Stohl et al. have recently published an article “The total release of xenon-133 from the Fukushima Dai-ichi nuclear power plant accident” (pdf) in the Journal of Environmental Radioactivity. They used xenon-133 measurements from a set of northern hemisphere monitoring stations of the CTBTO and compared it with estimates of xenon-133 inventory in rectors 1-3 of Daiichi. The amount of xenon in these reactors is known accurately, it is about 12.2 EBq (12.2E18 becquerels).
The total release of xenon was estimated from monitoring stations’ data in 3 ways. The first was a simple observation-based box model, and it yielded an estimate of 16.7 EBq. The other two ways were more sophisticated, they integrated results from the GFS (American) and ECMWF (European) meteorological reanalysis models. The GFS gave a figure of 14.2 EBq of Xe-133, while the ECMWF came up with 19.0 EBq. The authors chose the lowball figure of 14.2 EBq from the GFS, claiming it was a better fit.
This still leaves 2 EBq of Xe-133 unaccounted for. Since iodine-133 decays into xenon-133, they assumed that all the I-133 was released from these reactors. This equalled the missing 2 EBq.
The first problem is, why would all the I-133 be released? There is still I-131 being released, and it has been increasing over the past few months. It should be the same case with I-133.
The second problem is that the other two methods came up with higher amounts. The ECMWF in particular was 4.8 EBq higher than the GFS. It is strange that the authors would prefer the GFS over the ECMWF. It is well known, and can be demonstrated statistically, that the ECMWF is the superior model in forecasting. Every meteorologist and weather weenie knows this. The Capital Weather Gang noted this and quoted as to the American GFS model:
“It is patently unacceptable for the United States – given its extraordinary need for accurate weather and climate information across all sectors of society – to operate a global forecast system that lags well behind those of other nations…”
(This also extends to historical reanalysis, as studied in this pdf). It’s actually SHOCKING to see someone preferring the GFS over the ECMWF for anything.
So if you go with the ECMWF, you have 6.8 EBq (that is 6,800,000,000,000,000,000 becquerels) of Xe-133 unaccounted for. This is over 50% more than the inventory at reactors 1-3. Where did it come from? The spent fuel pools? Unlikely. They are emitting cesium and plutonium and the like, but you wouldn’t expect them to go critical. Reactors 5 and 6? Daini? It’s got to come from somewhere. And, in addition, there are isotopes like thallium-208 that were measured by the EPA, and in other papers, the existence of which suggest that it came from re-processing or fast breeding, which should not have been happening at Fukushima Daiichi. Where are the missing meltdowns?