Special San Fernando Earthquake Edition

SEISMOLOGIC AND CRUSTAL MOVEMENT INVESTIGATIONS OF THE SAN FERNANDO EARTHQUAKE

By Roger Greensfelder

SEISMOLOGY

Fault mechanism and aftershocks

The epicenter of the main quake was in the vicinity of Magic Mountain (map, page 62), about 10 km (6 mi.) north-northeast of Sylmar [2,3]. No seismograph stations were close enough to the epicenter to permit reliable determination of the quake's focal depth ; however a depth of about 12 km (7 mi.) is considered reasonable. The thousands of after-shocks that were recorded in the first few days following the earthquake were scattered over an area extending from the main shock's epicenter south to Sunland and southwest to Chatsworth [2,4] (map, page 62). About two dozen portable seismographs were deployed throughout the aftershock region during the week following the earthquake. Data collected with these instruments will permit the precise determination of locations (geographic position and depth) and focal mechanisms of many hundreds of aftershocks. This information will help to determine the position and mechanics of faulting associated with the main shock.

Preliminary solutions of the focal mechanism prepared by the National Center for Earthquake Research of the U. S. Geological Survey and the California Institute of Technology [2,3] indicate that the San Fernando earthquake was generated by oblique-slip reverse faulting on a plane probably dipping 50° to 60° N and striking between N64°W and due west. The hanging wall (upper San Gabriel Mountains block) moved southwesterly or up and west relative to the foot-wall (lower San Fernado Valley block); the ratio of reverse to left-lateral slip was between 1:1 and 2:1 (diagram, page 64). Fault rupture reached the earth's surface, and the observed surface displacements represent a continuation of reverse faulting along the south flank of the San Gabriel Mountains which began less than 2 million years ago [5].

On the basis of the above information, we can conclude that fault rupture began at a point (the hypocenter) about 10 km (6 mi.) NE of Sylmar and 12 km (7 mi.) deep, and progressed upward toward the south and southwest, apparently spreading out along several faults as it approached the surface as suggested by the presence of multiple, near-parallel fault traces in a zone about 1 km wide (map, page 76 and 77).

Local and regional seismicity

Seismic activity (seismicity) in the Newhall-San Fernando region, and throughout the western San Gabriel Mountains, has been remarkably low in comparison with the rest of southern California during recent times. A strain release map of southern California for the period 1934 to 1963 [6] shows that this particular area had seismicity equivalent to less than four magnitude-3 shocks per 100 square km; other surrounding areas have shown activity more than 1000 times as intense, chiefly in the form of a single earthquake of magnitude 7 or greater.

Only one strong, destructive earthquake is known to have occurred in historic time previously in the San Fernando-Newhall area. This took place in 1593 near Pico Canyon, about 5 miles southwest of Newhall, causing landslides, rock-falls, and ground fissures; in Newhall and Saugus, chimneys were knocked down, and an adobe house was destroyed [1,5]. The shock was felt strongly in Mojave, San Bernardino, and Ventura, lightly in Los Angeles and Santa Ana, and was perceptible in San Diego. An intensity f ("Intensity" refers to the strength of earthquake ground motion at a particular locality, and ratings are based on observable effects on man-made structures and the ground. It is not to be confused with magnitude, which is an instrvmental rating proportional to an earthquake's energy, and is indegendent of the point of measurement.)(Modified Mercalli Scale) of VIII to IX was assigned to this earthquake by Townley and Allen, apparently on the basis of its ground effects and felt area, as the reported minor to moderate structural damage suggests intensity VII to VIII. Severe structural damage caused by the San Fernando earthquake corresponds to an intensity of IX, and locally X, and so it appears that the 1893 shock was the lesser of the two.

Thus the San Fernando earthquake occurred in an area which has had relatively low seismicity and was not caused by movement on a major, historically active fault, such as the San Andreas or one of its branches. In fact, portions of the fault associated with this shock were not previously mapped, although they might have been partly inferred from the abrupt, linear scarp along the southern base of the San Gabriel Mountains between Big Tujunga and Lopez Canyons (see map). Several other destructive California earthquakes have occurred on faults which were either unmapped or not known to be seismically active. In 1892, the towns of Vacaville, Dixon, and Winters in the Sacramento Valley were severely damaged by an earthquake on an unmapped fault. In 1954, Eureka suffered moderate damage from a magnitude 6.6 shock, also on an unmapped fault. And in 1940, the Imperial Valley earthquake (magnitude 7.1), which caused moderate to severe damage in several towns, took place on a fault which was unknown before it ruptured and broke the surface at that time. The Kern County earthquake of 1952 (magnitude 7.7) took place on a tentatively mapped fault (the White Wolf) which was not known or suspected to be seismically active; other potentially destructive shocks have occurred similarly.

It is therefore evident that faults whose existence or seismic activity is unknown can be the source of destructive earthquakes. Indeed, since the great 1906 earthquake on the San Andreas fault, many destructive earthquakes in California have occurred on such faults. This clearly points out the need for the identification of mapped faults, and of those whose existence is only suspected, which could give rise to destructive earthquakes or surface rupture in and near urban areas. Better, more comprehensive information on the accumulation of strain could provide a better estimate of the frequency and magnitude of earthquakes in particular localities and on particular faults than can be obtained strictly on the basis of historical seismicity. While the historical sample of earthquake occurrence in a large region (tens of thousands of square kilometers) may be an adequate index of its future seismicity, apparently small regions (several thousand square km or less) often have not exhibited a valid sample of their long term seismic activity during historic time. More detailed studies of geomorphology (geologic analysis of topography) may allow this record to be extended far back into prehistoric time.

Magnitude and maximum ground motion

Many workers have studied the relationship between earthquake magnitude and intensity of ground motion, based on both instrumental and non-instrumental data; the rate of attenuation of ground motion with distance as a function of magnitude is a matter of special interest to structural engineers. Because very few strong-motion accelerograph recordings have been obtained near earthquake epicenters in the past, say within 25 km (15 mi.), we have little information on maximum acceleration as a function of magnitude.

The accelerograph at Pacoima dam, located about 5 km (5 mi.) south of the epicenter of the San Fernando earthquake, recorded a maximum horizontal acceleration somewhat greater than that of gravity (1.05 g on the S16°E horizontal component) [9]. This is the greatest acceleration ever recorded during an earthquake, and is from two to ten times that expected at the epicenter of a magnitude 6.6 shock. The Pacoima dam accelerograph rests on a reinforced concrete abutment founded in crystalline basement rock. Therefore, the accelerogram should closely represent hard-rock motion, which is typically smaller than that observed on alluvium or soft sediments. It has been previously estimated that a magnitude 6.5 shock should produce an acceleration of 0.1 g on "better ground" [30].

The Parkfield earthquake of 1966, which had a magnitude of 5.5, yielded a peak acceleration of 0.5 g at a distance of about 100 meters from the fault rupture near Cholame. Again, this acceleration is much greater than had been expected from a magnitude 5.5 shock. Apparently moderate earthquakes may result in quite strong and locally very damaging ground motion near their epicenters. In terms of potential damage, perhaps the chief difference between a shock of magnitude 6 and one of magnitude 8 lies in duration of shaking and size of the area affected, rather than in the intensity of ground motion very near the epicenter.

Structural damage

Death and injury in an earthquake are caused primarily by partial or total collapse of man-made structures. In the San Fernando earthquake, the number of deaths (65), of people sustaining serious injuries (over 1,000) and amount of major structural damage was almost entirely confined to portions of San Fernando and Tujunga Valleys, and the Newhall-Saugus area (map, page 62). Beyond these areas, damage was minor to moderate and mostly limited to old, unreinforced masonry structures as far away as downtown Los Angeles, 40 km (25 mi.) south of the epicenter.

Zones of extreme damage are closely correlated with those of surface faulting, lurching, compression-ridging, cracking, and other "permanent" ground deformation. Intense lurching, heaving, and cracking of ground generally parallel the contact of alluvium and bedrock along the southwestern base of the San Gabriel Mountains (see map, page 62).

At Olive View Hospital, a number of newly constructed reinforced concrete buildings were very seriously damaged or collapsed, causing the death of three persons. A two-story building "pancaked," its second floor dropping to ground level. Fortunately, no one was on the ground floor. Four five-story stairwell-wings pulled away from the main building, three of them toppling over (photo, on cover). Nevertheless, considering that several hundred people were inside the main building at the time of the earthquake, the building was relatively successful in terms of safety.

Forty-four people were killed at Sylmar Veterans Hospital when two unreinforced tile masonry buildings collapsed; these structures were built more than 40 years ago, before the general adoption of modern building codes which require earthquake-resistant design.

The majority of modern residential and commercial structures in the Newhall and northern San Fernando Valley areas escaped serious harm; however, it has been estimated that more than 1300 buildings and 1700 mobile homes suffered major damage. Old, unreinforced masonry walls, and new chimneys and concrete block fences were knocked down or cracked over a large portion of San Fernando and Tujunga Valleys, and in downtown Los Angeles. Had the earthquake occurred a few hours later, many people might have been seriously injured or killed by bricks falling into the street.

Schools built since the passage of the Field Act in 1933, which reqires earthquake resistant design and construction of school bulldings, performed well in the earthquake. However, a number of schools built before 1933 were hard hit, and in some cases had to be condemned.

New high-rise buildings in downtown Los Angeles and southern San Fernando Valley suffered no structural damage. However, this cannot be taken to be proof that these structures will not be damaged under maximum expectable earthquake ground motion at their critical periods of vibration. These periods are a function of a building's height and rigidity, and can roughly be said to fall in a range of from 1 to 5 seconds. The San Fernando shock did not produce very strong motion in this period range compared with that expected from a magnitude 7- to 8- earthquake. However, the numerous strong-motion accelerograph recordings obtained for the first time in many new high-rise buildings should be very useful in estimating their response to stronger ground motion.

The old, hydraulic earth-fill dam on Lower Van Norman Lake suffered major sliding on its upstream face, posing such a threat of failure that a large area below had to be evacuated for a four-day period. People were allowed to return to their homes only after the water level in the lake had been lowered and careful inspection had shown that immediate danger of failure had passed. The dam on Upper Van Norman Lake was also severely damaged, and it, as well as the lower dam, will have to be rebuilt.

Modern freeway roads and bridges in the Sylmar area were severely damaged. A number of bridge spans collapsed; one collapsing span killed two men in a pickup truck. Eight-inch thick concrete pavement slabs were compressed into ridges and thrust over one another, apparently as a result of land-shortening in a north-south direction. Lateral spreading affected some sections of freeway underlain by fill material. Practically all bridge collapses were in the Interstate Highway 5/210 interchange and pavement was ridged between there and the Interstate Highway 5/405 interchange, about 2 miles to the south. A fault disrupted the Foothill Freeway (1-210) just west of MacLay Street; preliminary measurements indicate 5 feet (1.5 m) vertical and 41/2 feet (1.3 m) horizontal (shortening) displacement across the main zone of rupture [12].

Utility lines--gas, water, sewer, telephone and electricity-were disrupted in the areas of most intense ground motion. Pipes failed where they crossed the zones of surface faulting.

Oil field facilities and related structures were relatively little damaged. There was minor damage to tanks, roads, pipelines, and a few wells in Aliso Canyon, Cascade, Castaic, New-hall, Newhall-Potrero, Oak Canyon, Placerita, and Ramona oil fields.

CRUSTAL MOVEMENT

Preliminary measurements of fault displacements across several traces indicate that the cumulative oblique reverse slip across the 1 km (0.6 mi.) wide zone exceeds 1 m (3 ft.). Left-lateral displacement appears to be smaller than vertical offset throughout most of the fault zone. This indicates that a western block of the San Gabriel Mountains moved south to south-westward and upward relative to the San Fernando Valley block (diagram, page 64).

Precise regional (geodetic) and local survey measurements are required to determine the magnitude and direction of crustal block movements. The City and County of Los Angeles, the U. S. Geological Survey (USGS), the National Ocean Survey (NOS), the California Division of Mines and Geology, and the State Division of Highways are engaged in triangulation, leveling, and laser ranging programs for this purpose.

In order to determine vertical movement, the City and County of Los Angeles will resurvey a first-order level line from the Los Angeles City Hall north to Newhall or Gorman, and another line east from San Fernando toward Pasadena. These lines were last surveyed in 1969, as a part of the Southern California Cooperative Level Net. In order to detect both initial and possible continuing vertical movements in the fault zone, the USGS is making repeated observations on several short level lines. Two lines are located in Lopez and Little Tujunga Canyons, and another goes through the town of San Fernando.

Horizontal movements will be determined by first-order triangulation and high-precision laser-ranging programs which are now underway. The area covered extends from the San Andreas fault on the north to the Santa Monica Mountains on the south. The NOS and County of Los Angeles are performing large-area triangulation, reinforced by Geodimeter (a laser ranging device) distance measurements on selected lines. A small trilateration network spanning the fault zone in the Sylmar-San Fernando area has been established and is currently being monitored by the USGS, using a Geodolite (also a laser ranging device). The California Division of Highways is re-observing triangulation networks which extend along Interstate Highways 5 and 210, and State Highway 14, from San Fernando Valley to the Newhall area.

A number of small survey figures were set up during the week following the earthquake in order to detect a possible continuation of fault movements, or creep, at the surface. At this writing, it appears that very little creep has taken place. At a site about 1 mile east of Lopez Canyon, vertical movement of about 5 cm (3 in.) was detected between February 11 and February 14; however, it is uncertain how much of this was due to fault creep [11] As yet, no significant movement at other sites has been reported.

It is of considerable interest to discover whether or not crustal movements accompanying the earthquake produced a measurable or significant strain change in the vicinity of the San Andreas fault from Gorman, 56 km (35 mi.) to the northwest, southeastward to San Bernardino. This portion of the San Andreas fault has apparently been "locked," showing essentially no seismic or aseismic slip since the great (magnitude 8+) 1857 Fort Tejon earthquake. It is believed that the fault slipped more than 6 m (20 ft.) and surface rupture extended more than 320 km (200 mi.) from Cholame Valley to Cajon Pass during that earthquake. Many geologists and seismologists believe that a great earthquake (magnitude 8+) is likely to occur on this portion of the San Andreas fault, and in the not too distant future-perhaps before the end of this century.

In consideration of this problem, the Division of Mines and Geology has reobserved 12 Geodimeter lines near or crossing the San Andreas fault between Gorman and San Bernardino. Eight of the lines form a small closed figure near Gorman, and their re-observation will allow calculation of the strain change in it since the last observation in 1966. The other four lines do not form closed figures, and re-observation of them will allow only rough estimation of strain change.

At this writing, the lengths of 10 of the above lines have been calculated from the field observations; however time has not permitted more than a cursory interpretation of the data. Nevertheless, it is significant that all lines except one are now longer than when last observed, irrespective of their azimuth. Three lines which cross the San Andreas fault between Little-rock, about 13 km (5 mi.) southeast of Palmdale, and San Bernardino were last observed in December 1970: one line (Tenhi-Ward) near Littlerock and another near Cajon Pass (Phelan-Sevaine) had become 2 cm (0.8 in.) longer when reobserved about 10 days following the earthquake; the line north of San Bernardino (Sevaine-Strawberry) showed no change. The change of +2cm on lines Tenhi-Ward and Phelan-Sevaine suggests regional dilatation (expansion) on the order of 1 part per million (ppm), and no movement on the fault. Another line (Sawmill-Thumb) which crosses and is nearly perpendicular to the San Andreas fault about 5 miles west of Lake Hughes is about 22 cm longer (1 ppm) than it was during the period 1960-1963. Preliminary data on five out of eight lines in the Gorman figure suggest dilatation on the order of 2 or 3 ppm since 1966.

On the basis of the above information, it appears that the San Fernando earthquake may have caused a small relaxation of stress in the region of the San Andreas fault between Gorman and Cajon Pass. The data do not indicate either slip or a significant change in shear strain on the fault.

REFERENCES

1. Allen, C. R., St. Amand, P., Richter, C. F., and Nardquist, J. M., 1965, Relationship between seismicity and geologic structure in the southern California region: Bulletin of the Seismologicol Society of America, vol. 55, no. 4, p. 782.

2. Lee, W. H. K., 1971, personal communication.

3. Nardquist, J. M., 1971, personal communication.

4. Allen, C. R., 1971, personal communication.

5. Oakeshatt, G. B., 1958, Geology and mineral deposits of the San Fernando quadrangle, California: California Division of Mines Bulletin 172, p. 92.

6. Allen et al., op. cit., plate 1.

7. Townley, S. D., and Allen, M. w., 1939, Descriptive catalog of earthquakes of the Pacific Coast of the United States, 1769 to 1928: Bulletin of the Seismological Society of America, vol. 29, no. 1, p. 92.

8. Perrine, C. D., 1894, Earthquakes in California in 1893, U.S. Geological Survey Bulletin 114. p. 13-16.

9. Hudson, D., 1971, personal communication.

10. Gutenberg, B., and Richter, C. F., 1956, Earthquake magnitude, intensity, energy, and acceleration (second paper): Bulletin of the Seismology Society of America, vol. 46, no. 2, p. 130.

11. Burford, R. o., 1971, personal communication.

12. Parmer, A., 1971, personal communication.

ACKNOWLEDGEMENTS

Thanks are due to Dr. Clarence Allen, Dr. Donald Hudson, and Mr. John Nordquist of the California Institute of Technology, and to Dr. Willie Lee and others of the U. S. Geological Survey, who kindly provided the basic seismological data presented in the report; also to Dr. Robert Burford and Dr. Robert Nason, who provided information on fault creep and displacements. Thanks are also extended to Mr. Al Parmer of the California Division of Highways for summarizing the damage to highways and bridges.