A few observations -
(a) this style of chart derives from an old DCA (antecedent of CASA) tech memo (in a previous life, inter alia, I produced quite a few of this genre for flight manuals).
(b) That's interesting, though, about not extrapolating any lines The equations in this chart are fairly simple and a small extent of extrapolation won't hurt in the real world. However, the rule is no extrapolation because the general user has no idea what the equations of a graph are and, hence, you have no idea what the shape of the line will be outside the printed values .. in which case, extrapolation becomes fraught with the possibility of significant error. The problem is not a concern with linear or, even, quadratic but higher order polynomials and other forms of equations can have all manner of twists and turns ..
(c) I don't know why they provide a line for 1100kg when the max weight permitted is 1090kg Solely to make life easier for interpolation.
(d) until I hit somewhere between 'long wet' and 'short wet' grass, which would clearly be just dangerous The surface correction grid simply provides several lines for different mu values assumed for the surface in question. Interpolation is rather meaningless and ought not to be done.
(e) They look like a credible mathematical device but so much is simply someone's guess No guesswork involved. The carpets are based on comparatively simple equations from the DCA TM. The usual way we used them either was to run some simple flight tests to get several TOD/LD values for known ambients, or read off a couple from the US OEM unfactored data, and then expand those points using the carpet equations to generate the final chart. Providing that the expansion is between sensible flight test data, the chart will be usefully accurate. Apart from the design standard required 0.5/1.5 wind grid factors, the one area where a bit of fudge work comes into things is with the factors presumed for surface mu values.
(f) At best the performance charts are a rough guide, that's why they have large margins of error built in For the concerns listed, fair comment. However, one needs to keep in mind that the charts are based on presumed coefficients etc. .. if actual conditions don't resemble that which is presumed, accuracy gets a bit rubbery. No different for AFM performance data in Mr Boeing's or Mr Airbus' aircraft operations. Apart from the wind grid, generally no margins included other than the CAO distance factors .. which, for lighties, are a tad slim for comfort.