Nice idea, well balanced, and workable.... but...
Neither of the Nedu profiles, hit any lasting limit in any model ! They don't get anywhere near close.
Oh, they do hit the ceilings depending on the conservatism settings. This is like asking the question: "If the test subjects in the NEDU study would've had a dive computer with them, or had planned their dives on a desktop computer, which settings should they have used (VPM conservatism level, or ZHL GF) so that the software had told them that A1 profile is OK but A2 profile is not?"
If I run the profiles in subsurface, I get:
Code:
-----------------------------------------------------------------
Setting A1 (good) A2 (bad) Class
-----------------------------------------------------------------
ZHL-16B GF 100/100 OK OK 3
90/ 90 OK fail 1
80/ 80 fail fail 2
70/ 70 fail fail 2
90/100 OK OK 3
80/100 OK OK 3
70/100 fail OK 4
80/ 90 fail fail 2
VPM-B +0 fail OK 4
+1 fail OK 4
+2 fail OK 4
+3 fail OK 4
+4 fail OK 4
-----------------------------------------------------------------
For ZHL+GF it seems that GF_high needs to be 90% or less to detect that A2 is bad.