Time-independent hazard-random process in time: used for building design, planning, insurance, probability

1
The Idea of Earthquake Prediction

• Time-independent hazard-random process in time
  used for building design, planning, insurance,
  probability
• Time-dependent hazard-a degree of predictability,
  enable authorities to prepare for an event-false
  alarms- loss of public confidence, China- shut
  down of public services during false alarms

2
Earthquake Predictions

• The 1960s- descriptions of physical changes
• 1970s- 1975, successful prediction 1976,
  250,000- people died in Tangshen, China
• 1980s- Jim Berkland, fluctuations of missing dogs
  in local newspapers, increase number of calls to
  satellite TV companies
• Greek scientists, electrical signals

3
Haichen, China, 1975

• An evacuation warning was issued the day before
• The local rural population was educated and aided
  in the monitoring
• changes in land elevation and ground water
• reports of peculiar animal behavior- snakes
  coming out of the ground in January, livestock
  not going into their barns
• an increase in small earthquakes

4
The Idea of Earthquake Prediction or Forecasting

• 1977- National Earthquake Hazards Reduction
  Program (NEHRP)
• 1. Precursor detection instruments at the site
• 2. Detect and recognize precursors
• 3. Get colleagues to agree, accept and put out a
  warning

5
Earthquake Forecasts

• Earthquake forecasting- some connection between
  the level of chance between observation and
  event probabilities and errors involved
• Scientists cannot predict earthquakes
• Evacuation before an event since 1975 has not
  occurred

6
Long-term Forecasts

• Retrofitting
• Building codes
• Legislation
• Insurance
• Emergency plans
• Education to the public
• Preparedness

7
Long-term Forecasting Methods
8
Paleoseismicity

• Identify fault trace
• Trench
• Analyze sedimentary layers and offsets
• Date organic material

9
Historical Seismicity Patterns

• Large Bay Area earthquakes are more frequent at
  some times more than others
• 1836-1911 1911-1979 (68 years of no events)

10
Recurrence Interval

• Statistical Method of historic earthquakes
• Shortest interval
• Longest interval
• Average 25.5 years
• Problems with this method?

• Parkfield Seismic Area
• 1988-90 probability for a M6 earthquake
• 1901
• 1922 21 years
• 1934 12 years
• 1966 32 years
• 2004 38 years

11
Paleoseismology and recurrence interval

• Shortest interval?
• Longest interval?
• Average?
• 187 years
• Strengths of this method
• Shortcomings?

• Pallet Creek offset
• 1857
• 1745 112 years
• 1470 275 years
• 1245 225 years
• 1190 55 years
• 965 225 years
• 865 100 years
• 545 320 years

12
Seismic Gaps -areas on a larger fault system that
lack seismicity
13
Variation of average movement on known faults

• Bay Area faults belong to the San Andreas fault
  system
• Movement varies
• Right-lateral strike-slip faults
• 1868- Hayward fault
• 1906-San Andreas fault
• 1989-San Andreas fault

14
Variation of average movement on known faults

• Bay Area faults move at different rates to
  accommodate the total amount of slip of the San
  Andreas fault system
• Total 34 mm/year
• SAF 19 mm/yr
• Hayward fault 9 mm/yr

15
Earthquake Prediction and Probability
Variation of average movement on known faults

• Each vector represents the amount of movement on
  each fault
• Variations may indicate an imminent earthquake

16
Pattern of Seismicity
17
Earthquake Forecasting and Probability

• Working Group on California Earthquake
  Probabilities
• Group of 100 geologists, engineers, government
  officials evaluated all information
• Came up with a probability of a magnitude 7 or
  greater in the next 30 years
• 1988. 1999, 2003, 2008

18
Earthquake Forecasting and Probability

• Assumes a random distribution of earthquakes
• In the Bay Area, 5 earthquakes greater or equal
  to M 6.75 occur approximately every 30 years
• 50 probability-just as likely to happen as not
  to happen

19
(No Transcript)
20
The Idea of Earthquake Forecasting

• Loma Prieta- 30 probability in 1988- eq, 1989
• Parkfield -90 probability in 1988-eq, 2004
• Landers, Northridge, Kobe- earthquake on a little
  known fault

21
Short-term Prediction based on precursors
22
All the described precursors sometimes occur but
as of yet do not occur in a manner that
successfully predict earthquakes
23
What should be expected?

• A reasonable time period
• The location- fault or fault segment
• Magnitude or amount of energy released

24
Ground Deformations

• Ground deformation
• Preceding the 1906 earthquake in Marin
  County-elastic rebound theory
• The Palmdale bulge- monitoring began in the
  1960s , associated with the SAF

25
Radon Gas Emissions

• Radon emission- a radioactive gas that sometimes
  shows an increase preceding an earthquake-
• There was a ten-fold increase 30 kilometers away
  from the epicenter 9 days preceding the
  earthquake.
• Problem with this method?

26
Parkfield Experiment

• Examining precursors or those events that
  sometimes occur before a main event
• Changes in physical properties in close proximity
  to the fault

27
Parkfield experiment

• Magnitude 6 earthquake occurred on average every
  22 years from 1857-2004
• Similar location
• Similar seismic wave pattern recorded

Hopes of documenting an earthquake before,
during, and after the main shock in order to
produce a short-term prediction technique.
28
Precursors expected to observe

• Foreshocks
• Ground deformation

Surface cracks associated with the 2004 earthquake
29
San Andreas Fault Observatory at Depth 2004
Examining physical changes to the rocks at depth
30
Animal Prediction

• Unusual animal behavior prior to a disaster
• Greece, 373 BC, rats, snakes and weasels deserted
  days before an earthquake
• Advanced vibrations?
• Change in electrical charges?
• Change in magnetic field?
• So far, unable to use unusual animal behavior to
  predict earthquakes

31
Chi Chi Earthquke,Taiwan, 1999

• Initial smaller earthquake during the first 4
  seconds have higher frequency waves than larger
  earthquake
• The warning could be sent
• Seismic waves travel at about 2 miles per second

32
Early Warning System

• Detect P-wave arrival
• Warning system
• Emergency centers, hospitals, railways
• Depends on distance to how helpful
• Japan-March , 2011, M9, Tokyo-stopped trains
  cell phone notification 8-10 minutes tsunami
• Kobe-fault under city
• Mexico-11 warnings, M6 or greater

33
Early Warning System U.S.

• 80 dollars over 5 years
• Arrival of P-waves
• 3-4 minutes Cascadia Subduction Zone
• 1 minute warning for SAF

• 30 seconds
• doctors stop surgery
• Duck and cover
• Trains stop