on final. 10 miles out, maybe, but I don’t consider that final.
And B) as a pilot, I know how to deal with unexpected windshear.
With the MB-only It’s the lack of availability of ADM prior to being on final - knowing it’s unexpectedly generated weather over a wide, inaccurate area.
I want to reiterate that I think that METARs are even too broad of a timescale to be used (though they’re better timescale than MB, just poorer in geographic coverage). I want radar. Weather events are too discrete and dynamic to have MB generate realistic radar signatures at this point in time. Give me all the cells propagating outflow boundaries, kicking up new storms downline and allow me to look at that in a loop to understand the development. Generate airmets, sigmets, convective signmets, center weather advisories. What’s that? Can’t do that at this time? Then it’s not useable.
I still think its just not a well coded engine all together that is kinda plastered together.
Meteoblue dynamic weather and metar sources should be intertwined together niceley and flows but instead in su7+ we got this “we will slap metar in there”
Look ill back not pulling metar from the sim due to the weather was so inaccurate at the airports prior to SU7 other than in western europe, but i just feel like the sim coming to these past reset SU(s) were just slapped together and now theyre finally making everything flow together
Gonna give them some more time on this to see the improvements as it has improved with every update slowly but surely.
I noticed that too when I did a quick load up at my local airport. I had spent most of the day sporadically loading in and out and the weather was generally always the same. Loading up after the new beta this evening and thought to myself when I saw it, where the heck did that come from. Obviously too early to tell anything yet. Plan to do some flights tomorrow and check it out.
Been hopping around to different airports. Had a little trouble trying to find some airports that had some decent layering to them. Have to say I think they’ve made improvements to the layering. Once you get up into the super high atmosphere or just broken or scattered clouds it still looks a little dull. But when you have either multiple layers or overcast and broken you end up with an image like this:
I definitely agree. Despite that though, from what I’ve seen on my end and especially with your screenshot, that looks pretty darn good from what we’ve become used to seeing. It’s encouraging that there’s at least some level of minute changes happening. Lower level clouds look great and seem improved with the layering. Now if they could just sort out the upper level clouds above FL250 or so and fix the density issue with the almost transparent scattered or few puffy clouds along with getting proper cirrus levels and coverage, I’d be a happy camper. Today at my local airport, the METAR would report clear but we had an entire blanket of high level overcast. Not enough to be deemed true overcast but thin enough to be interpreted as clear. If they fix that, I’m happy.
you all need to try and fly in a compact weather system to actually see if the weather engine has had any improvements. Pick a weather front in the world map and see if in game is made of broken CBs or strange artifacts…also how many layers are present..
