Good pilots, they just have an extreme amount of experience and they look at the sky and can judge the current weather much, much better than people who simply lack that experience. So even if I might understand how an alpine pump happens on paper, if you don't know which pass it's happening over and what the signs for it are—visually, how to recognize it—then it doesn't help me if I've learned it on paper; I really just have to get out there and fly and experience it, and that's why I believe there's still an incredible amount for me to learn in this specific area.
Podz-Glidz, the Lu-Glidz podcast.
Stories from the cosmos of paragliding.
Today with...
Stephanie Westerhuis.
And I'm Lucian Haas.
In paragliding, the weather is always the biggest unknown. Nowadays, you can study half a dozen weather models on the internet to get a rough idea beforehand of whether, where, and how safely you could take to the air on a specific day. But what actually awaits you on-site is sometimes completely different from the forecast.
Anyone who ends up disappointed and struggles with the weather models might lack some understanding of how these models actually work. In fact, the models still handle many things quite crudely. And that's to be expected. Their results are a great help, but one should keep in mind that they have a wide margin of error. A woman who has a good professional grasp of this is Stefanie Westerhuis. The 30-year-old works at Meteo Schweiz on the development of Meteo models. And she also flies a paraglider. Quite successfully, even. In 2022, she finished 4th in the women's ranking of the Swiss XC League. During her flights, she regularly feels and sees where the models are accurate and when they hit their limits in reality.
That's what this 106th episode of Podz-Glidz is primarily about. Dealing with the weather and forecast uncertainty. And why you should take off for ridge forecasts much more often.
If you enjoy Podz-Glidz, then please support my work as a patron. You can find out how to do that easily and without any obligation on the website of my paraglider blog Lu-Glidz, specifically on the Support page.
Stefanie, in your case, are meteorologists actually the better pilots?
Definitely not.
Why not? Actually, you already know a lot about the weather through your job or your studies or something like that. And if you know a lot about the weather... Or for flying, you need good weather knowledge. So why aren't meteorologists definitely the better pilots? I mean...
I think having a meteorological background definitely helps. But especially for paragliding, you have to gather so much knowledge on small scales. You don't learn that in your studies at all. And people always have this feeling like, "Well, you're a meteorologist, you know everything now." And I always think, "Oh help, I've only learned and acquired all of that in the last few years myself."
And I think the best pilots I know aren't really interested in meteorology at all in some cases. They can just digest some general forecast texts and then know pretty quickly that Wednesday is the day of days and that I can fly this and that.
And they don't look any further into the weather either; they just look at one forecast and see, I know, this large-scale weather pattern, that's just flying weather. And I just make the most of what it is. Yeah,
I'm not saying they don't look at it at all, but I know some good pilots who spend way, way less time than I do looking at weather maps. They have so much experience; they know much more about which flights they can execute in which weather conditions.
You're saying they use much less time than you. How much time do you spend on... if you were to say, I want to go flying today, what would you do in the morning... how much time do you actually still invest in going online and checking weather sites?
So in the morning, I hopefully don't invest any time at all or only very little, because if I want to do a good cross-country flight, I actually have to have everything prepared by the evening, and then I get on the train at six in the morning and the plan already has to be set, where I want to go and roughly what flight route I want to take, and that's why it's mandatory that I look at it by the evening at the latest, and that can easily take an hour or two with everything included. So I'm not saying that I prepare for two hours every flight day, not at all, but if I somehow know that I have the coming weekend off, I can go flying, then
It does happen that three or four days beforehand, I spend half an hour every day looking at weather maps, weighing things up, and then discussing them again with my flying buddies, and yeah, in addition to that, it might be two hours until Friday evening, and then on Saturday you have a plan.
How far ahead in the Vorrhein do you look? So, if you say you want to go flying on the weekend, would you say it's only really worth checking three days in advance because the models usually still change so much in their forecasts that it's not worth investing much time into it on Monday, or do you really look at the whole week and then sort of obsess over it?
There are like two decisions to make. One decision is which day am I going flying? I have the luxury of having pretty flexible working hours, which means I can already think about which day next week might be good, and then I check my agenda to make sure I don't schedule any meetings there, and for that, I sometimes have to look at the upcoming weather four or five days in advance, but I actually never make the plan for the next day that far in advance, because I'm aware that the weather forecasts are always still being adjusted, especially in the details.
If you had prepared well last night, do you also really check, for example, if you're heading to the train station now and then to the flight area, or if you're going up by train or by mountain railway, do you also check, for example, a satellite image again or do you look at...
real weather station data to see if it actually matches the forecast I put together last night?
Definitely, so looking at the measurements is omnipresent. I do almost all my paragliding trips by train, and you usually have one, two, or three hours until you're actually at the launch site, and right then, as soon as the sun has come up, you want to know if it's really sunny now or if it's still cloudy. We often check the apps for that. What I also definitely check again and again are the precipitation forecasts; especially when a day is coming up where many thunderstorms are predicted, I'm aware that it changes every time a new model run comes out, and I check that regularly as well.
But in the end, looking out at reality is actually almost more important than looking at the model values. For example, let's say I want to go flying in Fisch in Valais, and I know I want to go up the Goms towards the east first and then back down, and depending on how high the base becomes, I can go further towards the Furka, or depending on whether the Grimsel from the Bernese Oberland is already too strong, I have to turn early or not. I can't decide all those things in advance. So the model gives me an indication of how it might be. Should I expect a strong Grimsel or rather not?
But in the end, it doesn't matter if I stress about it too much, because I won't know until I'm actually there—whether there's that 20-knot north wind over the pass or not. So what I look at most are the satellite images or
Sunshine duration measurements. Especially on days when it's not so clear if it's actually going to be nice from the start.
Do you also check those kinds of apps again during the flight, or look for a satellite image or a rain radar to say, hey, is that just an isolated thunderstorm I'm seeing there or is it moving towards me? So, do you still use that kind of information, or do you really just look out with your eyes during the flight, see the weather, and that's all you pay attention to from then on?
No, clearly not. While flying, I look at two things. One is wind values. So, especially when I'm flying in the Alps and have to account for a lot of valley wind, I want to know: is it 10 km/h or 45 below me? Or if I'm flying over this pass now, what's waiting for me on the other side? Or, taking the Grimsel example, I'm already, let's say, 20 km ahead, and then I look at the measured value and decide continuously—as I get closer, I keep checking the measured value so I can get a picture of whether the pass is already flooded or not. So, the wind values are one thing, and the other is
Radars, so...
Precipitation data—and not the station data, but actually the radar measurements. On convective days, I find it difficult to be up in the air and have to decide around you, are these large cells or small cells, and are they moving fast or not moving that fast, and yes, what's developing there, because thunderstorms, for instance, can become quite critical for us, let's just say. I want to have a bit of a picture of what that looks like well in advance.
What apps do you use while flying to be able to check things like that quickly in real time?
Yes, I'm a Meteor-Schweiz kid, of course, so I primarily look at the Meteor-Schweiz app. It has wind measurements, and the radar data there is also quite well-presented. In addition to the Meteor-Schweiz app, I also use winds.mobi. A colleague from the Jura developed that. He simply collects all possible wind measurements and displays them on a map. It has many Holfui stations in Switzerland, or also private devices that were set up somewhere. It definitely has many more measurements on there than the Meteor-Schweiz app.
With your experience, on one hand your meteorological background, and then your experience with the apps you constantly check and all that, how often does it happen that you end up realizing, "Here, I've made a wrong decision from a meteorological perspective"?
Yes, all the time. I think it happens to me way too often that I let myself get intimidated. I'm actually flying and everything is fine, and then I see somewhere that 35 km valley wind was measured in Fiesb and then, yeah, where am I supposed to land now? That makes me uneasy; I don't want to be in that strong wind. And then I have to calm myself down a bit and say, yeah, just because this station right there is 35 km doesn't mean you can't still find good landing spots in the nearby area. I think that's more my problem than me completely misjudging something these days. Could
could it be that people who obsess too much over the weather and always pay too much attention to those kinds of values aren't necessarily the best pilots, like we had at the beginning?
I don't think it makes you a bad paraglider pilot or a Gleitschirm-Blog per se if you deal with a lot of measurement data. I think you just have to find a good way to handle it. I think you have to really engage with how representative the measurement I'm looking at actually is. After all, I want to know what the flight conditions are like, and just because a measuring station shows a numerical value, it's not that trivial to decide or imagine what the air is actually like around it.
What has been your, well, basically your biggest mistake in weather assessment in your flying career so far?
Definitely one specific incident where I really misjudged things. That was in the year 2000,
2017 or something. Yeah, I was in Innsbruck for half a year, doing an internship at the university there, and in Innsbruck, there's the Nordkette for flying, and that's not really the kind of place where you can just go flying every day, because there's still the airfield and the valley itself isn't known as a peaceful, happy-go-lucky flying area, but in February I wanted to test a glider from AirDesign; they have their production site nearby. I didn't live very far from it, I went there by bike, borrowed a glider, then biked back and up to Seegrube at the Nordkette, and I only wanted to do a slide, and I only had my small Hikenfly harness with me, and I'd even forgotten...
to take the carabiners I normally use to attach my reserve parachute, and I also thought to myself, yeah, should I just pack the reserve in my backpack? I'm only doing a glider, what can really go wrong? And then I thought, oh no, I'm not doing that anyway, and then I very laboriously put the reserve into the
I hung the carabiners where I usually hang them, so it wasn't really a good setup, and then I looked at the readings in the area; there's the Patscherkofel opposite the Seegrube, and there I saw, yeah, it's measuring some kind of, pff, I don't remember the details anymore, 40 km/h wind or something, and I already knew it was tending towards a south wind for this day, but the Patscherkofel is significantly higher than the launch site where I was, and at the launch site itself, I had some lift, some sink, some lift, no wind, and then I thought, yeah, it's like at home in the Jura, when the high station shows strong wind, then we just go 200 meters lower to the lower launch site and we'll be able to launch out of there without any problems.
and what happened next is, I took off and flew out, and at first I was still happy that my variometer was beeping because I was a bit scared of whether I'd actually make it to the landing field, simply because I'd never been there before, and then my variometer kept beeping more and more and beeping more and more and at some point I thought, yeah, actually I want to go down already and not just up, and in the end, yeah, it just blew me extremely high up because the wind was way too strong and I then tried, with this glider that I didn't know either, to somehow put on some brakes and do a baseball, and in the end, yeah, I had to, I'd resigned myself to it,
that it only goes up from now on and the wind gets stronger at the top, and then it blew me over the Arztlerscharte at 70 kilometers per hour and I ended up landing in the Karwendel, yeah, after it gave me quite a shake, and then—well, it was February—I landed at the back with a pulse of like 300, just glad to be alive, and in the end, I still had to call the helicopter because there was so much snow and then, once I had landed and parked and calmed down a bit, I called my colleagues at the university like, "Hey, can I walk out? I'm back there somewhere in the Karwendel, over there and there," and he said, "No, Steffi, that's not possible."
Either you call the helicopter or I call the helicopter, that's way too dangerous. And then the police helicopter picked me up and basically flew past the university to the airport, let off some pressure, and then I was back at work in time for lunch. Yeah, and so, this incident definitely shaped my later flights because I really completely misjudged the situation. So the assumption that the wind is stronger on the Patschkuffel, but the Patschkuffel is also higher and that I can still take off from there without any problems, that just didn't hold up at all. And even many years later...
I struggled with the fact that I'm sometimes afraid in the air when I experience something that I can't fully explain from a meteorological perspective. Because I kept thinking, wow, maybe I'm in that situation again where I think A is happening, but actually it's B, and B would be very unfavorable for me in this situation. And this experience of having misjudged something so, so badly stayed with me for a long time.
Was that also kind of embarrassing for you, because you're saying, "I'm a meteorologist, I should have been able to assess that better"?
Yes, maybe a little, but in the end I always say, I mean, I did study meteorology and climatology, but that's very far removed from the knowledge you acquire when you go paragliding. So, I attended lectures where it's about chemistry in the stratosphere and some climatological indices, but you don't learn at university how to assess small-scale meteorology in situ.
What conclusions did you draw for yourself from this experience?
That's a difficult question because it was quite some time ago.
What I took away from that is, as I said, first of all this fear that stayed with me for years—that I'd completely misjudge something again. And what clearly helped me against that was the weather briefings with my pilot colleagues. I have several people in my circle who also really enjoy dealing with meteorology, and I think even today I learn the most from looking at a forecast or observing a current phenomenon and then perhaps discussing it with these people that evening.
And I think I've simply also learned that there is still so much to learn.
Yeah, that's the beauty of meteorology, you never stop learning. And every time you think you've understood something, you're standing at the launch site again and suddenly something happens, like the wind picks up. You stand there thinking, why now? I don't understand it either. Did your interest in the weather actually only come with aviation?
I can't really answer that very well. It was like two parallel stories developing. Even as a little kid, I was allowed to accompany my uncles and great-uncles who flew paragliders back then. And for me, it was clear that I'd do it too someday. But I didn't turn 20 until I signed up for the flight school. And at the same time, independent of those interests, I had already signed up for my studies. My degree was in Environmental - Natural Sciences. That's pretty broad. And for me, it was always clear from the start that I definitely wouldn't do that specialization in climate, because everyone wants to do that. And somehow, I like trees. And then I attended the first dendrology lectures and realized that you have to memorize quite a lot there.
And memorizing things was never my thing. On the other hand, all the programming and the equations and the numbers, that always grabbed me and it suited me well. And that was exactly this specialization in atmosphere and climate. And then, pretty much at the same time as signing up for the paragliding school, I decided to do my master's in this field as well. So, maybe subconsciously, it already kind of fit that I'm interested in paragliding on one hand and at the same time chose my master's specialization in, well, meteorology. But in my active consciousness, these were two completely separate decisions.
Today you work, as you mentioned earlier, at MeteoSchweiz.
What are you working on today, or what's the topic you deal with on a daily basis?
So, I did my studies at ETH, and ETH is quite close to MeteoSchweiz because both institutions are in Zurich. I had a lecture I really liked where someone from MeteoSchweiz gave a few lectures. Then I had to decide on a master's thesis, and I thought, well, I'd actually be interested in doing something that doesn't concern basic research, but rather focuses on applied meteorology. So I wrote to this person and got the opportunity to do a master's thesis at MeteoSchweiz. At that time, I was dealing with the selection of ensemble members that we then use to drive the MeteoSchweiz model.
And I liked it so much that, during the course of the work, I asked if there was a possibility to also do a PhD. And then, during my master's thesis, we wrote a research proposal on the topic of improving fog forecasting with a meta-model. That was accepted, and that's why I had the opportunity to deal with fog forecasting for four years. And the topic of fog wasn't actually my suggestion; it was my supervisor's suggestion because he knew that fog forecasting is something that will challenge MeteoSchweiz in the future, or where they still have a
I liked it so much that, during the course of the work, I asked if there was a possibility to do a doctoral thesis as well. And then, during my master's thesis, we wrote a research proposal on improving fog forecasts with a meta-model. That was accepted, and that's why I had the opportunity to deal with fog forecasts for four years. And the topic of fog wasn't actually my suggestion; it was my supervisor's suggestion because he knew that fog forecasting is something that will challenge MeteoSchweiz in the future, or where they have a... let's say, a major shortcoming. Or the models currently have a lot of trouble making accurate fog forecasts. Personally, I would have preferred to deal with wind if you had asked me at that point. But then I thought, yeah, fog, why not? It's actually good too. And then I struggled with fog for four years. Then at the end of my dissertation, my team leader approached me and offered me the chance to stay with them and continue developing models. And what I'm doing currently is helping out in a project where we are replacing the Cosmo model with the ICON model.
So Cosmo is a model that many people in the German-speaking world probably know. It's been around for 20 years.
And the German rescue service has already switched to ICON. ICON is a newer model. And MeteoSchweiz still has to make that switch. And there are so many things to test and optimize. And I'm active in the team that deals with that.
Since we have a model development expert at the mic right now, we could go over some basic things regarding weather models—typical questions that pilots always ask when it comes to weather models. Maybe you can help clarify a few things there. A very simple question that I keep getting from people who say, "Ah, Lucian, and now Stefanie, which is the best weather model?"
Of course, ICON is the best. So, no, ICON is really a very good model. The German Weather Service has been calculating it globally and regionally for some time now. And they constantly compare it with the European model, the GFS, and a Canadian model. There are several scores that you look at. But that's not what interests us pilots first. Those are scores like, for example, the temperature at 1,500 meters or the 500 hectopascal geopotential height, or things like that.
But in the end, there's a good overall picture of a model's performance. And Icon actually finishes very well there. But if you then ask me, "Yeah, what's the best model for my flight day tomorrow?" then you have to remember that, for example, thunderstorms are very dependent on what initial conditions you give the model. And Meteo Schweiz is currently investing a lot of resources in this specific area to generate what they call a better analysis. Because we say, if the model's initial conditions aren't good, then the subsequent forecasts won't be good either.
No matter how high the resolution of the model is.
Let's talk about how the data gets into the model; let's go over that again right now. We're back at model selection. You just said ICON is good. But there are different types of ICON models. There's a global ICON model, there's a European ICON model, where all of Europe is mapped out. Then there's the ICON D2, calculated by the DWD, where Central Europe around Germany is mainly calculated with a 2km grid and so on. Can I basically assume that these more finely calculated models are always the more relevant ones for me as a pilot? Should I always look at the fine models, or are there situations or specific analysis scenarios where I'd say it's much more sensible to take a slightly larger model, i.e., with a larger
...model grid and then, for example, looking at an ICON EU instead of an ICON D2 or something, or just the ECMWF or something like that. Instead of the ICON D2, where I'm saying, okay, the grid, the point spacing is a bit larger. Can that still have advantages over the finer one?
I don't think so. If you're interested in tomorrow, I'd take the model with the highest resolution that's running for tomorrow. Because Icon D2, for instance, is a regional model, so it doesn't cover the whole globe, but the boundary conditions come from the global Icon. This means any information you find in the global Icon is then fed into the higher-resolution model. And MeteoSchweiz, as per tradition, doesn't use the global Icon, but still uses ECMWF as the global model, and all the information you find in the ECMWF is then indirectly fed into our regional high-resolution Cosmo, which we're still using, or in the future, Icon.
That's why I don't think you really get any added value by taking the time to look at the coarse-resolution models for a day like tomorrow or the day after. On the other hand, when it comes to making a decision, like "yes, I'll take next Tuesday or Wednesday off," you have no choice but to use the coarse-resolution models, because the high-resolution models only run for about 33, 48, maybe just under 50 hours, but they are so expensive that it's not worth running them for five days.
If you're looking far ahead, say seven days out, do you still say Icon? I think that's right at the limit for the Icon EU model at least, so I'd still check that, or I'd definitely take an ECMWF model as the major European model, which might provide the best data there.
Personally, for me, it hardly depends on the model at all, but simply which model is available to me in the graphics I like to look at. And especially when we're discussing where I want to fly to in seven days from today, you have to be aware that the forecast accuracy can be very low, depending on the weather situation. In that case, it might be worth looking at two or even three models, but not in detail. For example, Windy has excellent options for switching between models, and then you go out six days and take a look at the ECMWF, then the GFS, and then maybe the ICON.
And then you see, yeah, it's more high-pressure dominated, okay, maybe good weather, or very clearly low-pressure dominated in all three models. Yeah, that means I don't have to block my agenda now and try to reschedule meetings. I think people who don't deal with models that much are a bit too trusting of them. They can sometimes underestimate how large the inaccuracy actually becomes once you project more than three or four days into the future.
When is it actually like rolling the dice? From which day? Can you put it simply?
Yeah, you just can't say that so simply. If you have so-called Omega high-pressure blocked situations, then you can predict the weather more or less for the next ten days—namely, nice and sunny. But as soon as you have any peripheral lows, then part of the day, like tomorrow, is already pretty uncertain. What I do is put on a bit of a lens that works more and more accurately. So, for example, I'll look today at when I want to go flying sometime next week, and then I'll see which days could potentially develop well, and then I'll mark them in my head, and then the closer it gets, the...
I also look at the forecasts more closely. But it's not worth looking at the details of what the wind forecasts are at ten meters high somewhere in a valley five days into the future, because that's going to change ten times anyway; it's more about estimating the general weather situation. And I find it amazing every time how bad the models can be, even just predicting the weather for the day after tomorrow. Especially when it comes to details like the snow line or cloud cover, which are quite relevant for us, you're constantly surprised by how inaccurate the forecast actually was. And it's happened to me again and again that I...
I somehow knew, yeah, there were good flying days coming up this week, and then I could only take Wednesday off, so I just sort of accepted it, yeah, I'm going flying on Wednesday. And then on Monday and Tuesday, the others come into some chat groups and say, yeah, but Thursday looks much better in my case, and then I have to disappoint them. No, now I can't take anything off on Thursday, I'm just doing Wednesday. And then it turned out afterwards that Wednesday was the much better flying day because some clouds moved in at some point on Thursday. I think it's important to deal with where and when a good flying day could occur, but you mustn't let yourself be influenced too much by the forecasts; instead, always think a bit in terms of scenarios.
So, a plan A for an optimal scenario, a plan B maybe if there's a bit too much wind, and a plan C, and then just see what you can do on the fly. Because the models, they're good, but there's also a lot they just don't handle at all. For example, something I still find quite striking are all these dust layers. I mean, none of the models we normally look at take into account the Saharan dust being transported to us from the Sahara. And there are other models that look at exactly that, but when we consider what the implications are for us, there are days that practically scream for 300 kilometers and then it just
the densest cirrus clouds and it might even be flyable, but in the end you have to admit that the flight day on the model you looked at—the model you looked at—looked much, much better than expected, simply because the Sahara dust isn't being taken into account.
Why aren't these kinds of things simply included in such models, when we know they can be so weather-determining at times? Because, I mean, the weather services do pick up on it, and they'll sometimes say, "Yes, dust is coming here," and then it can be much hazier and so on, but the models themselves still output the much more optimistic forecast accordingly. Can't these kinds of dust forecasts simply be integrated into such models?
So, the work on that is ongoing, but Sahara dust specifically isn't that easy to get in. Because for Sahara dust to end up in our regional model from the Sahara, it already has to be in the global model. And calculating a global model is very, very expensive, and Sahara dust is something very, very small—it's on a scale that is never resolved, so it's not primarily worth calculating there. And if the Sahara dust isn't in the global model, it can't land with us in the first place. So, if you imagine how clouds are represented in the model, you realize that while Sahara dust...
It would be nice to have, but it's still a bit of a pipe dream because, currently, models are pretty dumb when it comes to clouds. They just know that my grid box, which is maybe a kilometer long, a kilometer wide, and 40 meters high, has water in it or not. So they know it has 0.4 grams per cubic meter of water or none at all, and then there's something called a parameterization or a representation for sub-grid scale clouds—so they try to make an assumption, based on temperature and humidity, of how many, let's say, cumulus clouds it might have.
But we are far from calculating droplet counts in our models. That's something that's already being done in research, and if you want to look at something in great detail, there are models that do take into account cloud droplet counts and the distribution—whether they are small droplets or large droplets—but our operational models simply know whether there is water or no water and how much water there is. And this grid box is huge, and if you then consider that you'd also need to have the Sahara dust, and the Sahara dust would then have to somehow trigger the condensation of water on these dust particles, that's something that will take a long time before it's really mapped out in detail.
Meteor Schweiz is one of the few weather services worldwide that also runs a very, very high-resolution weather model. The Cosmo 1 with a one-kilometer model grid spacing—for me, the question always arises: okay, you have such a nicely high-resolution model, but there isn't a weather station every kilometer to provide me with the data I can actually feed into the model to say, "these are my initial data points that I'm starting my calculations with." How do you actually populate such a fine model with the initial data, which is probably still quite coarse in the end?
That is a really complicated science. I don't work in so-called data assimilation myself, but that is one of the areas where Meteor Schweiz invests a lot of resources, because that's exactly the crux of it. You have complex orography, a few measuring stations, and you have to achieve the most accurate initial state possible.
What is done is, you have all these measurements—two-meter temperature, two-meter humidity, ground pressure—then you have primarily the radiosondes, and then a one-hour model is calculated, and then you correct this one-hour model run with all these data. This corrected model state is then the initial state for my operational forecast. And it's actually quite striking how few data points you need to still get a pretty good model state. But that's also based on the fact that the physical
...conditions are mostly correctly represented. I mean, the ratio of how much moisture condenses at a given temperature is correctly implemented; we don't need to discuss that. It's more about the difficulty of actually localizing all these measurements—meaning, you have to consider, for example, how this two-meter temperature in Fisp will affect the temperature profile in the neighboring valley, but at an altitude of 3,000 meters. So, the whole machinery behind it is quite sophisticated, quite complex, and also very resource-intensive.
But if you look at it from an overview perspective, what would you say?
...are ultimately just estimated or interpolated data, because you say, okay, I have two weather stations seven kilometers apart, and everything that happens in between, I still have five points in between there, and I just roughly fill them in.
I think far fewer measurement data are currently assimilated than people initially think. For example, the two-meter temperature and humidity weren't taken into account until not that long ago. Radiosondes have a very large influence, as they provide a vertical profile, and then there are also the physical conditions—simply the fact that the sun, let's say, rises at 7:00 a.m. and sets at 7:00 p.m., and the sun always has the same amount of energy acting on the ground, and all these processes are quite dominant and can be represented accurately.
But in that regard, one can certainly hope that many things will be improved in the future. I mean, MeteoSchweiz now has some measuring instruments designed so that you don't just consider two-meter values, but that you also take into account the boundary layer—the lowest few hundred meters—
scans, let's say, temperature and humidity profiles, and that these are then fed into the initial states of the models.
But how many of these scans are planned, for example, for a region like Switzerland alone? Is it five, is it 20, is it 150?
Yes, rather five, in terms of magnitude. So we have a radio sounding in Switzerland, there's one in Bavaria, and then the radio soundings of the surrounding countries are taken into account, and for not that long now, profile data from these so-called
microwave radiometers are taken into account, and they are additionally located in Gränchen and Schaffhausen, as well as in Bavaria. And specifically in the inner alpine region, we don't have any such devices being assimilated yet, because it's very complicated to ensure that the model doesn't end up worse than it already was before.
If you imagine it like an average person, you'd think, "Hmm, I've got a radiosonde, maybe three more microwaves, or whatever you just called them, looking upwards and providing me with data, and now they're calculating the weather for me to the nearest kilometer—whether it's good or bad or whatever." At first glance, it sounds like you'd think, "Well, you'd have to be lucky if they actually hit it in any way."
Yeah, that's how it is. So currently, we're looking at...
...give master's students a few mountain meteorology cases, and it was interesting to see, for example, that we start a model run, let's say at noon, and then it runs through the night and we simulate the following day. The night itself was simulated pretty poorly. There were cold air pools in the alpine valleys, and the models completely underestimated them. But as soon as the sun comes out and the valley walls are essentially heated up, the observation data and the model data match up much better again. And that's what I meant earlier by the basic physical conditions being
...properly represented by such models. So, you have the sun, then you have the ground warming up, and that leads to a low-pressure system—a heat low—over the Alpine region, which sucks air in from the surrounding countries. You can reproduce all these processes quite well, even on a relatively coarse resolution grid. And even if you didn't represent the cold air pool correctly at night in detail, or if the measurement wasn't there to correct for that cold air pool or something, because the broad determining processes are represented correctly, the forecasts aren't all that bad either.
One thing that meteorologists do a lot today, which you also mentioned earlier, are so-called ensembles, where a weather model calculates with various different initial data—so small changes—and then we simply run, say, 7 runs, 10 runs, 30 runs, or something like that and look at the parallel solutions. How much they vary. That's what the weather services do to be able to estimate a bit how certain we actually are about what we're telling people. But if I use Windy, for example, or something like that, I'm always only presented with one run. I never get ensembles displayed in Windy to be able to estimate, okay, it's showing me rain now, but maybe that was only in this one run and in 10 other runs it would be dry. I can't see something like that. Is there any way...
where I, as a user of Windy or other things where I only ever get individual model runs, can still estimate how likely it is that it will actually happen like that?
Yes, the dissemination of ensemble information is really shockingly poor, I'd say. That's certainly also a bit of a political discussion, because there's never really a consensus on what you're allowed to show users so that they don't get overwhelmed. For example, on the Meteo page of the Swiss Hang Gliding Association, we've shown some products from Meteo Switzerland where these model uncertainties are mapped out. And other sites I know where it's just freely available are, for example, the Weather Center; there, there's the option to...
to look at so-called spaghetti plots or also on
On Windy or the ECMWF website, you can look at so-called meteograms, and then you see, for example, a temperature forecast, but not just a single line, but also these boxes above and below it, which provide information on how uncertain the forecast is. But I completely agree with you; I would also hope that in the future there will be many, many more products like that, where you and I can simply access these model uncertainties on the freely available sites. And if you look at the MeteoSchweiz app, for example,
scolds you and you look at the point forecasts, you can see that there have actually been these uncertainty bars for not that long. So Meteo Schweiz is constantly trying to introduce this, to show the model uncertainty, but I'm also aware—I mean, the temperature uncertainty four days in advance at a point, at a single point, that doesn't help me assess whether this day is a good or bad flying day now, because the 2-meter temperature doesn't tell me anything in the end about whether the cloud cover I see is forecasted or not.
So there's still definitely room for improvement, I'd say.
One thing I like to do in the morning—which is why I asked earlier what you check in the morning or something—is that when I look at the weather in the morning, I also take a look, relatively broadly, for example, at Europe on Windy and then pull up the current rain radar to see where it's currently raining. I also pull up a current satellite image to see where those clouds are right now, and then I compare that with the respective model forecasts. So, I just switch back and forth between the rain radar and the rain forecast for the exact same time to see, okay, do they roughly overlap or are there big differences? Also, just to say, "Hmm, that's a huge difference." Actually, I probably can't trust this model run for today, at least for the next half day, because somehow the model is lagging two hours behind or is two hours ahead or whatever, so that I can also decide a bit which model might be the interesting one for the current day. Yeah, just checking the real situation against the model that is perhaps the best fit for me for the day, where I say, "This one represents the current weather development best."
Do you think that's useful?
Yeah, definitely. I personally rarely look at five different models in the morning on a day when I'm already on my way, but what I do do is, I still have the forecasts from the previous evening in my head, then I take the train somewhere and I see, oh, there are many more clouds than expected or, uh, it's somehow snowing or there's a lot, so right now also much more rain than expected, which can also be a real thermals damper, and I just let that flow into my planning for today. So I'm not the type to completely scrap my flight day now
throw out of the window, so by the time I'm already on my way, whether model A or model B was better isn't really that decisive for my flight day anymore, but I just take note of it—ah, you thought it would be like this and the base would be like that, but yeah, it's raining much more than expected, so maybe you have to account for taking off an hour or two later than planned, and whether a model got it right or not isn't really relevant anymore because I'm already past the point of looking at the models; I'm much more in the now-casting area, and that is definitely one of the points where I
have a lot to learn, even though on paper I'm a trained meteorologist and climatologist. So that's probably what I meant by, yeah, trained meteorologists aren't necessarily the best pilots, because good pilots just have an extreme amount of experience; they look at the sky and can judge the current weather much, much better than people who simply lack that experience. So even if I might understand how an alpine pump happens on paper, if you don't know which pass it's happening over and what the signs for it are—visually, how you recognize it—then it doesn't do me any good if I learned it on paper. I really just have to get out there and fly and experience it, and that's why
I think there is still an incredible amount for me to learn in this area as well—observing, analyzing, and making accurate assessments.
How often do you do the weather for others, like the weather forecast, where you say, "Guys, we're going flying today, and let me explain what to expect"?
It's incredibly rare. For example, my father also flies and he wants to know regularly if it's flyable tomorrow at the home mountain or not, and he always comes with the Meteo Blue meteogram because you can see the clouds so clearly there. Then he just looks at the grid point of our home mountain and says, "Yeah, there's a southwest wind there, that works, I can go flying today." And I'm like, "No, you can't go flying today, haven't you looked at the other maps?" I mean, it's possible that the Meteo Blue model predicted some southwest wind exactly at that grid point, but if you look at the big map, you see that there's simply way too much wind and much more west than southwest, and you can't handle that today.
So, I check the weather for him sometimes, and I now organize an XC League Women's Weekend twice a year for the Swiss Hang Gliding Association together with Gabriela Jakober. There, my main task is deciding where we go flying and maybe when, and I also try to give a weather briefing, but I have to say, I have extreme respect for people who organize and manage competition tasks and then make decisions for hundreds of pilots on when to start, which flight route, briefing, debriefing—it's pretty
nerve-wracking when you're not just responsible for yourself, but suddenly for others.
Are you extra cautious in situations like that? I mean, do you notice yourself thinking, "Hmm, I'm not entirely sure, I might go flying, maybe I can still assess that, I'll take a look," but out of the 30 women who might be with you at such a weekend, not all of them have as much weather experience as you do. Many might already fly a lot, others don't fly that much yet and don't have that experience; they lack the experience to recognize weather developments immediately, which you can see right away with experience. So, do you basically make a very cautious, yeah, a very cautious weather forecast to tell people to land earlier rather than something happening?
Mhm, so that's very clear. I just try to be as transparent as possible about everything I'm thinking myself. For example, we had a female pilots' weekend in Valais, in Viersch, about two years or so ago, and it was announced that there would be a south wind. I had the feeling it would be flyable, but you just have to keep checking the measurements so you aren't suddenly surprised that the wind has shifted in the valley and is suddenly way too strong. I presented it that way, and in the end, ten people decided to stay on the ground in the valley and went for a hike, and there were four of us in the air. We were also traveling together and...
constantly monitored the wind values, and in the end, we had a great flight, landed, and were happy. But I try to make people aware of things, like, I think the scenario will develop this way and that way, but I'm unsure here, and I'm unsure there, and I'll look at this, and I'll look at that, and now you have to decide for yourselves what you want to do. And that has actually always worked very well so far, and especially in our women's group, there was somehow never the problem of someone forcing themselves into the air when they didn't actually feel up to it, because, well, we always try to give everyone the feeling of, hey, we're just making the best of the day, and in the end, if you didn't fly, that's not a problem at all, because we still have so many
other, let's say, aspects that we look at together during this XC -Liga -weekend, and whether or not you flew those 10 minutes today is just the smallest part of it.
Do you get the impression that, especially at these women's meetups where the women are among themselves, they also—I mean, do women have a different feel for the weather than male pilots?
I don't think you have a different feel for the weather, but I think it shows that, on average, let's just say, women are simply a bit more cautious and reserved than the men. And I keep that in the back of my mind when I'm planning something like this—so if I'm going with my two buddies I've known for 10 years, and they're both men and a bit of the daring type, we plan different flight days than if I know, yeah, we're women, we're a women's group, and some of them have only had their pilot's license for six months. But then it's maybe less
not the fact that they are women, but rather that they simply have less experience and that I have to make sure that everyone still feels included.
Let's do one last topic change. It also has something to do with the weather. In 2019, you were a backup supporter for Kriegel Maurer at the X-Alps. Did you do the weather forecasts for him then too?
Very, very little. So, the people who asked me, "Yeah, how did you get into that? Did you do the weather for him?" they always think I was busy with weather forecasts all day. But actually, I did everything else. I was a substitute supporter because his original second supporter got injured at short notice, and I already had experience supporting those long hike-and-fly races. And yeah, what you basically do is, you primarily always have to make sure you're parked in the right spot with the car. Then you have to make sure there's always enough food. And thirdly, you have to make sure the clothes are washed and dried as much as possible, so that he can always change into his favorite shoes.
And there are situations where you sit down briefly the evening before and say, "Okay, so tomorrow the weather is this and that." And yeah, there are several options to keep moving in the right direction. And if it's like this, then we'll do that, and if it's like that, we'll do this. But Krigl is also one of those people who simply has a massive wealth of experience
Meteo situations and paragliding. You don't really have to tell him much. He just goes out himself and observes what the weather is doing around him. And then he makes the right decisions.
Doesn't he fly in all kinds of weather anyway?
That's another thing. He flies in all kinds of weather anyway. Yes, that's the other point. I mean, he flies in many conditions where many others would have already packed up their glider a long time ago. And what I can contribute as a meteorologically skilled supporter is that I...
...have already spent a bit more time looking into the weather during the day. And then I can show him in the evening, okay, look, tomorrow this, that, and that are relevant. And then he can look at the details himself again. And something I also definitely have under control is knowing where I can keep an eye out for potential thunderstorms while he's out there. So, there was once a stormy day and we agreed in the morning, yeah, keep an eye on the storms and give me an update every now and then. But overall, Triggl is someone who doesn't really want to be told what to do, but wants to make his own decisions. And that means, even as a meteorologist, your role is actually more about just preparing the information quickly and not really saying, you have to do this and you have to do that.
He doesn't need that at all.
You were also with Xpeer as a supporter at one point. Did you do more weather there? That wasn't for Triggl, but what's his name again?
Fabian Umbricht. Yeah, sort of. Fabian is a special case because he's a meteorologist himself and, unlike me, he's actually active in the forecasting service. I have to keep training myself to look at forecasts and then make flight decisions, but he does that every day as his job. That means he's actually more of a pro than I am when it comes to creating weather forecasts. And that's why, I did help out there, but he's the one who wants to look at it himself and then decide what he thinks will happen the next day. I'm more like a good discussion partner where I can say, "Yeah, and what do you think about that?" and "What's a decision point for you in this regard?" so that I can help out a bit.
But I didn't really have to look at it in such detail because he wants to do it himself anyway. But that's exactly where we noticed that it helps enormously when you're in the Alps, where you already know the weather much better or simply have much more training in looking at forecasts and then thinking about what it will actually be like in the air. Unlike the Pyrenees, where we just didn't have the experience with it—there were days with a lot of moisture in the air and we had to somehow estimate, yeah, is it worth hiking up now and maybe we can take off at the top, or will we just be in the fog anyway? In those cases, we couldn't do anything more than just look at the models and go out in the morning and hope for the best, unlike people who
...who have less experience with models, is that we at least know the models don't have it under control.
Yeah, then basically you can say, yeah, I'll take a rough look at the models and otherwise I'm just looking at the local sky to say, okay, I know the basic wind direction today is southwest, or maybe in the Pyrenees, in between it's some kind of north wind, which is annoying because I always get these slight Pyrenees foehn effects, so to speak, on the south side if you have to fly there, at the X-Pyr, but ultimately you just have to look at the sky on-site too.
Exactly, so we don't make the mistake of saying, "Well, my app says we won't have clouds until 17:00 and our launch site is nearby at 16:00, so it's definitely doable," because we're simply aware that the chance of it working might be 50-50 and that we have to think ahead about what plan B is if it doesn't work out the way we thought. So, just having a feel for the uncertainty of the models definitely helps there. But as I said, if we were at home in the Pyrenees now, we probably would have with fewer
weather model to watch, we probably would have been able to better assess how it would actually turn out by looking at fewer weather maps, simply because you can gain a lot of experience.
When you go flying, is there such a thing as a dream weather, or a favorite...
...or favorite weather conditions where you say, I definitely want to have those? Or does it also happen to you that you say, hey, I'm interested in the weather, I'm also happy if the weather doesn't have the super XC conditions, but something where you can say, I can learn something new or try something out, or I just have to learn how to deal with these perhaps not-so-optimal weather conditions accordingly.
I'd say one doesn't rule out the other. So, I clearly have...
favorite flying conditions, and primarily, if there's just little wind, I'm already happy. So, no accidents like getting swept over the Nordkette, and basically, I'm not much of a valley wind fan, or I'm still partly overwhelmed today by properly assessing valley winds because my home flying area is the Jura and it just doesn't exist to the same extent as in the Alps. So my favorite flying day is actually when there's just little wind, no matter where I am. But that doesn't rule out that I'm not gripped by learning new things. As I said, I find it a cool challenge to think about which is the best flying day next week and where is the best flying area and what could you get the most out of there.
But in the end, I'm not competitive enough to get any real satisfaction from having done the absolute best; I'm just someone who is very curious and enjoys discovering things. And when I have some free time and go flying tomorrow, the most important thing for me is often having two or three people come along whom I really like—so that I just had a good social outing and
then it really gives me a lot just to think about how it might be and then go out and explore, and I actually never do that—sitting at home and thinking, yeah, today definitely wasn't that perfect, yeah, I'm just not going today—instead, I pack up, I really want to know what it's like there, and sometimes I've even experienced that in a bit of a negative way, where I'm flying somewhere and then I think, yeah, there's definitely a lee back there, that's probably not very pleasant, but I still want to know what it feels like to fly into it, and I don't do that if I think it's truly reckless, but there are always so many unknowns and I'd love to uncover them all and find out what it would really be like if you were there right now and did it, and that's also one of the things I really like about paragliding.
If I were to...
Gymnastics, I used to do gymnastics tours, so I actually know what to expect every time. I go into a gym, it has rings, it has a high bar, I might go into a different gym sometimes and it also has rings and a high bar, and that's one of the most beautiful aspects of paragliding for me—that you never know exactly what to expect and that you can be surprised over and over again.
Do you also sometimes have flight days that are actually pretty bad on paper, thermally speaking or something, where you might say, okay, the wind is weak, you can fly safely, but I know the cloud base is very low, it might even be sitting right on top, I can start under it, but the air mass below is still so unstable that you can actually hold it very well. This isn't a day for big cross-country flying, but sometimes a day with such super magical conditions where you can fly extremely close to the clouds, where it lifts everywhere, where you can fly into almost every tiny valley—there's no big valley wind or anything, you can fly in and out everywhere, it lifts super well everywhere, but in the end, it's not the score at the end that makes you say, wow, that was a great flight, but simply just a scenic flight.
Do you experience that quite often too?
Yes, all the time. So, the way you just described it, that already sounds almost perfect for a flight day, because obviously, there are days when it's very clear that it's good for long-distance flights, but personally, I often find it difficult not to be dissatisfied when I'm standing on the ground after four hours on a good flight day. I mean, I have to drive home and I see that everyone else is flying even better and longer, and I'd love to still be in the air. I've often experienced being just frustrated, and now, the days you just described—those are actually the best days to counteract that, because you have zero expectations and yet you still just...
pure joy of flying itself, and I need that for my soul every now and then. So, if I've had too many botched, good cross-country flight days, I absolutely need to experience something while flying that is completely detached from a
pure joy of flying itself, and I need that for my soul every now and then—so if I've had too many botched, good cross-country flying days, I absolutely need to experience something while flying that is completely detached from an
posterior
unstable
a low-base day, or it could be flying into the evening sun somewhere on a ridge, or it could be a hike-and-fly on a day when there's too much wind in too many places but you can still manage a glide. I absolutely need that to keep my joy in flying over the long term.
But as pilots, we're always looking for great weather, or whatever we consider to be good flying weather, so we don't take the opportunity for those kinds of experiences very often because we actually always want the best weather. And in those not-so-great conditions, where you can still say it's safe to fly from a winter perspective, and it's just going to be a glide, I also just test out these weather conditions and go flying anyway, even if it looks suboptimal or even relatively bad on paper, just to experience that magic of bad weather in a positive way every now and then.
I think that's a good point you're making here, because when you're out there in the cross-country flight scene, you always see the great flights from the top pilots and you want to achieve that too, and you end up emulating them, and then sometimes you forget that there are also just beautiful soaring experiences. And then, on the side, I have my father, who's been flying for a few years now and is primarily out at the local mountain, or my younger sister, who's still in the process of getting her license, and sometimes they come home and say, "Hey, today was such a beautiful day, we just spent an hour circling thermals in the same tube," and then they're so happy, and that holds up a mirror to me, making me think, "Yeah, that's actually what makes me happy too," and
That's why I try to make sure that I somehow...
I can clear out cross-country flight days in my agenda, but I also have weekends where I just write down, yeah, paragliding, and then I just see where I can do something, and if it's just an hour of soaring, then it's an hour of soaring and I'm still happy. I'm not ambitious enough to only focus on the good cross-country flight days, and I also claim I'm not good enough, so if you've been doing cross-country for 20 years and only have a few flight days left because of family and work commitments, then I understand that those people just pick the five best days of the year.
to pick out, but I still feel like an incredibly small flight chick myself, and for me, it's also important that I go out in many conditions to test things out, get to know my glider, and get to know the surroundings, and I just really enjoy that.
Has it made you more humble over time?
Definitely, I mean the experience there at the Nordkette really intimidated me and it certainly made me more humble, but at the same time, I also learned that you just have to go out there and see. Kriegel is such a good example of that, especially in the Hike-Fly races anyway; you can't change the fact that you have to get from A to B, so you just have to make the best of it. And what I've learned over the years, especially in Hike-Fly, is that it usually flies somehow better than you initially expected, and that's why I also have days where I don't...
...spend hours obsessing over the weather; instead, I know it's going to be somewhat windy tomorrow, some cloud cover, it's definitely not top-notch, but I'm just going there anyway, we're doing a hike-fly and we'll make the best of it—and I'll tell you now, in 80% of cases, you fly better than you actually initially thought.
So we basically underestimate the real flying conditions based on the forecast?
Maybe it's not the real flight conditions that are underestimated, but maybe the possibilities that present themselves, even if the flight conditions are as the models say—just because it's cloudy doesn't mean there aren't thermals per se. There are always days where you can still get some good lift even though not that much sunlight is reaching the ground. And those are exactly the things you don't learn in a meteorology degree, you don't learn them at a flight school, and you don't learn them by listening to someone give a weather lecture; you only learn that by just going out and trying. And yes, I always tell myself, I don't think it's that bad if I try something somewhere and then say, "Well, I didn't fly today because there was really too much wind," but on the plus side, I had a good friend with me and we had a great day together, and the fact that we didn't fly is just one point, but instead we talked about the weather for an hour, learned something, and had a good time together, and I think that's also part of my positive paragliding experience, even if the flight itself didn't happen at all.
So, the essence is: people, don't let the weather, in whatever form it takes, stop you from having beautiful, positive experiences with flying. Definitely. Alright, Stefanie, thanks for your great, in-depth stories about the weather and paragliding weather, how models work, and how little you should sometimes trust them—or at least allow for a much larger margin of error than people normally think, assuming that exact point forecasts will always come out. Thanks for that, and I wish you much success in continuing to gain experience so that what you've understood theoretically and keep applying can be reinterpreted even more instantly with a practical look at the sky, to say, hey, I'm right and not the model.
Thank you so much.
That was Podz-Glidz episode 106, The Meteorologist with Stefanie Westerholtz. The Meteorologist with Stefanie Westerholtz. There are some further links on the Gleitschirm-Blog Lu-Glidz. My name is Lucian Haas. I am the producer of Podz-Glidz and the publisher of Lu-Glidz. If you enjoy these two offerings and use them regularly, then please consider becoming a supporter of my work. If you give, for example, one euro per podcast episode or two euros per reading month on Lu-Glidz, it certainly won't hurt. But you contribute an important part to me being able to continue and develop this project in the future. And that independently and ad-free. All the necessary information on how your freely chosen contribution reaches me can be found at www.Lu-Glidz.blogspot.com. There, you'll find the menu item Support.
Lu-Glidz is spelled Lu-Glidz. By the way, I'm always looking for other interesting conversation partners with noteworthy stories from the cosmos of paragliding. If you have a suggestion, tell me about it and, if possible, send me the contact details right away. Best sent via email to lugleitzkontakt at gmail.com. Until next time. Ciao.