alt Hacker NewsTo put a number on it, linear no threshold models predict ~130 deaths as a result of the radiation (and are known to over-estimate lethalities at low doses).
Around 50 people a year die while clearing snow in Japan, so it's ~ twice as dangerous as shoveling snow in worst-case predictions.
LNT is not known to over-estimate lethalities at low doses. The actual situation is that the predicted deaths at low doses occur at such a low rate that the signal cannot be detected above the noise. That doesn't mean the prediction was wrong, just that it cannot be verified. It's possible (as in, consistent with evidence) that LNT under-predicts deaths at low doses.