Yes, sorry askmelater, my post was a little too summarised.
In point (a) I am suggesting that some MDA's are not set at a height to guarantee terrain clearance on a 2.5% climb gradient, but instead are set because it is deemed unsafe to go lower on a non-precision approach.
For example, a certifying authority might decide that it does not want an aircraft descending below 300' AGL until the aircraft is visual to the runway - this is not the pilot's ADH, but is the specified MDA for the approach. For example, let's say that the certifying authority wants 296 feet of terrain clearance at circling MDA, so that is the MDA set.
If the missed approach takes the aircraft over water or flat terrain, then terrain clearance on the missed approach is not an issue, and there is no obstacle clearance related reason for the asymmetric aircraft to use a higher MDA to compensate for reduced climb gradient. The pilot of a light twin that has suffered an engine failure could quite happily go down to MDA, and climb back at a 1% climb gradient if necessary - he or she isn't going to hit anything, so why set a higher MDA than necessary? (I'm ignoring airspace and traffic flow problems in this scenario.)
In my post I am attempting to show that adjusting MDA to account for reduced climb performance is difficult, sometimes expensive, and unable to be done accurately by the pilot in the cockpit - there are too many variables.
Perhaps I should sell my CAA's half-baked effort at an AIP & buy a Jeppesen suite of manuals with some decent PANS-OPS reference material in it!
O8