A monazite sand placer deposit rich in thorium dioxide is slowly crushed into a sedimentary rock by plate tectonics. Water in it acts as a neutron moderator which ensures the ore formation doesn't blow itself apart before it forms. For some reason, the water vanishes, so the neutron spectrum hardens — i.e. neutrons can be more energetic because there's no moderator to slow them. Spontaneous fission by the 232Th that makes up all the thorium in the formation can now breed fairly large quantities of 233U; previously, the water slowed the neutrons down into thermal neutrons, but now they can be 1+ MeV fast neutrons capable of fissioning 232Th. The 233U — which, unlike 232Th, can sustain a nuclear chain reaction — absorbs spontaneous fission neutrons from the 232Th, starting a runaway chain reaction among itself and eventually destroying the entire ore formation in a gigantic (~750 gigatons/~3.138E21 joules) nuclear explosion.
This is comparable in energy to a supervolcanic eruption (for reference, a Yellowstone eruption would be in the ~875-gigaton/3.661E12 joule range). Due to the extremely low purity of its fissile material, it's highly inefficient — perhaps 1 in 1 million parts of the 233U fission, as opposed to the 1 in 100 to 1 in 10 of early human-made fission explosives) — but its sheer size makes up for that lack of efficiency, with nearly 38,300 metric tons of 233U managing to fission before the ore formation blows itself apart.
The 1.7-kiloton Plumbob Rainier was totally contained only when detonated 899 feet underground. I feel confident an explosion 8+ orders of magnitude more energetic wouldn't fail to punch through 9 times more ground and therefore that radioactive fission products from this will enter the atmosphere. However, I have no idea of the relationship between nuclear detonation magnitude and resultant fission product radioactivity produced, nor the relationship between impact energy and the mass of dust kicked into the atmosphere — a significant consideration, given the amount of energy involved here.
With this in mind, the question is: how bad is this going to be for the environment/biosphere?
There are many factors to consider:
- the initial shockwaves — expressed both as atmospheric overpressure and via earthquake
- debris rainout
- ash and dust injected into the atmosphere; if you can, provide a quantity in tons so I can compare it with the Chicxulub impact
- radionuclides injected into the atmosphere; if you can, provide a quantity in becquerels/curies/etc. so I can compare it with Chernobyl
Assume this is detonated on Earth. Where on Earth? I'm not sure it matters, this thing is going to be a problem for all of the Earth. If it does matter, assume it pops off right in the middle of Kansas City.
Bonus: how quickly afterwards would it be possible to safely colonize the multi-tens-of-kilometers-wide crater I assume would be left over? The radiation would die down within the year — the Sedan crater proved that much — but what about the heat?