Menu
Contact Us Search
Organization Title

  HAZUS Annualized Earthquake Loss Estimation for California

Rui Chen and Chris J. Wills
California Geological Survey
April 2011

Download Summary
Summary

 

Comprehensive estimation of the scale and extent of damage, social disruption, and economic losses due to potential earthquakes provides useful information for local and state officials in developing earthquake hazard mitigation strategies and preparing emergency response and recovery plans.  Accurate loss estimation, however, is difficult to accomplish because of changing earthquake risk due to continuing development, particularly in earthquake-prone urban areas, and because of changes in input ground motion values from evolving knowledge in seismic hazard assessment.  Significant research developments have occurred in the past few years in ground motion prediction, rupture source characterization, and our ability to more accurately account for the effects of near surface soil conditions on surface ground motions in probabilistic seismic hazard analyses (PSHA).  This study evaluates the effects of these latest developments on annualized earthquake loss (AEL) estimation, using HAZUS-MH, a geographic information system-based multi-hazard (MH) loss estimation tool developed by the Federal Emergency Management Agency.  AEL is a long term measure of earthquake risk.  We also estimated annualized building related loss as percentage of the building replacement value, or annualized percent earthquake loss (APEL).  Because it is normalized by the building replacement value, APEL is a better indicator of relative damage in different geographic areas than building-related loss.  It is also a better indicator of how estimated relative earthquake losses have been changing over time.  

Annualized loss estimation used probabilistic ground motions calculated using the United States Geological Survey (USGS) PSHA models developed for the 2008 Update of the United States National Seismic Hazard Maps and incorporating the effect of local soil conditions.  The 2008 USGS PSHA models use Next Generation Attenuation (NGA) relations and the earthquake source parameters of the 2007 Working Group on California Earthquake Probabilities.  AELs were estimated at three levels of geographic resolution: census tract, county, and Metropolitan Statistical Area (MSA).  Our calculations show that the lower PSHA ground motions predicted by the NGAs translate directly to substantially lower annual earthquake loss estimations in California.  Estimated annualized losses using PSHA ground motions incorporating site soil conditions are much higher than those assuming a uniform rock site condition.


We estimated the statewide average earthquake loss due to building damage to be approximately $2.8 billion per year, nearly 82% of this loss occurring in the top four MSAs with the highest estimated annual losses (Table 1).  These MSAs are Los Angeles-Long Beach-Santa Ana, San Francisco-Oakland-Fremont, Riverside-San Bernardino-Ontario, and San Jose-Sunnyvale-Santa Clara.  The ten counties with the highest seismic risk measured by estimated annualized earthquake loss are Los Angeles, Alameda, Santa Clara, Orange, San Bernardino, Riverside, Contra Costa, San Francisco, San Mateo, and Ventura.  Los Angeles County tops the list with an estimated annual loss of nearly one third of statewide loss and more than 3.5 times the loss in the 2nd-ranking county, Alameda (Table 2).  Statewide APEL is 0.103%.  San Benito, Imperial, and Alameda Counties rank the top three in terms of APEL, reflecting the relatively high ground motion hazards in these areas.  The distribution of estimated annual building losses on census tract level shows relatively higher values along the coast and in highly urbanized areas (Figure 1).  On the other hand, higher APELs are estimated along the San Andreas Fault System, generally reflecting higher ground motion hazards (Figure 2).  Estimated injuries are mostly not life threatening.  Fatality, however, could occur with an estimated annual rate ranging from a few individuals in some high risk counties up to 30 individuals in Los Angeles County.


 

Table 1. Ten California Metropolitan Statistical Areas with the Highest Estimated Annualized Building Loss and the Highest Annualized Percent Building Loss

  

Rank

Area

AEL ($M)

% of

State Total

Rank

MSA

APEL (%)

1

Los Angeles-Long Beach-Santa Ana

1102.8

39.0

1

El Centro

0.190

2

San Francisco-Oakland-Fremont

643.9

22.8

2

San Francisco-Oakland-Fremont

0.152

3

Riverside-San Bernardino-Ontario

316.4

11.2

3

San Jose-Sunnyvale-Santa Clara

0.146

4

San Jose-Sunnyvale-Santa Clara

243.5

8.6

4

Riverside-San Bernardino-Ontario

0.141

5

Oxnard-Thousand Oaks-Ventura

82.1

2.9

5

Oxnard-Thousand Oaks-Ventura

0.131

6

San Diego-Carlsbad-San Marcos

71.2

2.5

6

Santa Rosa-Petaluma

0.120

7

Santa Rosa-Petaluma

55.7

2.0

7

Los Angeles-Long Beach-Santa Ana

0.118

8

Santa Barbara-Santa Maria-Goleta

36.6

1.3

8

Santa Barbara-Santa Maria-Goleta

0.114

9

Sacramento-Arden-Arcade—Roseville

32.5

1.1

9

Santa Cruz-Watsonville

0.112

10

Vallejo-Fairfield

32.4

1.1

10

Napa

0.112

 

 

 

Table 2. Ten California Counties with the Highest Estimated Annualized Building Loss and the Highest Estimated Percent Building Loss

 

Rank

County

AEL

($M)

% of

State Total

Rank

County

APEL Loss

(%)

1

Los Angeles

903.1

31.9

1

San Benito

0.202

2

Alameda

247.9

8.8

2

Imperial

0.190

3

Santa Clara

235.4

8.3

3

Alameda

0.180

4

Orange

199.7

7.1

4

San Bernardino

0.155

5

San Bernardino

177.7

6.3

5

Humboldt

0.151

6

Riverside

138.7

4.9

6

San Mateo

0.146

7

Contra Costa

129.5

4.6

7

San Francisco

0.146

8

San Francisco

122.6

4.3

8

Santa Clara

0.144

9

San Mateo

111.2

3.9

9

Contra Costa

0.137

10

Ventura

82.1

2.9

10

Los Angeles

0.131

 

 


 

Figure 1. Distribution of seismic risk in California by estimated annualized building loss on census tract level based on ground motions calculated using the 2008 USGS PSHA models and incorporating site-specific VS30 values.



 

Figure 2. Distribution of estimated annualized building percent loss by census tract based on ground motions calculated using the 2008 USGS PSHA models and incorporating site-specific VS30 values.


 

An early study by FEMA (2008) using HAZUS default ground motions [i.e., the 2002 USGS PSHA data calculated using ground motion prediction equations (GMPE) and source models developed before 2003] indicates that the annualized earthquake loss to building stock in California is about $3.5 billion per year, 66 percent of the national loss ($5.3 billion).  It estimated an APEL of 0.145% for California.  California ranks number one in the nation in both estimated AEL and APEL.  In a study conducted in 2005, CGS quantified annualized building-related losses using the PSHA ground motions similar to those used in the FEMA 2008 study.  The 2005 study, however, used the HAZUS default data on built environment and demographics that are about 10 years older than the data used in both our current study and the FEMA 2008 study.  The CGS 2005 study estimated an AEL of $2.2 billion and an APEL of 0.138 for California.     

In comparison, our calculations yielded a statewide AEL that is about 19% lower than the FEMA 2008 estimation and 29% higher than the CGS 2005 estimation (see Table 3).  We believe the difference between our estimation and that of the FEMA 2008 study is primarily due to the use of NGAs.  The difference between our current estimate and that of the CGS 2005 study is attributed mainly to the compound effects of increase in building exposure and generally decreased ground motion predictions.  Specifically, increase in building exposure led to increase in AEL, whereas decrease in ground motion predictions led to decrease in APEL.  This postulation is supported by the comparison of APEL:  our current APEL estimation is about 28% lower than the FEMA 2008 estimation and about 25% lower than the CGS 2005 estimation.  To a lesser extent, these differences may also be attributed to the different methods used to incorporate site effects among the FEMA 2008, CGS 2005, and our current studies.

Table 3. Comparison of AELs and APELs from Various Studies
 

Study

Building Value ($M)

AEL

($M)

Difference1 (%)

APEL

(%)

Difference1 (%)

CGS 2010

2,721,165

2,829

-

0.104

-

FEMA 2008

2,416,5522

3,504

-19.3

0.145

-28.3

CGS 2005

1,590,000

2,200

28.6

0.138

-24.9

 

1Difference is computed using Equation 4 in this report

2Building value for FEMA 2008 study is back calculated using AEL and APEL reported in FEMA (2008)

 

Conclusions

·         Estimated AEL due to building damage is $2,829 million per year and estimated statewide APEL is 0.103%.  Los Angeles County ranks the top on the county list with an estimated AEL of nearly one third of statewide loss.

·         Generally lower ground motions predicted by the NGAs compared to older GMPEs translate to lower estimated statewide building loss (25% – 28% lower) as reflected by differences in estimated AELs from the current study and FEMA 2008 study; and by differences in estimated APELs between the current study and FEMA 2008 as well as CGS 2005 studies.  On the other hand, increase in building exposure lead to increase in AEL as reflected by differences in estimated AELs from the current study and CGS 2005 study.

·         Higher values of estimated annual building losses are distributed along the Pacific Coast and in highly urbanized areas, such as San Francisco Bay area and Los Angeles.  Higher percent losses occur along the San Andreas Fault System, generally reflecting higher ground motion hazards.


Ground motion text files can be downloaded from CGS ftp site: ftp://ftp.consrv.ca.gov/pub/dmg/rgmp/PSHA_2010/ or by clicking on the file names given below:

1.       Convert_for_HazardMap160.out: ground motions for VS30 = 160 m/s

2.       Convert_for_HazardMap216.out: ground motions for VS30 = 216 m/s

3.       Convert_for_HazardMap287.out: ground motions for VS30 = 287 m/s

4.       Convert_for_HazardMap377.out: ground motions for VS30 = 377 m/s

5.       Convert_for_HazardMap489.out: ground motions for VS30 = 489 m/s

6.       Convert_for_HazardMap609.out: ground motions for VS30 = 609 m/s

7.       Convert_for_HazardMap760.out: ground motions for VS30 = 760 m/s

8.       GMforHazardMap_ VS30.out: ground motions incorporating grid point-specific VS30 values. In this file, ground motion values are zero at locations where VS30 is 0 (e.g., in water bodies).

The first seven files contain ground motions for the seven VS30 values and the last one contains ground motions incorporating grid-point VS30 values.  Each file includes acceleration values for 4 spectra periods (peak ground acceleration and spectral accelerations at 0.2, 0.3, and 1.0 seconds) and 3 exceedance probability levels (2%, 5%, and 10% in 50 years).

See CGS 2009 Study for a complete set of selected HAZUS scenario results. These results can also be viewed by clicking the scenario names in Table 4.

Table 4. List of CGS 2009 Scenario Earthquakes

 

Northern California Scenarios

 Southern California Scenarios

N1

SAF_SAS+SAP+SAN+SAO

S1

Verdugo Fault

N2

SAF_SAS+SAP+SAN

S2

San Andreas Fault – Southern M7.8

N3

SAF_SAP+SAN+SAO

S3

Chino Hills Fault

N4

SAF_SAS+SAP

S4

San Joce -Carro Prieto Fault

N5

SAF_SAS

S5

Hosgri Fault

N6

SAF_SAP

S6

Elsinore Fault (Julian)

N7

SAF_SAN+SAO

S7

North Channel Slope

N8

SAF_SAN

S8

Puente Hills Fault

N9

SAF_SAO

S9

San Joaquin Hills Fault

N10

HRC_HS

S10

Elsinore Fault

N11

HRC_HN

S11

Raymond Fault

N12

HRC_HS+HN

S12

Whittier Fault

N13

HRC_RC

S13

Imperial Valley Fault

N14

HRC_HN+RC

S14

San Andreas Fault – Coachella Valley

N15

HRC_HS+HN+RC

S15

San Andreas Fault – Southern M7.4

N16

CLV_CS

S16

San Jacinto Fault

N17

CLV_CC

S17

Newport – Inglewood Fault

N18

CLV_CS+CC

S18

Palos Verdes Fault

N19

CLV_CN

S19

Santa Monica Fault

N20

CLV_CC+CN

S20

Rose Canyon Fault

N21

CLV_CS+CC+CN

S21

San Andreas Fault – 1857

N22

CGV_CON

 

 

N23

CGV_GVS

UCERF 2-Based Scenarios

N24

CGV_CON+GVS

U1

SAF_CO

N25

CGV_GVN

U2

SJ_SBV+SJV+A+C

N26

CGV_GVS+GVN

U3

SJ_A+C

N27

CGV_CON+GVS+GVN

U4

HRC_RC+HN+HS

N28

SGF_SGS

U5

HRC_RC

N29

SGF_SGN

U6

SAF_SAO+SAN+SAP+SAS

N30

SGF_SGS+SGN

U7

SAF_SAP+SAS

N31

GNV_GS

U8

Imperial

N32

GNV_GN

U9

Palos Verdes

N33

GNV_GS+GN

U10

Newport-Inglewood

N34

MTD

 

 

N35

GV