So Was it an RCC Failure? (more evidence)
RCC = reinforced carbon-carbon (the black D-shaped sections bolted on to a flat plate and which are the highest heat resistance surfaces constituting the wing leading edge). Loss of any of this RCC composite leading edge would be catastrophic.

This diagram has some words with it which seem to make a lot of sense out of what we have been hearing about the ablation pattern and the sequence in which the sensors failed. In addition the conclusions reached here (link) would seem to be supported by both this locational data (below) and the USAF photo (as represented by the "projection" on the left wing inboard section seen on that photo).

--- Quote from words on the diagram ----
... Wiring diagram shown is for the left wing. Scanners attached to the orange lines failed first, and six sensors on the main gear failed last.
----- End of quote -----------------------------

The orange wires take a 90º bend directly behind the leading edge of the wing where the sweep-back angle changes, shown in the NASA photo below of Columbia on the pad before launch of STS-107.

The Explanation of why it may have been an RCC failure

AW&ST Article Addressing the Leading Edge Failure Probability


 

 

The apparent malformation "defect" two feet out on the leading edge is "unexplained" at this time (but begs an explanation). It is apparent in both the plan and frontal view.
Regrettably, it now seems that the lo-res picture that has triggered so much debate in recent days really is all they took at Kirtland on that fateful morning - with a 3½" telescope and an "old" computer . . "in their free-time." Evidently, the precision elements of Starfire were not employed - nor indeed, requested. No comment.

Just to consolidate some of the plethora of references on this thread (please note download size before clicking!) here are links to the pdf (Acrobat) docs referred to - the big ones are for seriously interested folks only! (Shift+Left Click to view in a separate window) :
The NASA emails about possible gear and tyre damage are here (192kb).
The original Paté-Cornell/Fishbeck paper - 1990 (4.8mb).
The later Paté-Cornell/Fishbeck paper - 1994 (2.9mb).







Ms Paté-Cornell has been reported as saying that she agrees with Ron Dittemore's remark about a "missing link" at an earlier briefing following the accident. It is also now being suggested that one sensor indicated "improper" left gear deployment, just before breakup, although two other sensors did not. Gear extension at that speed, if so, would undoubtedly be catastrophic and does not, in my personal opinion only, explain the progressive drag increase observed prior to the main event.

Below is a diagram of the left wing structure, which I hope helps to visualize the layout and what it all looks like under the Thermal Protection System (TPS). Note the short distance between the LE and the wheel well, around the point where the wing-sweep changes. The exact position of the in-fuselage heat-rise is unknown - but there are obviously a number of direct metallic structure routes evident from the suggested failure point.

Enlargement .

This is taken from the earlier Ko-Fields Report Oct 1987 (1.3mb)

I remain inclined to believe that there may have been (also) a RCC LE panel contributing to the scenario. This would indeed be a formidable missing link. Picky's link of Columbia's LE (Page 10) clearly shows the grey RCC panels for those unfamiliar with the subject. Likewise, the video appears to show two separate pieces of debris fall from the ET at the 81-sec point during launch ---- two bits = two possible damage sites.