Christchurch Earthquake Christchurch Earthquake New Normal or Old - - PowerPoint PPT Presentation

Christchurch Earthquake Christchurch Earthquake ‐ New Normal or Old Normal, and Implications for Policy

Professor Paul Somerville Chief Geoscientist Risk Frontiers, Macquarie University

Outline Outline

• The New Normal –
• The New Normal

– Greater earthquake source strength? More frequent Canterbury earthquakes? – More frequent Canterbury earthquakes?

• Evidence about source strength from recorded

d ti ground motions

• Uncertainty about more frequent earthquakes
• Implications for policy

Definitions of an Earthquake q

• Engineer (and everyone else):

Engineer (and everyone else): “a shaking of the ground” G i ti t

• Geoscientist:

“a sudden movement on a fault”

• In this talk, Earthquake means a sudden

movement on a fault, which causes ground motions and other effects

Christchurch ‐ The New Normal? Christchurch The New Normal?

• Greater source strength ‐ Are the source
• Greater source strength ‐ Are the source

strengths (stress drops) of some categories

• f New Zealand earthquakes larger than we
• f New Zealand earthquakes larger than we

had thought?

• More frequent Canterbury earthquakes
• More frequent Canterbury earthquakes –

Are large earthquakes in the Canterbury Plain going to be much more frequent than Plain going to be much more frequent than before for decades to come?

Tectonic Setting and Seismic Hazards

Peak acceleration with 475 year ARP GNS Science Risk Frontiers

Faults and Tectonics beneath Wellington – the Hikurangi subduction zone

GNS Science

1995 Mw 6 9 Kobe Earthquake 1995 Mw 6.9 Kobe Earthquake

Improving Building Performance – Kobe Earthquake damage statistics Kobe Earthquake damage statistics

Reinforced Concrete Steel Reinforced Concrete Steel

AIJ

Building code changes in 1971 and 1982 were very effective

We Have Only Seen a Few of All the P ibl E h k i N Z l d Possible Earthquakes in New Zealand

Earthquake recurrence intervals are hundreds Earthquake recurrence intervals are hundreds

• f years to tens of thousands of years

Japan

• Earthquakes expected in the

i l k

New Zealand

• Earthquakes expected in the

it l W lli t capital – Tokyo

• Earthquake happened in Kobe

in 1995 capital – Wellington

• Earthquakes happened in

Christchurch in 2010‐11

• Previous earthquake occurred

in Kobe in 1596 G ll d f

• No previous surface faulting

Canterbury events in 15kyr G ll d f

• Generally good performance
• f new buildings in 1995
• Generally good performance
• f new buildings in 2010‐11

except for soil failure

Canterbury Plain Earthquake Sequence Canterbury Plain Earthquake Sequence

GNS Science

Canterbury Earthquake Sequence

• The earthquakes occurred on previously unidentified faults

that probably have not ruptured in the past 15,000 years h f h k h b ll l

• The aftershock sequence has been unusually long, consisting

progressive eastward propagation of seismic activity

• The 4 Sept 2010 Mw 7 1 Darfield earthquake produced
• The 4 Sept 2010 Mw 7.1 Darfield earthquake produced

expected levels of ground motions in Christchurch for that magnitude and distance, corresponding approximately to g , p g pp y 1/475 building code levels

• The 22 Feb 2011 Mw 6.1 Christchurch earthquake produced

ground motion levels in Christchurch much larger than expected, for reasons that relate to known seismic source and i ff Th l l d l propagation effects. These levels correspond to an annual probability of exceedance of about 1/2,500

ShakeMaps – Darfield & Christchurch ShakeMaps Darfield & Christchurch

USGS USGS

Ratio of Christchurch to Darfield Peak Acceleration

USGS

Coincidence of Ground Motion Intensity and Building Density – Christchurch Event

ShakeMap Building Density

USGS

USGS

Ground Motion Model Elements: Source, Path and Site

(GMPE) (PHYSICS‐BASED SIMULATION)

Intra- Event Ground Motion Variability

2004 Niigata Chuetsu Earthquake

H i ll F t ll Hanging wall Foot wall Hiroe Miyake

Inter‐Event and Intra‐Event Variability

Al Atik et al 2014 Al Atik et al., 2014

Did the Christchurch Earthquake have a Higher Source Strength? have a Higher Source Strength?

h h ( d ) ld

• Higher source strength (stress drop) would

produce a high inter‐event term

• Find out by comparing the recorded ground

motions with the predictions of a ground motion prediction model

• Use the pre‐Canterbury earthquake ground

p y q g motion prediction model of Bradley (2010)

Ground Motion vs Closest Distance Ground Motion vs Closest Distance

Darfield Christchurch

About the same as or a bit lower than the model Higher than model, < 10 km Lower than model, > 10 km

Christchurch Ground Motions Not Consistently High at All Distances

Inter and Intra Event Variability Feb 22 Christchurch Event

Al Atik et al., 2014

Brendon Bradley

Did the Christchurch earthquake have a Higher Source Strength? No.

• The event term is a source parameter, not a path
• r site parameter
• If the Christchurch event had a high event term,

we would expect its ground motions to be large at all distances, which was not the case

• This suggests that factors other than source, i.e.

ti th d it d th propagation path and site response, caused the unexpectedly large ground motions in the CBD

CBD Strong Motion Recording Sites

Averaged CBD Response Spectra Averaged CBD Response Spectra

Royal Commission

Mw 7 1 Sept 4 2010 Darfield Earthquake Mw 7.1 Sept 4, 2010 Darfield Earthquake

Ground Motion vs CBD Response Spectra Ground Motion vs Closest Distance CBD Response Spectra

Brendon Bradley 1 sec Spectral Acc vs R

Mw 6 2 Feb 22 2011 Christchurch Event Mw 6.2 Feb 22, 2011 Christchurch Event

Ground Motion vs Closest Distance CBD Response Spectra

Brendon Bradley 1 sec Spectral Acc vs R

Christchurch Ground Motions: due to Higher Source Strength? due to Higher Source Strength?

• The Mw 6 2 22 Feb 2011 Christchurch
• The Mw 6.2 22 Feb 2011 Christchurch

earthquake ground motions were unusually high within 10 km but at ordinary levels g y beyond 10 km, so cannot be attributed to high source strength (stress drop)

• Local conditions in Christchurch may have

increased the ground motion levels:

– Source: Rupture directivity effects – Path: Basin resonance effects Sit S il lifi ti ff t – Site: Soil amplification effects

Large Near‐Fault Directivity Pulses Recorded in both Darfield and Christchurch Events

Brendon Bradley

Directivity pulse recorded at Lyttelton in the Darfield Earthquake

Brendon Bradley

y y

• The directivity pulse is a shock wave analogous to

sonic boom sonic boom

• It only occurs close to the fault and is different from

source strength (stress drop) which affects ground ti t ll di t motions at all distances

Ch i h h l d S di B i Christchurch located on Sedimentary Basin

Brendon Bradley

Basin: Estuarine Sediments Overly the Lyttelton Volcano

Lawton et al. 2012

Trapping of Waves in Basins

Fl L B i Ed Flat Layers Basin Edge

Robert Robert Graves

Basin Waves – Christchurch Earthquake Basin Waves – Christchurch Earthquake

Schematic Geology Recorded Waveforms Schematic Geology Recorded Waveforms

Christchurch – on Quaternary basin Lyttelton – on volcanic rock Brendon Bradley

Lyttelton, on bedrock, has just the directivity pulse. ChCh also has basin waves

Basin Waves – Darfield Earthquake Basin Waves Darfield Earthquake

Christchurch - CHHC Lyttelton - LPCC Lyttelton, on bedrock, has just the directivity pulse. ChCh also has basin waves

Focusing of Energy by Rupture Directivity and Basin Effects

Puente Puente Hills Blind Thrust Thrust

Los Angeles - Directivity Long Beach – Basin Effect

Robert Robert Graves

Soft Shallow Soils Soft Shallow Soils

Tonkin & Taylor

Difference between Rock and Soil Response Spectra ‐ Lyttelton

The ground motion that enters the soil from below is presumably similar to that recorded on the adjacent rock site

Brendon Bradley

similar to that recorded on the adjacent rock site

Conclusions: Christchurch Ground Motions

• Local conditions in Christchurch may

h i d h d i have increased the ground motion levels:

– Source: Rupture directivity effects – Path: Basin resonance effects – Site: Soil amplification effects

Higher Source Strength ‐ The New Normal?

• Earthquake source strength ‐ Are source strengths
• f some categories of New Zealand earthquakes

uniformly higher than we had thought?

– No Thi l i i l ll f N Z l d – This conclusion is relevant to all of New Zealand – But we need to fully understand the conditions that caused the locally high ground motions in the Christchurch y g g earthquake and then assess where else in New Zealand such conditions may exist

More Frequent Canterbury Earthquakes? q y q

• Seismic hazard analysis usually assumes spatial

and temporal randomness in earthquake

• ccurrence

d h h h k

• Evidence shows that earthquakes occur in

spatial and temporal clusters GNS S i li d t l d ti l

• GNS Science applied temporal and spatial

clustering to time‐varying seismic hazard analysis for Christchurch – the first such analysis for Christchurch the first such application worldwide

GNS Time Dependent Hazard Model – p Christchurch (Gerstenberger et al., 2012)

• An increased rate of earthquakes is expected to

last for decades and far exceed the rate of h k h earthquake occurrence in the previous 170 years (by a factor of about 20) 1/475 k l ti i d f

• 1/475 year peak acceleration increased from

0.22g to 0.65g, higher than Wellington (0.4g)

• 1/475 building code peak acceleration increased
• 1/475 building code peak acceleration increased

from 0.22g to 0.35g; a compromise

G S d ildi C d S GNS and Building Code Spectra ‐ Christchurch

Gerstenberger et al., 2012

Alternative Seismic Source Models

• Gerstenberger et al. (2012)

g ( )

– Time‐varying; includes short term (aftershocks) and long term earthquake clustering

• Bradley (2015)

– Time‐varying; includes aftershock model

GNS (2014)

• GNS (2014)

– Time‐independent GNS model with added aftershocks derived from Gerstenberger et al (2012) g ( )

• Calculations using these alternative models (next

slide) produce very different results; this issue remains unresolved

Response Spectra for 475 year ARP Response Spectra for 475 year ARP Calculated from Alternative Models

Which model to use remains unresolved

More Frequent Canterbury Events q y ‐ The New Normal?

• More Frequent Canterbury Earthquakes – Are large

earthquakes in the Canterbury Plain going to be much more frequent than before for decades to come?

D ’ k – Don’t know – Has immediate implications for Canterbury Plain – Rather than simply apply a statistical earthquake forecast Rather than simply apply a statistical earthquake forecast model it would be preferable to seek physical evidence for changes (e.g. in stress level and orientation) and try to identify where earthquakes might occur identify where earthquakes might occur

Effect of More Frequent Canterbury E th k Earthquakes

• More frequent earthquakes increase the hazard level

by causing a higher probability of experiencing larger and closer earthquakes, and more severe ground motions (above the median ground motion level) for motions (above the median ground motion level) for that magnitude and distance

• Considerations related to existing buildings:*
• Considerations related to existing buildings:

– Building codes do not require design for the strongest possible ground motions – Well engineered buildings have the capacity to withstand ground motions beyond the design level ground motions – this was demonstrated in the Christchurch earthquake this was demonstrated in the Christchurch earthquake

*From this point I am expressing opinions about engineering; I am not an engineer

Goal of Earthquake Engineering

Maximise utility defined as maximising Maximise utility, defined as maximising total benefit, human‐centered on a moral f d b b l foundation, by balancing:

– Demand vs. Capacity p y – Cost vs. Benefit

Earthquake as Base Shear Demand

Joe’s

Beer! Beer! Food! Food!

W ZICS V 

2001 PEER Annual Meeting

R

Joe’s Bar and Grill courtesy of Ron Hamburger

Capacity of Building to Incur Drift (Lateral Displacement of Roof) due to Demand

Base Shear Demand

Joe’s

Beer! Beer! Food! Food!

Very rare events (2%/50yrs)

Demand

Beer! Beer! Food! Food!

Rare events (10%/50yrs) (2%/50yrs)

Structurally Operational

Occasional events (20%/50yrs)

y Stable Life Safe

Frequent events (50%/50yrs)

2001 PEER Annual Meeting

Lateral Deformation

PEER

Performance‐Based Earthquake Engineering Performance Based Earthquake Engineering (PBEE) – Probability Framework Equation

  

 ) ( | | | IM d IM EDP dG EDP DM dG DM DV G DV v 

Performance (Loss) Models and Simulation Hazard Impact

• DV - Decision Variable ($ loss, downtime, life-safety)
• DM - Damage Measure (condition, necessary repairs,… )

EDP E i i D d P t

(drift acceleration )

• EDP – Engineering Demand Parameter (drift, acceleration, ...)
• IM - Intensity Measure (Sa, Sv, duration …)

PEER PEER

Earthquake Source Strength – Implications for Policy

• Existing ground motion prediction models are

validated

• Look for conditions in other urban areas that

may resemble those that amplified the Ch i t h h d ti Christchurch ground motions

• Continue orderly measures to reduce building

vulnerability based on Christchurch data vulnerability based on Christchurch data, current knowledge and capabilities

Earthquake Source Strength – q g Implications for Policy

• The 22 Feb 2011 Christchurch earthquake was

quite small magnitude 6 2 quite small – magnitude 6.2

• Earthquakes this small typically do not break the

ground surface so the faults on which they could ground surface, so the faults on which they could potentially occur may be difficult to identify

• Focus on identifying such faults in urban regions

Focus on identifying such faults in urban regions

More Frequent Canterbury Earthquakes More Frequent Canterbury Earthquakes – Implications for Policy

• Building code officials need to clearly

understand the basis of time varying hazard and understand the basis of time‐varying hazard and consider whether its use is suitable for their purposes p p

• Abrupt large changes in seismic hazard level are

difficult to reconcile with desired continuity in building codes and with existing building stock

Residual Capacity – Policy Implications p y y p

• The Christchurch earthquake has shown again how

q g difficult it can be to relate ground shaking level to damage, and damage to residual capacity

• Focus research on assessment of residual capacity
• f damaged buildings to meet requirements for:

– Safety tagging and safety assessment – Insurance loss assessment – Decision to repair or demolish – Code mandated repair – Seismic certification of buildings

Damage Control – Policy Implications g y p

• The Christchurch earthquake has shown that

building codes aimed primarily at life safety have building codes aimed primarily at life safety have been largely successful in that goal, but have not been effective at preventing losses

• Focus research on design and retrofit innovations

aimed at reducing losses as well as enhancing life safety: – Performance based design: the owner specifies t bl d t t f i l l f acceptable damage states for various levels of annual probability Protective systems (base isolation) – Protective systems (base isolation) – Self‐centering structural systems

Cook Strait and Lake Grassmere Coo St a t a d a e G ass e e Earthquake Sequences, 2013

Earthquake Sequences Grassmere Ground Motions

GNS Science GNS Science

Fourth Announcement

10th Pacific Conference on

Earthquake Engineering

{& Annual Meeting of the AEES} BUILDING AN EARTHQUAKE‐RESILIENT PACIFIC Sydney, Australia 6‐8 November 2015

New Zealand Society for Earthquake Engineering Australian Earthquake Engineering Society Conference website aees.org.au/10pcee