The Evolution of Regional SeismicityBetween Large Earthquakes

David D. BowmanCalifornia State University, Fullerton

Geoffrey C. P. KingInstitut de Physique du Globe de Paris

Charles G. SammisUniversity of Southern California

All California Earthquakes M≥6.51950-1995

Slipping Fault Slipping FaultSeismic Slip Seismic Slip

Creep at Depth

FutureEarthquake

Calculate from motion on all adjacent faultsplus creep at depth

Stress Change From Loading a Locked Patch on a Simple Fault

FutureEarthquake

Creep at Depth

=

Negative Slip

Where are pre-earthquake stresses?

Accelerating Seismicity in Stress Accumulation Regions

slip 0 unitsslip 0 units slip 0 units

slip 0 units

slip 0.25 units slip 0 units

slip 0.25 units

slip 0.25 units

slip 0.5 unitsslip 0.5 units

slip 0.5 units

slip 0 unit

slip 0.5 units

slip 0.5 units

slip 0.5 unitsslip 1 unit

slip 1 unitslip 0.75 unit

slip 0.75 units

slip 0.75 units

slip 1 unit slip 1 unit slip 1 unit

slip 1 unit

EarthquakeStatic stress (Coulomb)

changes during the earthquake cycle

Stress FieldFrom Previous History of EQs

Stress FieldFrom Loading

CurrentStressField

= +

• Stress distant from the fault must approach,but not exceed, the failure stress

• Stress must be low along the strike of the fault(or rupture would continue)

• Coulomb field + background field can notexceed the failure stress (except locally)

• Stress distant from the fault must approach,but not exceed, the failure stress

• Stress must be low along the strike of the fault(or rupture would continue)

Coulomb field immediatelybefore the event

Coulomb field immediatelyafter the event

- Positive Coulomb stressesmust have a correspondingnegative value in thebackground field

• Coulomb field + background field can notexceed the failure stress (except locally)

• Stress distant from the fault must approach,but not exceed, the failure stress

• Stress must be low along the strike of the fault(or rupture would continue)

Background stress field• Stress must be low along the strike of the fault(or rupture would continue)

Characteristics of the Background Stress Field(for a simple model)

The start of theearthquake cycle

The end of theEarthquake cycle

Stress Change

Stress Level Relative to the Failure Stress

Tectonic Memory Stress

Eart

hquake

Eart

hquake

25

0

-25

bars

0

-50

bars

failurestress

Stress Change vs. Stress Through the Earthquake Cycle

An earthquake occurs when stress rises above the failure level

A BFault

Str

ess

Failure Stress

A BFault

Str

ess

Magnitude of the event depends on the size of the stress concentration

This allows the calculation of theFrequency-Magnitude relation

FailureStress

Creating a Synthetic Catalog

Immediately after the earthquake

25% of the cycle

50% of the cycle

75% of the cycle

Immediately before the earthquake

failure stress

-50 0bars

Stress and Seismicity Through the Seismic Cycle

fault

Aftershocks Stress shadows

Start of theEarthquake

Cycle

Approachingthe

Earthquake

Immediatelybefore theEarthquake

Immediatelyafter the

Earthquake

75% of theEarthquake

The end ofthe

Earthquake cycle

Earthquake

Seismicity in the Earthquake Cycle

• Accelerating Moment Release over a broad spatial region before large EQs

• a-value increase before a large event anddecreases after the event

• Evolution of the frequency-magnitude statistics (Gutenberg-Richter relation)

• Stationary (time-independent) b-value

• Off-fault aftershocks occur in regions of elevated static stress change due to the earthquake

• Main fault is seismically quiet for most of the seismic cycle

• “Mogi doughnuts”

Implications of Regional Stress Accumulation Model

• Region size scales with size of the “predicted” earthquake

California Seismicity 1912-2001 M>3.5

Earthquake

Late in theEarthquake cycle

Immediatelybefore theearthquake

Immediatelyafter the

earthquake

Build-up to the Earthquake

San Fernando EarthquakeModel

Cum

ula

tive

Be

niof

f S

trai

n

Time Time

Which cumulative moment release curve is fora REAL seismicity sequence?

Accelerating Seismicity

Earthquake

Late in theEarthquake cycle

Immediatelybefore theearthquake

Immediatelyafter the

earthquake

Build-up to the Earthquake

Evolution of Gutenberg-Richter Scaling Beforethe 1987 Superstition Hills Earthquake

126˚W

126˚W

124˚W

124˚W

122˚W

122˚W

120˚W

120˚W

46˚N 46˚N

48˚N 48˚N

50˚N 50˚N

Two large nearby events show accelerating moment release

The regions overlap, approximating the evolution of the seismicity over 2.5 cycles

Seismicity in the Pacific Northwest

Pacific Northwest Seismicity

Seismicity in the model & Pacific Northwest

Cumulative Benioff Strain

Evolution of the frequency-magnitude statistics

Evolution of the frequency-magnitude statistics

• Better calculation of the “noise” functions- incorporate stress transfer?

• More complex fault geometries- simulate real fault networks

• Additional Testing on real data- Test on earthquakes from other regions

(Greece, Turkey, China, etc)- False alarm rate?

• Relationship to Time-Dependent Hazard Analysis

Looking Forward:

Pre/re-prints available at:http://geology.fullerton.edu/faculty/dbowman