Microseismicity data forecast rupture area D' Schorlemmer

1
Microseismicity data forecast rupture areaD.
Schorlemmer S. Wiemer (2005)
http//www.le0ne.com/photos/parkfield1_med.jpg
2
Background papers

• S. Wiemer M. Wyss (1997) Mapping the
  frequency-magnitude distribution in asperities
  An improved technique to calculate recurrence
  times?
• D. Schorlemmer, S. Wiemer M. Wyss (2004)
  Earthquake statistics at Parkfield 1.
  Stationarity of b value

3
b-values

• log10 N a - bM .
• N cumulative number of earthquakes of magnitude
  M or greater
• a earthquake productivity of a volume
• b relative size distribution of earthquakes

Wiemer Wyss (1997)
4
Calculating b-values

• Along a vertical cross-section of the fault, the
  seismicity is sampled using overlapping
  cylindrical volumes of constant radius
• Volumes have 1 km spacing

Wiemer Wyss (1997)
5
Schorlemmer, Wiemer Wyss (2004)
6
NCSN
HRSN
NCSN
Schorlemmer, Wiemer Wyss (2004)
7
What do low b-values indicate?

• b-values are inversely related to stress (in
  laboratory experiments)
• Changes in b (both temporal and spatial) could
  provide information about stress in the crust
  (Wiemer Wyss 1997 and Schorlemmer, Wiemer
  Wyss 2004)
• b-values decrease as a function of depth at
  Parkfield (Wiemer Wyss 1997)
• Lowest b-value coincides with the Parkfield
  asperity (below Middle Mountain)

8
Probabilistic recurrence time

• Tr ?T / (10a-bM)
• Gives the probabilistic recurrence time for an
  earthquake with magnitude greater than or equal
  to a chosen M.
• ?T length of recording period which produces
  the a-value

9
Annual probability for 1 events of M6
Constant b-value, varying a-values
Varying both a- and b-values
10
Past predictions

• Major earthquakes in Parkfield are most likely to
  nucleate in the Middle Mountain asperity, where
  b-values are lowest.
• Probabilistic recurrence time for the Middle
  Mountain asperity
• 23 (-12/18) years (WW, 1997)
• 30 years (S et al, 2004)
• These differ greatly from the 72 years and 116.6
  years calculated over the entire Parkfield volume
  (from both papers, respectively).

11
Parkfield Mainshock, 2004
e) Cumulative slip distribution as a function of
b-value. The areas of low b-value contain 99 of
the slip. Areas of high b-value prevent rupture
propagation.
a) Topographic map with mapped b-values
c) Preliminary slip distribution for the 2004
mainshock
b) Microseismicity b-values from 1981 to 2004
d) 2004 aftershocks. Symbol size is proportional
to magnitude
f) Cumulative distribution of aftershocks. Low
b-value areas contain 95 of aftershocks.
12
Main conclusion

• Evidence indicates that b-values could be used to
  predict rupture areas.

13
Although the timing of earthquakes remains
unpredictable, precise forecasting of the
location and size of events is becoming a
possibility.