Conceptually these reactors are all mostly pressurized water reactors. The first domestic nuclear power reactor in fact, the reactor as Shippingport, PA, was simply a submarine reactor built on land. The reactor ran from 1957 until 1982 until he was decommissioned and decontaminated.
It was a tiny reactor, 90 MWe. The reactor operated as a light water breeder reactor on the Thorium fuel cycle from 1977 until shutdown. According to this link, the reactor contained 1.3% more fissile material than it did at start-up during the five year run.
http://www.atomicinsights.com/oct95/LWBR_oct95.html(I'm not sure I believe this totally, although there may have been elements of the design, probably a compact core, that made it possible to do this in spite of the relatively high neutron capture cross section of light water.)
The Shippingport reactor was shutdown mostly because it was too small to be economical. Because it required special fuel, it just wasn't worth running.
Seaborne reactors differ from commercial larger scale power reactors in one very important way. Because these reactors are compact, and because they are required to be comparatively lightweight, they use very highly enriched fuel. Because commercial reactors do not have to float, they can afford to have huge amounts of uranium that is relatively dilute. Commercial reactors use relatively low enriched nuclear fuels 2-3% U-235. A typical 1000 MWe nuclear reactor may contain up to 100 MT of fuel. (This amount of fuel typically provides a full load for as much as two years of continuous operation - during which very little of it is actually consumed.)
Several ships having nuclear reactors have been lost at sea, including the Thresher (1963) and the Scorpion (1968). Neither accident apparently involved the reactors, although the exact cause of the Thresher sinking is not precisely known. It is believed that the submarine, which was designed to reach unprecedented depths, failed during a depth trial due to an implosion resulting from the failure of its hull. The Scorpion was apparently torpedoed by another American submarine during a war games accident.
In spite of being radioactive, I don't think that the Thresher has ever been found or detected in any way. The Scorpion has been filmed however.
In the early 1960's, the navy of the glorious socialist people's republic of the USSR deliberately discarded 8 nuclear reactors, two of which had full fuel loads, into the Barents sea.
Curiously a joint Russian-Norwegian study in 1992 found that the area is not dangerously radioactive. In spite of the glorious people's workers soviet socialist anti-imperialist state's dumping of 85 Peta Beq (A petabequerel is equal to about 27,000 curies) the joint Russian-Norwegian expedition found that the radioactivity in the local water was not particularly high.
http://odin.dep.no/odinarkiv/norsk/dep/md/1995/eng/022005-990538/dok-bn.html#Little%20radioactivity%20and%20limited%20effectsAll seawater is radioactive naturally, since all seawater contains 3 ppm of uranium and its decay products. In addition, as I have discussed here before, the ocean contains about 500 billion curies of potassium-40, all of which derives from the supernovae out of which the earth formed.
A cubic meter of seawater weighs roughly 1030 kg, and contains about 36 kg of salt, of which about 400 grams is potassium. Of this potassium, about 0.0117% or 50 milligrams is radioactive potassium 40. Since the specific activity of K-40 is roughly 260,000 Beq/gram, the radioactivity from naturally occurring K-40 in a cubic meter of seawater is around 12,000 Beq.
(I will be happy to demonstrate to anyone who asks how to do these calculations from this data, but I hope that at least one person will be inspired to try such calculations on their own:
http://atom.kaeri.re.kr/ton/nuc3.htmlhttp://hypertextbook.com/facts/2002/EdwardLaValley.shtmlhttp://omp.gso.uri.edu/doee/science/physical/chsal1.htm)
The Norwegians and Russians, working in the vicinity of the old Soviet dumps, found, for comparison purposes, that the Cs-137 radioactivity in the sea in this area, most of which is said to be attributable to nuclear weapons testing on Novaya Zemlya, based on about 100,000 samples, ranges between 2 and 32 Beq/m^3. Thus at its worst, the contamination is an almost insignificant fraction of natural background.
This said, it is worth noting that all of the people who live near the Atlantic Ocean (in which the Thresher and Scorpion sank) will die.
In spite of the fact that there are hundreds of nuclear reactors at sea, none of which have containment structures, very few have actually caused deaths. In fact the last death associated with an
American military nuclear reactor occurred in 1961 as the result of a steam explosion. These deaths, so far as I am aware, are the only fatal nuclear reactor accidents in US history. (There were two people killed at the Rancho Seco nuclear power plant in a steam explosion in a non-nuclear portion of the plant.)
I personally believe that all warships, nuclear or otherwise, should be decommissioned, because I am a pacifist. That said, I believe that the world would be better served if most of the diesel engines that now drive merchant vessels were replaced with nuclear reactors.