17 December 2010

AGU Day 4: Debris flows and the Station Fire [improved in light of comments from Susan Cannon Cannon]

Posted by Dave Petley

NB the text below has been updated to reflect some very helpful comments from Susan Cannon.  Genuine thanks to her for her help.

The AGU Fall meeting continues like an unstoppable rhino, with a dizzying amount of sessions and a rapidly developing sense that my brain has become full.  However, I will highlight here just two talks yesterday from the public communication of natural hazards session, both about the potential for debris flows from the area affected by the 2009 Station Fire in California.  Now, I have covered these debris flows before – see my post here – so I’ll not go into great detail about the events.

There were two talks – the first by Sue Cannon from the USGS, who talked about issuing warnings for the debris flows, and the second by Eric Boldt of the Weather service talking about the meteorology of the warnings.

So let’s start with the USGS focus.  The key issues here were that the fires burnt in August and September, and the wet season starts in the Fall, such that there was an urgent need to provide information to the authorities and to residents.  The San Gabriel mountains are an area of known debris flow risk (in 1934 about 50 people were killed by debris flows in this area).  Remarkably, the USGS were able to issue an open file report on 15th October highlighting the potential risk and providing hazard maps.  The report is online and free to access.

Sue spent most of her talk highlighting the successes and failures of the assessment exercise.  The “hits” were:
  • The information was released before it rained;
  • Most public response agencies were receptive to, and appreciative of, the information;
  • There was an extensive public information campaign, and they found that there was strong demand for the  information
On the other hand the “misses” were:
  • Public response was low, decreased with each storm , probably because they didn’t understand the dangers posed by debris flows;
  • There were inconsistent messages from different agencies;
  • There is a poor level of understanding of the uncertainties associated with both weather and debris flow predictions
  • Peoples’ desire to protect personal property during storm often meant huge resistance to relocating.
Sue’s take home message was that in toto, each person wants to make own decision on how to act based on their knowledge and their personal experience.  This leads to a very fragmented response.

Second up was Eric Boldt of the National Weather Service, who talked about the specifics of the warning system for the debris flows.  This was established in 2005/6 using  rain rate thresholds for burned areas determined by the USGS.  Areas that have been burnt in the last three years are included, with the threshold being associated with rain rate and the time since the fire (i.e. there is a lower threshold for the first year than for the second, when revegetation will have started to occur.  Sitting alongside the warnings was a high level of preparedness, including the use of sand bags, fire teams on standby and standing evacuation orders.
The second part of the talk focused on what actually happened:
  • On 12th November 2009 there was a comparatively small fire that generated some small debris flows;
  • On18th to 22nd  January 2010 rain rates of about 25 mm an hour, and totals of up to 50 mm, occurred in the  mountains.  [EDIT: this section provided in a comment by Sue Cannon] There were signficant debris flows in response to the Jan 18 storm from nearly every burned watershed. There was heavy damage in areas that were not protected by debris basins, and those debris basins that did exist were mostly filled to the brim.  With significant rainfall projected for the following day, teams were frantically (and without much luck because the material was still wet) trying to clear the basins as rapidly as possible. This was an extremely dangerous situation, particulalry since a ‘dodged the bullet’ attitude prevailed, rather than a recognition of potential consequences should the coming storm live up to forecasts. Luckily, the cells of high-intensity rainfall moved to the north and south of the burned area, and catastropie was again diverted.[end of edited section].  The speaker expressed the view that as a result the public might have  lowered their guard.
  • On 6th February a large storm rolled in. The weather service had identified the threat a week ahead and had ramped up the warnings provided.  The night before the storm made landfall they warned that 50 mm of rainfall might occur. At 12:21 am they issued a warning that thunderstorms were forming.  At 3 am they issued a warning of debris flows, and the Ocean Boulevard flow happened 45 mins later.  Fortunately no-one was killed, but there was a great deal of property damage.

The speaker finished by pondering the key messages:

  • There are problems when debris flow happens in the night.  It may be possible to use mobile phones, which are left on at night, to provide warnings at this time
  • The forecasts were not perfect because of local weather effects.
  • There is a need to work with social scientists to understand how people respond to long and short term warnings
  • More work is needed on whether people should evacuate or shelter in place.

One of the key messages of the meeting for me has been the huge progress that has been made in developing combinations of landslide hazard assessments and real time warning systems.  However, the weak link remains understanding how to communicate this information, and how people respond to it.  This is a key research frontier now.