Journal of Seismology 6: 279 283, 2002. 2002 Kluwer Academic Publishers. Printed in the Netherlands. 279 The Arequipa (Peru) earthquake of June 23, 2001 H. Tavera 1, E. Buforn 2,I.Bernal 1, Y. Antayhua 1 & L. Vilacapoma 1 1 Centro Nacional de Datos Sismología, Instituto Geofísica del Perú, Peru; 2 Dpto de Geofísica y Meteorología, Universidad Complutense, Madrid, Spain Received 28 September 2001; accepted in revised form 2 November 2001 Key words: Arequipa earthquake, focal mechanism, intensity, tsunami Abstract The Arequipa earthquake of 23 June 2001 has been the largest earthquake (Mw = 8.3)occurred in the last century in southern Peru with a maximum intensity of VIII (MM scale). Focal mechanisms of main shock and three larger aftershocks have been studied, showing thrusting solutions for main shock and two aftershocks and normal motion for the event of July, 5. The rupture area has been obtained from distribution of aftershocks. The occurrence of the Arequipa earthquake is related with the convergence process between the Nazca and South America plates. Introduction On June 23, 2001 a catastrophic earthquake (Mw = 8.3) occurred at 20h 33m 14.2s off the southern coast of Peru, 82 km at NW of Ocoña city, in the Arequipa department. The epicenter was located at 16.20 S, 73.75 W, with focal depth of 38 km (Instituto Geofísico de Peru, IGP). Over 74 people were killed and 2.689 injured, with more than 80 villages severely damaged in the departments of Arequipa, Moquegua and Tacna (Peru) and Arica and Iquique (Chile) (Figure 1). About 217400 persons suffered the effects by the earthquake. More than 35000 houses were damaged and 17580 totally destroyed. The earthquake was followed by a local tsunami with wave height of 8 10 m in Camana city, that killed 23 people and a total of 64 persons missing. This earthquake is related to the subduction of the Nazca plate under the South America plate, with a convergence velocity of 8 10 cm/yr (Minster and Jordan, 1978). This convergence process has also caused the large earthquakes (M>7) in the central and southern regions of Peru in 1746, 1868, 1940, 1942, 1966, 1974 and 1996. The Arequipa earthquake is located 250 km SE of the 1996 earthquake (November 12, Mw = 7.7) and at the northern border of the fault associated with the large earthquake of 1868 (Mw = 9.0), which generated a tsunami with maximum height tides of 16 m. The southern region of Peru is shaken frequently by large earthquakes located between the Peru- Chilean Trench and the coast, many of which also generate tsunamis. Focal mechanisms of these earthquakes generally correspond to reverse motions with nodal planes oriented in a NNW-SSE direction and a near horizontal plane dipping to the east, that corresponds to the subducted plate (Tavera and Buforn, 2001). Isoseismal map The intensity map for the Arequipa earthquake (Figure 1) has been compiled using information from more than 60 towns. The isoseismal lines show an elliptical distribution. A maximum intensity of VIII (MM scale) was reached in the cities of Ocoña, Camana and Mollendo. The isoseismal IV has the major axis of 1150 km oriented in a NW-SE direction parallel to the coast. In Arequipa city many churches and historical monuments have been severely damaged, in the rest of the Arequipa, Moquegua and Tacna departments, major damage correspond to old adobe and quinche buildings. On the Panamericana Sur road (the main road in this area), several cracks developed, oriented parallel to the coast and several landslides occurred. Similar effects were observed in the cities of Arica
280 Figure 1. Isoseismal map for the Arequipa earthquake. The star shows the epicenter of the main shock. and Iquique, in Chile. In the interior of the continent, the earthquake was felt in Cuzco (Peru) and La Paz (Bolivia), 300 and 600 km from the epicenter respectively. The occurrence of a tsunami was not expected due to the location of the epicenter close to the coastline. However, the sea level increased 30 40 cm from Ocoña to Molledo (Figure 1). After 15 20 minutes, the sea retreated 100 200 m from the coast line between Chala and Mollendo (200m in Camana and Ocoña, 150 m in Chala y 100 m in en Ilo). This resulted in a local tsunami, with waves of 7 8 m, that produced damage in the towns located in the coast of Canamá, with floods in an area of 1.6 km 2,parallel to the coast This local tsunami may be related to the complexity of the rupture that produced the earthquake.
281 Table 1. Hypocentral coordinates of earthquakes studied (taken form NEIC) and focal mechanisms solutions (this study) Date Time Epicenter H Mw T axis P axis N Score Lat. Long. 062301 20:33:14.2 16.20 73.75 32 8.3 67±4 81±41 24±4 259±16 52 1.00 062301 20:33:13.0 16.15 73.40 33 062601 04:18:31.6 17.73 71.34 31 6.8 63±8 75±47 27±8 254±41 35 1.00 070501 13:53:49.7 15.56 73.45 69 6.6 15±7 56±10 50±9 163±15 21 0.95 070701 09:38:43.0 17.38 71.78 26 7.5 63±6 81±46 26±5 246±28 49 1.00 = Hypocentral determination by NEIC. H = depth in km. = plunge. = trend. N = number of observations. Score = proportion of correct observations. Distribution of aftershocks and focal mechanisms The hypocentral parameters for the main shock and the three larger aftershocks, are shown in Table 1 (NEIC). Focal depths have been obtained from arrival times of P and pp waves recorded on broad-band stations at teleseismic distances (greater than 20 ). A total of 15 seismograms have been used for main shock, and between 10 and 12 for the aftershocks. The focal depths shown in Table 1, have been obtained from IASPEI tables, assuming a P wave velocity on the focus of 8 km/s. These values are comprising between 26 km (for event of 7 July) to 69 km, for the event of July, 5. Shocks with epicenters offshore (Figure 2) have similar depths (between 26 32 km), whilst deeper depth corresponds to the event located inside the continent (July, 5). Distribution of 150 aftershocks (M 4.0), which occurred in the 18 days after the main shock are plotted in Figure 2 as white circles. These events, located using stations of the Seismic Network of Peru, are distributed in a rectangle of 370 160 km 2 oriented in a SE direction with the main shock in the northern end and the first aftershock (June 26) in the southern. The aftershock of July 5, may be considered situated at the limit of the rupture area inside the continent. Focal mechanisms of the main shock and three larger aftershocks have been estimated from polarities of P wave (Brillinger et al., 1980; Udías and Buforn, 1988). Data used correspond to stations at regional and teleseismic distances, with a maximum of 52 observations for the main shock and a minimum of 21 for the aftershock of July 5 (Table 1 and Figure 3). The solution obtained for the main shock corresponds to thrusting fault with nodal planes oriented in a N-S direction and the pressure axis near the horizontal and oriented in ENE-WSW direction. One plane is nearly vertical (68 ) and the second plane is nearly horizontal and dipping to the ENE. Similar solutions have been obtained for the aftershocks of June 26 and July 7. However, the aftershock of July 5 has a different focal mechanism, corresponding to normal faulting, with planes oriented in NS and NW-SE directions and dipping to the west and NE respectively. These solutions are similar to those obtained from Harvard CMT. Discussion The Arequipa earthquake is the largest shallow (h<60km) shock that has occurred in the last century in southern Peru and it may be associated with the subduction process of the Nazca plate under the South America plate. Aftershocks distribution decribe a rectangular area of 370 160 km 2, parallel to the coast, that may be considered as the dimensions of the rupture. This rupture area is located near the southern part of the aftershock area for the 1996 event (96.12.11), with a possible small seismic gap between both rupture areas (Figure 2). The geometry of the rupture area for the Arequipa event, delimited by the aftershocks, is in agreement with the isoseismal map, that shows a distribution of energy in an elliptical area with the major axis parallel to the coast. The occurrence of a local tsunami with a maximum height tides of 7 8 m, located between Ocoña and Mollendo cities, could be due to the complexity of the rupture process. This complexity can be observed in the broad-band records at teleseismic distances where three different pulses are observed, that
282 Figure 2. Distribution of aftershocks (white circles) for the Arequipa and for the 1996 (white squares) earthquakes, together with focal mechanisms for the Arequipa event and the three larger aftershocks. Focal mechanisms of the larger events in this area are enclosed. In top of the figure vertical cross section with pressure and tension axes for the Arequipa sequence. Figure 3. Focal mechanisms of main shock and three larger aftershocks of the Arequipa sequence. Lower hemisphere of focal sphere are shown. Black circles correspond to compressions and white circles to dilatations, P = pressure axis, T = tension axis.
283 may be identified as three different steps in the rupture process. Focal mechanisms obtained for the main shock and aftershocks of June 26 and July 7, correspond to thrusting solutions with a horizontal plane dipping to the ENE and horizontal pressure axes striking in ENE-WSW direction, in agreement with focal mechanisms of larger earthquakes which have occurred in Peru between the trench and the coastline (Figure 2). The stress regime obtained from focal mechanisms of the Arequipa sequence (on top of Figure 2) may be associated with the convergence motion between the Nazca and South America plates that extends to 60 km depth. The aftershock of July, 5 located inside the continent and at deeper depth (69 km), shows horizontal extension in a NE-SW direction, parallel to the convergence direction according to the stress regime for intermediate depth earthquakes in this area, where the lithospheric material is stretched and pushed at deeper zones (Tavera and Buforn, 2001) Acknowledgements The authors wish to thank to Prof Udías and Dr Mcintosh, Universidad Complutense de Madrid, and Dr Woodman and Montes, Instituto Geofísico de Peru for helpful comments on the manuscript and to I. Perez- Pacheco for the collaboration in the Figures. Special acknowledgements to the people in the area affected by the earthquake, especially the people of Camaná. This work has been supported in part by the Ministerio de Ciencia y Tecnología (Spain), project REN2000-0777-C02-C01. Publication 008-CNDG-IGP/2001. References Brillinger, D., Udias, A. and Bolt, B., 1980, A probability model for regional focal mechanism solutions, Bull. Seism. Soc. Am. 70, 1479 1485. Kennett, B., 1991, IASPEI 1991 Seismological Tables, Res. School. Earth. Sci., Camberra, 167 pp. Minster, J. and Jordan, T., 1978, Present-day plate motions, J. Geophys. Res. 83, 5331 5334. Tavera, H. and Buforn, E., 2001, Source mechanism of earthquake in Peru, Journal of Seismology (in press). Udias, A. and Buforn, E., 1988, Single and joint fault-plane solutions from first motion data, In: Doornbos, D. (ed.), Seismological Algorithms, Academic Press, London, pp. 443 453.