The total North Sea crude + condensate plot shows a perfect Hubbert Linearization trend, with a peak at 52% of Qt.   I think that you interject too much noise into the analysis when you focus on sub-basins.
I would like to see those graphs. Can you give a link or post them?

I have to admit that I am not as convinced by Hubbert Linearization as I was before. I am planning to post a comment soon about my criticisms of the Hubbert Linearization method. Basically the Hubbert Linearization method makes almost everything look like a line. Hubbert Linearization method plots cumulative production versus (real production)/(cumulative prod) and then a Hubbert curve gets transformed into a line. But if you plot cumulative versus real production the Hubbert curve gets transformed into a parabola that passes through the origin.
See Jean Laherrère (pages 11-14)
for some graphs of this type. I think that this method is much more reliable and robust to determine the peak.

I'll have to dig out the North Sea plot, but you can see some recent HL stuff at:  http://graphoilogy.blogspot.com/

We tested the reliability of the HL model regarding the Lower 48 in the following article:  http://www.energybulletin.net/13575.html

I do have a suggestion for an experiement of sorts.  It is remarkably easy to do an HL plot for the North Sea using the EIA data.  It would be interesting to see what kind of Qt that several people came up with.  Data (crude + condensate):  http://www.eia.doe.gov/emeu/ipsr/t41b.xls  

I sent the North Sea graph to Chris.  If you give me an e-mail link, I can send it to you also.

Jeffrey J. Brown
westexas@aol.com