Tuesday, March 15, 2011

Nuclear Earthquakes and Paleoseismology

2 PM Tuesday, Ides of March

It’s snowing out, which normally would make me happy, but now makes me even more concerned for refugees in northern Honshu. And… it makes me think of nuclear fallout. …not going to happen, really. I just watched a video that, while acknowledging the recent 3rd explosion and fire at #4 and leakage, made clear it’s not and won’t be Chernobyl. I am not in danger. But the stress levels for the country are so high and now higher.

Our Japanese hosts just decided to call off our workshop 28-30 March:

Dear Ben! Unfortunately the disturbed environmental condition by nuclear reactors accident at Fukushima has been becoming worse day by day. So we with Nakagawa-san, and Tezuka-san have to make decision calling off the workshop of end of this March. Please announce this decision to all participants quickly. Best regards, Amano T.

“the worst nuclear disaster except Chernobyl,” I just heard on a video news clip.

…. and on another clip [paraphrased] “What happened there can’t happen in the U.S. because we upgraded our systems since 9/11…(but then acknowledging that) the Japanese would have said much the same before this situation.”

I remember hearing that Japanese engineers came to Northridge, CA, after the heavily damaging earthquake there on 17 January 1994, to study the damage and then stating that such damage would not happen in Japan because of stronger earthquake engineering. Then Kobe happened …17 January 1995. The story as I tell it is second hand, but the statement “it can’t happen here” feels like a red flag to me. That said, lots and lots of buildings here withstood this Mw 9 earthquake.

So—this is why, as I’ve been saying to friends since Friday, nuclear regulatory commissions ask for “worst-case scenarios” from geoscientists and others, whereas community planning is typically for a typical flood or a 100-year event or something like that.

How does we generate “worst-case scenarios” for earthquakes and tsunamis? Giant earthquakes and tsunamis don’t happen that often, in any one place, commonly on the order of every 500 years, or more. And with events that happen rarely (extreme example—asteroid impact), we may not even know (or believe or comprehend) they can happen.

I have read an expert interviewed who said that these power plants in Japan were designed for a magnitude 8.2 earthquake (and presumably tsunami!). That’s probably because in recent history, which is long in Japan, 8.2 was as big or bigger than known to happen on this coast (see table).

Here are “selected earthquakes of general historic interest in Japan, in the instrumental historical record (USGS list) (I think there are likely more large events in the earlier record, but unless they had fatalities, they didn’t necessarily get onto a list) (underlines mean there is more information on the USGS NEIC website):

1891-10-27: Mino-Owari, Japan – M 8.0 Fatalities 7,273

1896-06-15: Sanriku, Japan – M 8.5 Fatalities 27,000 –a famous “tsunami earthquake” – deaths almost entirely from tsunami

1923-09-01: Kanto (Kwanto), Japan – M 7.9 Fatalities 143,000 – this is the most recent one that hammered Tokyo

1927-03-07: Tango, Japan – M 7.6 Fatalities 3,020

1933-03-02: Sanriku, Japan – M 8.4 Fatalities 2,990 – an outer-rise event

1943-09-10: Tottori, Japan – M 7.4 Fatalities 1,190

1944-12-07: Tonankai, Japan – M 8.1 Fatalities 1,223

1945-01-12: Mikawa, Japan – M 7.1 Fatalities 1,961

1946-12-20: Nankaido, Japan – M 8.1 Fatalities 1,330

1948-06-28: Fukui, Japan – M 7.3 Fatalities 3,769

1964-06-16: Niigata, Japan – M 7.5 Fatalities 26

1995-01-16: Kobe, Japan – M 6.9 Fatalities 5,502 – the Kobe earthquake was a shallow crustal event, not so large, but very damaging

2003-05-26: Near the East Coast of Honshu, Japan – M 7.0

2003-09-25: Hokkaido, Japan Region – M 8.3

2003-10-31: Off the East Coast of Honshu, Japan – M 7.0

2004-09-05: Near the South Coast of Western Honshu, Japan – M 7.2

2004-09-05: Near the South Coast of Honshu, Japan – M 7.4

2004-10-23: Near the West Coast of Honshu, Japan – M 6.6 Fatalities 40

2004-11-28: Hokkaido, Japan Region – M 7.0

2005-08-16: Near the East Coast of Honshu, Japan – M 7.2

2005-11-14: Off the East Coast of Honshu, Japan – M 7.0

2009-08-09: Near the South Coast of Honshu, Japan – M 7.1

I added: 2010-03-09: Off the Northeast coast of Honshu, Japan – M 7.2

[now add bunches more, mainshock and aftershocks, since 11 March 2011]

Systems like reactors and bridges are overdesigned to withstand forces greater than the maximum projected earthquake. Moreover, it looks like the earthquake engineering in this case worked ok, though I am not sure about cracks, etc. But there’s more than the earthquake! Here’s some of a scenario, real and envisioned, for an event like this, a scenario that I would HOPE had been considered for a coastal reactor site along Japan’s subduction zone before they were built:

  • The reactors shut down automatically during the quake.
  • But then they have to be cooled down.
  • All around the plant is destruction from the earthquake and especially the tsunami – roads damaged, access impaired, people injured,… on and on.
  • Power is out – ok, they have backup generators to pump water.
  • But generators are flooded by the tsunami and do not work – ok, they have a battery, but of quite short life. Not so easy to get back-ups.
  • Water supply lines broken --pump in sea water (ends working life of plant)
  • Cooling can’t take place fast enough, steam and other gases building up, etc. etc., explosions and fires, minor radiation leaks.
  • Partial meltdown, more serious radiation leakage, at least one crack in containment; other radiation sources (spent fuel, wastewater pools) are sources of radiation.
  • Meltdown, which in the best case at THIS point is contained.

I am quite sure that a scenario something like this was written. The situation now is not due to bad decisions at the present time, but probably limited information or analysis about “worst case scenario” more than 40 years ago or so when they were planning. Fukushima 1 is one of the oldest operating reactors, and, when you think about it, we were only beginning to understand plate tectonics and subduction zones around the time it would have been planned!



Net capacity


Commercial Operation

Fukushima I-1


439 MWe


March 1971

Fukushima I-2


760 MWe


July 1974

Fukushima I-3


760 MWe


March 1976

Fukushima I-4


760 MWe


October 1978

Source: World Nuclear Association

What about the historic record? When was the last magnitude 9 (or so) event in this area? Well, it’s hard to put even moderately specific magnitudes on events before instrumentation. Even now, scientists will disagree about magnitudes of earthquakes as recent as… all of them! Usually the disagreement are ±0.1 or so, but sometimes more, and for events in the earlier 20th century, many events are still not well quantified and perhaps can’t be. But as of Friday night, I heard my colleagues referring to the 869 A.D. Jogan earthquake and tsunami as the precursory event of this scale – approximately, of course, but the tsunami reached kilometers inland, unlike any more recent case till 11 March 2011.

Minoura, K., F. Imamura, D. Sugawara, Y. Kono, and T. Iwashita, 2001The 869 Jogan tsunami deposit and recurrence interval of large-scale tsunami on the Pacific coast of northeast Japan, Journal of Natural Disaster Science, v.23, no.2, 83-88.

Satake, K.; Sawai, Y.; Shishikura, M.; Okamura, Y.; Namegaya, Y.; Yamaki, S., 2007. Tsunami source of the unusual AD 869 earthquake off Miyagi, Japan, inferred from tsunami deposits and numerical simulation of inundation American Geophysical Union, Fall Meeting abstract #T31G-03

See Wikipedia article for additional references in the Japanese language http://en.wikipedia.org/wiki/869_Sanriku_earthquake_and_tsunami

Even though 869 A.D. is a historic event in Japan, understanding its scale comes partly from paleoseismology –studying the record from evidence such as tsunami deposits and liquefaction–evidence in the latter case of strong shaking. Minoura et al. (2001) estimated by tsunami modeling that the Jogan earthquake was a Mw (moment magnitude) 8.3. Satake et al. (2007) also modeled the tsunami and came up with an estimate of 8.1-8.3. They also predicted that such an earthquake was almost sure to happen (99% probability) in the next 30 years. Less than four years later (AGU meetings are in December), it happened, only it was more than 10 times bigger than that.

In Wikipedia, the article uses the term “869 Sanriku” because Jogan is the name of an era, not a place. Sanriku, however, is the northern end of this zone, so… so it goes. At least one of the Wikipedia authors is Japanese and cites the Japanese-language literature, also giving a “surface wave magnitude” to the reconstructed event of 8.6, but this magnitude saturates at about 8…. I haven’t tracked down the source of that number or the likely error in labeling it surface-wave magnitude.

Paleoseismology as a field of research basically has its roots in the 1980s, and these plants were designed in the 1960s and built in the 1970s.

Another nuclear power plant is involved in the story as well as in the development of paleoseismology.

http://pubs.usgs.gov/pp/pp1707/ -- The orphan tsunami (Atwater et al.)


It’s a long story, but here’s a short version of how the Cascadia paleoseimological story got started: In designing a nuclear power facility for SW Washington State (the Satsop plant, in Elma, WA, not directly on the coast but some distance inland), the US Geological Survey got involved in evaluating seismic potential of the Cascadia subduction zone (which some even questioned WAS a subduction zone, but that’s ANOTHER story). Was it aseismic, as some interpreted, or could it produce big earthquakes, even though there was no historical record of one? Seismologist Tom Heaton of the USGS evaluated the potential by comparing CSZ to other subduction zones and thought it could. Seismologist Lynn Sykes of Columbia University compared it to other subduction zones and thought it wouldn’t. USGS geologist Brian Atwater asked the question—given that the historic record is so short, what kind of evidence might be in the marshes of coastal Washington State? Answer—big earthquakes cause coastal subsidence, so drowned marshes and forests would be a big clue. Not only did he find such drowned vegetation, but he found a sand layer he thought was a tsunami deposit. This was one of the important early paleoseismological studies. Many people worked on many sites, and finally even a historical record of the tsunami was found in the Japanese literature.

But the WPPSS Satsop plant construction in 1983 was shut down for financial, not for seismic-risk, reasons. And the studies of paleoseismology actually were published after that. Now the location is a park.

Added 19 March: Muckraking piece on Japan's nuclear power policies:


No comments:

Post a Comment