(Warning: the post is very long. It's for my record.)
The post was being written for my Japanese blog in early August, but I may write it up for here as well for my record, in light of the off-handed remark of "300 tonnes of contaminated water leaking every day" by a career bureaucrat from the most powerful government ministry in Japan.
The remark was actually a mere assumption by the official and his office but was reported as "fact" by the reporters from Japanese and foreign media who were present at the press conference and who heard the official repeat the word "assumption" a number of times throughout the press conference.
It was not just established mass media but also some independent journalists who decided to go with the "300 tonnes of contaminated water" assumption, because it fits their narrative of hopeless situation at Fukushima (plant) and in Fukushima (prefecture).
The latest article (8/11/2013) by one of the independent journalists covering the nuclear accident presents an amazing story - a hearsay supposedly directly from workers at the plant that this "300-tonne-a-day contaminated water" is flowing like a river, no, it's worse, it's pouring out like a waterfall. Needless to say, this article in Japanese is being retweeted at a furious pace. No doubt it will be translated into English and spread throughout the world very soon, if not already.
After the remark by the METI bureaucrat was widely reported as "fact", the mental image held by people outside Japan (and many inside) seems to be that the "highly" contaminated water directly from the basements of the turbine building at Fukushima I Nuke Plant is flooding into the Pacific Ocean unchecked and polluting marine life, some experts claim as far away as in southern hemisphere.
The actual "highly" contaminated water, whose contamination is so high that it is expressed in Bq/cm3 instead of Bq/Liter (as in "millions of becquerels of radioactive cesium per cubic centimeter"), sits in the underground trenches coming from the turbine buildings. It's been mostly sitting there since April and May 2011.
These trenches may have been damaged by March 11, 2011 earthquake and tsunami, or by subsequent strong aftershocks, and part of this water may have been seeping into the layers of crushed stones beneath the trenches and into the surrounding soil, which then gets mixed with the groundwater that's been flowing from the mountain-side (west side) of the plant. Radioactive cesium, all-beta including strontium, tritium, and several other gamma nuclides have been detected in the sample water (probably this mixture of groundwater and the trench water) from the observation holes dug along the embankment.
The density of radioactive cesium in the sample water from the observation holes is expressed in liter instead of in cubic centimeter, as the contamination is far less than the water in the trenches, and as the contamination of seawater is measured in liter. Instead of millions of becquerels of cesium per cubic centimeter (or in liter, billions of becquerels) as in the water in the trenches, the sample water from the observation holes has between 10s and 1000s becquerels of cesium per liter.
And what about the seawater itself? The seawater samples have been taken almost every day since April 2011 from right outside the water intakes for the reactors, which is located along the embankment. TEPCO calls it "open culvert". Each water intake location is somewhat sealed off by silt screen, and the open culvert is also somewhat sealed off by silt screen from the rest of the plant harbor.
Diagram from TEPCO's "Enhancement of Monitoring Plan at the East Side of Unit 1-4 Turbine Buildings and in the Port at Fukushima Daiichi Nuclear Power Station", 6/26/2013, showing the plant harbor and the open culvert area (in gray) where the water intakes are located. Silt screens are indicated in curved lines in red:

As of August 1, 2013, the highest measurement of radioactive cesium in seawater was 66 becquerels/liter of cesium-134, and 140 becquerels/liter of cesium-137, inside the silt screen of Reactor 3 water intake in the open culvert. Outside the silt screen, the numbers drop to 13 Bq/L for cesium-134, and 23 Bq/L for cesium-137.
Maximum amount of cesium in discharge water allowed for Fukushima I Nuclear Power Plant BEFORE the accident, per liter:
From TEPCO's Nuclide Analysis Result Page (in Japanese only):
Fukushima I Nuclear Power Plant nuclide analysis of seawater inside the plant harbor, released on August 1, 2013
Now, going back in time, what was the density of radioactive cesium in the extremely contaminated water that was pouring out from the crack in the concrete wall near the Reactor 2 water intake screen in April 2011?
(Photo from TEPCO's Photos and Videos 4/2/2011)
What about the density of radioactive cesium in seawater when the above "highly" contaminated water was pouring out of the crack?
Nuclide analysis of seawater inside the plant harbor, released on April 5, 2011:
Unit: Bq/cm3 (cubic centimeter, or 1/1000 of liter)
Maximum density: seawater at the Reactor 2 water intake screen Cesium-134 35,000 Bq/cm3 (or 35 million Bq/Liter), Cesium-137 36,000 Bq/cm3 (or 36 million Bq/Liter)
The density of radioactive cesium in seawater then rapidly fell down. Nearly three months later, the density was low enough to be expressed in becquerels per liter, instead of becquerels per cubic centimeter.
Nuclide analysis of seawater inside the plant harbor, released on July 31, 2011
Unit: Bq/L (Liter)
Maximum: Seawater in front of Reactor 3 water intake screen (inside the silt fence) Cesium-134 1,300 Bq/L, Cesium-137 1,500 Bq/L
Expressed in liter, the density of radioactive cesium in seawater went from 71 million Bq/L in April to 2,800 Bq/L in July, 2011.
By the end of 2011, the density of radioactive cesium in seawater right in front of water intake screens dropped to three digits, from 4 digits in July.
Nuclide analysis of seawater inside the plant harbor, released on December 22, 2011
One year later in December 2012, the density further dropped to two digits, but still above the legal limit for discharge water from the normal nuclear power plant (Cs-134: 60 Bq/L, Cs-137: 90 Bq/L).
Nuclide analysis of seawater inside the plant harbor, released on December 31, 2012
In early 2013, the density dropped below the legal limit of radioactive cesium in the discharge water from the normal nuclear power plant (again, Cs-134: 60 Bq/L, Cs-137: 90 Bq/L).
Here's from April 1, 2013, one day before TEPCO started the construction of the impermeable wall by driving down sheet piles with vibratory hammer along the embankment.
Nuclide analysis of seawater inside the plant harbor, released on April 1, 2013
The data released on April 2 and April 3 do not show much difference. However, from the seawater sample taken on April 3, one day after the construction using sheet piles and vibratory hammer started, radioactive cesium was found again exceeding the legal limit for Cs-137.
Nuclide analysis of seawater inside the plant harbor, released on April 4, 2013
Depending on the locations, some samples show the density of radioactive cesium one order of magnitude larger than the samples taken before the start of the construction work, from two digits to three digits. Since then, the density of cesium in the seawater in front of water intake screens has been fluctuating between two-digit numbers and three-digit numbers.
Here's a chart plotting the density of radioactive cesium in seawater at the Reactor 3 water intake screen, inside the silt fence. It's a composite of the chart released on April 20 and the chart released on August 1. The thin red line indicates 10 to the power of 2 (1.0E+2). (Y-axis is in log scale.)

Toward the end of April, as the construction of the impermeable wall with sheet piles progressed, the density of radioactive cesium in seawater started to fluctuate wider.
Still, as far as radioactive cesium is concerned, the seawater in front of the water intake screens has not at all reached the level of contamination that obtained when the truly highly contaminated water from the Reactor 2 turbine building was pouring out into the ocean in April 2011.
It is safe, I believe, to assume that the highly contaminated water that still exists in the trenches and which has radioactive cesium in billions of becquerels per liter is not leaking in great quantity into the seawater yet.
If the density of radioactive cesium in the seawater in front of the water intake increases to 4 digit numbers per liter, I might start to worry. So far, it hasn't happened, despite TEPCO's (and METI/Agency of Natural Resources and Energy's) ill-conceived attempt to stop the groundwater flow with waterglass.