The New York City Consortium for Earthquake Loss Mitigation (NYCEM) lftlogolvl2
maintrcorner

 

NYCEM 1st -Year Technical Report
May 1, 1999 to April 30, 1999:

Earthquake Loss Estimation Study for The New York City Area

by

Guy Nordenson, Michael Tantala and Amanda Kumpf
Department of Civil Engineering and Operation Research
Princeton University

George Deodatis
Department of Civil Engineering and Engineering Mechanics
Columbia University

 

Table of Contents

Executive Summary
1.0 Introduction
    1.1 Seismicity of the New York Area
    1.2 HAZUS Methodology
    1.3 Sources of Information
2.0 Study Regions and Scenario Earthquakes
3.0 Region 1: Wall Street Census Tract
    3.1 General Building Stock for Wall Street Census Tract
    3.2 Soil Information for Wall Street Census Tract
    3.3 Economic Losses to Buildings for Wall Street Census Tract       
    3.4 Scenario Comparisons for Wall Street Census Tract
4.0 Region 2: Kips Bay Census Tract
    4.1 General Building Stock for Kips Bay Census Tract
    4.2 Soil Information for Kips Bay Census Tract
    4.3 Economic Losses to Buildings for Kips Bay Census Tract
    4.4 Scenario Comparisons for Kips Bay Census Tract
5.0 Region 3: Manhattan Below 59th Street
    5.1 General Building Stock for Manhattan below 59th Street
    5.2 Soil Information for Manhattan below 59th Street
    5.3 Economic Losses to Buildings for Manhattan below 59th Street   
    5.4 Scenario Comparisons for Manhattan below 59th Street
6.0 Region 4: Tri-State Region (NY, NJ and CT)
    6.1 General Building Stock for the Tri-State Region
    6.2 Soil Information for the Tri-State Region
    6.3 Economic Losses to Buildings for the Tri-State Region  
7.0 Conclusions
References
List of Figures
List of Tables

 

Executive Summary

A preliminary forecast of the type of losses that the New York City area could suffer after an earthquake is the subject of this study funded by FEMA Region II and coordinated by the Multidisciplinary Center for Earthquake Engineering Research (MCEER). The initial stages of this study involved fact-finding and assessment, with the development of preliminary soil maps and building inventories. The primary objective of this study was to carry out an initial risk characterization for Manhattan below 59th Street. The present report documents the findings of a preliminary study focusing on seismic risks in the New York City area. The vehicle for performing these loss estimations has been a software tool entitled Hazards US (or HAZUS). The Federal Emergency Management Agency, through the National Institute of Building Science (NIBS) and RMS, Inc., developed the standardized earthquake loss estimation methodology and computer modeling program HAZUS, which can be used to quantify regional seismic risks and to form the basis for a more coordinated national loss program. HAZUS uses geographic information systems to model the built environment against the backdrop of possible natural disasters. The tasks of this initial first year study were to:

  • Become familiar with earthquake loss estimation methodologies and the HAZUS program.
  • Perform HAZUS scenario runs in the New York City area using default soil and building information supplied by the HAZUS code.
  • Perform HAZUS scenario runs using two representative census tracts in Manhattan to examine the sensitivity of loss estimation to different soil conditions and different building inventories.

The preliminary results of this research indicate:

  • Dramatic differences in total loss estimates between runs done with default values and runs done with improved estimates of soil conditions and building inventories. Differences are more dramatic for smaller magnitude events.
  • Total loss estimates in the modified runs can differ significantly with those of the default (by more than a factor of 10).
  • The effect of switching to better estimates of building inventory can be as important as the effect of switching to better estimates of soil conditions.
  • Parts of New York City have the unique characteristic of a considerable percentage of tall buildings.
  • It is of paramount importance to establish better estimates for soil conditions and building inventory for the entire New York City Area.

Future work for this research is recommended to develop a more accurate loss estimate. Suggested future work includes:

  • Provide better data for building age, type, quality, height, square footage, and seismic design level and perform sensitivity analyses to determine their relative importance.
  • Upgrade soil and building inventory information for the entire New York City area.
  • Develop and upgrade more accurate fragility curves for the type of buildings unique to the New York City area.

Eventually, the aim of this loss estimation project will provide a framework for businesses and agencies to take mitigative action to reduce potential damage and losses which might be experienced after an earthquake.

 

1.0 Introduction

The past several decades have witnessed a series of costly and damaging earthquakes. Although earthquake losses in the United States have been predominantly in California, many significant earthquakes have occurred and many more are projected in the areas that have been inactive in the last century (Figure 1.1).

New York City’s seismic risk exposure is of increasing concern. The New York City metropolitan area has been classified by the United States Geologic Survey (USGS) to the moderate level for potential earthquakes. In order to be prepared for such natural disasters, it becomes essential to be able to estimate and predict the risk associated with these potential losses. Risk is typically defined by three components: a hazard (the earthquake), the assets involved and the fragility of those assets. For New York City, the probability of a large earthquake is moderate, however it becomes an area of high risk because of its tremendous assets and the fragility of its structures, which have not been seismically designed as most on the West Coast.

The present report documents the findings of a preliminary study which focused on seismic risks in the New York City area. The vehicle for performing these loss estimations has been a software tool entitled Hazards US (or HAZUS). The Federal Emergency Management Agency, through the National Institute of Building Science (NIBS) and RMS, Inc., developed a standardized earthquake loss estimation methodology and the computer modeling program HAZUS. It can be used to quantify regional seismic risks and form the basis for a more coordinated national loss program. HAZUS uses geographic information systems to model the built environment against the backdrop of possible natural disasters.

The objectives of this initial study were:

  • Become familiar with earthquake loss estimation methodologies and the HAZUS program.
  • Perform HAZUS scenario runs in the New York City area using default soil and building information supplied by the HAZUS code.
  • Perform HAZUS scenario runs using two representative census tracts in Manhattan to examine the sensitivity of loss estimation to different soil conditions and different building inventories.

 

1.1 Seismicity of the New York Area

Earthquakes are not unknown in the New York City metropolitan area and up to a Modified Mercalli Intensity VII (MMI VII) has been observed in historical times (e.g., the 18 December 1737 event, which reportedly caused chimneys to fall in New York City). Seismic hazard in the northeast United States is a subject involving considerable uncertainty. Major events in the New York City area include the 18 December 1737 and the 10 August 1884 earthquakes. The 1884 earthquake is the largest and probably best documented event for the New York City area. The earthquake was a strong shock, centered off Rockaway Beach about 17 miles southeast of New York’s City Hall, and felt over 70,000 square miles, from Vermont to Maryland. In New York City, the effects were strong but varied, frightening many but causing very little or no damage. In Manhattan, newspaper reports indicated general alarm and in many portions of lower Manhattan, crockery and bottles rattled but generally did not fall.

On the basis of different historical descriptions, it has been estimated that the general intensity of this pattern in Manhattan and northwest Brooklyn (not yet a borough) was Modified Mercalli Intensity (MMI) IV and approaching MMI V toward the southeast. Current best estimates of the magnitude based on felt area, Mfa; of the 1737 and the 1884 events are Mfa = 4.5 for the 1737 event (epicenter 41N, 73.75W) and Mfa = 4.9 for the 1884 event (epicenter 40.51N, 73.83W).

 

1.2 HAZUS Methodology

The HAZUS methodology involves three basic components: classification of different systems for inventory (in this study, building types and soil information), methods for evaluating the damage and calculating losses, and databases of information on demographics, building information and the regional economy. An earthquake loss estimate can be performed using HAZUS for any location in the nation using only the methodology and default databases, however, more accurate loss estimates can be generated by collecting and incorporating additional (modified) information.

The first step to perform a loss estimate in HAZUS is to select an area to be studied, which might be defined by political boundaries (for example, census tract or a city). Then select a magnitude and epicenter location of a scenario earthquake. This can be based on available knowledge of historic seismicity. Information on local soil conditions can be incorporated to facilitate the mapping of estimated shaking intensities and the probability of permanent ground deformation. Using building capacity and fragility curves, HAZUS estimates damages and loss from the given scenario earthquake. Given appropriately modified input information (e.g., for the building stock), more accurate estimates of loss may be determined.

 

1.3 Source of Information

This study produces loss estimates using default and modified building and soil information. The modified building information was determined using visual inspection of the study regions, engineering judgement and the 1998-1999 Sandborn maps. Sandborn maps are typically used by the insurance industry and provide information about the building height, size, location, occupancy and type. The modified soil information was provided by a parallel research group at the Lamont-Doherty Observatory of Columbia University (Jacob 1999). This research group was headed by senior researcher, Klaus Jacob.

 

2.0 Study Regions and Scenario Earthquakes

A HAZUS model requires a defined study area (composed of census tracts) and a scenario earthquake (defined at least by a magnitude and an epicenter location). The basic geographic unit of analysis is a census tract. The four separate study regions that were modeled (see Table 2.1) within and around the New York City area include: a single census tract around Wall Street, a single census tract around Kips Bay, a collection of 132 census tracts in Manhattan below 59th Street and a collection of about 5,200 census tracts in the surrounding Tri-State region. While the Wall Street census tract is representative of a commercial area, the Kips Bay census tract is representative of a residential area. The Manhattan study region below 59th Street provides a model of a small-scale impact assessment on a series of census tracts. The Tri-State study, shown in Figure 2.1, provides a large-scale impact assessment showing how a natural disaster can affect a large region.

Table 2.1-Regions and Scenario Earthquakes Studied
 
Fixed Location
Constant Probability
Earthquake at 1884 Historical Location De-Aggregated M-D for 2% in 50 years
  5.0 M 6.0 M 7.0 M 5.0 M 6.0 M 7.0 M
Wall Street Census Tract            
Default soils and default building inventory A1 A2 A3 A4 A5 A6
Modified soils and default building inventory B1 B2 B3 B4 B5 B6
Default soils and modified building inventory C1 C2 C3 C4 C5 C6
Modified soils and modified building inventory D1 D2 D3 D4 D5 D6
             
Kips Bay Census Tract            
Default soils and default building inventory E1 E2 E3 E4 E5 E6
Modified soils and default building inventory F1 F2 F3 F4 F5 F6
Default soils and modified building inventory G1 G2 G3 G4 G5 G6
Modified soils and modified building inventory H1 H2 H3 H4 H5 H6
             
NYC Below 59th Street            
Default soils I1 I2 I3 I4 I5 I6
Modified soils J1 J2 J3 J4 J5 J6
             
31 County Region (NY, NJ & CT) K1 K2 K3 K4 K5 K6
             

For each study area, different inventory information was modeled and compared. For the Wall Street and Kips Bay census tract studies, four runs were considered: using default building and default soil inventories, using modified building and default soil inventories, using default building and modified soil inventories and using modified building and modified soil inventories.

As shown in Table 2.1, for each study area with a specific inventory case, six different earthquake scenarios were modeled and examined. Three scenario earthquakes, magnitudes 5, 6 and 7, were modeled at the fixed location of the 1884 historic earthquake. An additional three scenario earthquakes, also magnitudes 5, 6 and 7, were modeled at locations that attempt to represent a constant probability of reoccurrence of 2% in 50 years. Figure 2.2 shows the epicenter location of the different scenario earthquakes in comparison to Manhattan and the distance of these epicenters to the Empire State building.

 

3.0 Region 1: Wall Street Census Tract

The first area of study is a single census tract containing Wall Street with a land area of 0.06 square miles and a population of 154 inhabitants. The Wall Street census tract is representative of a predominately commercial area with 63 buildings, containing approximately 40 million square feet of floor area. Figure 3.1 shows the location of the Wall Street census tract in lower Manhattan. Figure 3.2 shows a closer view of the Wall Street census tract (shaded) with each building marked by a star and labeled with the number of stories for each building.

 

3.1 General Building Stock for Wall Street Census Tract

HAZUS uses building square footage to calculate economic losses for buildings from a scenario earthquake. For Wall Street, a comparison of building area by general occupancy type (see Table 3.1) shows significant differences in the building area of the HAZUS default inventory with the actual. The HAZUS default inventory assumes 20 million square feet of building area for 710 buildings. The Wall Street census tract actually contains about double the square footage (40 million square feet) for only 63 buildings. The default HAZUS inventory characterizes Wall Street as a census tract with a large number of low-story buildings, when it actually contains mainly a small number of very tall buildings.

Table 3.1-Wall Street Census Tract: Comparison of Building Square Footage by Occupancy
 
Hazus Default
Wall Street Census Tract
  square feet count square feet count
Residential 201,800 10 991,914 5
Commercial 19,599,500 649 38,574,963 57
Industrial 567,800 31 - -
Agricultural - - - -
Religious 250,700 17 18,468 1
Governmental - - - -
Educational 52,800 3 - -

 

Total 20,672,600 710 39,585,345 63

Table 3.2 shows a comparison of the number of buildings by type for the HAZUS default inventory with the actual. Again, the HAZUS inventory greatly overestimates the number of buildings actually in the Wall Street census tract. For example, HAZUS estimates 163 unreinforced masonry buildings, when there is actually only one.

Table 3.2—Wall Street Census Tract: Comparison of Number of Buildings by Type
 
Hazus Default
(count)
Wall Street Census Tract
(count)
Wood 154 3
Steel 264 49
Reinforced Concrete 46 4
Precast Concrete 22 3
Reinforced Masonry 61 3
Unreinforced Masonry 163 1
Mobile Homes - -

 

Total 710 63

In addition to building square footage, HAZUS makes some other building assumptions to estimate economic loss from earthquakes (Table 3.3). This includes assumptions on the percent distribution of buildings in each census tract by age, quality, seismic design level and building height. HAZUS inventory assumes that 100% of the buildings in the Wall Street census tract were constructed after 1970. Although detailed information on the date of construction was not available for all of the buildings in the Wall Street census tract, engineering judgement and a sampling of some of the buildings indicate that half of the buildings were actually constructed between 1950 and 1970 and the other half were constructed after 1970.

Table 3.3 - Wall Street Census Tract: Building Assumptions
    Hazus Default Wall Street
Age    
  Pre 1950 0% 0%
  1950-1970 0% 50%
  Post 1970 100% 50%
Quality    
  Code 25% 75%
  Inferior 75% 25%
  Superior 0% 0%
Seismic Design Level    
  Low 0% 75%
  Moderate 0% 25%
  High 100% 0%
Building Height    
  Up to 3 Stories 100% 5%
  4-7 Stories 0% 14%
  Above 8 Stories 0% 81%

For building quality, HAZUS identifies buildings by three categories: inferior, built just to code requirements, and superior. The HAZUS inventory assumes that about 75% of the buildings in the Wall Street census tract are inferior and 25% of the buildings are built to code quality. Again, although detailed information on the quality of construction was not available for all of the buildings in the Wall Street census tract, engineering judgement and a sampling of some of the buildings indicate that three-quarters of the buildings were actually constructed to code and only one-quarter were constructed to an inferior quality.

For seismic design level, HAZUS categorizes buildings as low, moderate and high. The HAZUS inventory assumes that all of the buildings in the Wall Street census tract are designed for the high seismic design level. The available information for this East Coast census tract indicates that none of the buildings are designed for the high seismic design level. Engineering judgement suggests that three-quarters of the buildings are actually designed for the low seismic design level and only one-quarter of the buildings qualify for the moderate level.

The fourth building assumption that HAZUS uses to estimate losses is the percent distribution of buildings by height. Taller buildings have in general relatively longer natural periods, which means that they will have a relatively lower response when compared to shorter buildings. This can be inferred from the response spectra plotted in Figure 3.3 (it should be pointed out that this statement is strictly qualitative). Eventually, a lower response will lead to less damage and loss. As a result, the height of a building becomes an important factor in determining potential damage and losses. By refining the building inventory with the correct building height distribution, it is possible to get a more accurate estimate of loss.

For the Wall Street census tract, the default HAZUS estimate is that 100% of buildings are under 4 stories. However, the actual count of buildings indicates that only 5% of buildings are under 4 stories. 14% of buildings are between 4 and 7 stories and 81% above 7 stories. Figures 3.3 and 3.4 show the default and actual distribution of building heights for the Wall Street census tract, plotted versus a typical response spectrum. In these two figures, buildings are represented by their fundamental natural periods. Comparing the two figures, it becomes obvious that using the actual distribution of building heights is extremely important to estimate reliably the overall structural damage (rather than using the default distribution provided by HAZUS).

 

3.2 Soil Information for Wall Street Census Tract

Soil type by census tract is another classification used to estimate losses. HAZUS uses the 1997 NEHRP Provisions to classify soil into site classes A, B, C, D or E, as shown in Table 3.4. The classification scheme of the NEHRP Provisions is based, in part, on the average shear wave velocity of the upper 30 meters of the local site geology.

Table 3.4-NEHRP Soil Type Classifications
Site Class Site Class Description   Shear Wave Velocity (m/sec)
      Minimum Maximum
A

Hard Rock

 

Eastern United States Sites Only 1500  
B

Rock

  760 1500
         
C

Very Dense Soil and Soft Rock

 

Untrained shear strength us >=2000 psf (us >= 100 kPa) or N >= 50 blows/ft. 360 760
D

Stiff Soils

 

Stiff soil with undrained shear strength 1000 psf <= us <= 2000 psf (50 kPa <= us <= 100 kPa) or 15 <= N <= 50 blows/ft 180 360
E

Soft Soils

 

Profile with more than 10 ft (3 m) of soft clay defined as soil with plasticity index PI > 20, moisture content w > 40% and undrained shear strength us < 1000 psf (50 kPa) (N < 15 blows/ft)   180
F

Soils Requiring Site Specific Evaluations

 

  • Soils vulnerable to potential failure or collapse under seismic loading: e.g. liquefiable soils, quick and highly sensitive clays, collapsible weakly cemented soils.
  • Peats and/or highly organic clays: (10 ft (3 m) or thicker layer)
  • Very high plasticity clays: (25 ft (8 m) or thicker layer with plasticity index > 75)
  • Very thick soft/medium stiff clays: (120 ft (36 m) or thicker layer)
   

The default HAZUS soil type for the Wall Street census track is Class D, indicating that it is a stiff soil. The actual soil is Class C, a soft rock with a higher shear wave velocity (Jacob 1999). Comparing typical response spectra for soils in Classes C and D (see Figure 3.5), it becomes obvious that the assumption of soil Class D will lead eventually to higher overall losses compared to the assumption of soil Class C (everything else remaining unchanged). By refining the building inventory with the soil information, it is possible to get a more accurate estimate of loss.

 

3.3 Economic Losses to Buildings for Wall Street Census Tract

Table 3.5 shows the structural and total economic losses for the Wall Street census tract for the various cases considered. The economic losses are shown for different inventory information (default and modified soil and building information) and for different scenario earthquakes (magnitude 5, 6 and 7 earthquakes at fixed location and constant probability.

Table 3.5-Wall Street Census Tract: Damages for Different Scenario Earthquakes and Building Inventory and Soil Information
Cost Structural Damage
(in thousands of dollars)
EQ, Fixed Location
5.0M
6.0M
7.0M
  EQ, Constant Probability
5.0M
6.0M
7.0M
Default Soil, Default Bldg 3,369 41,249 187,388   Default Soil, Default Bldg 3,369 4,488 21,104
Modified Soil, Default Bldg 1,701 26,079 151,426   Modified Soil, Default Bldg 1,701 2,476 9,779
Default Soil, Modified Bldg 772 82,287 283,809   Default Soil, Modified Bldg 772 4,188 43,167
Modified Soil, Modified Bldg 218 46,296 228,635   Modified Soil, Modified Bldg 218 1,430 19,834
 

Cost Total Loss
(in thousands of dollars)

EQ, Fixed Location
5.0M
6.0M
7.0M
  EQ, Constant Probability
5.0M
6.0M
7.0M
Default Soil, Default Bldg 59,322 412,136 1,713,500   Default Soil, Default Bldg 59,322 32,969

151,485

Modified Soil, Default Bldg 24,530 260,632 1,403,898   Modified Soil, Default Bldg 24,530 15,605 62,739
Default Soil, Modified Bldg 13,735 445,526 1,651,316   Default Soil, Modified Bldg 13,735 18,431 203.018
Modified Soil, Modified Bldg 4,068 272,707 1,307,527   Modified Soil, Modified Bldg 4,068 6,236 79,541
 
Total Loss Per Square Foot
(in dollars)
EQ, Fixed Location
5.0M
6.0M
7.0M
  EQ, Constant Probability
5.0M
6.0M
7.0M
Modified Soil, Modified Bldg 0.10 6.89 33.02   Modified Soil, Modified Bldg 0.10 0.16 2.01

For a 5.0 magnitude earthquake at the 1884 historic epicenter, the Wall Street census tract has an estimated $59.3 million total loss using the HAZUS default inventories. As the default building inventory and soil information are modified to represent the actual site conditions, the estimated total loss becomes $4 million. Refining the HAZUS default information changes the loss estimate by a factor of about 15 for a 5.0 magnitude earthquake. As the magnitude of the earthquake increases from 5.0 to 7.0, the general trend is the same, however the loss ratio of completely default information to completely modified information decreases from 15 to about 1 or 2 (depending on whether it is a fixed location or constant probability earthquake). Therefore, to get an accurate estimate of loss, especially for smaller earthquakes, it is extremely important to have accurate building and soil information.

 

3.4 Scenario Comparisons for Wall Street Census Tract

As shown in Figure 3.6 for fixed location scenario earthquakes, the total economic loss to the Wall Street census tract increases as the magnitude of the earthquake increases. The general trend shows that for the loss difference between the default and modified site information decreases as the magnitude of the earthquake increases.

Figure 3.7 for constant probability scenario earthquakes shows similar results. The total economic loss to the Wall Street census tract generally increases as the magnitude of the earthquake increases. The trend also shows that for small earthquakes, the loss difference between the default and modified site information decreases as the magnitude of the earthquake increases.

 

4.0 Region 2: Kips Bay Census Tract

The second area of study is the Kips Bay census tract, between 1st and 3rd Avenues and 29th and 31st Streets, with a land area of 0.02 square miles and a population of 7,195 inhabitants. The Kips Bay census tract is representative of a predominately residential area with 230 buildings, containing approximately 5 million square feet of floor area. Figure 4.1 shows the location of the Kips Bay census tract in lower Manhattan. Figure 4.2 shows a closer view of the Kips Bay census tract (in bold).

 

4.1 General Building Stock for Kips Bay Census Tract

HAZUS uses building square footage to calculate economic losses for buildings from a scenario earthquake. For Kips Bay, a comparison of building area by general occupancy type (see Table 4.1) shows significant differences in the building area of the HAZUS default inventory with the actual. The HAZUS default inventory assumes 6.5 million square feet of building area for 458 buildings. The Kips Bay census tract actually contains about two-thirds of the square footage (4.9 million square feet) for about half of the buildings (230 buildings). The default HAZUS inventory characterizes Kips Bay as a census tract with a large number of residential buildings and this is true, however it actually only contains about half the number of buildings.

Table 4.1-Kips Bay Census Tract: Comparison of Building Square Footage by Occupancy
 
Hazus Default
Kips Bay Census Tract
  square feet count square feet count
Residential 5,127,800 383 4,046,300 205
Commercial 1,121,200 58 433,800 20
Industrial 102,800 5 2,500 1
Agricultural 6,200 1 - -
Religious 126,600 8 25,900 2
Governmental 14,400 1 283,700 1
Educational 30,700 2 95,900 1

 

Total 6,529,700 458 4,888,100 230

Table 4.2 shows a comparison of the number of buildings by type for the HAZUS default inventory with the actual. Again, the HAZUS inventory overestimates the number of buildings actually in the Kips Bay census tract, here by a factor of two. For example, HAZUS estimates 270 wood buildings, when there are actually only 11. This is a significant discrepancy.

Table 4.2—Kips Bay Census Tract: Comparison of Number of Buildings by Type
 
Hazus Default
(count)
Kips Bay Census Tract
(count)
Wood 270 11
Steel 38 10
Reinforced Concrete 18 63
Precast Concrete 2 26
Reinforced Masonry 34 60
Unreinforced Masonry 96 60
Mobile Homes - -

 

Total 458 230

In addition to building square footage, HAZUS makes some other building assumptions to estimate building economic loss from earthquakes (Table 4.3). This includes assumptions on the percent distribution of buildings in each census tract by age, quality, seismic design level and building height. HAZUS inventory assumes that 100% of the buildings in the Kips Bay census tract were constructed after 1970. Although detailed information on the date of construction was not available for all of the buildings in the Kips Bay census tract, engineering judgement and a sampling of buildings indicate that 15% of the buildings were actually constructed before 1950, 50% were built between 1950 and 1970 and 35% were constructed after 1970.

Table 4.3 - Kips Bay Census Tract: Building Assumptions
    Hazus Default Kips Bay
Age    
  Pre 1950 0% 15%
  1950-1970 0% 50%
  Post 1970 100% 35%
Quality    
  Code 25% 50%
  Inferior 75% 25%
  Superior 0% 25%
Seismic Design Level    
  Low 0% 75%
  Moderate 0% 25%
  High 100% 0%
Building Height    
  Up to 3 Stories 100% 26%
  4-7 Stories 0% 68%
  Above 8 Stories 0% 6%

For building quality, the HAZUS inventory assumes that about 75% of the buildings in the Kips Bay census tract are inferior and 25% of the buildings are built to code quality. Again, although detailed information on the quality of construction was not available for all of the buildings in the Kips Bay census tract, engineering judgement and a sampling of some of the buildings indicate that half of the buildings were actually constructed to code, one-quarter were constructed to a superior quality and another one-quarter to an inferior quality.

For seismic design level, the HAZUS inventory assumes that all of the buildings in the Kips Bay census tract are designed for the high seismic design level. The available information for this East Coast census tract indicates that none of the buildings are designed for the high seismic design level. Engineering judgement suggests that three-quarters of the buildings are actually designed for the low seismic design level and only one-quarter of the buildings qualify for the moderate level.

For the Kips Bay census tract, the default HAZUS estimate is that 100% of buildings are under 4 stories. However, the actual count of buildings indicated that 25% of buildings are under 4 stories, 68% are between 4 and 7 stories, and 6% above 7 stories. Figure 4.3 and 4.4 show the default and actual distribution of building height for the Kips Bay census tract, plotted versus a typical response spectrum. In these two figures, buildings are represented by their fundamental natural periods. Comparing the two figures, it becomes obvious that using the actual distribution of building heights is extremely important to estimate accurately the overall damage (rather than using the default distribution provided by HAZUS). A similar conclusion was also reached when studying the Wall Street census tract.


 

4.2 Soil Information for Kips Bay Census Tract

The default HAZUS soil type for the Kips Bay census track is Class D, indicating that it is a stiff soil. The actual soil is Class C, a soft rock with a higher shear wave velocity (Jacob 1999). Comparing typical response spectra for soils in Classes C and D (See Figure 4.5), it becomes obvious that the assumption of soil Class D will lead eventually to higher overall losses compared to the assumption of soil Class C (everything else remaining unchanged). By refining the building inventory with the soil information, it is possible to get a more accurate estimate of loss.

 

4.3 Economic Losses to Buildings for Kips Bay Census Tract

Table 4.4 shows the structural and total economic losses for the Kips Bay census tract for the various cases considered. The economic losses are shown for different inventory information (default and modified soil and building information) and for different scenario earthquakes (magnitude 5, 6 and 7 earthquakes at fixed location and constant probability.

Table 4.4-Kips Bay Census Tract: Damages for Different Scenario Earthquakes, Building Inventory, Soil Information
Cost Structural Damage
(in thousands of dollars)
EQ, Fixed Location
5.0M
6.0M
7.0M
  EQ, Constant Probability
5.0M
6.0M
7.0M
Default Soil, Default Bldg 337 4,414 21,524   Default Soil, Default Bldg 337 804 2,974
Modified Soil, Default Bldg 235 2,540 16,476   Modified Soil, Default Bldg 235 440 1,572
Default Soil, Modified Bldg 47 2,277 12,952   Default Soil, Modified Bldg 47 384 2,357
Modified Soil, Modified Bldg 27 1,222 9,210   Modified Soil, Modified Bldg 27 165 1,002
 

Cost Total Loss
(in thousands of dollars)

EQ, Fixed Location
5.0M
6.0M
7.0M
  EQ, Constant Probability
5.0M
6.0M
7.0M
Default Soil, Default Bldg 4,402 49,521 186,793   Default Soil, Default Bldg 4,402 7,703 26,358
Modified Soil, Default Bldg 2,381 29,497 151,169   Modified Soil, Default Bldg 2,381 3,507 11,902
Default Soil, Modified Bldg 813 20,462 85,509   Default Soil, Modified Bldg 813 2,922 15,158
Modified Soil, Modified Bldg 317 11,666 66,856   Modified Soil, Modified Bldg 317 1,248 6,549
 
Total Loss Per Square Foot
(in dollars)
EQ, Fixed Location
5.0M
6.0M
7.0M
  EQ, Constant Probability
5.0M
6.0M
7.0M
Modified Soil, Modified Bldg 0.06 2.38 13.64   Modified Soil, Modified Bldg 0.06 0.25 1.34

For a 5.0 magnitude earthquake at the 1884 historic epicenter, the Kips Bay census tract has an estimated $4.4 million total loss using the HAZUS default values. As the default building inventory and soil information are modified to represent the actual site conditions, the estimated total loss becomes $317 thousand. Refining the HAZUS default information the loss estimate by a factor of about 14 for a 5.0 magnitude earthquake. As the magnitude of the earthquake increases from 5.0 to 7.0, the general trend is the same, however the loss ratio of completely default information to completely modified information decreases from 14 to about 3 or 4 (depending on whether it is a fixed location or constant probability earthquake). Therefore, to get an accurate estimate of loss, especially for smaller earthquakes, it is extremely important to have accurate building and soil information. A similar conclusion was reached by examining the Wall Street census tract.



4.4 Scenario Comparisons for Kips Bay Census Tract

As shown in Figure 4.6 for fixed location scenario earthquakes, the total economic loss to the Kips Bay census tract increases as the magnitude of the earthquake increases. The general trend shows that the loss difference between the default and modified site information decreases as the magnitude of the earthquake increases.

Figure 4.7 for constant probability scenario earthquakes shows similar results. The total economic loss to the Kips Bay census tract generally increases as the magnitude of the earthquake increases. The trend also shows that the loss difference between the default and modified site information decreases as the magnitude of the earthquake increases.

 

5.0  Region 3: Manhattan Below 59th Street

The third area of study is the Manhattan region below 59th Street, which includes 132 census tracts, with a land area of 10.71 square miles and a population of 550,000 inhabitants. The Manhattan region below 59th Street includes both residential and commercial census tracts with 44,762 buildings, containing approximately 914,000,000 square feet of floor area. Figure 5.1 shows the area of lower Manhattan below 59th Street that is considered in this study (outlined).

Figure 5.2 shows the population distribution within Manhattan below 59th Street. This figure is part of the HAZUS output.

 

5.1 General Building Stock for Manhattan Below 59th Street

For this preliminary report, only default building information was used for the Manhattan model below 59th Street. HAZUS uses building square footage to calculate economic losses for buildings from a scenario earthquake. Figure 5.3 shows the default distribution of the number of buildings in the Manhattan region below 59th Street.

In addition to building square footage, HAZUS makes some other building assumptions to estimate building economic loss from earthquakes (Table 5.1). This includes assumptions on the percent distribution of buildings in each census tract by age, quality, seismic design level and building height. HAZUS default inventory assumes that 100% of the buildings in each of the census tracts in Manhattan below 59th Street were constructed after 1970.

Table 5.1- Manhattan Census Tracts Below 59th Street: Building Assumptions
    Hazus Default
Age  
  Pre 1950 0%
  1950-1970 0%
  Post 1970 100%
Quality  
  Code 25%
  Inferior 75%
  Superior 0%
Seismic Design Level  
  Low 0%
  Moderate 0%
  High 100%
Building Height  
  Up to 3 Stories 100%
  4-7 Stories 0%
  Above 8 Stories 0%

For building quality, the HAZUS default inventory assumes that about 75% of the buildings in the Manhattan census tracts below 59th Street are inferior and 25% of the buildings are built to code quality. For seismic design level, the HAZUS default inventory assumes that all of the buildings in the Manhattan census tracts below 59th Street are designed for the high seismic design level.

Finally for the Manhattan census tracts below 59th Street, HAZUS as a default estimates that all of the buildings are under three stories. At this juncture, it is pointed out that the above default values provided by HAZUS most probably are not representing accurately the actual conditions in Manhattan below 59th Street.




5.2 Soil Information for Manhattan below 59th Street

The default HAZUS soil type for all of the Manhattan census tracts below 59th Street is Class D, indicating that it is a stiff soil. As shown in Figure 5.4, the actual soil class varies throughout Manhattan below 59th Street. This information was provided by Jacob (1999). By refining the soil information, a more accurate estimate of losses is expected.

 

5.3 Economic Losses to Buildings for Manhattan Below 59th Street

In Table 5.2, the economic losses are shown for different inventory information (default and modified soil information) and for different scenario earthquakes (magnitude 5, 6 and 7 earthquakes at fixed location and constant probability). For the Manhattan region below 59th Street, only default building inventory information was used.

Table 5.2-Manhattan Below 59th Street: Damages for Different Scenario Earthquakes and Soil Information
Cost Structural Damage
(in thousands of dollars)
EQ, Fixed Location
5.0M
6.0M
7.0M
  EQ, Constant Probability
5.0M
6.0M
7.0M
Default Soil, Default Bldg 86,399 1,140,683 5,386,804   Default Soil, Default Bldg 86,399 145,883 648,811
Modified Soil, Default Bldg 69,630 539,581 4,299,301   Modified Soil, Default Bldg 69,630 95,513 385,950
 

Cost Total Loss
(in thousands of dollars)

EQ, Fixed Location
5.0M
6.0M
7.0M
  EQ, Constant Probability
5.0M
6.0M
7.0M
Default Soil, Default Bldg 1,339,320 11,460,330 44,882,943   Default Soil, Default Bldg 1,339,320 1,188,883 4,870,706
Modified Soil, Default Bldg 952,801 5,488,060 37,168,731   Modified Soil, Default Bldg 952,801 712,884 2,803,283

For a 5.0 magnitude earthquake at the 1884 histoic epicenter, the Manhattan region below 59th Street has an estimated $1.34 billion total loss. As the default soil information is modified to represent the actual site conditions, the estimated total loss becomes $952 million. Refining the HAZUS default information for soil changes of the loss estimate by a factor of about 1.4 for a 5.0 magnitude earthquake. As the magnitude of the earthquake increases from 5.0 to 7.0, the general trend is the same, however the loss ratio of default soil information to modified information decreases from 1.4 to about 1.2. To get an accurate estimate of loss, especially for smaller earthquakes, it is very important to have accurate soil information. As shown in the Wall Street and Kips Bay studies, it is extremely important to have accurate building and soil information. This Manhattan study below 59th Street, however, only used modified soil information. Finally, Figure 5.6 shows the total damage by census tract in lower Manhattan for a constant probability 7.0 magnitude earthquake.

 

5.4 Scenario Comparisons for Manhattan Below 59th Street

As shown in Figure 5.6 for fixed location scenario earthquakes, the total economic loss to the 132 Manhattan census tracts below 59th Street increases as the magnitude of the earthquake increases. The general trend shows that the loss difference between the default and modified site information decreases as the magnitude of the earthquake increases.

Figure 5.7 for constant probability scenario earthquakes shows similar results. The total economic loss to the Manhattan census tracts below 59th Street generally increases as the magnitude of the earthquake increases. The trend also shows that the loss difference between the default and modified site information decreases as the magnitude of the earthquake increases.

 

6.0  Region 4: Tri-State Region (NY, NJ, and CT)

The fourth area of study is a 31 county, Tri-State region of New York, New Jersey and Connecticut, which includes 5,238 census tracts, with a land area of 12,990 square miles and a population of 20 million inhabitants. The Tri-State region contains both residential and commercial census tracts with 4 million buildings, containing approximately 16 billion square feet of floor area. Figure 6.1 shows the Tri-State region census tracts considered in this study (outlined).

Figure 6.2 shows the population distribution within the 31-county, Tri-State region. The darker regions are the more dense areas. This figure is part of the HAZUS output.

 

6.1 General Building Stock for the Tri-State Region

For this preliminary report, only default building information was used for the Tri-State region model. HAZUS uses building square footage to calculate economic losses for buildings from a scenario earthquake. Figure 6.3 shows the default distribution of the number of buildings in the Tri-State region. The darker regions are more dense.

 

6.2 Soil Information for the Tri-State Region

The default HAZUS soil type for all of the Tri-State region census tracks is Class D, indicating that it is a stiff soil. Only the default soil information was used for this scenario.

 

6.3 Economic Losses to Buildings for the Tri-State Region

In Table 6.1, the economic losses are shown by state for default soil and building information and for different scenario earthquakes (magnitude 5, 6 and 7 earthquakes at constant probability, although the term constant probability loses its meaning here because of the large area of the region under consideration).

Table 6.1--Tri-State Region: Damages for Different Scenario Earthquakes
(All values in thousands of dollars)
Constant Probability 5.0M
    Structural Damage Cost Total Loss
  CT 664 2,037
  NJ 123,346 1,374,521
  NY 388,941 6,220,177
 
  Total 512,951 7,596,735
 
Constant Probability 6.0M
    Structural Damage Cost Total Loss
  CT 58,185 434,001
  NJ 229,714 1,641,331
  NY 1,310,984 11,645,806
 
  Total 512,951 7,596,735
 
Constant Probability 7.0M
    Structural Damage Cost Total Loss
  CT 416,990 3,217,089
  NJ 1,337,779 9,409,302
  NY 3,815,976 25,777,500
 
  Total 5,570,745 42,403,851

For magnitude 5, 6 and 7 earthquakes at the constant probability locations, the Tri-State region has estimated total losses of $7.6 billion, $13.72 billion and $42.4 billion respectively. As the magnitude of the earthquake increases from 5.0 to 7.0, the obvious trend is that the total damages increase. Figure 6.4 shows a typicaldistribution of total loss in the Tri-State region due to a 7.0 magnitude earthquake (the darker regions experience the more significant losses).

 

7.0 Conclusions

The preliminary results of this research indicate:

  • Dramatic differences in total loss estimates between runs done with default values and runs done with improved estimates of soil conditions and building inventories. Differences are more dramatic for smaller magnitude events.
  • Total loss estimates in the default-modified runs can differ significantly with those of the default (by more than a factor of 10).
  • The effect of switching to better estimates of building inventory can be as important as the effect of switching to better estimates of soil conditions.
  • Parts of New York City have the unique characteristic of a considerable percentage of tall buildings.
  • It is of paramount importance to establish better estimates for soil conditions and building inventory for the entire New York City Area.

Future work for this research is recommended to develop a more accurate loss estimate. Suggested future work includes:

  • Provide better data for building age, type, quality, height, square footage, and seismic design level and perform sensitivity analyses to determine their relative importance.
  • Upgrade soil and building inventory information for the entire New York City area.
  • Develop and upgrade more accurate fragility curves for the type of buildings unique to the New York City area.

Eventually, the aim of this loss estimation project will provide a framework for businesses and agencies to take mitigative action to reduce potential damage and losses which might be experienced after an earthquake.

 

References

Federal Emergency Management Agency (FEMA), A Nontechnical Explanation of the 1994 NEHRP Recommended Provisions, FEMA 99, September 1995, Page 3.

Coffman, J. L. and Hake, C. A., Earthquake History of the United States, Publication 41-1 (rev. ed.), U.S. Department of Commerce, Boulder, CO, 1982.

Bernreutter, D. L. et al., Seismic Hazard Characterization of the Eastern United States: Methodology and Interim Results for Ten Sites, Lawrence Livermore National Laboratory, Livermore, CA, 1984.

The New York Times, August 11th and 12th, 1884; The Herald Tribune, August 11 and 12th, 1884.

Scawthorn, C. and Harris, S. K., Estimation of Earthquake Losses for a Large Eastern Urban Center: Scenario Events for New York City, The New York Academy of Sciences Annals, 1988.

Building Seismic Safety Council, NEHRP- Recommended Provisions for the Development of Seismic Regulations for New Buildings, Part I, Washington, DC, 1985.

Federal Emergency Management Agency (FEMA), 1997 NEHRP Provisions, FEMA 222A, 1997.

Jacob, K. H., Site Conditions Effecting Earthquake Loss Estimates for New York City, Technical Report Prepared for MCEER, 1999.

 

List of Figures

1.1 Seismicity of the United States: 1899-1990
2.1 Tri-State Study Region Study (Census Tracts outlined)
2.2 Epicenters of Scenario EQs with respect to Manhattan
3.1 Location of the Wall Street Census Tract within Manhattan
3.2 Wall Street Tract: Buildings, Number of Stories for each Building
3.3 Wall Street Tract: Default HAZUS dist., building heights vs. response spectrum
3.4 Wall Street Tract: Actual HAZUS dist., building heights vs. response spectrum
3.5 Comparison of an EQ Response for Soil Classes C and D
3.6 Wall Street Tract: Damage sensitivity, Different M, Fixed Location EQ Scenarios
3.7 Wall Street Tract: Damage sensitivity, Different M, Constant Probability EQ Scenarios
4.1 Location of the Kips Bay Tract within Manhattan
4.2 Kips Bay Tract
4.3 Kips Bay Tract: Default HAZUS dist., building heights vs. response spectrum
4.4 Kips Bay Tract: Actual HAZUS dist., building heights vs. response spectrum
4.5 Comparison of an EQ Response of Soil Classes C and D
4.6 Kips Bay Tract: Damage sensitivity, Different M, Fixed Location EQ Scenarios
4.7 Kips Bay Tract: Damage sensitivity, Different M, Constant Probability EQ Scenarios
5.1 Region of Manhattan below 59th Street with Tracts Outlined
5.2 Population Dist. in the Manhattan Region Below 59th Street
5.3 Dist. of the Number of Buildings in the Manhattan Region Below 59th Street
5.4 Actual Soil Types in Lower Manhattan by Tract (Jacob, 1999)
5.5 Total Damage by Tract in Lower Manhattan from a Constant Probability 7.0M EQ
5.6 Manhattan Below 59th St.: Damage sensitivity, Different M, Fixed Location EQ Scenarios
5.7 Manhattan Below 59th St.: Damage sensitivity, Different M, Constant Probability EQ Scenarios
6.1 Tri-State Region with Tracts Outlined
6.2 Population Dist. in the Tri-State Region
6.3 Dist. of the Number of Buildings in the Tri-State Region
6.4 Total Tri-State Damage by Tract for Constant Probability 7.0M EQ

 

List of Tables

2.1 Regions and Scenario Earthquakes Studied
3.1 Wall Street Census Tract: Comparison, Building Sq. Footage by Occupancy
3.2 Wall Street Tract: Comparison, Number of Buildings by Type
3.3 Wall Street Tract: Building Assumptions
3.4 NEHRP Soil Type Classifications
3.5 Wall Street Tract: Damages for Scenario EQs, Building Inventory, Soil Information
4.1 Kips Bay Tract: Comparison, Building Sq. Footage by Occupancy
4.2 Kips Bay Tract: Comparison, Number of Buildings by Type
4.3 Kips Bay Tract: Building Assumptions
4.4 Kips Bay Tract: Damages for Different Scenario EQs, Building Inventory, Soil Information
5.1 Building Assumptions, Manhattan Below 59th Street
5.2 Manhattan Damages Below 59th Street: Different Scenario EQs, Soil Information
6.1 Tri-State Region: Damages for Different Scenario EQs

 

mainblcorner.jpg (1604 bytes)

Home | Background | HAZUS | Research Activities | Who We Are | Technical Documents | Contacts | Links

Last modified: 1/29/2002 9:40:33 AM
Suggestions and comments to the webmaster.