Stuart, you really spend a good bit of time with excellent graphs and charts.  I believe you come up with some great results.  Linearization works wonderfully well when depletion rates behave according to Hubbert's models.  That is, normal drilling and pumping.  When you factor in huge gas-water injection variables along with horizontal and bottle-brush drilling (especially from super large fields), the results using linearization becomes fuzzy.

I would like to see you try linearization on the North Sea Fields....and see how they behave.  Collapse rates that will occur in Mexico's largest field, as well as the Saudis, the Kuwaitis, and the Russians will be quite interesting to see if they behave nicely to linearization.

We have looked at the North Sea, and the linearization method does predict the high decline rates. You can see the UK graph in this post here, where linearization calls for K=13% (once it settles down after the dual peak). Indeed we have seen very high decline rates of 6-15% in recent years in the UK. Furthermore, it calls for very high decline rates in Norway also:

width="90%">