Thanks sandm. A reasonable voice.
Originally I was trying to offer Eric some advice with a degree of substantiation but without getting too technical. There is much more here than just 'boundary loading' or other regurgitated stuff that someone read on the net.
Think this through. This is real. There will be a test.
The speed of sound is 1128 feet per second. So let's say we gladly sacrifice all really deep base knowing we are in an open field environment without reinforcing planes. But if we are running a party barge with four 12s we at minimum expect some good Michael Jackson Thriller type bass. Good ole punchy bass. So let's focus on 55 Hz. You are not going to get deeper bass anyway in this situation. Okay, that constitutes a 21 foot long waveform. At half that length (10.5 ft.) when the perception point (your ears) is close to one group of woofers then the output of the two woofer groups are out of phase and directly cancel one another. The cancellation is not as steep on either side of the 55 Hz frequency but the negative impact is still impactful. Does that 10.5 ft. distance sound somewhat coincidental to the distance between the upper and lower decks? Then compound that with the comb filter (cancellations) effects of the water plane and various points of perception. All variables and 90 percent of them are bad. Again, reflected energy out of phase with the direct energy at a given point of perception. The woofers and driving amplifiers will work against a huge acoustic handicap over their entire range where you expect a contribution. Soooo, I recommend a single and collected woofer position as close to the water level as possible (lower deck). Make a controlled compromise. Change the design model a bit and you will be far happier in the end. Seems simple enough to me.

David