Sunday, September 12, 2010

San Bruno Pipeline Explosion

Fire consumes neighborhood in San Bruno, California, following rupture of a natural-gas pipeline. Photo by Paul Sakuma / Associated Press, courtesy Los Angeles Times.

Tragic news coming out of California as the death toll has now reached 7, with 6 people still missing, from the massive pipeline explosion and fire that consumed 37 homes in San Bruno, just south of San Francisco (map and photo gallery from Los Angeles Times). The 30" diameter high-pressure natural-gas pipeline, installed in 1956, ruptured shortly after 6pm local time on Thursday, September 9.

The explosion was so strong that USGS seismometers recorded it as a Magnitude-1.1 earthquake.

Map showing location of San Bruno explosion and fire. San Andreas Fault shown in purple. Speculative minor fault or fracture zone, interpreted by SkyTruth from Google Earth imagery, shown in yellow.

What caused this pipeline to rupture? Nobody knows yet, but this is of course a very seismically active region with notoriously unstable geology. The pipeline failure happened at the intersection of Earl Avenue and Glenview Drive, at the head of a steep ravine. Our analysis of the imagery in Google Earth (high-resolution aerial survey photography) suggests this ravine may follow the trend of a fracture zone in the bedrock that is roughly parallel to the San Andreas Fault, which lies just 1,000' to the west of the blast site. It is possible that stress was steadily and incrementally building on the rigid pipeline as strain accumulated across this minor fault or fracture zone, or as soil gradually slid down into the ravine. Consider this analysis to be preliminary and highly speculative.

UPDATE 9/13/10 @ 11am - Check out the Pipeline Safety Trust's web page for this disaster, and for more information on causes of pipeline failure, other incidents, and pipeline regulation and oversight.

3 comments:

  1. Thanks for investigating this. I don't know why the few news reports to comment on whether fault movement could have caused the pipeline rupture have dismissed the idea. It could be we are seeing accelerated fault creep prior to a major (mag. 7+) earthquake on one of the major faults nearby. If that's the case, we could see a major quake in the next week or two. I hope the USGS is watching this carefully. I wonder what the latest geodetic measurements across this region show?
    Michael R. Nelson
    Ph.D., seismology, MIT, 1988

    ReplyDelete
  2. Mike - thanks for the comment. The USGS has not reported any seismicity in the area except for the shock waves caused by the explosion itself. Accumulating fault strain might also cause cracking/buckling of roads and sidewalks, and cracking in building foundations, but we can't see that in the Google imagery.

    Rather than a precursor to a big quake, this could be gradual and incremental aseismic slip along a minor fault or fracture zone, or good old-fashioned slope-creep with soil moving into the ravine. Either scenario might slowly increase stress on the pipeline until it reached a failure threshold and suddenly ruptured.

    But I am most definitely speculating here. It could have been something else: simple failure due to metal fatigue in a 54-year-old steel pipe; failure caused by mechanical damage to the pipeline sometime during those 54 years (getting scraped by a backhoe excavating the road to install or repair other underground utiltites, for example).

    For more info on pipelines and what makes them fail, check out the Pipeline Safety Trust's web page on the San Bruno disaster.

    ReplyDelete
  3. Something as simple as corrosion has triggered many pipeline failures over the years. It still causes failures.

    Some deadly pipeline failures from corrosion have been looked at by the NTSB:

    http://www.ntsb.gov/publictn/1998/PAR9802S.pdf

    http://www.ntsb.gov/publictn/2003/PAR0301.pdf

    Here in the Bay Area, slow fault creep goes on along parts of the Hayward & Calaveras fault zones. But, I heard nothing about any creep on the San Andreas north of Watsonville.

    ReplyDelete