News blog

Poor science fails to stop Tokyo’s fish market move

Posted for David Cyranoski

The cleanup of the contaminated site that’s to be a new home for Tokyo’s fish market, the world’s biggest, is still dogged by controversy. As Nature reported back in April, the figures for concentrations of toxic chemicals at the site before the cleanup appear in key documents crossed out with black marker. Now the authorities have admitted in a report published last week (22 July), after consultation with experts, that this is not proper science. Yet that hasn’t stopped its plans for the move.

The government hopes to move the crowded Tsukiji Fish Market to a spot 2 kilometres away — an old gas plant site tainted by high concentrations of at least seven chemicals. From January, it was leading experiments to demonstrate that the chemicals could be reduced to safe levels. In March, it presented figures showing dramatic success.

The most egregious example was benzene levels. The experiment reduced the levels to 0.003 mg/litre. The government claimed this was a huge reduction from the previous level, proving its technology worked. But what was the previous level? The only one in the tables was was 430 mg/litre, which was actually found in a sample taken from the new site.

It would have been a huge reduction. But that was from soil taken by a different survey at a different time and different spot within the site. Apples and oranges. What was the actual pre-experiment level for the sample used in the experiment?

When Japanese Communist Party members of the city’s metropolitan assembly asked for the data, they got it. But they found a problem: the blotted-out “before” data. The government excused itself, saying that there was an inconsistency that needed to be clarified so that they didn’t confuse the public.

Some people, like me, might argue that they should have sorted out the inconsistency before presenting results at the March budget hearing. Or not presented them at all.

So in the end, were the data consistent with their interpretation? The answer, we found out late last week when the government presented its data at technical meeting held to plan for the move, is no.

It turns out that the benzene levels in the sample used in the experiment was 0.027. The reduction to 0.003 was minuscule compared to what would need to be done at those sites where the levels were actually 430 mg/litre.

The government explained that different samples from the site must have been mixed together, diluting the high concentration. In documents presented at the meeting, the authorities describe the mix-up as a cautionary procedure — withholding data until they could “hear expert opinion about the meaning of pre-experiment data” — rather than a cover up. The conclusion, after listening to their experts, was that “pre-experiment data should be included with data that was gathered during the experiment.” Indeed.

The government is undeterred. To prove its case, it claims to have carried out another experiment in which scientists artificially produced soil samples with benzene levels of 2000 mg/litre and reduced them to 0.005.

But will the artificially produced samples act the same as the real ones? And, having shown such a weak understanding of scientific data, is any one going to believe it without a thorough check by a thoroughly independent scientific group?


  1. Report this comment

    Uncle Al said:

    Benzene is leukemogenic over 20+ years of exposure. One fails to see any health hazard arising through 2030.

Comments are closed.