There was possibility 50-60 hours warning of the possibility. I sent an e-mail to a friend in Baton Rouge on Friday 8/26 at about 5:30 PM ET, stating that he was on the "western edge" of the zone. The problem with disasters is that there is usually very little warning.
The 5 AM Friday forecast predicted about 90 KT winds and a landfall around Pensacola. And said "it is certainly possible" that Katrina could strengthen to a major category storm (3-5).
The 11 AM Friday forecast predicted about 110 KT winds and said strengthening to a major storm was expect.
The 5 PM Friday forecast predicted about 115 KT winds and moved the forecast for significantly westward. This is at about 60 hours out.
The first forecast that predicted really high storm surge (15-20 feet) with 25 foot waves was 10 PM Saturday.
The following is kind of interesting if you know where the 17 Street Canal is. This simulation used a eye going west of New Orleans that pulled a storm surge up into Lake Pontchartrain, but the flooding appears to have come from either simulated breaches of the Mississippi levy or overtopping.
Hurricane Pam Simulation This shows a modeling of the storm surge for Katrina. The text indicates that it includes simulated breaches of the floodwalls (eg. the floodwall along the 17th Street Canal was reduced to ground level).
Hurricane Katrina Hindcast Notice the area west of Jefferson Parish. This must be like New Orleans and Jefferson looked like before the Lake Pontchartrain levees were built.
New Orleans elevations. July article in USN&WR
7/18/2005 US News & World Report A 2000 article from a insurance trade publication predicts $10B insurance loss from New Orleans hurricane.
Lost City of New Orleans