Ea Ph rthqu hilade uake Augu elphia (C) S ust 6, a, Pen Study 2017

2017 Summer S r Nation nal Meetting

Earthqu uake (C) S Studyy Grooup

ust 6, 2017 Augu Ph hiladeelphiaa, Pennnsylvvaniaa © 2017 Nationaal Association of Insurance Commissioners C s

© 2017 National Association of Insurance Commissioners

Date: 7/20/17 2017 Summer National Meeting Philadelphia, Pennsylvania

EARTHQUAKE (C) STUDY GROUP Sunday, August 6, 2017 11:00 a.m. – 12:00 p.m. Philadelphia Marriott—Grand Ballroom AB—Level 5

ROLL CALL Ron Dahlquist, Chair Emil R. Mackey Russ Galbraith/William Lacy Ken Allen Wanchin Chou Brett Gerger Angela Nelson

California Alaska Arkansas California Connecticut Illinois Missouri

Gennady Stolyarov John D. Doak David Dahl Raymond G. Farmer Mike Humphreys Tracy L. Klausmeier Stacy Middleton

Nevada Oklahoma Oregon South Carolina Tennessee Utah Washington

NAIC Staff Support: Anne Obersteadt AGENDA 1.

Consider Adoption of its Spring National Meeting Minutes—Ron Dahlquist (CA)

Attachment A

2.

Hear a Presentation, “RMS North America Earthquake Model Overview” —Matt Nielsen (Risk Management Solutions—RMS)

Attachment B

3.

Discuss Any Other Matters Brought Before the Study Group—Ron Dahlquist (CA)

4.

Adjournment

W:\National Meetings\2017\Summer\Agenda\Earthquake SG Agenda.docx


1

Attachment A Consider Adoption of Spring National Meeting Minutes


Draft Pending Adoption Attachment Six Property and Casualty Insurance (C) Committee 4/10/17 Draft: 4/18/17 Earthquake (C) Study Group Denver, Colorado April 8, 2017 The Earthquake (C) Study Group of the Property and Casualty Insurance (C) Committee met in Denver, Colorado April 8, 2017. The following Study Group members participated: Ron Dahlquist, Chair, Ken Allen and Rachel Hemphill (CA); Michael Ricker (AK); William Lacy (AR); Wanchin Chou (CT); Brett Gerger (IL); Angela Nelson (MO); Rajat Jain (NV); Tyler Laughlin (OK); Will Davis (SC); Brian Hoffmeister (TN); and Doug Hartz (WA). Also participating was: Charles Angell (AL). 1.

Adopted its 2016 Fall National Meeting Minutes

Ms. Nelson made a motion, seconded by Mr. Gerger, to adopt the Working Group’s Dec. 10, 2016, minutes (see NAIC Proceedings – Fall 2016, Property and Casualty Insurance (C) Committee, Attachment Six). The motion passed. 2.

Heard a Presentation from AIR Worldwide on the AIR U.S. Earthquake Model

Claire Pontbriand, Ph.D. (AIR Worldwide) discussed updates and enhancements to the AIR U.S. Earthquake Model, which will be released this summer. The model incorporates new information for regional tectonic hazards and the influences of human behavior in inducing seismicity in the central U.S. The model also updated the shake, fire-following and liquefaction components of the model and added two new sub-perils: tsunami and landslide. Dr. Pontbriand stated that limited and incomplete data present a challenge in seismic hazard modeling. In addition, hidden fault sources make it hard to monitor fault dynamic states. The updated 2014 U.S. Geological Survey (USGS) National Seismic Hazard Maps (NSHMP), 2016 USGS One-Year Seismic Hazard Forecast and updated 2015 Uniform California Earthquake Rupture Forecast (UCERF3) are the primary drivers in the model updates. Updates to the Stochastic Catalog and Intensity Module reflect mixed changes in the USGS NSHMP update. AIR’s updated forecast relies more on physical modeling instead of expert opinion. The greatest magnitude changes in the modeled seismic risk occur in California, with significant, but lesser-degree, changes in the Pacific Northwest. For California, new data and a multi-fault rupture view have resulted in modeled scenarios showing a decrease in earthquake magnitudes, excluding tail risk of magnitude 8 or greater earthquakes, which increased. Updates to the Cascadia Subduction Zone (CSZ) for new paleo records and previously unknown faults have resulted in an increase in hazard in Northern California. The rupture geometry, magnitudes and return periods of events for the New Madrid Hazard Zone were updated using USGS data and published research. The addition of rare large magnitude earthquake zones and catalog updates tends to increase the hazard for the New Madrid Seismic Zone. The updated model integrates the latest data and scientific opinion put forth in the USGS 2016 One-Year Seismic Hazard Forecast with research conducted by AIR scientists to address additional uncertainties beyond the USGS formulation, and thus provides a robust view of induced seismicity risk. The induced seismicity catalog can either be merged with the standard earthquake catalog or not; it is easily turned on and off. The updated model also features tsunami as a new sub-peril for the West Coast and the CSZ. Tsunami risk was added as new research indicates the potential for increased frequency of large earthquakes in this zone. The tsunami module was developed by enhancing the TUNAMI model—a numerical model that incorporates geographic data to determine the physical characteristics and propagation of a tsunami—to capture the complex nature of a tsunami’s development and behavior as it reaches land; fluctuations of astronomical tides are also considered. The newly added landslide module incorporates Digital Elevation Model (DEM), geological and precipitation data to capture the damaging effects of earthquake-triggered landslides for the contiguous U.S. Mr. Dahlquist stated he is concerned with the potential for waste water injection to induce a larger earthquake many years into the future and asked what the modeled frequency of earthquakes at the state level were for Oklahoma. Dr. Pontbriand stated AIR takes a longer-term view and uses different declustering methods because there are so many uncertainties with induced seismicity. She stated she would need to look into the frequency of earthquakes at the state level for Oklahoma.


1

Draft Pending Adoption Attachment Six Property and Casualty Insurance (C) Committee 4/10/17 Mr. Ricker asked how AIR model updates applied to Alaska. Dr. Pontbriand stated the updated model showed an increase in tsunami risk for Alaska, but did not include other updates as the USGS updates did not include this state or Hawaii. Annalise Mannix (Fair Insurance Rates in Monroe—FIRM) asked how damage done to dams and levees was incorporated in the model. Dr. Pontbriand stated that these were covered under the Flood model. Mr. Laughlin asked how the model would reflect Oklahoma’s recent limits on waste water injection. Dr. Pontbriand explained that it is hard to say which well is influencing induced seismicity and, due to the potential for stress transfer, reframing from injecting waste water at the surface does not necessarily have a corollary impact on induced earthquakes. However, the USGS comes out with one-year models, which allows AIR to easily update its modeled treatment of induced seismicity. Charles Burhan (Liberty Mutual) asked how aquifers were being treated in the model. Dr. Pontbriand stated that, according to the USGS, whether the acquifer is confined or unconfined will influence the degree of water-level fluctuations in the wells in response to seismic waves. Some acquifers may act as resonators, which may amplify the response. She also stated the model does not have a participation or climate component included for induced seismicity. Having no further business, the Earthquake (C) Study Group adjourned. W:\National Meetings\2017\Spring\Cmte\C\Earthquake\04-EarthquakeSGmin.docx


2

Attachments B Hear a Presentation, “RMS North America Earthquake Model Overview” Matt Nielsen (Risk Management Solutions—RMS)


8/2/2017

RMS VERSION 17.0 NORTH AMERICA EARTHQUAKE MODEL UPDATE Ashley Bernero NAEQ Product Manager August 6th, 2017

NAEQV17 PROJECT TEAM, COLLABORATORS & REVIEWERS • 100 person years of research and development • RMS Publications: 40+ EQ conference and peer reviewed journal publications, in 2014 (an average year for publications) • USGS and NGA-West 2 project involvement It takes a village..

Copyright © 2017 Risk Management Solutions, Inc.

All Rights Reserved.

Research Collaborators & Reviewers

August 2, 2017

2

1

8/2/2017

RATIONALE Objective of v17.0 NAEQ release: Enhance our client’s business.

•

•

Deliver significant advances in the application of earthquake science and engineering, providing the most up-to-date view of earthquake risk for the U.S., Canada and Mexico Driven by three key imperatives: •

To accurately implement the updated seismic hazard models such as the 2014 U.S. Geological Survey (USGS)

•

To optimize the model to accurately capture tail risk

•

To comprehensively re-calibrate the loss models, from site-specific hazards to damage and loss

Optimization >360,000 events reduced to 53,000 Preservation of: • • • •

AAL RP losses Total rate Mw distribution

Loss

•

Magnitude

Incorporate lessons learned from significant global earthquakes that have occurred since the v9.0 model update, most notably the 2010-2011 New Zealand Canterbury Earthquake Sequence, among others



August 2, 2017

3

SCOPE OF NAEQV17 AT A GLANCE

Canada

U.S.

Mexico



August 2, 2017

4

2

8/2/2017

SIGNIFICANT ENHANCEMENTS TO ALL NAEQ COMPONENTS

Seismic Sources

Ground Motions

Geotechnical Data

Secondary Perils

Vulnerability

 New/revised source models and rates

 USGS / NGAWest 2 updates

 Soil amplification (Vs30)

 Tsunami update

 Vulnerability updates

 Major event set characterization update for California

 RMS developed for Mexico and Hawaii

 Sprinkler Leakage update

 Very Tall Height band

 Basin updates  Liquefaction & landslide update

 Fire Following EQ update

 RMS developed Alaska and Hawaii Source model

 Inventory Distributions  Casualty  IFM updates

2014 USGS Copyright © 2017 Risk Management Solutions, Inc.

 BI updates


August 2, 2017

5

WHAT IS THE 2014 USGS UPDATE? v16.0 was based on 2008 NSHMP & UCERF2

UCERF3

2014

2014

NSHMP  Contiguous U.S.  Incorporated into building codes

 California only  Detailed research into faulting  Includes time-dependent rates

USGS – U.S. Geological Survey NSHMP – National Seismic Hazard Mapping Project UCERF3 – Uniform California Earthquake Rupture Forecast version 3 Copyright © 2017 Risk Management Solutions, Inc.


August 2, 2017

6

3

8/2/2017

WHAT IS THE 2014 USGS UPDATE? 2014

NSHMP

USGS – U.S. Geological Survey NSHMP – National Seismic Hazard Mapping Project UCERF3 – Uniform California Earthquake Rupture Forecast version 3


Seismic Sources



August 2, 2017

Ground Motions

Geotechnical Data


August 2, 2017

7

Secondary Perils

Vulnerability

8

4

8/2/2017

High

WHERE IS THE RISK? Low

Pacific Northwest (PNW)

New Madrid GR 250 Loss Ratio

California

3.0% V16

2.5%

V17 2.0%

1.5%

v17.0 NAEQ Ground-up Loss Cost

1.0%

0.5%

0.0% MX Copyright © 2017 Risk Management Solutions, Inc. .


California

August 2, 2017

New Madrid

PNW

Eastern CA Western CA

9

EQ 101: WHAT IS AN EARTHQUAKE? What is an Earthquake? Mechanics - an earthquake is the sudden slip along a fault in the Earth’s crust Site response - an earthquake is the ground motion or shaking felt due to vibrations in the soil or rock

Earthquake sources: Known Faults


Subduction Zones


August 2, 2017

Background Seismicity

10

5

8/2/2017

Primary goals of UCERF3: 1. Reduce frequency of Mw 6.5-7 events 2. Include multi-fault ruptures

Includes events connecting north and south (going through creeping section)

Courtesy of UCERF3 report Copyright © 2017 Risk Management Solutions, Inc. .


August 2, 2017

11

DIFFERENTIATED INSIGHT 1: UCERF3 COMPLETENESS OF EVENT SET

OpenSHA validation USGS vs RMS at 500 yr Majority of grid cells, ±2% Copyright © 2017 Risk Management Solutions, Inc.


August 2, 2017

12

6

8/2/2017

Seismic Sources


Ground Motions

Geotechnical Data


August 2, 2017

Secondary Perils

Vulnerability

13

EQ 101: GROUND MOTION ATTENUATION



Attenuation described by GMPEs

Decay of ground motion intensity with distance from the event source 



August 2, 2017

Rate of attenuation varies regionally and by source type

14

7

8/2/2017

GROUND MOTION PREDICTION EQUATIONS (GMPEs) UPDATE 

Significant recent advances in ground motion modelling (e.g., NGA-West2 and BC Hydro)

GMPEs updated for all regions

USGS 2014** NGAWest 2*

USGS 2014

*Pacific Earthquake Engineering Research Center **Mexico also includes a GMPE based on local data


Seismic Sources



August 2, 2017

Ground Motions

Geotechnical Data


August 2, 2017

15

Secondary Perils

Vulnerability

16

8

8/2/2017

EQ 101: GEOTECHNICAL DATA

Soil type (updated)

Sub-perils



Softer soils increase ground motion amplitudes



Amplification is greater in deep sedimentary basins







Liquefaction (updated) Saturated, sandy soils temporarily take on liquid-like properties and lose bearing capacity

Landslide (updated) Ground shaking may induce down-slope movement of soil / rock

August 2, 2017

17

v17 Vs30 map

SOIL DATA & SITE AMPLIFICATION



Model now uses Vs30 instead of RMS Soil Type

Vs30

shear wave velocity in the top 30m of the earth’s surface



Modeling of site amplification more accurate



New site amplification models for all tectonic settings –

New ground motion amplification models directly link Vs30 site conditions to new GMPEs and are the basis of the v17 site amplification factors



August 2, 2017

18

9

8/2/2017

DIFFERENTIATED INSIGHT 2: GEOTECHNICAL UPDATES QUANTITATIVE PARAMETER OF VS30

LOS ANGELES, CA

v17 RMS Current RMSSoil Soil



August 2, 2017

19

BASIN AMPLIFICATION BASIN AMPLIFICATION Vancouver Seattle  Impact of deep sedimentary basins highlighted in Mw 8.0 Michoacán EQ, 1985 (Mexico City)  Ground motion amplification depends on a building’s location within the basin, its natural period of vibration, and the frequency content of an earthquake

Mississippi Embayment Los Angeles Updated New

Deep Basin

Basins included in NAEQ v17.0 Copyright © 2017 Risk Management Solutions, Inc.


August 2, 2017

Mexico City 20

10

8/2/2017

LIQUEFACTION





New probabilistic liquefaction model incorporates lessons learned from NZ: –

Liquefaction at surface correlates strongly with ground water depth

–

Ground settlement and lateral displacement result in significantly different building damage

Model assesses: –

Probability of liquefaction occurring

–

Expected vertical and horizontal displacement if liquefaction occurs



August 2, 2017

21

DIFFERENTIATED INSIGHT 3: LIQUEFACTION

RiskLink MDR footprint

Observed MDR footprint

11

8/2/2017

DIFFERENTIATED INSIGHT 3: LIQUEFACTION

Claims Data

Modeled MDR footprint

Seismic Sources


Ground Motions

Geotechnical Data


August 2, 2017

Secondary Perils

Vulnerability

24

12

8/2/2017

UPDATES TO SECONDARY PERILS

Tsunami

Fire Following EQ



New accumulation footprints provided with installation data



Accumulations can be run for various Cascadia EQ scenarios (instead of one)





Improved high-resolution simulations to inform model



New building density data for metro areas



Improved resolution for Alaska and Hawaii

Sprinkler Leakage



Expanded to include Canada

Differentiation by MDR (LOB-specific) in addition to simple wet vs dry


Seismic Sources



August 2, 2017

Ground Motions

Geotechnical Data


August 2, 2017

25

Secondary Perils

Vulnerability

26

13

8/2/2017

VULNERABILITY UPDATES

New Height Band for Very Tall Buildings



Building Inventory & Vulnerability Regions

New height band to better capture performance of very tall buildings:



Updated inventory and vulnerability regions for Canada, Mexico & Hawaii



Tall (15-40 stories)





Very Tall (>40 stories)

Updates to inventory distributions (all regions)



Business Interruption Updates



Adjustments to facility and lifeline restoration functions based on 2010-11 CES BI claims



Variable impact on losses

August 2, 2017

27

INDUCED SEISMICITY Change in AAL vs. base model



Induced seismicity rates and GMPE based on 2016 USGS model



Can now turn on induced seismicity in DLM profile



RMS model developers are closely following the evolving science on induced seismicity Induced Seismicity



August 2, 2017

Earthquakes resulting from human activity (primarily wastewater injection during drilling)

28

14

8/2/2017

SUMMARY: NORTH AMERICA EARTHQUAKE VERSION 17

KEY MESSAGES 





KEY NEW FEATURES 

New probabilistic liquefaction and landslide model



Vs30 data product



High-resolution VRG for better risk differentiation



Ability to accurately model very tall buildings

General decreases in loss



Ability to model induced seismicity

–



Alternative ground motion & vulnerability



New hazard validation tool



New IEDs / ILCs

Accurately implemented 2014 USGS update –

Captures correlation of risk

–

Preserves richness of tail

Host of non USGS-related updates to improve whole model

More complex stories for California and NM



August 2, 2017

29

15

Any Other Matters


Adjournment


Ea Ph rthqu hilade uake Augu elphia (C) S ust 6, a, Pen Study 2017

Recommend Documents