I believe Katrina will come inland east of current projections of SE LA. I have noticed previous storm projections on storms that are in the beginning of the recurvature process tend to have a left bias. This may be partially due to persistance, and partially due to missing the exact moment the storm beginning their turn poleward. The storms are big ships and take a while to turn, but once they turn, they do so more than expected.
I believe model runs after 00z will begin to shift right. I also believe the storm will predominately spend it time to the right of the current forecast path. Ultimately landfalling between MS/AL border and Pensacola(again).
As for intensity, I believe there will be a rapid intensification period in the next 12 to 24 hours. There is a small window when the approaching troph creates an ideal environment as outflow is enhanced and shear is minimal. Then as quickly as these conditions appear, the troph gets too close, impenges on the outflow, and returns strong shear. This also is a recurvature event that is misunderstood. This is often blamed on eyewall replacement cycles, however many storms undergo these cycles and maintain a constant (but fluctuating level). These storms actually weaken from external, noy internal forces. Classic example of this would be Opal. However, we need not look any further back than Dennis, and to some extent, Ivan.
Katrina may reach as high as a weak Cat 5 and scare some people to death, and then back off before landfall a Cat or two. Still a strong and dealy beast, but there will be a bit of a let down for the media.