Fire Following Earthquakes

29
Fire Following Earthquakes

29.1 Introduction 29.2 Fires following Selected Earthquakes
The 1906 San Francisco Earthquake and Fires The 1923 Tokyo Earthquake and Fires The 1989 Loma Prieta, California Earthquake and Fires The 1994 Northridge, California Earthquake and Fires The 1995 Hanshin (Kobe), Japan Earthquake and Fires The 1999 Marmara, Turkey Earthquake and Fires Summary
29.3 Analysis
Ignition Fire Report and Response Fire Growth and Spread Fire Response and Suppression Final Burned Area Example: Vancouver, B.C.
29.4 Mitigation
Reduction of Damage Automatic Suppression Suppression by Citizens Communications Fire Department Assets Water Dedicated Fire Protection Systems
Charles Scawthorn
Consulting Engineer Berkeley, CA
29.5 Conclusion Dening Terms References Further Reading
29.1 Introduction
Earthquakes cause damage by a variety of damaging agents, including fault rupture, shaking, liquefaction, landslides, res, release of hazardous materials, tsunami, etc. Shaking is present in all earthquakes, by denition, and is the predominant agent of damage in most earthquakes. Occasionally, however, building characteristics and density, meteorological conditions, and other factors can combine to create a situation in which re following earthquake, or post-earthquake conagration, is the predominant agent of damage. Large res following an earthquake in an urban region are relatively rare phenomena, but have occasionally been of catastrophic proportions. In both Japan and the United States, re has been the single most destructive seismic agent of damage in the twentieth century. The res following the San Francisco 1906 and Tokyo 1923 earthquakes, which were both terribly destructive, rank as the two largest peacetime urban res in mans history. While not widely perceived today by the public or even many professionals in the earthquake or re service elds, re following earthquake is recognized by professionals specializing in this eld as continuing to pose a very substantial threat in both countries.
2003 by CRC Press LLC
29-2
Earthquake Engineering Handbook
Although re following the 1906 earthquake was the overwhelming cause of the damage to San Francisco and Santa Rosa, and has continued as a signicant cause of damage since, it has received relatively little attention in the United States. This is perhaps due to several factors: Earthquakes historically have been the professional concern of seismologists and structural engineers, who, as a class of professionals, are largely uninformed about re. Fire protection engineers and re service personnel have similarly ignored earthquakes, seeing their goal as the mitigation of chronic re losses by code implementation and other techniques, rather than as earthquake response. Major conagrations were a common occurrence in the United States prior to World War II, so that the 1906 experience was seen as more of a conagration than an earthquake phenomenon. The subsequent decline in U.S. urban conagrations, due to improved re and building codes and to improved re service response due primarily to the use of radios, has only increased this sense of it cant happen here. Less apparent, but perhaps even more important, factors contributing to the lack of attention paid to post-earthquake re have been: The United States has not experienced a major urban earthquake since 1906. It is little appreciated that it takes a great earthquake, striking a large urban region, to create the condition of dozens or hundreds of ignitions overwhelming the re service, thus creating a conagration. San Francisco 1906 and Tokyo 1923 fullled this condition. Earthquakes since 1906 (in the United States: 1933 Long Beach, 1964 Alaska, 1971 San Fernando, 1987 Whittier, 1989 Loma Prieta, and 1994 Northridge; and in Japan: 1968 Tokachi-oki, 1978 Miyagiken-oki, 1984 Nihonkai-chubu, and 1995 Kobe) have generally not fullled this condition. Note, however, that there were many ignitions in 1971, 1989, 1994, and 1995, and that there were conagrations of many acres in Kobe in 1995. There is a general lack of awareness of the existence of an analytical framework within which to analyze the many factors involved in post-earthquake re and to quantify these factors and the outcome many small res, or conagration? That large res following earthquakes remain a problem is demonstrated by ignitions following recent earthquakes such as the 1994 Northridge and 1995 Kobe earthquakes, as well as several recent large nonearthquake conagrations, including the 1991 East Bay Hills and 1993 Southern California wild res. While long a concern to re departments and the insurance industry, consideration of the re following earthquake problem has been subject to debate regarding the likelihood and severity of post-earthquake res in any future events. Until recently, perhaps the only group at all concerned with post-earthquake re has been the insurance industry, which, after 1906, became quite aware of the potential for catastrophic loss by way of this phenomenon. Steinbrugge [1982] presents probably the best summary of knowledge deriving from this eld. Scawthorn et al. [1981], based on work by Hamada [1951], Horiuchi [n.d.], Kobayashi [1979], and others in Japan, developed a probabilistic post-earthquake re ignition and spreading model that has subsequently been applied at two levels: 1. Jurdisdictional a detailed modeling, with ignitions, re loading, engine location, and other parameters modeled gridwise at about the 10-hectare level of resolution. Due to the sizable data collection and computational effort involved, this model has been applied to only one U.S. jurisdiction, the City of San Francisco [Scawthorn, 1984]. 2. Regional a coarser model based on approximations derived from the jurisdictional model. Applied to the San Francisco and Los Angeles areas and other regions [Scawthorn, 1987; Scawthorn and Khater, 1992; Scawthorn and Waisman, 2001], this model for the rst time permitted quantied estimates of the aggregate losses due to re following earthquake. This work has largely served the needs of the insurance industry. The fact that re following earthquake has been little researched or considered in North America is particularly surprising when one realizes that the conagration in San Francisco after the 1906 earthquake
29-3
was the single largest urban re in history to that date. It remains today the single largest earthquake loss in U.S. history, in terms of life and economic loss. The loss over 3 days of more than 28,000 buildings within an area of 12 km2 was staggering: $250 million in 1906 dollars, and more than 3,000 killed.1 That re has since only been exceeded in a peacetime urban re by the conagration following the 1923 Tokyo earthquake, in which more than 140,000 people were killed and 575,000 buildings destroyed (77% of those buildings were destroyed by re) [Usami, 1996]. Fires following large earthquakes are a potentially serious problem because of the multiple simultaneous ignitions that re departments are called to respond to, while, at the same time, their response is impeded by impaired communications, water supply, and transportation. Additionally, re departments are called to respond to other emergencies caused by the earthquake, such as structural collapses, hazardous material releases, and emergency medical aid. Most recently, the January 17, 1994 Northridge (M 6.6) and January 17, 1995 Kobe (M 7.2) earthquakes have again emphasized the importance of the problem of re following earthquake. To aid in an understanding of this problem, this chapter rst discusses experiences regarding res following selected earthquakes, next presents a methodology for analysis of the problem, and then discusses actions that can be taken to mitigate the problem.
29.2 Fires following Selected Earthquakes

This section discusses selected U.S. and foreign earthquakes, with emphasis on the res following these events. More general information on each event is provided in Chapter 1 of this book. Table 29.1 lists all U.S. twentieth-century events with post-earthquake ignitions.
29.2.1 The 1906 San Francisco Earthquake and Fires

The April 18, 1906 earthquake was the most devastating earthquake in U.S. history. The region of destructive intensity extended over a distance of 600 km (see Chapter 1, this volume), with strong shaking over a wide area. Despite the strong shaking, the vast majority of the damage in the entire earthquake, and especially in San Francisco, was caused by re. Scawthorn compiled data on the 52 known ignitions in the City of San Francisco (see Table 29.2). 29.2.1.1 Ignitions and Fire The conagration following the 1906 earthquake was a complex re, actually consisting of several separate major res that grew together until there was one large, burnt area that comprised the northeast quadrant of the city and destroyed more than 28,000 buildings. The progress of these res has generally been divided into four periods [NBFU, 1906; Bowlen, n.d.], although actual times for these periods differ between sources. Generally, the periods comprise the following times: 1. From the earthquake until mid or late in Day 1, when most of the south of Market section had been destroyed, but the higher value north of Market section still remained largely intact 2. The night of Day 1 and the early hours of Day 2, when the north of Market section was invaded by the re progressing from the west 3. Continued progress of the re to the north, and a bit to the south, during the remainder of Day 2 4. During Day 3, the last day, when the re progressed almost entirely to the north, around Telegraph Hill, and burned down to the bay
1 Exact number of fatalities is unknown until the 1980s, it was believed approximately 700 had been killed. Research by Gladys Hansen, San Francisco Librarian, indicated that far more people killed had not been accounted for. In painstaking research over many years, she slowly gathered evidence from letters of the time, gathered from all over the world, of many more deaths. Of particular interest was the fact that many minority fatalities, especially in San Franciscos large Chinatown, were known in 1906, but not included in the ofcial count. Her work is ongoing as of this writing, and the count is still increasing. See Hansen and Condon [1989].
29-4
TABLE 29.1 Twentieth-Century Post-Earthquake Ignitions

Year 1906 1906 1906 1906 1906 1906 1906 1933 1933 1933 1952 1957 1964 1969 1971 1971 1971 1971 1971 1979 1983 1984 1984 1986 1987 1989 1989 1989 1989 1989 1989 1989 1994 1994 M 8.3 8.3 8.3 8.3 8.3 8.3 8.3 6.3 6.3 6.3 7.7 5.3 8.3 5.7 6.7 6.7 6.7 6.7 6.7 6.4 6.5 6.2 6.2 5.9 6 7.1 7.1 7.1 7.1 7.1 7.1 7.1 6.8 6.8 City or Area Affected Berkeley Oakland San Francisco San Jose Santa Clara San Mateo Co. Santa Rosa Los Angeles Long Beach Norwalk Bakerseld San Francisco Anchorage Santa Rosa Burbank Glendale Los Angeles Pasadena San Fernando El Centro Coalinga Morgan Hill San Jose N. Palm Springs Whittier Daly City Berkeley Marin Co. Mountain View San Francisco Santa Cruz Santa Cruz Co. Los Angeles Santa Monica Ignitions 1 2 52 1 1 1 1 3 19 1 1 1 7 2 7 9 128 2 3 1 4 4 5 2 38 3 1 2 1 26 1 24 77 15 MMI VIII-IX VII-IX VII-X VIII VIII-IX VIII X VI-VII IX VII-VIII VIII VII X VIII VII VI-VII VI-VII VII IX VII VIII VII VIII VI-VII VI VI VI VI VII VII VIII VII-VIII VI-IX VIII
Note: Various sources, compiled by author.
Figure 29.1 shows some of the ignitions in the Central Business District, around the foot of Market and Mission Streets and their spread during the rst day, while Figure 29.2 shows the nal extent of the burnt area. Figures 29.3 to 29.5 show scenes during and after the re. 29.2.1.2 Fire Department Response The San Francisco Fire Department in 1905 protected approximately 400,000 persons occupying an urbanized area of about 21 mi2. The department consisted of 585 full-paid re force personnel (resident within the city and on duty at all times), commanded by Chief Dennis T. Sullivan and deployed in 57 companies (38 engines, 1 hose, 10 ladders, 1 hose tower, and 7 chemical) [NBFU, 1905]. The distribution of these companies was well conceived, being centered about the congested high-value district (i.e., the Central Business District, or CBD, known in San Francisco as the Financial District), with 24 engines, 8 ladders, 1 water tower, and 7 chemical companies within 2 mi of the center of the CBD. All but 2 of the 38 steam engine companies dated from 1890 or later and were rated at an average of 680 gallons per minute (gpm), although the 8 engines tested in 1905 averaged only about 70% of their rated capacity, and the ability of the men handling the engines was in general below a proper standard [NBFU, 1905]. The rated pumping capacity of the 38 rst-line and 15 relief and reserve engines totaled 35,100 gpm. In
29-5
TABLE 29.2 San Francisco 1906 Earthquake Ignitions in City of San Francisco
No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 Location 211 /ckat (100' W of Davis) 119-123 Clay St. 15 Fremont Drumm & Jackson Sts. (401 Drumm?) Dupont (Grant)@Pacic & Bdwy NW crnr William & OFarrell 6th & Howard (130-134 6th St) Montgomery bet Bush & Pine 307-11 Davis nr Clay Blk bet Calif, Mkt & Davis Sansome North of Pine Calif & Battery, NW corner 219 Front nr Sacramento Mission & 22nd St GG Ave & Buchanan (NW crnr) Fulton & Octavia (NW crnr) Hayes & Laguna Ashbury & Walter Masonic & Waller Ashbury Hgts Homestead Dist (FIRES?) Homestead Dist (FIRES?) Pacic @ Leavenworth (NW crnr) 17th & Clement 3rd & Clement Oak nr Stanyon 2625 1/2 Harrison (22nd & 23rd Sts) Butchertown (Hunters Pt?) Butchertown (Hunters Pt?) Polk bet Bush & Pine (1312 Polk) Bay & Powell Bay & Kearny Steuart (E) @ Mission & Howard Fremont (E) @ Mission & Howard Howard E of 3rd St 282 Natoma (E of 4th St) Bet 5th & 6th Sts, Mkt to Harrison Bet 5th & 6th Sts, Mkt to Harrison 395? Hayes (S) 75' E of Gough Davis bet Pacic & Bdwy Front bet Vallejo & Bdwy Davis & Vallejo (crnr) East St bet Bdwy & Vallejo Golden Gate Ave nr Divisadero Bush & Kearny Market & Kearny Geary & Stockton (239-241 Geary) Howard & 12th (NW crnr) Sutter & Polk (1215 Sutter) Minna * 5th Jessie nr 3rd Market & Beale E12 Responding Company/Comments
T2 extinguished E2; exting w/sand (Capt. Brown) E6 retreated E fr 6th bet Folsom & Clementina No re indicated; gas pressurized E12 extinguished Very rapid spread, frame bldg; E1 In basement, very rapid spread, frame bldg E1 respndd, no water Extinguished after 5 hrs; E13, 18, 24, 37, T7, E25 Extinguished after 5 hrs; E15, 23, 27, 30, 34 Extinguished after 5 hrs; E14, 21, 34, Chem4 Extinguished after 5 hrs; E14, 21, 34, Chem4, T6 E30, extinguished E30, extinguished E30 (Argonaut reports E30 exting 3 res Asbury Hgts) E32; district is S of Mission dist, nr county line E32; district is S of Mission dist, nr county line E31; exting w/sand (NB; not same as E2) E36; exting by bucket brigade E26; exting small re rear of store E22; exting? E25, T9; exting E11 E11 E3, T4 E28; stack collapse on kiln; exting in 2 hrs E20, no water; redirect to foot of Washington St re E1, 9, 38 Little water Spilled coals; E4 (re stn next dr), little water No. FIRES & locations unknown; T3 arrives at 0800 No. FIRES & locations unknown; T3 arrives at 0800 Ham & Eggs re; E14, 24, 34; see NBFU Fig 14; no water Exting by Chem Co Exting by Chem Co Exting by Chem Co Exting by Chem Co E30 (poor water, re conned to 1/4 blk; 12 hses lost) Noticed in neighborhood No water E2, exting, SAR E22, no water, attempted to draft fr sewer unsuccessfully T4, exting Soon merged with Jessie & 3rd re Soon merged with Minna & 5th re E1, but a dozen powerful streams
Source: Scawthorn, C. and ORourke, T.D. 1989. Effects of Ground Failure on Water Supply and Fire Following Earthquake: The 1906 San Francisco Earthquake, in Proc. 2nd U.S.Japan Workshop on Large Ground Deformation, July, National Center for Earthquake Engineering Research, Buffalo, NY.
29-6
FIGURE 29.1 San Francisco 1906 re: ignitions and spread, Central Business District. (From Scawthorn, C. and ORourke, T.D. 1989. Effects of Ground Failure on Water Supply and Fire Following Earthquake: The 1906 San Francisco Earthquake, in Proc. 2nd U.S.Japan Workshop on Large Ground Deformation, July, National Center for Earthquake Engineering Research, Buffalo, NY.)
summary, the department was rated by the National Board of Fire Underwriters [NBFU, 1905] as efcient, well organized, and, in general, adequate. The NBFU, however, concluded in 1905 that: In fact, San Francisco has violated all underwriting traditions and precedent by not burning up. That it has not done so is largely due to the vigilance of the re department, which cannot be relied upon indenitely to stave off the inevitable.
29-7
FIGURE 29.2 San Francisco 1906 re: nal extents. (From USGS. 1907. The San Francisco Earthquake and Fire of April 18, 1906 and Their Effects on Structures and Structural Materials, Bull. No. 324, U.S. Geological Survey, Washington, D.C.)
FIGURE 29.3 San Francisco 1906 collapsed building, re in background. (Courtesy the National Oceanic and Atmospheric Administration)
Within moments after the earthquake, Chief Dennis T. Sullivan was seriously injured as a result of the damage done to the re station where he was sleeping, and later died. Ten re stations sustained major damage [Tobriner, personal communication], although no engines were seriously disabled by the earthquake and all went into service [NBFU, 1906]. Street passage was, in general, not a problem, and a number of res were quickly suppressed, although many more could not be responded to. That is: Fires in all parts of the city, some caused directly by earthquake, some indirectly, prevented an early mobilization of re engines and apparatus in the valuable business district, where other original res had started and were gaining headway. [NBFU, 1906]
29-8
FIGURE 29.4 View down Sacramento Street during the San Francisco 1906 re. (Photo: Arnold Genthe, Steinbrugge Collection, courtesy of the University of California Berkeley Earthquake Engineering Research Center.)
FIGURE 29.5 San Francisco destruction by re: view looking west from Telegraph Hill, showing unburned houses on summit of Russian Hill. St. Francis Roman Catholic Church, with excellent brick walls, in foreground. (From USGS. 1907. The San Francisco Earthquake and Fire of April 18, 1906 and Their Effects on Structures and Structural Materials, Bull. No. 324, U.S. Geological Survey, Washington, D.C.)
The NBFU Conagration Report [NBFU, 1906] concluded: The lack of regular means of communication and the absence of water in the burning district made anything like systematic action impossible: but it is quite likely that during the early hours of the re the result would not have been otherwise, even had none of these abnormal conditions existed [sic]. That is, the NBFU concluded that, even under normal conditions, the multiple simultaneous res would probably have overwhelmed a much larger department, such as New Yorks, which had three times the apparatus. Nevertheless, Bowlen [n.d.] concluded that, by 1:00 p.m. (i.e., about 8 h after the earthquake), the re department, except that it was without its leader, was in fairly good shape, that is the men and horses were in good trim for reghting, the apparatus was in shape and could be worked where there was water. There is not one report of an engine or man going out of commission during the early hours of the re, and the department was hard at work all the time, even though there was little to show for its effort. 29.2.1.3 Water System Performance Several factors contributed to the initial ignitions rapidly growing out of control. While the weather was relatively hot and dry, undoubtedly the primary factor leading to the conagration was the failure of the water system [Scawthorn and ORourke, 1989]. In summary, in 1906, water to San Francisco was supplied from two series of reservoirs. South of San Francisco, water was impounded by earth and concrete dams
29-9
N
SAN FRANCISCO
Lake Honda Reservoir Lake Merced College Hill Reservoir
San Andreas Reservoir Legend: Fault Pipeline Lateral spread damage Bridge damage Fault damage Scale: 0 2 4 km
FIGURE 29.6 1906 San Francisco water supply (adapted from Schussler, 1906). (From Scawthorn, C. and ORourke, T.D. 1989. Effects of Ground Failure on Water Supply and Fire Following Earthquake: The 1906 San Francisco Earthquake, in Proc. 2nd U.S.Japan Workshop on Large Ground Deformation, July, National Center for Earthquake Engineering Research, Buffalo, NY.)
to form the San Andreas, Crystal Springs, and Pilarcitos reservoirs. Transmission pipelines conveyed water from these reservoirs to a second series of smaller reservoirs within the city limits. Water then was distributed throughout the city by means of trunk and distribution pipelines. Figure 29.6 presents a plan view of the 1906 San Francisco water supply adapted from maps prepared by Schussler. At the time of the earthquake, the San Andreas, Crystal Springs, and Pilarcitos reservoirs held a combined volume of 88.7 billion liters. They supplied nearly all water for the city of San Francisco in 1906, whereas today, they represent approximately one half of the local storage capacity in the San Francisco Bay area. Transmission pipelines conveying water from the southern reservoirs were built mainly of wrought iron. Within the city limits there existed approximately 711 km of distribution piping at the time of the earthquake, of which roughly 18.5 and 66.5 km were wrought- and cast-iron trunk lines, respectively. These lines were larger than or equal to 400 mm in diameter. The bulk of the system had been constructed during the years 1870 to 1906. Superimposed on Figure 29.6 are the approximate locations of transmission pipeline damage caused by the earthquake. Flow from all transmission pipelines stopped shortly after the earthquake. Because telephone service was out, emergency control information had to be obtained by dispatching personnel
Sa
Pilarcitos Pipeline
nA nd s rea Fa ult
City Line San Andreas Pipeline University Mound Reservoir
Crystal Springs Pipeline
Pilarcitos Reservoir Crystal Spring Reservoir
29-10
FIGURE 29.7 Failure of Pilarcitos 30-inch pipeline. (From USGS. 1907. The San Francisco Earthquake and Fire of April 18, 1906 and Their Effects on Structures and Structural Materials, Bull. No. 324, U.S. Geological Survey, Washington, D.C.)
into the eld, where maintenance crews reported on the damage. Right lateral strike-slip movement along the San Andreas Fault ruptured a 750-mm-diameter wrought-iron pipeline conveying water from the Pilarcitos to Lake Honda Reservoir (see Figure 29.7). More than 29 breaks were reported north of the San Andreas Reservoir, where the pipeline was constructed parallel to the San Andreas Fault. Fault movement near the San Andreas Reservoir was measured as 3.6 to 5.6 m. Pipeline ruptures were caused by tensile and compressive deformation of the line. More than 3 months were required to reconstruct the pipeline. Within 16 h after the earthquake, repairs were made to that part of the Pilarcitos conduit that was located within the city limits. Water then was pumped from Lake Merced through the Pilarcitos line into Lake Honda at a rate of approximately 25 million liters per day. Dynamic distortion of bridges was responsible for rupturing a 925-mm-diameter wrought-iron pipeline conveying water from the San Andreas Reservoir to College Hill Reservoir, and for rupturing, at three swamp crossings, an 11029-mmdiameter wrought-iron pipeline conveying water from Crystal Springs to the University Mound Reservoir. The wooden trestle bridges all were damaged by strong ground shaking, with no damage or misalignment observed in their timber pile foundations. Approximately 3 days were required to repair the 925-mmdiameter pipe, and over a month was required to restore the 1100-mm-diameter Crystal Springs Pipeline. Figure 29.8A is a map of the 1904 water supply within the San Francisco City limits. There were nine reservoirs and storage tanks, with a total capacity of 354 million liters. Approximately 92% of this total, or 325 million liters, was contained in the Lake Honda, College Hill, and University Mound Reservoirs. These and the pipelines linking them with various parts of the city were the backbone of re protection. All trunk lines 400 mm or larger in diameter are plotted in Figure 29.8. Trunk lines are shown connected to the Lake Honda, College Hill, University Mound, Francisco Street, and Clay Street Reservoirs; all others were connected to piping 300 mm or less in diameter. Superimposed on the gure are the zones of lateral spreading caused by soil liquefaction, as delineated by Youd and Hoose [1978]. Breaks in the pipeline trunk system crossing these zones are plotted from records provided by Schussler [1906] and Manson [1908]. It can be seen that multiple ruptures of the pipeline trunk systems from the College Hill and University Mound Reservoirs occurred in the zones of large ground deformation, thereby cutting off supply of more than 56% of the total stored water to the Mission and downtown districts of San Francisco. Two pipelines, 400 and 500 mm in diameter, were broken by liquefaction-induced lateral spreading and settlement across Valencia Street north of the College Hill Reservoir. These broken pipes emptied the reservoir of 53 million liters, thereby depriving reghters of water for the burning Mission District of San Francisco. Figure 29.8B shows a map of the San Francisco water supply and area burned during the re. All trunk lines of the College Hill and University Mound Reservoirs downstream of the pipeline ruptures are removed from this gure to show the impact and lack of hydraulic conductivity caused by severing these conduits. With the College Hill and University Mound Reservoirs cut off, only the Clay Street tank and
29-11
the Lombard and Francisco Street Reservoirs were within the zone of most intense re, and therefore capable of providing water directly to ght the blaze. The combined capacity of these reservoirs was only 21 million liters, or 6% of the system capacity. The usefulness of such limited supply was further diminished by breaks in service connections caused by widespread subsidence and burning and collapsing buildings. Schussler [1906] identies service line breaks as a major source of lost pressure and water. There were roughly 23,200 breaks in service lines, between 15 and 100 mm in diameter. Fallen rubble and collapsed structures often prevented remen from closing valves on distribution mains to diminish water and pressure losses in areas of broken mains and services. As is evident in Figure 29.8B, the Lake Honda Reservoir was able to provide a continuous supply of water to the western portion of the city. The re eventually was stopped along a line roughly parallel to Van Ness Avenue, where water still was available from the Lake Honda Reservoir. Moreover, the southern and southeastern extent of the re was bounded by areas south and southeast of the trunk system ruptures. It is likely that these unburned areas had water from the University Mound Reservoir. Recognition of the critical role played by damage to the water system led to the construction of San Franciscos Auxiliary Water Supply System, which is described later in this chapter.
29.2.2 The 1923 Tokyo Earthquake and Fires

This section summarizes the res that followed the 1923 Kanto earthquake, affecting Tokyo and nearby cities. The 1923 earthquake was followed by the largest urban conagration in history, in which approximately 140,000 persons perished. The earthquake and conagration are rst briey summarized, followed by a discussion of the effects on water supply components and a qualitative analysis of the contribution of the water supply damage to the occurrence of the conagration. 29.2.2.1 Seismological and Overall Damage Aspects The M 7.9 Kanto earthquake [Usami, 1987] occurred at 11:58 a.m. local time September 1, 1923, with an epicenter located just offshore in Sagami Bay, at 139.5 E, 35.1 N. Damage was extensive, with major crustal movements (maximum 2 m uplift), signicant numbers of engineered buildings sustaining structural damage [Freeman, 1932], destroying approximately 128,000 houses and damaging another 126,000 [Kanai, 1983], with extensive liquefaction [Hamada et al., 1992] (Figure 29.9) and landsliding [ASCE, 1929]. Shaking intensity has been estimated at JMA 6 (equivalent to MMI IX; Figure 29.10) [Kanai, 1983]. A major tsunami (4 to 6 m in height) affected the Miura and Boso peninsulas, destroying 868 houses [Kanai, 1983; Hamada et al., 1992]. Chapter 1 has additional details on this event. 29.2.2.2 Ignitions and Fire Because of a dense urban aggregation of wooden buildings, Tokyo had long been recognized as a major conagration hazard. Following a recent dry period and nearby typhoon, meteorological conditions were particularly adverse at the time of the earthquake, with hot (approximately 26C, 80F) dry winds of approximately 12.5 m/sec (28 mph) at the time of the earthquake. Winds grew continuously all day, reaching a maximum of 21 m/sec (48 mph) at 11:00 p.m. that evening. The re occurred just prior to lunchtime, when numerous small charcoal braziers were lighted for the noontime meal, resulting in approximately 277 outbreaks of re, about 133 of which spread [Okamoto, 1984] (Figure 29.11). The result was a major conagration with rapid respread (Figure 29.12), which burned for several days, causing approximately 140,000 deaths and destroying approximately 447,000 houses [Kanai, 1983; Hamada et al., 1992]. 29.2.2.3 Damage to Water Supply Components Water supply systems for the cities of Tokyo, Yokohama, Kawasaki, and Yokosuka were all generally similar in conguration, drawing their supply from rivers emanating from the mountains surrounding the Kanto Plain, conveying the water by gravity to terminal stilling basins via concrete or cast-iron (CI) aqueducts, and thence via distribution systems composed mainly of CI pipe. The performance of each component
29-12
Legend: Transmission Pipelines Trunk Pipelines ( 400 mm dia.) Zone of LiquefactionInduced Deformation Pipeline Break
Golden Gate Presidio Military Reservation Presidio Heights Reservoir
Francisco St. Reservoir Lombard Reservoir Clay St. Tank
k ar
et
St
Bay of Sa n Fra
8 Breaks
Golden Gate Park 13 Breaks N Lake Honda Reservoir Clarendon Heights Reservoir 9 Breaks Potrero Heights Reservoir
ncisc o
Scale:
Conduit
Valencia St.
du it
Ocean
1000 m
College Hill Reservoir University Mound Reservoir

on rea sC
Pacific
(A)
FIGURE 29.8 (A) Map of the 1904 San Francisco water system, with ground failures superimposed; (B) same map, showing nal burned area. (From Scawthorn, C. and ORourke, T.D. 1989. Effects of Ground Failure on Water Supply and Fire Following Earthquake: The 1906 San Francisco Earthquake, in Proc. 2nd U.S.Japan Workshop on Large Ground Deformation, July, National Center for Earthquake Engineering Research, Buffalo, NY.)
is detailed in an unpublished report by ASCE [1929]. Particularly relevant was the performance of the distribution system within the urbanized areas. Tokyos distribution system totaled 723 mi in length and included 255,000 linear feet of 16- to 60-in.diameter CI trunk line pipe, and secondary mains from 4 to 14 in., totaling about 700,000 lineal feet. No specic information is available regarding damage to the Tokyo distribution system, except that CI pipe throughout Tokyo suffered in places, especially in lled ground and along river banks. Yokohamas distribution system totaled 170 mi of CI pipe, ranging in size from 4 to 36 in. All pipe trenches were excavated to recaulk joints in general, smaller pipe was more damaged than larger, and ttings, tees, and elbows were the worst damaged. Wherever the pipe crossed bridges, pipes were broken. There were 83,600 services to houses 80% of these houses were burned and their services were destroyed, while services to the 20% unburned were undamaged. Yokosukas distribution system comprised 117,000 linear feet (lf), about 50% of which was on reclaimed land. About 50% of the joints required recaulking damage was greatest to smaller-diameter pipe and to ttings, etc., repeating Yokohamas experience, as did the experience with burned houses also having services destroyed. In summary, distribution systems were constructed of CI and sustained substantial damage, mostly in smaller diameters, and at tees, elbows, and other ttings and mostly in softer ground.
tal Crys gs rin Sp duit Con
Sa
nA
Lake Merced
Pilorc
itos
nd
29-13
Legend: Transmission Pipelines Trunk Pipelines (400 mm dia.) Zone of LiquefactionInduced Deformation Pipeline Break
Golden Gate Presidio Military Reservation Presidio Heights Reservoir
Francisco St. Reservoir Lombard Reservoir Clay St. Tank Burnt Sections of San Francisco
of Bay
Golden Gate Park N Lake Honda Reservoir Scale: Clarendon Heights Reservoir Potrero Heights Reservoir
San Fran cisco
Conduit
Ocean
1000 m
College Hill Reservoir

on du it
itos
Pacific
Pilorc
(B)
FIGURE 29.8 (CONTINUED)
A major impact was the additional damage to the distribution system as the re grew in size, due to damage to building services as the buildings burned. 29.2.2.4 Impacts of Water Supply Damage on Fireghting Analyses are not available as to whether Tokyo and Yokohama could have defended themselves against conagration if meteorological conditions had been more favorable. In the actual event, conditions were extremely unfavorable perhaps several hundred ignitions occurred almost immediately because of the lunch-hour timing of the earthquake, when thousands of small grills were being employed. The ignitions were fanned by high winds and grew rapidly in the densely built-up neighborhoods of almost exclusively wooden buildings, which had been made more ammable by a recent dry period. Tokyo and Yokohama, particularly the areas most heavily burned, are low-lying. Damage to the distribution systems in these areas was heaviest, so that hydrants were probably often dry. Another important factor was the impact of the re on both demand and capacity. That is, as the re grew, it impacted the water system in two mutually exacerbating ways: 1. Demand In general, re growth is exponential, for a plentiful fuel supply. While the perimeter that must be defended will grow as the square root of the area of the re, in general, the net result is that the reghting water demand increases exponentially over time. 2. Capacity As the re grows in area, relatively more of the distribution system is available for supply, so that it would be expected that water supply would increase. However, buildings within the re are collapsing, breaking their services. In general, the net result is that the capacity of the distribution system is actually decreasing as the re grows, due to the increased drain placed on the system by hundreds or thousands of broken services.
tal Crys gs n Spri duit Con
Sa
nA
Lake Merced
nd
rea
University Mound Reservoir
sC
29-14
EXPLANATION Soil Liquefaction 19 Investigated Sites 7 6 Kasukabe Urawa 4

TOKYO
3
TOKYO BAY
Kofu Kawasaki Yokohama Chigasaki 1
8 Chiba
Numazu
SAGAMI BAY Epicenter So 9 u (af rce TATEYAMA ter Ar An ea do ) 0 50 km
SURUGA BAY
FIGURE 29.9 Map of Kanto and Chubu Districts showing areas principally affected by liquefaction during the 1923 Kanto earthquake. (From Hamada, M. et al. 1992. Liquefaction-Induced Ground Deformations during the 1923 Kanto Earthquake, in Case Studies of Liquefaction and Lifeline Performance during Past Earthquakes, Vol. 1, Japanese Case Studies, NCEER Technical Report 920001, M. Hamada and T.D. ORourke, Eds., National Center for Earthquake Engineering Research, Buffalo, NY.)
This situation is typical of urban conagrations and undoubtedly existed in Tokyo, irrespective of the initial seismic damage to the distribution system. On the other hand, the portions of Tokyo and Yokohama most heavily burned are low-lying, with numerous canals and access to waterways. Therefore, secondary emergency water supply should have been available to the reghters. However, it is likely the equipment of the time was limited in capacity and could not furnish the volume of water required to contain the large res that quickly developed. Therefore, the primary factor leading to the conagrations in Tokyo and Yokohama was not the seismic damage to the water supply system, but rather the rapid growth of numerous simultaneous ignitions under a situation of adverse meteorological and dense wooden-building conditions. These conditions would have overwhelmed the re service and an undamaged distribution system, even had there been no earthquake. As in San Francisco in 1906, the city was waiting to burn. 29.2.2.5 Summary The 1923 Kanto earthquake resulted in strong shaking and widespread permanent ground deformations. In general, however, this did not severely damage the main water supplies to Tokyo and neighboring cities. The distribution systems, on the other hand, sustained numerous breaks, primarily in smaller pipes in low-lying ground. Notwithstanding the failure of the distribution system, however, the conagrations that developed, which were the main agent of damage in this catastrophe, were primarily due to non-seismic factors. These included a conagration-prone built-up environment and an extremely adverse ignition scenario and meteorological conditions. Under these circumstances, it is likely the re service would have proved inadequate even had there been no earthquake.
29-15
FIGURE 29.10 Distribution of earthquake intensity in Tokyo. (From Scawthorn, C. 1997a. The 1923 Kanto, Japan, Earthquake, in Reliability and Restoration of Water Supply Systems for Fire Suppression and Drinking Following Earthquakes, Report NIST GCR 97730, D.B. Ballantyne and C.B. Crouse, Eds., National Institute of Standards and Technology, Gaithersburg, MD.)
29.2.3 The 1989 Loma Prieta, California Earthquake and Fires

This section summarizes aspects of a signicant re following the 1989 Loma Prieta earthquake. The earthquake and damage are rst briey summarized, followed by a discussion of the effects on water supply components and an analysis of the interaction of water supply damage to the occurrence of res.
29-16
Sumida River
Ara River : Extinguished (144) : Spread (133)
Tokyo Bay
i
FIGURE 29.11 Outbreaks of res in the city of Tokyo (Kanto earthquake of 1923). (From Okamoto, S. 1984. Introduction to Earthquake Engineering, University of Tokyo Press, Tokyo. With permission.)
FIGURE 29.12 Firespread in Central Tokyo, showing direction and hourly progress of ame front, 1923 Kanto earthquake. (From Japanese government report.)
29-17
29.2.3.1 Seismological and Overall Damage Aspects On October 17, 1989 at 5:04 p.m. local time, an M 7.1 earthquake occurred due to an approximately 40-km rupture along the San Andreas Fault. The epicenter of the 20-sec earthquake was located near Loma Prieta in the Santa Cruz mountains about 16 km northeast of Santa Cruz, 30 km south of San Jose, and about 100 km south of San Francisco. Major damage included the collapse of the elevated Cypress Street section of Interstate 880 in Oakland, the collapse of a section of the San FranciscoOakland Bay Bridge, multiple building collapses in San Franciscos Marina District, and the collapse of several structures in Santa Cruz and other areas in the epicentral region. Damage and business-interruption losses were estimated as high as $6 billion. Human losses were 62 people dead, 3700 reported injured, and more than 12,000 displaced. At least 18,000 homes were damaged, 960 were destroyed, and more than 2500 other buildings damaged and 145 destroyed. See Chapter 1 in this volume for more details. 29.2.3.2 Water Supply and Fire Protection San Francisco possesses three water supply systems: The Municipal Water Supply System (MWSS), which is typical of water systems in any other city, is owned and operated by the San Francisco Water Department (SFWD), and serves both reghting and municipal (potable water) uses. The Auxiliary Water Supply System (AWSS) was rst developed following the 1906 earthquake and re and extended periodically thereafter (Figure 29.13). The Portable Water Supply System (PWSS), developed in the 1980s, is primarily a truck-borne large-diameter hose system (Figure 29.14). The latter two systems are specically dedicated to reghting use and are owned and operated by the San Francisco Fire Department (SFFD). They were built following, and in direct response to, the 1906 earthquake and re (see above). The greatest damage to the MWSS in the Loma Prieta incident consisted of approximately 150 main breaks and service line leaks. Of the 102 main breaks, over 90% were in the Marina, Islais Creek, and South of Market inrm areas. The signicant loss of service occurred in the Marina area, where 67 main breaks and numerous service line leaks caused loss of pressure (Figure 29.15). The AWSS consists of several major components: Static supplies. The main source of water under ordinary conditions is a 10 million-gallon reservoir centrally located on Twin Peaks, the highest point within San Francisco (approximately 750 ft elevation). Pump stations. Because the Twin Peaks supply may not be adequate under emergency conditions, two pump stations exist to supply water from San Francisco Bay each has 10,000 gpm at 300 psi capacity. Both pumps were originally steam powered but were converted to diesel power in the 1970s. Pipe network. The AWSS supplies water to dedicated street hydrants by a special pipe network with a total length of approximately 120 mi. The pipe is bell-and-spigot, originally extra heavy CI (e.g., 1 in. wall thickness for 12 in. diameter), and extensions are now Schedule 56 ductile iron (e.g., 0.625 in. wall thickness for 12 in. diameter). Restraining rods connect pipe lengths across joints at all turns, tee joints, hills, and other points of likely stress. Fireboat Phoenix.2 The pipe network has manifold connections located at several points along the citys waterfront to permit the reboat Phoenix to act as an additional pump station, drafting
Within days following the 1989 Loma Prieta earthquake, the San Francisco Fire Department purchased a reboat that Vancouver, B.C. had just discarded. Renamed the Guardian, the reboat is arguably the largest in North America, with 20,000 Igpm pumping capacity. The Phoenix and the Guardian are both active as of this writing, with each alternately in service for one to several months, and the other in reserve. Both are stationed near the foot of Folsom Street, close to the San Francisco-Oakland Bay Bridge.
2
29-18
from San Francisco Bay and supplying the AWSS. The Phoenixs pump capacity is 9600 gpm at 150 psi, about the same as Pump Station No. 2. Cisterns. Finally, in addition to the above components, San Francisco has 151 underground cisterns, again largely in the northeast quadrant of the city. These cisterns are typically constructed of concrete with a 75,000-gallon capacity (about 1-hour supply for a typical re department pumper).
(A)
Twin Peaks Reservoir

(40,000 tons (10 mill. Gallons)
240m
Asbury Tank Jones St Tank
Upper Zone
Pump Station
Lower Zone
(B)
FIGURE 29.13 San Francisco AWSS: (A) plan; (B) elevation (schematic).
29-19
Portable Hydrant Cleeson Valve Gated Wye Cleeson Valve
S. F. Hay or Lake
Portable Hydrant
Gated Wye
Portable Hydrant
Cleeson Valve
Cleeson Valve Cistern Cistern Portable Hydrant
FIGURE 29.14 San Francisco PWSS. (Source: C. Scawthorn)
1906 Waterfront
150 200 150 200 100 150 200 200 300
Marina
Blvd
15 0
Earthen Mole of San Francisco Gas Light Co. (1899)
100 150 200 150 Buchanon St. Shoreline
150 Beach 200 150 St.
150 200
200 300 150 300 150
200
150
100 100
100 150 300 Boy St. 1857 Chestnut St. Fillmore St. N 0 0
Baker St.
150
EXPLANATION Service repair Main repair Repair at or near gate valve Pipeline, with diameter (in millimeters)
Orvisodero St.
Scoll St.
200m 600 f1
FIGURE 29.15 MWSS Pipe Breaks, Marina District, 1989 Loma Prieta earthquake. (From ORourke, T.D. et al. 1992. Lifeline Performance and Ground Deformation during the Earthquake, in The Loma Prieta California Earthquake of October 17, 1989: Marina District, U.S. Geological Survey Professional Paper 1151F, T.D. ORourke, Ed., Strong Ground Motion and Ground Failure, T.L. Holzer, Coordinator, U.S. Government Printing Ofce, Washington, D.C.)
29-20
The AWSS is a system remarkably well designed to furnish large amounts of water for reghting purposes under normal conditions and contains many special features to increase reliability in the event of an earthquake. The Loma Prieta earthquake resulted in only moderate shaking for most of San Francisco, typically of MMI VI, although selected areas sustained much greater shaking, perhaps as much as MMI IX in the Marina District, where 69 breaks in the domestic water supply and more than 50 service connections to water mains quickly dissipated all domestic water supply in the 40 blocks of the district. The AWSS main serving the Marina District remained intact. However, as a result of the shaking, in locations other than the Marina, the AWSS sustained signicant damage. The most signicant damage to the AWSS occurred on Seventh Street between Howard and Mission Streets, where a 12-in. main broke. This location is on the boundary of Inrm Area No. 3 and, moreover, the AWSS pipe at this location crosses over a sewer line. Soil settlements in this area are thought to have occurred prior to and also as a result of the earthquake, causing the AWSS pipe to settle onto the sewer line and break. Other breaks included (1) a break in an 8-in. hydrant branch on Sixth Street between Folsom and Howard Streets (where the hydrant branch crossed up and over a sewer line) and (2) ve 8-in. elbow breaks, four within Inrm Area No. 3, including one on Bluxome Street where a portion of a building collapsed onto an AWSS hydrant. These breaks resulted in such major leakage that the Jones Street tank (controlling pressure for the Lower Zone) had completely drained in approximately 15 min. Leakage continued so that the rst engines to arrive at the Marina re found only residual water when they connected to AWSS hydrants. Due to uncertainty as to the number and location of AWSS breaks, valves connecting the Upper Zone to the Lower Zone were not opened, and Pump Stations 1 and 2, although available, were not placed in service immediately but only at 8:00 p.m., following identication and isolation of broken mains. As a result, all pressure in the AWSS Lower Zone was lost for several hours following the earthquake. The pump stations were operated at half capacity so as to ll the AWSS mains slowly, out of concern for entrapped air that was exhausted out of the Lower Zone through the Jones Street tank (air could be heard exhausting from the tank). This operation continued until 10:00 p.m., when full pressure was restored and the Jones Street tank had been lled with salt water. Other damage was conned to: One 75,000-gallon cistern at Fifth and Harrison Streets, which developed a leak at the cold joint between the roof and sidewall due to earthquake damage and lost 20% of its water, leaving just 60,000 gallons for re suppression purposes. One high-pressure hydrant destroyed by falling structures and another damaged. 29.2.3.3 Water System Performance and Fireghting Twenty-six res occurred in San Francisco as a result of the earthquake, 11 of them on October 17 [SFFD, 1990]. One of these res, which occurred in the Marina District, threatened to become a major conagration. Fireghting efforts were severely hampered by the lack of MWSS and AWSS service to hydrants caused by the severe liquefaction and resulting pipe breakage in the Marina and elsewhere. Fireghters were forced to resort to drafting from nearby lagoons, which, however, was inadequate, and the re continued to grow. Deployment of San Franciscos PWSS, in conjunction with the reboat Phoenix, provided the only adequate source of reghting water, and this was the only way the Marina re was extinguished. A re began in the four-story wood-frame building at 3701 Divisadero Street at the northwest corner of Beach Street. The building is a typical corner building in the Marina District. Built in the 1920s, it contained 21 apartments, with the ground oor being primarily a parking garage. The buildings lower two oors had collapsed in the earthquake, and the third and fourth oors were leaning southward several feet. The re, initially quite small, was in the rear of the building. This, combined with the confusion following the earthquake, resulted in a delayed report, meaning that the rst San Francisco Fire Department (SFFD) unit did not arrive until approximately 5:45 p.m. (all times estimated). The
29-21
source of ignition has not been denitely determined as of this writing. Wind speed was virtually zero. Arrival of SFFD Trucks 10 and 16 was closely followed by Engine (E) 41. Based on the appearance of black smoke, the re appeared to the ofcer in charge of E41 (Lt. P. Cornyn) to be a wood-structurefueled re. E41 connected to the AWSS hydrant directly in front of the building (which actually was leaning over the hydrant) and charged the pump, but found no water pressure. Because of radiant heat, E41 then withdrew across the street. At about 6:25 p.m., the building at the northeast corner (2080 Beach) ignited. At about 6:30 p.m. 3701 Divisadero exploded (the fuel source was most likely leaking gas), ames shot 100 ft or more into the air, buildings across both Divisadero and Beach Streets were either burning or smoldering, and the re had spread to the north and west neighboring buildings. At about this time, E41 and E10 were protecting the buildings across Divisadero and Beach Streets. E41 was being supplied by E16, which was relaying from a hydrant at Scott and Beach Streets, while E10 was being supplied by E21, which was drafting from a lagoon of the Palace of Fine Arts (located two blocks to the west) and relaying the water. By 6:50 p.m., the re was burning northward; E3, E36, E22, E31, E14, E25, HT25, and an additional re reserve engine also were on the scene. E22 attempted to draft from the Marina lagoon, but, due to low tide, was unsuccessful. Because the re was located only two blocks from the Marina, the reboat Phoenix was called for, arriving at about 6:30 p.m. At approximately the same time, PWSS hose tenders arrived at the scene and were able to connect to the Phoenix. In all, three PWSS hose tenders responded to the Marina re (a fourth was in the shop being outtted for service). Four major runs of hose (or portable water mains) were laid at the Marina re, with some 6000 ft of 5-in. hose being deployed, using nine portable hydrants. The Phoenix pumped 6000 gpm at 180 psi for over 18 h. Fire spread was stopped at about 7:45 p.m. by master streams from the monitors on the hose tenders, as well as ladder pipes and hand lines. Figure 29.16 indicates the deployment of various SFFD apparatus, including the reboat Phoenix, at about 6:24 p.m. 29.2.3.4 Summary The 1989 Loma Prieta earthquake provided a number of valuable observations and lessons, including: The Marina re was potentially very severe it was a very large re in a dense neighborhood of wood-frame construction an unusually calm wind was a very fortuitous circumstance. The re was within 500 ft of San Francisco Bay and the Pacic Ocean the largest body of water on Earth. However, these could not be drafted from by arriving re engines, and the water was inaccessible. The MWSS system had over 400 million gallons of storage within San Francisco, but the numerous breaks in the Marina prevented adequate pressure or volume at Marina hydrants; elsewhere in the city, MWSS performance was generally satisfactory. The AWSS is designed for earthquake ground motions, and did not sustain damage in the Marina despite widespread liquefaction. Nevertheless, breaks several miles away caused a loss of pressure in the Lower Zone. The backup to the backup, that is, the PWSS backing up the AWSS, which backs up the MWSS, provided reghting water for extinguishment at the Marina re. The PWSS exibility and portability proved adequate to the task.
29.2.4
The 1994 Northridge, California Earthquake and Fires
The Northridge earthquake was the largest earthquake to have occurred within a U.S. city in more than 20 years. The 4:31 a.m. January 17, 1994 MW 6.7 earthquake was centered under the Northridge section of the San Fernando Valley area of the Los Angeles region. It resulted in shaking intensities greater than MMI VIII over approximately 700 mi2 of the northern Los Angeles area. The population most heavily affected was in the San Fernando Valley, which is primarily protected by the Los Angeles City Fire Department (LAFD). Table 29.3 lists re departments signicantly affected by the earthquake and their summary statistics (see Scawthorn et al., 1997, for additional detail).
29-22
FIGURE 29.16 1989 Loma Prieta earthquake, Marina re, SFFD deployment. (Source: C. Scawthorn)
As Sepponen [1997] reports: At the time of the earthquake there were 788 LAFD personnel on duty. During the quake reghters were injured, and several re stations suffered major damage. Most of the city was without electrical power. Structural damage was noted in 35 station buildings. Although most were repaired quickly, Station 90 was closed for one month for roof repair, Station 70 was out of service for six months, and Station 78 was condemned and demolished. (Ward, 1995) Once equipment and personnel were safely outside of buildings, reconnaissance patrols were sent out as designated in the Earthquake Emergency Operational Plan. The timing of the earthquake at 4:31 a.m., was advantageous for the LAFD because their 24-hour shift change occurs at 6:30 a.m. each day. So, soon after the earthquake occurred the new shift came to work and the other old shift continued to work due to the emergency, doubling the number of personnel available. This enabled reserve equipment to be easily put into service.
29-23
TABLE 29.3 Fire Departments Affected by the January 17, 1994 Northridge Earthquake
Estimated Population (thousands) 3,400 2,896 700 97 94 132 166 25 34 41 12 Area (sq. miles) 469 2,234 126 8 17 23 30 3 6 5 2 Number of Stations 104 127 30 4 6 8 9 1 3 3 1 Fire Fighting Personnel 2,865 1,842 327 100 120 150 167 27 81 66 9 Number of Engines 104 144 40+/ 5 6 8 9 2 7 5 1
Fire Department Los Angeles City Los Angeles County Ventura County Santa Monica Burbank Pasadena Glendale South Pasadena Beverly Hills Culver City Fillmore
Source: Scawthorn, C., Cowell, A.D., and Borden, F. 1997. Fire-Related Aspects of the Northridge Earthquake, Report prepared for Building and Fire Research Laboratory, NIST-GCR-98743, National Institute of Standards and Technology, Gaithersburg, MD.
TABLE 29.4 Fire Following the January 17, 1994 Northridge Earthquake
Community Los Angeles City Los Angeles County Ventura County Santa Monica Burbank Pasadena Glendale South Pasadena Beverly Hills Culver City Fillmore Total Number of Earthquake-Related Fires 77 15 10 4 0 1 0 0 1 0 2 ~110
Source: Scawthorn, C., Cowell, A.D., and Borden, F. 1997. Fire-Related Aspects of the Northridge Earthquake, Report prepared for Building and Fire Research Laboratory, NIST-GCR-98743, National Institute of Standards and Technology, Gaithersburg, MD.
Approximately 110 res were reported as earthquake-related on January 17, as shown in Table 29.4. The Northridge earthquake reportedly caused or was a contributing factor in 77 res in the LAFD service area. These were among a total of 161 res that occurred on the day of the earthquake. The time line in Figure 29.17 shows all calls for assistance with res on the day of the earthquake. Among the earthquakerelated res, structure res predominate (86%). More than 70% (66) of the earthquake-related res occurred in single- or multiple-family residences, as might be expected from the building stock that is typical in the San Fernando Valley. The major cause of ignition was electric arcing as the result of a short circuit, although gas ame from an appliance was also a recurring source of ignition. The breakdown of the days calls into dispatch categories is shown in Figure 29.18. The Northridge earthquake affected the water supply for portions of the San Fernando Valley [Heubach, 1997]. Breaks occurred in at least six trunk lines and a large number of leaks occurred at other locations. The Department of Water and Power estimated that approximately 3000 leaks were caused by the earthquake, including two lines of the Los Angeles Aqueduct. The damage to the system
29-24
16 14 12 Frequency 10 8 6 4 2 0 430 530 630 730 830 930 1030 1230 1330 1430 1530 1630 1730 1830 1930 2030 2130 2230 2330 1130 EQ Fires Other Fires
Hours
FIGURE 29.17 LAFD res, 4:31 to 24:00 hours, January 17, 1994. (From Scawthorn, C. et al. 1997. Fire-Related Aspects of the Northridge Earthquake, Report prepared for Building and Fire Research Laboratory, NIST-GCR98743, National Institute of Standards and Technology, Gaithersburg, MD.)
Other 17%
EMS/Rescue 19%
Earthquake 1% Service Calls 7%
Good Intent Calls 10%
False Calls/Alarms 19%
Fires 12%
Hazardous Conditions 15%
FIGURE 29.18 LAFD incident response types, 1308 incidents, 4:31 to 24:00 hours, January 17, 1994. (From Scawthorn, C. et al. 1997. Fire-Related Aspects of the Northridge Earthquake, Report prepared for Building and Fire Research Laboratory, NIST-GCR-98743, National Institute of Standards and Technology, Gaithersburg, MD.)
resulted in the water pressure dropping to zero in some areas. On January 22, 5 days after the earthquake, between 40,000 and 60,000 customers were still without public water service, and another 40,000 were experiencing intermittent service. Scawthorn et al. [1997] have documented a number of specic res and re department operations, as well as all ignitions, in this event. We provide a synopsis of one event here. 29.2.4.1 North Balboa Boulevard Fire 29.2.4.1.1 Site Description This re scene is located in the Granada Hills area of the San Fernando Valley. It is a residential area with one- and two-story wood-frame single-family dwellings, many with swimming pools (see Figure 29.19A). A 56-in. water main under the street was broken, ooding the street and front yards of the homes (see Figure 29.19B).
29-25
(A)
(B)
FIGURE 29.19 North Balboa Boulevard incident, 1994 Northridge earthquake: (A) aerial view; (B) surface view. (Authors collection) Shown as Color Figure 29.19.
29.2.4.1.2 Location, Ignition, and Cause About 20 min after the earthquake struck, a broken 20-in. gas main under Balboa Boulevard was ignited by the driver of a nearby stalled pickup truck who was attempting to start his vehicle. Electric arcing in the ignition system ignited a large gas cloud, creating a reball and setting re to two dwellings on the east side of Balboa and three on the west side. Radiant heat from the gas re was a major factor in the spread of the re. Wind was 15 to 20 mph from the northeast. Five homes were destroyed, with minor damage to four others. 29.2.4.1.3 Fire Department Operations Fireghters from E8 and E18 were out on district survey, saw the re, and responded. E74 responded to a radio request for assistance from E8. Fireghters from E8 arrived rst and found Balboa Boulevard impassable because of the water owing from the broken water main. Captain Rust took E8 around the streets parallel to Balboa (Paso Robles and McLennan Avenues) and cross streets Lorilland and Halsey to check the re hydrants for water. They were dry. E8 reghters entered the alley west of Balboa to protect the structures on that side of the re. They located a swimming pool behind a home on Paso Robles and used it as a water source. Water from this swimming pool was also supplied to E18 at the south end of the alley. E18 reghters entered the alley west of Balboa and set up to protect the homes at the south end. The heat from the res was intense and forced reghters to operate from protected areas. A Los Angeles County brush re hand crew arrived on the scene as a mutual aid resource and was directed to cut and remove combustible shrubs, trees, fences, etc. around homes exposed to the re. Fireghters from E74 arrived on the scene, checked the hydrants on the north side of the re, found them dry, and entered the alley east of the re. The alley was made impassable by debris from collapsed block walls. Resident volunteers removed the debris, and E74 reghters proceeded south to use a swimming pool for a water source. They extinguished a re in the attic of an exposed one-story dwelling and continued to direct water streams onto the exterior of this building. A group of local citizen volunteers formed a bucket brigade on the northeast side of the re, again using a swimming pool for a water source. They protected the house exposed to the re at that location. Engine companies 8, 18, and 74 pumped water between 1 1/2 and 2 hours during the reghting operation. It took about 2 hours for the natural gas leak re to be reduced in size such that it presented a minimal threat from radiated heat. The incident commander at the scene, Captain Rust, directed operations on the west side between his company and E18, and coordinated efforts with E74 on the east side. Heavy radio trafc use made radio communications very difcult. An aerial photograph of the scene after the structure res were extinguished is shown in Figure 29.19A. Note that the ruptured gas main is still burning.
29-26
11/2 Siphon
Ejector
E18 1
11/2 Suppy 11/2
E8
Alley
Natural Gas Leak Fire Balboa Blvd Broken Water Main
11/2 Siphon
Ejector
1 E74 Volunteers with Buckets

16 Wi -20 nd M PH
Alley
FIGURE 29.20 LAFD Deployment 5:35 a.m., North Balboa Boulevard, 1994 Northridge earthquake. (From Scawthorn, C. et al. 1997. Fire-Related Aspects of the Northridge Earthquake, Report prepared for Building and Fire Research Laboratory, NIST-GCR-98743, National Institute of Standards and Technology, Gaithersburg, MD.)
29.2.4.1.4 Water-Related Aspects Breaks in the water mains rendered all surrounding re hydrants inoperative. Fortunately, several homes in the area had swimming pools that were used as water supply sources, providing approximately 70 min of water ow. E8 and E74 reghters used their 1 1/2-in. siphon ejectors, which can supply water at 92 to 115 gpm, to draw water into their tanks. Hose layout as of 5:35 a.m. and water usage for the incident are shown in Figure 29.20 and Table 29.5, respectively.
29.2.5 The 1995 Hanshin (Kobe), Japan Earthquake and Fires

The 5:46 a.m. January 17, 1995 MW 6.9 (JMA M7.2) Hanshin (ofcial name: Hyogo-ken Nambu) earthquake was centered under the northern tip of Awaji Island near Kobe, in the Kansai region of Japan. The event resulted in shaking intensities greater than MMI VIII over approximately 400 km2 of the KobeAshiya-Nishinomiya area. Population of the affected area (MMI VIII or greater) is approximately 2 million. The Kobe Fire Department (KFD) is a modern, well-trained re response agency, organized into Prevention, Suppression, and General Affairs sections and a Fire Academy. The city is served by 1298 uniformed personnel. Equipment includes two helicopters, two reboats, and 196 vehicles. Approximately 100 res broke out within minutes, primarily in densely built-up, low-rise areas of the central city, which comprise mixed residential-commercial occupancies, predominantly of wood construction. Within 1 to 2 hours, several large conagrations had developed. A total of 108 res were
29-27
TABLE 29.5 Water Usage in the Balboa Boulevard Fire

Engine 8 One 1 1/2-inch siphon ejector in pool supplying approx. 100 gpm One 1 1/2-inch supply line laid to Engine 18 for their water source
One 1 1/2-inch tip line with spray tip 125 gpm TOTAL: 8,750 gallons Engine 18 One 1 1/2-inch supply line in to ll tank One 1-inch line with spray tip 25 gpm TOTAL: 1,750 gallons Engine 74 One 1 1/2-inch siphon ejector in pool supplying approx. 100 gpm Two 1-inch lines/spray tips 50 gpm TOTAL: 3,500 gallons Total estimated water employed to control or extinguish res: 14,000 gallons Source: Scawthorn, C., Cowell, A.D., and Borden, F. 1997. Fire-Related Aspects of the Northridge Earthquake, Report prepared for Building and Fire Research Laboratory, NIST-GCR98743, National Institute of Standards and Technology, Gaithersburg, MD.
TABLE 29.6 Post-Earthquake Fire Ignitions

Post-Earthquake Fires in Time Order for January 1719 in Kobe City Kobe City by Ward Higashi-nada Nada Chuo Hyogo Nagata Suma Tarumi Kita Nishi Kobe City Total 1/17 ~6:00 10 13 8 11 13 4 0 0 1 60 ~7:00 1 0 4 0 1 4 0 0 0 10 ~8:00 2 1 2 2 0 0 0 0 0 7 ~9:00 1 1 1 2 0 4 0 0 0 9 ~24:00 3 2 2 2 3 1 6 1 0 23 1/17 Total 17 17 17 17 17 13 6 1 1 109 1/18 Total 2 2 2 4 1 2 0 0 0 14 1/19 Total 4 0 0 3 4 1 0 0 0 15 1/1719 Total 23 19 19 24 22 16 6 1 1 138
Post-Earthquake Fires in Time Order for January 1719 in Cities Other Than Kobe Cities Other Than Kobe Ashiya Nishinomiya Takaradzuka Itami Kawanishi Amagasaki Awajicho Osaka Toyonaka Suita Other Cities Total 1/17 ~6:00 4 11 2 2 1 3 1 7 3 1 37 ~7:00 4 11 0 2 1 2 1 4 1 1 26 ~8:00 1 1 0 2 0 1 0 1 0 0 5 ~9:00 0 1 0 1 0 0 0 1 1 0 4 ~24:00 0 10 2 0 0 2 0 2 0 0 16 1/17 Total 9 34 4 7 2 8 2 15 5 2 88 1/18 Total 2 4 6 1/19 Total 2 3 5 1/1719 Total 13 41 4 7 2 8 2 15 5 2 99
Note: " " means no post-earthquake re reported. (Data as of November 1995.) Source: Sekizawa, A. 1997. Post-Earthquake Fires and Fireghting Activities in the Early Stage in the 1995 Great Hanshin Earthquake, NIST Tech. Report NISTIR 6030, Building and Fire Research Laboratory, National Institute of Standards and Technology, Gaithersburg, MD.
29-28
FIGURE 29.21 Aerial view of burned area, 1995 Kobe earthquake. (Authors collection)
reported in Kobe on January 17 [Kobe Fire Department, 1995], the majority being in the wards of Higashi Nada, Nada, Hyogo, Nagata, and Suma. Fire response was hampered by extreme trafc congestion, collapsed houses and buildings, and rubble in the streets, rendering many areas inaccessible to vehicles. Table 29.6 lists time of ignitions for Kobe and other cities and Figure 29.21 is an aerial view of one of the burned areas. Fireghting water is taken primarily from the city water system, served by gravity from 30 reservoirs. Of these, 22 have dual tanks, with one tank having a seismic shutoff valve so that, in the event of an earthquake, one tanks contents are conserved for reghting. In this event, all 22 valves functioned properly, conserving 30,000 m3 of water, which, however, could not be delivered because of approximately 2000 breaks in the underground piping. Kobe has approximately 23,500 re hydrants, typically ushmounted (i.e., under a steel plate in the sidewalk or street) with one 150-mm-diameter hose connection. The city has provided underground storage of water for disaster reghting in 968 cisterns, generally of 40,000-liter capacity, sufcient to supply a pumper for about 10 min. All engines carry hard suction so that additional water can be drafted from Osaka Bay or the several streams running through Kobe. Kobe sustained approximately 2000 breaks in its underground distribution system. Water for reghting purposes was available for 2 to 3 hours, including the use of underground cisterns. Subsequently, water was available only from tanker trucks. KFD attempted to supply water with a reboat and relay system, but this was unsuccessful because of the relatively small hose used by KFD. The author ew over the area at about 5:00 p.m. on January 17 and was able to observe all of the larger res (about eight in all) from an altitude of less than 300 m. No re streams were observed, and all res were burning freely several with ames 6 m or more in height. No re apparatus was observed in the vicinity of the large res, although it could be seen at other locations (their activities were unclear from the air). Some residents formed bucket brigades (with sewer water) to try to control the ames. Selected aspects of the 1994 Northridge and 1995 Hanshin earthquakes are compared in Table 29.7. Several key observations include: Earthquakes in urban areas continue to cause multiple simultaneous ignitions and degrade emergency response due to impaired communications, transportation, and water supply. These events are replicable, as shown by comparison of the 1971 San Fernando and 1994 Northridge events [Scawthorn et al., 1995], and by comparison of the ignition rates and other factors in the Northridge and Hanshin events, providing some validation for simulation modeling and projections for larger events. Under adverse conditions, large conagrations are possible in modern cities, as shown by events in California (i.e., the 1991 East Bay Hills Fire and the 1993 Southern California wildres) and by the Hanshin earthquake in Japan.
29-29
TABLE 29.7 Hanshin and Northridge Earthquakes: Comparative Analysis

Aspect Event Factor Magnitude (MW) Date (winter) Time Population (MMI VIII) Density (pop/sq km) Number (total) Structural res Rate (MMI VII) Ign/pop FD communications Resources (ff/popul) Stations Trafc congestion Mutual aid Water system damage Cisterns Automatic shutoffs Northridge 6.7 Jan 17 0431 1~1.5 million 1,000~1,500 110 86% 14,719 Manual dispatch 1,338 104 Minor Available not needed Some Swimming pools Calm ? few% Minor 6.9 Jan 17 0546 2 million 4,000 108 97% 13,676 1,138 26 (Kobe) Major After 10 hr Total? 946, mostly 40 tons (10 min) Minor 70% ineffective due to structl collapse Major: 5,000 bldgs Hanshin
Region Ignitions
Response
Water Wind Gas Spread
29.2.6 The 1999 Marmara, Turkey Earthquake and Fires

The MW 7.4 Kocaeli (Izmit) earthquake occurred at 3:10 a.m. local time August 17, 1999 on the eastwest-trending northern strand of the North Anatolian Fault Zone (NAFZ), about 100 km southeast of Istanbul [Scawthorn, 2000]. The 125-km-long fault and high damage area follows or is close to the south shore of Izmit Bay and has predominantly 2.2 m right lateral displacement, from Adapazari in the east to Yalova in the west. Signicant vertical fault scarps of as much as 2 m occur at several locations. Peak ground accelerations of approximately 0.4 g were recorded near the fault and liquefaction and subsidence were observed on the shores of Izmit Bay and Lake Sapanca. Several million persons live in the Izmit region, which has experienced rapid growth and heavy industrialization in the last two decades. The predominant building type is mid-rise nonductile reinforced concrete frames with hollow clay tile inll, thousands of which collapsed in a pancake mode. Conrmed dead were approximately 17,000. Estimated population requiring short- to long-term shelter was 200,000. Lifelines generally performed well, with the exception of underground piping in the heavily affected areas, where major damage was reported. Electric power, highways, rail, and telephone were generally functional within several days following the earthquake and the Izmit Water Project (the regional water supply and transmission system) was only lightly damaged and fully functional. Fires occurred in a number of collapsed buildings but were generally conned to building of origin. Two res broke out at the Tupras oil renery, which burned for several days and which are of special interest. The most widely publicized and spectacular damage to any industrial facility, because of the tank farm res that burned out of control for several days [Johnson et al., 2000], occurred at the massive U.S. $3.5 billion renery near Korfez, owned by the state-owned oil company, Tupras. The rst re was initiated in a oating-roof naphtha tank. Naphtha, a highly volatile material with a low ashpoint, is easily ignited. The speculation is that the sloshing of naphtha in the tank caused the oating roof to breach its seal, allowing naphtha to spill. Sparks from the friction between the steel roof and tank wall likely ignited the naphtha. The renery receives its entire water supply through a dedicated pipeline from Lake Sapanca, some 45 km to the east. Due to multiple breaks in the pipeline, the renery quickly lost all water and all reghting capabilities. As the re spread to additional tanks, aircraft attempted to douse the res by dropping foam. After a few days, along with the aerial foam attack, the renery used diesel pumps to draw water directly from Izmit Bay to ght the re. The res were nally declared under control on Sunday, some 5 days after the earthquake.
29-30
(A)
(B)
FIGURE 29.22 Tupras Renery damage due to re following earthquake: (A) burned and collapsed tank; (B) top of stack collapsed into unit, severely damaging heater unit. Top of stack is in foreground of photo. (Source: Gayle Johnson, EQE International.) Part (A) shown as Color Figure 29.22.
While the re was burning out of control, an area within 2 to 3 mi of the renery was evacuated, including some areas where search-and-rescue operations were taking place in collapsed buildings. Train service was disrupted in the area because of the re. The re and heat eventually consumed numerous tanks in the tank farm. It was reported that at least 17 tanks were considered to be total losses. They were generally crumpled by the intense heat, with one tank expanding as if ready to explode. The other area of severe and spectacular damage in the renery occurred in one of their three crude units, when a 90-m-high reinforced concrete heater stack collapsed. The break appeared to occur at about the height of the large-diameter heater duct. Sifting through the rubble of the stack did not immediately reveal the cause of the failure. The top of the stack fell into the unit, destroying the heater, while the bottom portion fell into a pipeway running around the perimeter of the unit. The destroyed pipeway was heavily congested with
29-31
piping from all over the renery. Several months would be necessary to identify, isolate, and repair damaged piping in this area (Figure 29.22). One of the pipes broken by the stack collapse was a naphtha line from the original burning naphtha tank in the tank farm. A re started when the collapse occurred and, although it was extinguished relatively quickly, it ared up several times because of the new fuel from the broken pipe. The supply could not be stopped because the two block valves were at the tank, inaccessible because of the re and downstream from the crude unit.
29.2.7 Summary
The accumulation of experience based on observations of the above events and others, which space does not permit discussing here, leads to the conclusion that the potential exists for large conagrations following a major earthquake in an urban area, particularly in a region with a large wood-building stock. Under adverse meteorological and other conditions, these conagrations could burn for several days, replicating the events of 1906 in San Francisco and 1923 in Tokyo. Extensive well-drilled mutual-aid systems are required to mobilize large resources in response, but the deployment of these resources will be hampered by transportation difculties and, perhaps most tellingly, failure of reghting water supplies. Improvements in planning and infrastructure are absolutely essential to forestall this potential.
29.3 Analysis
This section presents a methodology for estimating re losses following earthquake, developed by the author in the 1970s [Scawthorn, 1981] and extended via research and applications in a number of projects since that time. The basic methodology can be seen in Figure 29.23, the re following earthquake process, which depicts the main aspects of the problem. Fire following earthquake is a process that begins with the occurrence of the earthquake. In summary, the steps in the process are: Occurrence of the earthquake This will presumably cause damage to buildings and contents, even if the damage is as simple as knocking things (such as candles or lamps) over. Ignition Whether a structure has been damaged or not, ignitions will occur due to earthquakes. The sources of ignitions are numerous, ranging from overturned heat sources to abraded and shorted electrical wiring, and from spilled chemicals having exothermic reactions to friction caused by things rubbing together. Discovery At some point, the re resulting from the ignition will be discovered if it has not self-extinguished (this aspect is discussed further below). In the confusion following an earthquake, the discovery may take longer than it might otherwise. Report If it is not possible for the person or persons discovering the re to extinguish it immediately, re department response will be required. For the re department to respond, a report to the re department must be made. Response The re department then must respond. Suppression The re department then has to suppress the re. If the re department is successful, it will move on to the next incident. If the re department is not successful, it will continue to attempt to control the re. However, the re spreads and becomes a conagration. The process ends when the fuel is exhausted that is, when the re comes to a rebreak. This process is also shown in Figure 29.24, which is a re department operations time line. Time is of the essence for the re following earthquake problem. In this gure, the horizontal axis is time, beginning at the time of the earthquake, while the vertical axis presents a series of horizontal bars of varying width. Each of these bars depicts the development of one re, from ignition through growth or increasing size (size is indicated by the width or number of bars).
29-32
Earthquake
Ignition
Discovery
Report
FD Response
Stop Y
Suppression
N Spread / Conflagration
Firebreak / Fuel Exhaustion

FIGURE 29.23 Fire following earthquake process.
Beginning at the left (that is, at the time of the earthquake) is the occurrence of various res or ignitions (denoted by the number of the re in a square box; see the legend at the bottom of Figure 29.24). Some of these res occur very soon after the earthquake, while others occur some time later (due, for example, to restoration of utilities). The mechanism of these ignitions following an earthquake is no different from ignition at other times, although the earthquake can create unusual circumstances for ignition to take place. The primary difference due to the earthquake is the large number of simultaneous ignitions. Following this ignition, or re initiation, phase, there is a period during which the re is undiscovered but grows. A typical rule of thumb in the re service is that the rate of growth of an unimpeded re in this phase will double each 7 sec. In Figure 29.24, the size of the re is denoted by its number of bars. That is, each bar for a particular re represents one engine required for control or suppression. Thus, if a re as charted in Figure 29.24 proceeds, with time, from one bar to two and then three, this denotes that the re is growing and now requires three class A re engines to control the re (class A re engines have approximately 1200 gpm of pumping capacity). The letter D denotes discovery of the re. Discovery under the post-earthquake environment is often no different from other times, although discovery may be impeded due to damaged detectors or distracted observers. Upon discovery, citizens may themselves attempt to combat the re and will sometimes be successful. We are concerned herein only with ignitions that citizens cannot or do not successfully combat,
29-33
Earthquake Occurs Earthquake Occurs
Fire Department Operations Timeline

Ignition Discovery Report FD Arrival Fireground Operations
Phase
1 D 2 3 4 5
R D R D D
6
? ? R ? ?
DR R D R D R D R
7 8
Time Legend
7
Fire Growth
Ignition
Fire Discovery
Suppression
R
Fire Report
Fire Response
Fire Growth
(1~3 Engine fire)
FIGURE 29.24 Fire department operations time line.
and that require re department response. The letter R denotes receipt of the re report by the re department. Under normal circumstances, res are reported to the re department by one of four methods: 1. 2. 3. 4. Telephone Fire department street boxes (voice or telegraph) Direct travel to a re station by a citizen (citizen alarms) Automatic detection and reporting equipment, usually maintained by private companies
With the exception of citizen alarms, these methods, under normal circumstances, communicate the occurrence of a re within seconds, which is critical in the timely response and suppression of structural res and in the size and design of modern re departments. In the critical minutes following an earthquake, review of earthquake experience has indicated that, at present, in areas of strong ground shaking, citizen alarms are likely to be the only feasible method for reporting res. The telephone system may or may not sustain damage, but almost denitely will be incapacitated due to overload. Fire department street boxes are generally no longer in use (in California, only San Francisco maintains street boxes). Automatic detection and reporting equipment account for only a fraction of commercial property. Such equipment may be damaged in an earthquake and will likely produce many false alarms, which leads to lack of response to real res because of the inability to discriminate the real from the false alarms. Several other less conventional reporting methods may be employed. These include: Amateur shortwave-radio operators Helicopter observation Ground reconnaissance by police or re personnel
29-34
With regard to the last of these methods, present post-earthquake damage reconnaissance planning on the part of several larger California re departments has been reviewed and, in general, found to be unrealistic and inadequate for identifying res at a sufciently early stage to prevent conagration. A fundamental aw in most of these plans is the performance of the post-earthquake damage reconnaissance by the re personnel themselves, employing re apparatus (i.e., engines and trucks).3 The aw lies in the fact that, following an earthquake, both personnel and apparatus will almost immediately be redirected from the reconnaissance to actual res or other emergency response. Helicopter observation and amateur radio operators similarly will typically be able to identify and report res only after they have reached greater alarm status (that is, when it is too late). Further, most re ofcers have no special training in aerial observation or command. Following the re departments receipt of the re report, apparatus will respond, if available (in Figure 29.24, arrival of apparatus at the re is denoted by the engine number within a circle note that herein we track only re engines, because only engines can suppress serious res). Response may be impeded by blocked streets due to collapsed structures or by trafc jams. Upon arrival, the re may be combated according to normal procedures or, if the general situation is sufciently serious, minimal tactics may be all that is possible. Minimal tactics may constitute: Deluge (ood and run tactics) Abandonment of the burning structure and protection of exposures only Recognition that exposed structures cannot be protected, with fallback to a defensible line (e.g., abandonment of a city block, with attempt to stop the re at a wide street) Total abandonment, that is, recognition that either little or nothing can be done, that the re will burn itself out at an identied rebreak with or without re department intervention, or that other situations are more critical and demand the apparatus Water supply is a critical element and earthquake damage to the water system may reduce supply, thus altering tactics. Due to the interconnectedness of the elements of a water supply system, earthquake damage at some distance from a reground may still result in reduced supply. In Figure 29.24, we denote increasing control of a re by the reduction in the number of bars (i.e., engines required for control). As suppression progresses and control of the re becomes near total, the incident commander will release engines for more pressing emergencies elsewhere. Movement of these released apparatuses is denoted by a diagonal arrow showing travel of an engine from one re to another. As res are controlled, engines eventually converge on one or several large res, or conagrations. The rate of growth and spread of conagrations is a function of building materials, density, street width, wind, water supply, and reghting tactics. In this methodology, the process depicted in Figure 29.24 can be coded in a computer program. An algorithm determines ignitions, assigns a number to each re, and tracks re growth. Algorithms also determine re-reporting time and re engine arrival. Each re engine is tracked from location to location. Damage to the water supply is determined on the basis of available information and, in the case of reduced water supply, reduction in re-suppression capability is estimated. Final burned area for each ignition is thus calculated as a function of re growth and applied re-suppression capacity. Each aspect of the problem is discussed in more detail next.
3 Basic re service apparatus consists of two types: (1) engines (or pumpers), which typically carry a pump of about 1200 gpm capacity, 2000 ft of hose, 500 gal of water, some tools and small ladders, and several reghters; (2) trucks (also termed ladders, aerials, hook and ladders, etc.), which carry large ladders of about 75 or 100 ft length, a much larger variety of tools and ladders and additional reghters. A ladder truck by itself cannot pump water, so that the pumper is one basic measure of reghting capacity. For typical non-earthquake structural res, however, the limiting factor is more often manpower than pumping capacity.
29-35
29.3.1 Ignition
Available twentieth-century U.S. experience of re following earthquake, provided in Table 29.1, has been reviewed so as to identify earthquake-caused ignitions requiring re department response. (A detailed review of these data is beyond the scope of this chapter.) Based on this review, it has been determined that post-earthquake ignitions are typically a random event due to: Excessive motions that result in overturning or breakage of building contents (e.g., ignitions due to open ames or chemical reaction) Excessive structural deections that result in abrasion or other damage to electrical wiring and consequent short-circuiting Building collapse is an extreme example of the rst mentioned cause and will greatly increase the probability of ignition. Gas piping may be ruptured by either incident. For example, a water heater may overturn, thus breaking the connection, or a building may slide on its foundation and shear connections. Less typical but observed modes of ignition are heat due to friction or sparking due to pounding of structures. Post-earthquake re ignitions have been normalized by building density and the data regressed against shaking intensity. Two equivalent measures of density and intensity are used: 1. For building density: SFED (single-family equivalent dwellings = 1500 sq ft of oor area) are chosen as a measure that is readily understandable by laymen (they can be thought of as detached houses). A large building of 1.5 million sq ft, for example, would be 1000 SFED. Millions of square feet of total building oor area. 2. For seismic density: Modied Mercalli Intensity (MMI), which is a common measure of shaking intensity. Peak ground acceleration (PGA), which is a measure more common in the technical community (see Chapter 4, this volume, for denitions of both MMI and PGA). SFED and MMI have been used by the author for several decades, for purposes of communication with laypersons and the re service (whose jargon differs from that of engineers). Figure 29.25 shows normalized re ignitions as a function of MMI vs. 1000 SFED, and PGA vs. a million square feet of building oor area. A clear trend of increasing ignitions with increasing intensity can be seen. This trend can be stated as shown in Table 29.8, which provides a simple rule of thumb for estimating the number of ignitions that a re department will have to cope with. For example, for a town of population 20,000 subjected to MMI IX, the number of single-family-equivalent dwellings (SFED) in the town can be roughly estimated to be 20,000/3.5 = 5700 for residential. Total SFED is then about two to three times this (to account for commercial and industrial space), which equates to a total SFED of about 14,000. If the same town were to be subjected to MMI IX, one would expect about ve ignitions. While only 1971 San Fernando and later data are employed in this relation, it would be about the same if all twentieth-century events had been employed. Interestingly, the trend is also quite similar to that observed in Kobe in the 1995 earthquake. Post-earthquake ignitions for a particular locality can thus be calculated as a random Poisson process with mean probability determined as a function of MMI or PGA and building inventory (i.e., SFED, or millions of square feet of building oor area). These rates of ignition do not account for possible intentional ignitions arising out of several motives (i.e., arson). It can be argued that arson will be a signicant problem, because property owners are, in general, aware that while they may not be covered for shaking damage by their insurance, their re coverage includes re following earthquake. As will be seen, it takes relatively few additional ignitions to overwhelm re department resources, with possible ensuing conagration. Countering this is the
29-36
0.30
Ign. per thous SFED
0.25 0.20 0.15 0.10 0.05 6 7 8 9 MMI 10
0.45 0.40 0.35
Ign. per MMsf
0.30 0.25 0.20 0.15 0.10 0.05 PGA 0 0.1 0.2 0.3 0.4 0.5 0.6
FIGURE 29.25 Post-earthquake ignition rate, based on San Fernando and later data: (A) ignitions per 1000 SFED vs. MMI and (B) ignitions per million square feet oor area vs. PGA.
TABLE 29.8 Approximate Number of SFED per Ignition vs. MMI

MMI 7 8 9 10 1 Ign. per SFED 12,000 7,000 3,000 1,000
29-37
point that, while past earthquakes have seen examples of arson, these appear to have been relatively few and did not occur immediately after the shaking but rather days or weeks later. That is, in the immediate post-earthquake period (the rst minutes to hours) of injury, emotion, confusion, and multiple simultaneous ignitions potentially leading to conagration, people will be too overwhelmed by events to consider arson. Later, when this occurs to the relatively few persons who would actually commit this crime, the res they set will occur in a period when the re departments are back on a more normal footing, have been reinforced from outside the stricken area, and presumably will be able to handle these res. We estimate that the latter scenario is more relevant and hence have not included an allowance for arson in the methodology. Arson ignitions can be included, if one wishes to do so, by simply increasing the number of initial ignitions. The author has specically investigated this aspect (i.e., intentional res) following major earthquakes during the 1980s and 1990s and conrmed the above observations.
29.3.2 Fire Report and Response

As discussed above, citizen alarms are likely to be the dominant or only method of reporting res in the minutes following a major earthquake. This means that, following discovery, a re will be reported by a citizen running or driving to the nearest re station. Thus, we determine time of reporting as a period of delay (herein termed earthquake delay, which accounts for initial confusion, trafc difculties, etc.) plus travel time from the location of the re to the nearest re station. In our model, travel time can be determined based on either direct or right-angle travel and vehicle speed is a variable. Time of re engine travel to the re following receipt of the report is similarly based on distance and vehicle speed. Under normal conditions, re engines average between 15 and 20 mph in responding to a re (i.e., if distance traveled is divided by total elapsed roll time, the result is in the range of 15 to 20 mph of course, higher speeds are attained during certain portions of the travel). Delays are possible due to detours or trafc jams. Debris blockage of streets is a potential issue and was observed to be a signicant factor in Kobe in 1995. This has been considered for typical U.S. urban situations, with the conclusion that typically this should not be a major impediment although it may occur in selected districts. Depending on the time of day, trafc jams may be a more critical factor. Based on a review of post-earthquake trafc conditions in U.S. earthquakes, typical delays can be accounted for by the abovementioned earthquake delay factor.
29.3.3 Fire Growth and Spread

Fire growth is a particularly complex phenomenon. Most research in the United States has concentrated on re growth within one room (so-called compartment res) and only very limited research has been performed on U.S. interbuilding urban re growth [Takata, 1968; Woycheese et al., 1999]. More work has been performed on wildlands re spread [Rothermel, 1983], but this work is of limited applicability to the urban situation. Of available urban re spread equations, the Hamada equations are most useful [Hamada, 1951, 1975; Horiuch, n.d.; Scawthorn et al., 1981; Terada, 1984]. These equations are based on Japanese experience in twentieth-century conagrations, both during peacetime (e.g., following the 1923 Kanto earthquake) and wartime. The equations provide an elaborate model of re spreading for urban Japan, accounting for buildup to a fully involved re, which is too extensive to detail here. From Hamadas model, conservatively using parameters for only a fully involved re, a result that can be used [Scawthorn, 1981] is that:4
2 N ( xt , V ) = (1.5 ) a0 K s K d + K u
] [
(29.1)
A similar form of the Hamada equations is provided in the HAZUS [2002].
29-38
where N(xt , V) is the number of low-rise buildings destroyed by re spread, per re outbreak; xt is time in minutes; V is wind speed, m/sec; a0 is average building plan dimension, m; is total plan area of wooden buildings divided by (total area minus water area) of the subject regional subsection; and K d = (a0 + d ) T4 xt K s = (a0 2) + d + (a0 + d ) Ts ( xt Ts ) K u = (a0 2) + d + (a0 + d ) Tu ( xt Tu ) and
(29.2a) (29.2b) (29.2c)
] [
] [
Ti =
(1 fb )[3 + 0.375 a0 + (8d (c 4i + c 5iV ))] + fb [5 + 0.625 a0 + 16 d (c 4i + c 5iV )]

c1i 1 + c 2iV + c 3iV 2
(29.2d)
where d is the average building separation, m; T4 is the time in minutes required by the fully developed re to advance to the next building in the downwind direction; Ts is similar in the side-wind and Tu is similar in the up-wind directions; fb is the ratio of re-resistant buildings; i = 4,s,u and cji (j = 1 to 5), given in Table 29.9.
TABLE 29.9 Constants for Equation 29.2 (ad)

i 4 s u c1i 1.6 1.0 1.0 c2i 0.1 0.0 0.0 c3i 0.007 0.005 0.002 c4i 25.0 5.0 5.0 c5i 2.5 0.25 0.2
The Hamada equations are based on the hypothesis of an urban area being a series of equal blocks of structures with equal spacing between structures. The plan dimension of the average structure is denoted a, with average separation distance between structures d. The built-upness, or building density ratio , is dened above. Figure 29.26 shows re spread as an irregular oval, which is typical of res burning through a relatively uniformly distributed fuel load, with constant wind velocity. In actual urban conagrations, patterns of re spreading are more variable and are a function of wind speed and changes in direction, varying fuel availability in differing directions, and of re suppression tactics. Burned area is approximately (Kd + Ku) * (2Ks). To explore the question of utility of these equations in the U.S. context, observed re spread in various U.S. twentieth-century res was compared with the re spread predicted using these equations [Scawthorn, 1987] (see Figure 29.27). It was found that, with the exception of spread by branding among wood buildings under high winds, estimation using the Hamada equations agreed well with observation. Floor-to-oor re spread in modern mid- and high-rise construction typically is slower, however, and we have modeled this by using the above equations reduced by a factor. Note that mid- and high-rise re rate of resistance varies substantially by jurisdiction, depending on local re codes and enforcement. These estimates are for re spread within one city block or a built-up district and do not account for re spread across rebreaks, such as streets. To develop estimates of the probability of a typical fully developed building re crossing a rebreak under various wind speeds and with and without active re suppression efforts (see Scawthorn, 1987 for details), World War II and other data were reviewed [Bond, 1946]. Fire growth and spread then is modeled using these equations, taking into account probability of
29-39
10 mins
Wind
1 hr
2 hrs
FIGURE 29.26 Typical re spread.

FIRE SPREAD: ESTIMATION VS OBSERVATION Method of Spreading B
40
35
Branding among wood bldgs in high winds
30
SFb B G
OBS. FIRE SPREAD (fpm)
25
20
15 SFg 10 G 5 0 0 S SFg SFg SFg G G G SFg SFg SFg G G SFg G G SFg S GU 2 4 6
Line
of p
er
Ag fect
reem
ent
SFg SFg D SFg 14 16 18 20
8 10 12 EST. FIRE SPREAD (fpm)
FIGURE 29.27 Fire spread estimated with Hamada equations vs. re spread observed in U.S. non-earthquake conagrations. (Source: Scawthorn, C. 1987. Fire Following Earthquake: Estimates of the Conagration Risk to Insured Property in Greater Los Angeles and San Francisco, All-Industry Research and Advisory Council, Oak Brook, IL. With permission.)
29-40
crossing a rebreak (with or without active re-suppression capability being present). Fires are tracked and merged where they meet (i.e., areas are not burned twice).
29.3.4 Fire Response and Suppression

As discussed above, re growth and re engine location are tracked concurrently. Under normal conditions, urban re department response to a structural re is usually a minimum of two re engines and one ladder truck (additional apparatus responds in high-value or extra-hazard areas). These normal responses will not be possible following a large earthquake because res may outnumber re engines. Based on review of actual operations following earthquakes and discussions with senior re department ofcials, it is likely that, following an earthquake, only one engine will respond to reported res initially, to suppress the re or size up the situation. Thus, initial response should be modeled in this manner, initially allocating one engine to each re and additional engines as available. Where res outnumber rst-line engines, the difference is termed excess res. A question arises whether re department resources will initially be totally and primarily devoted to re suppression, because it should be recognized that other demands (search and rescue, hazardousmaterial response, emergency medical treatment) will also be placed on these resources. For most analysis, it should be assumed that the re departments rst priority will be re suppression. Note that this assumption is optimistic, especially because building collapses and hazardous-material releases may involve large numbers of victims. This aspect has been reviewed with senior ofcials of several re departments and their opinion is that some re department resources will have to be diverted from reghting to these other services. However, experience has shown that serious res typically receive rst priority for the following reasons: Fire service training and tradition. Fires are dynamic while building collapses are relatively static that is, a re situation will worsen if neglected, while the building collapse and rescue situation can often wait several hours (indeed often must await the arrival of heavy equipment). Other services (police and others) can assist in building collapses, emergency medical treatment, and hazardous-material management (via isolation and evacuation), while only the re service is equipped to handle serious res. The impact of this assumption is to decrease total expected losses due to re following earthquake. In addition to each jurisdictions re-suppression resources (i.e., the departments rst-line and reserve engines, other equipment, and personnel), automatic and mutual aid need to be considered. These resources, of course, arrive somewhat later, from more distant locations. The size of the re at rst engine arrival time (in terms of actively burning SFEDs and water required for suppression) is the next parameter in the model and should be compared with that engines suppression capability. Engine-suppression capability should be modeled using guidelines appropriate for typical structural res under non-earthquake conditions and modied to take into account reduced available water (due to earthquake damage to the water supply) and likely re department use of minimal tactics (discussed above). That is, given a burning area, required re ow under normal conditions can be computed on the basis of 4 gpm for each 100 cubic feet (cf) of occupancy directly involved in the re or immediately exposed [Kimball, 1966]. For larger res, this volumetric calculation is based only on perimeter defense. Depending on construction and available manpower, a re engine typically can apply up to 1500 gpm (using the monitor or handlines note, however, if 1500 gpm is to be applied entirely by handlines, additional personnel will be needed). This may be reduced depending on the post-earthquake condition of the water supply system. Thus, one engine can typically attack a maximum burning volume of 37,500 cf (typically, 3000 to 4000 sq ft of oor area) if the monitor can be efciently used, or half of this (about one SFED) if handlines are used and additional personnel are not available. If minimal tactics are
29-41
employed (i.e., no attacks, perimeter protection only), then the capacity of one engine can be considered to be increased (e.g., up to 300 or 400 linear feet of perimeter). Damage to the water supply was, of course, of prime importance in the 1906 San Francisco, 1989 Loma Prieta, 1994 Northridge, and 1995 Kobe earthquakes and will likely be critical in future earthquakes. As discussed above, although the main water transmission lines into San Francisco were severed by the earthquake, this was not critical sufcient water for initial reghting needs existed in reservoirs, etc., within the city itself. Rather, hundreds of breaks in the local distribution network, due to large ground displacements and liquefaction, resulted in the system hemorrhaging water and losing pressure. To account for damage to the water supply, two approaches can be employed: 1. A detailed hydraulic model of the system with estimated damage due to shaking, permanent ground displacements, etc., is preferred. This approach, however, may be beyond the scope of some investigations. 2. An alternative method is to review the regional water supply systems and likely areas of liquefaction and permanent ground displacement. Based on this review, for areas where permanent ground displacements are determined to severely impact the water supply system, remaining water supply functionality and reduction in re-suppression capability can be estimated judgmentally. This approach is described in detail in ATC [1992]. At the time of rst arrival, a re (in terms of SFEDs) larger than the suppression capability (in terms of SFEDs) of the allocated (one or more) engines is termed a large re. Large res are assumed fought by available engines in a downwind perimeter defense, and to initially spread elliptically at down/up/ side wind rates through a uniformly spaced gridwork of buildings. This assumption of a uniform gridwork of buildings is reasonable, and has been used in all urban re modeling to date. The spacing, story height, etc., of this gridwork are a function of building density and built-upness (the latter is the ratio of property devoted to buildings contrasted to the total area, including streets, parks, etc.). Firespreading rates are decreased and the initial elliptical shape is altered as the available reghting capability increases. Eventually, for each re, one of three things happens: 1. Fires are suppressed. That is, reghting capability exceeds needs (either initially, or with build up of engines as other res are suppressed and their engines redirected or mutual aid engines arrive) and the re is surrounded, controlled, and suppressed. 2. The re is too large and capability is exceeded by demand. The re burns relatively freely within a city block. At each street or other (down/up/side wind) rebreak, the re crosses, with crossing probability as discussed above. This probability is a function of rebreak width, wind direction, and, especially, suppression capability. If sufcient engines and water are available, many res will typically be stopped at the rst wide street. Note that strong winds have an important effect on downwind rebreak-crossing probabilities. In this context, branding (windborne transmission of aming debris) is sometimes an important factor, especially where wood roofs are prevalent (these are banned in San Francisco, but are common in some other jurisdictions). In most cases, a re is not expected to cross more than a few typical streets, so that most large res will be stopped, or burn themselves out, within a few blocks. Again, moderate to strong winds will extend this stopping distance signicantly. 3. The re reaches an ultimate rebreak that is, a large expanse of water, the edge of the urbanized area, etc. Available engines are sufcient to defend exposures along the remaining perimeter, and the re is controlled and suppressed.
29.3.5 Final Burned Area

The above method is followed for each re. Fires need to be tracked and merged where they meet (i.e., areas should not be burned twice). Final burned areas are summed, to arrive at total nal burned area.
29-42
29.3.6 Example: Vancouver, B.C.

The re following earthquake methodology presented above has been applied in a number of localities for various purposes, including: Assessing the adequacy of the re service, in terms of the amount and location of resources, such as re stations, reghters, and equipment, as well as their earthquake planning. Examining the reliability of water supply systems to provide reghting water following a major earthquake. Planning a metropolitan region, to examine whether major thoroughfares could serve as rebreaks and how building codes might be employed to reduce the re following earthquake problem. Examining the potential nancial impacts a major earthquake may have on the insurance and nance industry. Note that in the United States and many other countries, re following earthquake is covered under the re insurance policy, not under the earthquake policy. Therefore, a major conagration following an earthquake is virtually fully insured, as opposed to major building damage due to shaking, which may have little or no impact on the insurance industry (see Chapter 32, this volume). The re following earthquake problem is a major concern to the insurance industry, which has supported research and studies in this area for several decades.
M6, M6.5 M7.5
1975 / 1997 epicenter M7.5
1909 epicenter M7.8
FIGURE 29.28 Scenario earthquake events; analysis of re following earthquake, lower mainland, British Columbia. (From Scawthorn, C. and Waisman, F. 2001. Assessment of Risk Due to Fire Following Earthquake, Lower Mainland, British Columbia, Report prepared for the Institute for Catastrophic Loss Reduction, Toronto, by EQE International, Oakland, CA. With permission.)
29-43
An example of analysis of re-following-earthquake potential is a recent study performed by the author and colleagues for the Vancouver, B.C. (Canada) metropolitan region [Scawthorn and Waisman, 2001]. In summary, the analysis of potential property loss due to post-earthquake res in the urbanized portions of the lower mainland of British Columbia included the municipalities and districts of Burnaby, Coquitlam, Delta, New Westminster, City of North Vancouver, District of North Vancouver, Port Coquitlam, Port Moody, Richmond, Surrey, Vancouver, and West Vancouver, which have a total population of 1.76 million, representing 96% of the Greater Vancouver Regional District (GVRD). The lower mainland of British Columbia is one of the highest earthquake risk regions in Canada. Seismological research indicates that there are several active seismic sources in the region that have caused destructive earthquakes in the past and are potential sources of future destructive shocks. The region is most affected by earthquakes originating (1) in the shallow crust, (2) on the Cascadia subduction zone, and (3) at depth under Vancouver Island and the Strait of Georgia. Based on the seismotectonics of the region, several scenario earthquakes were analyzed, including: A magnitude 9.0 megathrust event on the Cascadia subduction zone Three events of magnitude 7.5 to 7.8 occurring at various locations in the Strait of Georgia Magnitude 6.0 and 6.5 events randomly placed central to the region of interest (i.e., located beneath the City of New Westminster)5 Shaking intensity due to the scenario events varies signicantly across the region, ranging from MMI V in areas of good soil to MMI IX in areas of soft soils (e.g., around False Creek in the City of Richmond).6 Because each of these scenario earthquakes has a nite probability of occurrence and numerous other earthquakes with associated probabilities of occurrence can occur at other locations, all possible earthquakes were also analyzed in a probabilistic analysis, based on the seismotectonics of the region. The lower mainland of British Columbia is also one of the larger urban regions of Canada, with a high concentration of population and property. An estimate of total property values for buildings, vehicles, and infrastructure is approximately $260 billion.7 Wooden buildings represent approximately 58% of the building stock by value and 65% by total oor area. There are about 1200 high-rise8 buildings in the region, only about 20% of which are equipped with sprinkler systems. Of signicant concern, however, is that very few of the high-rise buildings in the region have on-site reghting water supply, in contrast to high-rise practice in the United States, where high-rise buildings in high seismicity areas typically have a 15,000-gallon or larger on-site dedicated water cistern. The analysis of potential losses due to res following earthquakes in this example followed the methodology discussed above, in which the process of (1) earthquake occurrence, (2) resulting intensity of shaking at each location in the region, (3) structural damage and resulting probability of ignition, (4) re discovery and reporting, (5) re service response and effectiveness given damage to water supplies as well as other demands on the re service, and (6) re spread is treated as a stochastic process, with selected key variables treated as random variables. The uncertainty of these random variables is quantied on the basis of data from earthquake and non-earthquake events. The geographic unit of analysis for the region was three-character postal code (e.g., V6E xxx) that is, the analysis treated each of the 80 three-character postal codes in the Greater Vancouver region as having uniform shaking intensity, building inventory, and other characteristics within that postal code area.
Magnitudes are measured on the moment magnitude scale. MMI is the Modied Mercalli Intensity scale, where VI is the initiation of damage to poorly built buildings and IX is collapse of such buildings, with signicant damage to average to good buildings. Corresponding peak ground accelerations range from about 0.02 to 0.45 g. 7 Unless otherwise noted, all monetary amounts are 1999 Canadian dollars. 8 Denition of a high-rise building varies in this analysis, high-rise was dened as seven stories or greater, which is a denition widely employed in the re service. This estimate is conservative and there are quite possibly more than 1200 high-rises in the study region.
6
29-44
To determine the availability and effectiveness of reghting and water resources, data were collected from a variety of sources, including interviews with key re service and water department personnel, Fire Underwriters Survey municipal grading summaries, and engineering analyses of water department seismic vulnerabilities, nding that: Fireghting The Greater Vancouver region is well protected under normal circumstances by a well-trained and equipped re service, largely professional but in part volunteer, with a total reghting stafng of about 2700, of whom about 400 are on duty at any one time. The total number of re engines in the region is 120, housed in 77 re halls or stations, a signicant fraction of which may be seismically vulnerable. In addition to re engines, additional pumping capacity is furnished by ve fast attack reboats sited around Burrard Inlet (equivalent to ve re engines) and, in the City of Vancouver, the recently constructed dedicated re protection system (DFPS discussed further below) and associated hose tenders and special pumps (from a pumping capacity perspective, DFPS can be equated to about eight re engines, although its effectiveness is much greater). Water The region is almost entirely served by the Greater Vancouver Water District (GVWD, part of the GVRD), which has three supplies: Capilano, Seymour, and Coquitlam. The Capilano and Seymour supplies are north of the Burrard Inlet. They supply Vancouver, Burnaby, and other areas via transmission lines that cross the inlet via submarine crossings that, however, pass through areas subject to liquefaction. The Coquitlam supply is northeast of Vancouver and no major marine crossings are required to reach the City of Vancouver or a number of other areas. Since the early 1990s, GVWD has been implementing a program of seismic improvements and is currently conducting a major review of that program to assure that appropriate goals are being met. Within each municipality, water is distributed via a gridded network of smaller pipelines that are anticipated to break in a large earthquake, such that in a number of locations (especially in areas of softer soils), re hydrants will initially be without adequate water pressure or supply. To prepare for such contingencies, the re service typically identies alternative water supplies and carries special hard suction hose and other equipment to access these supplies. Examples of alternative water supplies include water reservoirs, creeks, waterfront access to bays or lakes, and swimming pools. Among Burrard Inlet, the Fraser River, and other sources, the study region is generously supplied with opportunities for alternative supply. All re departments interviewed had equipment for accessing alternative water supplies, but only ve out of eight departments appeared to have adequately planned for such contingencies. As noted above, reboats and Vancouvers DFPS signicantly enhance emergency water supply in Vancouver and around Burrard Inlet. However, there are no reboats located on the Fraser River. For each scenario, the number of initial res was estimated based on shaking intensity, and re department response estimated taking into account effects of the earthquake on transportation, communications, water supply, and related factors (see Figure 29.29 and Table 29.10). Fire growth was estimated for each re, taking into account re suppression (or the lack thereof), to arrive at an estimate of nal burned area for each re and the sum of all res. Loss is estimated as the monetary amount corresponding to this nal burned area.9 Based on these data and analyses for the scenario events, it was estimated that the Greater Vancouver region would sustain res and losses as indicated in Table 29.10. Uncertainty associated with these estimates is substantial, with coefcients of variation on the order of 0.5 to 1.0. Taking into account all possible earthquakes together with their associated probabilities of occurrence, it was found that the annualized loss for the study region was $88.3 million, representing 0.04% of the value at risk. These losses are of similar relative magnitude to previous estimates for other North American urban areas [Scawthorn and Khater, 1992].
9 Note that shaking damage was not quantied in this analysis, and that shaking loss and post-earthquake re losses cannot be directly summed, as there may be some burning of the rubble.
29-45
FIGURE 29.29 Mean number of ignitions, M9.0 subduction zone event. (Source: Scawthorn, C. and Waisman, F. 2001. Assessment of Risk Due to Fire Following Earthquake, Lower Mainland, British Columbia, Report prepared for the Institute for Catastrophic Loss Reduction, Toronto, by EQE International, Oakland, CA. With permission.)
TABLE 29.10 Scenario Events, Fires, and Losses (Loss in Year 2000 C$ millions, % of total value at risk)
Georgia Strait M = 7.5 113 res Loss $2,795 % 1.04% 1909 Epicenter M = 7.8 31 res Loss $1,172 % 0.43% 1975/1997 Epicenter, M = 7.5 101 res Loss $2,685 % 1.00% New Westminster M = 6.0 66 res Loss $1,752 % 0.65% New Westminster M = 6.5 157 res Loss $8,529 % 3.12% Subduction Zone M = 9.0 155 res Loss $4,700 % 1.74%
With regard to the conditions found in the Greater Vancouver region, it was found that reduction of post-earthquake re risk could be accomplished via a number of measures, including: Development of a regional earthquake plan with a detailed post-earthquake re element, which would include procedures for rapid location and reporting of post-earthquake res, mutual aid response to larger res, identication of alternative water supplies, and related measures. Improved provision for alternative water sources, including extension of the DFPS on the other side of False Creek, installation of saltwater hydrants along shorelines in higher density areas, provision of reboats on the Fraser River, and acquisition of additional portable pumps by several of the regions re departments, similar to the HydroSub recently acquired by the City of Vancouver. Installation of water cisterns in sprinklered high-rises and retrotting of unsprinklered high-rises with sprinklers. A regional effort at reducing post-earthquake re ignitions for example: Vancouvers central business district (CBD) is an enormous concentration of value, yet it is the only major North American city with overhead electric transmission lines. Wood polemounted transformers abound in the CBD, in many cases only inches from commercial buildings (this is discussed further below). In past earthquakes, pole-mounted transformers arced and exploded. Because these were typically in residential areas, few ignitions occurred, but it is expected that many ignitions would result in Vancouvers CBD due to these electrical
29-46
sources. Undergrounding of the electrical transmission in the CBD would signicantly reduce post-earthquake res. Fire departments could institute public education as to post-earthquake res, the need for their rapid suppression, and, specically, the need to brace residential water heaters.
29.4 Mitigation
This section discusses mitigation of the re following earthquake problem. Fortunately, numerous opportunities exist, although, unfortunately, their application is sparse. A framework for the discussion of mitigation opportunities is shown in Figure 29.23, which has been enhanced in Figure 29.30 to show opportunities for intervention in the re following earthquake process. As indicated in the diagram, following the earthquake damage in the most general sense (even if it is only the overturning of a candle) is required for an ignition. Thus, reduction of damage is the rst opportunity for mitigation of res following earthquake. This ignition might be extinguished without human intervention if automatic sprinklers or other automatic suppression systems (e.g., Halon) are present, and if the automatic system is post-earthquake functional, and if the water, for example for the sprinklers, is available. If the ignition is not automatically extinguished, then the re remains to be discovered, and the citizens so discovering it may be able to extinguish it, if water is available. If those persons are not able to extinguish the re, they must report it to the re department. For the citizens to report the re, functional communications (i.e., telephones) are required (unless, perhaps, the re department observes the smoke column and selfreports). For the re department to respond, they must have available re department assets, and the assets (i.e., apparatus) must be able to travel to the re that is, transportation, in the form of roads, etc., must be functional. Upon arrival, the re department will be able to extinguish the re, if water is available and if the re is controllable, given the re department resources reporting to the scene. If not, the re will continue to grow (admittedly, perhaps at a smaller rate given the re departments efforts), until it meets a rebreak. The provision of rebreaks, via the urban planning process, is the last opportunity for mitigation of re following earthquake. We discuss each of these aspects in turn.
29.4.1 Reduction of Damage

In a general sense, reducing earthquake damage or effects due to shaking reduces the potential for ignitions following an earthquake. Obvious examples include restraining sources of ignition such as water heaters, lamps, and other heat sources, and containers holding potentially reactive chemicals. Many sources of information and products exist for bracing and anchoring of simple household and ofce goods (e.g., www.abag.org) a simple search of the Internet will uncover many. Gas appliances in the home are a concern to many persons, because overturned water heaters or broken gas piping (due to excessive displacements of a seismically weak house) can result in gas leaks that, in a few minutes, can produce a life-threatening explosive mixture of gas and air. Figure 29.31 shows how to restrain a gas water heater. Mroz and Soong [1997] evaluated the method depicted as well as other methods of restraint, and found varying effectiveness using only plumbers tape provided resistance against falling away from the wall, but not toward the wall thus use of rigid tubing is recommended. Similarly, Figure 29.32 illustrates the gas valve controlling service to a home, together with a gas wrench, afxed by the homeowner to the gas valve so as to be always available in the event of an emergency. Because persons may not be home when an earthquake occurs, and thus are unable to turn off the gas service in the event of a gas leak following an earthquake, automatic shutoff valves have been invented. These are of two types: (1) excess ow valves actuated by a change in gas pressure and (2) seismic valves actuated by motion or movement. Both are intended to actuate when subjected at the point of installation to an earthquake severe enough to cause damage to gas systems and appliances within the structures they protect. Figure 29.33 shows an example of an automatic gas shutoff valve installation.
29-47
Earthquake
Damage Reduction
Ignition
Automatic Suppression (e.g., sprinklers)
Discovery
Water Citizen Suppression Communications FD assets / Transportation
Report
FD Response
Stop
Suppression Y
Spread / Conflagration
Urban Planning
Firebreak / Fuel Exhaustion

FIGURE 29.30 Fire following earthquake process, with mitigation opportunities.
The above precautions are appropriate assuming the structure is relatively safe; however, should the house slide off its foundation, the gas service line will be sheared off, allowing pressurized gas to escape to the atmosphere one spark produces a gas torch of major proportions, which will quickly ignite the home. In such cases, if one can get to the pipe before the gas ignites, simply plugging the broken pipe with a tapered piece of wood effectively turns the gas off. Older homes with unbolted foundations or unbraced cripple walls are particularly common in the western United States and especially prone to sliding off their foundations. Figure 29.34 shows an example of how to bolt and brace cripple walls, while Figure 29.35 shows houses fallen off their foundations in the 1989 Loma Prieta earthquake. Another source of post-earthquake ignition is electricity, which can cause ignitions via a number of mechanisms, including wiring shorts, appliance damage, overhead line arcing, etc. A particularly egregious example of the latter is found in the CBD of Vancouver, which is the only major North American city with overhead electric transmission (see Figure 29.36).
29-48
Strapping Your Water Heater

4 Minimum
5/ 3 16
4 Minimum 24 Maximum
24 Stud Sheetrock
Log screw with washer
24 Maximum
1/ EMT (conduit) flattened at ends 2 Bolt (with nut and washers) 1/ 16 -gauge metal strap 1 2 through strap and flattened end of EMT 5/ 11/ Bolt (with nut and washers) 16 4 through bent ends of strap 5/ 3/ 16 4
(A)
Locate center of studs with a small finish nail Wall sheathing

1/ EMT (conduit) 2 flattened at ends
Water heater
11/216 -gauge metal strap
24 Studs behind sheetrock (B)
Flexible gas pipe connection (C)
FIGURE 29.31 Strapping water heater: (A) plan view, showing plumbers tape strap around water heater and attachment to framing using rigid tubing; (B) elevation view; (C) photo of actual water heater note plumbers tape strap around belly of the water heater. (Courtesy EQE International and C. Scawthorn)
Finally, another major source of post-earthquake res is chemical reactions. Glass and other containers containing chemicals can easily fall off shelves or tip over in an earthquake, spilling their contents and reacting with other spilled chemicals or materials such as ooring or furniture. The heat given off by the reaction (termed exothermic reactions) can be sufcient to ignite cloth or other materials, thus starting a re. High-school chemistry laboratories are a source of res in almost all major earthquakes. In California and other high-seismicity regions, it has become relatively common practice to put lips on
30
Water heater
Max
imum
30
M um im ax
29-49
Shutting of the Gas and Other Utilities

GAS SHUTOFF
(Click on diagram for a detailed drawing)

ON OFF
Switch off gas
Switch on gas Special Tool for turning off gas water on 3 gas and Gas Meter water (A crescent Wrench Switch off or on gas WI Work.)
1. Locate the main gas shutoff (usually outside the house) and all pilot lights. 2. Clear the area around the shutoff valve for quick and easy access in case of emergency. 3. A wrench (or specialty tool for turning off gas and water) should be attached to a pipe next to the shutoff valve or in another easily accessible but hidden location. 4. You may want to paint the shutoff valve with white or fluorescent paint so that it can be located easily in an emergency. 5. If you are concerned about your ability to turn off the main gas shuttoff valve, are unsure whether it is in proper working order (indication of rust, etc), or do not know how to relight your pilot lights, contact your local gas company. The gas company can send a service representative to your house to show you the proper procedure and check the valve and pilot lights to be sure that they operate properly. (A)
(B)
FIGURE 29.32 (A) EQE Web site explaining how to shut off gas valve and (B) actual gas service to a home, showing pressure regulator (at, round, plate-like object) and valve, with gas wrench wired to pipe and writing on side of house showing location. (Courtesy EQE International and C. Scawthorn)
shelves in laboratories, restrain chemical containers, put door latches on laboratory cabinets, and take other precautions to prevent spillage of chemicals.
29.4.2 Automatic Suppression

The re hazard in non-earthquake conditions is such that automatic suppression systems have become a standard part of most larger buildings. These also serve the same purpose following an earthquake and, by stopping many ignitions at their initiation, can be very effective in reducing the overall post-earthquake
29-50
TWINGATE SEISMIC VALVE
B-LINE B 437 CLAMP ON 1 1 STRUT
PLUG TEE
REMOVE EXISTING
GAS METER (A)
(B)
FIGURE 29.33 Example of automatic gas shutoff valve: (A) typical installation; (B) close-up of valve. (Courtesy Quake Defense, Inc.)
re problem. Thus, an item of critical importance for the problem of re following earthquake is the seismic performance of automatic sprinklers. Design and installation of automatic sprinklers in the United States are generally governed by NFPA 13,10 which includes provisions for seismic design of sprinklers. This design, which consists of
NFPA 13, Standard for the Installation of Sprinkler Systems (most recent edition), National Fire Protection Association.
10
29-51
1/ 2
Framing clips Rim joint Floor joints Rim joint connecting top or blocking extenior plate and rim joint Plywood of grade at 16 spacing plank subfloor structural 8d of nails at 8 plywood spacing at Top infernediate plates studs Exterior stud wall Anchor Bolt
Foundation grade beam 8d nails at 3 spacing around the perimeter of each plywood sheet Sill plate 3/8 from edge Nailing each side Holddown located of plywood joints or end of each 11/2 dia. ventilation Reinforcing plywood wall panel less than 8 holes top and bottom bars in length between studs 1995 by EQE International Materials Needed
1/ , STRUCTURAL I or C-DX plywood 2 8d common nails Simpson HD2A holddown or equivalent Simpcon A35 Framing clips with N8 nails or equivalent Anchor bolls (see project2)
Tools Needed Circular saw Jigsaw 1-1/2 hole saw Framing square Hammer Plywood blade Tape measure Chalk line Pencil
(A)
(B)
FIGURE 29.34 (A) Diagram and (B) example of cripple wall bracing using plywood. Note steel straps nailed through plywood into studs and horizontally bolted into concrete foundation. (Courtesy EQE International and C. Scawthorn)
lateral and longitudinal bracing for sprinkler mains at specied distances, is relatively simple. These provisions have generally proven to be adequate in most earthquakes [Fleming, 1998], although a aw in the design is the general overlooking of situations where the sprinklers are restrained by the bracing on either side of a building temperature joint. Large buildings are often built with joints, intended to limit stresses due to normal heating and cooling of the buildings during the day. Where sprinklers cross these joints, exible couplings have to be provided. Otherwise, in an earthquake, the two parts of the building on either side of the joint (effectively, that is, two buildings) will move differentially, thus
29-52
FIGURE 29.35 Watsonville in 1989 Loma Prieta earthquake; examples of houses fallen off their foundations. (Courtesy EQE International)
FIGURE 29.36 Vancouver, B.C. central business district, with overhead electric lines an obvious re hazard, especially following an earthquake. (Photo: C. Scawthorn)
29-53
FIGURE 29.37 Test of intumescent paint both buildings had crib res set at the same time. Building on the left is of bare wood and has ashed over and is fully involved at slightly less than 5 minutes following ignition, while building on the right, painted interior and exterior with intumescent paint, still has only a small re on the interior, not yet ashed over. (Photo: C. Scawthorn) Shown as Color Figure 29.37.
rupturing the sprinkler. Differential movement of piping supports, in general, is a seismic design issue often overlooked. One or a few failures of sprinkler mains in this manner can drain the sprinkler system and prevent functional post-earthquake re suppression. In addition to water-based sprinklers, other automatic suppression systems include halogenated agents (e.g., HALON), CO2, dry chemical, foam, and other systems, each of which can similarly be of great benet in the post-earthquake environment, and each of which should be reviewed to assure postearthquake functionality. In addition to automatic suppression systems, the post-earthquake functionality of re-detection and -alarm systems is vital following an earthquake, if for no other reason than to provide warning for evacuation. The author was involved in the development of an entire methodology for the assessment of such systems to assure post-earthquake functionality [Johnson et al., 1999], which is discussed in another chapter of this book. Finally, the application of intumescent paint, which is special paint for walls and other surfaces that is highly resistive to radiant heat, is emerging as a promising form of passive re protection in general. This paint may have special application for re following earthquake risk reduction. The resistance of intumescent paint is largely gained via chemicals in the paint that utilize the heat gain in expansion, resulting in a swelling or blistering of the paint rather than ignition, and the creation of a re-retardant foam. Ignition is thus much delayed and, in certain cases, prevented (see Figure 29.37). San Francisco has approved intumescent coatings for application in a number of historical buildings as an alternative or adjunct to other re-protection systems and is investigating its possible benets with regard to re following earthquake [Kobayashi and Korneld, 2002].
29.4.3 Suppression by Citizens

Because the essence of the re following earthquake problem is the multiple simultaneous ignitions following an earthquake, which overwhelm re department resources when they are most in demand, one approach to mitigating the problem is the use of additional personnel in the form of citizen volunteers. The concept of volunteers in re departments is not new; in fact, the rst re departments were all volunteers. The United States today still has large numbers of volunteer re departments outside the major metropolitan areas and many re departments are part paid and part volunteer. In this regard, several of the larger re departments in the United States have developed specialized training of citizen volunteers to be mobilized in the event of major disasters.11 San Francisco, for example, has trained 10,000 citizens in elementary reghting and search and rescue (see Figure 29.38).
11 For example, the San Francisco FD Neighborhood Emergency Response Teams (NERT), (http://www.sfnert.org/), and the Los Angeles regions Community Emergency Response Teams (http://www.cert-la.com/).
29-54
San Francisco Neighborhood Emergency Response Teams
[ Home ] [ NERT Auxiliary ] [ Our Mission ] [ Calendar ] [ Contact NERT ] [ Help NERT ] [ Links ] [ Search ]
NERTNews Calendar Change....... The next Coordinators Meeting will be June 26th, Archive 1011 Turk Street, 6:30 pm .... NERT Auxiliary Annual Meeting will be held Class Schedule May 22nd, 1011 Turk Street, 7:00 pm .... Training Class Content NERT Fall Class Schedule updated - 5/13/03 NERT Manual Battalions April Drill Photos Staging Areas Add SFNERT.ORG to your favorites NERT Forms HAM Operators: HAM Cram MS-Internet Explorer users click here Disaster Registry Practice your skills during Siren Netscape users hit CTRL+D to bookmark our site. First Responder Net Pictures CCSF announces Fall 2002 Terrorism Click here if you would like e-mail Bay Net First Responder Training notification of updates to SFNERT.org click here for details New Ham Cram - July 20, 2002
NERT Shirts Now Available
Let the U.S. Geological Survey know if you feel an earthquake
NERTNews is the official voice of the San Francisco NERT organization and is published quarterly by the San Francisco NERT Executive Board. NERTNews Editor is Lynn Jacklevich. NERTNews - Online Editor is Bill Travis. San Francisco NERT is a non-registered San Francisco volunteer group, registered with the Office of Emergency Service of the State of California as Volunteer Disaster Service Workers, and dedicated to preparing citizens of the City and County of San Francisco for potential disaster. To contribute to NERTNews, contact the Editor. The deadline for submitted materials is the first Friday of February, May, August, and November. NERTNews is under no obligation to print materials received, and submitted materials may be subject to editing for style, content, and length. For permission to reproduce or distribute any of the content of this newsletter, contact the San Francisco Fire Department for conditions and approval. Last Modified : 05/14/02 07:01 AM Copyright 1999-2002 SF NERT All rights reserved
Email webmaster
FIGURE 29.38 San Francisco Fire Departments NERT (Neighborhood Emergency Response Team) home page. (Source: http://www.sfnert.org/)
29-55
To quote the Los Angeles CERT (Community Emergency Response Team) Web site: Local government prepares for everyday emergencies. However, during a disaster, the number and scope of incidents can overwhelm conventional emergency services. The Community Emergency Response Team (CERT) program is an all-risk, all-hazard training. This valuable course is designed to help you protect yourself, your family, your neighbors and your neighborhood in an emergency situation. CERT is a positive and realistic approach to emergency and disaster situations where citizens may initially be on their own and their actions can make a difference. While people will respond to others in need without the training, one goal of the CERT program is to help them do so effectively and efciently without placing themselves in unnecessary danger. In the CERT training, citizens learn to: Manage utilities and put out small res Treat the three medical killers by opening airways Control bleeding and treat for shock Provide basic medical aid Search for and rescue victims safely Organize themselves and spontaneous volunteers to be effective Collect disaster intelligence to support rst responder efforts
The CERT concept emerged in 1985 as part of observations on Japanese neighborhood preparedness made by the United States during a cooperative United StatesJapan exchange of emergency response ofcials. The City of Los Angeles Fire Department under the leadership of Chief Frank Borden then developed a pilot program, which was so successful that, by 1993, the Federal Emergency Management Agency (FEMA) decided to make the concept and program available to communities nationwide. By 2002, 38 states and 6 foreign countries were using the CERT training.
29.4.4 Communications
The foregoing discussion emphasized the need for rapid reporting of res following an earthquake, and the fact that this is problematic due to saturation and possible damage to the telephone system, as well as overloading of the emergency dispatch (911) system. Fire department emergency dispatch centers have contingency plans for coping with a large number of calls in an emergency, but can still be overwhelmed at some point. An additional issue is damage to the emergency dispatch center itself, which should not only be structurally sound, but also have assured backup power and other resources.
29.4.5 Fire Department Assets

The re service is the rst line of defense against res following earthquakes. However, the re service is both vulnerable to earthquakes itself and sometimes not as well organized for this specic issue as it might be. This section discusses actions the re service itself can take to improve its preparedness for res following earthquakes. 29.4.5.1 Fire Stations As noted above (in the discussion of the 1994 Northridge earthquake, where 35 re stations sustained damage [Sepponen et al., 1997], re stations themselves can sustain damage in an earthquake, potentially injuring and rendering the assets they house (i.e., reghters and their equipment) useless in an earthquake. Many larger re departments in high seismic areas have reviewed their re stations structural seismic adequacy and subsequently strengthened inadequate premises. In the late 1980s, San Francisco, for example, undertook an analysis that found that, of 55 department facilities, 11 were collapse hazards and another 13 would sustain extensive damage in a major earthquake [EQE, 1989]. To strengthen and
29-56
(A)
FIGURE 29.39 (A) Example of single-bay unreinforced masonry re station, built in 1913. (Source: EQE, 1989. Report prepared for the San Francisco Fire Commission) (B) Fire station strengthened in the 1990s per indicated structural reinforcement scheme. (Source: C. Scawthorn)
replace re stations, the city embarked on a sweeping program that has since been completed. Figure 29.39, for example, shows a 1913-vintage unreinforced masonry single-bay re station that was analyzed by a team headed by the author and found to be a collapse hazard. A structural reinforcement scheme, also shown in the gure, was developed and constructed during the 1990s. In the Vancouver study discussed above, many re departments were found to have quite new re stations, or have reviewed and as required strengthened their re stations. However, there still remained somewhere between 14 and 30 re stations (i.e., 18% to 38% of the re stations in the region) that were seismically questionable. Thus, a review of re stations and other emergency facilities, for seismic
29-57
Center Core North and South Walls, Full Height, East and West Parapets Install Steel Diaphragm Bracing Beneath Ceilings, 1st and 2nd Story, Anchor Walls to Bracing Brace East and West Parapets
Gunite Front and Rear Walls
Infill Selected Windows Install Moment-Resisting Frame and Foundation 2 Places
(B)
structural adequacy and to assure post-earthquake functionality, should be undertaken by all re departments in seismic areas. 29.4.5.2 Mutual Aid Fire departments regularly practice mutual aid and, based on their daily experience, are quite condent of their ability to respond to emergencies. However, a major earthquake is a very rare occurrence, one that few senior re ofcers experience in their lifetimes. Earthquake planning for re departments is thus mandated. All re departments in a region should cooperate in the development of a regional earthquake plan with a detailed post-earthquake re element that should include procedures for rapid location and reporting of post-earthquake res, mutual-aid response to larger res, identication of alternative water supplies, and related measures. 29.4.5.3 Post-Earthquake Damage Surveys Most re services have a damage survey by re companies as part of their earthquake plan. This is very unrealistic, in that re companies self-dispatch or are dispatched immediately following the earthquake. An alternative to damage surveys by re companies needs to be developed. One possibility is that damage surveys should be performed by the police department, because they are not equipped to deal with res or building collapses, yet they know their neighborhoods, are mobile and reliable observers, and have reliable communications. Whether the damage survey is performed by police or others can be discussed, but tasking re companies with this role is unrealistic.
29.4.6 Water
Water for reghting is obviously a vital aspect of the re following earthquake problem. Earthquakes damage water systems and, thus, post-earthquake reliability of water supply is a necessary element of
29-58
post-earthquake re mitigation. Water supply systems in seismic regions should be assessed for their post-earthquake functionality. Several methodologies for this exist [Ballantyne and Crouse, 1997; ATC, 1992]. If the municipal water supply system cannot be made post-earthquake reliable, alternative water supplies are required. Alternative water supplies can be provided for in a number of ways, including: Identication of alternative water sources, such as rivers, lakes, swimming pools, or other supplies. This is a normal part of re department planning that is increasingly being overlooked as the re department increasingly takes on emergency medical treatment (EMT) as part of its central mission. Los Angeles Fire Department, in the response to the North Balboa Boulevard re, made excellent use of backyard swimming pools as an alternative source of reghting water. Thus, re ofcers need to be aware that the hydrant alone cannot be relied upon following a major earthquake, and they need to identify and regularly practice using alternative sources of water supply. It is remarkable that San Francisco burned down in 1906, despite being surrounded on three sides by the largest body of water on Earth. As part of its AWSS, San Francisco has installed saltwater hydrants (i.e., special connections to the bay for drafting) along its shorelines. It is even more remarkable that other similarly situated cities (e.g., Los Angeles, Tokyo, Osaka, Manila, Istanbul, etc.) have not done so. Acquisition of specialized portable pumps and large-diameter hoses similar to the HydroSub system recently acquired by the cities of Vancouver, Oakland, Vallejo, and Berkeley. It is not well understood that water can be practicably drafted only a vertical distance of about 26 ft (8 m). If a bridge, seawall, pier, or other access point is more than this distance above the water, a re engine cannot access the water it must get closer, which is often difcult. To overcome this problem, several systems such as the HydroSub are designed to lower a powered pumphead to the water and push rather than draft the water through the hose. In this manner, the HydroSub, for example, can access water more than 100 ft (33 m) below its location. In the United States, in recognition of the unreliability of municipal water supplies in earthquakes, high-rise buildings in higher seismicity zones are required to have on-site secondary water supplies. These requirements usually translate to a minimum of 15,000 gallons.12 Other high seismicity regions should adopt this requirement. Consideration can be given to development of special auxiliary high-pressure water systems similar to San Franciscos AWSS (discussed above) or Vancouvers DFPS (discussed below). Such highpressure systems can be justied only for major metropolitan areas in high seismic regions, but there is no shortage of these. It is noteworthy that Vancouver undertook construction of the DFPS in the 1990s.
29.4.7 Dedicated Fire Protection Systems

In the early twentieth century, as a result of major urban conagrations, a number of U.S. cities built special high-pressure systems to protect their central business districts. The 1871 Chicago, 1904 Baltimore, and 1906 San Francisco res are only the best known of many large urban conagrations that plagued the United States at that time. A number of factors contributed to these conagrations, but lack of reliable water supplies was a key factor that led to the construction of the high-pressure systems. For various reasons having to do with improved building stock and re service capability, most of these cities have since abandoned these systems, with San Francisco being the one major exception. Due to its high seismic
12 903.3.5.2 Secondary Water Supply. A secondary on-site water supply equal to the hydraulically calculated sprinkler demand, including the hose stream requirement, shall be provided for in high-rise buildings in Seismic Design Category C, D, E or F . The secondary water supply shall have a duration of not less than 30 minutes. (International Building Code, 2000 note that the 1997 UBC has similar requirements.)
29-59
Approved Layout Constructed to date

Burrard Inlet
Coal Harbour Pump Station

English Bay
False Creek Pump Station

FIGURE 29.40 Dedicated re protection system, City of Vancouver, B.C.
risk and the trauma of the 1906 re, San Francisco maintains its AWSS. A key aspect of San Franciscos ability to maintain and even extend this unique system is the fact that it is, by city charter, owned and operated by the re department. Recently, the City of Vancouver, as a result of analysis of its water supply systems seismic reliability (or lack thereof), embarked on the design and construction of a system similar to San Franciscos that Vancouver called its Dedicated Fire Protection System (DFPS). This section briey describes Vancouvers DFPS. The City of Vancouvers primary water supply is via the 1st and 2nd Narrows Crossings, which both cross liqueable deposits and were identied a decade ago as seismically vulnerable [EQE, 1990]. The only in-city water storage is at Little Mountain Reservoir in the center of the city the reservoir has been recently strengthened. Fully 77% of the domestic water distribution system in Vancouver is seismically vulnerable CI pipe more than 30 years old, while 30% of the larger steel trunk mains in the city are more than 70 years old and of highly vulnerable riveted construction [MTR, 1991]. The DFPS is a major asset for protection of the high-value district of Vancouver and consists of two seismically reliable pump stations (Coal Harbor and False Creek), each capable of 10,000 Igpm (see Figures 29.40 and 29.41). The two stations are connected by a loop around the CBD of 20-in.-diameter welded steel pipe (see Figure 29.41B). Special large-diameter hydrants connect to this loop, from which reghters can receive high-volume, high-pressure ows. The Vancouver Fire and Rescue Service is acquiring special large-diameter hose that, together with the DFPS, can be used to protect a 1000-ft radius from any hydrant off of this loop, thus blanketing the CBD.
29.5 Conclusion
The 1989 Loma Prieta and 1994 Northridge earthquakes, the 1991 Oakland Hills and 1993 Southern California res, Japans 1994 Kobe earthquake, and Turkeys 1999 Marmara earthquake all demonstrate the potential for res following earthquakes and the current, real possibility of conagration. In North America, all the elements that would hamper reghting capabilities are present: density of wooden structures, limited personnel and equipment to address multiple res, debris blocking the access of reghting equipment, and a limited water supply. Large cities in the United States and those in other countries, such as Tokyo, Osaka, Vancouver, Istanbul, and Wellington, are at substantial risk of re
29-60
(A)
(B)
FIGURE 29.41 Vancouver, B.C. DFPS. (A) False Creek pump station (foreground, nearing completion in 1995); (B) welded steel pipe under construction, 1995; (C) False Creek station proof test, 1995. (Photos: C. Scawthorn)
29-61
(C)
following earthquakes. A number of proactive steps can be taken to reduce this risk. Awareness, assessment, and mitigation of the risk of re following earthquakes are crucial if catastrophe is to be avoided.
Dening Terms
Conagration Large destructive re. In the re service, in the urban context, conagration usually
denotes a large re that spreads across one or more city streets, or a large re destroying a complex of buildings.
References
American Society of Civil Engineers. 1929. Report of the Special Committee on Effects of Earthquakes on Engineering Structures with Special Reference to the Japanese Earthquake of September 1, 1923, San Francisco. ATC (Applied Technology Council). 1992. A Model Methodology for Assessment of Seismic Vulnerability and Impact of Disruption of Water Supply Systems, Report ATC-251, funded by the Federal Emergency Management Agency. Prepared for Applied Technology Council by EQE International, C. Scawthorn and M. Khater, Applied Technology Council, Redwood City, CA. Ballantyne, D.B. and Crouse, C.B., Eds. 1997. Reliability and Restoration of Water Supply Systems for Fire Suppression and Drinking Following Earthquakes, Report NIST GCR 97730, Building and Fire Research Laboratory, National Institute of Standards and Technology, Gaithersburg, MD. Bond, H. 1946. Fire and the Air War, National Fire Protection Association, Boston, MA. Bowlen, Fred J., S.F.F.D. n.d. Outline of the History of the San Francisco Fire, Original manuscript. Chandler, C.C. et al. 1963. Prediction of Fire Spread Following Nuclear Explosions, Pacic Southwest Forest and Range Experimental Station, U.S. Forest Service, Berkeley, CA. Earthquake Engineering Research Institute. 1995. The Hyogo-ken Nanbu Earthquake, January 17, 1995, Preliminary Reconnaissance Report, Earthquake Engineering Research Institute, Oakland, CA. EQE. 1989. San Francisco Fire Department Facilities, Summary Report, prepared by EQE/AGS under contract to Department of Public Works, City and County of San Francisco, for the San Francisco Fire Commission. EQE Engineering. 1990. Emergency Planning Considerations, City of Vancouver Water System Master Plan, prepared under subcontract to CH2M-Hill, for City Engineering Department, City of Vancouver, B.C.
29-62
EQE International. 1995. The January 17, 1995 Kobe Earthquake: An EQE Summary Report, EQE International, San Francisco. Fleming, R. 1998. Analysis of Fire Sprinkler System Performance in the Northridge Earthquake, NIST-GCR98736, Building and Fire Research Laboratory, National Institute of Standards and Technology, Gaithersburg, MD. Freeman, J.R. 1932. Earthquake Damage and Earthquake Insurance, McGraw-Hill, New York. Hamada, M. 1951. On Fire Spreading Velocity in Disasters, Sagami Shobo, Tokyo (in Japanese). Hamada, M. 1975. Architectural Fire Resistant Themes, No. 21 in Kenchiku-gaku Taikei, Shokoku sha, Tokyo (in Japanese). Hamada, M., Wakamatsu, K., and Yasuda, S. 1992. Liquefaction-Induced Ground Deformations during the 1923 Kanto Earthquake, in Case Studies of Liquefaction and Lifeline Performance during Past Earthquakes, Vol. 1, Japanese Case Studies, M. Hamada and T.D. ORourke, Eds., NCEER Technical Report NCEER 920001, National Center for Earthquake Engineering Research, State University of New York, Buffalo, NY. Hansen, G. and Condon, E. 1989. Denial of Disaster: The Untold Story and Photographs of the San Francisco Earthquake and Fire of 1906, Cameron and Company, San Francisco. HAZUS. 2002. HAZUS99 SR2 Technical Manual, National Institute of Building Sciences, Washington, D.C. Heubach, W. 1997. The 1994 Northridge, California Earthquake Water Supply Effects, in Reliability and Restoration of Water Supply Systems for Fire Suppression and Drinking Following Earthquakes, Report NIST GCR 97730, D.B. Ballantyne. and C.B. Crouse, Eds., Building and Fire Research Laboratory, National Institute of Standards and Technology, Gaithersburg, MD. Horiuchi, S. n.d. Research on Estimation of Fire Disasters in Earthquakes, Disaster Prevention Section, Comprehensive Planning Bureau, Osaka Municipal Government, Osaka (in Japanese). Jensen, A. 1989. Report to the Board of Supervisors Concerning the Water Supply, Memorandum from acting general manager, San Francisco Water Department, to Public Utilities Commission, City and County of San Francisco, November 21. Johnson, G.S., Ascheim, M., and Sezen, H. 2000. Performance of Industrial Facilities, in The Marmara, Turkey Earthquake of August 17, 1999: Reconnaissance Report, MCEER Technical Report MCEER000001, C. Scawthorn, Ed., Multidisciplinary Center for Earthquake Engineering Research, State University of New York, Buffalo, NY. Johnson, G.S., Sheppard, R.E., Quilici, M.D., Eder, S.J. and Scawthorn, C. 1999. Seismic Reliability Assessment of Critical Facilities: A Handbook, Supporting Documentation and Model Code Provisions, Technical Report MCEER990008, Multidisciplinary Center for Earthquake Engineering Research, State University of New York, Buffalo, NY. Kanai, K. 1983. Engineering Seismology, University of Tokyo Press, Tokyo. Kimball, W.Y. 1966. Fire Attack! Command Decisions and Company Operations, National Fire Protection Association, Boston, MA. Kobayashi, M. 1979. A Systems Approach to Urban Disaster Planning, Ph.D. dissertation, Kyoto University, Kyoto, Japan (in Japanese). Kobayashi, M. and Korneld, L.M. 2002. Strategies for Reducing Post-Earthquake Fire Damage in Historic Wood Structures, in Proc. 7th National Conference on Earthquake Engineering, Boston, Earthquake Engineering Research Institute, Oakland, CA. Kobe Fire Department. 1995. Heisei 7 Hyogo ken Nanbu Earthquake Regarding Kobe City Damage and Fire Department Activities, Kobe City (in Japanese). Los Angeles City Fire Department. 1987. Earthquake Emergency Operational Plan. Los Angeles City Fire Department. 1994. 94 Earthquake After Action Report, March 9, 1994. Los Angeles County Fire Department. 1994. Northridge Earthquake Fact Sheet. Maher, T.J. 1933. Abstract of Reports Received Regarding the Earthquake that Occurred in Southern California on March 10, 1933, U.S. Coast and Geodetic Survey, Field Station, San Francisco.
29-63
Manson, M. 1908. Report on an Auxiliary Water Supply System for Fire Protection for San Francisco, California Board of Public Works, San Francisco. Mohammadi, J., Alyasin, S., and Bak, D.N. 1992. Investigation of Cause and Effects of Fires Following the Loma Prieta Earthquake, Report to the National Science Foundation funded under research grant BCS-9003557, Department of Civil Engineering, Illinois Institute of Technology, Chicago. Mroz, M.P. and Soong, T.T. 1997. Fire Hazards and Mitigation Measures Associated with Seismic Damage of Water Heaters, Report NIST GCR 97732, Building and Fire Research Laboratory, National Institute of Standards and Technology, Gaithersburg, MD. MTR. 1991. Pre-Design Report for Saltwater Pumping Stations and Dedicated Distribution Systems, prepared by MTR Consultants Ltd. and James M. Montgomery Consulting Engineers, for the Waterworks Engineer, City of Vancouver, B.C. National Fire Research Institute. 1995. Report on the Hyogo Ken Nanbu Earthquake and Urban Fires in Kobe City, Mitaka, Tokyo, Japan. Nawa, M. 1924. Quake Damage to Water Supplies of Japanese Railways, Engineering News-Record, 93(2.5), 986. NBFU (National Board of Fire Underwriters). 1905. Committee of Twenty, Report of National Board of Fire Underwriters on the City of San Francisco, CA, National Board of Fire Underwriters, New York. NBFU (National Board of Fire Underwriters). 1906. The San Francisco Conagration of April, 1906, S. Albert Reed, Special Report to National Board of Fire Underwriters Committee of Twenty, May 1906, National Board of Fire Underwriters, New York. NBFU (National Board of Fire Underwriters). 1933. Report on the Southern California Earthquake of March 10, 1933, National Board of Fire Underwriters, New York. Okamoto, S. 1984. Introduction to Earthquake Engineering, University of Tokyo Press, Tokyo. ORourke, T.D., Pease, J.W., and Stewart, H.F. 1992. Lifeline Performance and Ground Deformation during the Earthquake, in The Loma Prieta, California Earthquake of October 17, 1989: Marina District, U.S. Geological Survey Professional Paper 1551-F, T.D. ORourke, Ed., Strong Ground Motion and Ground Failure, T.L. Holzer, Coordinator, U.S. Government Printing Ofce, Washington, D.C. Porter, K.A., Scawthom, C., Honegger, D.G., ORourke, T.D., and Blackburn, F. 1991. Performance of Water Supply Pipelines in Liqueed Soil, in Proc. 4th U.S.-Japan Workshop on Lifeline Earthquake Engineering, Los Angeles, August 1923, National Science Foundation, Washington, D.C. Ree, A.C.D. 1941. Fire Department Operations during the Long Beach Earthquake of 1933, Bull. Seismol. Soc. Am., 31(1), 912. Rothermel, R.C. 1983. How to Predict the Spread and Intensity of Forest and Range Fires, General Technical Report INT-143, U.S. Forest Service, Ogden, UT. Scawthom, C. 1984. Simulation Modeling of Fire Following Earthquake, in Proc. 3rd U.S. National Conference on Earthquake Engineering, Charleston, SC, Earthquake Engineering Research Institute, El Cerrito, CA. Scawthorn, C. 1987. Fire Following Earthquake: Estimates of the Conagration Risk to Insured Property in Greater Los Angeles and San Francisco, All-Industry Research and Advisory Council, Oak Brook, IL. Scawthorn, C. 1997a. The 1923 Kanto, Japan, Earthquake, in Reliability and Restoration of Water Supply Systems for Fire Suppression and Drinking Following Earthquakes, Report NIST GCR 97730, D.B. Ballantyne and C.B. Crouse, Eds., Building and Fire Research Laboratory, National Institute of Standards and Technology, Gaithersburg, MD. Scawthorn, C. 1997b. The 1989 Loma Prieta, California Earthquake Water Supply Effects, in Reliability and Restoration of Water Supply Systems for Fire Suppression and Drinking Following Earthquakes, Report NIST GCR 97730, D.B. Ballantyne and C.B. Crouse, Eds., Building and Fire Research Laboratory, National Institute of Standards and Technology, Gaithersburg, MD.
29-64
Scawthorn, C. 1998. Reliability-Based Design of Water Supply Systems, in Proc. International Water Works Association Workshop on Anti-Seismic Measures on Water Supply, November, Japan Water Works Association, Tokyo. Scawthorn, C., Ed. 2000. The Marmara, Turkey Earthquake of August 17, 1999: Reconnaissance Report, MCEER Technical Report MCEER000001, Multidisciplinary Center for Earthquake Engineering Research, State University of New York, Buffalo, NY. Scawthorn, C. and Donelan, J. 1983. Fire-Related Aspects of the Coalinga Earthquake, in Report on the Coalinga Earthquake of May 2, 1983, Earthquake Engineering Research Institute, Berkeley, CA. Scawthorn, C. and Khater, M. 1992. Fire-Following-Earthquake: Conagration Potential in the Greater Los Angeles, San Francisco, Seattle and Memphis Areas, prepared for the Natural Disaster Coalition by EQE International, San Francisco. Scawthorn, C. and Khater, M. 1994. Fires Caused by Earthquakes: A Greater Threat than Many Realize, NFPA Journal, May/June, 8286. Scawthorn, C. and ORourke, T.D. 1989. Effects of Ground Failure on Water Supply and Fire Following Earthquake: The 1906 San Francisco Earthquake, in Proc. 2nd U.S.Japan Workshop on Large Ground Deformation, July, National Center for Earthquake Engineering Research, Buffalo, NY. Scawthorn, C. and ORourke, T.D. 1997. The 1906 San Francisco, California Earthquake, in Reliability and Restoration of Water Supply Systems for Fire Suppression and Drinking Following Earthquakes, Report NIST GCR 97730, D.B. Ballantyne and C.B. Crouse, Eds., Building and Fire Research Laboratory, National Institute of Standards and Technology, Gaithersburg, MD. Scawthorn, C. and Waisman, F. 2001. Assessment of Risk Due to Fire Following Earthquake, Lower Mainland, British Columbia, Report prepared for the Institute for Catastrophic Loss Reduction, Toronto, by EQE International, Oakland, CA. Scawthorn, C., Yamada, Y., and Iemura, H. 1981. A Model for Urban Post-Earthquake Fire Hazard, DISASTERS, Int. J. Disaster Studies Practice, 5(2), 125132. Scawthorn, C. et al. 1985. Fire-Related Aspects of the 24 April 1984 Morgan Hill Earthquake, Earthquake Spectra, 1(3), 675685. Scawthorn, C. et al. 1995. Fire-Related Aspects of the January 17, 1994 Northridge Earthquake, Earthquake Spectra, 11C, 419435. Scawthorn, C., Porter, K.A., and Blackburn, F.T. 1992. Performance of Emergency Response Services After the Earthquake, in The Loma Prieta, California Earthquake of October 17, 1989: Marina District, U.S. Geological Survey Professional Paper 1551-F, T.D. ORourke, Ed., Strong Ground Motion and Ground Failure, T.L. Holzer, Coordinator, U.S. Government Printing Ofce, Washington, D.C. Scawthorn, C., Cowell, A.D., and Borden, F. 1997. Fire-Related Aspects of the Northridge Earthquake, Report prepared for Building and Fire Research Laboratory, NIST-GCR-98743, National Institute of Standards and Technology, Gaithersburg, MD. Schiff, A., Ed. 1997. Northridge Earthquake: Lifeline Performance and Post-Earthquake Response, Report NIST GCR 97712, Building and Fire Research Laboratory, National Institute of Standards and Technology, Gaithersburg, MD. Schussler, H. 1906. The Water Supply of San Francisco, California, Spring Valley Water Company. Sekizawa, A. 1997. Post-Earthquake Fires and Fireghting Activities in the Early Stage in the 1995 Great Hanshin Earthquake, paper presented at 13th Meeting of the UJNR Panel on Fire Research and Safety, March 1320, 1996, NIST Tech. Report NISTIR 6030, Building and Fire Research Laboratory, National Institute of Standards and Technology, Gaithersburg, MD. Sepponen, C. 1997. Fire Department Emergency Response, in Northridge Earthquake: Lifeline Performance and Post-Earthquake Response, Report NIST GCR 97712, A. Schiff, Ed., National Institute of Standards and Technology Building and Fire Research Laboratory, Gaithersburg, MD, chap. 16. SFFD (San Francisco Fire Department). 1990. Report on the Operations of the San Francisco Fire Department Following the Earthquake and Fire of October 17, 1989, San Francisco Fire Department.
29-65
Steinbrugge, K.V. 1982. Earthquakes, Volcanoes and Tsunamis, Skandia-America Group, New York. Takata, A. and Saizberg, F. 1968. Development and Application of a Complete Fire-Spread Model, Vols. 1 to 4, Ofce of Civil Defense, Illinois Institute of Technology Research Institute, Chicago. Usami, T. 1996. Nihon Higai Jishin Soran (List of Damaging Japanese Earthquakes), University of Tokyo Press, Tokyo. USGS. 1907. The San Francisco Earthquake and Fire of April 18, 1906 and Their Effects on Structures and Structural Materials, Bull. No. 324, U.S. Geological Survey, Washington, D.C. Woycheese, J.P., Pagni, P.J., and Liepmann, D. 1999. Brand Propagation from Large-Scale Fires, J. Fire Protection Eng., 10(2), 3244. Youd, T.L. and Hoose, S.N. 1978. Historic Ground Failures in Northern California Associated with Earthquakes, Professional Paper 993, U.S. Geological Survey, Washington, D.C.
Further Reading
Comprehensive works on re following earthquake are rare. Steinbrugge [1982] provides a good qualitative discussion, while Scawthorn [1981] and the AIRAC report [Scawthorn, 1987] provide the rst integrated model of the problem and probably still the best overall discussion, followed by HAZUS [2002]. Sekizawa and other Japanese researchers cited in the references are well worth reading.

Fire Following Earthquakes

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Fire Following Earthquakes

Transféré par

Droits d'auteur :

Formats disponibles

29