Geotechnical Report from June 23, 2001 Peruvian Earthquake
2.0 SEISMOLOGICAL ASPECTS AND GROUND MOTION
The June 23 shock occurred in the interface between the Nazca and the South American plate, with the hypocenter located approximately 200 km southeast of the source of a Mw 7.7 earthquake that occurred in November of 1996. Preliminary seismological observations indicate that the June 23 earthquake appears to have involved rupture of part of the plate-boundary segment that produced an earthquake of magnitude approximately 9.0 in 1868 (USGS 2001). The June 23, 2001 earthquake is an inter-plate subduction event between the oceanic Nazca plate and the overriding South American continental plate as depicted in Figure 2.1. The angle of subduction along the southern coast of Peru is typically 12 to 15 degrees from horizontal as compared to the steeper angle shown in Figure 2.1. Similar subduction zones can also occur in the Pacific Northwest of the United States and in the Aleutian-Alaskan subduction zone.
Figure 2.1: Schematic diagram of a subduction zone. Picture obtained from the USGS site.
Various magnitude values have been reported (Mw 8.4 by the Harvard CMT, Mw 8.3 by the USGS, Mw 8.2 by the Earthquake Information Center, Tokyo). The USGS moment tensor solution can be found at http://neic.usgs.gov/neis/bulletin/01_EVENTS/010623203313/Q0106232033.html. The earthquake had a shallow focal depth; however, because a large portion of the plate interface ruptured, it is difficult to estimate a single representative focal depth value. Values reported for epicentral depth are 8 km (USGS), 25.7 km (Harvard CMT), and 30 km (EIC, Tokyo).
The main shock of June 23,2001, was recorded only by one ground motion station in the city of Moquegua. The ground motion station at Moquegua is owned by CISMID (Peru Japan Center for Seismological Investigations and Disaster Mitigation) and is located at S17.1868 W70.9287. The accelerometer is located in a flat region in the south side of an east-west trending river valley. The closest topographic feature is the valley wall located about 100 m from the accelerometer. The instrument is housed in a wooden shelter next to a one-story reinforced concrete building (Figure 2.2). A concrete block wall located immediately behind the instrument shelter collapsed during the earthquake without damaging the shelter. The strong motion station is presumably located on top of a coarse gravel alluvial soil of unknown depth. The characteristics of the soil, however, are also unknown.
Figure 2.2: Wood shelter housing the Moquegua strong motion station.
According to a printed report from CISMID, the recorded peak ground accelerations (PGAs) were 0.30 g and 0.22 g in the east-west and north-south directions, respectively. The recorded vertical PGA was 0.16 g. Figure 2.3 illustrates the predictions of the Youngs et al. (1997) relationship for inter-slab earthquakes using an approximate distance of 30 kilometers (estimated using a shallow rupture plane at an approximate depth of 29 km, consistent with initial depth estimates by the Peruvian Geophysical Institute, IGP). Further ground motion parameters can be obtained at the CISMID homepage (http://www.cismid.uni.edu.pe).
Figure 2.3: Comparison of predicted spectral accelerations (5% damping) with the recorded spectral accelerations at the Moquegua strong motion station (east-west direction). Note that the recorded spectral accelerations are approximated from a printed report by CISMID. The PGA is illustrated as the spectral acceleration at a period of 0.01 seconds. Attenuation relationships are given with a one standard deviation band. Observe that the recorded motion is within one standard deviation of the soil attenuation relationship. The predominant period of the recorded motion, however, is larger than that predicted by the attenuation relationship.
The peak ground accelerations predicted by the Youngs et al. (1997) attenuation relationship for inter-slab earthquakes are shown in Figure 2.4. Approximated distances for the main urban centers using a shallow dipping rupture plane render approximate median accelerations of 0.12 g, 0.15 g and 0.24 g for rock sites in Tacna, Arequipa, and Moquegua, respectively. For soil sites, the estimated median PGA are 0.20 g, 0.24 g, and 0.36 g for Tacna, Arequipa, and Moquegua, respectively. The recorded PGA in Moquegua matches with the predicted accelerations. Damage levels in Tacna, however, are larger than those in Arequipa. This must be reviewed further using more complete rupture plane solutions and including site amplification considerations.
Figure 2.4: PGA predictions by the Youngs et al. (1997) attenuation relationship.
It is interesting to note that damage levels in towns near the epicenter, such as Ocoña, were significantly lower than those observed in Moquegua and even in Tacna. Larger ground motions in Moquegua might be related to the slip distribution in the fault plane. A preliminary fault-slip solution published by the Earthquake Information Center in Tokyo shows a region of high slip located 150 km southeast of the epicenter (Figure 2.5). Similar observations linking ground motion parameters to slip distribution and stress drop were made regarding near-field ground motions in the Kocaeli and Duzce earthquakes (Anderson et al. 2000).
Figure 2.5: Preliminary slip distribution by Kikuchi and Yamanaka (http://wwweic.eri.u-tokyo.ac.jp/EIC/EIC_News/105E.html). Mean dislocation is 2.8 m with maximum value of 4.5 m.
Typical brick and concrete dwelling construction is similar in all the cities visited. These buildings are typically 1 to 3 stories high. Roofs and intermediate floors consist of reinforced concrete slabs (with hollow brick infill to decrease dead weights), supported on 25-cm-wide brick bearing walls, with no structural continuity between slab and wall. Non-bearing walls are generally 15-cm wide. Generally, in order to provide confinement to the walls, reinforced concrete columns are placed at the wall corners. Institutional buildings (such as schools) and buildings in excess of three stories are usually of the reinforced concrete frame type.
Buildings of the types described above are predominant in Tacna, Ilo, Arequipa, Camaná, and Ocoña. Damage distribution varied significantly from one city to another, and also between different areas of a given city. Observations regarding roads and mining facilities are presented in subsequent sections.
- – Tacna: Damage was minor in the central and southern sections of Tacna (typically cracks in non-bearing walls), and moderate to severe in the northern part (Cuidad Nueva and Alto de la Alianza).
- – Moquegua: In addition to brick and concrete buildings and dwellings, Moquegua has extensive sections with many adobe dwellings. The downtown area also has many colonial structures, typically over 100 years old. Damage was moderate to the brick and concrete structures, and severe to total collapse to the adobe and colonial construction.
- – Ilo: Damage was minor, concentrated on the higher portion of the city (Pampa Inhalámbrica).
- – Arequipa: Damage throughout the city was minor, with exception of old churches and other religious monuments that suffered severe damage, including the cathedral, a UNESCO World Heritage site located in Arequipa’s historic section.
- – Camana: Damage throughout the city was minor. Damage was total in the adjacent beach resorts of Cerrillos and La Punta that were swept by a tsunami.
- – Ocoña: Damage throughout the city was minor to moderate.
- – Youngs , Chiou, Silva, and Humphrey (1997). "Strong ground motion attenuation relationships for subduction zone earthquakes," Seismological Research Letters, 68(1), 58-85.
- – Anderson, J. et al. (2000). "Implication for Seismic Hazard Analysis,", in the 1999 Kocaeli, Turkey, Earthquake Reconnaissance Report, Earthquake Spectra, 16A, 113-137.