Physicist Ellery Schempp, a friend of sf writer Jim Morrow's and a fascinating character in his own right, sent in the following interesting analysis of the levee failures:

First of all, we recognize that the flooding in New Orleans was predicted in an article in Scientific American in October, 2001.   There have been many disputes between environmental groups and the Army Corps of Engineers as to how to manage the lower Mississippi.  It surely is not a black and white matter.

There is dispute as to when the first levee broke--some say it was ca. 10 AM on Monday on the 17th St canal.   Others say it was later that evening.  In either case, all the catastrophe and suffering occurred after the levees broke, which resulted in flooding the city.  Before that moment, the remaining residents were not in severe danger.  All the horrors occurred because of the flooding.  All the future expenses are due to the flooding.

It is truly amazing that no one seems to know when the first levee wall broke.  How could it be that people were not monitoring the walls?  Relatively simple and inexpensive measures could have been implemented in the months before.  The levees did not fail because the water was so high that it overran them; it wasn't the direct effects of a category 5 hurricane that destroyed New Orleans.  In fact, 99% of the miles of levees held perfectly well--there were two breaches, one about 50 ft wide on the 17th Street Canal and the other about 200 ft long on the Lake Pontchartrain levee.  Maybe there was a third one--no account seems to know.  How could people not know?

I emphasize that this was not a case of an overflowing bathtub.  Cement walls failed, not due to a category 5 hurricane, but because there was no system in place to monitor and support the levees.

This  is well explained in the graphics of the recent History Channel description.  No earthen levee or cement wall failed { via snap}; in every case the failure took place over hours.

What should have been done:
1. All the major levees should have been outfitted with electronic sensors to monitor them.  There are a host of simple means from simple water sensors on the dry side (shielded from rain), to breakable wires to monitor continuity, to tiltmeters--dozens of technologies geologists use to monitor volcanoes and earthquake zones.  These could all have been connected to redundant monitoring centers provided with multiple battery and diesel generator backup.  In the event of an incipient failure, alarms would be sent pinpointing the area.

2. The important levees could have been constantly patrolled on the parallel roads so any signs of trouble would be seen quickly.   Trucks filled with sandbags and crews ready to work would be stationed at intervals.  Mobile cranes, too.   

3. How did the levees fail?  The most probable scenario is this.  Levees are at risk of water infiltrating under the barrier, percolating through and undermining the footing.  This is well known, so engineers prevent this by carefully sealing the water side with mixtures of clay and grout to prevent any water infiltration.  However, the dry side is usually not so protected. 

Probably, water started to slosh over from waves.  As it did, it ran down the dry-side walls and began to undermine the easily-penetrated dry-side foundations, possibly somewhat weakened by the heavy rains.  Eventually, one section began to tilt slightly, allowing more water over, more erosion, and from then the result was foregone.  The first breach probably was over only one or two 8-foot sections.  But as everyone knows, as soon as water starts flowing fast through a breach, sections on either side go quickly.  So a small breach became 200 feet wide.

4. Had the sensor monitoring system been in place and had the patrols been in place, this could have been prevented.  Sensors and patrols would have noticed the water coming over and immediately two-feet of sandbags placed on top of the levee, preventing further undermining.  The levees would have been saved, and there would have been no massive flooding.

5. Assume that even with monitoring for an incipient situation and timely response, a cement wall breaks {snap?  there were no warning signs?}.  How to stop the rush of water?  Well, you have barges prepared for this--they move into place and block the flow enough.  Takes some engineering to make it work, but certainly doable.  Street-based crews provide back up.

6. In a worst case scenario, the levees would have failed despite these early warning and corrective activities.  In that case, the monitoring and patrols would have been able to raise the alarm.  In a well-organized emergency management system, immediately loudspeaker trucks and buses would fan out throughout the city.  "The levees have broken,  The levees have broken.  You must leave now.  A flood is coming.  You must leave now." 

At the time the levees broke, there was a 12-18 hour window to get into the city and evacuate people.  In New Orleans, when the levees broke, everybody was clueless,  there was no early warning, and there was no effort whatever to repair a small break before it became unstoppable.  There was no effort to evacuate the residents with warnings of the NEW danger, after the hurricane had passed. 

It is clear that a mere $50 million in monitoring technology and levee patrolling and response teams would have saved $400,000 million in future costs.

There is no excuse for the lack of a constant monitoring system; there is no excuse for not patrolling the levee walls; there is no excuse for not having a response plan to prevent a minor break from expanding to flood the entire city.  Incompetence and stupidity are high on the list.  I mean, why wasn't a simple monitoring system for failure in place?

(One minor point: The people swarming around this web site were able to verifiy three breaches in three different locations by matching media images to satellite images; and as far as I know we were the first "media" source to do so. Pretty sad, yes? I entirely agree with Schempp's general point about the shocking lack of information about the breaches.)

Schempp's CV concludes:

In a 1999 interview, Ellery said:  “I have had the good fortune to have worked 400 miles from the North Pole and then 800 miles from the South Pole thanks to Prof. Robert Nichols at Tufts; I enjoyed these unique places immensely and they were transforming in my life; I had the good fortune to have the Supreme Court agree with my understanding of the First Amendment, with lots of help from the ACLU and my parents.  I had the good fortune to do a PhD in physics at Brown, with support from Prof. Philip Bray and many professors.  I am happy to be acknowledged, and I guess I made some contributions, but I think many others are to be recognized.”

He is also the author of the widely circulated satirical essay, Warning: Gravity is “Only a Theory.” Of his gravity essay, he remarks, " Amazingly, I found a lot of responses that took me seriously.  I therefore had to write an annotated version that explains the science and jokes. "