Some Considerations in Spent Fuel Pool Fires

This morning, I read the Brookhaven National Labs report Severe Accidents in Spent Fuel Pools in Support of Generic Safety Issue 82 from 1987.  This is an analysis of the risk of large scale radioactivity release from the spent fuel stored onsite in a water pool at a nuclear reactor, in the case that an accident caused a loss of the water in the pool.  This is something that now may have happened (for the first time) in the pool at Fukushima-Daiichi.

The document is long and complex, and there are many very significant uncertainties - to the point where I judge that no-one will be able to reliably predict what may happen in this particular accident.  However, the key points seem to me as follows:

• The main risk is if the amount of heat released from the fuel is enough to cause the Zircaloy cladding of the fuel rods to catch fire.  In that case, much of the radioactive material in the burning fuel rods will be dispersed into the atmosphere (if the building doesn't contain it, which it might not given the existing building damage at Fukushima-Daiichi).
• If the rods don't catch fire, they will probably just sit in place and be intensely radioactive, but not release huge amounts of radioactivity over long distances.
• If the entire pool were to burn and be atmospherically dispersed (the absolute worst case), the Brookhaven report estimates that a couple of hundred square miles of territory would be rendered permanently uninhabitable (ie towards a Chernobyl scale event).
• The amount of heat produced by spent fuel decays with a half life of about a year.  Thus it is the most recent fuel-rods that are the greatest risk of starting a fire (nuclear reactors have their fuel changed out about every 18 months, so that is the frequency of new additions to the storage pool).  This is the concern with the pool at reactor 4, which has just had a reactor unload event.
• Whether a fire can start, and how far it can propagate into the older less hot fuel is very dependent on details of the fuel storage geometry that we have no information on at Fukushima.
• If the newest fuel becomes uncovered, a fire is likely to start fairly quickly if it's going to (certainly in the first day).  Given the lack of clear reporting in the present situation, it's possible such a fire already happened and was put out.  It's also possible there is still enough water to prevent the hottest fuel having it's Zircaloy cladding from catching fire, and only hydrogen fires have occurred.
• In the scenario in which the water is slowly boiling off, rather than having all drained at once, it would seem hard for the whole assembly to burn.

My sense is that the worst case certainly hasn't occurred at present (the entire lot burning up would be extremely visible in video of the plant), and probably isn't likely.  The most likely scenario to result in that would be a catastrophic failure of one of the pools - perhaps damaged in the (9.0!) earthquake, and then failing totally in a subsequent explosion.

It's probably worth highlighting - the fact that none of these pools drained when the plant was first hit by a 9.0 earthquake close by, and then a massive tsunami, is pretty remarkable and some kind of testament to the quality of the original engineering.