Brian, the how many times has the Mississippi flooded and how many earthquakes were there is the good point, and I suggested that someone could check this. A lot of the rest of what you say I partly, but not entirely, disagree with.
The Flow on the Mississippi is 1.8 million cubic feet/sec. now. I don't know what the flow is at an average stage but I think it would be a fraction. Yes, the water in the channel would be flowing faster but the mass of water in one area would likely be a couple times average now. So, is not a trivial mass. Yes, the New Madrid seismicity is probably quite deep; may even be >10 km. But, under certain conditions deep quakes can be triggered by reservoirs: the question is whether it can be by increased pore pressure at great depths. Logically, you could be correct. But, talking about water getting into the ground is not how it works in some areas: if you have an open fracture, just making the level higher at the surface will increase the pressure at depth. Think of a 1 km-high vertical pipe into the ground, sticking 100 m up into the air. If you change the water height in that pipe from ground level to the top of the pipe, you increase the pore pressure at 1 km depth by, I think, 10 atmospheres (is that the same as 10 bars?...I think so; atmospheric pressure is about 1000 millibars). Static stress changes known to trigger earthquakes can be a fraction or a couple of bars.

In many places the levees are pretty far apart: a good part of the flood plain is allowed to flood. So, I think you underestimate the depth of water outside of the normal riverbank (although I am not sure of this).

Chris