Most pilots actually don't manage to build up that much experience and do really long cross-country flights in their entire flying career. And having a tool like this helps immensely to see that it could work up ahead, but not many people have turned on at this corner. And only then can I start to consider why that might be. What could be the reason? And I think it can be a huge help as a learning aid, especially for beginners. And I don't think it's wrong to seek out support.
Podz-Glidz, the Lu-Glidz podcast. Stories from the cosmos of paragliding. Today with Ulrich
Scheller and
Lucian Haas at the microphone. Anyone who wants to undertake longer cross-country flights will soon realize that good planning can contribute a lot to XC success. Those who know in advance where thermals are frequently located in the terrain and at what altitude they need to take off elsewhere to reach these thermals safely will less often struggle with their decisions in the air. Good planning foundations can be found by the thousands in flight databases like XContest or the DAV XT. They are a true treasure trove of data. However, it is tedious to click through many individual flights and extract, so to speak, collective experiences regarding thermal spots, approach routes, and the best valley soaring opportunities. Unless, of course, you use a computer to help with that.
The Munich software developer Ulrich Scheller, who is also Gleitschirm-Blog, has programmed and made available his own tool for this: a dynamic thermals map called Thermalmap. Special calculation routines have automatically identified thermals spots in the IGC files of tens of thousands of flights. This data can now be visualized on zoomable maps, but intelligently filtered depending on the time of year and day, the prevailing wind direction, and the respective flights. This is the flight altitude from which you start searching for thermals. What advantages this approach offers over static thermals databases and how this flight wisdom could even be used live in flight in the future is only part of what Ulrich Scheller discusses in this 55th.
episode of Podz-Glidz tells.
Ulrich, you're a software developer. What are you working on right now?
Right now, I'm working on the Luca app, which you might have seen in the media. It's that COVID relief app where you can register each other.
and ultimately, where the health departments can trace who had contact with whom, if they want to, and can help accordingly with contact tracing during the Corona pandemic.
You're working on the Luca app. How many people work together on an app like that? Is it like a two-person team doing something like this, or are there 20 people behind it? Or how can one imagine that?
So, two people have worked on the app. I'm brand new to it. That means it's not that many. But in the background, many, many more people are working on it. For example, the whole topic of security concepts and data protection involves a much greater effort than what the user sees afterwards.
When you program something like Luca, do you also see it as being for your own benefit, so that maybe you can go fly around more freely again soon or something like that?
So, it has a different motivation than just the work and the money. Because I think that's a very good idea. Corona is affecting all of us, of course, and as a freedom-loving person, it affects me very much too. And the sooner we have the opportunity to go out again, meet up, and go to a restaurant, the better. So that motivation is definitely on top of it as well.
Are there, how should I put it, perhaps certain character traits of a software developer or a programmer that are advantageous in any way when it comes to flying, from your perspective?
That's a good question.
So, a software developer typically always sees the problems first, not so much the possibilities. And maybe that helps you in aviation too, by avoiding the big problems more often. But that can just as easily be a disadvantage as an advantage. I don't know if it's the personality traits. But I do see very, very many people from IT in aviation. What I can imagine, however, is that it's more about the personal freedoms that you might have more of in that field than in others. The
By temporal freedom, do you mean that you can organize your time a bit like that? I can work on a Luca app at night and go flying during the day.
Well, maybe it's not quite that free. But we certainly have more freedom of time than in many other professions, and also the financial means to do it if you want to.
What about analytical thinking, or acting in routines, or something like that?
Analytical thinking is certainly very good, especially for long flights or long-haul journeys where you need planning. Although that's also very much a matter of personality. I wouldn't go so far as to claim that it's any different because of software or IT specifically. Because other people, people from other industries, also have analytical thinking. I'd go as far as to say you're also a relative analyst, based on how I'd evaluate your lectures. And yet you come from a different field.
A long time ago, I also programmed in Basic and other languages. But I wouldn't say that's what I used to develop what I'm working on now. Ulrich, last year you started a special project with someone else—I think you could call him a computer scientist—Maximilian Koch. What is it about?
Yes, that's the thermals map. And the thermals map is special in that it processes the existing data that can be pulled from the XC databases in a different and much more precise way than the thermals maps currently available. For comparison, there was, for example, this KK7 map that many people will know, where various map layers were pre-calculated. With what Max and I have built now, we went through the specific DHV XC database, analyzed the flights, and then calculated, yes, where the thermals were at a certain time. And the map display of the thermals map represents that for a certain point in time and a certain time of day, but then calculates it down to the individual thermal.
That means you have many more options to optimize the display than with these pre-calculated maps, which basically just overlay a single graphic on the map once. And that's what was done with the thermals map. As a result, there's simply a much finer display with a lot more detail.
We'll get to the details in a moment. Let's take a step back—before this thermals map even existed. How did you two come up with the idea that we want to do something like this?
Basically, as a professional, you come up with that idea relatively quickly. I've already spoken with different people and many of them have had the thought of tapping into the DHV database to calculate various things. We actually did it, over Christmas and during the winter. And the first step was really just that we wanted to have this data and play around with it a bit. The fact that the thermals map came out of it just happened naturally. It wasn't planned from the start; it just moved in that direction step by step.
So, you basically said, okay, there's an incredible amount of data in these XC databases—specifically the DHV XC ones for now—and we're going to extract it, and then we'll figure out what to do with that data. Did I understand that correctly?
Exactly. And the process of extracting the data takes quite a long time. I believe there are 1.3 million flights in the DHV database. And the website isn't incredibly fast either, so it takes a while, meaning we were already experimenting a lot while the extraction was running and had already taken the first steps, and then with more data, we could do a bit more. And so, it was one step at a time.
Can you just access all those flights that are in there? I mean, can you really just pull all those IGC files down with a routine?
It's not that simple, and that's probably one of the reasons why not many people do it. The IGC files at the DHV are protected by Capture. Many people know this: if you want to download a foreign IGC file, you have to choose one of four forms. That's probably still relatively easy to solve. But the thing is, the KML files at the DHV contain exactly the same data. So those can be downloaded freely. However, the website is relatively slow, which means even if it's pretty open, it still takes quite a while. And there are also different databases. For example, XContest would be very interesting because it has worldwide coverage. And they've put a lot of brainpower into making sure it's not that easy. That means, technically speaking, it's been done very, very well, I'd say.
And you then have to invest a corresponding amount of work just to extract the data.
That means one software developer takes their hat off to another and says, okay, I recognize good work here. With the other, newer one, it's not as well protected. That's how I actually get to all the stories. Exactly. Now, if you say that in these XC databases, there are huge treasure troves of data lying dormant. Well, you've at least extracted part of that. How do you then figure out what insights you can gain from it? One thing was probably, okay, let's first see if we can get the stuff down? Now you have it on your hard drive. And then you say, wonderful, what can we even do with this treasure trove of data? What kind of data mining or whatever else can we do with it?
So Max basically started it. He just put the flights onto a map without doing much calculation. That worked quite well at the beginning. And with a kind of heatmap, you eventually get something that looks like the Skyways. But that's not very exciting either, because it just highlights the places where a lot of flying happens. And yeah, then you just keep going. What else nice can we calculate? There are different libraries that can, for example, calculate the thermals or the gliding distances from the flights. And that was the path I followed more. Specifically, calculating the thermals, putting them into a database, and then putting them back onto the map.
And the first steps of that were still relatively slow and not very dynamic, above all. That is, I could start a calculation, and then it would take about 20 or 30 seconds. And eventually, a map would come out that you could navigate a bit with. But in terms of practical utility, it wasn't very high yet.
So that was a map where, in a sense, all the thermals were plotted in some way?
Yes, exactly. And with the larger databases or data sets, that just took a very long time. And that one map was sometimes 60 megabytes, 100 megabytes in size. So for actual use on the internet, it was relatively impractical.
And then you said, okay. You have to somehow get something more dynamic out of that, you say. So how did it go from there?
So I've always looked at it a bit from the perspective of: I want to see on this map what would interest me during the flight. Just like the thermals map shows it now. And then, of course, I thought about how to get there. On one hand, it was just the sheer amount of work. You first have to set it up so that the data is displayed correctly, so that it can be requested via a web server, so that it's dynamic. And the other problem was, above all, making it fast enough. I mean, there's a reason why all the thermals maps have always been calculated statically offline until now. Because it's simply a lot of data and it takes relatively long to calculate. So I had to come up with something smarter to make sure it's still fast enough.
And it worked quite well because the amount of data is limited. I mean, you don't have to calculate the entire world map all at once; you can just pull out the exact points you need. And beforehand—in computer science, they call this hashing—you have to set that up accordingly so that the relevant data can be retrieved very quickly.
Just for the sake of understanding, so you have a database where all these IGC or KML files with all the flight data are stored. You've run a routine over that which extracts all the thermals that can somehow be identified in these flights, and you say, I'm putting them into their own database. And then it says here, there's a thermal, it started at an altitude of so-and-so, ended at an altitude of so-and-so plus X, so some 500-meter gain in altitude, and it was probably also at this strength or something during that time. All of that is now in your database.
Yeah, exactly. There are a few more pieces of data. For example, where exactly is the entry point, where exactly is the exit point. So not just the altitude, but also the position.
Yes, various smaller values. And nowadays, there are also a few wind calculations included, which aren't being used on the thermals map just yet. But there's certainly more potential there that could also be exploited in that direction.
Now that you have this thermals map, you can already access it online. I've briefly introduced it on Lu-Glidz and linked to it a bit. But explain to me now, how can I best use this thermals map if I were to say, for example, I want to go on a route somewhere? And now there's this thermals map offer. How can I use it to prepare a bit, for instance?
So, in terms of preparation, you can use it to fly over your planned route quite well. It starts with placing the pin where you want to start and setting approximately the right season and the exact day you'll actually be there. Then you can slowly narrow down the time to see when you can expect thermals at all.
At least, you always have to keep in mind what it's actually showing; it's not the thermals that will happen on that exact day, but it's simply a summary of the last few years. What happened in the last few years at all, what took place there? But it is already a very good indicator of when you should start at certain places, for example. And on the map, which might not be entirely obvious, you can also change the time in minute increments and not just with this selection dialog. That means, for example, at the border, you can now break it down exactly to the minute of when the first thermals were actually there.
That means I can set a start time and see, okay, there are, I don't know, 1,000 flights recorded from the border, and I can see that the first thermals people actually flew were already at 9:12 or something like that. Exactly.
So that's how you can see it in principle. Now, there's always a bit of uncertainty involved. I mean, the whole thing isn't an exact science; I have to have a weight. I have to take a certain interval in the calculation just to be able to see any flights at all. And it's like this: for the thermals map, I'm currently setting a 1-hour interval around this point. However, in the morning, that one hour is shifted backward, so from 9:00 to 10:00, for example. And in the evening, it's exactly the opposite, and throughout the day, this window shifts so that the results are actually still very, very good. And you can just pretend mentally that it's exactly this time.
So I can then assign it, okay, to this time—when I set it, I can see where in the terrain around me thermals might be forming that I can actually reach from my current position, if I've understood this correctly.
Exactly, the reachability is the next point. As pilots, we're actually only interested in the thermals we can really use. That means I first have to be able to get there based on my glide angle. Otherwise, I don't need to look at it at all. And the other thing is that I also need to have a bit of time to do it. And the other thing is, if I'm already very, very high in the air—for example, I'm at 3,500 meters—and the thermal, for whatever reason, only goes up to 2,000 meters and is relatively close below me, then I'm less interested in it too. So, those are two things that the thermal map can also do, which static maps just can't do because they no longer have the individual thermal in scope.
You said it's best to put this point on your own starting point at the beginning. But then you fly slowly from the border, through the whole Defragmental and so on, or something like that, or other corners. For further flight planning, I'd say, okay, then I'll just always set my starting point or my current point a bit further ahead and assume I've reached 2,500 meters or something like that, set a height, and then crawl along my planned flight path forward bit by bit, looking further and further ahead to see how it could continue? That's one way you can do it.
And that's how I do it myself. I mean, I use it for flight planning too. Usually, on really good days, like for example from the border, once you've mentally geared up, you can just keep pushing the pin forward and see every good thermal source that comes up again. Just by the clustering or a cluster of the thermal points, which eventually get shown in red due to the volume of thermals, you have a very good indicator of where you could gear up again.
That brings up the question of how flight preparation works. But that would also be, okay, I'm looking at a very specific route that I'm basically flying lengthwise. It could happen that I'm already somewhere along the way and didn't prepare well, or I did prepare but then chose a different route. Would it be possible to query such data dynamically during the flight? Like, I have my smartphone up in the cockpit and I can say, "Wow, there's that great thermals map." I'll just take a quick look at where I could reach the next one in my surroundings, or that something like that even happens automatically in the app?
Yes, there is not only the possibility, but that was actually one of my first thoughts on how to do it. And ultimately, the thermals map as it is currently available on the internet is the test mode for that. You have to be able to test it somewhere at home or at work. What actually happens if I were at this spot right now? And I actually told myself, I've built it into XCTrack exactly as you just described it, but I haven't made it public yet because it simply hasn't been tested much and because it might be relatively high in terms of server load if many people accessed it at the same time. But that is certainly a next step and there are a few people who are currently testing it. And if it proves reliable, if it works well, then I can certainly imagine rolling it out further,
that you have the possibility to see exactly that live during the flight.
From your point of view, is it actually, how should I put it, meaningful to rely on so much technology and trust these kinds of analyses? Or does it not also lose a bit of the essence of paragliding, if it's not about me looking at where—I mean, I still need to have the experience and everything else—where the next thermals are, but in the end, I basically have a computer telling me, "Hey, watch out, there are already 100 ahead of you who flew here, and there was always a thermal over there." You basically just have to fly to these points, and your altitude is usually enough to get you there. So that you don't fly based on your own feeling anymore, but actually just follow the machine. Doesn't something of the meaning—what do I mean by meaning, I mean what makes paragliding so free—get lost there?
I haven't had that feeling so far. I see it the other way around. I see that you have to build up this experience somewhere. And most pilots don't actually manage to build up that much experience and do really long cross-country flights in their entire flying career. So having a tool like that at hand helps immensely to see that it could work up ahead, but not many have kicked in at this corner yet. And only then can I start to think about why that might be. What could be the reason? And I think it can be a huge help as a learning aid, especially for beginners. And I don't think it's wrong to get support, because we all have an insane amount of support in the cockpit. We have a variometer, we have all kinds of planning apps, flight computers, and that's not anything bad or wrong either.
In the end, it's ultimately about having good flights, having fun, and for many people, being able to fly far—seeing success in the distance you cover. It's nothing wrong or bad, and everyone just has to fly in the way that works best for them and that they enjoy.
But in a certain way, it's always flying on well-trodden or flown-over paths, because the points that your software shows at least aren't you saying, okay, I'm learning from artificial intelligence where thermals lift best or whatever and I can extrapolate that to corners where no one has flown yet; instead, you're really showing the classic Skyways now calculated down to individual points where you say, that's where people mainly fly, I can pull data from that. I can then represent this data. That means people are being tempted again to really fly those thermals where you say, yeah, everyone has always turned it on there. Maybe 500 meters to the side is a much better one, but no one has ever discovered it yet, and everyone always just says, that's the good thermals over there, and everyone flies there.
So basically, the spirit of discovery is destroyed a bit by it. I mean, you could maybe call it a virtual spirit of discovery, where I place my pin on the route, slide the map back and forth, and say, let me see where the thermals appear there, and then I can think about why. And sure, you can learn a lot from that. But in the end, maybe a bit of creativity, trial and error, a bit of chaos, and all that other stuff that makes up aviation and flying in general, is lost because of it.
It just depends on what you do with it, of course. I've flown a few flights live with this tool now, and for me personally, it's always about looking at what's actually happening live first. It often happens that I don't look at my tool at all for an hour, simply because there are other pilots there. And we have this effect you mentioned in almost all major cross-country flight areas anyway. Because so many other pilots are out there, I don't go on a scouting course at the border, for example, because I see 20 others in front of me, 50 others behind me or something, and because of that, I have a very clear indication of where it's going and where it's not. And I believe that for cross-country flights specifically, scouting isn't that effective anyway.
You just have to stick to certain spots where you know things are relatively good. And that doesn't rule out the fact that you can still do some exploration flights.
How valid is this data, actually? I mean, there are probably huge differences in, at least regionally, I'll start with the quantity of thermals you have, and maybe also the quality you can derive from them. There's always a thermal or there was a thermal once, and now it shows up as a point somewhere on your map. You could say, oh, there it is, but maybe that was an absolute exception during that one flight that went along there, and now it's being shown to me anyway and I might be misled by it, while the better thermals might be at the next ridge where no one has flown along yet, or the thermals there weren't registered accordingly and are now included in your thermal data. Would you see big differences between the areas where you say a lot of flying happens and I can trust what your map shows immensely, and is there a way to recognize...
Ah, this region—when I'm planning there, I should basically be much more cautious with your data; maybe view it as a basic aid, but not as something I can rely on completely.
So I'm still testing that a bit, of course. Most of the time was winter, which means there weren't that many flight days where I could try it out yet. But I think you can already clearly say that it works particularly well in the Alps, for example specifically in the Northern Alps, where an extremely large number of flights were submitted. And based on my current experience, it's not very useful to look at in the lowlands because there are very, very few thermal points. I also practically never rely on a single thermal point on the map. So if I say I'm flying there now, I want to see a cluster of points, not just some single one, because otherwise the probability that it was a one-off flight for that thermal is quite high.
These are the data where you say, okay, you can read the thermals from here. From these data as you have them now, along with these thermals databases you already have or that one large thermals database, are there other things that could also be read directly from them that would be helpful?
I think there are a lot of things you could read out, but they aren't that easy to extract. For me, a certainly interesting point is not just the thermals, but also the glide paths. I wonder if there are things you can identify from those, like which glide paths always work well for valley crossings or wherever. And that's not that easy to calculate, and it's all very, very difficult to verify. But I think the potential of such things is extremely high. The only thing is the effort required to get the data and present it in a way that a user understands and, above all, can benefit from it while in flight—that's still difficult.
You mean, for example, that you can identify certain lines that hold particularly well, where you'd say, okay, if you're already cruising along here or if you want to decide where I can do a valley jump, for instance, it might be more sensible to do it between these two ridges on both sides of the valley instead of the adjacent ones. Because, for whatever reason, the sink rate in between was always less during flight, which the pilots would also have.
Exactly. I think there's still enormous potential there for flying. I believe it was Ferdinand Vogel who said he flew the entire Lechtal once without circling. That means there must be such lines that work incredibly well, where you obviously have to take every bit of lift along the way. But the fewer circles you fly somewhere, the more efficient you are on your way. And the more altitude you can take, the more height you retain, the easier staying up remains. I think there's very, very much potential there. Yes, if you had this so-called thermal vision, if you could read the air in some form, which will never work 100 percent. But if you had a few more insights like that, flying—staying up—would certainly become much easier.
Although, regarding the thermals, most people always say, "I want to see the warm air." And maybe these lifting lines don't necessarily always have anything to do with heat—so air rising due to heat—but sometimes maybe just through convergences, where air masses are pushed upwards, or ridge lift, because there are also simply dynamic ridge lifts where the wind hits it so well that you can glide along nicely, and so on. If you take your thermals database now, what defines a thermal for you, what do you read out of these IGC or KML data? What defines a thermal? Is it just a certain gain in altitude in a specific amount of time? Or does this pilot also have to have been circling? Or what belongs to it?
Yes, that's relatively complex. Ultimately, I also linked into a library that calculates that. And it's more than just the gain in altitude. This circling is taken into account. I find it very difficult... I
I have to interrupt for a second. "Library" is once again... that's a software technical term. There are certainly people who might not quite place it. What do you mean by a library in this context?
So a library is a software construct that you can reuse in different places and that takes a certain amount of work off your hands.
That means someone else has already programmed it. And you say, okay, it's basically already there. And I take it as a building block and somehow integrate it into my program. Exactly.
And this thermal detection—there's something there that isn't perfect, but it's defined accordingly with certain parameters. Those can also be adjusted. That means if I were to recalculate the entire database from scratch now and set the parameters completely differently, we'd have slightly different values and slightly different thermals, which might be displayed a bit differently. But the calculation works very well. I mean, I've taken out many examples and then looked at them in Google Earth, for instance, to see what they look like. That means they make a lot, a lot of sense. In contrast, for example, to the glide paths. The glide paths are also read out via this library. And I see more difficulties there. And that's why I also started with the thermals,
because you simply get good results.
So, to summarize briefly, what defines a thermal in that? Minimum heights, gain, and what else? Or what all factors go into that? There
The change in direction, the fact that there's a turn involved, does play a part. The height, clearly, the change in altitude. And yes, also simply the time. Because sometimes there are thermals where you have a climb, then a short sink, then another climb. That it's this thing, how do you say, that oscillates. That it's not a new thermal for every small climb, but rather looking at this entire climb over a longer period of time.
But that means if there's supposed to be a circle—a change in direction—for example, these climbing lines, which even today in, let's say, really performance-oriented slope soaring, where people try to circle as little as possible and instead say, I might not even fly that high, but I'll stay close to the slope and always try to catch that slight lift. That kind of thing just isn't detected in there at all.
At least not as a thermal at that spot. I mean, that would probably still be detected because there are people who don't just scrape along the ridge but also turn there. That would be detected. But that very, very performance-oriented drifting along the ridge and taking the altitude with you without much height gain or loss, that wouldn't appear in there.
Have you or have you guys ever thought about how to process this data you have now, let's say, with a kind of artificial intelligence, to see if you can recognize patterns that I could then simply transfer to less-flown regions, to say, look, these are the flown ones, and I'm giving you the simulated ones from that,
experience shows that thermals also position themselves. You might be able to recognize that these aren't flown, but you could still use that to better analyze completely different regions for yourself.
The idea definitely exists. The difficulty here is that you still need data. You have to have something to compare it against. A good example might be Paraglideable. Most people know that, where the expected cross-country flight conditions are simply displayed on the map. But for that, you need the corresponding values. What's the weather like? Or I don't know exactly which values are being looked at, but you have to trace it back to something. And that's what I'd be missing at the moment to make such a calculation. And the question, of course, is what exactly can you calculate? I mean, the thermals themselves are very, very small-scale effects. That means we're currently trying to look at exactly where the thermals are.
While all the weather forecasts do more of a large-scale calculation. And I think that's even simpler in that regard. Or, you need very, very high-resolution values, like detailed wind reports from the ground or maybe even the heat distribution on the ground. I don't know exactly, but you definitely need various of these features to apply an artificial intelligence and be able to compare what works well and what doesn't work so well.
As far as I know, Paraglidable also has historical data. So it calculates using the GFS model forecasts—the weather model—and they actually took those from the last ten years and then added the flights accordingly, and then said, okay, with which GFS model forecast values, whatever all flows into that, did good flights occur. Theoretically, you could go from that and say, look, I'm checking Paraglidable. If it shows me very good cross-country flight days, I can always look at what I might find associated with them. In your thermals database, you could also calculate that accordingly and say, okay, which thermals are the ones that are particularly good on the good cross-country flight days or something like that. You could probably do much, much finer analyses and everything else with that.
Definitely. I think there are still some things that can be improved. Now, the thing is that the thermals database is very, very good in different hotspots—specifically in Bassano or Kössen, we have an extremely large amount of data in there. And in those cases, it certainly makes sense to see if you can break it down more finely and show less; ultimately, the more precisely you calculate something, for example, if you factor in the wind or if you factor in a Paraglidable forecast, you can actually only exclude certain thermals. That means you show fewer data points, which are hopefully more relevant. That makes sense especially when I have a very, very large amount of data. At points where it's a bit more limited, where not that much is displayed, I can't gain that much from it either. Because then the hit rate would have to be very, very high. And that's where you have to find a trade-off: to show something that is already useful, but also not too little, otherwise at some point the data basis won't be enough.
I saw on your side that you also wrote a blog post about being able to display these thermals in relation to wind directions. So, basically, building in a filter where you say, okay, if I say it's a day with a northeast wind, where were thermals found on days with a northeast wind? So you can say, okay, everything that might be pushed up on west-facing slopes, which normally happens with a west wind, obviously isn't now. And they won't be displayed accordingly. Is that actually helpful as an addition?
It's also hard to judge. I really need to test it more over the course of the year. I think these wind calculations are very, very helpful, but I haven't built it in yet because I'm not sure how well it actually works.
On one hand, as I also wrote in the blog post, I was a bit surprised that the wind direction is actually almost only east or west with the thermals. This could be because, for example, the Alps are simply in the way and therefore like to deflect the wind to the east or west, so the prevailing wind direction is actually very, very often east and west and less north-south. But those are a few things that make me unsure about how well it really works. And you have to test it a bit first and put some energy into it to see if it runs well. The other thing is that with the wind, it's also not entirely easy to say, "we have this wind now," but we very, very often, especially in the Alps, in lower regions—meaning at lower altitudes—we have one wind, and then further up we have a completely different wind.
That means the classification isn't exactly easy, and theoretically, you'd have to classify the wind depending on the altitude. And then it also depends on how high the surrounding mountains are, for example. Whether the valley wind applies there or not.
That's why I think it's relatively difficult to just assume and calculate a single wind.
The wind you're taking into account now is a wind you've basically extracted from the thermals. Or where are you getting the wind data from?
Exactly, I'm currently getting it from the thermals' shift, because those are the data I have in that case.
And if you say that in the Alps, west or east winds mainly prevail, is it really like that—do most thermals drift either to the west or to the east?
Exactly, based on this calculation basis—assuming there are no major errors in it—it was very, very significant that there was really much more east and west than, for example, northeast or even north or south.
But that could really be related to the fact that with the east-west orientation of the Alps, there's somehow so much of an obstacle in the way from north to south that the wind would rather just veer off to the side. It might be interesting to look at that if you go to southern France, where you already have the Western Alps, which in many areas have a north-south orientation from their main core, to see if you might have a different thermal orientation mainly from the wind there. But those are data points you can probably just calculate. Do you actually have data from France and so on as well, or is it so far really just this DHV-XC, where you're saying, okay, this is the German perspective on the thermal data in the Alps?
So, the majority of the data does come from the DHV; that's about 400 gigabytes we have there now. Someone from the French community got in touch on Facebook regarding my post and gave me a French site that made the data very, very easy to extract. So you could download all the data relatively quickly. That added another 40 gigabytes of flight data. That means those are the two that now form the data basis for the thermals map.
Are you planning some kind of collaboration with XContest or something, where you say, "Hey, look, I have this great product, I can do some really cool calculations with it"? Is there any way I can get better access to your flight data?
So, if someone is listening who can make that happen, then I'd be happy to incorporate that, of course. At the moment, it's also a matter of time. I had a lot of time this winter and was able to put a bit more energy into it. Right now, it's become a bit less. But I'm of course looking forward to improving the project. And specifically, if the possibility were still there on the XContest side. That is certainly the best data basis we have.
You said you already use it yourself sometimes when flying. So you even have these dynamic data points displayed during the flight. What other instruments do you actually have in flight? What do you use?
I fly with the XC Tracer. That's still the two-seater, so the normal, classic one. And a phone, a phone with XC Track. And with XC Track, you have the option to display any websites. And one of these websites or one of these widgets is then the thermals map.
But you still have to navigate to it specifically. It's not like the thermals map works in a way where it automatically detects your flight location and always shows the surroundings directly. Based on the flight altitude that XC Track has already saved. Or can that really be automated to that extent?
Yes, I can already do that. So there's basically a secret side to it that allows you to do that. And then the map flies along with you. And it always picks up the new thermals, at least as long as there's an internet connection. And then you can see that live as well.
And has this map ever basically enabled a cross-country flight for you, or saved a cross-country flight that you might not have been able to continue otherwise because it showed you the saving thermals somewhere?
Not in the sense of saving me with thermals. But what it did do was during that flight in the Bavarian Forest this year. That was my first flight ever in the Bavarian Forest, and I didn't know much of the area. At the second turn point, it showed me again that there has to be something up ahead. There are several points there. And it wasn't until much later that I noticed some were circling. So there really is a point somewhere, there has to be something. There has to be a launch site. And that tempted me to pull the turn point out quite a bit further to catch the good thermals.
I used to use the Flyme software quite often as a flight instrument. I don't know if you know it. It already has thermals plotted as standard. A lot of people have tuned into those main thermals that it extracted from flight analyses. I have to say, that was very common and also accurate. So when you flew those points, it was always quite good. At the same time, if you have a bit of thermal or terrain understanding, you'd also see, yeah, clearly typical terrain there. If it breaks off somewhere, then it actually has to be up there. Nevertheless, this point was sometimes a quick orientation aid to say, where am I actually flying to? Do you think your dynamic map really still offers that many more advantages compared to something relatively static like that?
Yeah, how much of a difference that actually makes is hard to say right now because I still need to test it more myself. I do believe it still offers advantages. For one, these static points can only be static over the course of the day, as you already mentioned. So, just the differences between the east and west flanks, for example, can't be shown directly in them. Or simply the time of day, the season—that can't be factored in. Things like the snow line, which is often good for thermals, might not be included in the static assessment at all. But over the course of a season, you usually have a comparable snow line in the Alps. And that alone would already help a lot. So I do think it's better. Obviously, you can fly without it; you can hit all these things correctly based on experience or other factors.
But it's definitely another pointer, one more bit of help.
Are you
Are you actually a big commercial pilot? Yes, very much. I love flying commercial. I fly whenever it's possible. I've done some good flights, but of course, there are better pilots than me.
Has it always been like that? I mean, was your flying career always aimed at cross-country flying?
Yes, for a long time I was rather cautious and just flew the local mountain. I've actually always wanted to fly cross-country, but it took me quite a while until I actually did it and made some progress.
What held you back at the beginning? You said you were fearful. So, what kind of fear was it?
A bit of that came from my training and from a lot of what you hear in the community—the comments about all the risks there are in paragliding. Yes, I've taken that on board. I've also evaluated it for myself in a way that I was always very, very cautious. And because of that, it took me a bit longer than some others who maybe just tried everything as young wild ones. Maybe it was better that way. But maybe the long-distance flights could have worked out a bit sooner.
You said it took a bit longer. Maybe tell me briefly about your flying career. When did you actually start flying, and when did cross-country flying really come into the picture?
So I've been flying since 2010, and I actually started long-haul flying around 2016. That means for a long, long time I didn't even have a B-license. I didn't get that until, I think, 2015. And before that, there were many other things in my life that required my focus, where flying had to take a backseat.
So, you flew less, but at some point you said you wanted to do more again, and then cross-country flying came into the picture. Was there anything like a tipping point, a trigger, something that made the knot untie in your head, where you said, I'm really daring to fly away from my home mountain now and actually go cross-country?
So a particularly beautiful experience was the cross-country flight seminar at the Zugvögeln in the Dolomites.
We had excellent conditions every day, and flying there was just a dream. Especially under the guidance of the pros. That means I saw for the first time how someone who flies very, very well also thinks, and what they consider when the planning goes on for so long. And I saw what is possible. And that was a spark for me, just to say, I want to do more of this. It's incredibly fun. And it should work roughly like that. So I'm going to try making a plan like that myself and just fly around freely accordingly.
What was the most important thing you saw with the pros that really pushed you forward?
So for me, the most important thing was actually, first of all, to cut the rubber band that holds you back at the landing field.
And watching the pros, I think, it's this very clear view of the mountains, of the things that are happening, how the air is flowing, where the sun is shining. I can't pin it down to one single thing, because there are so many points you have to pay attention to. But I think just noticing how clear and structured someone proceeds, that helped a lot.
Now you have this, I mean, the cross-country flight seminar, that's probably a week you spent there in the Dolomites. After that, it's also about how can I implement this myself for my own flights? How did you do it and transfer that? Now, without the help of the professional, so to speak, lead pilots or yes, also weather tips and other people, where you say, I have to figure this out for myself now.
So it naturally required a lot of trial and error. I tried a lot and also ended up on the ground every now and then. I mainly stuck to the routes that are typically flown. So, what we just discussed, where many successful and good cross-country flights worked, from Kössen, on my Spreikboden, where you see other people. What always helped me particularly much is seeing people live in the air, so where I really notice, okay, a small group is flying in front of me now and I'm trying to keep up with them. And then you get a feel for it very, very quickly, which corners work well and which don't.
That means really just gaining experience, building that experience by basically copying other people, maybe flying behind them and so on.
Exactly. First, there's flying behind someone. What I've also done a lot of is flying together. Especially with Georg over the last few years. You pick up so, so much, which is also hard to describe. So things like, when you're turning in the thermals, when is it actually worth doing maybe one more search circle? When is it not worth it? When is it better to stay in? And you can see that incredibly well, also on-site, if you have someone right next to you who is out there with the same gear, roughly the same level, and so on, because you see immediately, okay, now it worked. Now he did something right and I should have done it better too. Or vice versa, what I just did was actually right after all. And that sharpens your senses very, very well for what works well when flying in thermals.
So, you guys went flying together often and really benefited from each other?
Yes, definitely. I believe we've both improved incredibly over these last two years by flying together, and that this waiting for each other—which you really, really don't want to do in the air—has paid off very, very often. You
they also really say they want to fly together. But that also means you're committed to saying we'll wait for each other so that it can continue together?
Within certain limits, of course, but yes, we definitely try to wait for each other.
If you were to summarize your entire development like that, is there anything in particular you would advise other pilots on how they can develop their XC skills?
The most important thing, of course, is to actually try it in the first place. I think it's very, very important,
yeah, taking a certain risk that you might be told not to take during your training. And it's always about that controlled risk—taking one step further, then evaluating again, was that good or was it not good. An example is flying; I fly into a leerlein. For me, as a long-haul pilot or a non-long-haul pilot, that was an absolute taboo. But if I don't fly into a leerlein, then I don't have much of a chance for longer routes. And there are many of these things—for example, in Föhn situations, where according to pure doctrine, you actually shouldn't fly at all. In many cases, it's still done anyway. And you have to figure out for yourself, and in general, how far is it okay? How far can you accept and justify it?
And then, of course, having that clear boundary of where it's no longer okay. And definitely not crossing it.
You mentioned earlier that it took a while before you started doing cross-country flying. Maybe that was actually for the best. How do you see that from today's perspective? Did the fact that you flew with anxiety for a long time and took such a slow approach to flying perhaps give you something, in the sense that you aren't just some fast-flying daredevil?
Well, at the very least, it's given me ten years of accident-free flying. And I'm very, very happy about that. And it's definitely true that the fun is at least as great as it was at the beginning. In fact, I have significantly more fun than I used to. And I think that's actually the most important thing you can take away from it. Almost no one in the paraglider community can make a living from it or is now financially very successful with it. That's why it's actually almost all about the fun. That's actually all just for the fun, I think. And if the fun remains, if it even grows, then you can say everything was done right.
You've had a great development from being a local mountain flyer to a relatively successful cross-country pilot with an ever-increasing theoretical background, also through the analysis of your thermal maps and other things. So, what are your next flying goals?
I don't really have any specific goals. Last year was so good that I'd like to confirm that again. So, I'm going to keep trying to fly long distances, assuming that's still possible now with Corona and the lockdown. I'd definitely like to fly at least one 200km route again this year. Maybe a 250km one will work out at some point. But the most important thing is to continue having fun with it.
And will you be using your XC-Track with your live thermal maps for all your cross-country flights now? Do you have a sort of advantage over others with that, because you can say, "Ha, I can still look at something here that you don't see right now"?
I will definitely use it. Whether there's an advantage and how big it is, only time will tell. So far, there hasn't been much time to evaluate that. It did do something good, as I said, in the Bavarian Forest. But I'm definitely going to keep using it.
Have you also ever flown in a competition, and could your software or this solution you have bring advantages to competition pilots, do you think?
I've been at the Newcomer Challenge before. I don't think it's that useful in that environment because there are just so many, many other indicators in the air. In competition, there are so many pilots in the air. The pilots show what is happening live right now. And if you can see what's happening at this very moment, it's still better than an analysis of the last few years. So in competition, I don't think it will take off in that form, because there are simply so many other options that are definitely correct right now.
Reality still beats any database.
Yes, absolutely.
Okay Ulrich, let's leave that as an outlook for now. Go flying a lot, keep an eye on the others. And if you don't have them, you can always still fall back on a database to help you prepare in some places. Ulrich, thanks for your explanation and great job on the work you're putting in there, providing us with a new resource, especially for flight planning and maybe even a real understanding of certain terrains that you can pull from. I'm curious to see what other additional info might come up eventually regarding glide paths or other possibilities you might find. Yes, definitely keep going. I'm excited to see what else comes out of it.
Thanks a lot, Lucian. And you definitely keep going. Your blog is very, very important and your podcast is always very interesting to listen to.
I'd love to keep going. And yes, now you're part of an episode too. Great.
That was Ulrich Scheller in conversation with Lucian Haas. If you're interested in the trend, the thermals database Thermalmap, you can find it on the internet at thermik.pumpt.net. On my blog Lu-Glidz, I've already described as an example how to use the tool. Ulrich Scheller also maintains his own blog, where he describes even more background information on this project. You can find the relevant links in the show notes for this podcast episode on Lu-Glidz. If you enjoyed this episode, please recommend Podz-Glidz. You can find this and all other podcast episodes on Lu-Glidz and Soundcloud, as well as in well-known audio catalogs like Spotify, iTunes, Google Podcasts, etc. To make sure you don't miss a new episode, you can also subscribe to Podz-Glidz right now in your podcatcher of choice.
And if you want to keep receiving more in-depth stories from the world of paragliding from me in the future, then please support my work as a patron. I don't fund Podz-Glidz and the associated Gleitschirm-Blog Lu-Glidz through advertising, hidden sponsorships, or subscription fees, but solely through voluntary contributions from my listeners and readers. Payments are very easy to make via PayPal or bank transfer. You can find all the necessary details on my blog Lu-Glidz, specifically on the Förder page. The contribution amount is completely up to you. If you're unsure how much you want to give, I suggest a non-binding one euro per podcast episode and two euros per reading month on Lu-Glidz. Even when added up and calculated for the year, that's significantly less.
than a subscription to a classic paragliding magazine costs. Stay healthy and land upright. Ciao.