Performance of masonry buildings during the Emilia 2012 earthquake

Abstract

The earthquake sequence started on May $20$th 2012 in Emilia (Italy) affected a region where masonry constructions represent a large part of the existing building stock and the construction of new modern masonry buildings is a common practice. The paper is focused on the performance of common architectural configurations, typical for residential or business use. The large majority of old masonry buildings is made of fired clay bricks. The seismic performance of these buildings is particularly interesting since major past earthquakes in Italy affected areas with mainly stone masonry structures. Apart from examples showing systematic or peculiar structural deficiencies governing the vulnerability of several buildings, the overall seismic performance of these structures to repeated shaking, with PGA as large as 0.25–0.3 g was rather good, despite the major part of them were only conceived for carrying vertical loads. In fact, seismic design is mandatory in the area only since 2003. Modern low-rise masonry buildings erected after this date and incorporating seismic design and proper detailing resulted in most cases practically undamaged. The examples reported in the paper allow an evaluation of the superior performance of seismically designed modern masonry buildings in comparison to older ones.

1 Introduction

The seismic sequence that began on the 20th of May 2012 in Northern Italy mostly affected the North-Eastern part of the geografical region of Emilia, i.e. part of the provinces of Modena, Ferrara and Bologna in the administrative region of Emilia-Romagna, the Southern portion of the province of Mantova in Lombardia and some municipalities of the province of Rovigo in Veneto. The magnitude of the two main events of May 20th and May 29th was $\hbox {M}_{\mathrm{w}}$ 6.0 and $\hbox {M}_{\mathrm{w}}$ 5.8, respectively. The earthquakes caused, directly and indirectly, a total of 27 victims, but only three of them were due to structural damage of residential buildings. The others were mostly caused by the collapse of precast industrial buildings. Several deaths occurred because of heart attacks and other illnesses. Two persons died for non-structural damages (collapse of a chimney and falling of rubble material) and a parish priest was killed during the collapse of a church in Rovereto sulla Secchia.

The affected territory is quite homogeneous in terms of geography, socio-economic development and built heritage. Most of the damaged municipalities are located in the alluvium plain formed by the Po river, just South of the river itself. The economic activities, which show a generally higher development in the provinces of Modena and Mantova, not only are associated with traditional sectors such as agriculture, breeding, textile and engineering industry and tourism, but also include sectors of excellence in food industry and specific districts like the biomedical one in Mirandola. The availability of the same raw materials and a similar history make the characteristics of the existing building stock rather homogeneous within the different municipalities. No large cities were affected but several medium sized municipalities were involved, including a number of medium-small towns with population between 10,000 and 35,000 inhabitants.

Although in 1570 the area of Ferrara was struck by a Mw 5.4 earthquake (Guidoboni et al. 2007), which also boosted some of the pioneering studies in the field of historical seismology (e.g. Breventano 1576), the knowledge on the seismic sources of the buried Northern Apennines thrust is still limited and only recent works have deepened the understanding of their seismogenic potential (e.g. Toscani et al. 2009). This only partly justifies the very recent seismic classification (OPCM3274 2003) of the municipalities affected by the 2012 events. Before 2003, structures were built without any seismic design concept and the seismic safety of existing buildings was not required to be assessed nor improved.

As shown in Table 1 for the settlements for which macroseismic intensity data were evaluated (Arcoraci et al. 2012), the large majority of the residential building stock in the area is made of masonry buildings (ISTAT 2001), generally low-rise. To some extent this fact can be considered as a circumstance which has probably counterbalanced the lack of a purposely introduced seismic protection. In fact, masonry buildings, although conceived for resisting to vertical loads only, possess an inherent additional capacity of resisting to horizontal actions, provided some good practice rules are followed (e.g. when walls are present along at least two orthogonal directions and a certain degree of connection between intersecting walls is guaranteed, contrasting the occurrence of out-of-plane failure of entire façades). This may not be the case for other structural types (e.g. precast r.c. framed structures), when “optimised” for a single loading condition. The use of fired clay bricks masonry with lime mortar was particularly common since many centuries in the flat areas of the Po valley, where clay is largely available and ancient Roman kilns can be found in several archeological sites. Stone masonry buildings are quite rare in this area.

Table 1 Observed macroseismic intensity (after Arcoraci et al. 2012), estimated PGA (Bozzoni and Lai 2012) and percentage of residential buildings with masonry structures and built before 1945 according to the 2001 Italian census data (ISTAT 2001) for some of the settlements mostly affected by the 2012 Emilia earthquake sequence (bold values indicate recorded PGA)

The most significant seismic events occurred in Italy in the last century affected areas where the buildings stock was mainly constituted of stone masonry buildings. This was the case for earthquakes occurred in mountain regions in the Apennines (e.g. $\hbox {M}_{\mathrm{w}}$ 7.0 1915 Avezzano, $\hbox {M}_{\mathrm{w}}$ 6.3 1919 Mugello, $\hbox {M}_{\mathrm{w}}$ 6.6 1930 and $\hbox {M}_{\mathrm{w}}$ 6.9 1980 Irpinia, $\hbox {M}_{\mathrm{w}}$ 6.0 1997 Umbria-Marche and $\hbox {M}_{\mathrm{w}}$ 6.3 2009 L’Aquila earthquakes), in the Alps $(\hbox {M}_{\mathrm{w}}$ 6.4 1976 Friuli) and Sicily $(\hbox {M}_{\mathrm{w}}$ 6.3 1968 Belice earthquake). Hence, the analysis of the seismic behaviour of brick masonry buildings is particularly interesting because the earthquake damage observation provides in this case a valuable and unique information.

As evident from Table 1 and from Arcoraci et al. (2012), another interesting aspect of the damage caused by the 2012 seismic sequence is that some settlements were mainly struck by just one of the two main events of the sequence $(\hbox {M}_{\mathrm{w}}$6.0 May 20th and $\hbox {M}_{\mathrm{w}}$ 5.8 May 29th events), while others, including the mostly affected centres of Mirandola, Cavezzo, Concordia, Novi and San Felice, cumulated a final damage resulting from the repeated shaking induced by the different events. The EUCENTRE Geotechnical Earthquake Engineering Section (Bozzoni and Lai 2012) estimated the values of peak ground acceleration felt at the different sites. Estimates are definitely more reliable for the May 29th event, because of the presence of a number of temporary stations which recorded the shaking. A shake map obtained from spline interpolation of the peak horizontal recorded values of acceleration is reported in Fig. 1. Data reported for the May 20th event were instead evaluated based on ground motion prediction equations (Bozzoni et al. 2012).

Fig. 1

Map of horizontal peak ground acceleration estimated for the May 29th 2012 event from interpolation of the recorded peak values of horizontal acceleration—stations are identified by blue triangles, while green dots indicate points of interest (Bozzoni and Lai 2012)

The construction of new structural masonry buildings, both unreinforced and reinforced, is still ongoing in Italy and, in the areas affected by the seismic events of May 2012, it is even more frequent than in other parts of the country. This may be also associated with the presence of several producers of modern blocks for masonry in the Emilia-Romagna region.

This paper aims to provide some insights into the seismic performance exhibited by both historical/old and modern masonry buildings in the area affected by the Emilia 2012 earthquake sequence. Special architectural forms such as churches, castles or towers (which constitute the major part of heritage buildings) are not considered herein, where the attention is focused on the most common residential/commercial/public building structural forms. Damages to churches and fortresses are reported in Sorrentino L, Liberatore L, Decanini LD, Liberatore D (2013) The prformance of churches in the 2012 Emilia earthquakes, Bulletin of Earthquake Engineering (submitted for this issue) and in Cattari S, Degli Abbati S, Ferretti D, Lagomarsino S, Ottonelli D, Tralli A (2013) Damage assessment of fortresses after the 2012 Emilia earthquake (Italy), Bulletin of Earthquake Engineering (submitted for this issue), respectively.

2 Structural damage in historical/old masonry buildings

Historical and old masonry structures are a significant part of the building stock in the earthquake affected area of Emilia.

Since the affected area was classified as potentially earthquake prone only very recently (since 2003), existing masonry buildings often show typical defects and lack of proper detailing which increase their seismic vulnerability. As an example, the presence of thrusting or unstable timber roofs caused a number of local collapses in residential and farm buildings (Sorrentino et al. 2013).

As it can be noted in Fig. 2, timber roof structures simply resting on top of perimeter walls, without any capacity of contrasting out-of-plane collapse of the upper part of the walls and in some cases transferring horizontal thrust components, can be often identified as causes of partial collapse. In some cases the collapse involved a higher portion of the façade, when other vulnerability causes were also present (e.g. the damaged building in Moglia, reported at the bottom left of Fig. 2, shows a significant distance between main cross-walls perpendicular to the façade and door openings systematically close to the walls’ intersections).

Fig. 2

Partial out-of-plane collapses of masonry façades triggered by thrusting timber roofs in the historical centre of Concordia sulla Secchia (top) and in Moglia (bottom) (survey on May 30th 2012)

Several strategic buildings (e.g. the Moglia city hall, bottom right of Fig. 2) suffered this type of damage, which made them unusable during the emergency period. As it is clearly shown in Fig. 3, the development of this relatively frequent damage mode ranged from a collapse limited to the upper part of a façade, far from its connection with perpendicular walls, to the complete collapse of the roof structure following the failure of the top part of all supporting masonry walls. In some cases (Fig. 3) overturning of gable walls with low vertical compression were also observed.

Fig. 3

Collapses due to wall-roof interaction in Medolla (top left), San Possidonio (top right and bottom left) and Cavezzo (bottom right) (survey on May 30th 2012)

In many other situations local failure modes resulting from the interaction of roof structure and perimeter walls were only triggered, inducing several damages without however reaching collapse conditions (examples in Fig. 4).

Fig. 4

Damage due to roof-wall interaction observed in Moglia (left) and Concordia sulla Secchia (right)

Buildings which did not suffer this sort of damage generally exhibited a mainly global behaviour governed by the in-plane wall response. This is confirmed by the damage observed in piers and spandrel beams.

During the post-earthquake surveys, a lot of damaged masonry spandrels were found. These elements, sometimes considered of secondary importance, appeared to be rather vulnerable. In-plane shear failure of spandrels was quite common and it was even observed in the absence of a steel tie rod applying an axial (horizontal) compression to the member. This observation suggests that a horizontal compression force could have been generated in the deformed spandrels when their axial deformation was inhibited by the presence of floor diaphragms (right part of Fig. 5).

Fig. 5

Damaged spandrel beams in Mirandola: shear (left) and flexural (right) failure modes

As already observed in laboratory tests (e.g. Beyer and Dazio 2012), an important role in governing the spandrel failure mode was also played by the type of lintel supporting it. Flexural damage was mainly observed for spandrels supported by timber lintels, while spandrels supported by masonry shallow arches generally failed in shear. Failures of spandrels supported by masonry arches (Fig. 6) could have been potentially much more dangerous for people inside and outside buildings due to potential falling of masonry portions.

Fig. 6

Damaged and collapsed masonry lintels supporting spandrel beams (Mirandola)

In several cases damages were concentrated in spandrel beams while the rest of the façades seemed to be almost undamaged, i.e. weak spandrels prevented the diffusion of shear damage to other structural members. On the contrary, in case of stronger spandrels, damage occurred mainly in masonry piers at the first storeys. Figure 7 reports examples of the two aforementioned cases together with an example of a building with both piers and spandrels damaged.

Fig. 7

Buildings with in-plane damage concentrated in piers (left, Novi di Modena), in spandrels (centre, Mirandola) and in both structural members (right, Rovereto sulla Secchia)

Existing masonry buildings with slight to moderate damage were frequently observed. Buildings with proper wall-to-wall, wall-to-floor and wall-to-roof connections, adequate wall thickness and sufficient masonry density (ratio between wall area and total floor area) generally presented limited damages. In the left part of Fig. 8, the damage state of the theatre of Novi di Modena is reported, which was undergoing renovation works and only suffered the local collapse of a parapet wall in the main façade, while other damages were limited by the presence of regularly distributed steel tie-rods. The building reported in the right part of Fig. 8 did not show any evidence of damage.

Fig. 8

Theatre in Novi di Modena (1926, left) and building in Cavezzo (right)

In some buildings heavier and anomalous damages were observed. This was generally related to some deficiencies of the structure causing specific vulnerability of the buildings. An example of the defects causing an increase in the seismic vulnerability of the buildings can be identified in the use of relatively slender walls with large unsupported clear length and height, and large wall-to-wall distances such as in the case of the Sant’Agostino city hall, Fig. 9.

Fig. 9

The heavily damaged city hall of Sant’Agostino (survey on May 30th 2012)

Other examples of damage concentration due to some form of structural irregularity are reported in Fig. 10, where the presence of inclined cracks involving the central part of masonry arches in buildings with colonnades at the ground storey is noticed, and in Fig. 11, where in-plane damage and collapse of masonry turrets rising over the roof gutter or ridge level are evident.

In some cases, multi-leaf clay brick masonry walls were also observed. The impressive collapse of a building in Concordia sulla Secchia (Fig. 12) shows the presence of two unconnected leaves in the collapsed façade rather than a solid brickwork, which could have benefitted from the total wall thickness. In this case, a single-headed brick wall was simply added externally to the pre-existing double-headed wall. Figure 13 presents a zoom of this detail together with other similar examples of partial collapses of the external veneers. Several cases were also observed of multi-leaf masonry walls originally constructed with heavy diaphragms resting on the internal leaf and only partially loading the external veneer. In these cases, instability of the internal leaf occurred, with pounding of the external one.

Fig. 10

Diagonal cracks starting from the central part of arcades in Cento (left) and Moglia (right)

Fig. 11

Damage concentration in turrets: in-plane damage to the turret masonry piers (Moglia, left and centre) and complete collapse of a central turret (San Possidonio, right)

Fig. 12

Building collapse in the historical centre of Concordia sulla Secchia. The picture shows that the building façade was made of two unconnected leaves of clay brick masonry

Fig. 13

Presence of multi-leaf clay brick walls and collapse of the external leaf: detail of collapse in Concordia sulla Secchia (left) and partial collapses in Moglia (top right) and Mirandola (bottom right)

Other structural deficiencies were particularly highlighted by the shaking of relatively high intensity which repeatedly struck the same buildings. An impressive example is reported in Fig. 14 for a masonry building in Cavezzo. The building was actually composed by two different structural units, erected in different periods. The more recent one was simply added against the pre-existing one, without any proper structural connection nor the realisation of a proper seismic joint. The May 20th event evidenced this situation by causing a vertical separation crack due to pounding. The May 29th events significantly increased the damage and caused the collapse of the added structural unit.

Fig. 14

Effect of repeated shaking on a portion of a building erected against a pre-existing building without proper structural connection nor seismic joint: the May 20th event caused the separation of the two independent structural units (left, courtesy of M. Messori) while the May 29th events determined the final collapse of the added one (right)

As also evident from Table 1, the final damage condition of the buildings in the affected area, and in particular in the Western part of the area damaged by the main event of May 20th, has to be associated with damage accumulation. Almost all events in the seismic sequence with magnitude higher than 5.0 induced peak ground acceleration values higher than 0.25 g at the recording station of Mirandola.

Lack of maintenance certainly contributed in many cases to increase the seismic vulnerability of existing masonry buildings, in particular for what concerns the effect of timber degradation in reducing the structural efficiency of roof and floor systems. However, the mechanical quality of traditional solid brick masonry is also rather sensitive to mortar quality, and in several of the most damaged structures the quality of mortar seemed to be rather low, partly due to weathering/degradation.

As already evidenced by past earthquakes in Italy, also in the case of the Emilia earthquakes several issues related to seismic vulnerability of historical centres arose, such as for example structural behaviour of building aggregates, negative interaction in the presence of irregularities in elevation between neighbour buildings, pounding of adjacent structures. Due to severe damages of a relatively limited number of buildings, to the difficulties of escape and rescue in narrow streets potentially occupied by debris and to the risk of falling of rubble material and non-structural elements, the historical centres of the affected cities were closed and the population evacuated. This also led in some cases to traffic detour, such as for the case of Rovereto sulla Secchia. Obviously all useful safety measures have to be implemented in these cases, but recent examples (e.g. Nocera Umbra, L’Aquila) also showed that the longer is the inactivity period of the city centre the higher are the socio-economic costs added to those directly associated with the earthquake damage.

3 Structural damage in modern masonry buildings

Typical modern masonry structures in this area of Italy are constituted by unreinforced masonry buildings usually not more than three storey high, with a prevalence of masonry erected with vertically perforated clay units, although also different types of concrete blocks are used. Several reinforced masonry buildings, mainly built with clay units or lightweight aggregate concrete (LAC), are also present in the area. The unreinforced clay masonry is usually constructed using medium to large thick units, usually in the range between 25 up to 45 cm, having a void ratio of about 45 % and mechanical properties which should guarantee, in accordance with the latest national norms, a sufficient robustness for avoiding local brittle failure. Both general purpose and, more recently, thin layer mortar bed-joints are used and head-joints are filled by mortar on the vertical face of the plain units or in the mortar pockets. For the erection of autoclaved aerated concrete (AAC) masonry, solid units and thin mortar joints are instead used. The reinforced masonry is usually constituted by units with proper holes which allow the casing of vertical steel bars, whereas the horizontal reinforcement is usually placed in the thickness of the bed-joints (general purpose mortar, 10–15 mm thickness).

Moreover, recent masonry buildings should have been conceived, designed and detailed to avoid the main sources of vulnerability of old masonry structures, following the national design provisions which promote the use of regular and robust units, the limitation of the slenderness of the walls, the effectiveness of the connections between intersecting walls and between floor/roof and walls through reinforced concrete ring beams at each floor (and roof) level to favour “box action”, a sufficient in-plane stiffness of the floor/roof diaphragms and the regularity of the structure. Conversely, prior to 2003, i.e. before the revision of the seismic classification and the enforcement of the seismic design code, modern masonry buildings erected in the area were only conceived to resist vertical loads and wind action. This often resulted in an inadequate building lateral strength due to an insufficient area of masonry walls in the two orthogonal directions, possible irregularities in plan and in elevation and differences in the required structural details (e.g. reinforced concrete ring beams, minimum wall length, intersections of orthogonal walls) and minimum mechanical properties of units and mortar (e.g. minimum compressive strength).

After the seismic events of May 2012, several post-earthquake surveys were carried out on different types of modern structural masonry buildings in order to evaluate the seismic performance of the different structural systems also as a function of the estimated or measured seismic action. In particular, inspections on 70 buildings were performed: 47 on unreinforced clay masonry buildings, 16 on reinforced clay masonry buildings, 4 on unreinforced AAC masonry buildings and 3 on reinforced lightweight concrete masonry buildings. In some cases, isolated r.c. vertical elements (normally r.c. columns or slender r.c. walls) were present in the internal parts of the buildings and also outside the buildings, for example to bear the portico. Among the structures inspected, the seismic response of five masonry buildings (an unreinforced clay masonry, a reinforced clay masonry, an unreinforced AAC masonry, a reinforced masonry with lightweight concrete units and an unreinforced clay masonry building constructed in the late ‘60s), taken as reference of the typical behaviour of the structural typologies, is described in the following. Moreover, the case of a residential district in Rovereto sulla Secchia, in the municipality of Novi di Modena, where some unreinforced masonry buildings have manifested relevant damages, is also reported. The shaking experienced during the two main events of May 20th and May 29th was estimated in order to provide a rough correlation of the performances of the masonry structures with PGA. Green dots numbered from 1 to 6 are reported on the map in Fig. 1 to identify the six sites. The first case is a two-storey unreinforced clay masonry building constructed in 2009 and located in Casoni di Sopra, a part of the municipality of Finale Emilia at about 9 km South/South West from downtown, at an epicentral distance of about 10 km for both events of the 20th and the 29th of May. The estimation of the PGA for this site is 0.20–0.30 g for the first event and 0.26 g for the second one. The building is constituted by 45 cm thick peripheral walls and 30 cm thick internal walls with vertically perforated lightweight clay units having a void ratio of about 45 %, thin layer bed-joints and mortar pocket head-joints. The intermediate floor and the roof are rigid in their plane and the structure is substantially regular in plan and in elevation. No cracks or any kind of damage was observed in the internal and in the external walls and in all other structural and non-structural components. The picture reported in Fig. 15a, taken just after the main seismic event of the 29th of May 2012, shows the external part of the building, exempt from any damage.

Fig. 15

Examples of undamaged buildings: a two-storey unreinforced clay masonry building in Casoni di Sopra, municipality of Finale Emilia; b three-storey reinforced clay masonry buildings in S. Felice sul Panaro; c single-storey unreinforced masonry building with AAC units in Carpi; d four-storey reinforced masonry building with LAC units in Vigarano Mainarda

The second case is an undamaged three-storey reinforced masonry building erected with vertically perforated clay units (see Fig. 15b). The external and the internal walls are, respectively, 30 and 25 cm thick and the structural configuration is very regular in plan and in elevation. This residential house, constructed in 2009, is located in S. Felice sul Panaro at a distance of 10 and 4 km from the epicentres of the two main seismic events. The value of estimated PGA which has stricken the structure at the 20th of May is 0.29 g (Bozzoni et al. 2012), whereas the acceleration registered after the 29th of May main earthquake at the seismic station of S. Felice sul Panaro, temporary installed at few hundred meters from the building, was equal to 0.22 g.

The third case is a single storey unreinforced masonry building with autoclaved aerated concrete (AAC) units and thin layer mortar joints having 40 cm thick walls, without plaster, built in 2006 and used as an agricultural warehouse; the timber roof is set on two different levels. The structure is located in the municipality of Carpi at about 4 km South East from downtown, at epicentral distances of 31 and 19 km and PGA of 0.01–0.10 g and 0.14 g as respect of the two main seismic events. No damage on this building is visible (see Fig. 15c).

The fourth case is a residential reinforced masonry building with LAC units, constructed in 2008 in Vigarano Mainarda, made up by 2 storeys with an attic floor. The structural thickness of the walls is 24.5 cm and the units are constituted by an assemblage of a load-bearing part, an insulation panel and an external non-structural masonry block. The vertical reinforcement is placed in sufficiently large holes of the units which allow to cast small r.c. pillars, whereas the horizontal reinforcement is placed in the mortar bedjoints. Making reference to the two main seismic events considered above, the epicentral distances are about 22 and 33 km and the values of the estimated PGA are 0.10–0.20 g and 0.00–0.02 g. Although the seismic action of the reference earthquakes was not particularly high, further rather intense earthquakes have stricken the vicinity of this area, in particular the magnitude $\hbox {M}_{\mathrm{W}}$ 5.1 event of the 20th of May at 3.18 pm (about 11 h after the main earthquake) with the epicentre located at only 1.5 km from this building. As illustrated in Fig. 15d, the building does not show any type of damage.

Figure 16 reports pictures of some of the damaged parts of a four-storey unreinforced masonry building constructed in the late ‘60s, located in Finale Emilia (epicentral distance equal to 8 and 16 km and PGA equal to 0.30 and 0.23 g for the events of the 20th and the 29th of May, respectively). The damage is mainly concentrated at the ground storey, where wide diagonal cracks have occurred in the external and in the internal structural walls and also in the partitions. Masonry walls with vertically perforated clay units having a thickness of 20 cm are connected to mixed r.c. joist-clay tile floors through r.c. ring beams.

Fig. 16

Four-storey unreinforced clay masonry building in Finale Emilia built without any seismic design

This building, although showing an adequate connection between structural walls and sufficiently rigid diaphragms, was not calculated for seismic resistance and does not fulfil all the modern concepts of seismic design for the lack of a sufficient amount or “density” of walls, especially along one of the two main directions. This deficiency leads to concentration of large vertical stress levels in particular on the slender and thin piers at the ground storey, causing a reduced in-plane deformation capacity of the walls for shear mechanisms with the occurrence of typical diagonal shear cracks. Moreover, a higher number of storeys is usually acknowledged as incrementing seismic vulnerability. This was one of the few cases in which this tendency was clearly observed. In most other cases of damaged masonry buildings the influence of the number of storeys was blurred by other main vulnerability factors such as lack of appropriate connections and presence of flexible diaphragms, confirming the outcome of previous post-earthquake observations in Italy (Rota et al. 2011).

A great extent of structural and non-structural damages on recently constructed infilled r.c. frames and structural masonry buildings is reported in a residential district located in the North/North West part of Rovereto sulla Secchia, in the municipality of Novi di Modena at epicentral distances of about 22 and 11 km from the two main seismic events (Fig. 17).

Fig. 17

Some examples of damage in the unreinforced masonry buildings in a district of Rovereto sulla Secchia, municipality of Novi di Modena

In this area, 12 three and four storey modern unreinforced masonry buildings with clay unit structural walls were surveyed. The majority of these buildings were built between the late ‘90s and 2003, one was in construction in 2003 and one was built after 2003; 6 buildings out of 12 are constituted by two different typologies of three identical buildings each. Among the 12 surveyed buildings, only two houses were found to be exempt by any visible damage or only slightly cracked and still in use. The other structures considered were unusable since affected by different levels of damage varying from moderate to very severe. Significant damages were concentrated mainly in the walls at the ground storeys where diagonal and bi-diagonal cracks, being in some cases very wide, and failures at the corners of L shape flanged walls occurred. Some bi-diagonal shear cracks were also found at the first storey of the buildings.

Most of the significant cracks were oriented in North/North East direction on the masonry piers parallel to the main street of entrance of the dwellings, where the walls, in some cases, possess large openings (such as the garage doors) and therefore lower amount of structural masonry, but also suggesting a possible “directivity” effect of the seismic action. The PGA of this site was estimated to be 0.10–0.20 g and 0.24–0.26 g for the two seismic events, according to the criteria defined in the introduction. However, possible local site effects of ground acceleration amplification may also have occurred. The earthquake has in fact hit strongly and almost indiscriminately different structural typologies (structural masonry and infilled r.c. buildings) in a limited area surrounded by adjacent zones with a much lower extent of damage. Therefore, the high level of damage to the structures constructed in this area could be attributed in part to their own vulnerability, considering also that the majority of these buildings has been built before the seismic reclassification of the site and most likely without the application of seismic design principles, but possibly also to a higher level of ground motion due to site effects.

4 Non-structural damage

As mentioned in the introduction, non-structural damage was responsible of two casualties in the May 29th Mw 5.8 event. One of the most common examples of non-structural damage in both new and existing buildings consisted of the damage occurring in masonry or precast chimneys. Figure 18 presents some examples of these damages, which tend to separate the chimney into different blocks. Rigid body rocking and sliding are then activated, often leading to the collapse of the non-structural element.

Fig. 18

Examples of damage to precast and masonry chimneys

Fig. 19

Examples of non-structural damages in existing masonry buildings: collapse of sunshade elements made of clay bricks (top left), collapse of a balcony stone balustrade (bottom left) and displacement of roof tiles (right pictures)

Some peculiar non-structural damages were also observed in existing masonry buildings. Figure 19 shows examples of damages to the roof tiles and to balustrades and clay brick sunshade elements. All these damages can be particularly dangerous for the safety of people walking just outside (or exiting) the building. Roof tile displacements may also cause seepage of rainwater and further degradation of timber elements.

5 Conclusions

The damage to old existing clay brick masonry buildings in the area affected by the 2012 Emilia earthquake sequence evidenced some recurrent vulnerability elements typical of masonry structures which were not conceived to resist to seismic actions. Nevertheless, majority large part of existing masonry buildings showed some inherent capacity to resist to horizontal actions. With respect to the damage surveys performed after Italian (e.g. D’Ayala and Paganoni 2011) and international (e.g. Javed et al. 2006) earthquakes affecting areas with a building stock mostly consisting of stone masonry buildings, the failure modes observed in the clay brick masonry buildings of Emilia showed the occurrence of out-of-plane failure modes limited to some specific cases, i.e. presence of thrusting roofs (or roofs with very limited diaphragm action and poor connections to wall edges) and/or combinations of large wall slenderness or unrestrained wall lengths and some structural irregularity. Mortar quality seems to be also an issue for some buildings, since in-plane capacity of old solid brick masonry is sensitive to mortar strength.

The outcome of the post-earthquake inspections on modern masonry buildings have evidenced, in the great majority of the cases, a good seismic performance without the occurrence of any significant damage in structural and non-structural elements, even in zones very close to the epicentres of the main seismic events. Such satisfactory seismic behaviour has been achieved on those low-rise (two- or three-storey maximum) modern masonry buildings characterized by a sufficient “wall density” or amount of walls, built with good quality materials and in accordance with the more recent detailing principles; similar situations have also been identified from the experimental and the past-earthquake experiences. After the 1992 Erzincan earthquake, Saatcioglu and Bruneau (1993) also noted that the extremely good performance exhibited by some URM buildings was mainly attributable to their high structural redundancy. In a more detailed analysis on the performance of masonry buildings in the same event, Sucuoglu and Erberik (1997) concluded that URM complying with seismic code requirements possess a remarkable lateral load resistance, in particular in case of regular structures with diaphragm action, which also provide significant shear redistribution among walls and overstrength. However, on the other hand, a limited number of recent unreinforced masonry buildings suffered a rather large extent of damage, in the form of shear diagonal and bi-diagonal cracks at the ground floor. A possible explanation for this could be that the majority of these damaged buildings had been built prior to 2003 with no seismic design (lack of a proper conceptual seismic design, possible deficiencies in the quality of the materials and in the structural details, low amount of structural walls), although the hypothesis of local soil effects should be further investigated for the site located in Rovereto sulla Secchia.

The performance of non-structural elements such as chimneys and balustrades was in many cases so poor that it compromised the safety conditions of the occupants of slightly damages structures, in particular in historical centres. Higher importance should be given to properly designing, detailing or retrofitting these components.