When hazards were identified and categorised at a meeting in August 2004, he said, the possibility of an explosion in the dry bay was graded in as "improbable".
Tuc,
Your conclusions above may prove correct, but the way I read the Air Cdre testimony is slightly different. It seems to me from what he says that they did infact incorrectly grade the hazard during the analysis (and indeed this must have happened also in the first place 20+ years ago when the SCP mod first went in under the prevailing safety analysis techniques). In this scenario the IPT would not have asked for more money etc. from above to fix it as you allude, but would have left the hazard there as they didn't pick up on it. Indeed, the IPTL would surely admit the latter if it were true as he could then pass the buck upwards as you say for lack of funds.
He did not however, and hence this was clearly therefore a key mistake. Ultimately he is correct as the IPTL to take responsibility for this mistake, though the error will have been 'made' or more properly not picked up by the various contributors/checkers of the Safety Case (a combination of IPT project officer(s) in his team, his contractor(s) (DA & safety case compiler / supplier), and his independant safety advisors who he contracted to review the Safety Case (probably Boscombe, maybe others)).
What is interesting here, and will surely form one of the key issues in the "procedural review" is that infact all/(most of?) the relevant extant safety processes and safety case procedures appear to have been followed*, and still a design flaw was left in the aircraft. Hence it is not so much a case here of "MoD not following its own procedures" that led to the loss (though I do not question whether this is potentially another valid issue that you Tuc raise regularly), but seemingly more a case of "MoD/suppliers followed procedures but despite this a design flaw still prevailed in the certified product".
For me the role of Boscombe (and other safety independants?) in this process failure will be interesting - after all are they not there to pick up on these kind of issues during Design/Safety Case reviews? Isn't this what they are contracted/paid for?
* this remains to be fully explored / understood.