In the African nation of Gabon, there is a town called Oklo that stands atop some rich deposits of uranium. Time and the geology of the area created a circumstance that has been observed only once. Some billions of years ago, the concentration of fissionable uranium in these deposits was high enough that they could form a natural nuclear reactor. By itself, though, the uranium was not sufficient. The spontaneous fission of a U235 nucleus releases a neutron too energetic to trigger the fission of another U235 nucleus. As in any human-made reactor, a "moderator" is required to slow the neutrons, and as in many human-made reactors, water served this role. At Oklo--and as far as we know, only at Oklo--groundwater flowed through the uranium deposits, slowing the neutrons, allowing a chain reaction, which then released energy.
However, this release of energy, as heat, caused the groundwater to boil. The reactor effectively killed itself--until it could cool down, and water flowed into it, bringing it back to life. Some modeling and experimentation suggests that the reactor would be "on" for about a half hour, and "off" for two and a half, and that this cycle lasted for thousands of years.
Today, tucked away in a corner of our main pasture, there stands a majestic compost pile. I'm kind of embarrassed by it; it is in a poor location, on soil rather than concrete, and it is not covered. It reminds me of the Oklo reactor because of its relationship to water. It has been accumulating raw material all summer, and with each new dose--which carried with it a deal of moisture--it would ferment, steam, fume and stink for a few days. It would cook off all the freshly added water, and the reactions would slow down and stop. After a light rain earlier this summer, it again seethed for a bit, and died.
The reactions, of course, are not nuclear. There's a lot of fermentation, to be sure--the sheep's manure and used bedding contains lots of reduced carbon. Where there's fermentation and anaerobiosis, there's going to be methanogenesis and acetogenesis, forms of respiration using carbon dioxide in the place of oxygen. There's also respiration going on using sulfate in place of oxygen; the stink of hydrogen sulfide attests to that fact. All these reactions produce energy for the microbes carrying them out. The second law of thermodynamics tells us that the universe can't allow any reaction to happen without a tax being paid to entropy, so a sizeable chunk of the energy from these reactions is lost as heat--heat that makes the compost warm to the touch, and that has dried the pile out every time it has been wetted.
Now we've just had a good rainfall, thoroughly soaking the pile. Hoooo-weee! It is reacting like mad, smelling and steaming. And so the cycle repeats, a natural reactor getting drenched with water and generating heat to turn itself off, then cooling down and starting over--though probably not for thousands of years, and while the stink might be annoying, it's not producing anything as bad as plutonium.