There's a serious unfolding nuclear situation happening, and it has been getting a lot of sometimes a little hysterical press coverage.
UPDATE: There's a really good post describing what's happened right here.
- Peter Ennis makes the point that this is not a Chernobyl-scale disaster. Any effect, even in the worst case, is local and relatively limited. Basically, the reactor cores were shut down properly, but the cores keep generate some heat, so they need active cooling until the temperature drops below water boiling point. It's that shutdown cooling that has been failing.
From this point of view, it is not a strong argument against nuclear power. The plants came through a disaster of historic proportions rather better than most structures. The burning oil refinery in Chiba alone has likely caused more health and environmental damage than these nuclear plant failures.
The major reason the cooling has failed is because the plants depend on active high-volume pumping systems. This is apparently a very old design (the plants ore more than 40 years old) and more modern plants rely on passive systems that wouldn't have failed in this situation. This really is an argument for replacing old outdated plants with modern ones that are much safer, more efficient and generate much less waste.
- Tokyo is preparing for rolling blackouts. About 30% of Japans energy comes from nuclear power, and most plants in Kanto and Tohoku are shut down since the earthquake.
A quaint feature of the Japanese power system is that there's two separate standards: All of Japan uses 100 volts, but eastern Japan, including Tokyo, runs on 50Hz while western Japan, including Osaka and Nagoya, runs on 60Hz. There was no national regulation in effect when Japan was being electrified, and different power companies settled on different standards.
This quaint difference turns into a serious issue now, however. Power plants in western Honshu and Kyushu are on line, but because of the different transmission frequencies it is difficult and inefficient to transfer power from one area to the other. Western Japan can't transfer enough power eastward to compensate for the disaster power losses, and so Tokyo is facing power outages. A large amount of power infrastructure will need to be rebuilt in eastern Japan, and it's certainly possible that they decide it's worth the pain to switch the entire country to 60Hz.
- This post-disaster situation shows how nuclear power is a bad match for a country like Japan. A reactor takes a few days to shut down in an orderly way, and another day to start up again. With any kind of emergency shutdown or disaster damage you will need extensive inspection, testing and possible repairs before you can even ask for permission to start up again. And according to media reports, the plants that have resorted to seawater flooding are almost certainly permanently lost, as the water (with the salt and impurities) damages a lot of sensitive materials in the reactor. It's going to take years to fully restore the capacity.
Gas, coal, oil, pellet and waste-burning plants are apparently fairly quick to restart; there's just a lot fewer sensitive components, there's less to go wrong, and the possible consequences of a fault are a lot less serious than for nuclear plants. Distributed systems like wind, solar and geothermal plants offer lower energy density but are of course very resilient by their distributed nature. Unless they're heavily damaged they can all be online again shortly after a disaster. I know Sweden keeps gas and oil plants mothballed, ready to start up whenever there's a sudden, serious shortfall of the regular power generating systems.
By contrast, nuclear plants have a long cycle time, and depend more than perhaps any other energy system on a distributed, well-functioning network of supporting resources to operate. With a large-scale disaster like this, you can have a major portion of the entire system - even units with little or no direct damage - effectively disabled for weeks or longer, exactly when you mostly need dependable energy to make up for the loss of other plants.
So, the nuclear accidents don't show that nuclear power is particularly unsafe for a country like Japan (even taking into account the rather bad safety record of Japanese nuclear operators). It does show that it's not a good fit for the kind of place that can experience large-scale disaster damage. And since the economics of fission-based nuclear power is quite unfavourable (nobody would build a nuclear plant without heavy public subsidy) and likely to become worse over time, it would seem wise not to put a long-term bet on nuclear fission over other energy sources in the future.
Oh, and it shows that putting critical infrastructure in the hands of private operators without enforcing nationwide standards is a bad idea that's going to bite you somewhere further down the line.