Improved Seismic Monitoring - Improved Decision-Making: Assessing the Value of Reduced Uncertainty (2006)

The proposals for improved seismic monitoring put forward by the U.S. Geological Survey (USGS, 1999) are based on the premise—described in the previous chapter—that the provision of enhanced information will improve decision-making by reducing the uncertainties associated with risk assessment, with risk perception, and with choice and risk management. The relative gains provided to society by improved monitoring information can be measured by the economic value of reduced decision uncertainty, assessed by comparing actions to be taken to manage the risks with and without improved monitoring. Benefit-cost analysis (BCA) is one tool that can be used to evaluate alternative risk management programs. This chapter presents the conceptual basis for the specific categories of benefits—as input to a BCA—that are discussed in greater detail in later chapters.

On several occasions over the last few decades, the USGS has been asked to undertake benefit-cost analyses in support of program initiatives to generate improved earth science information. Typically, these requests have originated from the Office of Management and Budget (OMB). The ensuing studies were focused on estimating the economic value of information that is derived from earth science initiatives, using a decision-making framework that is supported by the theoretical principles of BCA and accepted estimation techniques. Similarly, this report also focuses on the issue of valuing seismic monitoring information within a benefit-cost framework. In this chapter, the following five questions are addressed:

1. What is the appropriate conceptual framework for valuing the prospective benefits of seismic information?

2. What is the nature of the information that is produced?

3. What are the various categories of benefits?

4. How can the benefits be measured?

5. Finally, has the necessary information been collected for a complete benefit-cost assessment of a seismic network?

An estimation of the societal benefits derived from seismic monitoring information has a number of components:

1. The framework for valuing benefits is based on the premise that enhanced information from a seismic monitoring network will lead to reduced uncertainty regarding the earthquake risk.

2. The scientific information gained from seismic monitoring will generate a variety of derivative benefits. As a seismic monitoring network captures ground motion for multiple events over time and space, gains in knowledge will assist with better decision-making in the future. Such gains can reflect increased accuracy as well as reductions in uncertainty.

3. The information produced by a seismic network is a “pure public good”—the information is available to all, and its use by one party does not detract from its use by others.

4. The major categories of benefits can be distinguished temporally as “immediate, near- and long-term, and cumulative gains in knowledge.” In general, benefits may be derived from improved contributions to risk assessment, risk perception, and individual choice. An alternative perspective is that the benefits would include, but not be limited to, improved emergency response, enhancements in performance-based engineering, and increased potential for forecasting and predicting earthquakes.

5. A wide variety of hazard-prediction models and methods can be improved with enhanced seismic information, including ground motion and loss models.

The baseline for determining the economic benefits provided by improved seismic monitoring is the present situation in which the nation’s seismic monitoring capabilities are distributed among a patchwork of essentially independent regional networks (described in Chapter 1), but with the important realization that existing funding levels are insufficient even to maintain present capabilities. Accordingly, any description of the economic benefits of improved seismic monitoring has to consider the incremental benefits in terms of

• the existing baseline situation, where existing funding levels will result in a gradual deterioration of the nation’s monitoring capabilities;

• increased funding leading to maintenance of the existing networks and their integration into a revitalized U.S. National Seismic Network (USNSN);

• higher funding levels that permit the deployment of the Advanced National Seismic System (ANSS), as proposed in USGS (1999); and

• the ideal situation, in which the deployment of the ANSS is extended to provide instrumentation wherever it is needed.

Seismic monitoring data provide potentially derivable benefits that may be observed over multiple time frames. First, there are the immediate benefits after an earthquake (e.g., an informative ShakeMap can be used by emergency responders). This increment in more accurate information assists officials to deploy limited resources more rapidly and strategically to areas that have been identified as experiencing the greatest shaking (see Chapter 7). As a result, communities will experience more rapid—and consequently less expensive—restoration of services, resulting in reduced business interruption and cost savings.

Second, there are near- and long-term benefits from seismic monitoring information, related to the interval of time that allows society to react to the information in a strategic manner beyond the immediacy of an emergency response. The incremental benefits in this time frame principally reflect additional loss avoidance activities, beginning with property damage and running the course of all loss categories. Such loss avoidance would result either from information gained from a single event or from the accumulation of monitoring information over time. This accumulated knowledge can potentially result in an improved approach to the design and construction of infrastructure, the implementation of appropriate mitigation of existing structures, and/or the revision of building polices and regulations.

The third category of benefits is the accretion of knowledge. The accumulation of information from improved seismic monitoring potentially leads to a more complete understanding of the spatial and temporal physical processes associated with faulting and other sources of seismic activity. The accumulated record of weak and strong motion information could ultimately lead to some type of earthquake prediction capability (described in more detail in Chapter 4).

The limited time available for the committee to receive input, deliberate, and draw conclusions precluded the completion of a fully comprehensive BCA. Nevertheless, wherever possible, quantifiable benefits have been identified, evaluated, tallied, and compared to estimated project costs. Subsequent chapters demonstrate that there are numerous other economic

benefits that will require a substantial compilation and analysis effort, and to assist this process the general principles of BCA that should be applied to improved seismic monitoring are outlined in the following sections.

The economic efficiency of any project, including those owned and/or operated by the public sector, can be estimated by the application of benefit-cost analysis (BCA). Efficiency, as measured by a BCA, represents one dimension of any project’s desirability as a component of an overall project evaluation. Efficiency is important in that it reflects the notion of resource scarcity. That is, are available resources being used in the most beneficial combinations from society’s point of view, regardless of to whom the benefits or costs accrue? An elaborated BCA moves beyond a single efficiency criterion to address who pays and who benefits, thereby providing important information on the distribution of benefits and costs. While efficiency criteria are couched in dollar terms, when distributional aspects are added, the distributional measures are not necessarily comparable across projects. Thus, benefit-cost ratios (or differences) that describe alternative projects are scalars numerated in dollars that can be ranked, whereas distributional data are vectors of varying types of information and are not easily ranked within a project analysis or when comparing project evaluations.¹

BCA is an elaborate accounting that observes basic economic principles. In this case, three basic ledger principles must be observed:

1. The ledger of costs and benefits must be specified with care, so that a comprehensive and exhaustive itemization is achieved.

2. Double-counting must be avoided.

3. Entries must not be mislabeled (e.g., the relatively common political suggestion that project benefits include the jobs created by a project when labor, in fact, is a cost).

To give just one (of many possible) examples of double-counting in BCA, consider the situation in which a municipality reduces its emergency

services for post-earthquake disaster assistance (e.g., by reducing the number of ambulances) and, as a result of this action, lowers its taxes. Any suggestion that both the reduced costs and the lowered taxes should be claimed as savings would constitute double-counting—both measure the same benefit and only one should be included.

All ledger items must be valued in dollar terms using market prices. Most benefits and costs are expected to occur in specific future years; at best, a distribution of each value for each future year can be made. The various possible realizations must each be weighted by their associated probabilities of occurrence. The resulting prospective benefits and estimated costs cannot be added for different years until properly discounted to a present value. Because costs change incrementally, it is important that plausible increments of any project be identified and that benefit-cost ratios that describe these increments are developed and reported. Where market prices do not exist because the commodity is not traded (e.g., clean air, seismic information), alternative methods—nonmarket valuation—must be used to estimate prices. The two broad categories of nonmarket valuation are “revealed preference” and “stated preference.”² Of the revealed preference approaches, the “hedonic price method”³ is widely used. It is recognized, for example, that residential property values respond to (capitalize) seismic information. By carefully controlling for other determinants of residential property values, it is possible to derive the dollar value of this information. The field of “contingent valuation” best represents the stated preference approach—contingent valuation studies use hypothetical markets to determine the willingness to pay for changes in risk levels. Because valuation is dependent on science information, any detailed contingent valuation analysis undertaken in the future must use the best available science information as the basis for understanding changes in risk.

As noted above, the basic ground motion and structure motion information that will be provided by improved seismic monitoring is a public good—that is, a product or service that can be shared by many users

simultaneously without detracting from its value to any one of them.⁴ The nature of the good (or product) that emanates from improved seismic monitoring is essential for understanding why a public good framework is appropriate. Previous assessments of the societal value of geologic information—a similar type of informational good to the improved seismic monitoring goods—are applicable to improved seismic monitoring information; for example, Bernknopf et al. (1993) make the case that the information represented by geologic maps is in fact a public good (see Appendix A). The discussion contained in that report also applies more broadly to the information generated by improved seismic monitoring.

The important implication is that in identifying the various categories of incremental benefits in evaluating the economic benefits to be derived from improved seismic monitoring, summing the various benefit categories is appropriate because they are not mutually exclusive. As such, the appropriate evaluation of the economic benefits of improved seismic monitoring is its overall benefit based on the sum of its verifiable uses, rather than the benefits of any single independent use.

As noted in the preceding section, the theoretical basis for evaluating the benefits of geophysical information in general—and seismic information in particular—is efficiency gain. Improved seismic information provides the basis for better societal understanding and decision-making by reducing uncertainty. A number of studies have demonstrated the benefits and costs of improved seismic information in general (e.g., Bernknopf et al., 1990, 1997; Mileti et al., 1992; Olson and Olson, 2001). Although these studies have not specifically addressed improved seismic monitoring, they have included the benefits and costs of mitigation through building codes, microzonation information programs as enhancements to housing markets, and earthquake predictions.

The benefits that potentially are available as a consequence of improved seismic monitoring are quite varied. The fundamental role of seismic information is to reduce uncertainty over time and to increase the accuracy of emergency preparedness activities, loss avoidance regulations, and/or earthquake prediction. As increasing amounts of information are collected by improved seismic monitoring, the different vintages of seismic information (e.g., series of earthquake events—foreshocks, mainshocks, and aftershocks) will lead to a more complete understanding of geophysical processes, more realistic models, and better-informed risk assessments. As the improved seismic monitoring information evolves, it will be possible to generate improved ShakeMaps nationwide, to design

better safety and regulatory programs, and to improve our earthquake prediction capabilities. The brief discussions of the four broad areas listed below are presented to set the stage for more detailed descriptions in later chapters.

Illustration 1—Improvements in Forecasting and Prediction. Additional seismic information will eventually improve the ability to micro-zone urban areas. Several types of incremental benefits can be envisioned. If the risk can be sufficiently differentiated geographically, there will be improvements in the efficiency of the operation of insurance markets. The benefit will be a reduction in the uncertainty of risk assessments and improvements in the accuracy of the information. Benefits will also accrue because of better forecasting and the potential for prediction. These types of benefit are discussed in more detail in Chapter 4.

Illustration 2—Improvements in Loss Estimation Models. An essential element of loss estimation models is their ability to adequately represent the frequency and severity of the earthquake hazard, as the basis for improving the efficiency of engineering design and providing better cost estimates for loss reduction. Because models of building vulnerability are based on limited performance data, improvements in the seismic hazard and building performance input data will reduce uncertainty in the estimation procedure. This type of benefit is discussed in more detail in Chapter 5.

Illustration 3—Improvements in Performance-Based Engineering. As a seismic network collects more information, new analysis will lead to better construction design criteria and lower construction costs. Because maps of the spatial distribution of seismic hazard can be improved, design criteria and structural damage mitigation will be better matched to local risk. This reduction in the uncertainty of how buildings will behave in an earthquake can be anticipated to reduce property damage and other losses. These types of benefits are discussed in more detail in Chapter 6.

Illustration 4—Improvements in Emergency Response Capabilities. If all bridges contained seismic monitoring capability, a determination could be made immediately following an earthquake as to which bridges were most likely to be damaged (based on fragility functions) and should be closed pending inspection and which were likely to be undamaged and immediately usable. Further, strong motion data recorded on bridges, together with damage data from these bridges, could be used over time to improve the accuracy of the fragility functions that relate ground motion to expected damage. This is an example of reduced uncertainty. The net benefit stems from the incremental savings in resources due to the fact that only damaged bridges, and not all bridges, would have to be inspected. Further, the transportation system could remain open or be opened sooner after a hazardous event, resulting in less business interruption. This type of benefit is discussed in more detail in Chapter 7.

The reductions in uncertainty that will translate into reductions in various types of earthquake damage—from the implementation of more appropriate improvements in the timing, location, and design of mitigation—can be categorized according to the timing of their benefits.

The USGS (1999) carefully documented the immediate or time-critical qualitative benefits that would result from a revitalized USNSN-ANSS by demonstrating that real-time seismic monitoring systems offer the opportunity for society to take steps to reduce damage or loss of lives in advance of an earthquake. Even a few seconds of advanced warning before an earthquake may save lives by signaling people to “duck, cover, and hold” or to drive more slowly and avoid bridges or overpasses. Some machinery is more vulnerable to earthquakes while in operation, so real-time warnings could prevent damage by enabling the equipment to be turned off. Failure in one portion of a network can cascade into geographically widespread problems, as was demonstrated by the 2003 electricity blackout in the eastern United States. Chemicals or molten metals can harden, thereby spoiling the batch, damaging equipment if interrupted midprocess, and delaying the resumption of business. The avoidance of release of hazardous toxic materials from any medium is also a benefit. Table 3.1 presents examples of immediate benefits (the emergency response benefits are described in more detail in Chapter 7).

As information accumulates from seismic networks, ground motion prediction models and loss estimation models will improve. The incremental benefits that result from these better models will lead to better engineering design parameters and more cost effective design, more detailed seismic zonation and land-use regulation, and so forth. As a result, there will be potential reductions in direct and indirect property damages compared to the situation without this improved information. There will also be reductions in direct and indirect business interruptions, reduced environmental damage, reduced iconic losses (both built and natural), and reduced human impacts. Furthermore, there is the potential to reduce infrastructure damage through improved seismic and structural “health” monitoring programs.

The first two columns of Table 3.2 provide a list of categories (adapted from Rose, 2004) and examples of incremental benefits from all potential uses of improved seismic information, although with an emphasis on near- and long-term applications. Direct property damage to buildings, contents, and infrastructure will normally be caused by ground shaking, although some property damage can result from subsequent deterioration caused by exposure to the elements. Indirect property damage is best exemplified by the impacts of fires, often resulting when ground shaking ruptures conveyances of flammable materials and is often further exacerbated by disruption of water delivery systems. Since property damage also typically sets in motion the other loss categories, mitigating property losses is fundamental to the near- and long-term incremental benefits of seismic monitoring.

One category of losses that can benefit from improved seismic monitoring is business interruption—diminished output of economic goods and services over some period of time from commercial enterprises caused by the earthquake. Business interruption can emanate from damage to physical capital but also from a cessation of other activity flows.⁵ For example, a factory may be unscathed by an earthquake, but be forced to shut down if its electricity supply is cut off or its employees are unable to report to work due to earthquake-induced damage to transportation networks.

The time dimension also means that business interruption losses are highly dependent on private and public sector decisions and actions regarding recovery.

Additional losses stem from “multiplier” or “general equilibrium” impacts on chains of upstream suppliers and downstream customers of damaged businesses and those cut off from their utility lifelines or access by their employees or customers. These indirect effects can, in the case of large, highly interdependent, self-sufficient regional economies, be even larger than the direct flow losses (e.g., Webb et al., 2000).

In addition to more quickly identifying areas in which valves in pipelines carrying toxic materials might be shut off, there are several near- and long-term environmental benefits of improved seismic monitoring. Retrofitting wastewater treatment facilities and other sensitive structures will lead to decreased risk of drinking water contamination. Warnings that reduce the risk of fire will help avoid a deterioration of air quality. Performance-based engineering is also applicable to infrastructure other than buildings—the benefit of preventing physical damage to these systems is obvious and readily measured. However, interruption of the services that some of them provide is not so readily measured, since infrastructure services are typically not sold in the marketplace. For example, with the exception of toll roads, although the value of continued highway and bridge access is unpriced, this does not mean that it has no value (e.g., Gordon et al., 1998).

For many earthquake impacts that result from damage to structures and loss of businesses, there is a corresponding social impact that has economic implications (Heinz Center, 2000). Seismic monitoring information should result in benefits by reducing the impact on individuals, families, and communities as a consequence of reduced death and injury from improved building codes, improved dispatch of emergency services, and so on (Peacock et al., 1997; Enarson and Morrow, 1998). Broader social impacts that can be reduced include emotional stress and population dislocation, both of which have disproportionate impacts on marginal populations.

Prevention of the various types of losses already discussed can reduce government administrative costs, such as those associated with processing Small Business Administration loan applications and supervising emergency relief and recovery. Even if the government activity is a transfer, such as a loan, the administrative effort represents a real use of resources and is therefore a cost that improved seismic monitoring information can reduce.

The prominence of these various incremental benefit categories differs in terms of the uses of an advanced seismic monitoring system. The last

four columns of Table 3.2 provide some “ballpark” estimates of the potential scope of four major applications of seismic monitoring—earthquake warning, emergency response, performance-based engineering, and earthquake forecasting. For example, its use for warning will reduce death and injury. However, warning will not reduce damage to structures, since they are immobile, but it can reduce damage to contents by facilitating emergency shutdown procedures. Monitoring systems can also reduce business interruption losses by giving electric utilities the opportunity to take precautions to avoid cascading outages. Moreover, loss reduction strategies can be fine-tuned to produce the mix of benefits considered to be in the best interest of society.

Increasing the number of earthquake strong motion recordings will contribute significantly not only to loss avoidance as discussed above, but also to improving society’s capabilities—through advances in underlying knowledge—in performance-based engineering, seismic zonation programs, and earthquake prediction. The accumulation of knowledge is typically viewed as an ongoing and long-term process, and it is this accumulation of knowledge that makes improved loss avoidance and emergency response and recovery gains possible.

Two examples illustrate situations in which a more complete set of ground motion records will enhance knowledge benefits. First, current USGS probabilistic ground motion predictions near major active faults are adjusted before they are used in the National Earthquake Hazards Reduction Program (NEHRP) design ground motion maps, so that they depict ground motion levels that are more compatible with observed damage in earthquakes. The ground motion models used in USGS maps are based on the few recordings at close distances to large earthquakes and are thus highly uncertain predictions. Effectively, the lack of a more substantial set of near-fault strong motion recordings makes it difficult to reliably implement performance-based engineering, since the levels of damage to structures observed in the near-fault setting seem to contradict the high ground motion estimates currently used. Second, seismic zonation will be improved with more complete ground motion records. With an improved understanding of the underlying geological structure, narrow regions of higher vulnerability can be identified within broader regions that are now considered to have equal exposure to the hazard. Improved monitoring will provide improved definitions of seismic “hot spots” for design and land-use plans, thereby contributing to more rational (and less risky) urban development.

An evaluation of benefits resulting from the proposed improvements to the nation’s seismic monitoring networks must be based on the prospective benefits from the additional information that will be forthcoming. Estimating the value of information has always been a challenge because it is not a tangible commodity and because its benefits are often very subtle. This is compounded by both hazard-specific and more general limitations of various estimation methods. For example, in retrospective studies, it is difficult to isolate the contribution of seismic monitoring from other factors that influence the reduction in earthquake losses. In theoretical or simulation analyses, it is relatively more difficult to verify the projections of benefits. Although such limitations are not uncommon in many areas of public policy, we can still glean some insights from the few rigorous analyses, from the many studies that yield ballpark estimates, and from theoretical work. Public policy decisions generally have to be made despite such limitations.

Table 3.3 lists methods typically used to estimate the various categories of benefits resulting from reducing earthquake losses. The table summarizes the major benefits from seismic monitoring information and lists the ways in which these methods can be used in the future to more rigorously estimate seismic monitoring benefits.

In applying any of the methods discussed above, it is important to keep in mind several basic principles of benefit-cost analysis. The major ones include:

• Losses must be evaluated in terms of real resource costs and prices that reflect their competitive value. This excludes transfer payments, such as taxes, and may require adjustment in existing prices for various other distortions (e.g., monopoly pricing).

• Benefits are not limited to those activities with markets but should also include nonmarket effects such as externalities (e.g., pollution) or reduction in public goods (e.g., transportation services).

• Future benefits must be discounted to adjust for the “time value of money” (except perhaps in the case of the value of a human life).

• Flow measures of benefits, such as business interruption losses, should be evaluated over the time period during which individual businesses and the economy as a whole have not returned to the projected normal level of economic activity.

• Benefits should not be estimated in a context in which decision-makers are assumed to react passively or in a “business-as-usual” mode to an earthquake; rather, benefits should reflect inherent and adaptive resilience at the individual, market, and community levels (e.g., a signifi-

cant proportion of lost production resulting from electricity outages can be “recaptured” by subsequent overtime work in most sectors).

In the following chapters, the incremental benefits of improved seismic monitoring are discussed in detail. This discussion yields predominantly qualitative benefits, although a lower-bound estimate of the benefits that result from performance-based engineering enhancements is presented in Chapter 6. Why are there not more examples of quantitative benefits, especially given the wide range of potential benefits discussed above? The answer is the lack of certain types of information. Although the committee has argued that a wide array of incremental benefits results from an enhanced seismic network, this argument is forward looking. That is, only over time will these benefits be realized. Just as improved seismic monitoring will provide more critical and basic information to earth scientists and engineers, critical information has to be gathered from future earthquake events and provided to scientists in the behavioral sciences (economics, finance, sociology, political science, psychology, etc.) to fully assess the benefits.