One scenario I've not been mentioning lately that I can envision is quite literally a tail strike scenario as the pilots lose control, lose a lot of altitude, regain at least partial control in a dive, and are still trying to pull up when they meet the reality that they didn't have enough altitude to even lay the plane down on the water properly.
JD-EE,
from the damage pattern visible in the images shown by BEA this might even be one of the most plausible explanations because it would cover the vertical as well as the longitudinal decelleration.
Problem is maybe more with the overall scenario:
How do you get the plane so quickly from 35kFt high speed to maybe 1kFt at a a relatively slow speed without breaking up the aircraft during the dive/recovery briefly regain control, drop it again and crash it belly first, nose up during the attempted second recovery? (OK, one way this could happen would be if you lost the engines somewhere in the process).
Henra, in support of bearfoil's concept let's consider the weight distribution a little bit more. The whole bottom of the VS is heavier than the top
If you look at the following image and extrapolate the two belts and pull a vertical line from where the two belts join, you get a quite good idea about the Cg of the actual fin:
http://www.airdisaster.com/photos/aa587/4.jpg
That would be between 40 and 45% of the span, i.e. quite comparable to the areal center. So a homogenous model should fit. Probably the best model would be an unweighted foam model.