A large number of people, myself included, have a need for things to be simplified as much as possible, not to require too much mental effort for something, and for this meteor briefing not to hold you back too much. And then, of course, a finely detailed model that actually looks very real is perfect to help here. And I think you can only say that over and over again. These are great tools, but they can fail or not work at exactly the wrong moment, when you want to rule out a danger.
Podz-Glidz, the Lu-Glidz podcast.
Stories from the cosmos of paragliding. Today with Roger Oechslin. And I am Lucian Haas.
All models are wrong, but some are useful. This aphorism comes from the British statistician George Box. It is one of Roger Oechslin's favorite quotes. The trained meteorologist and flight instructor is considered an important influencer in the Swiss paragliding scene when it comes to the topic of flight weather and corresponding education and further training. This also includes the realization that many pilots today rely almost blindly on weather models and corresponding weather apps when assessing flight weather, without being aware of their limitations. According to Roger, we should instead focus on understanding the dynamic basic patterns behind phenomena such as, for example,
strong valley wind or shallow Föhn. Because it is important to be able to better assess the uncertainties of forecasts and take them into account when making your own flight decisions.
By the way, I can only present these special conversations to you in Podz-Glidz because there are supporters who financially support my work on the podcast. Join in too. You can find all the necessary information about this in the Gleitschirm-Blog Lu-Glidz, specifically on the Fördern page.
Roger, you're a meteorologist and you fly paragliders. Do you ever experience your meteorological knowledge getting in your way, so to speak, while flying?
Yes, that happens every now and then. It happened mostly right after my meteorology studies. When I was overthinking things and seeing ghosts everywhere, I've since developed a more relaxed approach to it and can look at it with more nuance. But yeah, that's naturally a thing whenever you know a bit more.
What kind of ghosts were they?
Yes, that I discovered lee situations everywhere when I looked out into the field. That even when the Föhn was only a minor issue somewhere, I analyzed everywhere where it could be a problem so intensely that I ended up flying less because of all that overthinking. Exactly, that shouldn't really be the goal, but that was just the way it was.
Do you ever experience the opposite situation today, where you say, "As a meteorologist, thanks to my deep knowledge, I can see things and connections that actually help you with your flying?" So, not just as a hazard warning, but that you're basically a better pilot because of it?
I think there are certainly situations where that's the case, especially when dealing with hazards. So, if you oversimplify the entire hazard assessment—I'll use the Föhn as an example again—you end up hitting thresholds, like this hectopascal rule that we probably all know, where you think you can't fly because the value has been reached, we have 4 hectopascal. But if you have a deeper understanding of the concept of the Föhn, how it works, and all the drivers involved, then you can still go out with a clear conscience on days when those simplified rules just don't work anymore.
And would you say that having a classic meteorology degree actually gives you an advantage there?
Yes, I think it's less about the meteorology degree itself. What drives us in our sport are perhaps more the connections from meteorology, from the theory. But then it's a lot of observation in the field, a lot of experience. And that's something anyone who never did a meteorology degree can acquire. I think the biggest background in terms of understanding the weather for us as paragliders, or in mountain sports in general, I gained during my time as a paragliding instructor, where I worked on it and actually had to constantly consider before school days what speaks for me being able to fly today. Based on which factors, on which data do I make the decision?
And then also really over and over again. I constantly observed outside what actually happened, how the weather behaved. And I think this repetition helped me the most throughout my career.
You're saying that as a meteorologist who is also a flight instructor. Can an average pilot, let's say, do that too? This repetition and really engaging with the meteorology so intensely, like going outside and seeing what's happening out there and stuff? Absolutely.
It certainly requires it. It
requires, of course,
a certain openness to the subject. What also helps are discussions with other pilots, flight instructors, and experts regarding the topics and observations. I think in the end, it's much easier than many people think. It's effectively just a type of training. I always describe it like sports. If you visualize an aerobatic maneuver in your head, for example, you can do the same thing with the weather. While you're reading the forecasts and looking at the weather maps, you can create a movie in your head: what do I expect today? And if you do that over and over again, you develop a critical approach to the whole thing.
So I mean, in a positive sense, critically. You notice when the weather starts heading in an unsuitable direction, when there's suddenly wind in a place you can't place, or it becomes gusty. If it's too thermally active, you have a model in your head that you can constantly cross-reference with. And of course, meteorology studies help, but that can be achieved by anyone with training.
Would you recommend telling everyone, okay, look, maybe create a routine, check the weather every morning, even if you're not going flying. So, take a quick look at the weather sites, the main ones you might also use for flying, and then imagine how it will be during the day... whether the weather should be like that, based on your expectations. And then keep looking out the window every two hours or something, even if you're sitting in the office. Maybe also check the wind station values or something and just compare. Did what I imagined it should be like actually happen?
Exactly, Lucian. So if people do this for themselves a few times, how shall I put it, if they've done it a few times and realized how easy it actually is. You don't have to invest three hours in it. You can read the general weather situation in the morning, like the small text. What's the big picture doing? That's also a topic—always going from the big to the small and then mentally projecting the weather into your own space, interpreting it. And either you look out the window, which you should do from time to time anyway while working. Or you look at the end of the day. Or you look at the measurements at the end of the day, as you said, or also webcam images. I mean, on Windy or many other weather portals, you can replay the whole day again today.
You can watch a radar loop of the day, and that way you can check your thoughts, your mental weather film.
How exactly do you mean I should check into that daily? I mean, I can jump into anything. For example, if I say I'd open Windy—and with Windy, I don't just have ECMWF as the base model; depending on where I'm looking, I basically have 13 others. Or seven, or whatever, I can select them all and look at everything. Do I have to look at all of them, or how do you go about it? Or do you really say, no, I have a favorite model, that's the one I mainly look at, and that's what I work with and it helps me?
...on. So I would, as I said, start with the big picture because it has the fewest details. And it's the least confusing because of that. These can also be surface pressure maps, like the ones from the national weather services. Meteo Schweiz has the general synoptic text. I think there's something like that from the DWD too, the overarching text. And I would start with that, because otherwise it gets very complex very quickly. And then I would take the next step, thinking into your own flight area, your own region. So what does that mean for me? If I have this strong west wind that I've had the last few days, where do I see it if a cold front comes, which might have been mentioned in the general text?
And I can take some time with that, which is already a good achievement. And then, of course, I can do more sensorially. I can start visualizing details, like where I feel the wind, similar to how it's done in sports. What does it feel like? And then, as a human being standing in the space, I consider what's changing. Also the day, in which direction. If you then want to get a bit more detailed with the models, as you asked, I would start with a global model again, just to avoid having too many details. It could be an ECMWF, it could be a GFS. In the end, it doesn't matter that much. So, while a GFS is more coarsely resolved, it remains a model.
And the mental effort, really thinking about the fine details, we should still do that ourselves. So you can maybe look out for a bit more wind detail at different altitudes. And constantly ask yourself, what does that mean for my valley? With this wind direction, does the Föhn break through more or less? Those are the detailed questions about meteorological hazards. You can take the fine-scale models then. They already look very realistic. That would be the final step for me. But if you've taken this path, from the coarse to the small, and constantly asked yourself what the weather is doing, you can look at the fine models much more critically afterwards, because even the fine models are models.
And models are only a reflection of reality, and despite all the achievements we already have—and they are immense, and they will be even more immense—there will always be errors in them. And with the way of thinking I'm describing here, you can approach these errors a bit more critically and recognize them, or better yet, recognize them more clearly.
But you'd still take that classic route—the way you're describing it now is like a classic, real meteorological approach, let's say. I first look at the large-scale weather situation and then move to the small-scale. I mean, nowadays with the websites that are so nicely prepared for us pilots—whether it's Burn Air, Paraglideable, Meteo Parapont, or whatever else is out there—you could also say I only look at those; I don't have to worry about the large-scale weather situation at all. I open up Burn Air once and check, roughly, in this thermals display, is everything somehow in the green or even blue area, or in that area where I want to go flying. Then I can see, everything actually looks quite good with that. Then I might call up the point forecast for my launch site and say, yeah, the wind there also looks flyable and there seems to be some thermals, and it's not raining.
So I can just go fly. Isn't that enough, actually? Yes. That
The problem is that in many situations—and I emphasize many situations—it's enough. It's actually amazing, as a meteorologist myself, and I've also worked in modeling, I'm fascinated by what's possible today. The problem is that in some situations it doesn't work, and those are primarily situations where small-scale weather phenomena are at play somewhere, which might have more of an effect in one valley than another, and the weather model still struggles with that fine differentiation. If you look too closely at the weather model, regardless of the platform, too closely at the detail to decide whether you can or can't,
then from my point of view, it's a dangerous application of these super tools. I mean, the tools are great, but they're being used incorrectly. If you go from large to small, then you already—if you go from large to small, then you already—if you go from large to small, then you already have the big picture and you already know, this could be dangerous today, that probably not. Then consulting such platforms is actually the last step, but with the necessary background knowledge, and then from my point of view, it works better again. So, first zooming in and looking at the wind arrow of the launch site and making a decision based on that—no, not zooming in first and looking at the wind arrow of the launch site and making a decision based on that, but going way out first and considering what themes I have today?
Could Föhn be a topic? Could thunderstorms be a topic? Strong valley wind, whatever? And only at the very end do I do a check with these sophisticated platforms. From my point of view, the platforms aren't the problem, that people might sometimes be misled. That's not the main problem. It's a matter of attitude—that of course, you want to keep it as simple as possible, you don't want to look at things for ages, so if there's an offer that actually takes this path, this mental work off your hands, then that's a natural human need. And I think I want to raise awareness a bit that it's a great tool, but it has to be used correctly with a critical attitude.
I'm going to open up a bit now,
I don't know if you can tell me anything about that, but Paraglidable was the first offering where an AI doesn't make a decision, but rather presents a weather situation to me. Basically, it takes the GFS and says, "We're combining these weather conditions and looking at where people flew in such conditions," and then the model learns the corresponding patterns. It sees that people flew in these specific weather conditions and ultimately outputs a percentage saying, "You can probably fly well here" or "not so well." Now, Paraglidable came out today—I think it was 2018 or something like that—so that was long before the big AI peak with the large language models and everything that has come out since then.
You could imagine that with AI, it would probably be possible to do much, much more. Do you think there's a path where we eventually say we have our flight AI, which was trained on really, really many flight data points from the last ten years and a huge number of weather situations from those, where we might also include station measurements and other things—so not just the GFS as a model, but really what was measured in reality on-site by individual items—and that we could even get something out of it where we just ask our AI, "Is this day flyable or not?"
I'm sure that we'll get a lot better here in terms of technology, whether it's with numerical weather models or the AI case you described. But that won't solve the problem if you take, for example, accident numbers as a metric; the accident numbers won't be reduced because the human reacting to it—the human, as soon as you can differentiate more, will push the limits further and will always try to find more of those green islands of flyable situations, somewhere in the model at the highest zoom level, "Ah, there's something, I'll deliberately do it a bit excessively and I can still fly over that mountain, but not the one next to it."
And when humans go along with that, we then have another problem, especially with AI, because of course we have pilots with very different skill levels and not every pilot is in the right place at the right time, while another one might be. And it's obviously difficult to separate that cleanly with such algorithms so that it provides a realistic picture for the people involved. So, that also leads to a false sense of precision, which in my view ultimately becomes a risk. I'm of the opinion there, too, that even if we eventually have such tools, it will follow the path I described at the beginning, from the large to the small,
...along with a conceptual understanding of our meteorological hazards; we won't be able to get around that either.
You also provide further training in meteorology yourself.
What are the points where you notice that the people coming to you understand the least? Or where can you get the least out of the models that are already there, and where is there simply a lack of knowledge and understanding?
So, what I generally notice is simply an oversimplification, a knowledge of meteorological hazards that is too schematic. To give some examples, it's deeply ingrained for many that the valley wind, the strength of the valley wind, is primarily driven by instability or by the thermals rising. Now you can ask such people, "Yes, but have you also experienced it being completely stable, high-pressure weather, and in the lower wall, for example, there was a very strong valley wind? How can you explain that? You couldn't fly thermals anywhere, but there was still a strong valley wind." And then they realize, "Ah, yes, somewhere that doesn't add up." And I think the problem is, even in the exams for the pilots, it's too much schematized in some parts.
that you just store knowledge somewhere too simply and think, "Ah, I can use this to solve the problems." And then a supposed solution suddenly becomes a problem because you're doing a hike and fly on a very stable day and end up somewhere in the afternoon and realize, "Ah, I'm flying backwards, what's going on?" The other thing, of course, is the whole story with the Föhn. There's this...
The textbook concept with this nice, first dry, adiabatically rising, then moist, where you can explain the warming and everything. And then you realize, hm, this schema, this model, it works, it works, but it still doesn't work everywhere. Then suddenly someone comes along and says, "shallow Föhn." Many people can't even place that word. They've heard it before and it's discussed widely in the scene, but many can't place it and don't know what it means for them when flying. Another example that I also find very interesting: "the strong valley wind protects me from the Föhn." For me, it's the opposite. If the valley wind unexpectedly becomes strong on a Föhn day, it means...
further back in the valley, the Föhn is at work there, causing a pressure drop that allows the valley wind to increase. That's the opposite logic. These are the kinds of things I try to go over with people step by step—not as much at once as I'm doing here in the podcast, of course—but I try to look at them with people step by step to help them organize and solidify the conceptual framework in their heads a bit better.
Now I think a few people listening here have questions. Yes, yes. Those who are starting to think, oh, now he's brought up this topic, I'm interested in that too. Maybe we can briefly cover that, but briefly explain the model behind it, the conceptual model behind it, how something like that works. So, first question: why can the valley wind still become so strong on such stable days?
Because it's not primarily the thermals rising and sucking air in that model, but rather it's effectively a stronger warming of the entire valley air because there is simply less air volume in the valleys—since the mountains take up volume compared to the old foothills. And this warming causes the entire air column over the valley to grow slightly vertically. If you look at that later in a cross-section from the Mittelland into the valley, you'll notice that the air level—or if you take a pressure surface, for example,
800 hectopascal pressure surface, that lies higher over the valley than over the old foothills. And that naturally creates a gradient in altitude, where the 800 hectopascal air in the valley is a bit higher than in the old foothills. And that's how the air starts to flow at high altitudes towards the Mittelland or towards the old foothills. At some point, you have too much air there and then there's a counter-current that flows back into the valley. That's the bigger picture. It might be difficult to explain here without a visualization, but it creates a bigger picture in your head. Larger forces are at work than if there were just a cloud, a thermal, sucking and pulling in the valley wind.
The system is multifaceted, much larger in scale than just this thinking about thermals. It's like a giant conveyor belt being started there, a sort of air conveyor belt. A circulation. It's actually a circulation that goes from the foothills into the valley. You don't notice it at high altitudes because, of course, it can
distribute vertically over a large area. And so you don't notice the wind moving out, but concentrated in the valley, it actually drives the valley wind to a large extent.
And it flows in there, but ultimately there could be days when you're flying and say, today it's rock solid, I can't get up at all, but if I sink lower, I hit a very strong valley wind and I just wonder, how can that be?
Exactly. And of course, what also plays a role is that you can always look at the whole thing—every phenomenon, every danger—like a recipe. You can think about what drives the valley wind. For instance, it's this different heating, where it warms up more in the valley than in the surrounding foothills. But of course, it's also when the wind—the supra-regional wind—enters the valley, pointing in that direction, inward towards the valley; that has a supporting character. If you have an inversion at exactly the right spot, where it also restricts this valley wind vertically a bit, that can lead to a Venturi effect, to an amplification. These are different components that play into it.
That was one topic. The other topic is shallow Föhn. Have you already addressed it? What is Föhn and what is shallow...
Föhn? Well, shallow Föhn is a type of Föhn. There are simply different types of Föhn that can't really be distinguished in black and white. That's the most important thing to understand. Maybe first the Föhn that the general public understands as Föhn, even outside of the paragliding scene. That's the Föhn you really notice in the mountains in different places, especially those classic Föhn plates, which are also usually mentioned in the text forecasts. It's quite pronounced vertically, coming over the Alps like a stream with a certain water level.
And in contrast, the shallow Föhn is actually, as the name suggests, a shallow overflow of these Alpine passes. So it mainly depends on how high the air comes over the mountains. Is it really like a river going over the Alps, or is it just a small trickle, a little stream, going over the mountains? That's the distinction. Then with the classic Föhn or this high-reaching Föhn, the wind at altitude is usually, in most cases, supported by a southern direction. With the shallow Föhn, it can also come from an eastern direction; it can come from a different direction.
But is a shallow Föhn, if it's just like a trickle flowing over the passes in some places, still dangerous?
It's still dangerous because it can often flow over places like the Rhine, if you imagine it as water, where you might otherwise be more protected in other Föhn conditions, where the whole... I already brought up the stream analogy earlier. A stream bed can take on a completely different flow pattern depending on the water level than when there's only a little water. And if then the direction changes slightly—I'm trying to imagine it like this—a tributary enters the stream bed from the right or left and brings in a bit of water from there, then we have completely different flow patterns. Of course, if you're in a place where the Föhn rarely occurs, and that's also in the text forecast,
nothing much is said about the Föhn, which can lead to surprise effects. From my perspective, that's a bit the problem with the shallow Föhn. No one talks about a Föhn storm, very few mention it, and it sometimes has different flow patterns than the classic Föhn.
How can I as a... pilot, who isn't a fully trained meteorologist, how can I recognize that today a shallow Föhn could become a problem? I mean...
There are several ways to do that. One is, of course, those fine weather models you mentioned, where you can already see the flow patterns quite well, but not perfectly. You just have to look at them critically and think, hey, I'm just going to see if shallow foehn could be a problem, yes or no. Not where or at what time, but whether it could be a problem. How do I see that then?
Because you're saying you can maybe see that in these models, but what is it? What should I be looking for?
So if you look at the height, at about 2,000 meters, as a model area, that in the valleys from a southern direction there are stronger winds coming over the passes and then, just like water, making their way towards the north. That would be one thing that
one could see there. I mean, I see
could see. So I see,
how strongly the tongues of wind that stick out there, the ones that are licking out. Yeah, exactly.
But, for example, shallow Föhn doesn't have to be in the dark red or whatever color scale is being used. It's simply a pattern. Ah, why is this coming from the south now? Strange. It's just different, perhaps, than on a typical sunny high-pressure day. You notice you have a flow from a southerly direction. What you could also see, partly, is of course the pressure difference, which is the classic way, but that can also be misleading. You don't necessarily need to have 4 hectopascals in the forecast for shallow Föhn to occur. And from my point of view, the most important things are the Previtemps, that's the thermals forecast,
so not the thermals forecast, the temp, so the temperature trend from the model. And we have that in Switzerland, we can compare north and south for different sections, and there you can also see at the height of the Alpine passes that the warmer air is in the north than in the south. So colder air comes from the south, which can then push over these passes. And that is an indicator that I use a lot. Of course, if you hear it like that now, wow, what exactly should I do? Sounds complicated. This is something you should have done a few times. Basically simple, you could also take that from a surface map; it must have the colder air in the south, and especially at pass height or slightly above, like 2500 meters, if there is colder air there, it creates higher pressure and leads to the air being pushed over the Alpine passes.
So, for example, I could go into Windy, pull up the temperature maps, set the altitude slider to 2000 meters, and if I see, okay, it's a bit bluer on the southern Alpine side and the colors are greener on the northern Alpine side, it's apparently a bit colder on the southern side than on the other side—those are the kind of days where I should be careful.
Yeah, the problem with the scale, especially with Windy as you mentioned, is that the temperature difference in these ranges, in these degrees, doesn't show up enough. So you have to tune the color scale—you can do that in Windy—and I would recommend using the finest models as a warning sign. It's important not to base such forecasts on just one value, but to look at different things. And I think the motto "in case of doubt" helps you with this difficult phenomenon,
where you can look up foot forecasts or find a final answer on how to deal with it. Regarding the topic of staying home, there are other sports as well, and if it's a potential shallow Föhn day, there are other things you could do.
Or other flying areas? I mean, isn't shallow Föhn mainly relevant for a few valleys where you can say it's most likely to pour in, which I just avoid, but are there certainly other areas where I—or would you say, no, there are such shallow Föhn conditions where a general avoidance in the Alps would make sense?
No, of course there's the distance to the main Alpine ridge, which already makes a big difference. I'm talking about the Jura now, where you're already further away. For me, I don't want to make a general statement, but in the meantime, you know a bit about which valleys are problematic for this shallow Föhn—like Valais or Upper Engelberg, there are several where it reacts, but my list isn't exhaustive. I think it's important to also contact local pilots, maybe even local flight instructors, and talk to them about what it means for you, rather than just taking a high-resolution model and saying "I can fly here and I can't there," because these models are often overwhelmed in terms of these questions regarding the Föhn.
And the other thing you can do, of course, is to be critical at the launch site. Humans have a bit of a tendency, which doesn't help them, that if they've perceived something once—a light, a green light, or just a feeling for themselves that today is a good day—they then go about with blinkers on and don't perceive the surroundings enough to realize that it could get worse, contrary to what they first planned, or that the wind suddenly becomes gusty from a weird direction. You have to stay open and pick up on those warning signs.
You've already mentioned these model limitations quite a few times. Is this actually not discussed enough in paragliding weather training? I mean, people always say, "Yeah, look at the models, then you'll see what the weather will be like," and you have the app where you can say, "Give me the exact wind values for x, y," and they're just spat out directly by the model or interpolated between two model points or whatever. It looks so real, but how do you get into people's heads to say, "Hey, this is just a model. It doesn't have to correspond to reality," or that it probably won't correspond to reality at the rarest points where you're actually on-site?
So, I'm quite sure that awareness of this is lacking. You can't say that as a general rule, though; it depends heavily on where you were trained, in which club, in what environment you were active, and how much sensitivity to this topic was there. Many people, or a large part of people—including myself—have this need for simplicity, to be able to simplify things as much as possible, not to need too much mental effort for something, and for the weather briefing not to hold you up too much. And then, of course, a high-resolution model that actually looks very real is perfect for helping out here.
And I think you can only say this over and over again, and that's my important message here: these are great tools, but at exactly the wrong moment—when you want to rule out a danger—they can fail or not work, and that is very deceptive. They look very good, they look like reality, but they are not reality and they will never be reality. There is a statistician who has spent his whole life dealing with models, George Box, who once said, "All models are wrong, but some are useful." So all models are always wrong in some way, because they are representations of reality.
And I think that's something you have to keep saying over and over again; that's also my key message here: these are great tools, but at exactly the wrong moment—when you want to rule out a danger—they can fail or not work, and that is very deceptive. They look very good, they look like reality, but it's not reality and it will never be reality. There's a statistician who spent his life working with models, George Box, who once said, "All models are wrong, but some are useful." So all models are always wrong somewhere, because they are representations of reality. And I think it's very important to know that somewhere. And the second need, of course, is also increasing in our sport. We always see that people are flying everywhere, and then you think, "Yeah, why can the other person do that? Probably they did the whole thing with the super-model." And the need today, "I want to fly somewhere"—you said you couldn't go to another location in the Föhn—that's the mindset. Yes, we are very mobile, and in the old area, we can quickly get to another location. I find a place where I can still fly today. And even there, I think everyone has to ask themselves again: does it really have to be? Or can we take a step back, especially for people who don't yet have a lot of
...have a flight backpack, in terms of experience and skill.
Then you said all models are wrong. I've experienced it many times before where I look at the weather, the model says strong gusts, forecast in the areas where I could fly, and I decide to skip them. Then I look later—which happens a lot—I check the live tracking or something and see, oh, people are actually flying there, and then I ask afterwards, what was it actually like on-site, what were the conditions? And they say it was super clean, like it wasn't a problem at all. The gusts were forecast at 35, 40, but nothing of that reached the bottom; it was actually all very calm. You could fly great. With my caution—and I'll be humble and say it—I hold back; with these gust values, I basically don't fly or something, and in the end, I've robbed myself of a flight.
So there are two sides to the model being wrong. It can trick me into thinking these might be good conditions, but then a shallow Föhn still washes in somewhere and knocks me down, but it can also trick me into thinking, hey, it's actually way too heavy, when in the end it's actually flyable. How can you, or is there a way to learn, to recognize, okay, which way is the model leaning today?
wrong? The problem is that this optimization, this drive toward optimization, and the fact that you're always seeing others, leads to you mixing up two things. One is that it can be flown, and the other is the immediate risk at that moment that it might tip over. On a day when the Föhn—the actual Föhn risk or the chance of it breaking through—is relatively high, I'm done. On many days it can be flown, but that doesn't mean it's sensible when it's already knocking at the door, like, "Hey, it just needs two more degrees of warming and I'm in." And now the pilots just got lucky that it didn't come.
So, purely in terms of learning, from a learning perspective, that's a dead end—a very dangerous dead end. And of course, the new media, all the livetracks we have, fuel this problem where you always see, "Oh, yeah, he still made it." On one hand, you don't know if maybe he was a much better pilot, and that simply makes a big difference. And you don't know how much more it would have taken for it to actually tip over. So that's also with other things, too—we're always talking about the Föhn, cold fronts, etc., where you used to just have a bit of... you used the word humility, a humbler attitude toward them, but today that's being softened because, "Yeah, I could handle that with my glider, I still made it."
And that might be a bit of a provocative statement, but that's how we're driving our sport into the ground.
Now I sometimes get asked by people who are observing this as well, who say, yeah, with climate change, many conditions are somehow becoming more intense. People are also noticing that thermals today are somehow different than they might have been in the past, because they sometimes develop much more intensely during the day or persist intensely into the evening, and all that other stuff. You can maybe question to what extent that has to do with climate change and all that, but things like this—isn't it something where we can eventually say, hey, if we want to keep practicing our sport, we also have to learn to deal better with such extreme situations?
Yes, you have to learn to deal with it. I'm thinking especially of these additional weather phenomena or hazards that climate change brings. For example, we had summers where we suddenly had phenomena that we might be more familiar with from holidays, like dust devils, or overheated air at take-off and landing sites that suddenly led to strong turbulence, to rotating turbulence. That means, perhaps purely in terms of daily planning, the afternoon could suddenly be off-limits, just like in other flying areas where people are already somewhat familiar with that.
Depending on the flight level, it just isn't fun or it's dangerous to fly. Somewhere in a large valley with very overheated air in the afternoon, maybe that's not the right place. And I think that's part of it, but it's more of an attitude thing. How do I deal with it? Would I rather be a bit defensive and not always want to fly just anywhere at any given time.
If a pilot is just starting out and getting into aviation, and he hears you say, okay wonderful, Roger says I should be humble, I should be careful, I shouldn't fly if the models are maybe in the yellow or something like that... how can he still become a good pilot if he never actually puts himself in difficult situations? That's...
That's not my message, so I might have to put that into perspective a bit, the way you just said it. I think you should push yourself. There's the flow model, between under-challenging and over-challenging. I think it's important that as a pilot, you're constantly being challenged and trying to set goals, training things. But training doesn't mean driving to Valais with everyone else as quickly as possible after your training on a strong day to do a big cross-country flight. Training means really thinking about which task will help me progress. Ground handling, doing a Siku, and then what I said at the beginning, visualizing the weather.
If you do this consistently and aren't just always in competition mode, but really in training mode, then I'm convinced you'll make progress and eventually reach a skill level where you can fly on days when a beginner simply can't do anything or maybe better stays on the ground. Also, those strong days I mentioned, those are often the days when you can fly very high over the Alps with a very high base. But in that moment, I also manage to be at the right altitudes and not expose myself to danger somewhere in the valley wind in that turbulent, overheated air; instead, I'm in the right place at the right time because I've trained the route like that.
Of course, if you put that into perspective, that would be the ideal path. I've also had moments where I was pushed back or became overwhelmed. We are in an outdoor sport, not in a gym. We have a very complex environment, and there are always situations where we suddenly think, "What did I do wrong?" and it can actually become dangerous at times. That's also part of the learning process. Ideally, we are in that learning zone between under-stimulation and being overwhelmed.
You said visualize the weather. How important is it to say, or is it also a possibility, to say, I basically just head out all the time. I take the glider with me, but I really go with the mindset that I'll only fly if I can see absolutely on-site that it's safe, whatever I mean by that. But that you still shouldn't just occupy yourself at home and ultimately maybe with some live tracks, wind values, and webcam images, but say, no, you actually have to experience the weather with all your senses. The best way is to really always go to the launch site, you just don't always launch.
So I think that's a very good way to do it, if I think about it that way. The only difficulty will be the human being themselves. You almost have to be a Zen master to actually make an objective decision. But that's also a kind of training, of course. Saying no, walking back down—for example, with a hike and fly, that's something you can also train for. And I do a lot of that, I go on a lot of hike and flies, I have my glider with me a lot, I just take it along. And that's also a kind of training, just to see, yes, is it compatible now? Is the flight I want to do, the weather, compatible with me as I stand there, with my gear, with what I can do, and should I take off?
Yeah, no. It's great training, but it takes a lot to actually be able to say, yeah, don't unpack, walk back down. But in terms of the learning effect, of course, it's very, very good. Also flight schools, for example, I've already heard the term "weather optimists" today, and then they go to the launch site—well, of course, you can do weather training in a classroom for a long time, that's where it happens, but what the pilot really has to be able to do is grasp the weather, perceive it: how is the wind coming, what is it doing, how gusty is it, and does it fit for us now? And if that's naturally well-moderated, it's a huge learning effect.
Do you know pilots who basically don't look at weather maps or anything like that, but just gauge the weather by gut feeling, yet still manage to stay safe in the end?
I'm sure those exist, and of course, intuition comes into play when you've been in an area for a long time where you've gained a lot of experience—how the wind changes when something happens, how I perceive the first signs of something that could become a problem for me. I don't think that's wrong. If someone does that for themselves in a place where they have a lot of experience, I just think that today, with the possibilities of this digital world and this information, if you use it meaningfully, plus the perception on the ground, you can be as safe as possible. And maybe once more, I don't think you have to do weather for half an hour somewhere every day; it's supposed to be a ritual that can also be efficient.
otherwise people don't do it. And I'm strongly of the opinion that if you do that repeatedly, you get an understanding of the processes, and you get faster and faster at it. Do you look at the weather every day, even if you're not going flying? Yes, I've kind of gotten into the habit of just following the general situation a bit every day, looking at it, and being a bit more conscious of what's changing out there when I'm in the office, looking out the window. How long does a basic assessment take you? Actually, it's my morning routine alongside my coffee, just opening the app briefly and going through it. Which app do you open then? For me, in Switzerland, we have a convenient setup with MeteoSchweiz, where, above all in the text forecast written by people, it provides high quality for a rough assessment.
You have a page, so also on social media, where you're on there as Cloud, I think you're also known as. And there you sometimes post very beautiful time-lapse shots of cloud developments in the valley, where you can also see how a valley wind flows or how clouds flow over a slope, over the ridge, and things like that. Shouldn't something like that actually be used a lot more to show people a bit how weather works in such valleys in terms of flow?
That would be exactly it. It's a great didactic tool for you where you don't have to explain much at all. Not as cumbersome as how I just explained the valley wind without anything to see. As a teacher or instructor, you can actually step back and just let people watch the whole thing, let it take effect. You don't have to add much because a time-lapse shows so beautifully how the clouds move, what's happening. You can, for example, visualize Foehn effects—so lee situations—superbly. I did that a few weeks ago with fog dynamics. And suddenly you see the cloud dissipating on the back side. That's this lee, this warming effect that you also have with the Foehn.
You can show a lot with that. And I even use it myself. We talked about the launch site earlier. I'm unsure about the time, what's happening out there, and I want to get a picture of it. And above all, of course, there have to be clouds, otherwise it gets difficult. Then I take out my phone, prop it up against a rock for 10 minutes, fifteen minutes, and it's already moving in front of me. And I get a much better impression of what's going on.
So you actually take a time-lapse on-site, and then you watch what happened in fifteen minutes in about 15 seconds, something like that. Exactly.
Most of the timelapses I make—I make a lot of them, but I don't upload them anywhere—are just for myself to improve my understanding. It can also just be in the area. Just being there and observing what happens. That doesn't require a camera, but it naturally takes more patience if you do it that way.
What always fascinates me about these time-lapses is seeing the dynamics build up sometimes; it looks like you're at the sea—not just a sea of clouds, but a real sea. It surges forward like waves, then washes over a ridge like so, and then it recedes, and the next surge comes over, and so on. Where does that even come from, that these kinds of waves can form at all? And you can actually say that there are perhaps situations where, if you were flying on the lee side of this ridge or were going to, you'd say, "I can maybe fly along for 5 minutes and it's safe because nothing is surging over right now, but in the next 5 minutes something will surge over, then it's crap, and the next 5 minutes it's safe again." Depending on when I get there, I might just be lucky to catch the right 5 minutes, but where does this wave-like, pumping motion of these situations even come from?
On the one hand, you can compare it quite directly to water if you take the word waves—the colder air underneath, the warmer air above, and then even without wind, we already have waves; we have a surge of water, it recedes again, we have these rather turbulent breakers that we might be able to imagine somewhere in the terrain, and then the wind comes over it. Often at air mass boundaries, at inversions, there are wind shears, and that can create different dynamics again. In extreme cases, like Kelvin-Helmholtz waves, which actually break—those instabilities—and you can compare that very well. I think it's also a good image to sensitize the pilots a bit: hey, as you said, right now it's fine, but in small scales, a problematic situation can form very quickly in the line.
I'd like to go back to – that
I just thought of something – going back to the models. You said all models are wrong. The interesting thing for us is, well, sometimes they are useful, as you also said, and the interesting thing is also to see that there are days when models match quite well for one point, and on other days not so well.
How can I check in a simple way, maybe for one day, if the model matches what's happening outside? I don't just mean by looking out the window and saying, "Well, the model says it's supposed to rain, but the sun is shining right now. Maybe it doesn't quite fit." But how can I check something more systematically—can I trust this model today, or do I have to say, "No, the last run probably calculated that a bit wrong for some reason, so the weather progression will be a bit different"? Are there ways for me to check that in a simple way?
So, your question about whether it's going to work or not, you can do that. Most models, or many models, are calculated from ensembles today, and on different platforms, you already have products where you can see a bit of that. An ensemble is nothing other than a model being run multiple times, and a small perturbation is introduced at the beginning, and you see afterwards how the model reacts to it. If the model reacts strongly to it and goes in a different direction, you notice that it's not generally very sure of its statement. So, if it's once 5 degrees temperature and with a small perturbation in the initial conditions it suddenly results in a 4-degree difference, a 10-degree difference, or whatever, somewhere after two days, then you notice that the model isn't sure about that.
And on different platforms, for example MeteoBlue, there are statements about accuracy, for instance. You can look at these classic spaghetti plots, where you can see how they diverge, or more and more often, instead of deterministic maps—yes, no, or which value—probability maps are being calculated. For example, how high is the probability of precipitation? With things like that, you can get a bit more out of it or see, as you asked, whether the model is confident. What I also find important is which model is suited for what, because there are
Models can be tuned for different things.
For example, in wind energy, there are models that are simply optimized very strongly for the wind at this specific altitude. And it's difficult for us as laypeople to see which one is for what. But we can track a bit which phenomena the weather model predicted well and which phenomena it always struggles with a bit. So it can predict high fog well, for instance, and I've also developed a bit of a feel for how to work with the model. But that's a high art; it's actually a bit of the meteorologist's craft in the forecasting office to take these differences and work with them meaningfully, because we are very strong in the area of optimization.
But do you do things like, for example, you looked into the model early in the morning and saw, okay, how the cloud cover is supposed to be at certain times, and then you're at the launch site—let's say you checked the weather at 8 a.m., you looked at how the cloud development was supposed to be, but you see, ah, a cloud band is supposed to get thicker at some point, it's supposed to move in at 10 a.m. and have moved on by 12 p.m. or something like that. Is it something where you then say, I'm at the launch site, I'll open my app again and really look at the current satellite image, where is this cloud band located, to then say, hey, the model said it should actually be lying across the Alps like this and that, but it's still further west or already further east or whatever, to say, well, apparently this model is lagging behind the development or it calculated it much faster than it's actually coming, so that you do a comparison like that to then say,
well, then I have to be a bit more cautious with these model values today. If the externally visible development is somehow different in terms of timing, then I can't really trust it all that much anymore.
Yeah, so I generally do it like this: at home, I look at the model and maybe some individual measurements, then while I'm out there, I take measurements and look outside—basically, what's the comparison? I mentioned the weather film at the beginning, like the script, the visualization of the weather, and if that's not the same as what I see on-site with the cloud bands or the wind, whatever it is, then I already know, ah, the film is running differently. Either I do another check with the model or, like you said, I say, yeah, maybe it's not exactly as the model thinks today. In the case described, that's not a problem yet.
At most, there might be some shading for a cross-country flight, and if I don't like that, I'll wait a bit before taking off, but I always do this check to see what the model says, what reality is like, and how what I've planned out as a film corresponds to reality.
And how do you deal with that on-site when you see that, actually, I can't trust the model's timing or what it's calculated. For example, the model shows you rain, but you see the sun shining right now and for longer, and I look at the satellite image and see, no, there's a very large gap in the clouds, so it's probably not going to rain that quickly because it shouldn't be a thunderstorm either or anything else that's overdeveloped.
How do you then—or how would you say, how should one proceed—when you basically say, I can't trust the model today, it's not right. What do I do then as an average pilot who isn't exactly trained as a super meteorologist and who usually trusts his app, but now just has to realize it's wrong?
So I think about what the danger is, or what I've considered—which could be a topic today—and I try to answer the questions in my head: is there really something there, or is it just a ghost I'm seeing? And only once I've answered the questions and can tell myself with a clear conscience, "Hey, there's really nothing there," then it might be that the model was just being pessimistic, and if I can explain that to myself, then have a nice flight. So...
if everything is in the green with the measurements and everything you see, then maybe not for a complete beginner, but for someone who already has a bit of conceptual knowledge, you can still conjure up ghosts.
Is there a weather situation where you'd say, "That's my favorite weather situation," where you know you just have to read the weather report on Meteo Schweiz and you already know today is just safe to fly? I can just go out, I don't have to deal with the weather at all anymore.
Yes, that depends a bit on the season. So, what I really like—and that's why I like the "Wiese" [meadow] conditions when they come—is not in a strong form, but just "Wiese" and not too unstable, so not exactly spring. That provides some really good flying areas near me, where I just get happy and then go a bit crazy. Exactly.
And then you don't even look deeper into the weather anymore; you just see meadow conditions, no overdevelopment or anything. Not so unstable. Today, I can go flying.
Yeah, exactly. So the "Wiesenlage" is usually coupled with high pressure, where all the disturbances, cold fronts, and low-pressure systems are a bit further away. Just the classic "Wiesenlage." Exactly. Feel-good weather.
And finally, have there been weather conditions where you—or what do you mean by weather conditions, situations where you completely misjudged the weather despite all your knowledge and ended up putting yourself in danger?
So there were certainly many situations. Specifically, I know of one incident I had in South Africa many years ago. I was in South Africa for a few weeks and we were in Cape Town that day at Signal Hill, and I didn't really know the flight area very well. Everyone there said they could read the water outside, you know, how it is with the wind. I flew there and managed to get myself into the lee of a high-rise building because I simply had a completely wrong concept of what the wind does down there. And then I also flew low over a road towards the beach. A lot of things came together there. Certainly not just insufficient knowledge, but also a bit of the human factor.
I've seen myself doing that over a beer with colleagues at the beach. There's a lot going on there. It happens to meteorologists too. In the end, it's always the human who makes the decision and has to deal with the information. And in certain areas, knowledge only helps to a certain extent.
So even you as a meteorologist sometimes forget the meteorology?
I forget them or maybe my focus is elsewhere for some reason. And I think, besides meteorology, that's the big topic where pilots can work on themselves. How do I handle my focus while flying? What kind of situational awareness do I have? Do I really know what's around me? What's going on? What's the weather doing? Where are the other pilots? Where are the clear spots? That's also a lot of work. It's just as complicated as the meteorology. And certainly a huge learning field where you have to be able to become a safe and also a satisfied pilot. And not, as you said earlier, only fly in completely safe conditions.
So if you have both components—meteorology and the human factor—well under control, then you can actually fly a lot. And in conditions where you can truly say, "Yes, I've put myself in the right place and I'm operating with deep risk." And that's fun.
Okay. Maybe we'll talk about the human factor in a later podcast episode at some point. For today, let's leave it at that. Roger, thank you so much for sharing. I hope some people took something away from this and also learned that humility is sometimes appropriate. At the same time, you also have to push yourself sometimes to get into the flow and say, a bit of pushing is always required so that you can learn a bit more about the weather, and then be able to act accordingly in those slightly borderline weather situations from a flight perspective. Thanks for your time. Thank you very much. That was
this was episode 173 of Podz-Glidz with Roger Oechslin. In the show notes, you'll find links, among other things, to a series of interesting weather time-lapse videos. Take a look. If you want to hear more inspiring stories from the world of paragliding, there are two important steps. First, subscribe to Podz-Glidz in your podcast app of choice so you don't miss an episode. And second, become a supporter of my work. On my Gleitschirm-Blog Lugleiter, under the "fördern" section, you'll find all the necessary information and links on how you can continue to provide a basis for this project with a freely chosen, large or small financial contribution.
I want to thank all the supporters. See you soon. Yours, Lucian.