My take on this is that there are three issues (at least):

1. Accuracy of GPS altitude. Good discussion in the first paper. Thanks, Tom!

2. Remaining clear of controlled/restricted airspace. U.S. FARs are clear in spec'ing pressure altitude.

3. Managing final glides. This was my original focus, in particular in the Western U.S. where both altitudes and temperatures can be high (leading to greater differences between observed pressure altitudes and GPS altitudes) and final glides are often much longer. The latter can increase the odds that, owing to both distance and time, a pilot may transition from an area with lift to an area where there is less or no lift, making it difficult to salvage a final glide that is falling short at the end.

Pressure altitude is conservative in the sense that it often under reads the geometric altitude and--at least out West--will therefore provide a cushion against unforeseen sink. But I already have an explicit arrival height safety margin. Layering that with an uncertain additional cushion isn't where I wanted to go.

That said, the GPS vs. pressure altitude cushion has the virtue that it tends to increase with altitude. Final glide computers I've used allow entering an arrival (safety) altitude of X feet, without regard to MC setting or altitude or length of glide. There are ways to make it proportional (e.g., % bugs or % risk) but using pressure vs. GPS altitude can do the same thing.. Of course, the differences will vary from day to day so that introduces more uncertainty into an area that already has plenty of it. I'd rather nail the altitude (subject to the uncertainties of GPS determination) and factor in the safety factor(s) explicitly myself.

Thanks for all the input. I learned some interesting things.

Chip Bearden
ASW 24 "JB"