All shook up: if the 1906 San Francisco earthquake occurred today, insured losses would likely reach $80 billion; however, there are many scenarios capable of producing even greater losses.

Key Points

* The 1906 San Francisco earthquake remains one of the most significant earthquakes in U.S. history, particularly because the actions that followed had a far-reaching impact on the understanding of earthquakes and their effects.

* AIR Worldwide estimates a recurrence of the 1906 San Francisco earthquake would result in nearly $80 billion of insured property losses, based on total property losses approaching $300 billion.

* Leading companies are increasingly employing catastrophe models to minimize the risk of extreme events.

This month marks the 100th anniversary of the Great San Francisco Earthquake and Fire of 1906, a catastrophe that impacted thousands of lives, caused hundreds of millions of dollars in damage and generated aftershocks in the scientific. engineering and insurance communities that continue to resonate today. Enormous advances have been made in scientists' understanding of what causes earthquakes and in builders' ability to erect structures that have a better chance of withstanding them. Nevertheless, if an earthquake of similar magnitude were to occur in a major metropolitan area today, the losses would be staggering.

A Terrific Trembling

At 5:12 a.m. on April 18, 1906, residents throughout the San Francisco Bay area were rudely awakened by the unmistakable rumblings of an earthquake. The foreshock proved to be just a preview of the main event Less than 30 seconds later, a moment magnitude 7.9 earthquake struck with an epicenter located just seven miles from the center of San Francisco. Moment magnitude measures total energy released by an earthquake.

The quake ruptured approximately 290 miles of the San Andreas fault. It was felt from Oregon to south of Los Angeles and as far east as central Nevada. A horizontal displacement of more than 20 feet was recorded in Marin County.

While the event has come to be known as the San Francisco earthquake, ground shaking of at least equal severity was experienced at many, though less populous, locations along the length of the rupture. To the south, the business district of San Jose was left in ruins, while some 200 miles to the north, massive landslides occurred in and near Petrolia. Santa Rosa was located in a pocket of very intense ground shaking, which reduced its downtown to rubble.

The shaking lasted almost a minute and ruptured water and gas mains in San Francisco. Fires broke out and, without a reliable source of water for fire suppression, merged into a full-scale urban conflagration. The city's fire department and the U.S. Army ultimately resorted to dynamiting buildings that had survived the earthquake in an attempt to create fire breaks. The fires raged out of control for three days and destroyed nearly five square miles of homes and commercial buildings.

According to data from the U.S. Geological Survey, an estimated 28,000 buildings were destroyed in all. Total property damage is estimated at $400 million in 1906 dollars. Out of a population of 400,000, some 225,000 were left homeless. While official estimates of the number of fatalities caused by the quake at the time were on the order of 700, a more recent study concluded the number was probably closer to 3,000.

A Call to Action

The 1906 San Francisco earthquake remains one of the most significant earthquakes in U.S. history, particularly for the actions taken in response to it--actions that would have a far-reaching impact on the understanding of earthquakes and their effects.

Three days after the earthquake, California's governor appointed a State Earthquake Investigation Commission to formalize and unify the efforts of scientists already examining the event. Hundreds of people were interviewed and photographs taken from every area damaged by the quake, which established a pictorial record against which tectonic and geologic changes in the San Francisco Bay area since 1906 can be compared. The final report, known as the Lawson Report after the commission's head, became a pioneering model for investigating and documenting earthquakes, a paradigm that is still widely followed today.

The commission found that building damage was strongly correlated with construction type and design, and the extensive damage to adobe and brick buildings was exacerbated by poor construction practices. The report also concluded that the failure of cripple walls--short, often inadequately braced stud walls that extend from the top of the foundation upward to carry the floor joists--led to many incidents of wood frame buildings having been "thrown from their foundations."

The team meticulously recorded and mapped the observed effects of ground shaking along the entire length of the rupture and found a strong correlation between damage and the local geology. The extensive damage in the China Basin and what is now the Marina District prompted them to postulate a relationship between soil types and seismic wave amplification.

By performing triangulation surveys, the team found the largest ground displacements occurred near the fault and decreased away from it. The observations led commission member H.F. Reid, a professor at Johns Hopkins University, to postulate that the surface of the Earth gradually--and elastically--distorts from the accumulating strain of relative ground motion until the strain is suddenly and violently released in the form of an earthquake. Known as the "elastic-rebound" theory, it is perhaps the commission's most important contribution and continues to form the basis of understanding of how earthquakes behave.

Risk Landscape Changes

The science of seismology has advanced considerably since the groundbreaking work of the commission, including the theory of plate tectonics, which only gained wide acceptance in the 1950s and 1960s. Strong-motion earthquake data gathered by networks of sophisticated sensors, supplemented by data on prehistoric earthquakes from fault trenching and, more recently, GPS data on the relative movement of the Earth's crust have given scientists a much better understanding of the likely frequency and severity of future earthquakes.

Despite the increased understanding of earthquakes and their causes, a major earthquake affecting a major metropolitan area will undoubtedly cause enormous damage and loss. The decreased vulnerability of the building stock, as compared with that of 1906, has been offset to large degree by the increase in the value and number of exposed properties, and changes in the earthquake insurance market.

Building for the Future

Through methods ranging from sophisticated computer simulation techniques to full-scale shaking table tests, engineers now have a better understanding of how buildings respond to earthquakes. As a result, today's buildings are far more resistant to the lateral loads imposed by ground shaking than they were in the early part of the century.

The change, however, has been gradual. In the urgency to rebuild San Francisco in the immediate aftermath of the 1906 earthquake, structures were erected with little regard to their ability to withstand future such events. Over time, more stringent building codes were adopted. These efforts eventually resulted in the creation of the Uniform Building Code, which provides standards for designing structures in six defined levels of seismic hazard across the country.

For commercial properties, the current mix of construction types also ameliorates potential losses. There are far fewer masonry buildings than in 1906 and a much higher percentage of steel frame buildings, which perform much better under earthquake loads. Many older buildings--particularly commercial structures--have undergone earthquake retrofit to bring them in line with current standards. In addition, today's building materials are far more fire-resistant than they were in 1906.

Counterbalancing the increased resistivity of the building stock is the increase in the number and value of insured properties. The total population affected by a recurrence of the quake has grown significantly, as the population in the San Francisco metropolitan area has grown to more than 1.7 million according to the 2000 Census. AIR Worldwide estimates the current value of residential and commercial properties within the damage footprint of the 1906 quake at more than $1.6 trillion.

Insured losses depend not only on the severity of the earthquake and the vulnerability of the exposure, but also on policy conditions (deductibles and limits) and take-up rates (the percentage of properties actually insured against the earthquake peril). Take-up rates increased after both the 1989 Lorna Prieta earthquake and the Northridge earthquake of 1994--reaching about 35% by 1995--but have since declined. Today it is estimated that only about 13% of residential policies currently carry earthquake coverage. While take-up rates for commercial properties are higher, in the decade since the Northridge earthquake, complacency has clearly set in.

The Cost Today

Using its U.S. Earthquake Model, AIR estimates a recurrence of the 1906 San Francisco earthquake would result in nearly $80 billion of insured property losses, based on total property losses approaching $300 billion.

In addition to property losses, there is the potential for significant losses to other lines, such as life and workers' compensation. The methodology developed by modelers for estimating the damage to structures can be leveraged to estimate the number of injuries and fatalities that result from catastrophe events. AIR estimates close to $7 billion in workers' compensation losses if the 1906 earthquake were to recur during working hours. As many as 5,000 fatalities could occur and more than 50,000 injuries.

For a catastrophe of this magnitude, total economic losses could easily exceed $400 billion.

While estimating losses for a recurrence of a historical event is an interesting and even informative exercise, an exact recurrence of the 1906 earthquake is highly improbable.

It is important to remember that it is not the risk of individual events that needs to be managed, but rather the risk of losses of various sizes, since many different events can cause the same size loss.

Many scenarios are capable of producing an insured loss of $80 billion or greater. One such scenario is a large magnitude earthquake on the Hayward fault, which lies almost parallel but to the east of the San Andreas. Significant damage could be expected in Oakland, all the way south to San Jose and west to such areas as Palo Alto. The last major earthquake on the Hayward fault occurred in 1868, when a magnitude 7.0 earthquake caused considerable damage, including to buildings in San Francisco.

In southern California, a major rupture of the Puente Hills fault is capable of producing an even greater loss. Located directly beneath Los Angeles, the fault was discovered in 1999. AIR estimates insured losses resulting from a significant rupture of this fault at more than $140 billion due to the high number and value of affected properties. Fortunately, the estimated return period associated with such a scenario is quite long. Other scenarios, such as a large magnitude earthquake occurring on the nearby Sierra Madre fault, have a higher probability of occurrence. AIR estimates that a rupture of this fault would result in an insured loss of about $80 billion, similar to a recurrence of the 1906 earthquake.

Risky Business

Catastrophe models enable companies to anticipate the likelihood of losses of various magnitudes so they can adequately prepare for their impact. AIR estimates that an industry loss similar in size to one resulting from a recurrence of the 1906 earthquake has about a 0.2% annual probability of occurrence. While this annual probability may appear small at first glance, it should not be dismissed. It translates, for example, into a 6% probability over the next 30 years and a 10% probability of an $80 billion or greater insured loss over the next 50 years.

To maximize the chances of surviving large losses, such as the scenarios discussed in this article, companies should use models to understand their entire modeled loss distribution, including losses well beyond the 100- or 250-year return period. Leading companies are increasingly employing catastrophe models to minimize the risk of extreme events through portfolio management--a process that encompasses actions such as reducing over-concentrations of exposures, evaluating various risk transfer options and developing underwriting guidelines to allow growth in premiums while controlling loss potential. By identifying ZIP codes and even individual policies that have the largest impact on losses in the tail of the distribution, companies can use catastrophe models to help them reduce their probable-maximum-loss-to-premium ratio and improve profitability.

RELATED ARTICLE: Widespread danger

WIDESPREAD DANGER: A large magnitude earthquake on the Hayward fault, which lies almost parallel but to the east of the San Andreas fault, could cause significant damage from Oakland to San Jose and as far west as Palo Alto.

RELATED ARTICLE: >Los Angeles loss prediction.

LOS ANGELES LOSS PREDICTION: AIR estimates that a major rupture of the Puente Hills fault, located directly beneath Los Angeles, could result in insured losses of more than $140 billion.

RELATED ARTICLE: The Seismic setting of the continental United States.

Based on geological criteria, the United States is divided into two broad regions for purposes of modeling seismic risk. These two regions can be roughly categorized as corresponding to the Western United States and the intraplate region of the Central and Eastern United States.

Seismicity in California is primarily controlled by the boundary of the Pacific and North American plates. This transform boundary consists of a complex web of faults. The principal member of this fault system is the San Andreas fault. Seismicity in the Pacific Northwest is dominated by the subduction of the Juan de Fuca Plate beneath the North American Plate at a rate of about 42 mm/yr. The tectonic stresses here are capable of producing great (>M8.0) earthquakes.

Unlike the plate boundary regions of the Western United States, seismic sources in the Central and Eastern United States are not as well understood. Most scientists now believe that these intraplate regions are characterized by traces of ancient geological deformations, or zones of weakness, that resulted from the collision and rifting of tectonic plates over hundreds of millions of years ago. These zones of weakness can be reactivated by present-day stresses.

Source: AIR Worldwide Corp.

Contributor Jayanta Guin is vice president of research and modeling at AIR Worldwide.

Probability of an $80 Billion or Greater Insured Loss (similar in
size to one resulting from a recurrence of the 1906 Earthquake)

Annual                   0.2%
Over the next 30 years     6%
Over the next 50 years    10%

Source: AIR Worldwide