Paul F. Fratessa
There were 63 deaths, $10 billion in direct and indirect losses and over 27,000 (California SSC, 1991a) structures damaged, and the knowledgeable consensus of experts is that there were few if any surprises. Those same experts caution against overconfidence, because the earthquake did not, in fact, provide a true test of earthquake resistant structures. Assuming these experts are correct, then results of a major earthquake close to a major urban area would be grim, and the Loma Prieta earthquake should be treated as a wake-up call.
Because there were few surprises, it seems logical that knowledge of building performance in earthquakes has advanced to the point where we can forecast building performance in the next event based on lessons learned not only from Loma Prieta but from all past earthquakes. If this level of knowledge is available, then it is clear that the mitigation of the hazard depends on the will and the means to get it done. This represents a challenge to the public to find the way to better safety. However, there is an equal challenge to the design and research community to be responsible to the public's trust and deliver the tools to make the mitigation effective.
The Loma Prieta earthquake also pointed out that the community was lucky and cannot be overconfident when viewing the prospects of a full service event. But there were indications of some potentially fatal flaws in the extension of the observed performance beyond the one event. Clearly the performance was good without surprises, but equally as clear is the perspective on what is not known and what the impact of that lack of knowledge might have on areas outside of California.
This paper explores the lessons learned from the Loma Prieta earthquake as
they apply to buildings and then extrapolates those lessons to areas outside of California. The paper is based on lessons learned from both research and observation. These lessons were extracted from the research report summaries and from discussions with practicing professionals familiar with the Loma Prieta earthquake. The paper does not focus on the details of research or observations but on what are perceived to be the major lessons.
Much has been written about the fact that there were "no real surprises" in the Loma Prieta event. Well-designed and well-constructed buildings performed well, while poorly designed and constructed structures did not. Implicit in this statement is the thought that someone can define "poorly designed and constructed." We can place unreinforced masonry buildings in this category, and they have long been identified as a problem. It is not as easy to identify other hazardous buildings as a class of structures. Some insight into what might be hazardous can be found by looking at damage statistics, which show that of the over 27,000 structures damaged, only 900 were reported to be of unreinforced masonry. The overwhelming majority of the damaged structures, classified by type, were wood-frame residential, a classification long ignored and presumed not to be potentially dangerous or the source of economic loss.
When considering the true impact of the event, economic loss due to loss of functionality at the work place became a dominant focus. While the loss of building function did not surprise the engineering community, building owners and public agencies were surprised by the impact that loss of functionality had on communities. This was so much so that for the first time in California history, the governor acknowledged the need for functionality in state-owned buildings after an earthquake and recognized several key matters necessary to ensure seismic safety. In Executive Order D-86-90, Governor Deukmejian stated in part:
... The Director of the Department of General Services shall prepare a detailed action plan to ensure that all facilities maintained or operated by the State are safe from significant failure in the event of an earthquake and that important structures are designed to maintain their function following an earthquake...
The order goes on to state that the plan should among other things:
... (c) Seek independent review of structural and engineering plans and details for those projects which employ new or unique construction technologies; and
(d) have independent inspections of construction to ensure compliance with plans and specifications...
For the first time, functionality was introduced as a performance goal, and quality control of design and construction were identified as keys to attain performance goals.
Within this one document (EO-D-86-90) are the themes of some of the Loma Prieta earthquake's real lessons. Engineers must learn to recognize the need for post-earthquake functionality and the fact that our codes are intended as life safety provisions only. Legislation responding to the Loma Prieta earthquake required the California Seismic Safety Commission, in cooperation with the State Architect, to "develop a State policy on acceptable levels of earthquake risk for new and existing state-owned buildings and submit their policy to the Legislature for consideration by January 1, 1991." That document, Report No. SSC 91-1 (California SSC, 1991b) was submitted to the legislature in January of 1991 and has yet to be acted upon.
It needs to be understood that code concepts cover normal structures within the intent of the performance goals of the codes. Special considerations need to address special structures in order to obtain performance. Since keys to oversight of that special consideration are independent peer review and independent inspection, these should become mandatory for such projects.
For a magnitude 7.0 event, strong ground shaking is usually expected to last about 20 seconds. The Loma Prieta earthquake was an unusual event as it had about 10 seconds of strong shaking and no clear surface-fault rupture. Seismologists are studying this phenomenon as they study the ground shaking that triggered numerous measuring devices yielding a plethora of valuable data. These data, when properly synthesized, will certainly yield valuable information toward predicting ground motion and damage distribution in future earthquakes. Certain anomalies are generally found in damage distribution and become the focus of intense study primarily by the earth sciences community. When these studies show that spectral design curves can be reliably modified from those currently in use, then this will be of distinct value to the design profession.
Despite the expressed lack of surprises, there was still much to be learned from the earthquake. While damage to types of structures and the effects of soft soils on building performance were predictable, there lingers some question as to whether longer and more-intense shaking generated by a similar magnitude earthquake closer to an urban area could have dramatically different results. Use of the ground-motion records and better understanding of the fault mechanisms should lead to better knowledge of earthquake risk outside of the Loma Prieta area.
Focus of Research
Disasters tend to generate a frenzy of activity in the research community. There is always a renewed interest in solving the problem and generally a limited amount of resources to fund the needed research. After the Loma Prieta earthquake, the August 1990 Earthquake Engineering Research Institute newsletter reported that the National Science Foundation and the U.S. Geological Survey awarded approximately $4.1 million in grants to do various studies. While learning from earthquakes, researchers should be studying the effects that are unique to that earthquake, yet it is interesting to note the focus of those grants. Twenty-four percent were focused on soil related topics, with an additional 10 percent on site-response issues. In sum, about 35 percent focused on geological and seismicity issues. Five and one half percent addressed evaluation and retrofit issues, and 2 percent addressed unreinforced masonry. Of note is that 2 percent of the awards money went to housing risk observations. The largest of the grants, 3.3 percent of the money, went to Risa Palm and Michael Hodgeson at the University of Colorado for work on "The purchase of Earthquake Insurance in California." Finally, 78 percent of the grants went to individuals associated with a university, and 10 percent went to practicing design professionals.
This summary of research awards is not a critique but is presented to focus the attention of the reader on evaluating what research seems to be needed as a result of the earthquake and where and how that effort might be best spent.
For research to be effective, it is essential that it be focused on the goals of a program. If one views research as the advancement of knowledge, then the studying of a subject for better understanding will in time be a benefit to the analysis and design of buildings. However, designers often have little patience with the research efforts, because the results do not translate directly into code language. There are short-term demands for research, such as the California Department of Transportation (CALTRANS) retrofit studies, which must be accomplished on a priority basis. Beyond those types of studies, there needs to be a more direct effort with a nationwide focus that will lead to advancements in knowledge in areas where it will have the most impact.
This paper pursued two avenues of study in order to develop an overview of the lessons learned for buildings from the Loma Prieta earthquake. First, lists of the research papers available through October 1992 were reviewed. These lists were available through the National Information Service for Earthquake Engineering/Loma Prieta Clearinghouse Project. The lists were augmented by personal referrals to ongoing research, in some cases not directly related to the Loma Prieta earthquake. Second, a select group of practicing engineers from throughout the United States was polled for their thoughts on the significant lessons that can be learned from the earthquake.
The results of these two avenues of study (although not all inclusive) were then synthesized into broad categories and findings as they related to the earthquake. Viewing these findings, the appropriate lessons were then extracted as they apply throughout the United States. The paper is first organized into a discussion of the relationship of research to observation and codes and then into a series of findings, which emerged from the study. Within each finding, there are not so much lessons but discussion and opinions on the subject topic. The discussion draws on the study as well as the author's opinion. The paper concludes with the lessons extracted from the various topics that apply throughout the United States.
THE ROLE OF RESEARCH IN THE PRACTICAL LESSONS PROCESS
Numerous research programs have sprung up, often focusing on the same subject and each claiming to be that which will pull together and coordinate the activity. Along with those programs come the questions of how effective they are and how the research will be used. Then there is the latest buzz word in the technical world, "technology transfer." When someone has developed technical information containing knowledge on a subject, that knowledge is not automatically transferred by presenting a paper on the subject. Unfortunately, the mere availability of the knowledge does not necessarily transfer the knowledge or, in terms of seismic design, become useful in solving a design problem. Not all research is accepted out of hand, and it generally goes through a consensus process before it becomes an accepted theory. So how is research viewed by those who are in some ways the benefactors?
Researchers seek analytical solutions or definitions of parameters from a postulated problem.
Engineers seek a better definition of performance of materials.
Building officials seek a prescriptive document, which is primarily black and white.
Building owners wonder what they are paying for and expect a functioning building in the post earthquake environment.
Webster's Intermediate Dictionary (1977) says:
research ... investigation or experimentation aimed at the discovery and interpretation of facts, revision of accepted theories or laws in the light of new facts, or practical application of such new or revised theories or laws.
As clear a definition as this may be, it is still heard differently depending on the listener. Researchers may view it as a way of life, while engineers complain that there is not enough information to provide precise answers. Funding agencies have a tendency to look at the complexity and sophistication of the research as an indication of its usefulness. And then, of course, the ultimate passing of the buck comes when information is not disseminated, or technology transferred.
In fact, research in the broadest sense is the backbone of our learning curve. Research does not always lead directly to an improvement in codes. Research, when viewed through experienced eyes, often improves the understanding of the earthquake phenomena by improved methods of analysis, improved understanding of capacities, and the confirmation of theorized performance characteristics. In looking back to the demand for research to mitigate the earthquake hazard, to be effective the research must have the potential to improve the understanding of performance but should not necessarily lead to a more complex answer to a simple problem. It should also be applied to the areas where the most impact can be gained.
Illustrations of the role of research generated by the Loma Prieta earthquake were the CALTRANS tests of retrofit confinement for concrete columns. Research was done for the need for short-term information for a limited number of special conditions (Roberts, 1990a, b; CALTRANS, 1991). Those tests have opened the research door to ways in which concrete can be confined in existing non-ductile frames. A consensus is not yet on the table for more wide use of that information. However, in this case, research was driven by a short-term, limited scope for a specific need. The extension of those findings has yet to come.
More broadly, research comes from the need to solve a yet unsolved problem or to expand our understanding of some aspect of seismic design. Engineers can often be impatient with the research community when researchers appear to be off on highly technical but broadly inapplicable subjects. Researchers must find a way to share their knowledge so that we are building a knowledge in a meaningful way.
There is no greater example of this than the potential windfall of information generated by the Strong Motion Instrumentation Program recording devices. A significant number of records are available for study through the California Division of Mines and Division of Geology. Not only do these records give researchers needed data for studies of the anomalies, but they give an opportunity to test results against current practice to see if it is really on the right track for a full-blown earthquake.
When viewing the broad potential of studying ground-motion records, there is a need for practical thinking. How should the spending of our research dollars be prioritized when considering what areas of research will provide the most return in the mitigation of the earthquake problem? As an example, much research attention is given to new construction and more advanced special designs of special structures; however, these structures are not the majority of the inventory in this country. These structures are also not the ones that generated the losses in the Loma Prieta earthquake. Should research follow the demand to mitigate? And if so, how does this theory relate to the uneconomical costs of retrofitting?
Is it appropriate for research funding to go to the specialty structural issues when they do not represent the vast majority of the structures subject to seismic
loading? Perhaps it is time to rethink this issue. Should not the cost of providing the research for specialty construction fall on those who will most directly benefit from it, such as the owner? CALTRANS in essence has set that example. They are building high-use specialty structures and are doing the research to support their design concepts.
The Loma Prieta earthquake has left many opportunities for research. The emergence of research coordinating committees is a reflection of the concern for the need to coordinate efforts. As these efforts are coordinated with the intent to provide a forum for the dissemination of the results, it will be important to weigh the priorities of the research such that the results will truly be of maximum benefit in solving real problems and really mitigate the earthquake hazard. Patience on the part of practicing engineers will be rewarded if the coordinated research effort starts yielding usable results that, with the test of time, can improve the designer's understanding of the earthquake phenomena.
THE ROLE OF OBSERVATIONS IN THE PRACTICAL LESSONS PROCESS
The backbone of the California practice of seismic design is based on the observation of the effects of earthquakes on structures. These observations led to a concept of how buildings ought to be designed and eventually to the first code provisions. It is important to note that the observations needed to be tempered with research in order to bring about the advancement of rational design provisions. The ongoing process of observations tempered with research has continued to advance seismic codes. The downside of this process has been that as codes become more and more based on research, and as they become more complete, they also become more like a cookbook. This may be fine for those only concerned with the ''does it meet code'' syndrome; however, that approach does not relate back to the prime teacher of good performance. That teacher is not the code but observation.
Many people who had never experienced an earthquake or really seen the destruction of even a moderate event became observers during the Loma Prieta earthquake. Like the media, their attention was riveted on what failed or what looked spectacular. Observation of what did not happen can be as informative as viewing the obvious disasters. Much attention was focused on pounding between adjacent buildings. An example of that focus is the excellent work of Kasai and Maison (1990) in defining the parameters to assess the impact of pounding. The report discusses pounding damage and defines types of damage. It locates within a limited study area the various types and locations of observed pounding damage. What is not contained in the report is a true statistical analysis of the entire building stock within the study area and data on what pounding did not occur. It is also useful, and sometimes puzzling, to view separations between buildings that did not pound.
Observations tend to lead to quick code fixes in an effort to get rid of an obvious problem. The Loma Prieta earthquake showed numerous examples of soft-story structures. Soft-story structures are known to perform poorly. However, there were many soft-story structures that did not suffer damage or suffered damage in the soft story only, without damage to the upper stories. Is the appropriate response then to prohibit soft stories? Or should engineers evaluate the damping effect the softness has on eliminating force levels above? The codes, while not outlawing, have reacted by putting limits on soft stories that require additional analysis. In designing retrofits, many designers are ignoring the impact that stiffening the base has on the upper stories.
The Marina District of San Francisco was a media favorite during the earthquake. Observations were easy to make. There was liquefaction and there were soft stories. As noted above, a quick reaction to the soft stories was to outlaw them, however, the observation, astute as it may be, does not solve the problem of what to do with the existing soft-story structures that may have not been damaged. Was it the soft story or the liquefaction? Extending this line of thinking to liquefaction, we know we cannot outlaw it; however, what design provisions could be developed in areas of liquefiable soils? In fact, code provisions for liquefaction are not what is needed. Mitigating the impact of liquefaction on existing structures would be useful, but for undeveloped sites, liquefaction is a land-use planning matter much like the Alquist Priolo is for fault offset.
In summary, observations are excellent at focusing attention on critical issues. Often, short-term measures are warranted to mitigate the impact of what those observations discover. More often than not, the observations need some thought and discussion with the research community to see if the observations match past knowledge or if they are pointing at an anomaly. With time, the observations will lead to better knowledge and are often the trigger for research with a high potential for useful results.
Sixty-three people are dead, and engineers are saying that structures performed about as expected. In excess of 27,000 buildings are damaged, and engineers are saying that is about what was expected. Public schools survived virtually unscathed, and again engineers say that is about what they expected. This was the first real confrontation with the performance intent of the modern seismic code. Before the Loma Prieta earthquake, "earthquake proof" was still a term that was used. Post-earthquake things have changed. The realization that buildings could be damaged and still be in complete compliance with the building code was now at hand. What were the expectations of the general public and the building owners?
Regardless of the expectations, there was a new reality. Buildings were not designed to be functional after earthquakes, they were designed only to be life safe. Building owners weighed the consequences of that, and soon the financial and insurance industries also saw the broad implications of such performance. It also became clear that retrofits for less than code levels were targeted for life safety only and that preservation of a historic resource was not the intent of applying the less than code approach. Retrofitting of historic buildings for less than preservation performance risks the loss of the assets.
Acceptable Risk Policy
The California Seismic Safety Commission was requested by a 1989 state law to develop a recommended policy on acceptable levels of risk in state-owned buildings (Fratessa and Turner, 1991). This law (SB 920) was signed by the governor about one month before the Loma Prieta earthquake. The commission responded with its recommended policy (California SSC, 1991b). This policy has had widespread attention, especially where it might apply as a general policy for all building owners.
The essence of the policy was to involve the building owner in the decision as to what level of seismic risk the owner wishes to accept. Clearly, nothing below life safety in new construction would be acceptable to the public. However, improved life safety for existing buildings was put forth as a possibility. A time frame of re-occupancy as a barometer of judging degrees of functionality was proposed. The bottom line was that engineers have traditionally assumed the role of the determiner of acceptable risk and it was now time for the owners to rightfully make those determinations. The engineer is responsible for the proper definition of those risks.
Clearly the Loma Prieta earthquake became the risk-enlightenment quake of the time. For the first time, owners were faced with the real decisions as to what level of risk they were willing to take and inevitably pay for if the cost was different. When this was extended to the post-earthquake repair and retrofitting of damaged buildings, it became not only the owners' issue but also that of the agency, financial interests, and the insurance company.
In first discussing acceptable risks, it became apparent that terminology was a key issue for understanding the subject. It seems that when the topic of seismic risk is raised, it is viewed from the perspective of the individual who often would place the individual's concerns or special interests at the top of the issue list.
To clarify the issue, the California Seismic Safety Commission defined the terminology. Later in this paper, this definition will become useful in discussing the impact of the earthquake, so it is developed briefly for reference.
Definition of Seismic Risk
Basically, the concept of risk is broken into two components: environmental risk and building performance risk. Environmental risk is that peculiar to the site, including the geology and seismicity. The type of shaking (spectral content), the intensity, the duration, and any special effects are considered. Liquefaction would also be an environmental risk, as would seismically induced landslides or soil instability. For any site, the seismological and geological communities must be able to define this risk.
Given the environmental risk, which is site specific, the performance risk has to be evaluated. This is based on the structure and involves the determination of the performance characteristics of that structure. The Loma Prieta earthquake moved that evaluation from just the safety or collapse mitigation to the determination of the level of damage and, therefore, post-earthquake occupancy potential. This latter functionality requirement puts a strain on the level of knowledge of the engineering and research community. Not only is the character of the ground motion critical but the performance characteristics of structural members in the inelastic range must be confidently known.
No single realization could have more significance on the design community not only in California but throughout the United States. As discussed later, this concept places the need and responsibility to make judgments on levels of acceptable risk on a clear platform.
Within California there is a considerable variation in capabilities of building departments. Depending on the size of the service area, there may be anywhere from a single, multipurpose individual who is expected to know and see everything to the large agencies with significant staff and, in some cases, even seismic-safety divisions. The degree to which the departments exert influence and control over projects also varies—often by choice but also by expertise, especially when the expertise is limited by the size of the department.
Tools and Technology Transfer
Departments are given tools to work with, which are called codes and technology transfer. In the specialized area of seismic design, it is particularly difficult to interpret, much less enforce, the code if one does not have a background in seismic codes. Literature is available and design seminars are offered in certain locations, but considering the other responsibilities of the job, it is virtually impossible for many officials to become completely conversant in seismic code matters. The ideal solution would be to have the code as clear as a cookbook and easy to interpret.
Building Damage Assessment
Immediately after the earthquake, various affected agencies responded to the crisis quite well considering the magnitude of the event and some of the pockets of localized damage. The state of California's Office of Emergency Services was called to test its well-developed response system and, as with any such event, was able to advance its knowledge. Given the competent initial response, the dust had to settle, after which new issues quickly arose. With the help of damage assessment volunteers and outside assistance from other agencies, the tagging of buildings for continued occupancy was accomplished. The tagging procedure and the lessons learned are discussed in Structural Engineers' Association of California (SEAC) (1991).
After this initial sorting out of the dangerous structures, a much more formidable task needed to be faced. That task was determining what to do with the structures that were tagged or structures that were in need of retrofitting regardless of the earthquake performance. In summary, the principal issues were
What were the criteria for allowing the demolition of a building?
What were the criteria for the repair of earthquake damage?
When did the damage warrant repair, and when should it mandate retrofitting?
What were the criteria to allow for the removal of red or yellow tags and therefore re-occupancy of the building?
How does one deal with the sometimes conflicting requirements of the disaster assistance agencies, where the issue may be to rebuild the way it was?
There is a political issue surrounding the subject of tagging, but the California Office of Emergency Services demonstrated that there is a manner in which the actual tagging procedure can be effectively managed. Several outstanding papers are available on the experiences of emergency volunteers and their efforts to survey and tag potentially damaged buildings. The entire emergency response effort will not be reviewed here, although there are several practical spin-offs that were faced and can be learned from. Given the diversity of the assessment teams and the lack of knowledge of how they should be deployed, some local officials will be faced with a wide disparity in tagging consistency. Records show that a particular building in San Francisco was tagged four times by different individuals. Twice it was tagged yellow, then it was tagged red, and finally it was tagged yellow again before the tag was removed. In the interim, only some temporary shoring in one location had been placed. The Office of Emergency Services program is clear on training the volunteers and the authority under which their work is done. What happens after the initial response may depend on the municipality affected. The issue really is, "what is next after the tagging?"
The Initial Assessment
The initial pass in the damage assessment effort is to get hazardous buildings tagged so that they do not continue as a life-safety threat. The intent is to have more-detailed inspection where the structure is complex or the evaluation cannot be made in depth due to various constraints. The initial pass is most effectively made by teams trained to accomplish this task. This model is well established in California.
The second level of review is most effectively done by engineers experienced in seismic design and observation. With such experience, the evaluation should establish safety or use goals agreed upon in advance by the local jurisdiction. It is paramount that those who have had to deal with this before share their information with those who may someday face the issue.
A Question of Perspective
Clearly the Loma Prieta earthquake introduced to the engineering community a genuine anomaly when it was found that a building could be not a "good" seismic performance structure although it was not heavily damaged. Although a building may not comply with the modern code without significant damage, there is no criteria to mandate the retrofitting of the building. An engineer might be tempted to over-tag a building that the engineer feels "has no bracing," because the engineer knows the building is hazardous even though there was little damage. Those dealing with this matter should read the above referenced papers and have a workable plan of action and methodology in place prior to an earthquake.
Once the tagging program is in place, the criteria for the removal of the tag are as essential as the understanding of the consistent meaning of the tag to enforcement officials and the general public.
Building departments were generally ill-prepared to deal with the post-earthquake repair or retrofit issues. Although some agencies within the state had in place ordinances dealing with these issues, this was not common, especially for the smaller communities in the Santa Cruz and Watsonville areas. The engineering community and the various state commissions had these issues on the list of work, but there was no consensus document available for each of these issues. As a result, emergency ordinances were hastily put in place to serve as a rational level of order. A prime example of this was the city of Oakland, which put in place emergency Ordinance 11217. This ordinance set forth the criteria to be used to determine whether a building was sufficiently damaged in the earthquake to warrant retrofitting as opposed to simple repair of the damage. If the damage
was greater than a specified percentage, the building was required to be repaired in a manner to bring the building in compliance with the 1988 Uniform Building Code. If it was not damaged greater than that specified level, simple repairs were all that were required. The intent of the ordinance was to catch buildings that had major structural damage, and there was an intent to offer leniency for hardship as well as for structures falling reasonably close to the value of damage. The value set by the city was a loss of lateral capacity greater than 10 percent of its pre-earthquake capacity. The ordinance was supplemented by a procedures document (City of Oakland Memorandum, 1990), which set out the procedure for making the evaluation of both pre- and post-earthquake capacity.
Interestingly enough, the city has been witness to some of the most intriguing engineering manipulations by competent engineers for finding ways to make beneficial use of that procedure. However, like the manipulation of the building-period calculation in the more creative seismic design efforts, the procedure requires that one use the same logic before the earthquake as after the earthquake, thus resulting in consistent solutions.
Description Of Need
Although Oakland's response is only part of the major puzzle, clearly some tools were found to be needed by all agencies. Some are technical issues while others are political issues involving the determination of acceptable risk for communities. Generally, they fall into the following categories:
emergency standards and criteria for the demolition of a building for safety reasons;
criteria for the determination of whether a building was damaged to a degree that it should be retrofitted rather than repaired;
standards for the level of repair if repair only is mandated;
standards for the repair for buildings to be retrofitted; and
determination of the level of acceptable risk for the community involved.
In the immediate hours following a damaging earthquake, designated officials may be faced with a safety hazard from a building that is badly damaged and could collapse in aftershocks. Often, the older structures most likely to be in this situation may also be historic buildings. The options include clearing the hazardous area, shoring the building (with the possible danger to workers), or demolition. Most of this is political in nature; however, the potential danger for injury is part of the purview of the engineering community and is the subject of virtually no study or dissemination of knowledge.
The development of a consensus document for the evaluation of the level of damage that rationally should trigger the upgrading of a building is needed. The city of Oakland's program might be a good starting point. Because of the thought that return periods of significant events vary throughout the country, the proce-
dure for evaluating the percentage of damage could be universally adopted while leaving the triggering percentage of damage to the assessment of risk to the responsible public agency.
Given the triggering plateau or the simple need on the part of the owner to retrofit, the standards of retrofit need to be developed. Although standards are a first step, the techniques needed to implement the standards are also needed if the standards are to be effective. An example of this is unreinforced masonry. The standards adopted in the Uniform Building Code are now supported by commentary developed by the Structural Engineers' Association of California. The commentary assists the designer in determining what principles of detailing and analysis are anticipated in the code document. Still to be determined is the real effectiveness of these provisions. This will be learned only in an earthquake that shakes some of these retrofitted buildings. As in the past, observations will assist in modifying the document to get better or more-consistent performance. Holmes's study on unreinforced masonry structures in San Francisco is an excellent reference on the background of a city's issues on the subject (Holmes, 1990).
The broader field of other potentially hazardous buildings is not as far along as that of unreinforced masonry (URM) structures. This effort is underway by the Applied Technology Council (ATC, in press) and the state of California (California SSC, 1991c, 1992). The understanding of performance limitations of less than code-complying buildings is a difficult issue, since it involves the anomaly of the thought of having less than fully ductile elements or connections within a structure that might be subject to significant ground motions. A further issue is that nonstructural elements attached to those same structures may have the less than code level of detailing. A need to develop the tools for performance-oriented evaluations is in fact the demand. Unfortunately, the response to the issue seems to be an attempt to develop code standards as opposed to developing the analytical tools and the material properties that will give predictable performance values.
The development of the URM model code for bearing-wall buildings and the subsequent commentary was mandated by the need of building officials to have some standard to set forth as the minimum acceptable level of compliance. The user of the provisions is obligated to inform the building owner of the performance expectations implied by the URM code.
Evaluation of Risk: A Challenge
The Seismic Safety Commission's acceptable risk could clearly be used as the scaling factor for the political bodies in determining what levels of risk are acceptable in the future, again taking into account return periods. Remembering
that the risk policy is based on building performance, including functionality, and that the engineering community is responsible for adequately predicting performance, it would appear that a major challenge in front of the research and engineering communities is the development of a methodology to predict levels of damage and extending that to imply levels of post-earthquake functionality. Of interest is the work of Freeman et al. (1993). This work is moving the design community past the strict compliance with code routine into real performance.
Once the engineering and research communities can develop the basis for the solutions to these issues, it will be necessary to determine the appropriate form in which to present this information to the building official so that the official can perform the fiduciary responsibility of protecting public safety.
Plan Checking and Site Inspections By Building Officials
Once the research has been developed and the industry brought up to speed, we are left with the dilemma of the building official. The official must enforce an ordinance that by its nature is inexact. Ideally, if all building designs were fully competent and flawlessly detailed the official would have no real problems. Follow that with detailed inspection by highly qualified inspectors fully familiar with seismic design as well as construction materials. That would make the official's job easier and might even diminish the need for anything as detailed as current seismic codes. This, however, is not the reality of the modern project, nor of the staffing capabilities of the average building department.
The building official, without even considering the Loma Prieta earthquake, faces a difficult problem with both the plan checking and site inspections of buildings. Certainly, budget plays a role in this. However, the public should demand that fees be charged that are adequate to provide proper levels of plan checking and site inspection. The Office of the State Architect (OSA) in California has a reputation for high performance for both plan checking and site inspection. Despite some complaints, the observed seismic performance record of structures reviewed and inspected by this agency is outstanding. Even so, it can be improved. In a legislative hearing about the Loma Prieta earthquake, surprise was expressed by a state senator when testimony indicated that plan check by any local building official was not the equivalent of that of the Structural Safety Division of the OSA.
Quality-Control Check Points
In all designs, there needs to be a series of check points to assure good quality control. The check points for design are the internal quality control of the engineer, the competence of the plan check, and the thoroughness and competence of the site inspection. Peer review is appropriate, especially when the design is complex.
Good construction implies that the design concepts go on the drawings so that a contractor can interpret them and then build in accordance with the drawings. The Loma Prieta earthquake uncovered situations where shear walls were left out of buildings and critical shear transfer elements were simply never installed. In addition, buildings were found to not comply with the seismic codes under which they were designed. So one can immediately blame the contractor or the designer, and yes, that is appropriate. But what about the last line of defense—the plan check for which each owner must pay a fee? For a moment, look at the situation for the building departments and the track records of success. In the Loma Prieta earthquake, post-Field Act school buildings had an outstanding record of performance. Anyone who has gone through the OSA process knows this is not an accident. Plan check with OSA is like having molars removed. One feels better about two weeks after the molar extraction. There is no doubt that OSA has gotten the job done. Yet in the California legislature, there have been recent efforts to eliminate the Field Act and OSA.
Some, but certainly not all, building departments in the state of California have plan check and inspection programs as well organized and effective as those of OSA. Given the present structuring of financial responsibility, it may be impossible, without new mandates, to improve laggard building departments. The lesson is that building departments, their inspectors, and plan checkers all have varying levels of capabilities. Given what will be called limitations, it is also appropriate to opine that there is no change from site to site in the need to get it right. As a result, there is a need to set some bottom lines for performance and find a way to help those agencies that need support when faced with quality-control issues involving plan checking and site inspection.
Some Action Is Possible
The California Building Officials, the International Conference of Building Officials (ICBO), and SEAC should already be working together to determine how to implement a statewide program for quality control of plan checking and site inspection for seismic issues. The result would be of mutual benefit to each organization and to the public.
Strong Motion Instrumentation
The Strong Motion Instrumentation Program was the recipient of a major payoff in the Loma Prieta earthquake. The investment in instrumentation yielded numerous strong-motion records. The raw data developed from these recordings are available for study from the California Division of Mines and Geology (1990), and sets of recorded data have been utilized in the analysis of the performance of several buildings (McClure, 1991). Given this valuable information, it is interesting to note that virtually every engineer responding was concerned with the effective use of these data.
Knowing that the data were needed to advance the understanding of earthquakes and earthquake effects on buildings was enough to get the original Strong Motion Instrumentation Program well established. There is an equally tough challenge ahead to optimize the use of the data and research in a manner that will be most beneficial to the mitigation of the earthquake hazard. What might be the true benefit of research utilizing these data?
Purpose And Use of the Data
Clearly, if the data could be used to confirm or even postulate ground-motion characteristics in other regions, such characterizations could be used to assist in the overall evaluation of seismic risk. It would seem that an essential question must be addressed before the true value of the data will affect building design. That question is, what are the attributes that make a difference in structural performance, and are we now fully knowledgeable in the understanding of those attributes?
There seems to be a tendency to try to extract from the data a higher degree of accuracy than currently assumed for predicting the ground motion at sites affected by the ground shaking. The basic points of the effects of soft soils are understood in principal and addressed in code provisions. Small degrees of improvement in the accuracy of ground motion will not have a meaningful impact on the design process at its present state of knowledge except for on the most exotic designs. If the focus of the research is to identify conditions that are currently not recognized as critical issues in design, then the research should receive high priority. To the engineer who recognizes that this 7.0 event with 10 seconds or so of strong shaking was an anomaly, it will be interesting to see how the findings from this event will be extrapolated to other events not only in California but in other areas of the United States. What appears to be of major significance is the continued assessment of the relative, and to some degree, actual, risk of an event. Such information is vital to the total assessment of seismic risk to a community and would be useful in establishing priorities for funding as well as for mitigation programs.
Regarding perspective on the interface between seismic design and the seismological experts, there does not seem to be any research being generated to identify the attributes mentioned above. As an example, do researchers understand what is more important in the successful design of a structure—the amplitude of the motion or the shape of the response spectra? Similarly, do earthquakes have predictable levels of energy input to various sites, and if so how would this be usable in the prediction of building performance? This bonanza of data should yield significant advancements in the mitigation of the earthquake
hazard. Some thought regarding the use of the results could yield a planned program with tangible benefit.
Strong-motion records are being reviewed and analyzed to extract lessons on building response. Studies that correlate the data are useful for certain levels of ground motion, however, most studies are based on an elastic model and have difficulty accounting for inelastic or non-linear response. Again, the level and duration of the ground motion may make the lessons in response somewhat limited in many cases. Although there was a proposal at one time to fund a program for the utilization of the ground-motion data, the program has not yielded results. ATC 35 may fill that gap. There does not seem to be any systematic or organized program to extract the most important building-response data and apply them to useful lessons for seismic design.
Some work is appearing that uses the Strong Motion Instrumentation Program records to extract correlation, or lack thereof, from the current provisions with measured and observed performance. The work of Werner et al. (1992) is reported to offer some insight into what may be an anomaly relating to drift.
In general, buildings performed pretty well, in accordance with the designs used for the original construction. Of course, the media focused on the spectacular news to the point that many viewers on the East Coast thought that San Francisco was completely engulfed in fire and that Santa Cruz had been leveled. Yes, there was damage and death, and that cannot be ignored. However, there is cause to reflect on what did in fact perform well and focus on what was accomplished in California through the volunteer efforts of countless structural engineers when they, without funding, developed the first seismic codes. This same group, the SEAC, has continued that effort year to year basically without funding and has continued to perfect a workable and effective code. It is not perfect, but while reflecting on the Loma Prieta earthquake and what did not suffer damage, it is appropriate to understand that the level of seismic safety was achieved not by accident but by the unselfish efforts of the engineering community with the help of active researchers. Without that effort, the damage and loss of life would have been many times that experienced.
Still, there is a nagging view that despite the success, the earthquake did not truly test the designs, so one cannot extrapolate that we can relax now that we have experienced this event. Additionally, it is evident that although there were few surprises in performance, there was damage, and that implies that we have sufficient current knowledge to predict where the damage will occur in future earthquakes. If that is the case, we are faced with the traditional problem of having the will and the means to mitigate the hazard.
The Loma Prieta earthquake allows a number of observations about where the lessons of importance to future seismic mitigation may be.
Owners And The Public Expectations
In viewing the Loma Prieta earthquake where there were few real surprises, the clear lesson was that the owners and public were surprised, or at least given new insight, into what the seismic profession views as performance expectations. The public and owners had pre- earthquake expectations that unreinforced masonry buildings were hazardous, but they had the perception that earthquake-resistant design protected them from damage that might interrupt their businesses. There was also the perception that URM buildings would perform well if they had some limited level of retrofitting. The lesson learned was simple. The public is now aware that it takes more than minimum code compliance to have one's building operational after a significant event.
The first steps from the design community in response to this issue have been to address the need to be better at predicting damage and, concomitantly, the levels of response. Traditionally, the design focus had been life safety, which really just brought designs over the threshold to preventing collapse and tying a building together well. As noted above, that step saved countless lives, but a new demand has emerged and that is one of functionality. The Loma Prieta earthquake may have triggered that awareness, but it may well be misleading when judging damage potential from a magnitude 7.0 earthquake. Results would have been significantly different if the duration of shaking had lasted twice as long or if the quake had been centered in the northern section of the Hayward fault. Clearly, there is a mandate to develop better tools to predict performance as it relates to control of damage and functional requirements desired by the owner. Certainly, engineers designing structures in the post-Loma Prieta earthquake environment have a new obligation to inform the owners of projects as to the limitations of the minimum requirements of the code with regard to damage potential.
Performance, Design And The Code
In the Loma Prieta earthquake, well-designed, well-constructed buildings all performed very well. Evidence indicates that this performance is not an anomaly. Good design, good construction, and good quality-control do result in high levels of performance despite the inadequacies some may claim exist in our knowledge of earthquakes.
Breaking this down, it appears that for the Loma Prieta earthquake the requirements of the Uniform Building Code yield results consistent with its intent. With this in mind, the Loma Prieta earthquake was not necessarily a true test for all buildings.
As a result, there is a higher awareness of the true intent of the code on the part of the owners, and it is projected that in the future owners will ask designers to focus not on the minimum requirements of the code but on the need for post-
earthquake functionality. If nothing else, this is music to the ears of the original seismic code writers who have been saying for years ''remember these are minimums not maximums.''
Safety Versus Post-Earthquake Usage
Tradition has basically led to the focus on the collapsed buildings. The Loma Prieta earthquake taught the public that they can be affected dramatically by the simple closure of a building, which takes away their home, place of worship, or place of work. Given this perspective, functionality has become a new issue. Since the codes are life-safety only, mere conformance to code no longer meets the demands of the public.
Focus of Research
A review of the types of research triggered by the Loma Prieta earthquake is interesting. The obvious attraction was the Bay Bridge and freeway structures. But this paper is on buildings. Again the focus was on the damaged structures that were generally unreinforced masonry or soft-story marina structures. The attention on masonry was obvious and direct, but it was known prior to the earthquake that these structures would not perform well. Being able to predict the poor performance should therefore direct the research attention to understanding how to repair or retrofit the buildings. Most attention has been focused on the URMs, however, it needs to be noted that wood-frame structures in the Marina District performed poorly. Some blamed this on liquefying soils. Whether it was the liquefaction or the soft story, or a combination of the two, little attention seems to have been paid to soft-story wood-frame structures, which were not damaged when subjected to similar ground motions.
Another perspective would be if one looks at the inventory of buildings suffering the most losses in the earthquake. In this case, it would clearly be wood-frame residential structures. This leads to the observation that to mitigate the maximum impact of the quake based on number of structures affected, one would focus on masonry and wood-frame structures, including residential structures.
The residential problem was identified as a series of definable issues. The legislature mandated that the Seismic Safety Commission develop the Homeowner's Guidebook to Earthquake Safety (California SSC, 1992), which is basically a disclosure document that must be handed to the buyer of a pre-1960
house prior to sale. The document does a good job of identifying, in lay person's terms, the various hazards often found in homes. It is disappointing that the mitigation of those identifiable hazards is not mandatory prior to the sale.
The hazards to be mitigated for residential structures are relatively straightforward. The consequences of not mitigating the simple hazards were pointed out in post-earthquake testimony. In testimony before the Seismic Safety Commission, a fire marshal was commenting on the need to brace water heaters and the potential for disaster if this is not accomplished. In the Marina District, gas fires were posing a serious threat. The fire marshal pointed out that it was lucky the gas problem was handled quickly before a fire storm situation developed and spread out of control. This was a dramatic testimony emphasizing the need for a simple mitigation measure.
Given that over 27,000 residential structures were damaged during the Loma Prieta earthquake, one would extract the logic that residential structures and the mitigation of the observed earthquake residential hazard would be high on the investment priority list of funding agencies. Two percent of the National Science Foundation (NSF) funding went to residential research, although the previously mentioned top award was on the topic of earthquake insurance in California. Only the California legislature took action in passing legislation to mandate the preparation and distribution of the homeowner's guide. It would seem obvious that the residential hazard is within reach of direct mitigation, and with some financial stimulus from governmental agencies and insurance interests, it is possible that this hazard could be wiped out.
The Profile of The Damaged Building
This brings the discussion to where the focus of research on buildings is, especially in the post-Loma Prieta earthquake time frame. The hazards are clear, and the majority of the hazards are identifiable as unreinforced masonry, non-ductile concrete, and residential construction anywhere in seismic zone 4. It should be noted that although the effects of soft soils were predictably more noticeable on soft soils for low-rise construction, there is no mandate based on observation to change the design approach to such buildings. Additionally, the profile of damaged buildings does not indicate that the problems were focused on new, tall, or exotic buildings even though this may not have been a significant event to test the true performance. It is also true that the majority of buildings constructed are not of that same category. The majority of the building stock in fact is low rise, and it was that stock that took a major hit during the Loma Prieta earthquake. If practicing professionals know where the problem lies, and generally agree that California's code provisions reasonably address the design of the majority of the problem buildings, then should not funding be likewise focused?
The Majority Of Structures
Having said that, it is postulated that code-conforming buildings in the class being considered performed very well. If that is the case, then for the ordinary buildings without complexity, the proper application of the code may in fact provide a rational level of functionality. This leads to the conclusion that the functionality problems may result from buildings of such complexity or irregularity that they cannot be designed by code in the first place. Given that the code was intended for life safety only but for the majority of structures (the simple buildings) has yielded good functionality performance results, it may be inferred that all but the complex or irregular buildings are probably going to perform well even when conceding post-earthquake functionality issues.
For this majority building population the Loma Prieta earthquake did confirm that several factors are general predictors of performance. They are
age of structure designed without codes;
construction deviations; and
improper application of the basic design concepts of earthquake resistant design.
Old materials put the issue of performance into the ballpark of how and if the materials can be saved while an effective system of lateral resistance is installed. The historical and social aspects of this debate are the subject of other papers, however, the engineering approach has been crystallized as a result of the Loma Prieta earthquake. As noted, there is a new awareness of functionality, not only on the part of owners but also in the insurance and banking industries.
To do the job right and bring a building up to expectations of performance can prove to be uneconomical. A response to this has been to allow some slack in strict code compliance and find a way, using all the materials, to make a building safe. Pre-Loma Prieta earthquake this was accepted, however, it is no longer professionally acceptable to fail to inform the owner of the realistic performance expectations of a design. Given this reality, owners are finding it difficult to retrofit at a functionality level, and without some financial incentives, the stock of truly historic buildings will be gone in the next major earthquake.
Clearly, making a retrofit safe only for the occupants to evacuate now has a limited appeal, especially to bankers and insurers. Of the people killed by unreinforced masonry in the Loma Prieta event, all were killed outside of the building that caused the problem. No occupants were fatalities. This suggests that safety considerations go beyond the envelope of the building.
As expected, during the Loma Prieta earthquake, unreinforced masonry did not perform well. No research is needed to understand the basics of this. Faced with partially damaged structures, the methods of repair and retrofit become an issue. Two major areas of concern are apparent. First, what are the performance characteristics of the masonry, and what are the characteristics when the masonry is put into a framework? Second, when approaching retrofits, what will the performance be of the retrofit buildings, especially when the buildings have been stiffened and some of the high damping characteristics that, to a limited degree, have helped the performance have been removed?
The work by the SEAC Hazardous Buildings Committee in developing standards and commentary for unreinforced bearing-wall buildings has led to a broad introduction of the concept that the older materials, as frail as some are, do have some limited capacity for energy absorption. The performance of buildings utilizing such materials is not likely to meet the functional performance standards of a fully informed owner, however, as suggested by the Seismic Safety Commission policy on risk, improved safety may be the only economically viable alternative in some cases.
As noted elsewhere in this paper, observations of damage conclude that the damage to unreinforced masonry was as expected. Equally interesting and often unreported are the undamaged unreinforced structures. The supposedly successful performance of these buildings can often not be attributed to anything more than circumstances beyond the control of the structure, such as interaction with adjacent buildings or damping of a soft story.
The retrofitting of such structures, whether it be to repair damage, to meet safety requirements, or to meet performance expectations is not a simple code prescriptive matter. The understanding of the dynamics of seismic response and influence of inelastic action is essential to obtain a retrofitted structure with the desired performance characteristics. Equally as important is the attention to interstory drift at both the elastic and inelastic levels. Correlation with the recorded Loma Prieta earthquake results of actual building measurements should help with insight into this subject.
A more difficult issue is facing the structures that have little or no engineered design and do not comply with the letter of the code. The code, or even our performance projections, are based on how engineers would like things to be, including idealized projections of material properties and building configuration. Unfortunately, existing pre-design buildings do not comply with those criteria, and engineers are forced to evaluate structures that have energy absorption characteristics that do not necessarily fit the mold. To address this, some new analytical concepts are being utilized that could offer significant insight into performance. These concepts focus not on capacity alone but on deformation and damping in the inelastic range.
No potentially hazardous category of building has had more attention focused on it then less than fully ductile concrete structures. Ban them? Use lower R values? Or as an alternative, perhaps understand their performance better? What risk is appropriate when fully considering the seismic risk of the locality? Studies by Gergeley have been aimed at projecting performance and specifically at revising the R values contained in code documents. Strictly from a performance point of view, it is essential that the capacity be understood and that the damage levels associated with these demands be developed.
The broader concern for non-ductile concrete, both existing and potential low-seismic-region design, has great broad impact. Clearly the Loma Prieta earthquake pointed to concern over those types of buildings. Building frames have been tested. These results, compared with the findings of the Loma Prieta earthquake ground-motion data, should give a better understanding of the effectiveness, or lack thereof, of non-ductile frames. Extending the ground-motion data into other parts of the United States will prove useful.
Research And Codes
It still remains that most of the applied research does not address the majority of the buildings but rather the exceptional buildings. However, only the knowledgeable should be the exceptional buildings, and the day to day fundamentals seem to already be well addressed by the basic codes. If that is true, then a lesson can be extended outside: the fundamental understanding of earthquake design is as important as any code. Codes have been getting more detailed and are covering issues that are not only difficult to understand but must be carefully applied in only the correct set of conditions.
Engineers experienced in seismic design agree that adherence to basic principles such as the providing of complete load paths and the rational transfer of forces in connections is essential to good seismic design. The code should support those basic principles by providing material parameters to allow the accurate evaluation of the strength of the materials used. The limits on the use of those parameters are equally important. The user of the code should then be able to read the code, determine the material strengths as well as the limitations of use of the parameters. When the limitations become more complex, the reader may lose site of the basics and become absorbed in the complexity of the code provision.
Shear stress in concrete is an example. Historically, shear in concrete was a simple calculation that added the shear carried by the concrete to the shear carried by the reinforcing steel to give the total shear capacity of the section being
analyzed. This was a simple straightforward computation and gave a basic sense of how loads were transferred in concrete. The modern code provisions for shear stress in concrete offer much more complex methods for evaluating shear capacity. Without study or research, the reader of the code may not fully understand the basics on which the provision is founded.
When viewing the shear wall damage, it is apparent that basic application of shear wall principles would have been a more effective mitigation measure than the application of more-complex code provisions. Stated another way: would the application of more-complex concrete shear wall provisions have improved the performance of buildings in the Loma Prieta earthquake? Would the nuances of the more complex provisions really have a significant effect on the performance when compared with the other performance parameters?
In short, codes may be getting too complex, may be trying to cover too much, and may be trying to cover all situations in a prescriptive manner. In principle, the code should address the largest inventory of buildings with basic provisions that cover the simplest of structures. As structures become more exotic, a higher level of design expertise is required. These more exotic structures should not be the focus of the codes but should require more sophisticated analysis, something the codes will never be able provide in a prescriptive manner. The review of such structures should be accomplished by independent peer review and not dumped on the building official.
Soils And Building Performance
A second lesson is clear. Loma Prieta was not the earthquake that typified the San Andreas fault, and it gave us a little wrinkle. It was a Richter magnitude 7.1 with a relatively short duration. However, Bruce Bolt has developed a model that correctly identifies the pockets of damage experienced during Loma Prieta earthquake (Lomax and Bolt, 1992). It remains to be seen if such a model applied with sources located at other sites along the San Andreas or other faults will lead to any insight into the need to make revisions to the current static-force procedure or whether it will assist in providing more-accurate site-specific information for those using more sophisticated analyses. Equally important would be the effectiveness of such a model for fault conditions outside of the California scenario.
While attention is being focused on site-specific data as it is traditionally defined, there are perhaps other issues in need of study. The earthquake gave us some newer terminology such as "focused energy" and "near field effects." In today's design approaches, there is little opportunity to address either of those notions. In fact, it is clear that stepping from such concepts to meaningful design parameters is a major issue when considering the effectiveness of research in the mitigation of hazards. While the parameters of sites are described in terms of peak accelerations and return periods, a designer must focus attention on the
spectral shapes as much as the ordinate of the spectrum. The complete understanding of earthquake design must include the understanding of inelastic response and how the inelastic softening of a building affects performance and the control of damage. The future may hold yet another parameter; the total site energy related to a given event. There is need for the environmental side of the risk equation to equate with the building-performance side so that the data from the Loma Prieta earthquake is effectively utilized for building performance in a manner that will advance our state of knowledge and not just provide interesting information.
It would be unfortunate to have the data of this limited event extended to rationalize modifications in the basic approach to static design, which in fact demonstrated good results in the Loma Prieta earthquake.
For any location within the United States, there are data that describe the environmental earthquake risk as best we know it today. The national effort to improve that data base was enhanced by the Loma Prieta earthquake. It is imperative that the National Earthquake Hazards Reduction Program (NEHRP) take a lead role in coordinating that effort such that there is a continuing improvement of that data and its application to seismic mitigation. The resultant knowledge must then be transferred not only to the design profession but also to the decision makers.
Such decision makers should then be looking to the design professional to define the building-performance risk associated with a given site. That performance must be reasonably predictable for the normal buildings, and may well require special studies for the complex or irregular buildings. It is opined that code compliance will not and never was intended to deliver that level of information. The Loma Prieta earthquake has alerted the owners, and they should now be demanding rational predictions of building performance based on post-earthquake functionality.
This rationale places a new burden on the research and engineering communities. However, in so doing, the earthquake has brought new awareness and, therefore, new responsibility to the developers or owners of buildings. This concept is that of acceptable risk, and it is a universal issue wherever earthquakes could affect communities. Some education of the public on this subject is mandated.
Fundamentals, Not More Code Provisions
When the consensus that there were no surprises was voiced, some may have wondered how much was known or maybe whether the code was right after
all. Don't count on it. The message was that well tied together, rationally designed buildings worked fine, perhaps despite the code. Basic fundamentals of seismic design seem to prevail where good performance is found. Where such basics are missing, poor performance is evident. Clearly, older buildings constructed prior to the understanding of seismic performance generally lacked the basic elements of good seismic design.
The lesson in this case is the same lesson as learned before in other earthquakes. The simplest of codes could say: tie the building together, provide complete load paths, and design it to meet requirements. But codes have become more complex, and perhaps in the effort to cover all types of buildings, they have obscured the fundamental of the majority of the constructed buildings.
The lesson is that it is necessary to keep the basics simple and understandable. Professionals (doing the design of the majority of the structures) need education in the simple principals of seismic design. Perhaps research efforts, along with the code bodies, ought to consider solving the majority problems on simple and basic terms and leave the more complex issues to be addressed by those proposing to design such structures.
Peer Review For The Complex Projects
Considering the more complex structures, if the above approach were to be applied, perhaps most everyone would be happy except the building officials. If the more complex structures use concepts not specifically stated in the code, how are officials to enforce compliance? Them were numerous structures damaged in the Loma Prieta earthquake, which for legal reasons will go unnamed, and in which there were design or construction errors that technically should have been found within the purview of the building-department program. This is not a criticism of the competence of the individuals but a statement of a problem with priorities that do not allow agencies to fund the necessary staff to accomplish a fundamental step in the construction process.
If the system of enforcing compliance for the majority of even the basic structures is imperfect, then how can building departments be expected to review the more complex structures? This paper will probably not stop the continued advancement of the complexity of the code, but it does suggest that the solution to the more complex structural work is independent peer review. Such review would open the door for the knowledgeable designer to use proven state-of-the-art knowledge in the analysis of complex buildings.
The lesson learned is that the more complex structures that do not fit into the category of simple regular buildings will never completely fit into code provisions, and at worst the code will be terribly messed up if the basics are obscured while attempts are made to cover all the exceptions. A better solution is to require independent peer review for the exceptional buildings. There needs to be a balance between prescriptive codes and performance-code requirements.
While attention is focused on complex designs and soil-structure interaction, another key link in the performance reliability is overlooked. This link is quality control and, in particular, inspection. There were numerous reports of damage directly attributable to structural elements being either improperly installed or missing altogether. The range of these defects extends from residential structures to multi-story facilities. One may question how the inspector could have missed that. Clearly it is time to do better. There needs to be a national mandate to get a qualified inspector on site, properly paid, and with enough time to confirm that all the seismic elements are properly installed.
That simple owner-paid investment seems warranted if the owner expects to be eligible for relief funds should the building ever be damaged.
It is no surprise that URM structures performed poorly. It is, however, instructive to note that there were no deaths to occupants of unreinforced structures. All the deaths associated with URM poor performance occurred outside of the structure, either on the sidewalk, street, or in an adjacent building. Whether or not this is an anomaly or not is not the issue. A URM building is a danger to more than just the occupants, it is a danger to the adjacent activities.
The Seismic Safety Commission vigorously discussed the value of posting hazardous buildings and, in particular, URM structures. The findings published in a Report on Posting of a hazardous building might have some impact on the persons wishing to enter the building, but how would the hazard be reduced for those adjacent to the URM structure?
The lesson is simple and well known; unreinforced masonry buildings are potentially hazardous. The ICBO Code for Building Conservation proposes standards for evaluating such buildings, but clearly the intent of that ordinance is for life safety only.
Need For Interaction Between Research And Practice
Returning to the relationship between research and design, there seems to be a tendency to demand that researchers immediately translate research into new design rules or code provisions. On the other side, researchers tend to be somewhat narrowly focused on the issue in front of them with less involvement on the design meaning on a broad scope. It would be wise for neither to become impatient with the other. The cycle is clear—research advances knowledge and must be seasoned with the testing of the design community in order to draw the applicable lessons for use in practice. This test should also be used in reverse to assist the research community in looking at research that will support the needed
advancements in practice. The Earthquake Engineering Research Institute, SEAC, and the Seismic Safety Commission all have research committees that are trying to develop a national research strategy to deliver the most bang for the buck in terms of usable product.
Direction And Funding Of Research
Given the parameters of effective earthquake mitigation, it seems appropriate to focus research in areas where it will be most effective in improving seismic hazard mitigation. The inventory of damage to buildings should provide direction on prioritizing research. If it is true that the majority of the building stock is not high rise or complex structures, then this stock of buildings should be a large focus of the research effort. Equally clear then would be the suggestion that funding for this majority would be in the best interest of the funding agencies working to mitigate the hazard. For those complex and special structures making up a minority of the stock, it is time to consider a mechanism by which those specialty projects carry their own burden of research needs. The Japanese model of research funding coming through construction of buildings to create the demand for research should be considered.
Ground Motion Data
The focus on the instrumental data from the Loma Prieta earthquake is credible, but the research effort needs to be directed toward results that will benefit the knowledge of the designers of buildings throughout the United States. To this end, the prediction of the spectral characteristics of sites throughout the country would be a direct and needed goal. The understanding of the energy input and how it varies throughout the country is needed as the profession starts to look at energy concepts for seismic design. However, of immediate use would be the correlation of predicted building response to recorded building response. This correlation needs to address the traditional use of elastic models with a magic word, "damping," thrown in to resolve the differences between actual and theoretical performance.
Attention to the unusual effects of site soils in moderate earthquakes seems warranted and has broad implications for areas of lower seismicity.
The Loma Prieta earthquake certainly should be viewed as a wake-up call mandating seismic-mitigation activity throughout the state of California. Unfortunately, action wanes as time separates the reality of the event. Outside of California, there may be less of a sense of urgency, less of a sense that there is a risk. Each community in the country should at least take the step to understand
what their real risk is and deal with the judgments necessary to determine what level of risk is acceptable. If there are non-believers, then their attention should be directed to the 27,000 damaged residences resulting from the Loma Prieta event and the deaths resulting from URM falling onto adjacent sites. Implementation of a mitigation program responsive to the decision on the acceptable level of risk starts with the understanding of the simple fundamentals and not with the adoption of prescriptive codes.
ATC. In press. ATC-33. Preparation of Guidelines for Seismic Rehabilitation of Buildings, ATC, Redwood City, California.
California Division of Mines and Geology. 1990. CSMIP Strong-Motion Records from the Santa Cruz Mountains, California Earthquake of October 17, 1989. Report OSMS 89-06. As presented in the Proceedings from the SEAC 59th Annual Convention.
California SSC. 1992. The Homeowner's Guide to Earthquake Safety. Sacramento, California, SSC 92-02, October.
California SSC. 1991a. Loma Prieta's Call to Action. Report on the Loma Prieta Earthquake of 1989, Sacramento, California.
California SSC. 1991b. Policy on Acceptable Levels of Earthquake Risk in State Buildings. Sacramento, California, SSC 91-1, January.
California SSC. 1991c. Breaking the Pattern, A Research and Development Plan to Improve Seismic Retrofit Practices for Government Buildings. SSC 91-05.
California SSC. 1992. Phase II Research and Development Action Plan, Proposition 122, Sacramento, California, October 7.
CALTRANS. 1991. Proceedings of the First Annual Seismic Research Workshop. California Dept. of Transportation, Sacramento, California.
City of Oakland Memorandum. 1990. Earthquake Damaged Buildings—10% Criteria. March 13 (revised April 4).
Fratessa, P. F., and F. Turner. 1991. The New Policy on Acceptable Levels of Earthquake Risk in State Buildings. In Proceedings of the 60th Annual Convention, Structural Engineers Association of California, Sacramento, California.
Freeman, S., J. A. Mahaney, T. F. Paret, and B. E. Kehoe. 1993. The Capacity Spectrum Method for Evaluating Structural Response During the Loma Prieta Earthquake. Preprinted for the 1993 National Earthquake conference, Memphis, Tennessee, May 9.
Holmes, W. T. 1990. San Francisco Unreinforced Masonry Buildings Study. In Proceedings of the 59th Annual Convention, Structural Engineers Association of California, Sacramento California.
Kasai, K., and B. F. Maison. 1990. Structural Pounding Damage Due to Loma Prieta Earthquake. In Proceedings of the 59th Annual Convention, Structural Engineers Association of California, Sacramento, California.
Lomax, A., and B.A. Bolt. 1992. Broadband Waveform Modeling of Anomalous Strong Group Motion in the Loma Prieta Earthquake Using Three-Dimensional Geological Structures. Geophysical Research Letters. Vol. 19, pp. 1963-1966, October 2.
McClure, F.E. 1991. Analysis of a Two-Story Oakland Office Building during the Loma Prieta Earthquake. SMIP91: Seminar on Seismological and Engineering Implications of Recent Strong-Motion Data. California Division of Mines and Geology, 13-1-13-11.
Roberts, J.E. 1990a. Putting the Pieces Together: The Loma Prieta Earthquake One Year Later. In Proceedings, Bay Area Regional Earthquake Preparedness Project, October 15-18, San Francisco. California. Bridge seismic research sponsored by CALTRANS.
Roberts, J.E. 1990b. Recent advances in seismic design and retrofit of bridges. In Fourth U.S. National Conference on Earthquake Engineering Proceedings . EERI, Palm Springs, California.
SEAC. 1991. Reflections On the Loma Prieta Earthquake October 17, 1989. Ad Hoc Earthquake Reconnaissance Committee, Chapter 9, Response of the Disaster Emergency Services Committees.
Webster's Intermediate Dictionary. 1977. Merriam-Webster Inc. Publishers, Springfield, Massachusetts.
Werner, S.D., A. Nisar, and J.L. Beck. 1992. Assessment of UBC Design Provisions Using Recorded Building Motions from Morgan Hill, Mount Lewis, and Loma Prieta Earthquakes. Dames and Moore, Oakland, California, April.
The following individuals and written material have been used in formulating this paper. The conclusions drawn from these references do not necessarily reflect the views of the individuals or authors listed.
Responding Professionals: Robert Burkett, Pat Campbell, Lloyd Cluff, Edward Diekmann. Neville Donovan, Robert Dyson, Eric Elsesser, Robert Englekirk, Sig Freeman, S.K. Ghosh, Mel Green, James Hill, Ronald Hamburger, William Holmes, George Housner, Wilfred Iwan, John Kariotis, James Libby, Charles Lindberg, Jack Meehan, William Menta, Jack Moehle, Rawn Nelson, Joe Nicoletti, Chris Rojahn, Mete Sozen, Don Strand, Tom Tobin, Ajit Virdee, Ted Zsutty.
Ad Hoc Earthquake Reconnaissance Committee. 1991. Reflections On the Loma Prieta Earthquake October 17, 1989. Structural Engineers Association of California, Fair Oaks, California.
California SSC. 1991. Loma Prieta's Call to Action, Report on the Loma Prieta Earthquake of 1989. Sacramento, California.
California SSC. 1990. Report to Governor George Deukmejian in response to Executive Order D-86-90, November 30, 1990. Sacramento, California. Report SSC 90-06.
National Clearinghouse for Loma Prieta Earthquake Information:
Catalog April 1991
Catalog November 1991
Catalog April 1992
Catalog November 1992
National Information Service for Earthquake Engineering/Loma Prieta Project. Earthquake Engineering Research Center, 1302 South 46th Street, Richmond, California 94804-4698.
Proceedings, 59th Annual Convention Structural Engineers Association of California. 1990. Structural Engineers Association of California, P.O. Box 399, Fair Oaks, California 95628.
Proceedings, 60th Annual Convention Structural Engineers Association of California. 1991. Structural Engineers Association of California, P.O. Box 399, Fair Oaks, California 95628.
DISCUSSANTS' COMMENTS: BUILDINGS SESSION
James Beavers, Martin Marietta Energy Systems
The real lesson I've learned from the Loma Prieta earthquake, as an observer primarily, is that we need to know more about the performance of buildings during earthquakes, especially in moderate seismic zones. We've been focused on retrofit policy issues—how to understand performance, how to evaluate these facilities. As Paul Fratessa mentioned, only 2 percent of Loma Prieta research dollars was for unreinforced masonry. Yet workshops I have attended have cited the critical need to understand better the performance of unreinforced masonry, as well as non-ductile concrete.
We have been doing a major program on understanding the performance of hollow clay tile, a common wall type in Tennessee. Some have said hollow clay tile walls don't have ductility, but tests showed differently—significant capacity out of plane. One of the key lessons we've learned through our tests is to understand completely the performance from a structural engineering standpoint. How are these components going to perform during an earthquake? These are critical questions that must be addressed in areas such as east Tennessee.
We all know that equipment in buildings when properly restrained or anchored has exhibited satisfactory earthquake performance. Where failures do occur, they can be attributed to inadequate or missing anchorage. Good performance of equipment subjected to past earthquakes is not surprising considering that equipment failures observed during shake table testing have been rare. We believe that it seems quite apparent that rather than qualification testing of equipment to prove that it can withstand earthquakes—emphasis should really be placed on good fabrication, good assemblage practice, quality assurance, and inspection. Thank you.
Stephen Mahin, University Of California, Berkeley
The Loma Prieta earthquake provided the first major opportunity in two decades for the earthquake engineering profession in the United States to assess its design procedures. While the region of intense ground motion was relatively remote, significant motions did develop in the vicinity of many major engineered structures. Especially important was the presence of numerous strong-motion instruments installed in buildings and at free-field sites.
The earthquake reemphasized many lessons from past earthquakes. These lessons include:
the importance of soil conditions to the intensity of motion at a site and to structural damage;
the significant damage to older (and some newer) residential construction, and the personal and social consequences of this damage;
the need to retrofit structures, such as unreinforced masonry buildings, non-ductile reinforced concrete, and so on (many retrofit structures suffered significant structural damage under moderate shaking; studies to determine appropriate retrofit procedures are needed);
the need to develop effective methods to assess the seismic resistance and structural integrity of damaged structures and to perform emergency as well as long-term repairs;
life safety includes consideration of falling hazards as well as structural collapse;
the need to develop methods for characterizing the importance of pounding of adjacent buildings and to mitigate the consequences of this pounding; and
the increasing importance of functionality and repairability following a major event.
All of these items require additional research. In addition, the earthquake did not provide compelling evidence regarding the behavior of structures on soils that liquefy; the performance of buildings, nonstructural components, and contents when subjected to design-basis shaking; the performance of new construction types; and the efficacy of current retrofit and repair techniques.
More detailed quantitative investigations of the Loma Prieta earthquake have been aided by the numerous records obtained during the event. However, these records remain under utilized due to limitations of funds to investigate their structural implications. While detailed engineering studies have been carried out on most damaged structures, much of this data has not entered the public domain due to legal considerations and the unavailability of a suitable forum. Similarly, public access to structures and information regarding structures, particularly modern engineered structures, has been poor. Similarly, statistical summaries of this information in terms of failure/damage rates for different types of structures, down time, repair and disruption costs, and so on have not been systematically gathered. This information would be invaluable in assessing the performance of structures in other earthquakes in the Bay Area and elsewhere. While short-term reconnaissance style studies are carried out, few financial and personnel resources are available for sustained and directed engineering studies of these types.
The Loma Prieta earthquake also raises many questions. For example, what structural attributes allow some older buildings to remain undamaged, while nearby, newer structures are damaged; why nearby buildings of similar construction have dissimilar performance; how to design structures in a region where the intensity of motion may only be locally severe (microzonation for near field, soil amplification and ground hazard effects); and how to design structures to remain functional or reparable following major earthquakes. These and other questions call for concerted, focused efforts by practicing engineers, building-code officials, owners, policy makers, and researchers.
Gerald Jones, Kansas City Building Department
Perhaps the most difficult concept to explain to the average citizen is that of acceptable risk. Society readily accepts the deaths of 300 to 500 people each weekend on our highways in vehicle accidents, but has little tolerance for any loss of life in building-related incidents. The same citizens are the first to complain when new code requirements raise the cost of their proposed new buildings or the cost of maintaining an effective building department. One of the greatest challenges is defining ''how safe is safe.'' When society reaches agreement on "how safe we should be," it is necessary to carefully describe the expectation level the code and regulations can deliver. Minimum legal levels must be differentiated from desired levels.
A building code is based on three primary issues—health, safety, and general welfare. Few people would disagree with the desire to cast codes in more performance-oriented formats. The issue becomes one of measurability—how does one know when the desired level of performance has been achieved without certain prescriptive yardsticks? This issue also raises the questions of competency for the designer, the builder, and the building official and of how the owner (the ultimate source of funding) can make the economics pay for the increased costs. The role of the building department is to act as the conscience of the community in the building process. I cannot agree more with Tom Tobin that there is a major difference between having a code and enforcing a code.
All in all, this symposium is an excellent opportunity to learn from each other. Thank you.
William Holmes, Rutherford & Chekene
We have heard many comments how, since the Loma Prieta earthquake, engineers have had to deal with policy makers or make policy themselves, much more so than ever before—such as how to define performance objectives and acceptable risk.
We need to settle some of these policy issues on where to put our mitigation resources. First, we need to look more closely at the assumed (non-engineering) deficiencies for residencies, and ask what the motive is for fixing these—life safety or economy. Water heaters should differentiate between gas and electric for life safety; the benefit in cost ratio is high. Chimneys are not life-threatening, and it costs as much to remove them as to repair the damage. The fixing of cripple studs and foundation bolt problems has a very high benefit in cost ratio. The more serious structural problems, such as side-hill sites and soft stories are life-safety hazard, but it is necessary to establish a standard to estimate a cost in benefit ratio and to mandate repair.
Second, we must look at how to decide when a damaged building can be re-occupied. Engineers have to identify the motive for repairing deficiencies in
damaged buildings. The policy choices are (1) to repair buildings to a level where the re-occupancy risk is no greater than it was before the earthquake; (2) if the earthquake revealed an obvious hazard, to mitigate it before occupancy was allowed (repair building elements); (3) to require retrofitting for buildings shown to be vulnerable by the earthquake above some trigger level; and (4) to look at damage patterns, deciding which class of buildings are hazardous and passing ordinances requiring all buildings in that class to be fixed, even if not damaged.
Problems for engineers include the methods and effectiveness of repairing elements to their original conditions. What can be done? How can the trigger be set for repair versus retrofit requirements? It is difficult to identify the extent to which the building has really been damaged.
Lastly, the Loma Prieta earthquake brought to light a whole set of adjacency hazards. These include pounding where two buildings collide, vertical dynamic irregularities when a high building is adjacent to short one, local instabilities caused by unaligned floors of adjacent buildings, hazards caused by a shared wall, and components of one building falling on the other. As we have no control over adjacent properties, the best we can do is notify the owner that there is a potential problem and perhaps design the strengthening of the building with the adjacent hazard in mind.
To solve these problems, engineers must get the policy makers and lawyers involved. Thank you.