Katrina hindsight

Instapundit points us to an initial analysis on the Popular Mechanics Science Blog that rips the draft report of the Senate Committee investigating the Hurricane Katrina response, finding it "riddled with poor logic, internal contradictions and exaggerations":

For now, though, here’s a quick overview of what seems to be the report’s most troubling shortfall: consistently blaming individuals for failing to foresee circumstances that only became clear with the laser-sharp vision of hindsight.

For example, the report states:

"Fifty-six hours prior to landfall, Hurricane Katrina presented an extremely high probability threat that 75 percent of New Orleans would be flooded, tens of thousands of residents may be killed, hundreds of thousands trapped in flood waters up to 20 feet, hundreds of thousands of homes and other structures destroyed, a million people evacuated from their homes, and the greater New Orleans area would be rendered uninhabitable for several months or years."

This statistic is referred to often, and refers to computer modeling of a direct Category 5 hurricane landfall in New Orleans. However, it's also a distortion. According to the data the Committee itself examined, 56 hours prior to landfall, Katrina was a relatively weak Category 3 storm, heading west in the Gulf of Mexico. Over the next few hours, it began its turn north, but where the storm was going to make landfall along the Gulf Coast was any weatherman's bet (the average 48-hour margin of error is 160 miles). In fact, it was not until the next day, Saturday, that it became more of a certainty that the hurricane was heading toward New Orleans. Furthermore, hurricane forecasters and emergency managers tell PM that until about 24 hours before landfall, hurricanes are too unpredictable to warrant the sort of blanket evacuation orders the report describes.

Indeed, at 56 hours before landfall (which was around 6am CDT on Aug. 29), the National Hurricane Center was forecasting an aggregate probability of only 17% that Katrina's center would strike near New Orleans - meaning there was an 83% chance that it would not do so - hardly an "extremely high probability", as the Committee report would have it.

Moreover, as you can see from the NHC's tracking map at the time (which is a composite of several different computer forecasting models), the potential strike zone covered about 500 miles of the Gulf Coast. At 56 hours before landfall, any slight jog in the storm track could have sent it far to the east or west of New Orleans.

When my state reserve unit in Georgia was activated for Katrina response, I spent part of the time working shifts in the National Guard's Joint Operations Center at Dobbins Air Reserve Base, where there were TV monitors showing the cable news coverage of the storm's aftermath. Those of us working in the JOC that week would either shake our heads in disgust or simply laugh out loud at the inaccuracies and distortions being propagated by the media; now nearly half a year after the disaster it appears the distortions continue.