Where Are the Ghost Gliders? or: How Dangerous is Soaring Really?

Preface: I wrote the article below in January of 2018 and I have since done more detailed research about the same question.  My most recent effort uses slightly different (and probably more reliable data) but the overall conclusion remains valid.   If you’re interested in seeing the results of my most recent research that also considers the relative risks of other air sports such as General Aviation, Hang Gliding, Paragliding, and Skydiving, you can find it here.

The tragic death of Tomas Reich during the last day of the most recent Sailplane Grand Prix final in Santiago de Chile and the ensuing debate about the safety in soaring competitions brought – once again – a key question to the forefront of my mind: How Dangerous is Soaring Really?

When I started soaring in 1983 at the age of 16, I often heard people say that “the most dangerous aspect of gliding is the drive to the airport”. Intuitively this never felt right to me and several people have since pointed out that it is indeed far from the truth. (See, e.g., the speech Safety Comes First, delivered by Bruno Gantenbrink).

But just how dangerous is it? To get a better sense we need a reference point. I believe the best way to think about the dangers of soaring is to compare it to the dangers of other relatively dangerous activities we might indulge in: e.g. we could go on a road trip, ride a bike, or ride a motorcycle.  And I think the best way to make such a comparison is on the basis of participation hours (rather than on the basis of miles traveled for example). E.g., when we have an afternoon to spend we may want to know: is it more dangerous to spend that time riding our bike or to go fly our glider? We have all seen the white-painted “ghost bikes” on the side of the road marking the spots where a cyclist was killed but we haven’t seen any “ghost gliders”.  However, we would be kidding ourselves if we thought that gliding was somehow less dangerous. (Spoiler Alert: the comparison is not even close.)

Unfortunately, good, global statistics about the dangers of soaring are hard to come by. In most countries, a comprehensive and reliable database of gliding accidents does not exist. Nor is there a reliable global record of the number of flights or hours flown that would provide a good reference point.

However, while the available data is not globally comprehensive, there is enough out there to draw these comparisons – at least directionally.

My analysis of gliding accidents is based on data from Germany: the German government keeps meticulous track of all flights and even separates out glider flights and flights in motor gliders. It also maintains a database of all flight accidents and reports on an annual basis the the number of fatalities, and the number of persons injured.  Now, one might think that using German data is rather limiting.  But that is not quite true because gliding is much more popular in Germany than elsewhere.  In fact, according to a report presented to the International Gliding Commission in 2010, Germany accounts for approx. one third of all glider flights worldwide.  If there is a limitation to using German data, it might be that it actually underestimates the dangers of soaring elsewhere simply because Germany has such a particularly well developed soaring and safety culture. But, since I can’t prove that, let’s assume the German stats do a fair job of representing the dangers of soaring in general.

So here is what I found.  The result is – unfortunately – rather sobering.

On average, soaring pilots have an accident every 10,000 flights (this is based on all flights in Germany from 2002 through 2016 – the exact number is 10,070). Fortunately some of these accidents only damage the glider or some other property. But once every 60,000 flights someone (usually the pilot and/or passenger) is seriously injured, and once in 83,000 flights the pilot and/or passenger dies.

If you consider that the average glider flight takes about 38 minutes (arguably my least generalizable assumptions since it is simply based on the flightlog of all club flights of members at the Soaring Club in Boulder between 2002 and 2017) this means that soaring pilots can expect to get seriously injured every 40,000 hours and die every 50,000 hours.

Wow! Fortunately we do also other things in life because these stats mean that we would die every 6 years if we did nothing else but fly gliders!

So how does this compare to other activities? Well, not favorably to say the least.  On a “per hour” basis, gliding is about 35x more dangerous than driving; 70x more dangerous than riding a bike, and still 3x more dangerous than riding a motorcycle.

Risk of dying per hours of engaging in a particular activity. Note that the comparison is directional because the data for the various activities are from governments in different parts of the world. (Gliding is based on German data, bicycling and motorcycling are based on UK data, and driving is based on data from the AAA in the United States).

Another way to look at this is to say that 1 hour of gliding is about as dangerous as going on a 35 hour road trip in a car, e.g. from Denver to San Francisco and back again. Or as dangerous as riding a bicycle from Denver all the way to Minneapolis (70 hours). Or as dangerous as riding a motorcycle from Boulder to Salida (3 hours).

Is this an acceptable risk to take? I think that is a question we all have to answer for ourselves. But the important thing is that we should all ask that question and think hard about what we can do to minimize the risk in our own flying decisions.  And no one should kid themselves into believing that those stats don’t apply to them because they are simply a better pilot.  (Instead, they should remind themselves that it’s often the best pilots, like Tomas Reich, who make up the sad statistic.)

With sincere condolences to the family and friends of Tomas Reich.

Rotor Fun – First Flight of 2018

Unseasonal warmth greeted me this morning as I stepped out onto our porch to film the clouds in the rising sun. Wearing only shorts and a t-shirt I felt as comfortable as I would on a mild summer’s day. A gentle breeze whisked around the corner as I mounted my camera onto the tripod, pointing it east towards the horizon.

The clouds told a story of winds aloft, but where I stood, in the lower foothills, 400 vertical feet above the valley, and 5,800 feet above sea level, the movement of the air was gentle and kind.

The night before, the outlook had already looked promising for my first soaring flight in the New Year: TopMeteo projected westerly winds of 30 kts at 12,000 feet, increasing to 40 kts at 18,000 feet.  Meteoblue projected a stable layer between 11,000 and 15,000 feet – right around the tops of the mountains. Dr. Jack’s cross-section chart for Boulder indicated multiple wave bars with modest climb rates even though it projected the wind to have a pronounced southerly component. Based on past experience, I decided to – once again – dismiss the Soaring Forecast from the National Weather Service, which predicted good thermals (very unlikely in the flat January sun despite the unusually high temperatures) and poor wave conditions.

But the best indicator for good soaring conditions was right in front of me: beautifully turning rotor clouds – as always an unmistakable indicator of mountain wave.

On my way to the airport I reflected upon my most recent wave flight, which was characterized by extreme turbulence below 13,000 feet. I braced myself for the possibility of earning another set of bruised shins even though I was hopeful that the comparatively modest wind speed might be a mitigating factor.

One decision was made for me already: I had learned at my club’s monthly meeting that a recent attempt to open the Arapahoe Wave Soaring Area (which allows flights above 18,000 feet within a pre-defined area) had failed because Air Traffic Control was completely unaware of its existence. This would be clarified in an upcoming meeting with ATC but until then it would be better not to put in further requests. This meant that I would have to stay below 18,000 feet and not have a chance to earn Diamond Altitude (which requires a 5,000 meter (16,400 feet) altitude gain in soaring flight after release from tow).  It also meant that I would not need to bring the more sophisticated oxygen equipment required for flights further aloft; and there was one more benefit: the risk of freezing my toes off would be much reduced 😉

At 11:00am local time I was the first pilot of our club ready to launch.  There were two beautiful lines of rotor clouds in the sky, indicating the positions of the primary and the secondary wave. There were also some isolated rotor clouds from the tertiary just to the north of the airfield.  I asked the tow pilot to take me to the upwind side of the secondary, which seemed to promise the opportunity for a longer flight along the wave bar.

Lines of rotor clouds indicate the position of wave lift. The lift is always on the upwind side of the rotor clouds, i.e. on the side facing the mountains.

After two initial turns near the airfield to gain altitude I followed the towplane toward the northwest. Soon after we had passed underneath a small rotor cloud from the tertiary we encountered the first pockets of strong lift.

When the third pocket of lift had lasted more than a few seconds I felt comfortable to release from the tow. I would try to climb in the tertiary and then attempt to push forward into the secondary without the help of a tow plane.

Release from tow in rotor lift from the tertiary wave at 7,900 feet (2,600 AGL). (You can see the tow plane turning left underneath.)

After releasing my first focus was to stay in the area of lift to reach a more comfortable altitude. (2,600 ft AGL may sound unproblematic but where there is strong lift there is also strong sink, and 2,600 feet may only equate to two minutes of remaining flying time if I were to encounter a major downdraft.)

I looked at my GPS (mounted to my right and not visible in the pictures) and quickly worked out the the crab angle necessary not to drift further away from the mountains. The wind speed was considerably weaker than during my previous wave flight. There was some turbulence but nowhere near as pronounced as during my prior wave flights in Colorado.

The area of lift in the tertiary was not very large. However, it was surprisingly calm even though I never reached a truly laminar air flow. The lift was moderately strong, varying from 5 to 10kts. Within 20 minutes after takeoff I climbed through 17,000 feet.

At 16,600 feet I was still in rotor lift from the tertiary wave. The airflow became gradually smoother the higher I climbed but it never turned completely laminar.

After some time in the tertiary I decided to try to move forward into the secondary. I looked for a gap in the clouds along the secondary rotor line and pushed forward into the wind. Vividly remembering my prior wave flight where I lost more than 6,000 feet during a wave bar transition I prepared for the potential of a similar loss in altitude.

This time however, the transition turned out to be easy and smooth.  There was some modest sink along the way but the entire push into the wind did not take more than three minutes during which I only lost 1,500 feet.

Transition from the tertiary into the secondary: my flight path took me around the end of the secondary wave bar. This had multiple advantages: less turbulence, less sink, and a flight path not obstructed by clouds.

Having arrived in the secondary, the lift was clearly stronger, and within two minutes I was back at just under 18,000 feet. Although the line of clouds was interspersed with blue skies it was fairly easy to locate the area of lift. I increased the airspeed of the Schweitzer 1-34 to 80mph and flew south where I could see the next clouds to the west of the Flatirons. I looked at the shape of the Continental Divide to my right and sought to maintain a more or less constant distance to the mountains, accounting for the direction of the wind, blowing from WSW.

Following the line of lift through a “blue” stretch. I had just left a line of rotor clouds from the secondary wave behind and followed a line towards the next rotor clouds in the south. To gauge the best flight path, I tried to maintain the same distance to the mountains on the right while staying to the right of an imaginary line that connected the visible rotor clouds. I also watched the vario for changes in the vertical air movement and corrected the flight path by adjusting the crab angle as necessary.

West of the city of Golden I turned around and retraced my route to the north, again following the line of lift. Without a single turn I continued to fly straight for over 40 miles until I was just west of Carter Reservoir. The lift in this area (north of Lyons) was the strongest of my entire flight: I had to fly at 90 mph with the air brakes fully extended in order to neutralize the lift and keep the plane below 18,000 feet. Next to me was an imposing rotor/lenticular cloud, its western side almost vertical, extending many thousand feet above and below my flight level. Based on my location and the direction of the wind, I assumed that the airflow forming this massive cloud was likely triggered by the steep downslopes of Mount Meeker and Longs Peak.  (I noticed that this area lies outside the boundaries of the Arapahoe Wave Soaring Area so I could not have used this location to climb above 18,000 feet even if the wave window had been active.)

From there I flew back towards the south. I briefly contemplated a push forward into the primary but at this point my feet had become quite cold and I decided to call it a day and return to the airport.

The frozen lakes surrounding the Boulder airport reminded me that it was the middle of winter.  Obviously, they were of no help detecting the wind direction on the ground.  However, the windsock, once in sight, was easy to read, showing a stiff breeze straight from the west.

I entered the landing pattern at 1,500 feet AGL and turned onto final at the end of the runway, still almost 1,000 feet above ground.  I pushed into the wind, flying the final approach at an airspeed of 80mph.  Seconds later, I touched down gently at a very low ground speed, just fast enough to roll the remaining 100 feet right up to the parking position.

Here is a link to the flight track.

For those interested, I have also compiled a lot of information about wave flying that you can find here.

Lessons Learned

  • A high tow may not be necessary to reach wave lift. I released at 7,900 feet and had no problem at all to climb into the tertiary. Today, the first good climb on tow was at 7,300 feet. It would have probably been sufficient. To practice, I need to be willing to release early and risk having to take a second tow. This is especially important with respect to reaching Diamond altitude. With today’s release altitude I would have had to climb to 24,400 feet to accomplish a gain of 5,000m (16,404 feet).  If I could release even earlier, I would not have to fly all that high.
  • Rotor turbulence can be gentle. Today’s rotors were very different from those I encountered during my most recent wave flight. I attribute today’s conditions to the much lower wind speed at altitude (about 25 kts versus 50 kts). On Nov 16 the winds were so strong that I struggled to make progress along the wave bar because most of my air speed was needed to push into the wind, whereas today the necessary crab angles were fairly modest. The flight on Nov 16 offered a bigger challenge. Today’s offered more pleasure.
  • Laminar air flow may only start above 18,000 feet. During today’s flight I never encountered a fully laminar air flow. That tells me that the rotors extended well above 18,000 feet. There was only little moisture at altitude; however, I did see a few lenticular clouds high above the rotors  (my guess is above 30,000 feet). Today would have likely been a great day to reach Diamond altitude.
  • Good climb rates are possible even when wind speeds are moderate. It does not take a howler to produce good climb rates in wave conditions. Today’s climb rates were between 5-10kts in the tertiary and reached well above 10kts in the secondary. At one point the average netto climb rate was 14kts at an altitude of just under 18,000 feet (demonstrating the great potential of today’s lift.)
  • Wave transitions don’t have to cost a fortune (in altitude). Rotor clouds can be super helpful in identifying the best locations for a (forward) transition from one wave bar to another. Today I deliberately picked a spot “in the blue” to push forward into the wind. This did not only reduce the risk of getting sucked into a cloud, it also dramatically reduced the sink rates encountered and therefore the amount of altitude lost during the transition.
  • New rotor clouds develop within seconds.  While I experienced no close encounters with developing clouds I observed numerous times how new clouds can form within seconds. It is critical to always be aware of your location relative to the line where new clouds could possibly form. Especially when a strong crab angle is required it may be difficult to spot that you are about to be engulfed in (newly developing) clouds from behind.