Just to tie this up, I've dug out one of the presentations from the Glasgow Congress I attended. This is from one of the big wigs in Radiation Protection, Abel Gonzalez
Biographical Sketch - Abel Julio González
He presented the the following limits for various products (NB a Bq is a measure of the radioactivity - the number of atomic transformations per second):
Water - according to the WHO Drinking water quality guidelines - 10 Bq/l for Caesium
Juice - as a food stuff 1000 Bq/l for Caesium.
Rice - 1000 Bq/kg as a food stuff
Rice paper - 100 Bq/kg based on an IAEA safety guide.
For the public these are vastly different levels and they appear inconsistent. Which is "safe" water vs juice why can we drink juice but not water? Is the juice dangerous? Obvious and natural questions if you don't have any background. The various levels are set by modelling intake and dose arising over an extended time period in order to keep people's exposure below a certain, low dose level.
The key thing is that these number are for long term exposure. So over a short term you could eat/drink something that exceeds the levels and still be safe. Take the water level at 10 Bq/l, drinking the average 2 l per day. That would be a limit of 20 Bq per day and 7300 Bq per year as a total intake. Over a short term it wouldn't be unsafe to drink say 40Bq in a day.
Anyway the fish limit has been reduced from 1000 Bq/kg to 500 Bq/kg - as best I can make out from the presentation. Basically to rationalise the various "safe" levels into a more consistent set of limits. There is now 1 level for Grains, fruit and veg, meat, eggs fish and etc. and 1 limit for Water, milk and dairy products.
There was a couple of interesting talks in Glasgow about the levels. They have moved to rationalise levels, which might explain the reduction. One of the findings was the public were confused about levels and that many similar things had different levels. For example juices as a foodstuff had a different level to water. I can't remember the exact numbers but it was like one had a level of 1000 and the other a level of 100. This one talk listed a whole range of discrepancies and basically made the point that this is bad for the public who are looking for a "safe" level and don't have the professional knowledge to know why there are different levels.
Not sure if this makes sense as such. I'll try to remember to pull out a summary or something.