When you talk to the old-timers about ridge soaring, they often say you just gotta believe in the ridge lift. I’ve always been a bit bothered by this statement. We are strapping ourselves into a 600 pound mass of fiberglass, metal or wood that has a considerable amount of energy and we’re insisting on taking this aircraft into some places where there is nowhere to land, counting on updrafts ahead to keep us up. In the places where you can land, many of the fields are tricky and have a much higher probability of damaging a glider and/or hurting yourself should you have to use them. Perhaps it’s my agnostic upbringing, but I am not one who readily counts on faith when piloting a sailplane.
So, if we discount “belief” in ridge lift, how can we manage this risk? First off, we have to define the risk we are trying to manage or minimize, which is the likelihood of falling off the ridge. How does this happen? This occurs when the conditions or topography do not generate enough lift to support the glider. Furthermore, we are especially concerned with the prospect of falling off a ridge in a place where there is no safe landing option.
As such, we have to make a prediction about the ridge lift we are likely to encounter. We can never be certain what the air is going to do, so how can we go managing this uncertainty? This goes into the realm of modelling and the estimation of probabilities. In doing so, we seek to understand and minimize the uncertainty associated with the ridge lift. Uncertainty is always there; it can never go away. Can you ever be sure that anything will work out in a particular way in an aircraft without an engine (except that you’ll eventually come down to Earth)? However, we can establish bounds to our uncertainty through experience and through the information we collect. And if the bounds have a sufficient margin that the likelihood of falling off a ridge is extremely unlikely, maybe then we can justify taking the gamble of flying the ridge and taking advantage of this wonderful energy source.
What are these bounds? We build a distribution of the expected lift conditions ahead. The typical way we think about this distribution is the spread between how high and/or fast we expect to go. Blue Mountain is approximately 1500ft-1600ft MSL and we can define how well it is working based on our energy state. If we are driving down on the trees, we track our airspeed. For instance, if we are flying at 100 mph, we ask how what is the likely change to how fast we will be able to go in the subsequent section? All things being equal, we can usually expect that this trend will continue. Maybe we will hit a little bit of sink and we will slow down, or the conditions will get even more solid ahead and we can drive a bit faster. But if we settled in to 100 mph, then the range of speed will fall in some sort of range, suppose +-10 mph.
However, what if we are just barely hanging on the ridge, going 55 mph? While we might still be able to keep going, any small changes, even temporary ones can have a catastrophic effect. We have no margin on the bottom end. Lose another several mph and you will be sinking below ridge top. And once you sink below ridge top, all bets are off and you will likely get bounced off the ridge.
As such, I believe it’s fairly intuitive to track the performance of the lift based on the state and the variability in our energy. Energy of course can also be measured in altitude. A stronger ridge will typically have a higher band and pilots who prefer to stay in the higher band will pay close attention to their altitude changes as they fly. If a pilot is able to maintain 500ft, +-100ft, the ridge is working quite well. However, if they are flying slowly at 100ft above the trees, then they might be in trouble.
But a big concept in this discussion is the variability of the lift. If the lift is consistent, one can reasonably suspect that it will remain consistent for some time into the future. On the SE ridge, we can often have relatively soft conditions that persist for a long time. Sometimes it is possible to fly a soft SE ridge where you can only maintain 60 mph, but there is almost no variability to the speed whatsoever. And it is sometimes possible to even get comfortable with that kind of ridge and go cross country!
As such, it follows that the thing we are concerned about from a safety risk perspective is the bottom bound of uncertainty. We track the average trend, let’s say in speed which lets us have a baseline to work with. And we track the changes from this speed. The better and more consistent the lift is, the less likely it is that it will go away. The said, even going at 100 mph, there is a slim possibility that the next similar section will not have enough lift to sustain the glider. We use the variability in the trend to estimate the likelihood of this happening. And if the ridge has been consistent, then we can rightfully judge that the likelihood of such an event is extremely low and we may even consider flying into an unlandable area.
What are the variables that let us predict the strength of ridge lift ahead?
1) The size and shape of a mountain. If it is steep and high, the ridge is more likely to work. If we are heading to a worse shaped ridge, we usually expect that the lift will get weaker. We must factor this into our model.
2) Wind speed and direction. Without wind, the ridge doesn’t work! The stronger and more perpendicular the wind is, the better. We must keep an eye on how these variables change as we move further along the airmass and as the ridge changes direction.
3) Convection. Thermals along the way add energy to the glider. A ridge with no thermals will usually have less energy, which if weak can be a problem.
4) Anabatic lift. If the sun is facing the ridge, the heating of the surface reinforces the ridge lift. Conversely, if it is in shade, it will need more wind in order to work.
5) Stability. If the air is more stable, the air has a less of a tendency to go up, but rather to speed up over the ridge or go through the gaps.
6) Wave suppression. This is usually overblown, but if the ridge is on the down-part of the rotor, this can suppress the lift. If there are obvious features of wave, such as foehn gaps or lenticulars, this is worth paying attention to.
7) Thermal suppression. When there is powerful streeting, there is also strong sink at the edge of the streets. If the ridge lift is weak, this can be enough to wash out the ridge.
Probably the most demanding aspect of managing risk in ridge flying is flying in unlandable regions. This requires gambling on the ridge working ahead. The difficulty is that the farther ahead we have to estimate how well the ridge works, the more uncertain we are about the outcome. This demands a greater confidence in the lift we are currently experiencing.
There are several ways to minimize this uncertainty. The first is to minimize the distance needed to be flown in unlandable areas. This sounds obvious, but scouting out landing options that halve the distance have a massive impact on the uncertainty one has flying in these regions. Next, one needs to increase the accuracy of their model of the ridge lift. This is best done by testing the ridge before committing to the unlandable area, which is best done by driving down on the trees and seeing how fast you can go. At this point, you then ask if the strength and variability of the lift is sufficient to gamble on X minutes that the ridge will work?
The interesting thing is that minimizing uncertainty can sometimes be inconsistent with maximizing altitude. Some pilots flying unlandable regions will try to have as much altitude as possible before entering such a place, with the hope of getting some more energy along the way to minimize their exposure. If the pilot is high enough such that they can get across, then undoubtedly this is the best option. Or if they are high enough to always have a landing option at their disposal. However, if a pilot is willing to commit to the unlandable area, this actually is counter to minimizing uncertainty. Unless the pilot had tested the ridge before climbing up, if the pilot encounters sink or fails to find lift along the way, they can be in really serious trouble. If he doesn’t know whether the ridge is working, the bounds on his distribution of expected lift strength are very wide. If the distribution is so wide that there is a possibility of the ridge not working, then the pilot is taking excessive risk!
So in summary, pay attention to the trend and the variability in the trend. And make predictions about how the trend is likely to change in the future. With more experience, pilots can make these predictions more accurately and need less margin. However, do not simply depend on belief. Statistical reasoning is a much better way going about gambling.
Soaring Economist 