...Seismic Risk Cover Page Probability-Based Guidelines ...
Scenario Earthquakes for Urban Areas Along the Atlantic Seaboard of the United States
Scenario Events: Historical Examples, Global Analogs and Hypothetical Scenarios
In this section, we explore historical earthquakes in the eastern U.S. and from around the globe that – with some limitations – are marginally transferable to eastern U.S. urban regions for the purpose of providing some insight into the loss potential for a large east coast city. For the most part, we concentrate on New York City. We consider earthquakes as small as M = 4 and as large as M = 7.5. Ideally, the global analogs should be for contemporary cities, with a similar size of population, similarly complex infrastructure, inventory of built assets, economic systems, and from a similar seismotectonic environment. It is virtually impossible to find regional or global analogs that can combine all these conditions. Therefore, the estimates based on the analogs must be taken with a grain of salt. Often the loss estimates are not more than reasonable extrapolations. This sparseness of quantitatively based estimates points to the need that tools, data bases and efforts are urgently needed to compute loss scenarios with some measurable level of confidence. It should be noted that all projections for New York City assume a direct hit (unless otherwise noted), i.e. that the epicenter is located near the center of the city. Any epicenter location near or beyond the periphery of the city will drastically reduce all estimated losses, injuries, etc.
Case 1: On October 19, 1985, a magnitude M = 4.0 occurred at a depth of about 5 km near Ardsley, NY, in Westchester County, some 20 miles north of New York City. The earthquake was widely felt, had locally a maximum felt intensity on the Modified Mercalli Intensity scale of MMI = V (some dishes and windows broken, cracked plaster), but caused, if any, only negligible damage. If the event had occurred directly under New York City, damage – if any – would have been probably less than $1 million; essentially no injuries or fatalities are expected, although freak occurrences are always possible. An expected adverse effect would probably be phone gridlock and related business losses, if the event occurs during business hours.
Case 2: On January 16, 1994, a very shallow-depth ( 2km) M = 4.6 earthquake occurred near Wyomissing Hills and Reading, PA; it caused light (nonstructural) damage to roadways, buildings and other facilities, with maximum intensities of MMI = VI-VII. According to unverified County emergency officer’s accounts, total damage is claimed to be about $3 million. It is conceivable, but highly speculative, that if a similar event were to occur in New York City, damages (mostly nonstructural) on the order of perhaps $10 million may result. A few injuries or even fatalities (say from falling objects) are possible but unlikely.
Case 3: On January 31, 1986, a magnitude M = 5.0 occurred near Painesville, Ohio, in an area of small towns/suburbia. It produced maximum intensities of MMI = VI-VII to distances of about 15 km from the epicenter. Peak accelerations of 18% g combined with very short duration were measured at a nearby newly built, not yet operating nuclear power plant. Loss estimates are unknown, but 17 people were treated for minor injuries. We speculate that an equivalent event centered directly on New York City would cause minor structural damage and considerable nonstructural damage on the order of $0.1 to 1.0 billion, and on the order of 100 injuries, with perhaps even a few fatalities. However, a freak collapse of even a single poorly maintained masonry tenement building could change the fatality or injury figures dramatically. Note that the M = 5.0+ 1884 earthquake off-shore Jamaica Bay, about 25 miles from downtown Manhattan, caused an unknown amount of damage from falling chimneys, parapets, broken windows, but did apparently not cause any serious injuries and no fatalities. This points to the importance of the epicentral distance from an urban center.
Case 4: On Dec. 28, 1989, Newcastle, a mining and industrial town on the east coast of Australia, was directly hit by a M = 5.5 earthquake which affected about a quarter of a million people in and around the city. The dominant construction is unreinforced masonry a few stories high, not unsimilar to the stock of buildings constructed around the turn of the century in the eastern U.S. The damage reported for the Newcastle quake was on the order of $3 billion (Australian) of which nearly $1 billion were insured losses. Eleven fatalities and more than 100 injuries occurred. But since the quake happened during the Christmas/New Year holiday season in the Australian summer when everybody tries to head for the beach resorts, the town was quite empty and all schools and many offices were closed. Since walls, parapets, and ceilings collapsed in many school and office buildings, the toll could have been substantially higher than 100 if the quake had occurred during school/working hours. We only can guess that a New York City-centered equivalent of the same magnitude (M = 5.5) could cause losses in the order of $1-10 billion (U.S.); tens to hundreds of fatalities; injuries in the order of 1,000 to 10,000; and at least similar numbers of homeless people needing shelter (see also Case 5). Reminder: if not centered directly on New York City, losses should be smaller.
Case 5: An earthquake of a similar magnitude (M = 5.5) struck the city of San Salvador on Oct. 10, 1986. The difference from Newcastle (Case 4) is that the earthquake affected a city of about 1.4 million people with collapse of several large buildings. About 1,500 citizens perished, about 10,000 were injured, and 150,000 were reported homeless. Losses were estimated to be about $1.5 billion (U.S.) of which about 2/3 was in building losses. See Case 4 for New York City equivalent estimates.
Case 6: On February 29, 1960, the city of Agadir in Morocco sustained a direct hit by a M = 5.9 earthquake that killed 12,000 and caused (in 1960) an estimated $0.3 billion (U.S.) which would correspond to at least $2 billion in 1995 dollars. Note the high death-to-dollar ratio, (also observed in Case 5) which can be attributed to relatively low construction standards and (on average) a non-affluent population (by U.S standards). For this reason we do not draw a New York City parallel.
Case 7. On Oct. 12, 1992, a magnitude M = 5.9 struck about 30 km from downtown Cairo, Egypt. An estimated 8,300 buildings were damaged including several collapses; 561 people were killed and 6,500 injured. The Cairo metropolitan area is home to about 12 million people, similar to the greater New York City metropolitan area (the five boroughs of New York City proper have 7.3 million inhabitants). Dollar losses were not given. Expected analog numbers for the New York City metropolitan area, provided the event occurs at a similar distance from New York City, could be comparable.
Case 8a-c: This case involves three hypothetical New York City scenarios for a magnitude M = 6.0 earthquake occurring at three different distances from City Hall (17, 11, and 5 miles) and constitutes the only study published to date that has ever attempted to at least semiquantitatively estimate losses for New York City (Scawthorn and Harris, 1989). The study is limited in the sense that only the building losses are estimated, and only from shaking. No losses from fires are included, no losses to infrastructure or building contents or from business interruption; nor are fatalities, injuries, and homelessness evaluated. Hence the numbers given below may need to be doubled and perhaps quadrupled to include the losses other than those from only shaking to buildings alone. Total building assets were determined in that study to be about $0.4 trillion. The damage percentage in terms of asset values of buildings for the three distances were estimated to be about 3, 4, and 6%, respectively. The total losses for the three assumed distances from City Hall amount to about $11, 18, and 26 billion, respectively. The latter of these numbers could thus imply total losses (including indirect and secondary losses) on the order of $100 billion.
The Scawthorn and Harris (1989) study for New York City did not address the human toll since the predictability of fatalities, injuries and displaced is even more uncertain than damage estimates. But the public needs to know what to expect. Short of adding speculations on fatalities, injuries, and people needing shelter for all three cases, we only attempt a vague estimate for the shortest epicenter distance (5 miles from city hall). Therefore, keep in mind that for larger distances, lesser human toll ought to be possible. We conclude (by extrapolating from examples from elsewhere that are hardly comparable, however) that the toll from a direct hit may include ‘numbers’ that are probably in excess of 1,000 fatalities, 10,000 injuries of various degrees of seriousness, and 100,000 displaced many of whom may need temporary shelter. Be reminded that 100,000 people means only 1.4% of the official population of New York City. Such ‘numbers’ imply a burden on emergency providers for which, at this time (and to the author’s knowledge), no realistic plans exist to meet these challenges.
Case 9: In 1990, a limited study for the greater Boston area (inside the Route 128 ring) was prepared (for a summary of the study see Ebel, 1993). The scenario assumed that the area was to be hit by an approximate repeat of the M = 6.4 off Cape Ann earthquake of Nov. 18, 1755. The epicenter is approximately 50 km from downtown Boston. Ignoring all damage outside the Route 128 ring, the study asserts a loss ranging between $2 and 10 billion, not considering secondary losses from business interruption. Hundreds of fatalities and thousands of injuries and displaced are quoted in the study, although the authoring committee pointed to the study’s severe limitations in terms of availability of a sufficiently quantitative data base. It is also obvious that the quoted numbers should rise as the quake’s location moves closer to downtown Boston.
There are no good eastern U.S. nor suitable global analog case earthquakes of recent times that occurred near major cities in areas of low seismicity for the magnitude range M = 6.3 to 6.8. Therefore, let us proceed to events with magnitudes M 7, although their occurrences near a major east coast city have low probabilities.
There are two historic events along the coast of eastern North America: Charleston and Grand Banks; in addition, we will mention the Tangshan, China earthquake. The recent Kobe (Great Hanshin) M = 7.2 earthquake of January 17, 1995, is discussed elsewhere in this volume. It occurred not far from an active plate boundary rather than in the midst of a plate (where New York City is located). Hence the Kobe experience can be translated to New York City only with upward adjustments of losses accounting for the even lower preparedness and higher vulnerability of the assets of New York City, compared to Kobe.
Case 10: On Nov. 18, 1929, the Grand Banks earthquake of Mw = 7.4 0.2 occurred about 200 km offshore Nova Scotia and Newfoundland on the Atlantic Shelf of Canada. Because of its distant offshore location, its severest impact on land was not from shaking but from a tsunami (seismic sea wave) which caused 52 deaths as it ran ashore, and an unreported dollar loss. It is also famous for causing an offshore underwater landslide and related mud flow which interrupted all existing trans-Atlantic telephone cables at the time that connected North America with Europe. This earthquake is important for reminding us of the possibility of large offshore earthquakes that may occur also along the U.S. portions of the Atlantic coast.
Case 11: On August 31, 1886, a magnitude Mw = 7.6 0.3 occurred centered on Summerville, about 20 km from Charleston, SC. Only about 60 fatalities occurred in Charleston, probably due to the fact that a large number of houses were wood frame construction. Walls in many brick buildings collapsed. About 14,000 chimneys were destroyed in Charleston. In that city alone the loss, measured in 1886 dollars, was about $5 million, corresponding to almost $1 billion in 1996 dollars for the 1886 assets, not considering that since then the stock of built assets has increased greatly not only in Charleston but elsewhere on the nearby coastal plain. This event is also known for the widespread soil liquefaction it caused, and for isolated high-intensity effects more than 200 km from the epicenter. The occurrence of a similar earthquake even at a distance of 100-200 km from New York City could be near catastrophic because it may cause the failure of many New York City high-rise buildings, especially if located on soft soils. A situation could arise that, in principle but not in detail, is analogous to what Mexico City experienced in 1985 from a M = 8 earthquake almost 300 km away, causing the death of an estimated 10,000 people mostly from the collapse of high-rise buildings on deep soft soils; or what the city of Leninakan, Armenia, experienced from a M = 6.8 earthquake in 1988 where 127 of 133 poorly designed high-rise buildings on deep soft soils collapsed. Leninakan was located about 30 to 100 km from the various portions of the earthquake fault, on a former volcanic lake bed, a condition similar to Mexico City. A total of more than 30,000 people perished in the Armenian quake. Similar construction on rock in the town of Kirovakan was damaged but did not collapse. The combined lessons from Mexico City, Leninakan, Loma Prieta, CA, and Kobe, Japan, together with similar experiences elsewhere point to the potential risk for long-period (high-rise) structures on soft soils especially during large earthquakes (M > 6.8). In some cases, this applies even when the high-rise buildings on soft soils are located at distances of 100 km or more from the earthquake rupture. This also corroborates the findings we arrived at during the Tappan Zee Bridge study described earlier.
Case 12: Nearly 20 years ago, on July 27, 1976, a magnitude M = 7.7 earthquake together with its principal aftershock of M = 7.2 destroyed the city of Tangshan in northeast China and killed, according to official accounts, 240,000 people (the London Times reported much later, on January 6, 1977, an unconfirmed account that over 650,000 may have been killed). The fatalities are attributable predominantly to the massive collapse of unreinforced masonry buildings. The total population of the city was an estimated 1.4 million before the quake. The tectonics and seismic wave attenuation properties of this region are marginally akin to that of the eastern U.S. in that it is an intracontinental region not directly associated with a currently active plate boundary. The preponderance of unreinforced masonry buildings is another marginal similarity to major U.S. east coast cities, although the details of construction are likely to be somewhat different. This earthquake should be considered the ultimate, but not very likely analog for a New York City event. Although the magnitude is similar to the estimated magnitude of the Charleston, SC, earthquake, it is not known whether faults near New York City are capable to break in such large quakes or not. On the other hand, New York City has a 5 times larger population than Tangshan, and therefore if such a magnitude would ever be deemed a possibility at close proximity to New York City, the Tangshan event may have to be considered the ultimate planning scenario. Whatever planning for New York City will be undertaken in the future, such an event clearly borders on the unmanageable. Because New York City’s function as an international center for finance, commerce and the arts, such an event would be a catastrophe not only for the region but of lasting national and global impact.
Home | Background | HAZUS | Research Activities | Education/Outreach
Who We Are | Technical Documents | Contacts | Links
Last modified: 4/18/2002 10:07:50 AM
Suggestions and comments to the webmaster.