http://switchboard.nrdc.org/blogs/mmckinzie/fukushima_radiation_risks_from.html
Assessing the Radiation Risks Associated With the Leaks at Fukushima
Thomas B. Cochran, Ph.D., September 2013
To keep the risk below 1x10-5 the consumer must limit his/her dietary intake to less than 10,000 Bq of cesium-137. Therefore this risk limit would be reached after eating about 0.7 kg of fish meat. While this is a conservative estimate of what is required to achieve a low risk, one could make a good case for quarantining fishing off the Japanese coast near Fukushima, which of course is what the Japanese government has done.
Near the west coast of the United States the maximum projected concentration is about 30 Bq/m3 some three years after the initial release. This is 5,000 times less than the 150,000 Bq/m3 concentration we have assumed near the Japanese coast. Therefore, to keep the risk below 1x10-5 the consumer must limit his/her dietary intake to less than about 3,000 kg (3 tonnes) of fish. In other words, do not worry about eating fish taken from US coastal waters. Since the concentration projected for waters near the Hawaiian Archipelago are even less than that projected for the West Coast, the same admonition applies to Hawaii.
...
Case 2: Chronic Leakage into the Sea from Fukushima.
From samples of seawater, Jota Kanda of Tokyo University of Marine Science and Technology estimated last year that about 0.3 TBq of radioactive material are leaking into the sea each month. And in this article in New Scientist, Ken Buesseler says the Kanda estimate is probably the best he is aware of, and closely matches figures released on 21 August by TEPCO, of 0.1 to 0.6 TBq per month for cesium-137 and 0.1 to 0.3 for strontium. At an average leak rate of 0.3 TBq/month,
it would take more than 6,000 years to equal the 22 PBq release assumed under case one above. Consequently, the current chronic leaks do not increase the risks associated with consuming fish caught in waters off the west coast of the United States or Hawaii.
Case 3: Recent leaking tanks. It has been reported in the press that a storage tank containing 24 terra-Becquerels (TBq = 1x1012 Bq = 1,000 billion Bq) of radioactivity in in 300 cubic meters of water was leaking.
Let’s be exceedingly conservative by assuming all the radioactivity is cesium-137 and all of it leaks into the groundwater and from there into the sea over a very short period of time. The source term, 24 TBq, is about 1,000 times smaller (and a couple of years later) than the source term used in modeling by Rossi, et al. Clearly, this does not change the conclusions regarding eating fish caught near the West Coast or Hawaii.
CONCLUSIONS
For the foreseeable future, one should avoid eating fish caught near Fukushima. Buesseler says that during his own sampling survey in waters 30 to 600 kilometres from Fukushima in June 2011, three months after the meltdown, the highest levels he found were 3 Bq/liter of cesium-137. This suggests that the consumption of fish caught in these waters would not represent a significant risk to individuals. There is not a significant radiological risk to individuals associated with consuming fish caught near the West coast of the United States and Hawaii.
Ken Buesseler notes that the north Pacific contains an estimated 100 PBq of cesium-137 from H-bomb testing in the 1960s, so the fallout from Fukushima is adding only a fraction of that. Total discharges from the Sellafield nuclear plant in the UK released 39 PBq over 40 years of operation, according to Buessler.