INTRODUCTION
In disaster planning, much attention is paid to the role of buildings – how will they perform in a major earthquake? How long will they take to repair? Will people be able to stay in their homes after a quake, or will they need temporary shelter? Less attention is paid to the role of the infrastructure systems that support urban life, which we call our "lifelines." By "lifeline," we mean the utility systems that bring us our water, electricity and natural gas and the transportation systems that allow us to get around, including public transit, ports and airports, and road infrastructure. As with buildings, lifelines are critical to our ability to recover from an earthquake. If our buildings are not "serviceable," nobody can live or work in them. San Francisco's capabilities for response to, and recovery from, an earthquake are highly dependent on the condition of lifelines in the wake of such a disaster.

The importance of reducing the risk to lifelines cannot be understated. Imagine what would happen if even one of our lifelines seriously failed in an earthquake. How would people be able to shelter in place without drinking water? What happens if our natural gas lines cease to work – or worse – stoke the flames of a massive fire? How will emergency workers get to our city if the bridges fail? How will our economy recover if we can't move people or goods around the region?

Lifeline owners in both the public and private sectors have made significant investments in designing, constructing, and retrofitting their facilities and systems to reduce the risk of damage in an earthquake and to facilitate restoration of services to their customers. However, the seismic performance standards for lifelines vary widely and are not tied to generally applicable public policies for reducing risk or for ensuring community resilience in the face of a major earthquake.

We need to know how our lifelines are going to perform in an earthquake. And we need to set performance targets based on "resilience" – i.e. our capacity to recover quickly and effectively from a major earthquake. To promote the city's recovery from an earthquake, the services provided by lifelines should be restored as quickly as possible – within hours or days. However, as things now stand it may take months or even years for some systems to be restored to full operation due to the uncertainties associated with potential damage, the lack of clearly articulated goals for restoration, and the lack of consistent standards for achieving those goals.

The San Francisco Planning and Urban Research Association (SPUR) has developed this policy paper raise awareness of the City's vulnerability and to encourage steps necessary to increase the City's resilience to a major earthquake, with respect to lifelines.

This paper is a component of SPUR's Seismic Hazard Mitigation Initiative. The initiative's goals are to:

• define the concept of "resilience" in the context of disaster planning;
• establish performance goals for the "expected" earthquake that supports the definition of resilience;
• define transparent performance measures that help the City reach these performance goals; and
• suggest next steps for implementing policies that improve the City's resilience.

IMPORTANCE OF LIFELINE SYSTEMS TO A RESILIENT SAN FRANCISCO
There are a number of reasons why the ability of lifelines to survive or recover from an earthquake is of the highest importance:

• Because lifeline systems spread across the region, they have weak links resulting from site specific conditions that increase the risk of partial or complete system shutdowns. For example, the City's water system includes specific elements that may withstand earthquake damage. However, the City's water is delivered through the Hetch Hetchy system, which transports water from the Sierra Nevada through 280 miles of pipeline and 60 miles of tunnels that cross several earthquake faults, including the Hayward Fault. Further, the water distribution system within the City includes areas subject to differential settlement that could damage mains and service lines.
• The impact of damage to lifelines is compounded by the interdependency among lifeline sectors, both within San Francisco and in the region as a whole. For example, electric power is required to run water and wastewater systems, and the airport depends on jet fuel brought in via pipeline from petroleum refining and distribution facilities.
• In addition to supporting first responders, the expedient restoration of lifelines reduces the need for evacuation and sheltering of victims who would otherwise be without critical services.
• The recovery of communities and the economic base is dependent on the re-establishment of lifelines. If the services provided by these systems cannot be restored expeditiously, the recovery will be delayed as residents and businesses struggle with the lack of critical services and the inability to move people and goods around the region.

DESCRIPTION OF LIFELINES
"Lifelines" are defined as those essential utility and transportation systems that serve communities across all jurisdictions and locales. These systems share the attributes of being distributed systems, rather than isolated facilities; and of providing products or services that are transferred through networks that often cross legal and jurisdictional boundaries (ALA, 2005).

LIFELINE SYSTEMS AND COMPONENTS
Lifelines include:

Water systems	Highways and roads
Wastewater systems	Ports and waterways
Electric power systems	Transit systems
Oil and natural gas systems	Railroads
Telecommunications systems	Airports

Hospitals and other medical facilities are often considered to be lifelines; however, these facilities are addressed in the Existing and New Buildings papers.

In general, a lifeline system incorporates a wide range of elements necessary for system operation, including linear components; mechanical, electrical, and electronic equipment; buildings containing system components; operating centers; and other supporting elements. The circumstances under which individual elements may fail vary widely, as do applicable design guidelines and standards. The performance of the entire system is as critical as the performance of individual elements; however, damage to individual elements may be sufficient to shut down part or all of the system.

Lifeline systems are also distinguished by their interdependency. The continued operation of a lifeline system, such as the communications network, may be dependent on the operation of another system, such as the power system. Similarly, the ability for system owners to restore their respective systems following an earthquake may be dependent on the condition of highways and other transportation elements.

LIFELINES IN SAN FRANCISCO

As a major city, San Francisco is served by all major categories of lifelines, as described in Table 1 (right). (The table includes lifelines that serve the City itself; it does not include lifelines serving the Bay Area region as a whole.)

While some of these systems (such as MUNI and the Port) are wholly contained within San Francisco, most are part of larger regional networks, such as the regional freeway system; and the failure of these larger systems would paralyze San Francisco as well as the region. Additionally, San Francisco is critically dependent on lifelines within the region, but outside of the City. For example, finished petroleum products are either delivered by pipeline from outside the region or manufactured in the region from crude oil imported to the region; and damage to these systems would result in a shortage of gasoline, diesel, and aviation fuel. Interdependencies among systems will increase their respective vulnerabilities to shutdowns, regardless of level of damage or location.

POTENTIAL RISKS TO LIFELINES
The potential risk to lifelines varies widely, depending on the sector, the relative age of the system, the location of individual system elements and site-specific conditions at those locations, and the extent to which the system in question is vulnerable due to damage to lifelines in other sectors. Site-specific conditions that increase vulnerability to lifelines in the Bay Area include the following.

• Ground shaking: Ground shaking due to the earthquake may damage system elements or associated features (such as buildings in which lifeline elements are located).
• Liquefaction: Many of the Bay Area's lifelines, such as ports and airports, are located in areas on the perimeter of the San Francisco and San Pablo bays that are susceptible to lifeline-damaging conditions related to liquefaction, such as lateral spreading and differential settlement.
• Displacement along faults: Key lifelines, such as the Hetch Hetchy system and major transportation elements, cross the Hayward Fault and other faults in the Bay Area.
• Landslides: Lifeline components may be displaced or closed due to earthquake-triggered landslides.

The 1989 Loma Prieta Earthquake provides the most recent example of widespread impacts of an earthquake to lifeline systems in the Bay Area. Examples of impacts to lifelines are described in Table 2 (above right).

ADDITIONAL HAZARDS ASSOCIATED WITH NATURAL GAS SYSTEMS
As described in Section 1 above, the failure of lifeline systems can have widespread and enduring effects on services critical to the response, public health and safety, re-establishment of the community, and economic recovery. Just as significant are the risks of additional damage and injury resulting from lifeline failures – in particular, the potential for fires caused by damage to natural gas systems. Gas leaking from a damaged line may ignite or explode, either because the gas supply is not shut off or is restored despite the existence of leaks that have not been detected and repaired. Approximately 35 percent of fires in San Francisco following the 1989 Loma Prieta earthquake were gas-related; and 55 percent of fires in the Los Angeles area following the 1994 Northridge Earthquake were gas-related (CSSC, 2002). Gas supplies may continue to flow to users following an earthquake, unless the provider shuts down the system or individual customers shut off the supply at the interfaces between the supply system and buildings (such as closing a shutoff valve at a residence).

Currently, San Francisco does not require automatic shutoff valves for gas lines as they enter buildings; and in many cases in San Francisco the gas pressure step-down is located adjacent to, or within the footprint of, the building. Consequently, the potential exists for gas leaks within damaged buildings, often under high pressure, increasing the risk of ignition.

THE RISK OF MUCH STRONGER EARTHQUAKES
With a magnitude of 7.1 and a fault rupture in a remote location 60 miles from San Francisco, the Loma Prieta Earthquake had far less impact on San Francisco than large earthquake events that are anticipated to occur closer to the City. As described in the Overarching Policy Paper, a reoccurrence of the magnitude 7.9 1906 San Andreas Fault Earthquake could result in regional losses of $150 billion, or 10 times the total losses from the Loma Prieta Earthquake. Although many lifeline owners have made substantial investments in seismic retrofits and upgrades since Loma Prieta, it is likely that damage to lifelines, and the and corresponding disruption to immediate response and long term reconstruction would be far greater – even to the point of inhibiting San Francisco's ability to ever fully recover. The possibility of such an earthquake in the near term heightens the urgency with which the City should address the vulnerability of lifelines.

SEISMIC PERFORMANCE OF LIFELINES
Guidelines, standards, and code requirements for the seismic performance of lifelines vary widely. The range of functions and designs of these systems, as well as the range of potentially damaging hazards, necessitates sector- and hazard-specific approaches to reducing damage, ensuring safety, and facilitating system restoration. Consequently, development of these standards occurs among numerous code development entities, other professional organizations and private sector entities, and Federal, state, and local government agencies. These entities have made great strides in developing standards to reduce risk to lifeline systems in all sectors.

Most sectors have progressed to system-based approaches in order to assess risk and reduce disruptions the performance of systems and delivery of services to customers. Nevertheless, achieving a consistent level of resilience is complicated by the many different regulating bodies to which system operators must answer. The general tendency toward sector- and hazard-specific development of standards results in the following problems:

• A lack of commonly understood definitions for acceptable seismic performance.
• Different standards for performance among different sectors.
• A lack of inter-sector coordination for the development of standards, setting of priorities, and implementation of mitigation.
• Limited understanding by political leadership and the general public of the potential performance of lifelines during an earthquake – and whether the performance of lifelines will meet expectations.

The sector-specific natural hazards provisions are generally based on varying levels of risk (for example, in terms of the design earthquake or probability of occurrence). Additionally, most sectors do not have standards for reliability – that is, practices that have been developed to ensure system restoration in accordance with goals set by stakeholders. According to the ALA, such standards have been developed only for highways/roads, ports, and railroads (ALA, 2004).

RESILIENCE IN THE CONTEXT OF LIFELINES
As described in the Overarching Policy Paper, SPUR defines San Francisco's "seismic resilience" as the City's ability to:

• Contain the effects of earthquakes when they occur.
• Carry out recovery activities in ways that minimize social disruption.
• Rebuild following earthquakes in ways that mitigate the effects of future earthquakes.

EXPECTED PERFORMANCE OF LIFELINES FOR RESILIENCE
SPUR recommends the establishment of clear, readily understood performance goals that define resiliency in infrastructure. Goals for the restoration of service are expressed in terms of the percentage of customers that have service after an earthquake.
These goals assume the occurrence of the "expected" earthquake, defined as an earthquake that can reasonably be expected to occur once during the useful life of a system. For San Francisco's buildings, this earthquake is defined as having a 10 percent chance of occurrence in 50 years. As described in the Overarching Paper, "The Resilient City: Defining what San Francisco needs from its seismic mitigation policies", a magnitude 7.2 earthquake on the peninsula segment of the San Andreas Fault would produce this level of shaking in most of the City. Since lifeline systems generally serve cities and regions for well over 100 years, a larger "expected" earthquake should be considered.

Categorizing lifelines in a meaningful way will require an assessment of those lifelines that are most critical to response and recovery. For example, Category I lifelines, which must be restored to service within 4 hours, would include those systems that are critical to the immediate response to an earthquake such as:

• water supply and electrical power to critical response facilities;
• water supply for firefighting;
• emergency communications systems; and
• critical transportation elements, such as roads to critical facilities and ferry landings.

This categorization may take into account the partial restoration of systems necessary to achieve the performance goals, rather than the complete restoration of those systems. For example, back-up systems may be implemented to restore services to critical facilities within the desired time frame and redundancies or bypasses in areas where ground failure is likely would allow for the restoration of services once control of the system has been established.

Additionally, the restoration of the transportation system may be achieved by shifting capacity among modes of transportation and among different providers. If damage to roads and bridges cannot be repaired within the timeframes described above, the capacity of other forms of transportation, such as buses, ferries, and rail systems, may be increased beyond pre-earthquake capacities to accommodate the additional load. In addition to measures necessary to reduce potential damage to transit facilities, this may require such mechanisms as modifications to equipment, establishment of service in areas previously not served by specific systems, or implementation of incentives for mass transit use and restrictions on vehicle travel that encourage use of transit systems.