Damage Assessment and Seismic IntensityAnalysis of the 2010 (Mw 8.8) MauleEarthquake

Maximiliano Astroza,a) Sergio Ruiz,b), c) and Rodrigo Astrozad), e)

The MSK-64 seismic intensities inside the damage area of the 2010 Mauleearthquake are estimated. With this purpose, field surveys of damage to typicalsingle-family buildings located in 111 cities of the affected area were used.Cities located close to the north part of the earthquake rupture suffered higherdamage, but most of this damage concerned adobe and unreinforced masonryhouses. Minor and moderate damage was noted in modern low-rise engineeredand nonengineered constructions, especially in confined masonry buildings.Despite the large length of the rupture, which reached more than 450 km,only one intensity value equal to IX was determined, and 21% of the valueswere greater than VII. The attenuation of seismic intensity was controlledby the distance to the main asperity more than to the hypocenter, whichwould be an important characteristic of the megathrust earthquakes, and itshould therefore be considered in the seismic risk of large subduction environ-ments. [DOI: 10.1193/1.4000027]

INTRODUCTION

The 2010 Maule earthquake is the fifth-largest instrumentally recorded earthquake; thus,the study of damage and seismic intensities associated with the earthquake is useful to under-stand the effects of historical great thrust earthquakes and the seismic hazard in other seismiczones, where this type of earthquake is expected and there are no strong ground motionrecords of large earthquakes, as in Northern Chile (Kelleher 1972) or the Cascadia subductionzone (Heaton and Hartzell 1987).

The great Chilean interplate or thrust earthquakes are characterized by a few zones ofhighest slip that controlled their rupture (Ruiz et al. 2011), which will be called asperities,following the original idea of Kanamori and Stewart (1978). The rupture process of theMaule earthquake was characterized by more than one asperity (Lay et al. 2010, Delouiset al. 2010, Tong et al. 2011, Lorito et al. 2011); however, the asperity with higher slip

Earthquake Spectra, Volume 28, No. S1, pages S145–S164, June 2012; © 2012, Earthquake Engineering Research Institute

a) Department of Civil Engineering, Universidad de Chile, Blanco Encalada 2002, Santiago, Chileb) Department of Geology, Universidad de Chile, Plaza Ercilla 803, Santiago, Chilec) Départament des Géosciences, Ecole Normale Superieure, 24 rue Lhomond, 75005 Paris, Franced) Faculty of Engineering, Universidad de Los Andes, San Carlos de Apoquindo 2200, Santiago, Chilee) Ph.D. student, Department of Structural Engineering, University of California, San Diego, La Jolla, CA, 92093

S145

(main asperity) is located to the north of the seismic gap, approximately in the same rupturearea of the 1928 Talca earthquake (Ruiz et al. 2012).

On the other hand, the 2010 Maule earthquake caused damages in a zone inhabited bymore than the 50% of the Chile’s population (INE 2002). The prevailing building types in thecities facing the 2010 earthquake are old traditional adobe constructions (6.1%) and masonryhouses (51.9%) built in the rural zone and the urban area of cities located in the ChileanCentral Valley; this means that constructions are concentrated in a few vulnerability classes(Medvedev and Sponheuer 1969, Grünthal et al. 1998). Most of these buildings were locatedon alluvial and fluvial deposits and many of them were old and weak nonengineeredstructures.

In this paper, we describe the damage in the most common type of low-rise buildingsfound in 111 cities visited after the earthquake, and we estimate the MSK-64 seismicintensity values in those places based on the observed damage due to ground shaking.These intensity values are an important tool to establish a link between historical andmodern large earthquakes, for which it is not possible to make an instrumental correlation,allowing a comparison from different points of view, for instance, in the size of theboundary of the damaged region or the distribution of damages (local site conditions).With these results, a correlation between the distance to the main asperity and theMSK-64 seismic intensity (IMSK‐64) is made, and a seismic intensity attenuation formulais proposed for the 2010 Maule earthquake. The formula is valid for the intensity rangeV ≤ IMSK‐64 ≤ IX, and it will be useful to compare the damage in Maule earthquake withprevious large earthquakes in Chile (Mw ≈ 8.0) for which strong-motion records are notavailable.

SHAKING EFFECTS

This earthquake caused widespread damage and was felt strongly throughout 500 kmalong the country, between 33.5°S and 38°S, and many towns and villages were affected.

This study does not consider the damage in cities decimated by the tsunami that followedthe earthquake, because it is very difficult to identify the damage associated only to the shak-ing effects. However, a field visit to the coast zone affected by the tsunami did not indicatevery high damage levels due to the ground motion, and many traditional and newer single-family buildings are intact or with light or moderate damage.

Light to moderate damage was also observed in other zones located in hard soil and rockconditions, despite the earthquake was felt most conspicuously in recent Quaternary deposits,mainly alluvial and fluvial, where shaking amplification have occurred historically.

Different type of construction suffered very different degrees of damage during the earth-quake. Structures built using weak materials (adobe) and old-fashioned buildings (unrein-forced masonry buildings built before the 1940s) were more affected, and modernsingle-family, confined masonry dwellings suffered no damage, as was observed in manyhouses built in the last 30 years in the areas damaged by the Maule earthquake.

The region affected by the earthquake was exposed to several major earthquakes in thepast, including the 1985 Central Chile earthquake (Mw 8.0). However, the performance of

S146 M. ASTROZA, S. RUIZ, AND R. ASTROZA

the buildings during past earthquakes was based on single-family dwellings of one and two-stories, built with adobe and masonry. In the 2010 Maule earthquake, reinforced masonry(RM) and confined masonry (CM) buildings of three and four stories were exposed to severeground shaking for the first time. The construction of RM and CM apartment buildings inSantiago began in the 1970s and, in the 1990s, in other main urban centers located in the areasdamaged by the 2010 Maule earthquake (e.g., Rancagua (34.16°S, 70.74°W), San Fernando(34.58°S, 70.99°W), Talca (35.42°S, 71.66°W), Constitución (35.34°S, 72.41°W), andConcepción (36.82°S, 73.03°W)). Most of the engineered masonry buildings of threeand four stories remained undamaged; nevertheless, a few of them were severely damaged,and three three-story buildings collapsed. More details about the deficiencies of modernmasonry construction are described by Astroza et al. (2012).

DAMAGE IN SINGLE-FAMILY BUILDINGS

The majority of the building stock in the central-south region of Chile reveals thatthe dominant buildings are nonengineered adobe and brick masonry one-story structures(Morales and Sapaj 1996), and field observations suggest that the damage was concentratedin this type of construction. Typical buildings are described in turn below.

ADOBE BUILDINGS

Most of adobe buildings are old buildings rebuilt after the 1928 Talca (Mw 7.7) and 1939Chillán (Mw 7.7) earthquakes, and they are single-story buildings constructed following tra-ditional techniques and using locally available materials and labor based on experience with-out a technical support (Figure 1a and 1b). The walls are made of sun-dried clay brick in mudmortar, both stabilized with straw and made by hand. Foundation and plinth are built withbrick or stone units bonded with mud or cement/lime mortar (Figure 2a). The roof is directlysupported on the load-bearing walls (Figure 2b), and it is made with timber beams or timbertruss covered by clay roofing tiles put over a thick mud isolation layer.

Adobe houses have an overall height between 2.5 m and 4.0 m. The wall thickness rangesfrom 300 mm to 600 mm. Most walls are plastered either with cement/lime or mud mortar,with a thickness of about 40 mm. This plaster is bonded with an expanded thin steel perfo-rated sheet to the wall. Wood lintels are put above the window and door openings (Figure 2c).Furthermore, there is no other reinforcement in the walls like vertical supporting columns ortie-beams running on the top of the walls. The partitions are built with a timber bracing framestructure infilled with adobe units and covered with a mud plaster (Figure 2d), without con-nections with the load-bearing adobe walls.

The collapse of adobe buildings is due to a weak connection among the walls, andbetween the walls and the roof; the heavy weight of walls and roof; very low strength;and slender unreinforced walls. Also, the old age, low quality of construction and materials,and the lack of maintenance contribute to damage in some cases. The presence of buttressesin order to reduce the unsupported wall length can be proved as a good seismic reinforcementfor this type of construction.

Typical damages are: vertical cracks at the wall corners resulting in the separation ofthem (Figure 3a), diagonal cracks (Figure 3b), and collapse through out-of-plane failure

DAMAGE ASSESSMENT AND SEISMIC INTENSITY ANALYSIS S147

mechanism (Figure 3c), with subsequent collapse of the roof due to the loss of the wall sup-port (Figure 3d).

A few adobe buildings survived the 2010 Maule earthquake with moderate damage levelin moderate intensity area, IMSK < VII (e.g., Rancagua) and many buildings suffered partiallyor completely collapse in high intensity area, IMSK ≥ VII (e.g., Talca, Curepto (35.08°S,72.02°W), and Constitución).

Figure 1. Photos of typical adobe structures in the affected area: (a) adobe house in Marchigue(34.40°S, 71.61°W) and (b) adobe house in Paredones (34.66°S, 71.89°W).

S148 M. ASTROZA, S. RUIZ, AND R. ASTROZA

UNREINFORCED MASONRY BUILDINGS

Low-rise unreinforced masonry buildings are mostly old construction built in the 1920swith burnt clay bricks in cement/sand mortar. They are located in the urban area of the maincities affected by the 2010 Maule earthquake, and some of them are part of Chile’s patri-monial heritage (Figure 4a and 4b).

The wall thickness ranges from 300 mm to 450 mm. Building heights vary between 3 and4 m for one-story buildings. In two- and three-story buildings, the floor is a timber diaphragmthat works as a one-way slab where loads are transferred to the wood joist and then to theload-bearing walls. Usually, these buildings do not have reinforcement except for a RC lintelabove the window and door openings. The roof is a corrugated iron sheet resting on a timberframe or truss. In some cases, there are RC tie-beams running on the top of the walls, wherethe roof is bore. The partitions are built with a timber-bracing frame infilled with adobe unitswithout connections to the walls of the structural resistant system.

Damages sustained in this type of construction are similar to adobe houses, due to lack ofreinforcement in the wall corners and in the walls, and an inadequate wall-to-floor anchorage.These conditions would explain the vertical separation of the intersecting walls, in-planediagonal shear cracking, out-of-plane wall collapse, flexural cracking of a free-standing

Figure 2. Photos of details in adobe houses located in the central-south part of Chile: (a) plinthbuilt with brick, (b) support of roof, (c) wood lintel above the window, and (d) partition.

DAMAGE ASSESSMENT AND SEISMIC INTENSITY ANALYSIS S149

parapet and gable walls and loss of a wall segment due to intersecting diagonal cracks. Out-of-plane damage is a very common damage pattern, especially in tall historical buildings aschurches (Figure 5a) and in upper portions of buildings.

In general, unreinforced masonry buildings sustained significant damage (Figure 5b),a confirmation of their fragility and that they should not be used in seismic areas. In thecities located within the areas damaged by the 2010 Maule earthquake, this sort of construc-tion was not affected seriously by the 1985 Central Chile (Mw 8.0) and the 1960 Concepción(Mw 8.2) earthquakes, for example, Rancagua, San Fernando, Curicó (34.98°S, 71.22°W),Talca, Linares (35.85°S, 71.59°W), and Los Ángeles (37.47°S, 72.35°W).

CONFINED MASONRY BUILDINGS

This type of construction is widespread over the central-south region of Chile. Typically,buildings are single-family, partially confined masonry dwellings of up to two stories high.They have had a good performance, including during the 1939 Chillán (Mw 7.7) earthquakeand the 2010 Maule earthquake (Mw 8.8), where most of them did not experience anydamage, with the exception of a few cases that suffered moderate damage.

Figure 3. Typical damage in adobe structures and their corresponding grade of damage: (a) ver-tical cracks at wall corner (G3), (b) diagonal crack in wall (G3), (c) wall collapse through out-of-plane (G4), and (d) collapse of the roof (G5).

S150 M. ASTROZA, S. RUIZ, AND R. ASTROZA

Confined masonry buildings are built with shear masonry walls and reinforced concrete(RC) confining elements (tie-columns and tie-beams), where the walls are constructed first,followed by casting of concrete in RC tie-columns. Walls are built using a variety of masonryunits (e.g., handmade solid clay unit and machine-made perforated and hollow clay bricks) incement/sand mortar. The RC tie-columns are usually placed at wall intersections, and inmany cases a RC tie-column is not provided at openings. The spacing between verticalRC confining elements is about 3 m to 3.5 m. A toothed interface between the walls andthe tie-columns was observed in new and existing buildings, as shown in Figure 6a and 6b.

The RC tie-beams are provided at the floor and roof levels. The floor system is plankingon timber beams and the roof is a corrugated iron sheet on timber trusses, which is anchored

Figure 4. Photos of unreinforced masonry structures (patrimonial heritage): (a) building of TalcaIntendencia and (b) San Fernando governorship building.

DAMAGE ASSESSMENT AND SEISMIC INTENSITY ANALYSIS S151

to RC tie-beams running at the perimeter on the top of the walls. Many buildings have smallplan dimensions and they are attached between them (townhouse-style).

Performance of single-family low-rise confined masonry is excellent, large majority ofbuildings were not damaged at all. The most common damage pattern observed in masonrywalls were in-plane shear cracking (Figure 7a) and damage to the RC confining members,

Figure 5. Typical damage inunreinforcedmasonry structures: (a) Putu church (35.22°S, 72.28°W)and (b) building in Curicó (34.98°S, 71.22°W).

S152 M. ASTROZA, S. RUIZ, AND R. ASTROZA

particularly RC tie-columns (Figure 7b). Out-of-plane wall damages were observed in para-pet and gable walls (Figure 7c and 7d).

The most severe damage in this type of buildings can be attributed to inadequate qualityof mortar and construction process, deficiencies in detailing of reinforcement in RC confiningelements, absence of reinforcement around the wall openings, and geotechnical and topo-graphics issues (Astroza et al. 2012).

SEISMIC VULNERABILITY

Considering that the single-family buildings used to estimate the seismic intensity werebuilt without the assistance of a professional (engineer or architect), there is no informationabout mechanical properties of them, but they represent the quality of the typical constructionthat had been affected by Chilean earthquakes in the last 100 years in this zone of thecountry, and consequently the observed response during previous events might be consideredcomparable.

Figure 6. Details of connection between walls and the tie-columns in confined masonry building.

DAMAGE ASSESSMENT AND SEISMIC INTENSITY ANALYSIS S153

The building stock in the region can be grouped, according to ability to resist damagefrom earthquakes, into four vulnerability classes depending upon the wall materials and thereinforcement (Monge and Astroza 1989). The vulnerability assignment of the single-familybuildings is conducted in terms of the MSK-64 macroseismic scale (Medvedev andSponheuer 1969) and EMS-98 scale (Grünthal et al. 1998; Table 1); it is an attempt to cate-gorize, in a manageable way, the strength of structures, taking building type and other factorsinto account.

Figure 7. Damage observed in confined masonry buildings: (a) in-plane shear cracking,(b) damage to the RC tie-column, (c) damage in parapet, and (d) damage in gable wall.

Table 1. Vulnerability class of single-family Chilean buildings

VulnerabilityClass (VC) Basic type of structure (TS)

VC-A Adobe. Masonry stone in mud mortar.VC-B Unreinforced masonry. Braced wooden frame infilled with

adobe unit.VC-C Confined masonry.VC-D Confined masonry designed according to NCh2123

Chilean Code.

S154 M. ASTROZA, S. RUIZ, AND R. ASTROZA

MSK INTENSITIES OF THE 2010 MAULE EARTHQUAKE

The intensity values in 111 cities in the earthquake’s damage area are estimated from thefield surveys considering the distribution of the observed damage in adobe and masonrysingle-family buildings of one and two stories. For this purpose, the buildings are classifiedin one of the vulnerability classes (VC) indicated in Table 1, and the grade of damage isassigned using the description of the MSK-64 macroseismic scale (Medvedev and Sponheuer1969), adapted to the Chilean houses by Monge and Astroza (1989).

The damage is classified by six grades as follows: Grade 0 (G0)-No damage; Grade 1(G1)-Slight damage (fine cracks in plaster; falling of small pieces of plaster); Grade 2 (G2)-Moderate damage (fine cracks in walls; vertical cracks at wall intersections; horizontal cracksin chimneys, parapets and gables; spalling of fairly large pieces of plaster; falling of parts ofchimneys; sliding of roof tiles); Grade 3 (G3)-Heavy damage (large and deep diagonal cracksin most walls; large and deep vertical cracks at wall intersections; some walls lean out-of-plumb; falling of chimneys, parapets and gable walls; falling of roof tiles); Grade 4 (G4)-Very heavy damage (partial or total collapse of a wall in the building; collapse of buildingpartitions); Grade 5 (G5)-Collapse or destruction (Collapse of two or more walls in thebuilding).

The seismic intensity value (above MSK-64 intensity V) is calculated using the relation-ship between the intensity degree and the distribution of damage grades for each basic vul-nerability class proposed by the MSK-64 scale, as adapted by Karnik et al. (1984), definingthe number of houses damaged in the following terms: few (5%), many (50%), or most (75%;Table 2). Karnik et al. (1984) completed the description of the MSK-64 scale so that the totalof 100% is reached for each intensity degree and vulnerability class. In order to include VC-Dclass in Table 2, we used the percentage proposed by Karnik et al. (1984) and the masonrybuildings classification and distribution of the damage grade proposed by the EuropeanMacroseismic Scale (EMS-98; Grunthal et al. 1998).

The damage field surveys were made in several blocks around the main square of eachcity (Astroza et al. 2010), taking care that the sample size was greater than 20 cases (Mongeand Astroza 1989). The use of Table 2 reduces the ambiguities because the effects on anumber of the most common type of buildings in a limited area are considered as sensors.

The estimated intensity values are shown in the Table 3 and Figure 8. Seismic intensitieswere plotted as point data and as isoseismal map, which provides a good overall descriptionof the extent of the damage. These isoseismal curves consider only the sites on hard soil(compact and old alluvial or fluvial deposits) and ignore damage due to the tsunami, land-slides, or liquefaction; in this way, the proposed isoseismal map represents the intensityvalues in a hard soil with shear-wave velocities between 500 m/s and 900 m/s, accordingto soil type II, in the new Chilean seismic code of buildings (INN 2010).

From Figure 8, it is possible to appreciate that the VIII isoseismal curve of the 2010Maule earthquake is related with the Northern asperity and the length of the rupture areais related with the length of the VII isoseismal curve, confirming that the damage is closelyrelated with the asperity distance (Ruiz et al. 2011), rather than with the hypocenter distance.The isoseismal curves are elongated in the north–south direction, according to the earthquakeslip distribution or the shape of the main asperities (Figure 8).

DAMAGE ASSESSMENT AND SEISMIC INTENSITY ANALYSIS S155

The highest intensity value is due to amplification effect of some Quaternary deposits.The highest MSK-64 intensity was IX in Constitución, located to the north of the epicenter ona silty sand soil. At locations where the MSK-64 intensity was VIII, almost all the adobehouses became uninhabitable (Damage grade 3 or more), and many of them (55%) totally orpartially collapsed.

Several cities, like Talca, Parral (36.15°S, 71.82°W), and Constitución, located in theepicentral zone, did not show a homogenous damage distribution, and the highest intensitywas found on volcanic ash deposit at Talca and Parral and silty sand soil at Constitución

Table 2. Damage grades (DG) and damage ratios (N, in %) for individual intensity degreeof the MSK-64 scale (Modified fron Karnik et al. 1984)

Intensitydegree

VC-A VC-B VC-C VC-D

DG N (%) DG N (%) DG N (%) DG N (%)

VG1 5 G0 100 G0 100 G0 100G0 95

VI

G2 5 G1 5 G0 100 G0 100G1 50 G0 95G0 45

VII

G4 5 G2 50 G1 50 G1 5G3 50 G1 35 G0 50 G0 95G2 35 G0 15G1 10

VIII

G5 5 G4 5 G3 5 G2 5G4 50 G3 50 G2 50 G1 50G3 35 G2 35 G1 35 G0 45G2 10 G1 10 G0 10

IX

G5 50 G5 5 G4 5 G3 5G4 35 G4 50 G3 50 G2 50G3 15 G3 35 G2 35 G1 35

G2 10 G1 10 G0 10

X

G5 75 G5 50 G5 5 G4 5G4 25 G4 35 G4 50 G3 50

G3 15 G3 35 G2 35G2 10 G1 10

XI

G5 100 G5 75 G5 50 G5 5G4 25 G4 50 G4 50

G3 35

G2 10XII G5 100 G5 100 G5 100 G5 100

S156 M. ASTROZA, S. RUIZ, AND R. ASTROZA

(Morales and Sapaj 1996). This pattern would be associated with the surface geologicalconditions (Astroza and Monge 1991). At Constitución, the situation was similar tothat observed in Valparaíso (33.05°S, 71.60°W) after the 1906 Valparaíso (Saragoniand Carvajal, 1991) and 1985 Central Chile earthquakes (Acevedo et al. 1989), withhigh intensity values in the flat zones of the city (silty sand soil) and lower values athill zones. Although in hill areas the seismic intensity was usually low, when the structureswere located on the hilltop ridge, topographic effects were remarkable. At O’Higgins Hill,located in Constitución, the structures next to the ridge suffered important damage whilestructures located far away from the ridge did not suffer any damage (Figure 9a and 9b).A similar pattern was identified at Canal Beagle for the Chile Central 1985 earthquake(Çelebi 1987).

Table 3. Estimated MSK intensities (Astroza et al. 2010). The identification number ofeach city is shown in Figure 8

MSK Intensity Cities

≥ VIII Peralillo (1), Pumanque (2), Licantén (3), Curepto (4), Constitución (5), Talca (6),Cauquenes (7), Parral (8)

VII ½ Navidad (9), Doñihue (10), Coínco (11), Esperanza (12), Coltauco (13),Pichidegua (14), Población (15), Santa Cruz (16), Chépica (17), Lolol (18),Curicó (19), Gualleco (20), Batuco (21), Concepción (22)

VII El Convento (23), San Enrique (24), La Boca (25), Matanza (26), Rengo (27),Litueche (28), Machalí (29), La Estrella (30), Peumo (31), San Vicente de TaguaTagua (32), Las Pataguas (33), Teno (34), Rauco (35), Hualañe (36), Molina(37), Cumpeo (38), Pencahue (39), Bobadilla (40), Empedrado (41), Chanco (42),Retiro (43), Cobquecura (44), Quirihue (45), Ninhue (46), San Carlos (47),Quillón (48), Bulnes (49), Penco (50), Tomé (51), Florida (52), Lota (53),Santiago (106)

VI ½ Pomaire (54), Talagante (55), El Monte (56), Melipilla (57), San Pedro (58),Rapel (59), Rancagua (60), San José (61), Marchigue (62), San Fernando (63),Paredones (64), Nancagua (65), La Huerta (66), Pelarco (67), San Clemente (68),San Javier (69), Villa Alegre (70), Linares (71), Longaví (72), Coelemu (73),San Nicolás (74), Chillán (75), Ñipas (76), Cabrero (77), Los Ángeles (78),Angol (79)

VI San Antonio (80), Codegua (81), Las Cabras (82), Requínoa (83), Chimbarongo(84), Romeral (85), Yerbas Buenas (86), Trehuaco (87), San Ignacio (88),General Cruz (89), Pemuco (90), Monte Aguila (91), Yumbel (92), Yungay (93),Arauco (94), La Laja (95), Curanilahue (96), Nacimiento (97), Lebu (98),Los Álamos (99), Tres Pinos (100), Valparaíso (107), Viña del Mar (108)

V Renaico (101), Cañete (102), Los Sauces (103), Contulmo (104), Purén (105),Los Vilos, Illapel, Salamanca

DAMAGE ASSESSMENT AND SEISMIC INTENSITY ANALYSIS S157

INTENSITY ATTENUATION

A classical regression analysis was considered for fitting the intensity data collected inthe 111 cities visited after the earthquake. We consider the intensity attenuation of the form:

EQ-TARGET;temp:intralink-;e1;50;162IðΔÞ ¼ C1 þ C2 ⋅ Rþ C3 ⋅ log10ðΔÞ (1)

where the constants C1, C2, and C3 can be associated to different stress drop, absorption andscattering, and geometric spreading respectively (Bakun and Joyner 1984). These constantsare calculated by the least-square method considering the mean or median value of the

Figure 8. Intensities of each city and the isoseismal map inside the damage area of the 27February 2010 earthquake. Roman numerals refer to the MSK scale and the Arabic numeralsare associated with the cities in Table 3. The slip distributions of Maule 2010 earthquake modifiedfrom Tong et al. (2011) are also shown.

S158 M. ASTROZA, S. RUIZ, AND R. ASTROZA

distance Δ for each intensity level. The variable Δ can be the hypocentral distance or thedistance to the asperity centroid, computed as

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

R2 þ h2p

, where R is the distance between thesite and the point on the earth’s surface vertically above the hypocenter, or the distancebetween the site and the point on the earth’s surface vertically above the asperity centroidrespectively, and h is the depth of the hypocenter or the asperity centroid respectively.

Using 945 values of Mercalli Modified Intensity (MMI) from 73 events (especially inter-plate or thrust) with M ≥ 5.5 and focal depth less than 120 km whose stroke Chile between1906 and 1977, Barrientos (1980) established the following intensity attenuation formula forChilean earthquakes:

EQ-TARGET;temp:intralink-;e2;71;316IMMIðΔÞ ¼ 1.3844 ⋅Mw − 3.7355 ⋅ log10ðΔÞ − 0.0006 ⋅ Δþ 3.8461 (2)

The database considered by Barrientos (1980) includes 13 events with M > 7.0, whichrepresents a high number of large earthquakes (including the 1906 Valparaíso Mw 8.5, 1928Talca Mw 7.7, 1939 Chillán Mw 7.7 and 1960 Valdivia Mw 9.5). However, only 25 eventshave intensity MMI > V. In this way, the Maule earthquake (Mw 8.8) allows, for the firsttime, the study of the intensity attenuation of a complete database for a megathrust earth-quake in the Chilean subduction. Furthermore, it is important to note that for the Maule earth-quake, we are considering only values of intensity equal or higher than V, because few of thevisited cities showed intensities lower than V, and such data are clearly incomplete; conse-quently, to consider them in the attenuation would induce statistical bias.

Figure 10 shows the intensities estimated for the 2010 Maule earthquake versus thehypocentral distance and Barrientos’s (1980) formula for Mw ¼ 8.8. Two datasets areshown, where the black triangles are the intensities corresponding to locations to thesouth of the epicenter and the gray circles are the intensities corresponding to locationsto the north of the epicenter (36.29°S, 73.24°W). Clearly, the Barrientos’s hypocentralattenuation formula does not reflect the intensity attenuation for the 2010 Maule earthquakeappropriately.

Figure 9. Buildings of O’Higgins Hill–Constitución: (a) Collapse of building next to the ridge(first-floor collapse), (b) building located far away from the ridge (without damage).

DAMAGE ASSESSMENT AND SEISMIC INTENSITY ANALYSIS S159

However, the rupture process of this event shows one main asperity where the highestslips were concentrated. The centroid of main asperity is located to more than 100 km of thehypocenter. Coordinates and depths of the centroid of the main asperity inferred from Tonget al. (2011) and the hypocenter are summarized in Table 4.

Considering that the most important high frequency energy was released from the onemain asperity (Ruiz et al. 2012), it is natural to expect that the highest intensity (IMSK ≥ VIII)is closer to this asperity rather than to the hypocenter (Figure 8), which would imply that thegreater damage produced by the 2010 Maule earthquake on the structures that were consid-ered to evaluate the intensity (low-rise, rigid structures) were controlled by this asperity.

Figure 11 shows the intensities estimated for the Maule earthquake versus the distance tothe main asperity centroid (ΔA) without accounting for site effects explicitly. From this data, aregression analysis is carried out in order to obtain the intensity attenuation curve. Also, theconfidence bounds of the mean values for a coefficient of 95%, equal to approximately two

Figure 10. MSK-64 intensity versus hypocentral distance plot for the 2010 Maule earthquake.Two sets of data are shown: Black triangles are the intensities corresponding to locations tothe south of the hypocenter, and the gray circles are the intensities corresponding to locationsto the north of the hypocenter.

Table 4. Coordinates and depths of the main asperity centroid (inferred fromTong et al. (2011) and hypocenter of SSN (2011)

Latitude Longitude Depth (km)

Centroid Main Asperity 34.8°S 72.6°W 20Hypocenter 36.29°S 73.24°W 30.1

S160 M. ASTROZA, S. RUIZ, AND R. ASTROZA

times the standard deviation if we consider normally distributed errors, are computed. Fromthe regression analysis, the formula describing the intensity attenuation for the 2010 Mauleearthquake is as follows:

EQ-TARGET;temp:intralink-;e3;71;612IMaule−2010ðΔAÞ ¼ 19.781 − 5.927 ⋅ log10ðΔAÞ þ 0.00087 ⋅ ΔA (3)

From Figure 11, it can be noted that the data fit very well if we consider the distance to thenorthern asperity centroid. Consequently, this asperity would have controlled the damageproduced by the earthquake on the structures if it took into account the estimation of theintensity values (low-rise rigid structures), which would be the result of the high-frequencypulses generated in this asperity (Ruiz et al. 2012). Formulas that use the shortest distance tothe surface projection of the fault plane gave a low correlation, due to the higher release ofenergy was concentrated in the Northern zone of the fault rupture (see Figure 8).

From Figure 11, it is also clear that the average trend of the intensity decay is well repro-duced by our relations and since the uncertainties have been estimated, the error in a newintensity estimate using the proposed relations is approximately of 0.5 intensity units, whichis basically due to the nature of intensity assessment and other effects, for example, siteconditions.

Comparing Figures 10 and 11, it can be noted that the data are better grouped together,reducing the scatter considerably if we consider the distance to the main asperity instead ofthe hypocentral distance. This characteristic is very clear when the intensity value is equalto V. This implies that for the study of the seismic hazard of megathrust earthquakes havingan extensive area of rupture, as in the 2010 Maule earthquake and the Valdivia earthquake of

Figure 11. Intensity versus distance to the main asperity centroid plot, comparing the estimatedintensities with the Equation 3 for the 2010 Maule earthquake data.

DAMAGE ASSESSMENT AND SEISMIC INTENSITY ANALYSIS S161

22 May 1960 (Astroza and Lazo 2010), the intensity attenuation should be defined as a func-tion of the distance to asperities, instead of the hypocentral distance or the nearest distance tothe rupture area, as had been assumed so far.

This results in an important conclusion about megathrust earthquakes by giving a guide-line for future work related to intensity attenuation curves for thrust earthquakes of largemagnitude.

DISCUSSION AND CONCLUSIONS

Despite the large rupture length of more than 450 km, only one intensity value equal to IXwas determined, and 21% of the values are greater than VII. The 2010Maule earthquake low-intensity values are in agreement with previous studies of other big interplate earthquakes,such as the 1906 Valparaíso earthquake (Saragoni and Carvajal 1991), the 1960 Valdiviaearthquake (Astroza and Lazo 2010), and the 2004 M 9.0 Sumatra-Andaman earthquake(Martin 2005). The most important damage occurred in the VI and VII Chilean Regions(34°S to 36°S), but most of this damage concerned traditional construction (adobe houses)and old masonry buildings (unreinforced masonry buildings). Minor damage, such as smallcracks in walls, and moderate damage were noted in modern engineered and nonengineeredbuildings (confined masonry houses and buildings). In these cases, the damage may be attrib-uted mainly to poor-quality construction and surface geological conditions, despite the factthat the most of these buildings do not have seismic design.

The 2010 Maule mega-earthquake confirms the importance of the main asperities in thedamage distribution, since the best fitting of the intensity attenuation formula is obtainedconsidering the distance to the main asperity, instead of hypocentral or to the nearestfault distance.

The characteristics of the 2010 Maule earthquakes should be considered in the seismicrisk of a large subduction environment, such as the Northern Chile or the Cascadia zone,where megathrust earthquakes have not occurred in the last century.

Macroseismic intensity is a valuable tool for supplementing seismic hazard assessmentswhen a complete data set is available and there are not many strong motion records available,because of its direct relation to the damage produced by earthquakes. Finally, to investigatehistorical mega-earthquakes effects, it is required to establish a comparison between availableseismic intensity data and the macroseismic data of the 2010 Maule earthquake (Mw 8.8).This is the only way we can fully exploit the wealth of historical information available.

ACKNOWLEDGMENTS

The authors would like to thank Professors Maria Ofelia Moroni, Leonardo Massone, andElizabeth Parra and the graduate students Francisco Cabezas and Felipe Cordero who col-laborated in the damage inspection. The field trip was funded by the Department of CivilEngineering of the Universidad de Chile. Sergio Ruiz has been partially funded by theMillennium Nucleus Program “Montessus de Ballore–IERC,” Ministry of Planning andCooperation (Mideplan), Chile.

S162 M. ASTROZA, S. RUIZ, AND R. ASTROZA

REFERENCES

Acevedo, P., Astroza, M., Canales, J. C., Monge, J., and Perretta, C., 1989. Relación entre lasunidades geotécnicas y los daños producidos por el sismo del 3 de Marzo de 1985 en la ciudadde Valparaíso, in Proceedings, V Jornadas de Sismología e Ingeniería Antisismica, Volume 1,Santiago, Chile, 515–523 (in Spanish).

Astroza, M., and Lazo, R., 2010. Estudio de los daños de los terremotos del 21 y 22 de Mayo de1960, in Proceedings, X Jornadas de Sismología e Ingeniería Antisismica, Santiago, Chile,CD-ROM (in Spanish).

Astroza, M., and Monge, J., 1991. Regional seismic zonation in Central Chile, in Proceedingsof the Fourth International Conference on Seismic Zonation, Volume III, Stanford, CA,487–494.

Astroza, M., Cabezas, F., Moroni, M. O., Massone, L., Ruiz, S., Parra, E., Cordero, F., andMottadelli, A., 2010. Intensidades sísmicas en el área de daños del terremoto del 27 deFebrero de 2010, http://www.eqclearinghouse.org/20100227-chile/wp-content/uploads/2010/04/Informe-de-Intensidades-M.-Astroza-y-otros.pdf, last accessed 31 October 2010 (in Spanish).

Astroza, M., Moroni, M. O., Brzev, S., and Tunner, J., 2012. Seismic performance of engineeredmasonry buildings in the 2010 Maule earthquake, Earthquake Spectra 28, this issue.

Bakun, W. H., and Joyner, W. B., 1984. The ML scale in central California, Bull. Seism. Soc. Am.74, 1827–1843.

Barrientos, S., 1980. Regionalización Sísmica de Chile, M.S. Thesis, Departamento de Geofísica,Universidad de Chile (in Spanish).

Çelebi, M., 1987. Topographical and geological amplifications determined from strong-motionand aftershock records of the 3 March 1985 Chile earthquake, Bull. Seism. Soc. Am. 77, 1147–1167.

Delouis, B., Nocquet, J. M., and Vallée, M., 2010. Slip distribution of the 27 February 2010Mw= 8.8 Maule Earthquake, central Chile, from static and high-rate GPS, InSAR, andBroadband teleseismic data, Geophys. Res. Lett. 37, L17305, doi:10.1029/2010GL043899.

Grünthal, G., Musson, R. M., Schwarz, J., and Stucchi, M., 1998. European Macoseismic Scale-1998, Vol. 15, Cahiers du Centre Européen de Geodynamique et de Séismologie, Luxem-bourg.

Heaton, T. H., and Hartzell, S. H., 1987, Seismic hazards on the Cascadia subduction zone,Science 236, 162–168.

Instituto Nacional de Estadísticas (INE), 2002. XVII Censo Nacional de Población y VI deVivienda, Instituto Nacional de Estadísticas, Santiago, Chile, www.ine.cl (in Spanish).

Instituto Nacional de Normalización (INN), 2010. NCh433: Norma Técnica de Emergencia parael Diseño Sísmico de Edificios, Ministerio de Vivienda y Urbanismo, Chile (in Spanish).

Kanamori, H., and Stewart, G. J., 1978. Seismological aspects of the Guatemala earthquake ofFebruary 4, 1976, J. Geophys. Res. 83, 3427–3434.

Karnik, V., Schenkova, Z., and Schenk, V., 1984. Vulnerability and the MSK Scale, Eng. Geol.20, 161–168.

Kelleher, J., 1972. Rupture zones of large South American earthquakes and some predictions,J. Geophys. Res. 77, 2087–2103.

Lay, T., Ammon, C. J., Kanamori, H., Koper, K. D., Sufri, O., and Hutko, A. R., 2010. Tele-seismic inversion for rupture process of the 27 February 2010 Chile (Mw 8.8) earthquake,Geophys. Res. Lett. 37, L13301, doi:10.1029/2010GL043379.

DAMAGE ASSESSMENT AND SEISMIC INTENSITY ANALYSIS S163

Lorito, S., Romano, F., Atzori, F., Tong, X., Avallone, A., McCloskey, J., Cocco, M., Boshi, E.,and Piatanesi, A., 2011. Limited overlap between the seismic gap and coseismic slip of thegreat 2010 Chilean earthquake, Nature Geoscience 4, 173–177.

Martin, S. S., 2005. Intensity distribution from the 2004 M 9.0 Sumatra-Andaman earthquake,Seismol. Res. Lett. 76, 321–330.

Medvedev, S. H., and Sponheuer, W., 1969. Scale of seismic intensity, in Proceedings IV WorldConference on Earthquake Engineering, Santiago, Chile, A-2, 143–153.

Monge, J., and Astroza, M., 1989. Metodología para determinar el grado de intensidad a partir delos daños, in Proceedings, 5as Jornadas Chilenas de Sismología e Ingeniería Antisísmica,Santiago, Chile, CD-ROM (in Spanish).

Morales, G., and Sapaj, D., 1996. Riesgo siísmico de las ciudades de Curicoó, Molina, Talca, SanClemente, Constitucioón, Linares, San Javier, Parral y Cauquenes, Memoria de Ingeniería,Universidad de Chile (in Spanish).

Ruiz, S., Kausel, E., Campos, J., Saragoni, G. R., and Madariaga, R., 2011. Identification of highfrequency pulses from earthquake asperities along Chilean subduction zone using strongmotion, Pure and Applied Geophysics 168, 125–139, DOI: 10.1007/s00024-010-0117-x.

Ruiz, S., Madariaga, R., Astroza, M., Saragoni, R., Lancieri, M., Vigny, C., and Campos, J.,2012. Short-period rupture process of the 2010 Mw 8.8 Maule earthquake in Chile, EarthquakeSpectra 28, this issue.

Saragoni, G. R., and Carvajal, A., 1991. Microzonation considering large earthquake, in Pro-ceedings Fourth International Conference on Seismic Zonation, Volume III, Stanford, CA,495–502.

Servicio Sismológico de la Universidad de Chile (SSN), 2011. Servicio Sismológico, Universidadde Chile, Departamento de Geofísica. http://www.sismologia.cl/seismo.html, last access 27January 2011 (in Spanish).

Tong, X., Sandwell, D., Luttrell, K., Brooks, B., Bevis, M., Shimada, M., Foster, J., Smalley Jr.,R., Parra, H., Báez Soto, J. C., Blanco, M., Kendrick, E., Genrich, J., and Caccamise, D. J.,2011. The 2010 Maule, Chile, earthquake: Downdip rupture limit revealed by space geodesy,Geophys. Res. Lett. 37, L24311, doi:10.1029/2010GL045805.

(Received 26 February 2011; accepted 14 March 2012)

S164 M. ASTROZA, S. RUIZ, AND R. ASTROZA