Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Evaluation of structural earthquake damage to buildings in Southwestern B.C. Blanquera, Ardel 1999

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1999-0250.pdf [ 7.7MB ]
Metadata
JSON: 831-1.0064091.json
JSON-LD: 831-1.0064091-ld.json
RDF/XML (Pretty): 831-1.0064091-rdf.xml
RDF/JSON: 831-1.0064091-rdf.json
Turtle: 831-1.0064091-turtle.txt
N-Triples: 831-1.0064091-rdf-ntriples.txt
Original Record: 831-1.0064091-source.json
Full Text
831-1.0064091-fulltext.txt
Citation
831-1.0064091.ris

Full Text

E V A L U A T I O N OF STRUCTURAL E A R T H Q U A K E D A M A G E TO BUILDINGS IN SOUTHWESTERN B.C. by ARDEL BLANQUERA B. S., Rutgers University College of Engineering, 1995 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF T H E REQUIREMENTS FOR T H E DEGREE OF MASTER OF APPLIED SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Civil Engineering) We accept this thesis as conforming to the reflujjed standard. THE UNIVERSITY OF BRITISH COLUMBIA May 1999 © Ardel Blanquera, 1999 In presenting this thesis in partial fulfillment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of <^iViL. Bsld <^ The University of British Columbia Vancouver, Canada Date / f e / C 30} ABSTRACT Seismic risk to the general building inventory is difficult to quantify. Risk analyses study the regional distribution of estimated damage that a building inventory will experience given a level of seismic intensity. This thesis presents a methodology for the evaluation of structural earthquake damage to buildings in Southwestern British Columbia. The steps in the risk analysis i addressed in this methodology are: (1) development of building inventory, (2) selection and application of motion-damage relationships and (3) regional distribution of damage using a geographic information system. Damage is measured as the ratio of dollar loss to replacement cost, or mean damage factor. The seismic hazard is characterized in terms of the Modified Mercalli Intensity. Only damage due to ground shaking hazard is considered. Several motion-damage relationships from prior expert opinion based studies were compared to select appropriate damage functions for use in this methodology. A case study is presented of the City of New Westminster, British Columbia. The seismic risk analyses was performed on a building inventory consisting of approximately 8,000 residential, commercial and industrial buildings. Two separate analyses were performed for comparison. The first uses a structural classification system based on 15 building types and the second is based on 31 building types. The regional distribution of results are aggregated and presented using the geographic information system, Maplnfo®. ii TABLE OF CONTENTS Abstract n Table of Contents i n List of Figures v List of Tables VII Acknowledgements VIII C H A P T E R 1 I N T R O D U C T I O N 1 1.1 Background 1 1.2 Objective 2 1.3 Scope 2 1.4 Organization of Thesis 3 C H A P T E R 2 L I T E R A T U R E R E V I E W 4 2.1 Overview 4 2.2 Regional Seismic Hazard and Risk Analysis 5 2.3 Definitions of Damage 8 2.4 Characterization of Ground Motion 10 2.5 Development of Building Damage Functions 11 2.6 Earthquake Loss Estimation in British Columbia 16 2.7 Conclusion 18 C H A P T E R 3 S T R U C T U R A L C L A S S I F I C A T I O N 20 3.1 Overview 20 3.2 ATC-13 22 3.3 Preliminary BC Survey 26 3.4 Southwestern B C Building Seismic Fragility Study 27 3.5 Conclusion 46 CHAPTER 4 DAMAGE ESTIMATION 48 4.1 Overview 48 4.2 ATC-13 49 4.3 F E M A - M B S 52 4.4 Preliminary BC Survey 56 4.5 Southwestern B C Building Seismic Fragility Study 59 4.6 Proposed Methodology for this study 70 4.7 Proposed Motion-Damage relationships for this study 72 4.8 Conclusions 83 CHAPTER 5 IMPLEMENTATION OF METHODOLOGY FOR SOUTHWESTERN BC 85 5.1 Overview 85 5.2 Structural Inventory 86 5.3 Regional Damage Distribution 97 5.4 Conclusions 98 CHAPTER 6 CASE STUDY: THE CITY OF NEW WESTMINSTER 99 6.1 Background 99 6.2 Seismic Hazard 101 6.3 Structural Inventory Development 102 6.4 Damage Estimation 105 6.5 Conclusions 109 CHAPTER 7 CONCLUSIONS 120 7.1 Summary 120 7.2 Conclusions 122 7.3 Future Work 125 References 127 IV LIST OF FIGURES Fig. 2.1: Regional Seismic Hazard & Risk Analysis 8 Fig. 4.1: FEMA-NIBS methodology for damage estimation 55 Fig. 4.2: Comparison of Mean Damage Factors for BC15 WLFR, BC31 WLFR and BC31 WPB Building Types 65 Fig. 4.3: Comparison of Mean Damage Factors for BC15 WLFCI, BC31 WLFCI and BC31 WPB Building Types 66 Fig. 4.4: Comparison of Mean Damage Factors for BC15 WLFLR, BC31 WLFLR and BC31 WPB Building Types 66 Fig. 4.5: Comparison of Mean Damage Factors for BC15 SBMFLR, BC31 SBFLR and BC31 SM-FLR Building Types 67 Fig. 4.6: Comparison of Mean Damage Factors for BC15 SFCWMHR, BC31 SFCWMR, BC31 SFCWHR, BC31 SFCI and BC31 SFMI Building Types 67 Fig. 4.7: Comparison of Mean Damage Factors for BC15 CFLW, BC31 CFLR and BC31 RCMFLR Building Types 68 Fig. 4.8: Comparison of Mean Damage Factors for BC15 CFMHW, BC31 CFMR, BC31 CFHR, BC31 RCFIW, BC31 RCMFMR and BC31 RCMFHR Building Types 68 Fig. 4.9: Comparison of Mean Damage Factor for BC15 RMLR, BC31 RMLR and BC31 RMMR Building Types 69 Fig. 4.10: Comparison of Mean Damage Factor for BC 15 PC, BC31 PCLRand BC31 PCMR Build-ing Types 69 Fig. 4.11: Wood frame (low rise) 77 Fig. 4.12: Unreinforced masonry with bearing wall (low rise) 77 Fig. 4.13: Unreinforced masonry with load bearing frame (low rise) 78 LIST OF FIGURES Fig. 4.14: Reinforced masonry shear wall without moment-resisting frame (low rise) . . . . 78 Fig. 4.15: Reinforced concrete shear wall without moment-resisting frame (low rise) . . . . 79 Fig. 4.16: Reinforced conrete shear wall without moment-resisting frame (medium rise) . 79 Fig. 4.17: Reinforced conrete shear wall without moment-resisting frame (high rise) . . . . 80 Fig. 4.18: Braced steel frame (low rise) 80 Fig. 4.19: Braced steel frame (medium rise) 81 Fig. 4.20: Braced steel frame with perimeter frame (low rise) 81 Fig. 4.21: Tilt-Up 82 Fig. 5.1: Southwestern BC Damage Estimation Methodology : 86 Fig. 5.2: Sample Building Inventory Form 96 Fig. 6.1: Predominant Material Type by Block 114 Fig. 6.2: Predominant Building Type by Block 115 Fig. 6.3: BC31 MMI VII Average Mean Damage Factor (%) by Block 116 Fig. 6.4: BC31 MMI Vlll Average Mean Damage Factor (%) by Block 117 Fig. 6.5: BC15 MMI VII Average Mean Damage Factor (%) by Block 118 Fig. 6.6: BC15 MMI Vlll Average Mean Damage Factor (%) by Block 119 VI LIST OF TABLES Table 2.1: Modified Mercalli Intensity 10 Table 3.1: Earthquake Engineering Facility Classifications for Buildings (ATC, 1985) . . 23 Table 3.2: ATC-13 Social Function Facilities 25 Table 3.3: Most Prevalent ATC-13 Classes in British Columbia 26 Table 3.4: BC31 Building Classification 29 Table 3.5: BC15 Building Classification 40 Table 4.1: Structural Damage States in terms of % Replacement Cost 51 Table 4.2: Damage Probability Matrix for Facility Class 1 - Wood frame (low rise) 51 Table 4.3: Example of Structural Damage State Definitions 54 Table 4.4: Comparison of BC Preliminary, ATC21-1 Studies 58 Table 4.5: Damage Probability Matrix for Wood Light Frame Residential (WLFR) 63 Table 4.6: BC15 Mean Damage Factors 63 Table 4.7: BC31 Mean Damage Factors 64 Table 4.8: Reduction of BC31 Classification to BC15 Classification 65 Table 4.9: Comparison of BC Preliminary facility classes to BC15 building types 76 Table 6.1: Example Structural Inventory Database Attributes I l l Table 6.2: Primary Use Classifications 112 Table 6.3: 3Ito 15 Database 113 A CKNO WLED GEMENTS Primary funding of this research was provided by NSERC, to who I am grateful. Partial funding was provided by the Insurance Industry of British Columbia and the City of New Westminster. I am especially grateful to my advisor, Dr. Carlos Ventura, for his guidance and support throughout my graduate studies at the University of British Columbia. I would also like to thank my second reader, Dr. Liam Finn, for his valuable input and guidance during this research. I would like to thank Mr. Leon Bell of Sandwell for his technical expertise in developing the vulnerability functions and training the assessment team. M y appreciation goes to several graduate and undergraduate students for their help in developing the Vancouver and New Westminster building inventories: Ding, Haidan, Ansie, Manuel and Eunice. I would like to thank Tuna Onur and the graduate students of the earthquake lab and structures lab, who helped make my experience at the University of British Columbia enjoyable. I want to especially thank Shane Cook for his continued friendship and his input on every aspect of this research. Finally, I would like to extend my appreciation to my parents and the rest of my family, for their love, wisdom and support throughout both my undergraduate and graduate studies, and to Chris Moyer, for his love and encouragement over the years. CHAPTER 1 Introduction 1.1 Background The uncertainty associated with a thorough understanding of the potentially devastating effects of earthquakes in Southwestern British Columbia is being reflected in higher insurance premiums and deductibles, thus reducing the market for insurance. This could result in significant losses in the province of British Columbia in the aftermath of a severe earthquake. The B C Insurance Industry and the University of British Columbia ( U B C ) have undertaken a joint program in collaboration with Carleton University of Ottawa to estimate the direct monetary losses of a major earthquake on the province. The direct benefit of this study to the insurance industry is a probabilistic risk assessment used to estimate economic loss and better manage property exposure. This study also is useful for municipal planners for earthquake preparedness and emergency response in the region. Methodologies for assessing regional seismic hazard and risk analysis for estimating consequences of an earthquake are well established. The main components of all methodologies are: Earthquake Hazard, Structural Inventory and Loss Prediction. In recent studies, the common tool used to combine these components is a geographic information system (GIS). A GIS integrates cartographic capabilities, a relational database management system ( D B M S ) to store graphic and non-graphic data, and spatial analysis modules 1 Chapter 1 Introduction (Holdstock, 1998). The user must supply the graphic and non-graphic data, and spatial analysis modules. In a risk analysis, the earthquake hazard consists of the combined effects of ground shaking, soil amplification, liquefaction and landslides. This hazard is input as graphical data (hazard maps) into the GIS for a given region. Non-graphic data would include a structural inventory. Given the seismic hazard and a detailed structural inventory, motion-damage relationships are used to calculate estimated damage. 1.2 Objective The main objective of this research is to develop a methodology to evaluate the regional distribution of structural earthquake damage to the southwestern British Columbia building inventory. The steps in the risk analyses addressed in this methodology are: (1) development of the building inventory, (2) selection and application of motion-damage relationships and (3) regional distribution of damage using a geographic information system. 1.3 Scope This research focuses on the review of regional seismic hazard and risk analysis methodologies using geographic information system technology. While most methodologies are applicable to any region, this thesis w i l l focus on customizing a methodology for the general building stock of Southwestern British Columbia. The seismic risk analysis primarily considers hazards due to ground shaking, while effects of local site conditions are not included at this time. Structural classifications and motion-damage relationships are developed specific to the region. Finally, a methodology for estimation of building damage 2 Chapter 1 Introduction due to ground shaking and a case study is presented. 1.4 Organization of Thesis This thesis presents a methodology for conducting a regional seismic risk analysis specific to the building inventory in Southwestern British Columbia. Chapter 2 is a literature review which gives an overview of existing methodologies, motion-damage relationships and development of building damage functions. Chapter 3 reviews existing classifications and develops a structural classification system suitable for the study region. A brief description is given for each structural type in the final classification system. Chapter 4 develops motion-damage relationships appropriate for use in the study region. Chapter 5 explores the implementation of a methodology to the study region. Discussion focuses on the development of a structural inventory and performing the damage analysis using a geographic information system. Chapter 6 presents a case study. A damage analysis was performed for the City of New Westminster, British Columbia. A description is given of the structural inventory development. The motion-damage relationships and classification systems of Chapters 4 and 5 were then applied to the inventory using the GIS software, Maplnfo®. Results of the analysis are presented. Finally, Chapter 7 presents conclusions and recommendations for future work. 3 CHAPTER 2 Literature Review 2.1 Overview At an international level, the past few decades have seen an increasing number of regional hazard and risk analyses being performed. Recent earthquakes in Japan and the United States have shown the tremendous losses, both social and economic, that countries must sustain in the aftermath of an earthquake. On a larger scale, losses of billions of dollars to any one country's economy have a ripple effect on the global economy. Concerned government and private agencies are realizing the benefit of accurate loss estimates. The common purpose of all these studies is to analyze the risks and assess mitigation alternatives. Besides hazard mitigation, they are useful for response, recovery and emergency preparedness planning. By the mid 1980's, a number of methods had been developed with the aim of estimating the damage or effects of future earthquakes (Reitherman, 1985). The scope and nature of these methodologies vary greatly depending on the end-user. This review is limited to those methodologies whose intent is the estimation of regional damage and not the prediction of expected damage and loss at a specific site or for a specific facility. Chapter 2 Literature Review 2.2 Regional Seismic Hazard and Risk Analysis An overview of the steps in regional seismic hazard and risk analysis procedure are illustrated in Figure 2.1. The minimum input data required to perform a regional damage estimate is the seismic hazard and a structural inventory. Applying motion-damage relationships to this information, the seismic risk to the structural inventory from the hazard can be estimated. The seismic risk is quantified in terms of damage, thus some literature will interchange regional damage estimation with risk analysis. It is commonly measured as a percentage of replacement cost. If the replacement costs are known then the estimated damage can be transformed into a monetary loss estimation, measured in dollars. The impacts of an earthquake are not only measured in monetary amounts, there are also non-monetary losses which will be discussed later. The following paragraphs, based on King and Kiremidjian (King, 1994), discuss each of these steps in more detail. The first part of any loss estimation is determining the seismic hazard. Seismic hazard analyses estimate the amount of ground shaking at a particular site. Seismic hazards may be analyzed deterministically as when a particular earthquake scenario is assumed or probabilistically in which uncertainties in earthquake size, location and time of occurrence are explicitly considered. In either case, possible earthquake sources that can affect the region are identified. A deterministic scenario analysis models the exact size, location and time of occurrence of an earthquake, either being a repeat of a historical earthquake or the maximum earthquake that the given seismic source is capable of generating. A probabilistic scenario models the occurrence of earthquakes on these sources based on one of several methods. In 5 Chapter 2 Literature Rev iew general, these models either have spatial or temporal dependence, meaning they consider the history of seismicity of each source. These methods are useful in modeling the behavior of large, rare earthquakes (King, 1994). To model small, frequent earthquakes, a poisson model is used which assumes earthquakes occur randomly in time, space and magnitude (Cornell, 1968). A complete review of these methods is not presented here. Once the expected behavior of the fault is understood, then the bedrock motion in the region of interest may be determined. The following question still remains. How does the motion at the source travel to the region of interest? The most common method of determining this involves the use of empirical attenuation relationships. These relationships are usually specific to ground motion, magnitude, distance and site conditions. In this thesis, the seismic hazard as measured from the bedrock will be referred as the ground shaking hazard. Obviously, the entire building inventory is not located on bedrock. Local site effects must also be considered. From wave propagation theory, we know that the motion at the bedrock level is amplified as it moves to the ground surface due to local soil conditions. Besides soil amplification, other local site effects are liquefaction, landslide and surface fault rupture. The ground shaking hazard with local site effects all contribute to seismic hazard. Once the seismic hazard is known, it is then necessary to develop a structural inventory of the region. That inventory typically includes buildings but could also be expanded to include lifelines and other non-building structures. It is the end-user who decides which facilities are to be studied. It has been found that the more detailed the inventory the 6 Chapter 2 Literature Review more accurate the loss estimation, though development of a regional inventory is time-consuming and expensive (Reitherman, 1985). Many sources give methodologies for the compilation of inventories (Vasudevan, 1992, ATC-13, 1985). For the purpose of this thesis, we wil l limit discussion to development of a comprehensive building inventory for Southwestern British Columbia. This discussion begins in Chapter 3. Once the inventory is complete and the seismic hazard calculated, then the damage to the building inventory can be ascertained. Damage can be divided into both structural and non-structural facility damage. In order to quantify the damage, motion-damage relationships are developed. These relationships define the extent of damage to a specific facility type given the estimated ground shaking hazard. There are numerous ways to measure seismic hazard and an equal number of definitions of damage. These relationships will be explored in more detail in Chapter 4. After completion of a regional damage estimate, it is then possible to determine expected losses in the region. Types of loss can be loosely divided into two categories: Economic Loss and Social Loss. In the literature they are sometimes referred to as Monetary and Non-monetary loss. Monetary losses stem from structural damage, non-structural damage, loss of business revenue, relocation expenses, contents and financing of repairs. Most building damage estimates typically will include costs associated with structural damage. Non-monetary loss estimates usually include casualties and loss of function. With all the steps in a regional hazard and risk analysis identified, this thesis will only focus on regional building damage estimation. Development of a classification system, 7 Chapter 2 Literature Review motion-damage relationships and performing a structural damage analysis will be further discussed. A seismic hazard analysis for the region has been performed and the results are available for use but the details of that analysis are only briefly discussed in this study. A loss estimation will not be performed at this time. Seismic Hazard Analysis Structural Inventory Data ) Ground Motion Local Site Effects > I Motion Damage Relationships Regional Damage Distribution i Repair and Replacement Cost Model i Regional Loss Distribution Fig. 2.1: Regional Seismic Hazard & Risk Analysis 2.3 Definitions of Damage Damage is the "physical impact of an earthquake on a facility (FEMA-249, 1994)." There are several ways to measure damage. When looking at a specific structure, damage can be measured with a "damage index". Damage indices characterize local or global damage 8 Chapter 2 Literature Review based on response parameters such as ductility ratio, inter-story drift, and dissipated energy (King, 1994). Typically, damage indices are structure specific, and therefore inappropriate for regional damage estimation. Early loss studies measured damage with a "damage ratio." This is defined as the number of structures damaged divided by the total number of structures. The most widely used measure of damage is the "damage factor." It is the ratio of dollar loss to replacement cost. The mean damage factor for a group of similar structures is defined as: n MeanDamageF actor (MDF) = - • X (,DollarLoss)i n (ReplacementValue), i = 1 where n is the number of structures in the sample (ATC-13, 1985). Motion-damage relationships are probabilistic distributions of damage, typically measured in either damage factor or damage ratio, at specified ground motion intensities. They are commonly expressed using damage-loss curves, fragility curves, damage probability matrices or expected damage factor curves. Damage-loss curves show mean damage ratios as a function of ground shaking intensity for various building types. Fragility curves describe the probability that a specified damage level will be exceeded for a given intensity of ground motion. Damage probability matrices (DPMs) show the probability that a structure is in a specified damage state given the level of ground shaking intensity. Expected damage factor curves display the same information as the DPMs or fragility curves. 9 Chapter 2 Literature Review 2.4 Characterization of Ground Motion Besides quantification of damage, another factor that differentiates motion-damage relationships is the measure of seismic intensity. Attenuation relationships define the behavior of ground motion as it travels from the source to site. These give results in terms of peak ground acceleration (PGA) , peak spectral response velocity (PRV) or in terms of an intensity using a subjective scale. Various scales are in usage around the world, common ones are: Modif ied Mercal l i Scale ( M M ) , Japan Meteorological Agency Scale ( J M A ) Medvedev-Sponheuer-Karnik Scale ( M S K ) , Rossi-Forrell Scale (RF) and the G E O F I A N scale. The subjective scale to be used in this study is the Modified Mercal l i Scale. The severity o f ground motion using this scale increases continuously from M M I I through M M I XI I . Figure 2.1 describes a portion of the scale for which significant damage occurs, M M I V I through M M I XII . (ATC-13 , 1985) Table 2.1: Modified Mercalli Intensity MMI Description of Effects V I Felt by all, many frightened. Some heavy furniture moved. A few instances of fallen plaster. Damage slight. VII Damage negligible in buildings of good design and construction; slight to moderate in well-built ordinary structures; considerable damage in poorly built structures. Some chimneys broken. VIII Damage slight in specially-designed structures; considerable damage in ordinary substantial buildings with partial collapse. Damage great in poorly-built structures. Fall of chimneys, factory stacks, columns, monuments, walls. Heavy furniture overturned. IX Damage considerable in specially-designed structures; well-designed frame structures thrown out of plumb. Damage great in substantial buildings, with partial collapse. Buildings shifted off foundations. X Some well-built wooden structures destroyed; most masonry and frame structures with foundations destroyed. Rails bent. X I Few, i f any masonry structures remain standing. Bridges destroyed. Rails bent greatly. XII Damage total. Lines of sight and level are distorted. Objects thrown into air. 10 Chapter 2 Literature Review 2.5 Development of Building Damage Functions Although damage functions can be expressed in several forms, the development of these relationships can be loosely divided into two groups: expert opinion and engineering parameters. Expert opinion based damage functions rely on the judgement and expertise of local engineers to derive motion-damage relationships. Engineering parameter studies quantify damage in buildings based on the seismic behavior of a building. Generally, empirical data and computational models are used to predict damage for each specific facility class. The trend in current research has been towards the use of engineering parameter studies as this is considered a more rational approach to loss estimation. 2.5.1 Expert Opinion In 1985, the Applied Engineering Technology Council presented a report entitled ATC-13, Earthquake Damage Evaluation Data for California. An advisory project engineering panel provided the expertise necessary to develop consensus damage functions in the form of damage probability matrices. The project identified the following primary tasks: 1. Identification of the most appropriate ground motion characterization for estimating earthquake damages and losses. 2. Development of a comprehensive building classification for California. 3. Development of damage and loss estimates based on the facility classes and ground shaking characterization. 4. Development of inventory data and methodology that are associated with the facilities and their social function class. It is a significant contribution to the field of damage and loss estimation and a detailed review of this study is performed in Chapter 4. 11 Chapter 2 Literature Review Since then, this data has been applied to various regions. Technology advancements since that time have resulted in a methodology to perform a regional loss estimate in a GIS environment (King, 1994). Case studies evolving from this study have been performed in regions including Quito, Ecuador (Fernandez, 1996), Portland, Oregon (Rad and McCormack, 1997), Salt Lake County, Utah (Rojahn, 1997) and Palo Alto, California (King, 1997). Rad and McCormack proposed a methodology for Portland, Oregon (1997). Their study surveyed the seismic hazards for 4,500 non-residential buildings in Portland. The goal was to develop an earthquake damage and loss estimation model for buildings to estimate the potential earthquake damage and loss for the surveyed buildings and also to perform a casualty estimate. This study applied the ATC-21 Rapid Screening Procedure for 12 building types. Performance Modifiers were combined with the Basic Structural Hazard to give a Structural Score for each building surveyed. Using ATC-13, the structural score was transformed into a damage factor, 60% being the accepted threshold of major damage. An ongoing project of the Applied Technology Council is the translation and update of ATC-13 damage matrices for application in Salt Lake County, Utah (Rojahn, 1997). The project, known as ATC-36, is a comprehensive methodology for earthquake damage and loss estimation. In this project, damage functions are in the form of expected damage factors versus MMI. They were developed by expert opinion by modifying the ATC-13 DPMs for Utah facility class construction, and by mapping between the classes in Utah and those defined in ATC-13. The methodology defined by ATC-25 (1991) was used as a guide in the translation of the damage functions. ATC-25 is a methodology developed for translating ATC-13 lifeline motion-damage relationships to regions outside of California. 12 Chapter 2 Literature Review Simultaneous research has been focused on assessing seismic hazard and vulnerability in the Central United States. A pilot study for Memphis, Tennessee was conducted by Shinozuka, et. al. (1997). The study emphasized the synthesis of local data in developing a loss methodology. Damage functions are in the form of fragility curves. The development of these curves varies in the use of computational models and expert opinion based on the available data. A report entitled "Expected Seismic Performance of Buildings" was presented by the EERI Ad Hoc Committee on Seismic Performance in 1994 (EERI, 1994). This study focuses on damage due to ground shaking in Zone 4 of the 1991 Uniform Building Code. Damage functions in the form of damage probability matrices are derived using expert opinion for eight building types. The ground shaking in this study is characterized using MMI. In an effort to improve the motion-damage relationships of ATC-13, a report "NCEER-ATC Joint Study on Fragility of Buildings" was presented in 1995 (Anagnos, et. al, 1995). In this study, the motion-damage functions of ATC-13 were adjusted in the following ways: (1) providing more detailed descriptions of the 40 building facility classes, (2) modifying the damage functions based on data gathered from recent earthquakes, and (3) developing fragility curves based on the ATC-13 damage probability matrices. This study was in response to deficiencies identified in ATC-13 in the decade since it's release, and to an increased amount of seismic activity in the region from which experimental data could be used to calibrate the original estimates. An important conclusion to the study was that "Damage data that are being collected after earthquakes in most cases do not include the types of information that is needed to update or develop damage motion relationships." This clearly 13 Chapter 2 Literature Review identifies the need for a systematic form of data collection following an event. 2.5.2 Engineering Parameter Models There have been numerous studies which derive damage functions based on expert opinion. Generally, this is because there is very limited data to calibrate these functions to actual observed damage, so expert opinion is used. An alternative method has been explored in the form of engineering parameter models (Thiel and Zsutty, 1987; Liu and Jones, 1996; RMS and CUREe, 1993). These models use engineering parameters to develop more reliable damage functions. By comparing building capacity data to demand spectra, damage functions are developed. Several models have been proposed which use ductility ratio, interstory drift or hysteresis energy dissipation as measures of damage (RMS and CUREe, 1993). The difficulty with this type of methodology has been the limited data available and the computational manpower required to determine building behavior based on structural characteristics. One commonly used methodology was developed by Thiel and Zsutty in 1987 (Thiel and Zsutty, 1987). This methodology models the damage for specific building types based on parameters that represent the building's structural system, the site and building periods, site materials and site acceleration. A simple Markov distribution relates the probability of occurrence of discrete damage states for a specific building type. The distribution and necessary parameters were chosen based on observed building damage from several California and Chinese earthquakes. At the time of development, however, sufficient damageability data was only available for brick and wood frame structures. In order to develop damageability models for a broad range of structural systems, the study then used the 14 Chapter 2 Literature Review results of ATC-13 to determine building parameters for those structural types which lacked sufficient damage data from previous earthquakes. In doing so, the study introduced the results of an expert opinion based study into what was originally intended to be a methodology based on site and building parameters. An important result of this study is that it is possible to quantify damage based on previous earthquake performance of structures using engineering parameters provided there is sufficient damageability data for the structure type. A natural progression from the methodology by Thiel and Zsutty (1987) is to determine the building parameters with computer modeling of structural performance as exemplified in current research. In 1997, a comprehensive damage and loss estimation methodology was presented by the Federal Emergency Management Association (FEMA) and the National Institute of Building Science (NIBS) which relies on engineering parameters to develop damage functions (HAZUS, 1997). Motion damage relationships for both non-structural and structural damage are expressed as the probability that a specified damage level will be exceeded for a given intensity of ground motion in the form of lognormal fragility curves. It has been described as a "significant step forward in the prediction of earthquake impacts." (Kircher, 1997) These functions use quantitative measures of ground shaking and building response. This is a landmark study in the field of loss estimation and a detailed review is presented in Chapter 4. A study by Liu and Jones (1996) predicts statistical distribution of damage factor for a class of building in an inventory by using simulated time histories of ground motion coupled 15 Chapter 2 Literature Review with a realistic structural model. Advantages include: several key physical modules are used to calculate the building damage, uncertainty at each step can be incorporated into the model in a statistically consistent manner. The structural modeling is performed by converting multi degree of freedom structures to a single degree of freedom by considering only their first mode. A damage factor formula is based on pre- and post-yielding stiffness, ductility and hysteretic energy. A paper by Singhal and Kiremidjian (1996) presents a method for estimating probability of damage by "quantifying the response of a structure subjected to a significant ensemble of ground motion with a wide range of parameter variations." It varies engineering parameters needed for a nonlinear dynamic analysis in a Monte Carlo simulation. For each building type, damage indices based on ductility ratio or dissipated energy are computed using the results of time-history analyses. Cumulative distribution function vs. damage index for a given spectral acceleration is plotted and fitted with lognormal distribution. These are then used to obtain probabilities of reaching or exceeding a damage state, presented in the form of fragility curves. 2.6 Earthquake Loss Estimation in British Columbia Southwestern British Columbia is located in the most earthquake-prone region of Canada. Evidence of historic earthquakes causing regional damage have been reported as far back as 1700. Such high seismicity can be attributed to the Cascadia subduction zone located to the west of Vancouver Island, extending from the north tip of the Island to northern California, and also from shallow crustal earthquakes occuring in the continental crust. The Geological Survey of Canada attributes activity in the Cascadia subduction zone for "the 300 16 Chapter 2 Literature Review or so small earthquakes that are located in southwestern British Columbia each year, and the less-frequent (once per decade, on average), damaging crustal earthquakes (e.g., a magnitude 7.3 earthquake on central Vancouver Island in 1946)" (GSC, 1999). With such high seismicity, the risk of economic loss to southwestern British Columbia in the event of an earthquake is high. The insurance industry has taken the lead in determining the effects of a major earthquake in this region. Two studies which have addressed the economic impact to the region are reviewed here. The first study was conducted by the National Research Council of Canada for the Canada Mortgage and Housing Corporation (CMHC) (NRC, 1989). The study is entitled "Assessment of Earthquake Effects on Residential Buildings and Services in the Greater Vancouver Area." The ground shaking hazard from two separate earthquakes are considered: 1) the "design level earthquake" of the National Building Code of Canada, 1985 ed.; and 2) a major subduction earthquake along the west coast of British Columbia. Damage estimates for the following structure types are predicted: • single family homes of wood frame construction, one or two storeys; • unreinforced masonry buildings; • low and medium-rise buildings, i.e. 2 to 4 story residential buildings or various small buildings like banks and small shopping area; 5 to 9 story residential and office buildings; • high-rise buildings, i.e. residential and office buildings > 10 storeys; • schools and hospitals. Estimates are also made for lifelines such as gas and water supply, electricity, communication systems, transportation routes, harbour facilities and airport structures. A simplified prediction of earthquake performance is made based on the performance of similar structures 17 Chapter 2 Literature Review in previous events in California, Alaska and Washington. The study presents results in terms of loss ratio for major structural damage based on the best estimate of the authors and do not reflect a statistical analysis of vulnerability of the regional building or lifeline inventory. The second study was performed by Munich Re-Insurance of Canada and is entitled "A Study of the Economic Impact of a Severe Earthquake in the Lower Mainland of British Columbia." (Munich Re-Insurance, 1990). The scenario earthquake evaluated is an earthquake of magnitude 6.5, at longitude 123W, latitude 49N and depth 10 km in the Strait of Georgia. Several submodels and corresponding analytical methodologies compose the model for economic impact. These modules are: indirect and direct building damage due to ground shaking; fire; offsite damage; inundation; landslides; onsite injury and loss of life; and infrastructural losses. The direct damage to buildings module presents results in terms of total replacement cost (in dollars) and utilizes two analytical methodologies. The first estimates losses based on total square footage in the GVRD for residential, commercial and industrial construction. The second, complementary methodology extrapolates losses for a detailed inventory of the downtown Vancouver core/West End to estimate losses for the rest of the region. Both methodologies utilize loss ratios aggregated by residential, commercial and industrial construction for MMI VII-VIII. These ratios are based on Munich Re-Insurance's assessment of damage generated by several recent large earthquakes, especially Mexico City in 1985. 2.7 Conclusion The literature reviewed in this chapter is just a portion of the extensive literature available regarding seismic hazard and risk analysis. In reviewing literature pertaining to 18 Chapter 2 Literature Review specifically the analysis of seismic risk to a regional building inventory, it is important to be familiar with not only the steps in the analysis but also the many assumptions made in performing the analysis. Different definitions of damage, characterization of ground motion and the development of motion-damage relationships are significant assumptions which can vary between different studies. These variables were chosen for discussion in this chapter to set the background for development of a methodology for southwestern British Columbia. Similarly, the previous studies by Munich Re-Insurance (1990) and the Canada Mortgage and Housing Corporation (1989) are examples of methodologies that have been found applicable to the region. Both studies acknowledge the lack of a detailed study of seismic performance of representative types of buildings in the GVRD. They show the importance of loss estimation to the economy in southwestern British Columbia and provide some indication of the advantages and limitations of a comprehensive damage estimation methodology. 19 CHAPTER 3 Structural Classification 3.1 Overview An inventory database is an essential component of any damage estimation. It describes the "physical and financial exposure subject to direct and indirect earthquake loss" (FEMA-249, 1993). As the accuracy of the inventory improves, the results of the damage estimation become more meaningful. Often, development of the inventory is the most difficult aspect of damage estimation. Ideally, a building-by-building in-depth structural analysis would be performed on each structure to develop a damage estimate for the regional inventory. Realistically, database development is limited by the time and resources available. These constraints neccesitate the use of existing databases, except the building information needed for damage estimation can be incomplete or inaccurate. Inference schemes are then used to determine missing information. Classification systems categorize the general building stock into groupings with similar risk characteristics. In general, buildings behave differently according to the strength and weight of their construction material, height, configuration, design and construction quality, and age (ATC-13, 1985). Therefore each structure has a unique reaction to ground motion based on these characteristics. Classification systems group buildings to represent the average characteristics of buildings in that class. Damage functions relate estimated damage 20 Chapter 3 Structural Classification to ground motion for each grouping. Therefore, each grouping must have a corresponding motion-damage relationship. There are two major types of classification systems, those based on construction class and those based on building usage (e.g. residential, commercial, industrial). Loss studies can use one or both types of systems depending on the purpose of the study. Studies based on structural type are useful for calculating structural damage, casualties and loss of function. Building usage, or occupancy type, is important in determining economic loss (Kircher, 1997). A classification of the inventory based on both occupancy and structural type yields a more comprehensive loss estimation. Due to differences in regional construction and resource availability there is no single standard classification system. While many systems exist, the most commonly used system was developed from fire insurance methodologies by Steinbrugge in 1982 (FEMA-249, 1993). This study was performed for the Insurance Services Office and 19 classes were chosen so the inventory could be easily categorized from available data. This study is widely recognized as the seminal classification system from which many systems have evolved. In determining classification systems the goal is to choose enough types to cover the entire building inventory. More recent classification systems have been derived from the NEHRP Handbook for the Seismic Evaluation of Existing Buildings (BSCC, 1997). The FEMA-NIBS methodology (FEMA, 1997) expands on the NEHRP classifications to include mobile homes, and building height. In this study 36 structural types and 28 occupancy classes are used. 21 Chapter 3 Structural Classification 3.2 ATC-13 While Steinbrugge only classified the general building inventory, a more detailed classification system was developed for ATC-13. Developed in 1985, this study contains 78 classes of structures, 40 of which are buildings. It was the first comprehensive inventory classification for both structural type and occupancy class not limited to the general building stock. Structural type classes are referred to as "Earthquake Engineering Facilities." The classes were chosen based on their prevalence in the California structural inventory and also because of their unique seismic risk characteristics. The buildings fall into 17 construction categories according to structural framing type and structural material. They can be further categorized into 40 classes if height (number of stories) is included. Building classes are listed in Table 3.1, the entire classification is available in ATC-13 (1985). There are 35 occupancy classes, referred to as "Social Function Facilities." Broad classes are: residential, commercial, and industrial buildings, critical facilities and lifelines. These classes were selected to conform to existing databases and to correlate to those used by the U. S. Department of Commerce. Table 3.2: ATC-13 Social Function Facilities lists these. 22 Chapter 3 Structural Classification Table 3.1: Earthquake Engineering Facility Classifications for Buildings (ATC, 1985) Facility No. Wood Frame (Low Rise) 1 Light Metal (Low Rise) 2 Unreinforced Masonry (Bearing Wall) Low Rise (1-3 Stories) 75 Medium Rise (4-7 Stories) 76 Unreinforced Masonry (with Load Bearing Frame) Low Rise 78 Medium Rise 79 High Rise (8+ Stories) 80 Reinforced Concrete Shear Wall (with Moment-Resisting Frame) Low Rise 3 Medium Rise 4 High Rise 5 Reinforced Concrete Shear Wall (without Moment-Resisting Frame) Low Rise 6 Medium Rise 7 High Rise 8 Reinforced Masonry Shear Wall (without Moment-Resisting Frame) Low Rise 9 Medium Rise 10 High Rise 11 Reinforced Masonry Shear Wall (with Moment-Resisting Frame) Low Rise 84 Medium Rise 85 High Rise 86 Braced Steel Frame Low Rise 12 Medium Rise 13 High Rise 14 Moment Resisting Steel Frame (Perimeter Frame) Low Rise 15 Medium Rise 16 High Rise 17 Moment Resisting Steel Frame (Distributed Frame) Low Rise 72 Medium Rise 73 High Rise 74 23 Chapter 3 Structural Classification Table 3.1: Earthquake Engineering Facility Classifications for Buildings (ATC, 1985) Moment Resisting Ductile Concrete Frame (Distributed Frame) Low Rise 18 Medium Rise 19 High Rise 20 Moment Resisting Non-Ductile Concrete Frame (Distributed Frame) Low Rise 87 Medium Rise 88 High Rise 89 Precast Concrete (other than Tilt-up) Low Rise 81 Medium Rise 82 High Rise 83 Long-Span (Low Rise) 91 Tilt-up (Low Rise) 21 Mobile Homes 23 24 Chapter 3 Structural Classification Table 3.2: ATC-13 Social Function Facilities Residential Permanent Dwelling Temporary Lodging Group Institutional Housing Commercial Retail Trade Wholesale Trade Personal and Repair Services Professional, Technical and Business Services Health Care Services Entertainment and Recreation Parking Industrial Heavy Fabrication and Assembly Light Fabrication and Assembly Food and Drugs Processing Chemicals Processing Metal and Minerals Processing High Technology Construction Petroleum Agriculture Mining Religion and Non-Profit Government General Services Emergency Response Services Education Transportation Services (Freight and Passenger) Highway Railroad Air Sea/Water Utilities Electrical Water Sanitary Sewer Natural Gas Telephone and Telegraph Communication Flood Control 25 Chapter 3 Structural Classification 3.3 Preliminary BC Survey In the fall of 1996, the University of British Columbia conducted a preliminary study of seismic vulnerability to the general building stock (Ventura and Rezai, 1997). Using A T C -13 as the benchmark, the study conducted a survey of local engineers. Based on the results of the survey and a review of technical literature, an assessment was made of the vulnerability of different types of building in BC. The survey used the same Earthquake Engineering Facility Classification as ATC-13. Participants were asked to provide their opinion on the prevalence or number of facilities that currently exist in the province of British Columbia. The number of occurrences was categorized as low (0 to 10 occurrences), medium (10 to 50 occurrences) or high (>50 occurrences). Facilities were chosen for further statistical analysis based on histograms of prevalence data for each class. Of the 40 original building classes, 11 were chosen as the most prevalent ATC-13 classes in British Columbia. Those classes are shown in Table 3.3. Table 3.3: Most Prevalent ATC-13 Classes in British Columbia Wood frame (low rise) Unreinforced masonry with bearing wall (low rise) Unreinforced masonry with load bearing frame (low rise) Reinforced concrete shear wall without moment-resisting frame (low rise) Reinforced concrete shear wall without moment-resisting frame (medium rise) Reinforced concrete shear wall without moment-resisting frame (high rise) Reinforced masonry shear wall without moment-resisting frame (low rise) Braced steel frame (low rise) Braced steel frame (medium rise) Braced steel frame with perimeter frame (low rise) Tilt-up (low rise) 26 Chapter 3 Structural Classification 3.4 Southwestern BC Building Seismic Fragility Study In a follow-up study completed in 1998, a local expert on earthquake engineering and building construction was contracted to develop a classification system and damage probability matrices for Southwestern British Columbia (Bell, 1998). An overview of that study is presented here. A classification system including 31 building types was developed from ATC-13 to encompass the local building inventory (Bell, 1998). The study gives a detailed description for each building type in terms of system description, prevalence, examples, identifying features, expected performance and general comments. Section 3.4.1 includes a brief summary of engineering characteristics for the BC31 building types. During a preliminary application of the BC31 classification system, it became apparent that due to limited resource availability, data collection would be simplified if the number of building types could be reduced. The resulting system includes 15 building types and was accomplished by the following (Bell, 1998): 5. Deleting those building types that are not represented in the BC building inventory. 6. Combining medium and high rise construction into one building type. 7. Combining building types that constitute insignificant portions of the inventory with building types that exhibit similar seismic behavior. 8. Combining building types that exhibit similar seismic behavior. This reduction does not significantly affect the global conclusion of the damage estimation to the building inventory, as is shown in the case study presented in Chapter 6. Section 3.4.2 27 Chapter 3 Structural Classification includes Table 3.5: BC15 Classifications, and a brief summary of engineering characteristics for the BC15 building types. Included in the general description of each building type is the combination of BC31 classes that produce the BC15 classification. 3.4.1 BC31 Building Classification Table 3.4 lists the 31 building types chosen for this study. They are categorized according to structural framing type, construction material and building height and where applicable have been cross-referenced to their corresponding ATC-13 facility class. Some building types (marked "n/a" in Table 3.4) do not correspond to any of the ATC-13 facility classes due to the decision of the local expert to expand the classification system to include all building types predominant in Southwestern British Columbia and which may not have been prominent in the California building inventory during the development of ATC-13. Alternately, some building types correspond to more than one ATC-13 facility class where the behaviour of the BC31 building type is similar to each of the ATC-13 facility classes listed. Also in this section is a detailed description of the 31 building types. It should be noted that all descriptions are made for average construction. Inherent in the definition of average construction are the following assumptions (Bell, 1998): 1. The building type is very nearly regular in shape. Any arms such as T, E , F , C, etc. are very small, being less than 20% of the length of the core structure. 2. The building type was designed and constructed to meet the quality of work at the time. 3. Buildings were designed to a code prior to 1990. 28 Chapter 3 Structural Classification Table 3.4: BC31 Building Classification Material BC31 Building Type BC31 ATC-13 Wood Wood Light Frame, Residential WLFR 1 Wood Light Frame, Commercial/Institutional WLFCI 1 Wood Light Frame Low Rise W L F L R 1 Wood Post and Beam WPB 1 Steel Light Metal Frame L M F 2 Steel Moment Frame Low Rise SMFLR 15,72 Steel Moment Frame Medium Rise SMFMR 16,73 Steel Moment Frame High Rise SMFHR 17,74 Steel Braced Frame Low Rise SBFLR 12 Steel Braced Frame Medium Rise SBFMR 13 Steel Braced Frame High Rise SBFHR 14 Steel Frame Concrete Walls Low Rise SFCWLR 6 Steel Frame Concrete Walls Medium Rise SFCWMR 7 Steel Frame Concrete Walls High Rise SFCWHR 8 Steel Frame with Concrete Infill Walls SFCI n/a Steel Frame with Masonry Infill Walls SFMI n/a Concrete Concrete Frame with Concrete Walls Low Rise C F L R 6 Concrete Frame with Concrete Walls Medium Rise CFMR 7 Concrete Frame with Concrete Walls High Rise CFHR 8 Reinforced Concrete Moment Frame Low Rise R C M F L R 18,87 Reinforced Concrete Moment Frame Medium Rise R C M F M R 19,88 Reinforced Concrete Moment Frame High Rise RCMFHR 19,89 Reinforced Concrete Frame with Infill Walls RCFIW 79 Masonry Reinforced Masonry Shear Wall Low Rise R M L R 9,84 Reinforced Masonry Shear Wall Medium Rise R M M R 10,85 Unreinforced Masonry Bearing Walls Low Rise U R M L R 78 Unreinforced Masonry Bearing Walls Medium Rise U R M M R 79 Tilt Up Tilt Up T U 21 Precast Precast Concrete Low Rise PCLR 81 Precast Concrete Medium Rise PCMR 82 Mobile Mobile Homes M H 23 29 Chapter 3 Structural Classification 1. Wood Light Frame Residential fWLFRt General Description: One or two storey, single family detached homes between 70 to 350 square meters. Wall systems are wood stud, roof/floor span systems are wood joist and rafter construction. Exterior walls are sheathed with wood or vinyl siding, stucco, plaster, brick or metal. Typically has many interior walls sheathed with plaster. A common detail is an unbraced cripple wall connection to concrete or brick foundation. 2. Wood Light Frame Low Rise Commercial/Institutional (WLFCI) General Description: One or two storey commercial and institutional buildings between 80 to 600 square meters. Wall systems are wood stud, roof/floor span systems are wood joist or wood truss construction. Exterior walls are sheathed with wood or vinyl siding, stucco, plaster, brick veneer or light metal. Typically has extensive areas of glazing creating a storefront. 3. Wood Light Frame Low Rise Residential (WLFLR^ General Description: Low Rise (up to 4 stories) wood frame multi-family residential buildings up to 1500 square meters. Wall systems are wood stud, roof/floor span systems are wood joist and rafter construction. Exterior walls are sheathed with wood or vinyl siding, stucco, brick veneer or metal. Typically has many interior walls sheathed with plaster. Since late 1970's, common to have underground parking and/or concrete retail level. 4. Wood Post and Beam fWPB) General Description: One or two storey commercial/institutional structures and some homes of the 50's and 60's. Roofs are wood frame on wood beams or trusses. Floor span 30 Chapter 3 Structural Classification systems are wood joist, tongue and groove decking, or laminated wood studs spanning onto timber beams. Lateral loads are transferred by roof and floor sheathing to wood clad walls, steel or wood cross bracing, or wood laminated or masonry stairwells. 5. Light Metal Frame (LMFl General Description: Lightweight pre-engineered industrial and agricultural buildings with rigid frame in the short direction and tie-rod cross-bracing in the other. Roof is lightweight steel purlins. Building typically has an open interior and exterior walls are clad with metal decking though some use asbestos board. 6. Steel Moment Frame Low Rise (SMFLR) General Description: Steel moment frames in one or both directions form perimeter or are clustered; 1 to 3 storeys in height. Roof is generally metal decking. Floors are concrete infilled metal decking on open webbed steel joists or reinforced concrete slabs. Exterior walls are metal studs and metal siding, brick veneer siding or precast concrete panels. Extremely rare construction type in southwestern B C . 7. Steel Moment Frame Medium Rise (SMFMR) General Description: Steel moment frames in one or both directions form perimeter or are clustered; 4 to 7 storeys in height. Roof is generally metal decking. Floors are concrete infilled metal decking on open webbed steel joists or reinforced concrete slabs on steel beams. Exterior walls are generally brick veneer siding with extensive glazing. Extremely rare construction type in southwestern B C . If the structure includes concrete or masonry stairwells, building type should be considered as S F C W M R . 31 Chapter 3 Structural Classification 8. Steel Moment Frame High Rise (SMFHRJ General Description: Steel moment frames in one or both directions form perimeter or are clustered; over 8 storeys in height. Roof is generally metal decking. Floors are concrete infilled metal decking on open webbed steel joists or reinforced concrete slabs onto steel beams. Exterior walls are generally brick veneer siding with extensive glazing. Extremely rare construction type in southwestern BC. 9. Steel Braced Frame Low Rise (SBFLR) General Description: Steel framed structure with cross or chevron bracing in both directions; 1 to 3 storeys in height. Roof is generally metal decking but can be wood or concrete. Floors are concrete infilled metal decking on open webbed steel joists but can be wood or reinforced concrete slabs spanning onto steel beams. Exterior walls are clad in metal or wood siding, stucco or masonry veneer. If the structure includes concrete walls, building type should be considered as SFCWLR. 10. Steel Braced Frame Medium Rise (SBFMR) General Description: Steel framed structure with cross or chevron bracing in both directions; 4 to 7 storeys in height. Roof is generally metal decking. Floors are concrete infilled metal decking on open webbed steel joists but can be wood or reinforced concrete slabs spanning onto steel beams. Exterior walls are clad in marble or masonry veneer. If the structure includes concrete or masonry stairwells, building type should be considered as SFCWMR. 32 Chapter 3 Structural Classification 11. Steel Braced Frame H i g h Rise ( S B F H R ) General Description: Steel framed structure with cross or chevron bracing in both directions; greater than 8 storeys in height. Roof and floors are generally concrete infilled metal decking on open webbed steel joists or reinforced concrete slabs spanning onto steel beams. I f the structure includes concrete or masonry stairwells, building type should be considered as S F C W H R . 12. Steel F rame with Concrete Wal ls L o w Rise General Description: Steel framed structure with concrete shear walls providing lateral resistance in one or both directions; 1 to 2 storeys in height. Roof is generally metal decking but can be wood. Floors are concrete infilled metal decking on open webbed steel joists. Exterior walls are clad in metal, stucco or masonry veneer. Shear walls are located interior (i.e., around stair and elevator shafts) or peripherally. Also included in this category is steel moment frame construction with concrete shear walls. In this case, the shear wall is the much stiffer structural element and w i l l carry most of the lateral loads (at least 75%). 13. Steel F rame with Concrete Wal l s M e d i u m Rise ( S F C W M R ) General Description: Steel framed structure with concrete shear walls providing lateral resistance in one or both directions; 3 to 7 storeys in height. Generally, roof is metal decking. Floors are infilled metal decking on open webbed steel joists or reinforced concrete floors spanning onto steel beams. Exterior walls are clad in steel panels, precast panels, masonry or marble veneer or extensive glazing. Shear walls can be located interior (i.e., around stair and elevator shafts) or peripherally. Also included in this category is steel moment frame construction with concrete shear walls. The shear wall is the much stiffer 33 Chapter 3 Structural Classification structural element and carries most of the lateral loads (at least 75%). 14. Steel Frame with Concrete Walls High Rise (SFCWHR) General Description: Steel framed structure with concrete shear walls providing lateral resistance in one or both directions; greater than 8 storeys in height. Generally, roof and floors are infilled metal decking on open webbed steel joists or reinforced concrete slabs spanning onto steel beams. Exterior walls are clad in steel panels, masonry or marble veneer or extensive glazing. Shear walls are usually located around stair and elevator shafts. Also included in this category is steel moment frame construction with concrete shear walls. In this case, the shear wall is the much stiffer structural element and carries most of the lateral loads (at least 75%). 15. Steel Frame with Concrete Infill Walls (SFCD General Description: Steel framed buildings with concrete floors and roof spanning onto steel beams. Concrete infill walls used as fire separation along the sides and sometimes the rear of the building but are not considered to be shear walls. Additional exterior walls are non load bearing brick often with plaster coatings. Numerous windows on front facade. Fairly common for older commercial construction (pre-1950's). 16. Steel Frame with Masonry Infill Walls (SFMI) General Description: Steel frame buildings with concrete floors and roof spanning onto steel beams. Masonry infill walls used as fire separation along the sides and sometimes the rear of the building but are not considered to be shear walls. Additional exterior walls are non load bearing brick often with plaster coatings. Numerous windows on front facade. 34 Chapter 3 Structural Classification Fairly common for older commercial construction (pre-1950's). 17. Concrete Frame with Concrete Walls Low Rise (CFLR) General Description: Concrete framed structure with concrete shear walls providing lateral resistance in one or both directions; 1 to 3 storeys in height. Roof and floors are reinforced concrete slabs and beams. Shear walls are located interior (i.e., around stair and elevator shafts) or peripherally. 18. Concrete Frame with Concrete Walls Medium Rise (CFMR) General Description: Concrete framed structure with concrete shear walls providing lateral resistance in one or both directions; 4 to 7 storeys in height. Roof and floors are reinforced concrete slabs and beams. Shear walls are located interior (i.e., around stair and elevator shafts) or peripherally. Exterior walls are clad in marble or masonry veneer, steel/ glazing panels, precast panels or extensive glazing. 19. Concrete Frame with Concrete Walls High Rise (CFHR) General Description: Concrete framed structure with concrete shear walls providing lateral resistance in one or both directions; greater than 8 storeys in height. Roof and floors are reinforced concrete slabs and beams. Exterior walls are clad in marble or masonry veneer, steel/glazing panels, precast panels or extensive glazing. Shear walls are located interior (i.e., around stair and elevator shafts) or peripherally. 20. Reinforced Concrete Moment Frame Low Rise (RCMFLR) General Description: Concrete moment frames provide lateral resistance in one or both directions; 1 to 3 storeys in height. Roof and floors are reinforced concrete slabs and 35 Chapter 3 Structural Classification beams. Exterior walls are clad in brick veneer, glazing, precast panels or metal siding. Extremely rare construction type in southwestern BC. 21. Reinforced Concrete Moment Frame Medium Rise (RCMFMR) General Description: Concrete moment frames provide lateral resistance in one or both directions; 4 to 7 storeys in height. Roof and floors are reinforced concrete slabs and beams.. Exterior walls are clad in brick veneer, precast panels, glazing or metal siding. Extremely rare construction type in southwestern BC. 22. Reinforced Concrete Moment Frame High Rise (RCMFHR) General Description: Concrete moment frames provide lateral resistance in one or both directions; greater than 8 storeys in height. Roof and floors are reinforced concrete slabs r and beams. Exterior walls are clad in brick veneer, precast panels, glazing or metal siding. Extremely rare construction type in southwestern BC. 23. Reinforced Concrete Frame with Infill Walls (RCFIW^ General Description: Concrete framed buildings with concrete floors and roof spanning onto concrete beams. Masonry infill walls used as fire separation along the sides and sometimes the rear of the building. The infill walls will act as a shear wall and behave poorly. Once they are damaged, lateral loads will be transferred to the concrete frame. Additional exterior walls are non load bearing brick often with plaster coatings. Numerous windows on front facade. Fairly common for older commercial construction (pre-1950's). Chapter 3 Structural Classification 24. Reinforced Masonry Shear Wall Low Rise (RMLR) General Description: Load bearing walls of reinforced masonry along building perimeter; 1 to 3 storeys in height. Roofs and floors can be wood joists or tongue and groove decking spanning onto timber or glulam beams; steel decking on open webbed steel joists and steel beams; or reinforced concrete beams and slabs. Common form of construction in southwestern B C after 1973 when building code required all masonry to be reinforced. 25. Reinforced Masonry Shear Wall Medium Rise (RMMR) General Description: Load bearing walls of reinforced masonry both inside the building and along building perimeter; 4 to 7 storeys in height. Roofs and floors are reinforced concrete beams and slabs. Very few built in southwestern BC. 26. Unreinforced Masonry Bearing Wall Low Rise (RMLR) General Description: Lateral loads resisted by bearing walls of unreinforced masonry (clay brick, concrete block or hollow clay tile) along building perimeter; 1 to 3 storeys in height. Roofs and floors can be wood joists or trusses spanning onto timber or steel beams; steel decking on steel beams; or reinforced concrete beams and slabs. Common form of construction in southwestern BC prior to 1973 when building code required all masonry construction to be reinforced. Quality of construction varies in the study region. 27. Unreinforced Masonry Bearing Wall Low Rise (RMLR) General Description: Load bearing walls of unreinforced masonry (clay brick, concrete block or hollow clay tile) along building perimeter. Some buildings may have side walls as masonry load bearing walls and concrete or steel framing for front and rear walls. 37 Chapter 3 Structural Classification Typically from 4 to 6 storeys in height. Roofs and floors can be wood joists spanning onto timber or steel beams or trusses; or reinforced concrete beams and slabs. Common form of construction in southwestern BC prior to 1940. Quality of construction varies in the study region. 28. Tilt Up (TIP) General Description: Low rise structure with concrete walls cast on site on top of concrete slab on grade and "tilted" into place. Precast panels act as exterior bearing walls; generally up to 2 storeys in height. Roof framing usually consists of steel decking on open webbed steel joists but may also be wood construction. 29. Precast Concrete Low Rise (PCLR) General Description: Low rise structure with concrete components cast off site and brought to site for installation; generally up to 3 storeys in height. Roof framing usually consists of steel decking on open webbed steel joists but may also be wood construction. Commonly used in commercial construction during the 1960's and 1970's. Is now primarily used for parking structures. 30. Precast Concrete Medium Rise (PCMR) General Description: Medium rise structure with concrete components cast off site and brought to site for installation. Roof framing usually consists of steel decking on open webbed steel joists but may also be wood or precast hollowcore slab construction. Exterior cladding is precast panels or concrete or masonry veneer and metal/glazing. Construction seldom used in southwestern BC. 38 Chapter 3 Structural Classification 31. Mobile Homes (Mm General Description: Single storey wood framed prefabricated buildings. Includes residential units and school portables. Lateral loads are transferred from roof/floor diaphragm to walls. Generally supported on concrete or timber pads. Anchorage to foundation may be non-existent. 3.4.2 BC15 Building Classification Table 3.4 lists the 15 building types chosen for the BC15 Building Classification (Bell, 1998). They are categorized according to structural framing type, construction material and building height and where applicable have been cross-referenced to their corresponding A T C -13 facility class. Also in this section is a detailed description of the 15 building types. It should be noted that all descriptions are made for average construction as described in Section 3.4.1. The BC15 classification includes 3 wood building types, 4 steel, 2 concrete, 3 masonry, and 1 building type each for tilt up, precast and mobile home construction. The wood building types remain relatively unchanged except for the inclusion of WPB construction where appropriate. The number of steel building types has been reduced from 12 to 4. L M F and SFCWLR remain unchanged. Four building types (high and medium rise braced and moment frames) were deleted because they were not found in the building inventory. The 6 remaining steel building types were condensed into 2 classes, SBMFLR and SFCWMHR. Concrete construction was reduced from 7 building types to 2. C F L W is a combination of CFLR and RCMFLR, valid because moment frame construction is so rare in this region. CFMHW combines CFMR, CFHR, RCFIW, RCMFMR and RCMFHR. 39 Chapter 3 Structural Classification Reinforced masonry now includes both low and mid rise buildings while unreinforced masonry building types did not change. Precast building types did not change except for the combination of low rise and mid rise construction. Finally, mobile homes remain unchanged. Table 3.5: BC15 Building Classification Material BC15 Building Type BC15 ATC-13 Wood Wood Light Frame, Residential W L F R 1 Wood Light Frame, Commercial/Institutional WLFCI 1 Wood Light Frame Low Rise W L F L R 1 Steel Light Metal Frame L M F 2 Steel Braced and Moment Frames, Low Rise S B M F L R 15,72 Steel Frame with Concrete Walls Low Rise SFCWLR 6 Steel Frame with Concrete Walls Medium, High Rise SFCWMHR 7,8 Concrete Concrete Frame with Concrete Walls Low Rise C F L R 6,18,87 Concrete Frame with Concrete Walls Medium, High Rise C F M H R 7,78,79,19 20,88,89 Masonry Reinforced Masonry Shear Wall Low Rise R M L R 9,10,84,85 Unreinforced Masonry Bearing Walls Low Rise U R M L R 78 Unreinforced Masonry Bearing Walls Medium Rise U R M M R 79 Tilt Up Tilt Up T U 21 Precast Precast Concrete Low, Medium Rise PC 81,82 Mobile Mobile Homes M H 23 1. Wood Light Frame Residential fWLFR) General Description: One or two storey, single family detached homes between 70 to 350 square meters. Wall systems are wood stud, roof/floor span systems are wood joist and rafter construction. Exterior walls are sheathed with wood or vinyl siding, stucco, plaster, brick or metal. Typically has many interior walls sheathed with plaster. A common detail is an unbraced cripple wall connection to concrete or brick foundation. Includes W L F R and WPB residences from BC31 Classification. 40 Chapter 3 Structural Classification 2. Wood Light Frame Low Rise Commercial/Institutional (WLFCD General Description: One or two storey commercial and institutional buildings between 80 to 600 square meters. Wall systems are wood stud, roof/floor span systems are wood joist or wood truss construction. Exterior walls are sheathed with wood or vinyl siding, stucco, plaster, brick veneer or light metal. Typically has extensive areas of glazing creating a storefront. Includes WLFCI and commercial/institutional WPB construction from BC31 Classification. 3. Wood Light Frame Low Rise Residential fWLFLR) General Description: Low Rise (up to 4 stories) wood frame multi-family residential buildings up to 1500 square meters. Wall systems are wood stud, roof/floor span systems are wood joist and rafter construction. Extensive interior load bearing walls. Exterior walls are sheathed with wood or vinyl siding, stucco, brick veneer or metal. Typically has many interior walls sheathed with plaster. Since late 1970's, common to have underground parking and/or concrete retail level. Includes WLFLR and appropriate WPB construction from BC31 Classification. 4. Light Metal Frame (LMFt General Description: Lightweight pre-engineered industrial and agricultural buildings with rigid frame in the short direction and tie-rod cross-bracing in the other. Roof is lightweight steel purlins. Building typically has an open interior and exterior walls are clad with metal decking though some use asbestos board. 41 Chapter 3 Structural Classification Building type definition same as BC31 Classification for LMF. 5. Steel Braced and Moment Frames. Low Rise (SBMFLR) General Description: Steel moment frames in one or both directions form perimeter or are clustered. Steel braced frame structures have cross or chevron bracing in both directions; 1 to 3 storeys in height. Roof is generally metal decking but can be wood or concrete. Floors are concrete infilled metal decking on open webbed steel joists or reinforced concrete slabs spanning onto steel beams. Exterior walls are metal studs clad in metal or wood siding, stucco, brick veneer or precast concrete panels. Moment frame construction is extremely rare in southwestern BC. If the structure includes concrete walls, building type should be considered as SFCWLR. This building type includes both SBFLR and SMFLR from the BC31 Classification. 6. Steel Frame with Concrete Walls Low Rise (SFCWLR) General Description: Steel framed structure with concrete shear walls providing lateral resistance in one or both directions; 1 to 2 storeys in height. Roof is generally metal decking but can be wood. Floors are concrete infilled metal decking on open webbed steel joists. Exterior walls are clad in metal, stucco or masonry veneer. Shear walls are located interior (i.e., around stair and elevator shafts) or peripherally. Also included in this category is steel moment frame construction with concrete shear walls. In this case, the shear wall is the much stiffer structural element and carries most of the lateral loads (at least 75%). Building type definition same as BC31 Classification for SFCWLR. 42 Chapter 3 Structural Classification 7. Steel Frame with Concrete Walls Medium. High Rise (SFCWMHR) General Description: Steel framed structure with concrete shear walls providing lateral resistance in one or both directions; greater than 3 storeys in height for S F C W M R and S F C W H R . SFCI and S F M I , any height construction. Generally, roof is metal decking. Floors are infilled metal decking on open webbed steel joists or reinforced concrete floors spanning onto steel beams. Exterior walls are clad in steel panels, precast panels, masonry or marble veneer, or extensive glazing. Shear walls can be located interior (i.e., around stair and elevator shafts) or peripherally. Also included in this category is steel moment frame construction with concrete shear walls. The shear wal l is the much stiffer structural element and w i l l carry most of the lateral loads (at least 75%). Combines S F C W M R , S F C W H R , SFCI and S F M I from BC31 Classification. 8. Concrete Frame with Concrete Walls Low Rise (CFLWl General Description: Concrete framed structure with concrete shear walls and/or moment frames providing lateral resistance in one or both directions; 1 to 3 storeys in height. Roof and floors are reinforced concrete slabs and beams. Exterior walls are clad in stucco or masonry veneer., Combines C F L R and R C M F L R from the BC31 Classification. 9. Concrete Frame with Concrete Walls Medium. High Rise (CFMHWl General Description: Concrete framed structure with concrete shear walls providing lateral resistance in one or both directions; greater than 4 storeys in height. Roof and floors are reinforced concrete slabs and beams. Exterior walls are clad in marble or masonry veneer, 43 Chapter 3 Structural Classification steel/glazing panels, precast panels or extensive glazing. Combines CFMR, CFHR, RCFIW, RCMFMR, RCMFHR from the BC31 Classification. 10. Reinforced M a s o n r y Shear W a l l L o w Rise ( R M L R ) General Description: Load bearing walls of reinforced masonry along building perimeter; 1 to 7 storeys in height. Roofs and floors can be wood joists or tongue and groove decking spanning onto timber or glulam beams; steel decking on open webbed steel joists and steel beams; or reinforced concrete beams and slabs. A common form of low rise construction in southwestern BC after 1973 when building code required all masonry to be reinforced. Combines RMLR and RMMR from the BC31 Classification. 11. Unreinforced M a s o n r y Bearing W a l l L o w Rise ( U R M L R ) General Description: Lateral loads resisted by bearing walls of unreinforced masonry (clay brick, concrete block or hollow clay tile) along building perimeter; 1 to 3 storeys in height. Interior partitions may contribute to resistance of lateral loads. Roofs and floors can be wood joists or trusses spanning onto timber or steel beams; steel decking on steel beams; or reinforced concrete beams and slabs. Common form of construction in southwestern BC prior to 1973 when building code required all masonry construction to be reinforced. Quality of construction varies in study region. Building type definition same as BC31 Classification for URMLR. 44 Chapter 3 Structural Classification 12. Unreinforced Masonry Bearing Wall Mid Rise (URMMR) General Description: Load bearing walls of unreinforced masonry (clay brick, concrete block or hollow clay tile) along building perimeter. Some buildings may have side walls as masonry load bearing walls and concrete or steel framing for front and rear walls. Typically from 4 to 6 storeys in height. Roofs and floors can be wood joists spanning onto timber or steel beams or trusses; or reinforced concrete beams and slabs. Common form of construction in southwestern BC prior to 1940. Quality of construction varies in the study region. Building type definition same as BC31 Classification for URMMR. 13. Tilt Up (TU) General Description: Low rise structure with concrete walls cast on site on top of concrete slab on grade and "tilted" into place. Precast panels act as exterior bearing walls; generally up to 2 storeys in height. Roof framing usually consists of steel decking on open webbed steel joists but may also be wood construction. Building type definition same as BC31 Classification for TU. 14. Precast Concrete (PC) General Description: Low rise structure with concrete components cast off site and brought to site for installation. Generally up to 3 storeys in height but some residential highrises use precast hollow core panels. Roof framing usually consists of steel decking on open webbed steel joists but may also be wood construction. Quality of connections from roof/floor diaphragms to shear wall is generally poor. Commonly used in commercial 45 Chapter 3 Structural Classification construction during the 1960's and 1970's, it is now primarily used for parking structures. Combines PCLR and PCMR from BC31 Classification. 15. Mobile Homes (Mm General Description: Single storey wood framed prefabricated buildings. Includes residential units and school portables. Lateral loads are transferred from roof/floor diaphragm to walls. Generally supported on concrete or timber pads. Anchorage to foundation may be non-existent. Building type description same as BC31 Classification for M H . 3.5 Conclusion Classification systems categorize the general building stock into groupings with similar risk characteristics. In determining which classification system to use in a regional risk analysis, the goal is to choose enough types to adequately characterize the entire building inventory for the region. The forty building facility classes found in ATC-13 are based on California construction up to the time the report was written in 1985. One of the main objectives of this thesis is to apply appropriate motion-damage relationships specific to Southwestern BC, and as such the ATC-13 classification system can not be directly applied to this study. The eleven Preliminary BC Survey facility classes are the ATC-13 facility classes most prevalent in British Columbia. However, by using ATC-13 as a benchmark, the Preliminary B C Survey limited the facility classes to a subset of those found in California. It is the opinion of this author that the subset of ATC-13 classes do not sufficiently represent the entire B C building inventory. The Southwestern BC Seismic Fragility Study expands upon 4 6 Chapter 3 Structural Classification the ATC-13 facility types based on engineering judgment regarding the local building inventory. Of the available building classifications, The BC15 building classifications, and to a greater extent the BC31 classification, better characterize the Southwestern BC building inventory. These classifications break down wood construction, the most prominent construction material in Southwestern BC's general building stock, into more building types than available in the ATC-13 facility classes. Also, when compared to the Preliminary BC Survey, the BC15 and BC31 classifications include building types for Precast, Light Metal Frame and Mobile Home construction. Both Precast, and Light Metal Frame construction are common in the industrial buildings in the region while Mobile Home construction is typical to satellite teaching facilities at schools. Therefore, based on this author's knowledge of the general building inventory in Southwestern BC, the BC31 and BC15 building classifications found in the Southwesten BC Building Seismic Fragility study are the most comprehensive classifications available and should be used in performing a risk analysis for the region. 47 C H A P T E R 4 Damage Estimation 4.1 Overview With the classification of the building inventory and hazard analysis complete, the next step in the loss estimation is to quantify damage given some ground motion. There has been voluminous research on deriving accurate motion-damage relationships (Algermissen and Steinbrugge, 1984; ATC-13, 1985; Kircher, 1997). Researchers have yet to come to an agreement on how to measure both the ground motion and damage. ATC-13 is an influential study which uses MMI, a subjective scale, to measure the level of ground shaking. Motion damage relationships prior to and evolving from this study relate the probability of being in a certain damage state given the occurrence of a given MMI. Damage may be measured as a percentage of replacement cost. FEMA-NIBS (HAZUS, 1997) developed a methodology which uses engineering parameters to measure ground motion and physical damage to the building stock. The following sections explore ATC-13 and FEMA-NIBS. Two additional sources of motion-damage relationships are reviewed, the BC Preliminary Study and the Southwestern BC Fragility Study. The chapter concludes with a comparison of these methodologies and the selection of motion-damage relationships for use in the B C damage estimation methodology. 48 Chapter 4 Damage Estimation 4.2 ATC-13 In 1985, the Applied Technology Council published a report entitled ATC-13, Earthquake Damage Evaluation Data for California. The objectives of this report were outlined in Section 2.5.1. One of the primary tasks was to develop earthquake damage and loss estimates based on the facility classes and ground shaking characterization. Loss estimates were calculated for the following: 1. Expected physical damage caused by ground shaking 2. Expected losses from collateral hazards such as ground failure, inundation, and fire 3. Loss of function or usability for each facility class 4. Expected casualties The following discussion focuses on the development of the expected physical damage directly caused by ground shaking, i.e., not including aggravation of damage due to soil type, ground failure, inundation or fire. At the time of development, comprehensive earthquake damage and loss data relationships for all facility types in California were not available in the literature. Therefore, A T C chose to draw on the experience and judgement of an advisory panel of specialists in earthquake engineering. Ultimately, 71 experts were selected for the Project Engineering Panel (PEP). Each participant was asked to provide best, low and high estimates of motion-damage relationships for the 40 building classes discussed in Section 3.2 and an additional 38 facility types. A self rating system and review by the PEP ensured the level of expertise involved in estimating damage probability matrices (DPM) for each facility class. 49 Chapter 4 Damage Estimation An interative questionnaire method called a Delphi procedure was used to develop a consensus on the final mean damage factors (ATC-13, 1985). Weighted mean values, standard deviations and standard errors were computed for the low, best and high estimates in each round of questionnaire. Responses were weighted based on level of expertise with each facility class. Typically, those responses lying outside of one standard deviation could be attributed to responses with low experience levels and therefore were weighted less than responses with a high experience level. Prior to an earthquake, quantifying the amount of damage to a specific structure within a facility class is not possible given the variability of structural response to ground motions. Therefore the damage factor is considered a random variable with a corresponding probability distribution at every level of ground shaking. The survey results most closely fit a beta probability distribution. Parameters of the distribution were calculated using the mean value of damage factor and the low and high estimates. Damage factor probability distributions were developed for each level of ground shaking. Damage probability matrices were obtained by discretizing the distribution for each damage state and then repeating the procedure for each facility class. Damage states defined for ATC-13 are listed in Table 4.1. An example DPM for wood frame low rise construction is shown in Table 4.2. 50 Chapter 4 Damage Estimation Table 4.1: Structural Damage States in terms of % Replacement Cost DAMAGE STATE DAMAGE FACTOR RANGE CENTRAL DAMAGE FACTOR (%) 1 - None 0 0 2 - Slight 0-1 0.5 3 - Light 1 - 10 5 4 - Moderate 10-30 20 5 - Heavy 30-60 45 6 - Major 60-100 80 7 - Destroyed 100 100 Damage States: 1 - None 2 - Slight 3 - Light 4 - Moderate 5 - Heavy 6 - Major No Damage. Limited localized minor damage not requiring repair. Significant localized damage of some components generally not requiring repair. Significant localized damage of many components warranting repair. Extensive damage requiring major repairs. Major widespread damage that may result in facility being razed, demolished or repaired. 7 - Destroyed Total Destruction of the majority of the facility. Table 4.2: Damage Probability Matrix for Facility Class 1 - Wood frame (low rise) Central Damage Factor Modified Mercalli Intensity VI VII Vlll IX X XI Xll FACILITY CLASS = 1 , 0.00 3.7 *** *** *** *** *** *** 0.50 68.5 26.8 1.6 *** *** *** *** 5.00 27.8 73.2 94.9 62.4 11.5 1.8 *** 20.00 *** *** 3.5 37.6 76.0 75.1 24.8 45.00 *** *** *** *** 12.5 23.1 73.5 80.00 *** *** *** *** *** *** 1.7 100.00 *** *** *** *** *** *** *** 51 Chapter 4 Damage Estimation 4.3 FEMA-NIBS Historically, the ground motion in loss estimation methodologies has typically been characterized using MMI. The rationale of using a subjective scale to describe ground shaking has come under close scrutiny in recent years. In response to recommendations made by FEMA-278, a methodology was developed which relates damage to engineering parameters. The FEMA-NIBS methodology uses quantitative measures of ground shaking to analyze building types in a manner similar to the evaluation of damage in a single building. The loss estimation software, HAZUS, was developed to implement this methodology for both scenario and probabilistic events. The following discussion of this methodology is based on the user and technical manuals available with the HAZUS software (1997). A description of the methodology should begin with an explanation of the damage states. Damage states describe the nature of damage exhibited by the components of the building. HAZUS describes damage with five damage states: none, slight, moderate, extensive or complete. The methodology uses 16 common building types (36 including height) and a description of damage states are given for each of these. A generic definition of damage states was rejected by this study in order to better assess each building type based on structural system and material. An example of the definition of damage states for light wood frame buildings is presented in Table 4.3. Based on the description of structural damage states, fragility curves were developed for each building type. Using a lognormal distribution, they describe the probability of being in a specific damage state as a function of lateral building displacement. Given a spectral displacement, the probability of being in a certain damage state can be derived from the 52 Chapter 4 Damage Estimation fragility curves. These probabilities are utilized in the loss estimation module of the methodology. In order to interrogate the fragility curves, the spectral response must be given. HAZUS performs several steps to determine the spectral displacement to be used in the analysis. It first considers the capacity of the building to resist lateral loads. Next the inelastic response spectrum is constructed. Building response, in terms of spectral displacement, is determined by the intersection of the building's capacity and response spectrum. Typically, building resistance is measured by plotting a pushover curve, static-equivalent base shear versus building displacement at the roof. For this application, capacity curves are constructed by converting the pushover curves for each building type to a plot of spectral acceleration versus spectral displacement. Capacity curves represent the nonlinear behavior of each building type and are characterized by two points: the yield capacity and the ultimate capacity. The yield capacity represents the lateral strength of the building and accounts for design strength, redundancies in design, conservatism in code requirements, and expected strength. The ultimate capacity represents the maximum strength of the building when the global structural system has reached a full mechanism. Separate capacity curves are available for three levels of seismic design and construction standards. The levels of seismic design are: High-Code, Moderate-Code and Low-Code. These correspond to 1994 Uniform Building Code lateral force design requirements. The construction standards are: Pre-Code, Code, and Special design. The year which delineates Pre-Code and Code construction is 1973. Special design applies to essential 53 Chapter 4 Damage Estimation facilities that are known to be of superior design and construction. After the building's capacity to withstand lateral forces is known, another curve is developed to measure the demand a given level of ground shaking requires of the structure. Spectral accelerations and spectral displacements are calculated for the region and presented in the form of a 5%-damped elastic response spectrum. Non-linear behavior at higher levels of building response is then taken into consideration by reducing the response spectrum. The amount of reduction is a function of the effective damping of the building. The modified response spectrum is called the demand spectrum. It intersects the building's capacity curve at the point of peak response displacement and acceleration. The peak spectral displacement is then used to interrogate the fragility curve for damage state probabilities. Figure 4.1 shows the steps in the methodology leading to discrete damage state probabilities. Table 4.3: Example of Structural Damage State Definitions Wood, Light Frame Slight: Small plaster or gypsum-board cracks at corners of door and window openings and wall-ceiling intersections; small cracks in masonry chimneys and masonry veneer. Moderate: Large plaster or gypsum-board cracks at corners of door and window openings; small diagonal cracks across shear wall panels exhibited by small cracks in stucco and gypsum wall panels; large cracks in brick chimneys; toppling of tall masonry chimneys. Extensive: Large diagonal cracks across shear wall panels or large cracks at plywood joints; permanent lateral movement of floors and roof; toppling of most brick chimneys; cracks in foundations; splitting of wood sill plates and/or slippage of structure over foundations; partial collapse of room-over-garage or other soft-story configurations; small foundation cracks. Complete: Structure may have large permanent lateral displacement, may collapse, or be in imminent danger of collapse due to cripple wall failure or the failure of the lateral load resisting system; some structures may slip and fall of the foundations; large foundation cracks. 54 Chapter 4 Damage Estimation Damage States: N-None, S - Slight, M - Moderate, E - Extensive, C-Complete Fig. 4.1: FEMA-NIBS methodology for damage estimation 55 Chapter 4 Damage Estimation 4.4 Preliminary BC Survey A study of the seismic vulnerability of structures in the Lower Mainland region of British Columbia was conducted in the fall of 1996 (Ventura and Rezai, 1996). The objective of the study was to develop preliminary earthquake damage factors for BC. Similar to the ATC-13 approach, the experience and judgement of local engineers was used to assess the vulnerability of various facility types. A damage vulnerability survey was conducted and statistical analysis of the responses produced BC Preliminary damage factors. These were then compared to the recommended ATC-21-1 modification factors applied to the ATC-13 mean damage factors for facilities built outside of California. The vulnerability survey was divided into four categories: facility prevalence in BC, related experience with each facility class, general vulnerability to damage of facility and facility damage factor in percent at different levels of earthquake intensity. The estimates of mean damage factors were made assuming the following: 1. The facility was assumed to be in good condition, designed and built according to the current design code. 2. The facility was located on firm ground; i.e. the foundation does not aggravate damage. 3. Fire, inundation, fault rupture and liquefaction do not aggravate damage. Using the ATC-13 definition of damage factor, experts were asked to provide their best estimate of damage factor at MMI VI through XII for each facility class. As a guide, the ATC-13 mean damage factors for each facility class for the respective MMI scale were given on the survey. Based on prevalence data, eleven building classes were chosen for further statistical analysis. Those classes are listed in Table 3.3. 56 Chapter 4 Damage Estimation In determining the BC Preliminary damage factors, average and standard deviation values were computed for the selected facility classes at each MMI scale. It was observed that there were large differences among the responses at most intensity levels. This was primarily attributed to the range of experience and knowledge of some of the participants in assessing damage as a function of the earthquake intensity levels and structural response. Survey averages of the mean damage factors were heavily influenced by responses outside the range of average plus/minus one standard deviation. Values outside this range were discarded and modified average values were calculated with the remaining responses. BC Preliminary damage factors are the modified average values for each facility class, calculated for MMI VI through XII. The BC Preliminary damage factors were then compared to the recommended A T C -21-1 modification factors applied to the ATC-13 Mean Damage Factors. ATC-21 is a study which developed modification factors to extend the ATC-13 recommendations to non-California buildings. Again, expert opinion was used to develop the modification factors. Factors are defined regionally for buildings in NEHRP (National Earthquake Hazard Reduction Program) Map Areas 1 through 6. The west coast of Canada, including southwestern BC, is classified as NEHRP Map Area 5. The BC Preliminary mean damage factors together with ATC-21-1 recommendations are shown in Table 4.4. In the results of the study, it was observed that, in general, the computed BC modified averages of the mean damage factors are slightly lower than the A T C -21-1 values. The BC Preliminary study concluded there is insufficient observational data on the province of British Columbia to classify earthquake losses with confidence. Precluding an 57 Chapter 4 Damage Estimation actual event, the best approach for estimating losses for the region is to rely on local expert opinion. The BC Preliminary study represents a general idea of expected performance of BC buildings and experimental data is necessary to verify or improve these estimates. Table 4.4: Comparison of B C Preliminary, ATC21-1 Studies Facility Class Mean Damage Factor for each MMI scale (%) VI Vll VIII IX X XI XII Wood frame (low rise) ; • ATC-21-1 2.4 4.8 6.0 13.2 25.2 31.2 46.8 BC Preliminary 1.9 5.3 8.7 15.4 27.2 32.9 46.2 Unreinforced masonry with bearing wall (low rise) ATC-21-1 5.0 12.0 24.0 43.0 67.0 78.0 88.0 BC Preliminary 6.2 13.1 27.6 46.7 66.5 80.0 91.5 Unreinforced masonry with load bearing frame (low rise) ATC-21-1 3.0 7.0 16.0 30.0 46.0 62.0 76.0 BC Preliminary 4.1 8.2 20.7 36.3 50.6 65.9 77.7 Rieinfereedeeoncreteisheariwall without moment-resistingWrame;(ldw rise) ATC-21-1 1.1 5.5 7.7 17.6 26.4 41.8 52.8 BC Preliminary 1.5 5.5 8.4 17.3 25.7 37.0 48.2 Reinforced concrete shear wall without moment-resisting frame (medium rise) ATC-21-1 2.2 5.5 11.0 20.9 34.1 45.1 56.1 BC Preliminary 1.9 5.7 10.7 21.2 30.9 41.4 53.2 Reinforced concrete shear wall without moment-resisting frame (high rise) ATC-21-1 3.3 6.6 15.4 28.6 44.0 60.5 77.0 BC Preliminary 2.6 6.3 15.0 27.8 40.0 57.0 70.2 Reinforced masonry shear wall without moment-resisting frame_(low rise) ATC-21-1 2.2 5.5 7.7 16.5 27.5 47.3 59.4 BC Preliminary 2.1 5.9 9.5 18.1 27.5 45.2 57.4 Braced steel frame (low rise) ' ATC-21-1 1.1 4.4 6.6 13.2 19.8 30.8 45.1 BC Preliminary 1.4 4.1 7.3 14.3 20.9 29.6 41.5 Braced steel frame (medium-rise) ATC-21-1 2.2 5.5 8.8 15.4 24.2 35.2 47.3 BC Preliminary 1.9 4.8 9.9 15.5 24.4 33.9 46.5 Braced steel frame with perimeter frame (low rise) ATC-21-1 2.2 4.4 5.5 7.7 18.7 25.3 37.4 BC Preliminary 1.6 4.0 7.3 11.2 19.9 26.6 37.4 • Tilt-up (low rise) ATC-21-1 3.9 6.5 16.9 26.0 39.0 59.8 85.8 BC Preliminary 2.7 6.2 14.0 22.1 36.5 49.6 73.3 5 8 Chapter 4 Damage Estimation 4.5 Southwestern BC Building Seismic Fragility Study In addition to development of the BC classification system detailed in Section 3.4, a local expert on earthquake engineering and building construction developed damage probability matrices for Southwestern British Columbia with the results presented in the report "Southwestern BC Building Seismic Fragility Study" (Bell, 1998). Two classification schemes were developed, one with 31 building types. The second scheme condensed the number of building types from 31 to 15. Damage probability matrices were developed that estimate the percentage of structures falling within each of the damage states for each MMI event. This study was conducted independent of the BC Preliminary study and the results of that study were not used in the development of the S/W BC damage matrices. As in the BC Preliminary study, the ATC-13 damage matrices were used as a guide. The assumptions made in developing the DPM's are similar to those of ATC-13 but were more explicitly defined for the S/W B C Seismic Fragility study. The estimates of vulnerability were made assuming the following: 1. The building is founded on firm material. 2. The building is of average construction as described in Section 3.4. 3. Collateral hazards such as ground failure, fault rupture, inundation, and fire do not aggravate damage. 4. Adjacency hazards do not aggravate damage. 5. Loss of function not considered. This study differed from the BC preliminary study in that a single expert was 59 Chapter 4 Damage Estimation consulted to develop motion damage relationships. Also, damage probability matrices as opposed to mean damage factors are used to relate ground motion to damage. The expert relied on his judgement to modify the ATC-13 matrices to local BC construction. When compared to the ATC-13 damage matrices, the BC31 and BC15 damage matrices contain similarities as well as significant differences. The main reason for the extent of those differences is the inherent variations in regional construction. There are differences between the regions with regards to commonly used construction details, building code requirements, and quality of construction materials and workmanship. Secondary reasons for differences are (Bell, 1998): 1. The knowledge base upon which the judgement for ATC-13 DPM's was developed was state of the art in 1985. Subsequent earthquakes have increased our knowledge of the damageability of building types. 2. The southwest BC study considered all buildings, including special buildings such as schools, hospital, firehall, etc. 3. The 15 building types are combination building types and thus expected to differ from a stand alone building type. An example DPM for Wood Light Frame Residential construction is presented in Table 4.5. A convenient way of summarizing the information contained in the DPMs is to calculate the corresponding mean damage factor for each intensity level. A summary of the BC15 and BC31 mean damage factors is presented in Table 4.6 and Table 4.7. Section 3.4 briefly discussed the reduction of the BC31 classification to the BC15 classification. Table 4.8 is a summary of how the BC31 classes were combined to create the BC15 classification. The figures which follow are a comparison of the mean damage factors 60 Chapter 4 for the combined classes. Damage Estimation The damage matrices for wood remain relatively unchanged. The percentage of WPB buildings in the building inventory is very small relative to the other wood construction types. Therefore, the inclusion of WPB construction has little or no effect on the original BC31 damage functions for WLFR, WLFCI and WLFLR building types. This is shown in Figure 4.2, Figure 4.3, and Figure 4.4, respectively. Typically, WPB construction in southwestern British Columbia would fall under the BC15 WLFCI building type. Several steel building types were deleted because they rarely occur in the local building inventory. The two building types that remain the same in both the BC31 and BC15 classifications are L M F and SFCWLR. These have a high prevalence in the building inventory. BC15 SBMFLR is a combination of SMFLR and SBFLR. This combination is valid because there are very few braced or moment frames in the building inventory. The ones that do exist are typically SBFLR. Therefore, one would expect the SBMFLR mean damage factor curve to closely resemble that of SBFLR, the trend shown in Figure 4.5. The final BC15 steel building type is SFCWMHR. It incorporates SFCWMR, SFCWHR, SFCI and SFMI. Figure 4.6 shows the mean damage factors for each of these building types. It can be seen that BC15 SFCWMHR is closely related to the SFCWMR, SFCWHR, and SFCI curves. SFMI, however is a fairly common form of construction for older steel buildings in southwestern BC. This building type is not expected to perform well because the connection between the masonry infill and the steel frame is generally poor. The 61 Chapter 4 Damage Estimation BC15 SFCWMHR mean damage factors do not reflect the difference in performance for SFMI buildings. Concrete construction was reduced from 7 building types to 2. BC15 C F L W is a combination of CFLR and RCMFLR and is shown in Figure 4.7. Because R C M F L R is so rare in this region this combination is valid. BC15 CFMHW combines CFMR, CFHR, RCFIW, R C M F M R and RCMFHR. RCMFMR and RCMFHR are not a significant portion of the building inventory so inclusion in this building type is valid. Of all these types, CFHR occurs most often in the building inventory. Figure 4.8 shows that BC15 CFMHW is most closely related to CFHR mean damage factors, as is expected. Reinforced masonry now includes both low and mid rise buildings while unreinforced masonry building types did not change. There are a handful of RMMR buildings in the region so inclusion in the BC15 RMLR building type has little effect on the mean damage factors, as seen in Figure 4.9. Precast building types did not change except for the combination of low rise and mid rise construction. Again, there are very few PCMR buildings in the region so inclusion in the BC15 PC building type has a small effect on the mean damage factors, as seen in Figure 4.10. 62 Chapter 4 Damage Estimation Table 4.5: Damage Probability Matrix for Wood Light Frame Residential (WLFR) Central Damage Factor Modified Mercalli Intensity VI VII Vlll IX X XI Xll 0.00 8.0 4.0 1.0 *** *** *** *** 0.50 75.0 28.0 6.0 1.0 *** *** *** 5.00 17.0 64.0 86.0 69.0 10.0 2.0 *** 20.00 *** 4.0 5.0 20.0 76.0 69.0 42.0 45.00 *** *** 2.0 10.0 12.0 25.0 50.0 80.00 *** *** *** *** 2.0 4.0 6.0 100.00 *** *** *** *** *** *** 2.0 Table 4.6: BC15 Mean Damage Factors Material Facility Class Mean Damage Factor for each MMI scale (%) VI VII Vlll IX X XI Xll Wood WLFR 1.2 4.4 7.4 12.0 25.4 29.9 37.7 WLFCI 1.2 5.5 9.4 15.1 28.3 36.9 44.1 WLFLR 1.0 3.8 4.9 11.6 18.8 28.1 37.4 Steel LMF . 0.5 2.7 4.1 7.0 18.8 23.9 36.7 SBMFLR 0.8 2.4 6.2 11.5 22.3 30.1 40.0 SFCWLR 0.9 4.5 6.2 16.1 21.4 36.0 46.5 SFCWMHR 1.4 5.0 8.7 21.3 30.7 47.0 55.0 Concrete CFLW 0.9 4.9 6.2 14.1 21.2 38.4 50.7 CFMHW 1.1 3.8 10.7 21.3 27.5 40.9 52.8 Masonry RMLR 0.7 4.0 5.9 16.6 31.5 43.4 58.3 URMLR 2.8 10.2 23.4 34.9 51.7 65.8 80.0 URMMR 4.3 12.2 26.9 38.2 53.8 70.0 83.7 Tilt Up TU 0.8 3.7 9.0 18.8 34.0 50.5 65.6 Precast PC 2.5 5.0 11.5 26.3 39.9 52.8 66.9 Mobile Home MH 1.8 5.6 13.5 18.8 31.8 45.0 56.7 63 Chapter 4 Damage Estimation Table 4.7: BC31 Mean Damage Factors Material Facility Class Mean Damage Factor for each MMI scale (%) VI Vl l VIII IX X XI XII Wood W L F R 1.2 4.4 7.4 12.0 25.4 29.9 37.7 WLFCI 1.2 5.5 9.1 14.5 27.4 36.9 44.1 W L F L R 1.0 3.8 4.9 11.6 18.9 28.1 37.4 W P B 1.4 6.4 11.8 18.9 31.6 39.1 45.9 Steel LMF 0.5 2.7 4.1 7.0 18.8 23.9 36.7 S M F L R 0.6 3.2 5.0 6.3 17.3 23.4 36.1 S M F M R 0.7 3.7 5.1 8.7 20.6 31.7 42.8 S M F H R 0.7 4.5 5.8 17.2 23.6 37.4 44.8 S F C W L R 0.9 2.6 6.9 12.3 22.4 31.4 40.8 S F C W M R 1.6 4.5 10.1 14.8 22.1 32.5 38.3 S F C W H R 1.6 5.9 10.5 16.0 23.8 39.6 48.4 SFCI 0.9 3.6 7.9 16.8 23.8 39.1 51.2 SFMI 3.1 7.5 16.5 36.2 45.8 64.0 69.2 Concrete C F L R 0.9 4.7 5.0 13.9 21.0 36.9 49.4 C F M R 0.9 3.6 7.9 16.8 23.8 39.1 51.2 C F H R 1.1 4.0 11.3 22.9 30.4 43.2 54.2 R C M F L R 3.0 5.5 13.8 21.0 37.9 49.9 54.5 R C M F M R 3.0 5.8 13.6 22.3 41.0 55.3 60.3 R C M F H R 3.4 4.9 15.7 25.5 41.6 60.1 67.4 RCFIW 2.9 7.7 15.6 30.4 39.6 60.6 67.5 Masonry R M L R 0.7 4.0 5.9 16.6 31.5 43.4 58.3 R M M R 0.9 4.6 8 26.7 35.3 47.8 67.3 U R M L R 2.8 10.2 23.4 34.9 51.7 65.8 80.0 U R M M R 4.3 12.2 26.9 38.2 53.8 70.0 83.7 Tilt Up TU 0.8 3.7 9.0 18.8 34.0 50.5 65.6 Precast P C L R 2.3 4.9 11.3 25.0 39.2 51.7 66.6 P C M R 2.7 6.1 13.0 28.4 38.0 53.0 69.0 Mobile Home MH 1.8 5.6 13.5 18.8 31.8 45.0 56.7 Chapter 4 Damage Estimation Table 4.8: Reduction of BC31 Classification to BC15 Classification Material BC15 Building Type BC31 Wood W L F R Wood Light Frame Residential WLFR, W P B WLFCI Wood Light Frame, Commercial/Institutional WLFCI, W P B WLFLR Wood Light Frame, Low Rise WLFLR, W P B Steel LMF Light Metal Frame LMF S B M F L R Steel Braced and Moment Frames, Low Rise SMFLR, S B F L R S F C W L R Steel Frame with Concrete Walls, Low Rise S F C W L R S F C W M H R Steel Frame with Concrete Walls Mid, High Rise SFCWMR, S F C W H R , SFCI, SFMI Cone. C F L W Concrete Frame with Concrete Walls Mid, High Rise CFLR, R C M F L R C F M H W Concrete Frame with Concrete Walls Mid, High Rise CFMR, CFHR, RCFIW, RCMFMR, R C M F H R Masonry RMLR Reinforced Masonry Shear Wall Low Rise RMLR, RMMR URMLR Unreinforced Masonry Bearing Walls Low Rise URMLR URMMR Unreinforced Masonry Bearing Walls Mid Rise URMMR Tilt Up T U Tilt Up T U Precast P C L R Precast Concrete Low, Mid Rise PCLR, P C M R Mobile MH Mobile Homes MH Fig. 4.2: Comparison of Mean Damage Factors for BC15 WLFR, BC31 WLFR and BC31 WPB Building Types 65 Chapter 4 Damage Estimation VIII IX X Modified Mercalli Intensity (MMI) Fig. 4.3: Comparison of Mean Damage Factors for BC15 WLFCI, BC31 WLFCI and BC31 WPB Building Types Fig. 4.4: Comparison of Mean Damage Factors for BC15 WLFLR, BC31 WLFLR and BC31 WPB Building Types 66 Chapter 4 Damage Estimation Fig. 4.5: Comparison of Mean Damage Factors for BC15 SBMFLR, BC31 SBFLR and BC31 SMFLR Building Types Fig. 4.6: Comparison of Mean Damage Factors for BC15 SFCWMHR, BC31 SFCWMR, BC31 SFCWHR, BC31 SFCI and BC31 SFMI Building Types 67 Fig. 4.7: Comparison of Mean Damage Factors for BC15 CFLW, BC31 CFLR and BC31 RCMFLR Building Types Fig. 4.8: Comparison of Mean Damage Factors for BC15 CFMHW, BC31 CFMR, BC31 CFHR, BC31 RCFIW, BC31 RCMFMR and BC31 RCMFHR Building Types 68 Chapter 4 Damage Estimation Fig. 4.9: Comparison of Mean Damage Factor for BC15 RMLR, BC31 RMLR and BC31 RMMR Building Types 100 VI vii VIM IX X XI Xl l Modified Mercalli Intensity (MMI) Fig. 4.10: Comparison of Mean Damage Factor for BC15 PC, BC31 PCLR and BC31 PCMR Building Types 6 9 Chapter 4 Damage Estimation 4.6 Proposed Methodology for this study In a comparison of damage estimation methodologies, it is important to recognize the difference between methodologies and motion-damage relationships. Two methodologies have been reviewed, ATC-13 and FEMA-NIBS. These represent the most comprehensive methods for regional damage and loss estimation to date, and extensive coverage of these methodologies is available in the literature. They have each been applied in various regions, both separately and as a comparison. Within these methodologies are motion-damage relationships. Choosing a methodology for application in a region does not necessitate the use of the motion-damage relationships within that study. ATC-13 was developed for California construction, so a direct application of the damage functions from that study to southwestern BC is inappropriate. The BC Preliminary Study and the Southwest BC Fragility Study, however, use expert opinion to modify the relationships in ATC-13 to local construction. In order to develop a methodology for southwestern BC, the initial step is to choose a methodology and then the motion-damage relationships to be applied in the methodology. The building damage functions within ATC-13 and FEMA-NIBS were developed using separate philosophies towards the characterization of ground motion and building response. The DPMs developed for ATC-13 combine expert opinion to relate building response to MMI. This approach was chosen because comprehensive earthquake damage and loss data were not available at the time. FEMA-NIBS relates damage to engineering parameters. This is in response to criticism of using a subjective scale to estimate damage. Task 1 in the development of the FEMA-NIBS methodology was to assess existing loss estimation data and provide recommendations for a standardized methdology for the 70 Chapter 4 Damage Estimation United States. In 1993, "Assessment of the State-of-the-Art Earthquake Loss Estimation Methodologies, FEMA-249" was released (RMS, 1993). That report states, "The use of MMI is somewhat irrational because of the circular logic of estimating damage based on ground-motion estimates derived from damage." The report goes on to question the validity in using a methodology which introduces a significant amount of uncertainty relating peak ground acceleration to MMI, and one which produces questionable results at higher levels of intensity. The introduction of the FEMA-NIBS methodology as state of the art has not dismissed the use of methodologies using qualitative descriptions both in Canada and in the United States. The southwestern British Columbia studies by Munich Re-Insurance (1990) and Canada Mortgage and Housing Corporation (1989) both use MMI to characterize the ground motion. Despite recommendations to quantify building response using engineering parameters, FEMA-249 acknowledges, "Nonetheless, these [qualitative] methodologies have gathered more consensus than other more analytically based approaches." While the debate continues, several advancements have been made since the 1985 release of ATC-13. Recent California earthquakes have provided the knowledge and data necessary to improve the original vulnerability functions and facility class definitions (Anagnos, 1995). Several studies have adopted ATC-13 to regions outside of California (ATC-21, ATC-36). Also, because the ATC-13 methodology is so widely approved for use, simultaneous analyses with the F E M A -NIBS methodology have been performed in order to compare the methodologies (McCormick, 1997, King, 1997). In general, Modified Mercalli Intensity Scale is still commonly used to characterize ground shaking in current research. 71 Chapter 4 Damage Estimation In light of the ongoing support for both methodologies, choosing a philosophy for this study must be based on the project scope and limitations. Recently, there have been no significant earthquakes in southwestern British Columbia. Regional damage functions are not available based on damage data from past earthquakes. The studies by Munich Re-Insurance (1990) and C M H C (1989) addressed this issue by relying on expert opinion of structures performance in past earthquakes to develop damage functions. A comprehensive set of damage functions based on local expert opinion are available from the BC Preliminary and Southwest B C Seismic Fragility Study. The common factors of the existing sets of locally applicable motion-damage relationships are: reliance on expert opinion to develop damage functions for local application and the use of MMI to characterize ground shaking. After considering the available information, the following conclusions were made for this study: (1) MMI scale is to be used to characterize ground shaking; (2) the ATC-13 damage estimation methodology is to be adopted with modification of damage functions to suit building types in southwestern British Columbia. 4.7 Proposed Motion-Damage relationships for this study ATC-13 damage probability matrices were developed for California construction while the FEMA-NIBS fragility curves are also for United States construction. Neither contains damage functions which are directly applicable to southwestern British Columbia. Therefore, further comparison must be made of the damage functions available for application in the region. The Munich Re-Insurance study (1990) uses motion damage relationships based on 72 Chapter 4 Damage Estimation residential, commercial and industrial construction. The study is based on a scenario event so available damage functions are limited to an event of magnitude VII to VIII. The damage estimates in the report by C M H C (1989) are also for a limited number of facility classes and two scenario events. In summary, both the Munich Re-Insurance study and the C M H C study use expert opinion based damage functions developed specifically for the lower mainland. However, neither of these studies presents a comprehensive set of motion-damage relationships for a range of earthquake intensities. The BC Preliminary study combines local expert opinion to modify the ATC-13 damage functions to regional construction. This report also contains a comparison to modified ATC-13 functions based on the ATC-21-1 methodology. The Southwestern BC Fragility study is a separate study also based on local expert opinion to modify the ATC-13 damage probability matrices to regional construction. The BC Preliminary study, including the ATC-21-1 comparison, considers the ATC-13 facility classes prevalent in the study region. The Southwestern BC Fragility study developed building types based on ATC-13 but which are unique to this study. Because the studies use distinct building classifications and corresponding damage functions, a direct comparison of damage functions for all the building types is not possible. The building types in the preliminary survey and their corresponding building types as defined in the BC15 classification are listed in Table 4.9. It should be noted that some BC15 building types correspond to more than one facility class. BC15 building types not listed do not have a corresponding facility class in the BC Preliminary study, and are not included in this comparison. The facility class, Braced steel frame (medium rise), does not have a corresponding BC15 building types but does correspond to the BC31 SBFMR 73 Chapter 4 Damage Estimation building type. Figure 4.11 to Figure 4.21 are plots comparing the mean damage factors recommended by the BC Preliminary study, the ATC-21-1 methodology for the west coast of Canada, and the corresponding BC15 building types listed in Table 4.9 for each ground shaking intensity. Comparisons between the BC Preliminary and ATC-21-1 values are described in the BC Preliminary study (1996). This comparison will focus on differences between the BC Preliminary and BC15 damage functions. From these plots, there are many similarities and differences between the two studies. From Figure 4.11 and Figure 4.12, BC15 mean damage factors are slightly higher for wood frames, and unreinforced masonry with bearing wall buildings. The difference increases at higher levels of magnitude. It should be further noted that there is very little data available to estimate damageability during a MMI XI or XII earthquake thus the results from both studies at these intensities are extremely subjective. The BC15 U R M L R mean damage factors are in close agreement with the BC Preliminary unreinforced masonry with load bearing frame values as seen in Figure 4.13. However, the description of the BC15 URMLR building type more close resembles that of unreinforced masonry with bearing wall (low rise). Figures 4.14 to 4.16 show the BC15 mean damage factors are all in close agreement with the BC Preliminary values for reinforced masonry and reinforced concrete shear wall construction. However, the BC15 study estimates improved behaviour for high rise reinforced concrete shear wall construction as seen in Figure 4.17. That improvement is more 74 Chapter 4 Damage Estimation significant at higher intensities. According to the BC15 building type description, CFMHW is a fairly common form of construction in the study region and is expected to behave well, particularly if the walls are distributed to minimize torsional effects. For steel and tilt up construction, the BC15 mean damage factors are all in close agreement with BC Preliminary braced steel frame and tilt up construction as seen in Figure 4.18 to Figure 4.21. Figure 4.19 shows a significant improvement for the behavior of the BC31 SBFMR building type at MMI XII, but again differences at higher levels of intensities are due to varying opinions of building behavior at higher intensities between the two studies. In general, for the classes chosen by local experts as most prevalent in the study region, the BC Preliminary mean damage factors are in agreement with the BC15 values. Two classes for which better performance was predicted by the BC15 damage functions at higher intensities are unreinforced masonry with bearing wall (low rise) and reinforced concrete shear wall without moment-resisting frame (high rise). As will be discussed in the case study of Chapter 6, the seismic hazard for the study region corresponding to current seismic design codes is between MMI VII and MMI VIII. At these intensities, there is little difference between the BC15 damage functions and the BC Preliminary values. From the BC Preliminary report, mean damage factors from that study are slightly lower than the ATC-21-1 values but are still in good agreement. After considering that the BC31 and BC15 classification systems are a more comprehensive representation of the local building inventory, and corresponding damage functions are in agreement with the BC Preliminary study, especially at the prescribed 75 Chapter 4 Damage Estimation intensity levels; the motion-damage relationships to be used for damage estimation in this study are the BC15 and BC31 mean damage factors. Table 4.9: Comparison of BC Preliminary facility classes to BC15 building types Material BC Preliminary Facility Classes BC15 Wood Wood frame (low rise) WLFR Masonry Unreinforced masonry with bearing wall (low rise) URMLR Unreinforced masonry with load bearing frame (low rise) URMLR Reinforced masonry shear wall without moment-resisting frame (low rise) RMLR Cone. Reinforced concrete shear wall without moment-resisting frame (low rise) CFLW Reinforced concrete shear wall without moment-resisting frame (medium rise) CFMHW Reinforced concrete shear wall without moment-resisting frame (high rise) CFMHW Steel Braced steel frame (low rise) SBMFLR Braced steel frame (medium rise) SBFMR (BC31) Braced steel frame with perimeter frame (low rise) SBMFLR Tilt Up Tilt Up (low rise) TU 76 Chapter 4 Damage Estimation Fig. 4.11: Wood frame (low rise) Fig. 4.12: Unreinforced masonry with bearing wall (low rise) 77 Chapter 4 Damage Estimation o co L L cu CO CD E CO Q c co co 100 90 80 70 60 50 40 30 20 10 0 — • • BC15 URMLR • - • - A - . . . ATC-21-1 - • -X BC Prelim yv* 't - ^r^. .A.:'. ... I I VI VII Vl l l IX X Modified Mercalli Intensity (MMI) XI Xl l Fig. 4.13: Unreinforced masonry with load bearing frame (low rise) 100 90 80 70 o co L L CD cn CO E CO Q c CO CD BC15 RMLR . . . . A - • • • ATC-21-1 x- — BC Prelim Vl l l IX X Modified Mercalli Intensity (MMI) Xl l Fig. 4.14: Reinforced masonry shear wall without moment-resisting frame (low rise) 78 Chapter 4 Damage Estimation Fig. 4.15: Reinforced concrete shear wall without moment-resisting frame (low rise) Fig. 4.16: Reinforced conrete shear wall without moment-resisting frame (medium rise) 79 Chapter 4 Damage Estimation Fig. 4.17: Reinforced conrete shear wall without moment-resisting frame (high rise) Fig. 4.18: Braced steel frame (low rise) 80 Chapter 4 Damage Estimation Fig. 4.19: Braced steel frame (medium rise) Fig. 4.20: Braced steel frame with perimeter frame (low rise) 81 Chapter 4 Damage Estimation 100 90 — 80 3 7 0 £ 60 50 h 40 30 20 10 -o K - • BC15TU ATC-21-1 BC Prelim 0 VI VII Vlll IX X Modified Mercalli Intensity (MMI) XI Xll Fig. 4.21: Tilt-Up 82 Chapter 4 Damage Estimation 4.8 Conclusions Although there have been several light to moderate size earthquakes felt in southwestern British Columbia over the past few decades (M4.9, 1975; M5.3, 1976; M4.9, 1990; M5.0, 1996; M4.6, 1997; Natural Resources Canada, 1999), there is very little damageability data available for the region. Consequently, no local motion-damage relationships exist based on observed building behavior. In light of this, the available options are to derive fragility curves based on building parameters or to utilize existing relationships based on local expert opinion. As discussed in the comparison of the ATC-13 and F E M A -NIBS methodologies, both philosophies have advantages and disadvantages. The main advantage of using an engineering parameter methodology is to erase the subjective nature of using Modified Mercalli intensities to characterize ground motion typical to expert opinion based models. The advantage of using expert opinion based models is that, where analytically derived ground motion-damage relationships are not available for southwestern British Columbia, comprehensive sets of expert opinion damage functions already exist for the region. Several studies have been presented which use expert opinion based functions. Therefore, the decision to use an expert opinion based methodology over an engineering parameter model is based on the availability of motion-damage relationships for the region and the ongoing support from the engineering community of the use of this type of methodology. In Chapter 3, the BC31 and BC15 building classifications found in the southwestern BC Building Seismic Fragility study were identified as the most comprehensive systems for describing the local building inventory. The validity of using the corresponding motion-83 Chapter 4 Damage Estimation damage relationships for these systems has been presented in this chapter. The damage functions from the previous British Columbia loss estimation studies (Munich Re-Insurance, 1990; C M H C , 1989) were dismissed because they do not contain a robust classification system nor are they applicable for a variety of intensities. The ATC-13 and FEMA-NIBS damage functions were not chosen because they apply to California and United States building construction. The BC Preliminary study damage functions are based on a statistical analysis of responses from local engineers to a damage vulnerability survey, similar to the ATC-13 approach and using ATC-13 DPM's as a guide. The damage probability matrices found in the Southwestern BC Building Seismic Fragility study were developed based on a single experts judgement of local building performance, again using the A T C DPM's as a guide and developed independent of the BC Preliminary study. It would therefore be expected that these two studies would yield similar damage functions. Typically, the BC Preliminary Study damage functions and the BC31 and BC15 DPM's do exhibit similar trends, and the validity of where they differ has been discussed. The BC15 and BC31 damage probability matrices from the Southwest BC Seismic Fragility study are in good agreement with other available motion-damage relationships. These damage functions correspond to a more comprehensive classification system than the BC Preliminary study. Therefore, the motion-damage relationships to be used for damage estimation in this study are the BC15 and BC31 mean damage factors. 84 C H A P T E R 5 Implementation of Methodology for Southwestern B C 5.1 Overview The general steps in a damage and loss estimation methodology were reviewed in Chapter 2. For this study, a methodology for structural damage estimation is suggested for use in southwestern British Columbia. The steps in a loss estimation which are addressed in this research are highlighted in Figure 5.1. They are: Building Inventory Data, Motion-Damage Relationships and Regional Damage Distribution. Chapters 3 and 4 discussed the selection of a building classification and corresponding motion-damage relationships. Due to resource limitations, a comparison of existing methodologies and damage functions was performed in order to select the values most appropriate for use in the region. The building classification and damage probabilities of the Southwest BC Fragility study are recommended for this methodology. As stated earlier in this thesis, the Repair and Replacement Models and Loss Distribution steps will not be addressed by this study. Also, it is assumed that the Seismic Hazard Analysis, including ground motion and local site effects, has been performed prior to the implementation of this methodology. The following discussion focusses on the remaining steps: Structural Inventory Data and Regional Damage Distribution. 85 Chapter 5 Implementation of Methodology for Southwestern BC Seismic Hazard Analysis -Structural Inventory Data v Ground Motion f Local Site Effects \ Motion Damage Relationships 1 Regional Damage Distribution l l l l l l i j I i l l lM T Repair and Replacement Cost Model r Regional Loss Distribution Fig. 5.1: Southwestern B C Damage Estimation Methodology 5.2 Structural Inventory After choosing a classification system, the next step in the regional damage estimate is to develop a database of all the buildings in the region. There are various methods for inventory development, each with an associated cost and level of detail. It is important to identify the existing resources and amount of time and funding available prior to beginning the inventory. This will dictate which method is most appropriate for use in developing a structural inventory. 86 Chapter 5 Implementation of Methodology for Southwestern BC The most common, and cost effective, method of creating an inventory is to integrate existing structural databases. If the required information to perform a damage estimate does not exist for each structure in the inventory, those databases can be augmented with information obtained through a field survey or inference schemes. If existing databases are not available, a field survey should be performed for each structure in the inventory. The level of detail performed in a field survey varies. It can range from a rapid visual screening taking only a few minutes to a detailed seismic analysis of each structure. The following sections outline a recommended sequence for developing a structural inventory in Southwestern British Columbia. An emphasis is placed on the minimum information required to perform a damage assessment but also discusses additional information that would be useful in calculating a more accurate damage estimate. Because the region under study is so large, the damage assessment is being performed one municipality at a time. This approach to covering the entire region was chosen to control database management and computing time for damage analyses. 5.2.1 Required Inventory Information The minimum information required to perform a structural damage assessment is geographic location and structural properties of each building. Location is important for showing the geographic distribution of damage in a region. It is used to determine the ground shaking hazard and local soil conditions at the building site. Location can be described specific to one building or by attributes shared by many buildings. For a single building, the most precise location can be stated by the longitude and latitude of 87 Chapter 5 Implementation of Methodology for Southwestern BC its' centroid. Another possible unique identifier is street address. Groups of buildings are identified by postal code, block number or, at a larger scale, census tract or neighborhood. The postal code can be specified with either 3 digits (Forward Sortation Area) or 6 digits (Local Delivery Unit) and is a common reference type for all municipalities. Certain structural properties are required in the database to determine the prototype of the building. Required building attributes are: construction material, type of framing, height (number of stories), and total floor area. This is the minimum information needed to classify the building according to the BC31 classification system. Construction material is either Wood, Steel, Concrete, Masonry, Tilt Up, Precast or Mobile Home. Type of framing is the appropriate structural prototype according to the BC31 classification described in Section 3.4.1. The database should include number of above and below grade stories. This information is not always straightforward, i.e., when the structure is built on a slope. In this situation, use the largest number of floors. Total footprint area is required for weighting the damage assessment according to floorspace. There are a number of other attributes commonly acquired in building inventories. They include use classification, building irregularities and soil type. Depending on the methods used for inventory development, the ease of obtaining these attributes varies. The personnel conducting the survey should weigh the benefit of a more detailed database over the costs of identifying these attributes. Use classifications are generally required in building inventories for several reasons. Classifying buildings according to their use is helpful for determining prototype. For 88 Chapter 5 Implementation of Methodology for Southwestern BC example, if a wood frame building is in a residential zone, then it would be categorized as wood light frame residential (WLFR). This information is also helpful for studies on non-structural damage or mitigation. Special use buildings such as post-disaster, hospital or schools could be identified for a risk assessment separate from the regular building inventory. Relations between use and occupancy could be used to estimate casualties. As was discussed in Section 3.4.1, the classification system made certain assumptions to define average construction. Attributes that indicate the building is not of average construction are considered modifiers. Modifiers are applied to the damage assessment to reflect non-standard construction. They either increase or decrease the damage of standard or average construction. Positive modifiers are those characteristics of the building that serve to improve the overall seismic behavior as compared to prototypes of average construction. For example, if all other factors are equal, a building that has been seismically retrofitted would in general behave better than one that has not. Or, a building that was built in accordance with current seismic requirements would behave better than the same prototype building built to previous codes. In the same sense, there are negative modifiers which impair the overall seismic behavior of a building. Examples of negative modifiers are storefront, pounding, poor building condition, adjacent building types and building irregularities. A storefront or "soft storey" refers to a reduction in lateral stiffness due to discontinuous shear walls or extensive openings. SEAOC defines it as a storey "in which the 89 Chapter 5 Implementation of Methodology for Southwestern BC lateral stiffness is less than 70% of that in the storey immediately above or less than 40% of the combined stiffness of the three stories above" (ATC, 1988). An example of this would be the extensive glazing used on the ground floor of retail construction. Pounding occurs when adjacent buildings are constructed with inadequate clearance between them. The deflection of the buildings during the earthquake causes them to pound against each other. This type of damage can lead to partial or total collapse if the floor of one building pounds into the column of an adjacent building (ATC, 1988). Adjacent buildings can also be considered a negative modifier due to the debris of one building causing further damage in an adjacent building. The degree of building regularity is another modifier that can significantly affect building behavior. Irregularity can be categorized in terms of floor plan, elevation, or internal properties such as mass, stiffness and damping distribution (Swiss Re). The most prominent result of building irregularity is the increase of torsional effects in the seismic response of the structure. A secondary effect is that various "wings" of a building act independently, resulting in differential movements, cracking and other damage (ATC, 1988). Irregular building footprints such as "L", "T", "E", "C", and "Z" are all examples of horizontal building irregularity. Soil type is the final modifier that will be discussed in this study. The National Building Code of Canada states that as the foundation soils become softer, the design motions which are for firm site will be amplified (NBCC, 1995). The BC31 vulnerability relationships were made with the assumption that the prototypes are founded on firm material. The 90 Chapter 5 Implementation of Methodology for Southwestern BC amplification of ground motion hazard could be accounted for if soil type is included in the structural inventory. 5.2.2 Sources of Inventory Data The approach of this study is to collect a structural inventory for each municipality and perform a damage estimate of each inventory separately. The results from all the municipalities can then be combined to produce an overall view of the risk to the building inventory in Southwestern British Columbia. The sources of inventory data available for each city will vary. Initial development of the structural inventory should begin with existing inventories. If those do not contain the required information to perform a damage estimate, e.g. location and prototype, inference schemes or field surveys should then be used to augment the original database. Development of the inventory entirely by field survey is costly and labor intensive. This method is only recommended as a last resort if no existing databases are available. Though this study focuses on preliminary damage estimates based on the minimum information required, as much information as possible should be obtained at the time of inventory development. This will eliminate the need for repeating work should more resources become available to perform a more detailed damage estimate (i.e. including building modifiers, or a casualty analysis). Sources of inventory data are available from both the private and public sector. They will vary in terms of completeness, accuracy, type of information included, media format and cost to obtain (King, 1994). Private databases are typically developed for a specific purpose. 91 Chapter 5 Implementation of Methodology for Southwestern BC Insurance companies have listings for buildings that they insure and may not cover all the buildings in the region. Real estate listings of property also have the same self-interest. Therefore, it is better to approach the public sector for existing inventory databases. Because each municipality is treated as a separate component of the region, it is easier to focus obtaining existing inventories from the local government as opposed to federal or provincial jurisdictions. Municipal building databases are typically kept by the Office of the Tax Assessor, the Building Department and the Planning Department. Tax assessor files may contain information about the age and location of the building, the total area and the number of storeys. Typically, they focus on ownership and property value, data attributes which are helpful in determining economic loss. It is important to note that the age of the building may be the year of assessment and not necessarily the year of construction (ATC, 1988). The Building Department generally keeps files of a building's permit, plans and and/or structural calculations on microfilm. These files are typically kept current for new construction but may be missing information on older buildings. They can be used to obtain street address, prototype, number of stories and floor area. Because obtaining information from this source involves reviewing structural plans, it is time consuming to use for all the buildings in a region. The use of this source is appropriate for the portion of the inventory where structural prototype is not easily deduced from inferences schemes or field survey. The Planning Department may have the most useful information for a damage assessment. Their files typically contain location information such as street address, census 92 Chapter 5 Implementation of Methodology for Southwestern BC tract, neighborhood, and lot number. Building information included is number of floors, footprint, building age, heritage structure designation, building condition and use. If a building is flagged as a heritage structure, then the local historical society may have additional files on those buildings. Zoning maps are available from the planning department which can be used to determine primary use classifications. Finally, it should be determined whether other studies have developed inventories for a specific structural type. Often, a separate database exists for a site with multiple buildings. Fire stations, police stations, hospitals, and schools are all examples of building types that may have facility specific databases. 5.2.3 Inference Schemes Once all necessary data attributes have been extracted from existing inventories, there are a few methods for obtaining missing information. As much information as possible should be determined using inference schemes. Inference schemes for this study are rules to determine missing data attributes based on opinions of local experts. Rules should be developed individually for each municipality to reflect the local construction history. The number of required schemes is dependent on the amount of missing data. Some examples taken from the case study of Chapter 6 are as follows: (a) IF (structure type = Concrete Block) and (age < 1973) and (number of stories = 1 ) T H E N (building type = unreinforced masonry low rise) (b) IF (use = residential) and (footprint area < 1500 sq. ft.) T H E N (building type = wood light frame residential) 93 Chapter 5 Implementation of Methodology for Southwestern BC Inevitably, there will still be missing information. At this point, building department plans can be consulted to confirm inference schemes. Also, a field survey can be conducted to determine the final pieces of missing data. 5.2.4 Field Survey If no existing databases are available than a field survey can be conducted to determine the required data. Because municipalities typically consists of thousands of buildings in the inventory, this method of data collection can be extremely costly and time consuming. It may be decided that only a portion of the building inventory should be collected in this way, and the rest can be adequately developed using the methods previously outlined. For example, a field survey could be performed for the commercial and industrial portion of the building inventory and residential construction could be developed using inference schemes. The level of detail performed in a field survey can vary from a rapid visual screening to a detailed structural analysis. A rapid visual screening is preferred for the level of detail needed in this study. For trained personnel, this procedure takes approximately 5 to 10 minutes per building. A building inventory form is used in the field to standardize the data for inventory management. Once the area of study is completed, then the data from each form is collated in a computerized database. Figure 5.2 is an example Building Inventory Form. This form is considered a guide for a building damage study but may be modified depending on the needs of specific users. This inventory form is modeled after the methodology in FEMA-154, Rapid Visual Screening 94 Chapter 5 Implementation of Methodology for Southwestern BC of Building for Potential Seismic Hazards: A Handbook (ATC, 1988) and has been customized for use with the BC31 structural classification system. As was described in Section 5.2.1, required inventory data falls into the categories of location, use and structural properties. Location information includes: Street address, Postal Code, Zone and Building Name. Use classification is listed in the Primary Use field. Structural properties include: Number of storeys, Year built, Footprint area, and Building type according to the BC31 Classification. Space is also allotted for building modifiers such as Building shape, Storefront, Pounding, Retrofit, Adjacent building types, and Soil type. Any additional information may be annotated in the comments section. Also, a photo or sketch should be taken of the building for reference purposes. Finally, the inspector should record their name and date of inspection on each form in case there is a need to review the inventory form. 95 Chapter 5 Implementation of Methodology for Southwestern BC Fig. 5.2: Sample Building Inventory Form Building Inventory Form Reviewer Address: Postal Code: Date: Building Name: Zone: Primary Use: Photo: No. of Storeys: Year Built: Footprint Area: Shape: Rect L T E C Z Other Storefront: Y N Wood W L F R Wood Light Frame, Residential Pounding: Y N WLFCI Wood Light Frame, Commercial/lnst. Retrofit: Y N W L F L R Wood Light Frame Low Rise Adjacent Building Types: W P B Wood Post and Beam Soil Type: Steel LMF Light Metal Frame Photo/Sketch: S M F L R Steel Moment Frame Low Rise S M F M R Steel Moment Frame Mid Rise S M F H R Steel Moment Frame High Rise S B F L R Steel Braced Frame Low Rise S B F M R Steel Braced Frame Mid Rise S B F H R Steel Braced Frame High Rise S F C W L R Steel Frame Concrete Walls Low Rise S F C W M R Steel Frame Concrete Walls Mid Rise S F C W H R Steel Frame Concrete Walls High Rise SFCI Steel Frame with Concrete Infill Walls SFMI Steel Frame with Masonry Infill Walls Cone. C F L R Concrete Frame with Concrete Walls Low Rise C F M R Concrete Frame with Concrete Walls Mid Rise C F H R Concrete Frame with Concrete Walls High Rise R C M F L R Reinforced Concrete Moment Frame Low Rise R C M F M R Reinforced Concrete Moment Frame Mid Rise R C M F H R Reinforced Concrete Moment Frame High Rise RCFIW Reinforced Concrete Frame with Infill Walls Masonr R M L R Reinforced Masonry Shear Wall Low Rise R M M R Reinforced Masonry Shear Wall Mid Rise U R M L R Unreinforced Masonry Shear Wall Low Rise U R M M R Unreinforced Masonry Shear Wall Mid Rise Tilt Up TU Tilt Up Precast P C L R Precast Concrete Low Rise P C M R Precast Concrete Mid Rise Mobile MH Mobile Homes Comments: 96 Chapter 5 Implementation of Methodology for Southwestern BC 5.3 Regional Damage Distribution With a complete inventory and motion damage relationships, it is now possible to estimate damage to the study region. It is assumed that the user has already determined the probabilistic levels of ground motion for the region, in terms of Modified Mercalli Intensity. A geographic information system (GIS) is the platform used for performing the damage analysis. A hazard map of MMI levels is input into the GIS. Each building in the inventory is then mapped according to its geographic location. The MMI corresponding to the building's geographic location is assigned to the building. For the building prototype, the mean damage factor is determined from the motion-damage relationships defined in Chapter 4. The GIS repeats this procedure for every building in the inventory. It is important to note that applying the motion-damage relationship to a particular building does not reflect the predicted damage for an individual structure. Motion-damage relationships estimate the average damage for a group of buildings of similar construction in the entire region. Finally, the GIS can be used to map the distribution of damage in the region. Results can be aggregated according to the specific needs of the user. For example, maps can be generated showing the distribution of unreinforced masonry in a region, or average damage factor (%) for MMI VII in each census tract. The case study in Chapter 6 gives further examples of presentation of results made possible by use of a geographic information system. 97 Chapter 5 Implementation of Methodology for Southwestern BC 5.4 Conclusions Chapters 3 and 4 focussed the discussion on selection of motion-damage relationships for southwestern British Columbia. Damage functions are a single component in estimating earthquake damage to the region's building inventory. This chapter addressed two other components of the methodology: Structural Inventory Data and Regional Damage Distribution. The collection of structural inventory data and database management can be the most time consuming and labor intensive part of conducting a damage estimate for a large region. The emphasis in this chapter has been discussion of ways to efficiently create a building inventory based on the available resources. Minimum information, such as geographical location and prototype, required for each building in the inventory was provided as well as additional information that could be useful in a more detailed study. A good rule of thumb when obtaining information is to obtain as much information as possible within the given resources. Begin database development with existing databases from the public and/or private sector. Complete the database using inference schemes. If geographical location and prototype are still missing for a portion of the inventory, use rapid visual screening techniques to perform a field survey to determine the missing information. It was suggested that separate inventories be developed for each municipality. This serves as both a starting place for obtaining information and as a means of disaggregating the building inventory into a set of smaller, more manageable databases. Damage estimates can then be performed at a municipal level and results from all municipalities aggregated in a GIS environment to show estimates of damage distribution for the entire region. 98 C H A P T E R 6 Case Study: The City of New Westminster 6.1 Background Southwestern British Columbia is composed of the Greater Vancouver Regional District (GVRD) and the Capital Regional District. The GVRD includes the city of Vancouver and the surrounding suburban municipalities. It is geographically bounded by the Canada/US Border to the south, the Straight of Georgia to the west, and mountains to both the north and east. The Capital Regional District is located on the southern tip of Vancouver Island located to the southwest of Vancouver and encompasses Greater Victoria, Sooke, Saanich and the southern Gulf Islands. Combined, this region contains almost 60% of the population of British Columbia. They serve as hubs for both economic and government sectors in British Columbia industry (BC Stats, 1999). The southwest coast of Canada is one of the most seismically active regions in Canada. Southwestern BC can be affected by both crustal earthquakes in the Continental North American Plate and a large subduction earthquake that may occur as the Juan de Fuca Plate subducts under the Continental North American Plate (Rogers, 1983). The location of a significant portion of the BC population in an area of high seismic risk could result in signicant losses to the British Columbia economy in the event of a 99 Chapter 6 Case Study: The City of New Westminster moderate or larger earthquake occuring within the region, i.e. Richter Magnitude greater than or equal to 5.0 (Munich Re-Insurance, 1990). The methodology developed thus far estimates regional structural earthquake damage to buildings in Southwestern BC and covers structural inventory development, motion-damage relationships and regional damage distribution. Applying this methodology to the entire region is a step towards quantifying potential economic losses due to seismic activity. The approach of this study is to produce regional damage estimates for each municipality in the region. The results from all the municipalities can then be combined to produce an overall view of the risk to the building inventory in Southwestern BC. Working on a municipal level primarily serves to break down the sizeable building inventory located in Southwestern BC into smaller, manageable building inventories. Regional damage estimation on a municipal level has an added benefit to local government agencies who can use damage estimates for earthquake preparedness, emergency response, and risk mitigation. A case study of a single municipality, the City of New Westminster, is presented here as an example of methodology implementation. The City of New Westminster, part of the GVRD, was chosen for their contributions in kind and their exchange of the city's building and land use information. The municipality has a population base of 49,350 people living within a 15.4 square kilometer area. The city's building inventory consists of over 8,000 structures, many of which date back to the late 1800s and turn of the century. The municipality is representative of the other cities in the region, containing a cross section of use classification and both historical and new building construction. 100 Chapter 6 Case Study: The City of New Westminster This case study presents preliminary estimates of structural damage to buildings in New Westminster that may be expected as a result of ground shaking consistent with the ground motions specified for design in the National Building Code of Canada (1995). A probabilistic seismic hazard analysis was performed for crustal earthquakes only. The analysis was performed on a building by building basis and the results are aggregated at the block level. The development of the structural inventory and ensuing damage estimation follows the methodology outlined in Section 5.2 and Section 5.3. An alternate analysis was performed using the BC15 classification to explore the trend in reducing the number of building types. 6.2 Seismic Hazard The methodology outlined in Chapter 5 is limited to damage estimation, one of the many steps in a risk analysis. It assumes the user has already performed the primary step in a risk analysis, determining the levels of ground motion for the region. This section will give a brief overview of how the seismic hazard analysis was performed for the case study. Seismic hazard analyses estimate the amount of ground shaking at a particular site. Seismic hazards may be analyzed deterministically as when a particular earthquake scenario is assumed or probabilistically in which uncertainties in earthquake size, location and time of occurrence are explicitly considered (King, 1994). For this study, a probabilistic seismic hazard analysis is performed. The level of ground shaking is characterized here in terms of the Modified Mercalli Intensity (MMI). As discussed in Section 4.6, this scale, though a subjective scale, was 1 0 1 Chapter 6 Case Study: The City of New Westminster chosen due to the availability of damage relationships for the region using this scale and also the general acceptance and widespread application of methodologies which characterize ground motion using MMI. The distribution of MMI for Southwestern British Columbia is calculated from the peak ground accelerations for a return period of 475 years, or a 10% chance of exceedance in 50 years. This is the probability level used for establishing the design ground motions in the NBCC (1995) and corresponds to a peak ground acceleration of 0.2g for New Westminster. Using area sources developed by the Geological Survey of Canada for current earthquake codes (Adams, J., et. al., 1996) the peak ground acceleration (PGA) at New Westminster is statistically calculated to be 0.196g. Using the relationship between PGA and MMI developed by Newmann, this corresponds to MMI = 7.5 (Newmann, 1954). However, MMI is defined using ordinal units so a damage estimate is performed for both intensity VII and intensity VIII. Finally, it should be noted the ground motions have been calculated for firm ground. The effect of soil conditions on ground motions has not been taken into account. 6.3 Structural Inventory Development The New Westminster building inventory was completed using the Planning Department building database provided by the city. The Planning Department maintains a comprehensive database of every structure in the municipality including location, structural properties, zoning and land use information. A new database was created for this study using many of the same fields. A sample entry of data attributes is listed in Table 6.1. The minimum information required for this analysis are structural properties and geographic location. 102 Chapter 6 Case Study: The City of New Westminster Structural properties required in the database are construction material, type of framing, height (number of stories), and total floor area. The planning database already contained information on number of stories and footprint. It also contained a field for structure type. This field was not compatible with the BC31 classification but still contained useful information for determining construction material and type of framing. Inference schemes relating structure type to BC31 building type were developed using expert opinion and were spot checked using building plans on file with the city. The following are examples of inference schemes developed based on structure type: (a) IF (structure type = Concrete) and (number of stories > 9) THEN (building type = concrete frame high rise) (b) IF (structure type = Concrete Block) and (age < 1973) and (number of stories = 1) THEN (building type = unreinforced masonry low rise) (c) IF (structure type = Brick) and (number of stories = 5) THEN (building type = unreinforced masonry medium rise) (d) IF (structure type = Heavy Timber) THEN (building type = wood post and beam) (e) IF (structure type = Frame) and (number of stories < 5) and (area < 1500 sq. ft.) THEN (building type = wood light frame residential) (f) IF (structure type = Frame) and (number of stories <5) and (area > 1500 sq. ft.) THEN (building type = wood light frame low rise) Building name is a field that could be used to cross check these inference schemes. Apartment, court or manor names with a structural type = frame were assumed to be wood light frame low rise. Frame construction with religious or commercial names are classified as wood light frame commercial/institutional. In some cases, inference schemes were insufficient for assessing the building type for each structure. For example, structures listed as Laminated, Metal or Therm-wall do not 103 Chapter 6 Case Study: The City of New Westminster easily fall into any of the BC31 building types. In these cases, building plans on file with City Hall were reviewed to determine building type. A spot check of various buildings was performed to confirm inference schemes. Additional information that was pursued in the development of the structural inventory was use classification. Three primary use classifications were developed from ATC-13 and are listed in Table 6.2, with the first being general land use information and classifications becoming more specific for primary use 2 and primary use 3. City zoning maps and land use information in the existing database were used to determine primary use 1 classifications such as residential, commercial, or education. Zoning maps could also be used to confirm inference schemes. For example, a building classified WLFR because of it's structure type and area would be in a residential zone and primary use 2 and 3 would be inferred as a permanent dwelling and house respectively. The final use of the city zoning map for structural inventory development was to identify the commercial and industrial areas of the city. After applying the inference schemes and reviewal of city hall building plans, the structural inventory was still incomplete. A portion of the building inventory was selected for field survey. Commercial and industrial areas were chosen for further field study because of the various building types in these areas. This decreased the number of buildings for individual field investigation from approximately 8000 to just over 1000. By completing as much information as possible in the database prior to the field survey, the survey personnel were able to work quickly and efficiently. The database already contained street address, number of storeys and area. Field personnel performed a "windshield survey" checking the inferred building types and also completing 104 Chapter 6 Case Study: The City of New Westminster any required information still missing from the database. With the database complete, regional distribution of structural characterstics of the building inventory can be mapped. Results were aggregated according to the finest geographical region that would produce maps neither too complex nor too general. Maps aggregating damage over neighborhood, forward sortation area (3 digit postal code), or by cencus tract were deemed too large a reference for such a small municipality. Point maps produced by mapping results according to the latitude and longitude of each building were deemed too precise for the resource constraints of this study. A larger region the planning department uses as a reference for every building is block number. Blocks are typically bounded by streets. This was deemed an appropriate geographic reference for mapping purposes. Every structure in the inventory was assigned the latitude and longitude corresponding to the centroid of it's block. Characteristics of the New Westminster structural inventory can be summarized at this time. Figure 6.1 and Figure 6.2 present the building inventory on a block-by-block basis using two types of building description. Figure 6.1 shows the predominant material type used in construction on a block-by-block basis. Figure 6.2 shows the prevalent building type in each block. These maps could be used with the damage maps to locate more hazardous areas and their corresponding construction type. 6.4 Damage Estimation For a given level of shaking the damage will depend on the vulnerability of the building, which in turn depends on the kind of construction and the material used in the 105 Chapter 6 Case Study: The City of New Westminster construction. In the Southwestern BC Building Fragility Study, the vulnerability of a building to a given level of shaking is specified in a damage probability matrix (Bell, 1998). In A T C -13, the mean damage factor is treated as a random variable with a beta distribution at every ground motion intensity level. The beta distribution is discretized to give the probability of being in a certain damage state for a given ground motion (ATC, 1985). Damage probability matrices for the Southwestern BC Building Fragility Study were developed from ATC-13 for each of the building types in the as described in Section 4.5. A summary of mean damage factors corresponding to each DPM for the BC15 and BC31 classifications are listed in Table 4.6 and Table 4.7. Within the geographical information system (GIS) platform, the structural inventory is combined with the other database tables. The structural inventory database contains structural information and geographic location specific to each structure in the inventory. Each individual building type is related to a database table of building types and their corresponding mean damage factor (MDF) for varying levels of ground shaking. The GIS software, Maplnfo was used to produce maps of the geographical distribution of structural damage for MMI VII and MMI VIII. For a given MMI, each building is assigned its corresponding MDF by building type. The MDF is then averaged over each block. The average MDF for each block can be interpreted as the average damage costs for the buildings on a block as a percentage of their replacement cost. This map shows which blocks have a higher percentage of expected damage than others. Using the BC31 classification system, Figure 6.3 shows the distribution of structural 106 Chapter 6 Case Study: The City of New Westminster damage on a block-by-block basis for MMI=VII. The damage is defined by seven categories of Average Mean Damage Factor: 30%-100%; 20%-30%, 15%-20%; 10%-15%; 5%-10%; 2%-5%; 0%-2%. These ranges where chosen to show the gradation in damage in the region. At MMI=VII, the highest mean damage factors are for unreinforced masonry construction ranging from 10.2% to 12.2%. Therefore at MMI=VII, no block will have an average mean damage factor higher than 12.2%, i.e. a block composed entirely of unreinforced masonry construction. The map indicates three blocks falling within the 10% to 15% damage range. Therefore, these blocks are made up of predominantly U R M construction and are expected to sustain the most damage in the event of an earthquake. Typically, most of the blocks have an average M D F in the range of 2% to 5% indicating a composition of building types experiencing light damage (mean damage factors less than 5%) for an event with MMI=VII. Figure 6.4 is for MMI=VIII. The effect of a one-step increase in intensity is to move the damage estimate from the range of 2%-5% to 5%-10% over most of the city area. An alternate analysis was performed with the damage probabilities matrices for the BC15 classification for both MMI=VII and MMI=VIII. A separate database table was included in the GIS for relating the BC31 and BC15 classification. Those relations were developed by expert opinion and are presented in Table 6.3. The methodology was repeated for the New Westminster building inventory using the BC15 mean damage factors. Figure 6.5 and Figure 6.6 show the distribution of damage on a block-by-block basis for MMI=VII and MMI=VIII. Again, a one step increase in intensity moves the damage estimate from the range 2%-5% to 5%-10% over most of the city area. The effect of moving from the BC31 classification to the BC15 classification is minimal. Less than 3% of the blocks have a change 107 Chapter 6 Case Study: The City of New Westminster in damage estimate. From these results, it can be concluded that using the BC15 classification instead of the BC31 classification does not produce a significant change in damage estimates. Therefore, the BC15 classification system can be used for future structural inventories with little effect on the accuracy of the damage estimate. The advantage of using a GIS program to map regional damage distribution is the capability of producing maps showing varied information. Based on the level of detail in the structural inventory, the maps can be manipulated to show a variety of aggregate building damage characteristics. The example maps in this chapter show average mean damage factor for each block. However, in developing this map, a mean damage factor has been assigned to each structure. A more detailed map could be made showing the distribution of M D F for a particular block indicating the MDF for each building. This would require the database to include latitude and longitude for each building and the maps would indicate a very fine damage mesh. Alternately, the MDF could be averaged over neighborhoods or forward sortation areas (3-digit postal code) and produce a much larger mesh. Another way to highlight different information using damage maps is to weight the average MDF by square footage. A damage map using a weighted average would reflect the influence of large square footage buildings with a higher vulnerability. This is of particular interest when a loss module is included in the methodology. Replacement costs are typically available as a fraction of square footage. For a given MMI, each building would be assigned a structural MDF. The estimated damage costs in dollars could then be calculated for each structure by multiplying the MDF x total square footage x replacement costs. The GIS program could then be used to generate maps which indicate the regional distribution of dollar 108 Chapter 6 Case Study: The City of New Westminster loss associated with structural earthquake damage. 6.5 Conclusions The methodology developed in this thesis covers structural inventory development, motion-damage relationships and regional damage distribution. The case study of the City of New Westminster is an example of implementation of this methodology using the GIS program, Maplnfo. Though only briefly discussed here, the seismic hazard for southwestern British Columbia was calculated for the same probability level (10% chance of exceedance in 50 years) as the ground motions designed for in the NBCC (1995). For the City of New Westminster, the peak ground acceleration was found to be 0.196g which corresponds to an MMI=7.5. Next, a structural inventory for New Westminster was developed using the methods outlined in Section 5.3. The structural database began with a database from the Planning Department and was supplemented with information from inference schemes, zoning maps, building plans on file with City Hall, and a field survey. Finally, the regional distribution of structural damage to buildings in New Westminster was determined. The structural classification and motion-damage relationships used are those found in the Southwestern BC Building Seismic Fragility Study (Bell, 1998). Damage maps were produced showing average MDF for MMI = VII and MMI = VIII using both the BC15 and the BC31 structural classifications. These intensities were predicted based on ground motions for firm ground and do not include hazards such as soil amplification and liquefaction. It should also be noted that the expected PGA for this area corresponds to an MMI = 7.5, so the damage maps for MMI = VII and MMI = VIII would be more accurately interpreted as a lower and 109 Chapter 6 Case Study: The City of New Westminster upper bound for expected damage in the region. There are several conclusions which can be made based on the damage maps presented here. Aggregating information based on the blocks produces maps shows a level of detail in damage distribution which is neither too precise nor too general. Though the use of average M D F does not reflect the influence of square footage on the damage estimate, these maps are helpful for indicating expected regions of higher damage. Maplnfo could just as easily be used to generate a number of helpful maps, some examples are aggregate building characteristics, damage maps reflecting square footage, and the location of U R M buildings within the City. The most important conclusion to be inferred from the damage maps of New Westminster is that there is negligible difference between the damage distributions using the BC31 and the BC15 structural classifications. There is a significant amount of resources required to classify the structural inventory according to the BC31 structural classification compared to the BC15 structural classification. Therefore, the BC15 structural classification and its corresponding mean damage factors is recommended for future implementation of this methodology. 110 Chapter 6 Case Study: The City of New Westminster Table 6.1: Example Structural Inventory Database Attributes STRUCTURE ID 7824 B L O C K NUMBER S-006 CENCUS NUMBER 933209 NEIGHBORHOOD E STREET NUMBER xxxx STREET N A M E GLENBROOKE DRIVE POSTAL CODE V3L FOOTPRINT 1314 N U M B E R OF STOREYS 2 BUILDING A G E 1983 HERITAGE N STRUCTURAL TYPE F R A M E BUILDING TYPE WLFR C H E C K E D N PRIMARY USE 1 Residential PRIMARY USE 2 Permanent Dwelling PRIMARY USE 3 Houses BUILDING N A M E not applicable 111 Chapter 6 Case Study: The City of New Westminster Table 6.2: Primary Use Classifications Primary Use 1 Primary Use 2 Primary Use 3 Residential Permanent Dwelling Houses/Apartments Mobile Homes Temporary Lodging Hotels/Motels Group Institutional Housing Dormitories Commercial Retail Stores Wholesale Warehouses Financial Services Banks Personal Services Service Station/Shops Professional Services Offices Entertainment & Recreation Rests./Bars/Theater Parking Garages Industrial Manufacturing Factories Agriculture Farm Buildings Mining Mine Buildings Religion/Non-Profit Churches/Offices Government General Services Offices Police Stations Fire Stations Medical Facilities Hospitals Hospitals Ambulance Services Garages Nursing Homes Convalescent Centers Health Care Services Clinics Education Elementary Schools Secondary & Jr. Colleges Schools Colleges & Universities Schools Transportation Freight Passenger Terminals 112 Chapter 6 Case Study: The City of New Westminster Table 6.3: 31tol5 Database Material BC31 Building Type BC15 Building Type Wood WLFR WLFR WLFCI WLFCI WLFLR WLFLR WPB * Steel L M F L M F SMFLR SBMFLR SMFMR ** SMFHR ** SBFLR SBMFLR SBFMR ** SBFHR ** SFCWLR SFCWLR SFCWMR SFCWMHR SFCWHR SFCWMHR SFCI SFCWMHR SFMI SFCWMHR Concrete CFLR C F L W CFMR CFMHW CFHR CFMHW RCMFLR CFLW RCMFMR CFMHW RCMFHR CFMHW RCFIW CFMHW Masonry RMLR RMLR RMMR RMLR URMLR U R M L R URMMR U R M M R Tilt Up T U T U Precast PCLR PCLR PCMR PCMR Mobile Homes M H M H * Buildings classified as WPB were assigned to either WLFR, WLFCI or W L F L R as appropriate. ** Very few buildings of this type found in the B C building inventory. Deletion of these building types building type does not significantly affect damage estimate. 113 Chapter 6 Case Study: The City of New Westminster 0 05 1 • all others (0) Fig. 6.1: Predominant Material Type by Block 114 6 Case Study: The City of New Westminster Fig. 6.2: Predominant Building Type by Block 115 Chapter 6 Case Study: The City of New Westminster • 5 to 10 (32) I 2tO 5 (343) • Oto 2 (0) Fig. 6.3: BC31 M M I VII Average Mean Damage Factor (%) by Block 116 Chapter 6 Case Study: The City of New Westminster • Oto 2 (0) Fig. 6.4: BC31 M M I VIII Average Mean Damage Factor (%) by Block 117 Chapter 6 Case Study: The City of New Westminster 0 0.5 1 • 5 to 10 (34) • 2 to 5 (341) • Oto 2 (0) Fig. 6.5: BC15 M M I VII Average Mean Damage Factor (%) by Block 118 Chapter 6 Case Study: The City of New Westminster 0 0.5 1 • Oto 2 (0) Fig. 6.6: BC15 M M I VIII Average Mean Damage Factor (%) by Block 119 C H A P T E R 7 Conclusions 7.1 Summary There is no question that southwestern British Columbia is at risk of significant losses due to major earthquakes. While it is impossible to predict the exact time and location of an event, it is plausible to insure ourselves for the financial loss and to improve the quality of construction of our built environment to resist the damaging effects of an earthquake. In order to do so, a loss estimation methodology which relates ground motion to monetary and non-monetary measures of loss needs to be developed for the region. Results of an economic loss study could be used for emergency response, recovery and preparedness as well as prioritizing regions for risk mitigation. One of the steps towards quantifying losses is to determine the regional distribution of damage. This thesis presents a methodology for the evaluation of structural earthquake damage to buildings in southwestern British Columbia. The steps in a risk analysis performed by this methodology are: (1) development of building inventory, (2) selection and application of motion-damage relationships, and (3) regional distribution of damage using a geographic information system. These steps do not constitute an entire loss methodology as they do not include a seismic hazard analysis or the modules necessary for determining a regional loss distribution. 120 Chapter 7 Conclusions Chapter 2 reviews the overall steps in a loss methodology, focusing on those steps relevant to this thesis. Different definitions of damage and characterization of ground motion were explored in order to understand the various ways motion-damage relationships are defined. In general, motion-damage relationships are probabilistic distributions of damage at specified ground motion intensities. Some studies rely on expert opinion to derive relationships while others use engineering parameters to quantify damage. Also included in this chapter was a review of previous loss studies performed for Southwestern British Columbia. Chapter 3 reviews existing building type Classifications suitable for the study region. Classification systems categorize buildings into groupings with similar risk characteristics. Groupings can be based on the structural framing or the use classification. This chapter does not conclude with the structural classification system chosen for this methodology as that decision should be made with consideration of the available motion-damage relationships discussed in the next chapter. Chapter 4 reviews existing damage functions and chooses those most appropriate for application in the study region. Two methodologies and their respective damage functions are reviewed, ATC-13 and FEMA-NIBS. The former is based on expert opinion and the latter relies on engineering parameters to define motion-damage relationships. Two additional sources of damage functions were reviewed, the BC Preliminary study and the Southwestern BC Seismic Fragility Study. A comparison of appropriate damage functions is performed by comparing mean damage factors for each level of seismic intensity and a classification system and corresponding damage functions for southwestern BC are chosen based on the results of 121 Chapter 7 Conclusions that comparison. The implementation of the methodology is discussed in Chapter 5. Structural Inventory development on a municipal level is discussed with a focus on the minimum information required to proceed with the damage estimate. Possible sources of inventory information are identified. The chapter concludes by explaining the methodology for application of the damage functions and classification system chosen in Chapters 4 and 5, to the building inventory of southwestern British Columbia. The methodology suggests inputting the building inventory and damage functions into a geographic information system to perform the regional damage estimation analysis. Chapter 6 presents a case study. A damage analysis was performed for the City of New Westminster, British Columbia. A description is given of the structural inventory development. The motion-damage relationships and classification systems of Chapters 4 and 5 were then applied to the inventory using the GIS software, Maplnfo. Results of the analysis are presented. Finally, Chapter 7 presents a summary, conclusions and recommendations for future work. 7.2 Conclusions The available literature on earthquake damage in Southwestern British Columbia indicates the potential for high economic losses in the event of an earthquake. To better assess those losses, the literature indicated the need for a more detailed study of seismic performance of representative buildings in the region. The primary goal of this study was to develop a 122 Chapter 7 Conclusions detailed methodology for estimating the seismic performance of the southwestern British Columbia structural inventory. Damage functions are ideally derived from observed damage due to earthquakes in the region of interest. However, very little damageability data is available for Southwestern British Columbia. Consequently, motion-damage relationships must be inferred from either expert opinion, as in ATC-13, or from engineering parameters, as in FEMA-NIBS. The main advantage of using an engineering parameter methodology is to erase the subjective nature of using Modified Mercalli intensities to characterize ground motion. The main advantage in using an expert opinion based model is the availability of several comprehensive sets of damage functions already in existence for the region. For this study, an expert opinion based methodology was chosen over an engineering parameter model due to the availability of motion-damage relationships specific to the region. Damage probability matrices from the Southwestern BC Seismic Fragility Study were chosen for use in the methodology. These functions were chosen over other available sets of motion-damage relationships because of (1) the comprehensive classification system used to characterize the BC building inventory, and (2) the general agreement with damage functions from similar studies, especially at certain levels of intensity. The building types in this study were chosen based on their distinct damage characteristics and, though ATC-13 was used as a guide in their development, they are not a subset of California construction types. Of the available damage functions, the BC31 and BC15 classification systems and corresponding damage probability matrices best characterize the vulnerability of the local building inventory. 123 Chapter 7 Conclusions The importance of the case study was to test the application of the methodology. New Westminster is a small municipality in southwestern BC with a vested interest in determining the estimated losses in their city if an earthquake were to occur. Using the steps outlined in Chapter 5, a structural inventory was developed based on Planning Department records but made complete using other techniques detailed such as inference schemes and a field survey. The methodology then uses the BC15 and BC31 damage functions to assess the damage distribution of the New Westminster building inventory for intensities VII and VIII. By performing the damage analysis at these levels of intensity, these maps show the range of expected damage associated with ground motions having a 10% chance of exceedance in 50 years. They provide a visual reference of where greater damage is expected in New Westminster. Two main conclusions can be made based on the New Westminster damage maps. First, aggregating damage by blocks produces maps where the level of detail in damage distribution is neither too precise nor too general. Second, there is negligible difference between the damage distributions using the BC31 and the BC15 structural classifications however the resources required to develop the structural inventory based on BC15 building types is much less that those required to use the BC31 building types. The sequence of steps recommended for performing a structural damage estimate for southwestern BC be as follows: 1. On a municipal level, develop a structural inventory using the BC15 classification system. Inference schemes should be developed using expert opinion specific to the municipality. The minimum information required in the database for each building is geographic location and structural framing type. 124 Chapter 7 Conclusions 2. From the results of a seismic hazard analysis of ground shaking, determine the levels of ground motion to be considered in the damage estimate. 3. For each building in the inventory, apply the appropriate BC15 motion-damage relationships for the MMI prescribed in Step 2. 4. Using a GIS program, aggregate the results of the damage analysis on a block level. Produce maps showing the geographical distribution of damage. 7.3 Future Work This research focused on the review of regional seismic hazard and risk analysis methodologies and customizing a methodology for estimating earthquake damage to the general building stock of Southwestern British Columbia. The seismic risk analysis primarily considers hazards due to ground shaking, while effects of local site conditions are not included at this time. Structural classifications and motion-damage relationships are developed specific to the region. Finally, a methodology for estimation of building damage due to ground shaking and a case study is presented. While the objectives of the report have been met, there are still considerable improvements that can be made to the methodology. Improvements can be made in the following steps of the methodology: seismic hazard, motion-damage relationships, and loss estimation. In the case of seismic hazard, this study only considered hazards due to ground shaking. Effects of local site conditions such as soil type, liquefaction, landslide and fault rupture are all major factors in the overall seismic hazard to a region. The effects should be considered in future studies. 125 Chapter 7 Conclusions In the area of motion-damage relationships, current literature debates the use of subjective scales to characterize ground motion. In response to this, state of the art methodologies have been developed which use engineering parameters to define demand, response and fragility curves. These methodologies were'dismissed within the scope of this study but are recommended for further investigation should more resources become available. Also, there have been no major recent earthquakes in southwestern British Columbia. The damage functions used in this study are based on expert opinion, and should be verified and calibrated with observed damage from an earthquake occurring locally. Finally, a continuation of the methodology to include loss estimation is highly recommended. Losses modules including monetary and non-monetary damage should be used thus addressing the overall impact of an earthquakes on southwestern British Columbia. 126 REFERENCES Adams, J., Weichert, D. H. , Halchuk, S., Basham, P. W. (1996). "Trial Seismic Hazard Maps of Canada - 1995: Final Values for Selected Canadian Cities". Geological Survey of Canada, Open File 3282. National Resources Canada. Algermissen, S. T. and Steinbrugge, K. (1984). "Seismic Hazard and Risk Assessment; Some Case Studies." The Geneva Papers on Risk and Insurance. Volume 9, Number 30. Applied Technology Council (1985). Earthquake Damage Evaluation Data for California, ATC-13. Redwood City, California. Applied Technology Council (1988). "Rapid Visual Screening of Buildings for Potential Seismic Hazards, ATC-21." Federal Emergency Management Agency (FEMA) Publication 154. Washington, D.C. Applied Technology Council (1989). A Handbook for Seismic Evaluation of Existing Build-ings, ATC-2 2. Redwood City, California. Applied Technology Council (1989). Seismic Evaluation of Existing Buildings: Supporting Documentation, ATC-22-1. Redwood City, California. Anagnos, T., Rojahn, C , and Kiremidjian, A. S. (1995). "NCEER-ATC Joint Study on Fragility of Buildings." National Center for Earthquake Engineering Research Technical Report NCEER-95-0003. Buffalo, New York. Arnold, C , et al. (1989). "Establishing Programs and Priorities for the Seismic Rehabilitation of Buildings, Supporting Report." FEMA Publication 173. Washington, D.C. BC Stats (1999). "Greater Vancouver Regional District - Community Facts"; Source: http:// www.bcstats.gov.be.ca/data/dd/facsheet/cfl 70.pdf. Material dated: March 4, 1999; retrieved April 1999. BC Stats (1999). "Capital Regional District - Community Facts"; Souce: http:// www.bcstats.gov.bc.ca/data/dd/facsheet/cf040.pdf. Material dated: March 4, 1999; retrieved April 1999. 127 BSCC (1997). NEHRP Guidelines for the Seismic Rehabilitation of Buildings, by the Applied Technology Council, ATC-33, for the Building Seismic Safety Council and F E M A , F E M A -273, Washington D.C. Bell, L. (1998). Southwestern British Columbia Building Seismic Fragility Study. University of British Columbia, B.C. Canada Mortgage and Housing Corporation (CMHC), (1989). "Assessment of Earthquake Effects on Residential Buildings and Services in the Greater Vancouver Area." Report prepared by the National Research Council of Canada, Ottawa. Cornell, A. (1968). Bulletin of Seismological Society of America, Vol. 58. pages 1503-1606. EERI Ad Hoc Committee on Seismic Performance (1994). "Expected Seismic Performance of Buildings." Earthquake Engineering Research Institute. Oakland, CA. F E M A (1997). "HAZUS - Technical Manual Volume I" National Institute of Building Sci-ences Document Number 5201. Federal Emergency Management Agency. Washington, D.C. F E M A (1997). "HAZUS - Technical Manual Volume II" National Institute of Building Sci-ences Document Number 5201. Federal Emergency Management Agency. Washington, D.C. F E M A (1997). "HAZUS - Technical Manual Volume III" National Institute of Building Sci-ences Document Number 5201. Federal Emergency Management Agency. Washington, D.C. F E M A (1997) "HAZUS - Users Manual" National Institute of Building Sciences Document Number 5201. Federal Emergency Management Agency. Washington, D.C Fernandez, J., Hoefer, G., Tucker, B., Ventura, C , and Villacis, C. (1996) Proceedings from the Eleventh World Confererence on Earthquake Engineering. Paper No. 2064. Holdstock, D. A. (1998). Journal of Computing in Civil Engineering, Vol. 12, Issue 1, pages 1 to 4. King, S. A., and Kiremidjian, A. S. (1994). "Regional Seismic Hazard and Risk Analysis Through Geographic Information Systems." The John A. Blume Earthquake Engineering Center Report No. 111. Department of Civil Engineering, Stanford University. Stanford, California. 128 King, S. A., Kiremidjian, A. S., Basoz, N., Law, Kl , Vucetic, M . , Doroudian, M . , Olson, R A . , Eidinger, J. M . , Goettel, K.A. , and Horner, G. (1997). "Methodologies for Evaluating the Socio-Economic Consequences of Large Earthquakes." Earthquake Spectra. Volume 13, Number 4, pages 565-584. Kircher, C. A., Nassar, A. A., Kustu, O., and Holmes, W. T. (1997). "Development of Building Damage Functions for Earthquake Loss Estimation." Earthquake Spectra. Volume 13, Number 4, pages 663-682. Kircher, C. A., Reitherman, R. K., Whitman, R. V., Arnold, C. (1997). "Estimation of Earthquake Losses to Buildings." Earthquake Spectra. Volume 13, Number 4, pages 703-720. Maplnfo Corporation (1997) "Mapinfo Professional GIS Software" McCormack, T. C , and Rad, Franz N. (1997). "An Earthquake Loss Estimation Methodology for Buildings Based on ATC-13 and ATC-21." Earthquake Spectra. Volume 13, Number 4, pages 605-622. Munich Reinsurance Company of Canada and Rescan Consultants, Inc. (1990). "A Study of the Economic Impact of a Severe Earthquake in the Lower Mainland of British Columbia". A publication of Munich Reinsurance Company of Canada. National Research Council of Canada (1995). "National Building Code of Canada". Ottawa. Natural Resources Canada (1999). "Geological Survey of Canada - Some Significant Earthquakes Felt in the Vancouver and Victoria Areas". Source: http://www.pgc.nrcan.gc.ca/ seismo/hist/vanlist.htm. Material dated February 2, 1998; retrieved April, 1999. Newmann, F. (1954). "Earthquake Intensity and Related Ground Motion." University Press, Seattle, Washington. Reitherman, R. (1985). "A Review of Earthquake Damage Estimation Methods." Earthquake Spectra. Volume 1, Number 4, pages 805-847. Risk Management Software, Inc. (RMS) and California Universities for Research in Earthquake Engineering (CUREe) (1993). "Assessment of the State-of-the-Art Earthquake Loss Estimation Methodologies." FEMA Publication 249. Washington, D.C. Rodgers, G. C. 1983. "Seismotectonics of BC." University of British Columbia. Vancouver, British Columbia. PhD Thesis, 227 p. 129 Rojahn, C , King, S. A., Scholl, R. E. , Kiremidjian, A. S., Reaveley, L. D., and Wilson, R. R. (1997). "Earthquake Damage and Los Estimation Methodology and Data for Salt Lake County, Utah (ATC-36)." Earthquake Spectra. Volume 13, Number 4, pages 623-642. Shinozuka, M . , Chang, S. E. , Eguchi, R. T., Abrams, D. P., Hwang, H. H. M . , Rose, A. (1997). "Advances in Earthquake Loss Estimation and Application to Memphis, Tennessee." Earthquake Spectra. Volume 13, Number 4, pages 739-758. Singhal, A. , Kiremidjian, A. (1996). "Method for Probabilistic Evaluation of Seismic Structural Damage." Journal of Structural Engineering. Volume 122, Number 12, pages 1459-1467. Thiel, C , Zsutty, T. (1987). "Earthquake Characteristics and Damage Statistics." Earthquake Spectra. Volume 3, Number 4, pages 747-792. Tiedemann, H (1992) "Earthquakes and Volcanic Eruptions - A Handbook on Risk Assessment." Swiss Reinsurance Company. Zurich, Switzerland. Vasudevan, R., Kiremidjian, A. S., and Howard, H. C. (1992). "An Integrated Inventory Methodology for Seismic Damage Assessment." The John A. Blume Earthquake Engineering Center Report No. 102. Department of Civil Engineering, Stanford University. Stanford, California. Ventura C.E. and Rezai, M . (1997). "Classification and Damage Estimation of Buildings in British Columbia." Report to IRAS Insurance Users Group, Dept. of Civil Engineering, University of British Columbia, Vancouver, B.C. Whitman, R. V., Anagnos, T., Kircher, C , Lagorio, H. J., Lawson, R. S., and Schneider, P. (1997). "Development of a National Earthquake Loss Estimation Methodology." Earthquake Spectra. Volume 13, Number 4, pages 643-662. 130 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0064091/manifest

Comment

Related Items