Performance of unreinforced masonry buildings during the 2010 Darfield (Christchurch, NZ) earthquake *

Transcription

1 1 Performance of unreinforced masonry buildings during the 2010 Darfield (Christchurch, NZ) earthquake * J Ingham Department of Civil and Environmental Engineering, University of Auckland, New Zealand M Griffith School of Civil, Environmental and Mining Engineering, University of Adelaide, South Australia, Australia ABSTRACT: The 2010 Darfield earthquake caused extensive damage to a number of unreinforced masonry buildings. While this damage to important heritage buildings was the largest natural disaster to occur in New Zealand since the 1931 Hawke s Bay earthquake, the damage was consistent with projections for the scale of this earthquake, and indeed even greater damage might have been expected. In general, the nature of damage was consistent with observations previously made on the seismic performance of unreinforced masonry buildings in large earthquakes, with aspects such as toppled chimneys and parapets, failure of gables and poorly secured face-loaded walls, and in-plane damage to masonry frames all being extensively documented. This report on the performance of the unreinforced masonry buildings in the 2010 Darfield earthquake provides details on typical building characteristics, a review of damage statistics obtained by interrogating the building assessment database that was compiled in association with post-earthquake building inspections, and a review of the characteristic failure modes that were observed. It was observed that structures that had been seismically retrofitted appeared to perform well, with further study now required to better document the successful performance of these retrofit solutions. 1 INTRODUCTION At 4.35 am on the morning of Saturday 4 September 2010, a magnitude 7.1 earthquake occurred approximately 40 km west of the city of Christchurch, New Zealand, at a depth of about 10 km (GNS, 2010), having an epicentre located near the town of Darfield. The ground motion had a peak ground acceleration (PGA) of about 0.25g and a spectral acceleration in the plateau region of about 0.75g, which corresponds well with the design spectra for a site class D soil site in Christchurch for spectral periods greater than 0.2 s. In general, the earthquake represented 67% to 100% of the design level event, depending upon the spectral period being considered (see figure 1), with most of Canterbury reporting damage consistent with MM8 on the Modified Mercalli intensity scale. Hence, significant damage to earthquake prone buildings and to non-structural components might * Reviewed and revised version of paper originally presented at the 2009 Australian Earthquake Engineering Society Conference, December 2009, Newcastle, NSW. Corresponding author A/Prof Michael Griffith can be contacted at mcgrif@civeng.adelaide.edu.au. have been expected, with associated fatalities and major casualties. Instead, the single most striking statistic was that there were no fatalities directly associated with the earthquake (although there was one heart attack fatality and one person hospitalised due to a falling chimney (NZ Herald, 2010)); and the overall impression is that damage in the central business district (CBD) was reasonably contained, restricted primarily to unreinforced masonry (URM) buildings, and damage to windows in taller steel and concrete structures. The absence of fatalities and more extensive damage is attributed to the comparatively high level of seismic design capability in New Zealand, and the fact that the CBD, which is the region containing the highest density of URM buildings, was almost completely unoccupied at 4.35 am. Despite the number of buildings that had received some form of seismic improvement, a question currently remains as to why a greater number of URM buildings did not suffer significant damage in the earthquake. The soil conditions in Christchurch have three separate material types: river outwash gravels, sands, and marshy ground in former swamp Institution of Engineers Australia, 2011 Australian Journal of Structural Engineering, Vol 11 No 3

2 2 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith 5% damped Response Spectrum Acceleration (g) NZSEE Class D Deep or Soft Soil Median of Soft Soil Sites Botanical Gardens CBGS-H1 Botanical Gardens CBGS-H2 Cathedral Collage CCCC-H1 Cathedral Collage CCCC-H2 Hospital CHHC-H1 Hospital CHHC-H Period T(s) Acceleration(g) Acceleration(g) N26W component Cathedral College site Time(s) N64E component Cathedral College site N26W N64E Time(s) Fig ure 1: Details of earthquake ground motion acceleration spectra, and acceleration time-history record (source: GeoNet). areas, with the central city built mainly on gravels, although there are pockets of sand and soft soil in former marsh deposits. The earthquake ground motion characteristics shown in figure 1 reflect the underlying soil condition, with the long period nature of the motion resulting from the soft soils upon which Christchurch is founded. These soft soils effectively act as a filter and remove high frequency ground motion (leading to smaller PGA values than on rock sites), but amplify long period motion, resulting in significantly larger longer period motion at about 2.5 s on several of the softer soil sites. As most URM buildings have a fundamental period of s, the underlying ground conditions appear to have assisted in reducing the seismic demand in this period range to approximately 70% of the current NZS (Standards New Zealand, 2004) design level loading for a site class D soil site in Christchurch. As can be seen in figure 1, strong ground shaking in the CBD had a duration of approximately 30 s, with similar amplitudes in the two orthogonal recording directions. This lack of distinct directionality probably explains why parapet failures were observed in streets running in both the north-south and east-west directions. 2 BACKGROUND It is well established that URM buildings perform poorly in large magnitude earthquakes, with a brief selection of relevant prior earthquakes that have caused major damage to URM buildings being detailed below: The M Hawke s Bay (New Zealand) earthquake caused widespread devastation to URM buildings in the city of Napier (see figures 2 and 2), with 256 fatalities. This earthquake remains the worst disaster of any type to occur on New Zealand soil (Dowrick, 1998). The M Newcastle (Australia) earthquake resulted in 13 fatalities and over 160 casualties. The earthquake damaged approximately 50,000 buildings (80% of these were homes) with URM buildings most widely affected (Page, 1991) and to date is the only earthquake to have caused any fatalities in Australia (see figures 2 and 2). The M Gisborne (New Zealand) earthquake caused damage to numerous URM buildings, including the collapse of 22 parapets (Davey & Blaikie, 2010). Examples of damage to URM buildings are shown in figures 2(e) and 2(f). The M Maule (Chile) earthquake caused extensive damage to older houses, churches and other buildings constructed of URM or adobe, as shown in figures 2(g) and 2(h) (EERI, 2010). The M Kalgoorlie-Boulder (Australia) earthquake occurred near Kalgoorlie-Boulder, causing damage to historic buildings in the city (see figures 2(i) and 2(j)). There were no fatalities, but two people were treated at Kalgoorlie Hospital for minor injuries resulting from the earthquake (Edwards et al, 2010). An initial evaluation procedure (IEP) is provided in NZSEE (2006) as a coarse screening method for determining a building s expected performance in an earthquake. The purpose of the IEP is to make an initial assessment of the performance of an existing building against the standard required for a new building, ie. to determine the Percentage New Building Standard (%NBS). A %NBS of 33 or less means that the building is assessed as potentially earthquake prone in terms of the Building Act (New Zealand Parliament, 2004) and a more detailed evaluation will then typically be required. A %NBS of greater than 33 means that the building is regarded as outside the requirements of the Act, and no further action will be required by law, although it may still be considered as representing an unacceptable risk and seismic improvement may still be recommended (defined by the New Zealand Society for Earthquake Engineering (NZSEE) as potentially earthquake risk ). A %NBS of 67 or greater means that the building is not considered to be a significant earthquake risk. NZSEE noted that: A %NBS of 33 or less should only be taken as an indication that the building is potentially earthquake prone and a detailed assessment may

3 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith 3 (e) (f) (g) (h) (i) (j) Fig ure 2: Representative examples of damage to URM buildings in past earthquakes overlooking Napier at the buildings destroyed by the 1931 earthquake and fires (source: Alexander Turnbull Library); view down Hastings Street, Napier, after the 1931 earthquake (source: Alexander Turnbull Library); out-of-plane wall failure in the 1989 Newcastle earthquake; parapet and cavity wall damage in the 1989 Newcastle earthquake; (e) toppled parapet in the 2007 Gisborne earthquake; (f) out-of-plane failure of a gable wall in the 2007 Gisborne earthquake; (g) 1940s URM house, Maipu Street, Concepcion, Chile (source: Andrea Garcia); (h) Barros Arana Institute, Barros Arana corner with Angol, Concepcion, Chile (source: Andrea Garcia); (i) damage to parapet and awning in Boulder; and (j) toppled parapet in Boulder.

4 4 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith well show that a higher level of performance is achievable. The slight skewing of the IEP towards conservatism should give confidence that a building assessed as having a %NBS greater than 33 by the IEP is unlikely to be shown, by later detailed assessment, to be earthquake prone. Recent research has suggested that there are approximately 3750 URM buildings in New Zealand (Russell & Ingham, 2010), with their distribution throughout New Zealand being aligned with the relative prosperity of communities during the period between approximately 1880 and Following the Darfield earthquake, it was reported that the Canterbury region had approximately 7600 earthquake-prone buildings, with 958 of these buildings being constructed of URM (Christchurch City Council, 2010; Wells, 2010). 3 ARCHITECTURAL CHARACTER The architectural features of the URM building stock in the Canterbury region of New Zealand are consistent with the general form of URM buildings throughout both New Zealand and Australia (Russell & Ingham, 2008; 2010). These buildings can be characterised as typically 2 or 3 storeys in height, with 2-storey buildings being most common, and being either stand-alone or row buildings, but with row buildings being most common (refer figure 3). In particular, similarities in architectural character between URM buildings in New Zealand and Australia are evident by comparing the images shown in figure 2. A common architectural feature is for passing pedestrians to be covered by an overhead awning that is located at the height of the first floor, which is typically braced back to the piers of the second floor. Figure 3 shows this detail. Post-earthquake inspection of building performance led to 595 URM buildings being assessed, out of the 958 URM buildings reported to exist in Christchurch. It is believed that the majority of unassessed URM buildings were undamaged and outside the primary inspection zone associated with the CBD and arterial routes extending from the central city. General features of the 595 assessed URM buildings are reported in figure 4, indicating that the majority of buildings were either 1- or 2-storey, consistent with prior findings by Russell & Ingham (2010). Figure 4 shows that the most common occupancy type was commercial or office buildings, and hence the majority of buildings were unoccupied at the time of the earthquake, significantly contributing to the lack of direct earthquake fatalities. The survey forms contained a field to record the estimated gross floor area of the building, and hence the estimated building footprint could be determined once accounting for the number of storeys (see figure 4), but the data is unfortunately incomplete as only 301 entries were recorded for the 595 separate buildings assessed. It is not possible to establish from the database whether individual entries belonged to a stand-alone or a row building. Fig ure 3: Pre-earthquake examples of typical URM building architecture in Christchurch 188 Manchester Street; corner of Sandyford and Colombo Street; 118 Manchester Street; and corner of Colombo and Tuam Street.

5 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith 5 3+ Storey 17% 1 Storey 29% % % 50 11% % Industrial 2% Other residential 7% Heritage Listed 3% Other 3% Public assembly 5% Religious 2% Dwelling 11% 2 Storey 54% % Commercial /Offices 67% Fig ure 4: Building characteristics derived from interrogation of the inspection database storey height (595 entries); footprint area (m 2 ; 301 entries); and occupancy type (595 entries). Fig ure 5: Masonry rubble from collapsed wall masonry rubble showing clean bricks; and weak mortar crumbles between fingers. Fig ure 6: Large sections of masonry intact after fall from buildings solid section of masonry gable; and solid section of parapet. 4 MATERIAL PROPERTIES The general observation from the debris of collapsed URM walls was that the kiln-fired clay bricks were generally of sound condition, but that the mortar was in poor condition. In most cases the fallen debris had collapsed into individual bricks, rather than as larger chunks of masonry debris (refer to figure 5). When rubbing the mortar that was adhered to bricks it was routinely found that the mortar readily crumbled when subjected to finger pressure (refer figure 5), suggesting that the mortar compression strength was very low. However, it appears that superior mortar was often used in the ornate parapet above the centre of the wall facing the street, as this segment of the collapsed parapet often remained intact as it collapsed (refer figure 6). 5 BUILDING DAMAGE STATISTICS In general, the observed damage to URM buildings in the 2010 Darfield Earthquake was consistent with the expected seismic performance of this building form, and consistent with observed damage to URM buildings both in past New Zealand and Australian earthquakes and in numerous earthquakes from other countries (see figure 2). As part of the emergency

6 6 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith response to this earthquake, the authors spent 72 hours assisting Christchurch City Council with building damage assessments, tagging buildings with either a green, yellow or red placard depending, respectively, upon whether a building was safe for public use, had limited accessibility for tenants/ occupants, or was not accessible (refer to figure 7). Many examples of earthquake damage were observed during this exercise, as well as many examples of seismic retrofits to URM buildings that had performed well. The results from the interrogation of damage assessment information are reported in figure 8. Figure 8 reports the useability assignment of the 595 URM buildings assessed. In consultation with staff of Christchurch City Council it is believed that the complement of the 958 URM buildings thought to exist in Christchurch probably had a green tag usability rating, and so this theorised damage distribution for the entire URM building stock is shown in figure 8. Figure 8 reports the level of damage for the 595 buildings that were surveyed by the Rapid Building Assessment teams. No explicit definitions were provided to the assessment teams for what represented and range (%) damage. The values recorded by the teams for each building surveyed were simply estimates (excluding contents damage). Despite the known vulnerability of URM buildings to earthquake loading, 395 of the 595 buildings (66%) were rated as having 10% damage or less, with only 162 (34%) of the buildings assessed as having more than 10% damage. It was also possible to study the distribution of damage dependent on storey height (figure 8), with the data indicating no definitive trend and a comparatively uniform level of damage assigned to buildings in each height category. 6 CHIMNEYS Unsupported or unreinforced brick chimneys performed poorly in the earthquake (figure 9), with numerous chimney collapses occurring in domestic as well as small commercial buildings and some churches. Many examples of badly damaged chimneys that were precariously balanced on rooftops were also seen (figure 9) and it was reported that one week after the earthquake, 14,000 insurance claims involving chimney damage had been received, from a total of 50,000 claims (NewstalkZB, 2010). Emergency services personnel were in significant demand, being deployed to remove damaged chimneys in order to minimise further risk and eliminate these falling hazards (figure 9). In contrast, figure 9 shows an example of a braced chimney that performed well. Note that figure 9 shows further evidence of the poor performance of mortar during the earthquake. 7 GABLE END WALL FAILURES Many gable end failures were observed, often collapsing onto or through the roof of an adjacent building (refer figures 9 and 10). However, there Fig ure 7: Building assessment notices Urban Search And Rescue personnel applying an assessment notice; green tag; yellow tag; and red tag.

7 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith 7 Red 21% Red 13% Green 47% Yellow 17% Yellow 32% Green 70% Percentage of Buildings Surveyed 30% 24% 18% 12% 6% Unknown 3+ Storeys High 2 Storeys High 1 Storey High Number of Buildings Surveyed 0% 0 none 0 1% 2 10% 11 30% 31 60% 61 99% not spec Extent of Damage to Building Fig ure 8: Damage statistics distribution of placard assignments (595 entries); theorised damage distribution for entire building stock (958 entries); and extent of building damage. Fig ure 9: Examples of chimney performance during the Darfield earthquake two damaged chimneys and gable wall; unstable damaged chimney; emergency service workers remove chimney; and braced chimney performed well.

8 8 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith Fig ure 10: Examples of gable end wall failures 93 Manchester St; 816 Colombo St; Montreal-Armagh street corner; and Kilmore-Montreal street corner. were also many gable ends that survived; many more than might have been expected, with the majority having some form of visible restraints that tied back to the roof structure. These examples are shown and discussed later (refer figures 17-18). 8 PARAPET FAILURES Numerous parapet failures were observed along both the building frontage and along their side walls, and for several URM buildings located on the corners of intersections, the parapets collapsed on both perpendicular walls (refer figure 11). Restraint of URM parapets against lateral loads has routinely been implemented since the 1940s, so while it is difficult to see these restraints unless roof access is available, it is believed that the majority of parapets that exhibited no damage in the earthquake were provided with suitable lateral restraint. In several cases, it appears that parapets were braced back to the perpendicular parapet, which proved unsuccessful. 9 ANCHORAGE FAILURES Falling parapets typically landed on awnings, resulting in an overloading of the braces that supported these awnings and leading to collapse. Most awning supports in Christchurch involved a tension rod tied back into the building through the front wall of the building. Many of these connections appear to consist of a long, roughly 25 mm diameter rod, with a rectangular steel plate (about 5 mm thick) at the wall end that is about 50 mm wide 450 mm long, and fastened to the rod and positioned either inside the brick wall or in the centre of a masonry pier or wall. In most cases the force on the rod exceeded the capacity of the masonry wall anchorage, causing a punching shear failure in the masonry wall identified by a crater in the masonry (refer figure 12). 10 WALL FAILURES Out-of-plane wall failures were the first images to appear on television directly after the earthquake. Inspection of this damage typically indicated poor or no anchorage of the wall to its supporting timber diaphragm. Several examples of wall failure are shown in figure 13. Figure 13 shows a corner building that had walls fail in the out-of-plane direction along both directions. Figure 13 shows a 3-storey building where walls in the upper two storeys suffered out-of-plane failures and figure 13 shows similar damage for a 2-storey building. In all three of these instances, it appears that the walls were not carrying significant vertical gravity loads, other than their self weight, due to the fact that the remaining roof structures appear to be primarily undamaged. In contrast, figure 13 shows an outof-plane failure of a side wall that was supporting the roof trusses prior to failure. As shown in figure 14, several examples of face load wall failure closely resembled observed damage in dry stack masonry experiments (Restrepo-Velez & Magenes, 2009), providing further support to the

9 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith 9 Fig ure 11: Examples of typical parapet failures multiple front wall parapet failures; corner of Sandyford and Colombo Street (see also figure 3); side wall parapet collapse onto roof; and corner Columbo and Tuam Street (see also figure 3). Fig ure 12: Anchorage failure of awning brace due to parapet collapse anchorage failure; and close-up of failed anchorage detail. Fig ure 13: Examples of out-of-plane failures in solid masonry walls corner Worcester and Manchester streets (see also figure 3); 118 Manchester Street (see also figure 3); 179 Victoria Street; and out-of-plane failure of a side wall that was supporting the roof trusses prior to failure.

10 10 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith Fig ure 14: Failure mechanism comparisons (observed earthquake damage versus experimental simulation) wall damage at 140 Linchfield Street; and high speed photograph of a dry-stacked masonry wall failing during a tilt test. Fig ure 15: Examples of out-of-plane failures in cavity walls cavity wall failure in a residential building; 832 Columbo Street; butterfly wall ties still intact; and metal wall ties badly deformed. supposition that many of the wall failures were partly attributable to poor mortar strength. Cavity wall construction is generally believed to be much less common in New Zealand than is solid multi-leaf (or multi-wythe) construction. However, cavity wall construction can be extremely vulnerable to out of plane failure in earthquakes in situations where the cavity ties were poorly installed, or more commonly have corroded over time, as the wall is then comparatively slender and less stable than for solid construction. Figures 15 and 15 show examples of cavity wall buildings that suffered outof-plane wall failures. Figures 15 and 15 show that cavity ties were present but were insufficient to prevent the outer leaf from failing. In some cases wall-diaphragm anchors remained visible in the diaphragm after the wall had failed, again indicating that failure had occurred due to bed joint shear in the masonry (refer figure 16). Figure 16 shows a situation where a diaphragm anchor had been embedded within the wall. It can be seen that the anchor successfully prevented the restrained wall from failing, but was not able to prevent toppling of the parapet that was located above the anchor. 11 SUCCESSFUL WALL ANCHORAGE A significant feature of the earthquake was the number of occasions where anchored walls performed

11 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith 11 Fig ure 16: Wall-to-diaphragm anchor details gable end wall failure despite anchor (see also figure 13); and wall anchor still intact (see also figure 6). Fig ure 17: Successful gable end wall and side wall anchorages on an arts centre building. Fig ure 18: Successful wall-floor and wall-roof diaphragm anchorages front elevation; and side elevation. well during the earthquake. Photographs showing this are presented in figures A typical wall-todiaphragm (roof or floor) anchor typically consists of a long 20 mm bolt with a large circular disc of about mm diameter between the wall exterior and nut that clamped the disc to the wall. This detail is shown quite clearly in figure IN-PLANE WALL FAILURES Where walls exhibited some damage to in-plane deformation the cracks were mostly seen to pass vertically through the lintels over door or window openings. Examples are shown in figure 19. However, this type of damage was not widely observed. 13 PARTIAL WALL FAILURES Another interesting feature of this earthquake is the observation of walls that only partly failed, allowing for identification of the specific failure mode at its onset. Several excellent examples are described below. The first of these is a 2-storey URM building on Ferry Road (see figure 20), where the front, street facing, wall of the building had started to fail out-of-plane despite the presence of wall-roof diaphragm anchors. As is shown, the anchors were on the verge of pulling through the masonry wall. Internal inspection of the building revealed that the front wall had separated from the long side walls of the building and moved approximately 50 mm

12 12 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith Fig ure 19: Examples of in-plane wall damage above window openings extensive vertical cracking above window openings; vertical crack above window opening; vertical crack through spandrel; and diagonal crack extending from window opening. Fig ure 20: Wall-roof anchorage failure and partial wall failure building overview; detail of partial anchorage failure; onset of anchorage failure; and internal view showing wall separation. towards the road with respect to the ceiling/roof diaphragm (figure 20). It is believed that due to the nature of strength degradation of the brickwork at the onset of a punching shear failure, the anchorage has effectively failed and offers little residual resistance against further shaking. The only reason the wall did not completely collapse is probably due to the earthquake not imposing sufficient displacement on the wall after the anchorage failure. A similar style of partial failure was observed in another building on Ferry Road (figure 21) but the authors were only able to observe the building externally. It should be noted that the buildings were not in close proximity to each other. An example

13 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith 13 Fig ure 21: Partial bed joint shear failure surrounding anchorage detail wall-roof anchorage failure; and gable end anchorage failure. Fig ure 22: Examples of yielded wall anchors overview of wall anchors; and close-up view of yielded anchor. of a gable end partial failure is shown in figure 21, which can be compared to the comparable anchorage detail shown in figure 16 that resulted in complete failure. There were frequent examples of wall-diaphragm anchors that had deformed plastically. In these photographs (figure 22), the circular plate can be seen to be slack due to plastic stretching of the anchor rod. 14 DIAPHRAGM DEFORMATIONS There was one instance where it was clear that diaphragm deformation, relative to the in-plane walls, contributed to partial failure of an out-of-plane wall. Figure 23 shows several views of a building that suffered out-of-plane parapet failure along its long, side walls. In figure 23 it can be seen that the roof joists have tilted towards the front of the building. This suggested that the front wall of the building was driven forward at its top. Careful inspection of the front wall (figure 23) revealed a substantial outwards curvature that was most pronounced at the top. 15 RETURN WALL SEPARATION Many buildings exhibited substantial cracking between their front wall and side (return) walls. This damage is not necessarily a catastrophic problem if stiff horizontal diaphragms are well connected to the walls in both directions, but where there is not good diaphragm connectivity, there is the potential for complete out-of-plane collapse of one or both walls. Figure 24 shows some examples where major cracking was observed between the side return walls and the front parapet and wall. 16 POUNDING Several instances of damage due to buildings pounding against each other during the earthquake were observed. Figure 25 shows how the shorter building in the centre, which has different floor heights than the building to the left, damaged the column of the taller building at its top storey. 17 SPECIAL BUILDINGS 160 Manchester Street is a 7-storey office building that is reported to consist of load bearing masonry and is the most significant masonry building, at least in terms of height, in Christchurch (figure 26). It is a registered heritage building and is a significant part of the fabric of the Christchurch city landscape. Unfortunately, the building suffered significant damage in the earthquake. The bottom two storeys

14 14 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith Fig ure 23: Example of diaphragm deformation causing out-of-plane wall failure overview of building; side-view of tilted joists; and front wall curvature. Fig ure 24: Examples of wall separation at corners of buildings. Fig ure 25: Example of building pounding damage building overview; and and close-up of column.

15 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith 15 are reported to be reinforced concrete, while the top five storeys are reported to have load bearing URM piers around the exterior of the building and a steel frame internally (columns spaced roughly at 5 m) with timber floors throughout (New Zealand Historic Places Trust, 2010). The masonry piers, having dimensions of approximately mm, were badly cracked at levels 3 and 4 (figure 27). This damage was most likely due to the transition from concrete to masonry at level 3 and the fact that the adjoining 2-storey building located along the southern wall side stopped providing lateral support at that level. It appears that the lift core had received some strengthening previously, as well as the roof, perhaps in the late 1980s as reported by the New Zealand Historic Places Trust (2010). Close up photographs of the masonry piers at levels 3 and 4 show the primary damage that concerned the assessment teams (figure 27). Further inspection by the assessment team exposed the internal face of one pier on the western face of the building to reveal that the external cracking continued through the entire pier thickness. Two days after the main earthquake, structural engineers met with Urban Search and Rescue Team leaders and city officials to determine a strategy for making the structure safe enough for building contractors and engineers to enter to determine more fully the extent of damage and the viability of repair. Four days after the main earthquake, the building had survived one M5.4 and three M5.1 aftershocks. At the time of writing, no decision had been made as to whether the building will be demolished or repaired. Currently the building poses a significant falling hazard for people, traffic and buildings located within 70 m in all four directions. This falling hazard has resulted in all the buildings within this distance being closed to the public (red-tagged) as well as both main streets (Manchester and Hereford) that run adjacent to the building. St Elmo Chambers is also a 7-storey building, reported to be a reinforced concrete frame building with external clay brick masonry piers. Owing to the absence of control joints between the masonry and concrete frame, it appears that the masonry piers attracted sufficient seismic in-plane forces to cause shear failure (refer figure 28). However, once the masonry cracked the seismic loads were transferred to the concrete frame. Judging by the extent of cracking in the brickwork, it appears that the storey drifts were less than 1%, implying that the concrete frame was not pushed to its maximum capacity (strength or drift). The authors were not able to inspect the building from inside. From external inspection, the building was judged to pose a falling hazard to pedestrians near the building, but was in little danger of the building itself collapsing in an aftershock. Fig ure 26: Manchester Courts Building (view from northwest). 18 BUILDING DAMAGE DUE TO GROUND DEFORMATION Perhaps the most striking aspect of this earthquake overall was the extensive amount of liquefaction Fig ure 27: Damage to masonry piers at levels 3-5 north wall piers, levels 3-4; and west wall piers, levels 4-5.

16 16 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith Fig ure 28: Views of St Elmo Chambers building in Montreal Street overview of building; and close-up of damage to brickwork. Fig ure 29: Damage to buildings having masonry veneer over timber frame, due to ground deformation and liquefaction damage to masonry veneer due to ground deformation; wide cracks due to ground deformation; masonry veneer damage to recently built residence; and damage due to liquefaction. and ground deformation that occurred. These phenomena were not seen to a significant extent in the Christchurch CBD region containing the highest density of URM buildings, but did impact on a number of timber-framed structures with masonry veneer. As shown in figure 29, several cases of extreme ground deformation were observed, and there were numerous cases where large crack widths formed in residential timber framed structures having a masonry veneer (figure 29). There were also cases where ground liquefaction had resulted in masonry structures having sunk into the ground (figure 29). 19 CLOSING REMARKS On the few occasions that building owners or occupants were in attendance it was possible to gain access to the interior of URM buildings and often observe that some separation had occurred between the floor and/or roof diaphragms and the masonry walls (in the out-of-plane direction). This damage was not easy to detect from the outside of a building, so that the damage reported from building surveys in the first 72 hours can safely be assumed to be a lower bound estimate of structural damage to URM buildings.

17 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith 17 There were many instances of buildings that were structurally sound themselves, but had suffered damage or were yellow or red-tagged owing to falling hazards from neighbouring buildings. In some instances it was clear that a parapet or chimney from a neighbouring building had fallen onto or through the roof, being the only damage to the structure. In other instances, a building abutting a taller building with damaged parapet or gable side walls or chimney was given a yellow card (no public access) due only to the falling hazard posed by the structure next door. These examples of collateral damage and risk, such as that posed by 160 Manchester Street, and the associated business interruption costs, will surely make the financial impact of this earthquake much greater than just the cost of rebuilding. ACKNOWLEDGEMENTS Rodney Henderson-Fitzgerald, a data analyst for Christchurch City Council, is thanked for his assistance with securing the building inspection data reported herein. Lisa Moon and Najif Ismail are thanked for their analysis and graphing of the building damage statistics. Dmytro Dizhur, Yi-Wei Lin, Alistair Russell, Charlotte Knox and Claudio Oyarzo-Vera are thanked for supplying some of the photographs that have been used herein. REFERENCES Christchurch City Council, 2010, Review of earthquake-prone, dangerous and insanitary buildings policy, www1.ccc.govt.nz/council/agendas/2010/ March/RegulatoryPlanning4th/4.Earthquake.pdf, 4 March, retrieved 20 September Davey, R. A. & Blaikie, E. L. 2010, Predicted and observed performance of masonry parapets in the 2007 Gisborne earthquake, 2010 Annual Conference of the New Zealand Society for Earthquake Engineering, Wellington, March, db.nzsee.org.nz/2010/paper07.pdf. Dowrick, D. 1998, Damage and Intensities in the Magnitude Hawke s Bay, New Zealand, Earthquake, Bulletin of the New Zealand Society for Earthquake Engineering, Vol. 30, No. 2, pp Edwards, M., Griffith, M., Wehner, M., Lam, N., Corby, N., Jakab, M. & Habili, N. 2010, The Kalgoorlie Earthquake of the 20 th April 2010: Preliminary Damage Survey Outcomes, Proceedings of the Australian Earthquake Engineering Society Conference, Perth. EERI, 2010, The Mw 8.8 Chile Earthquake of February 27, 2010, Earthquake Engineering Research Institute Newsletter, June, images/eeri_newsletter/2010_pdf/chile10_insert. pdf, retrieved 16 September GNS, 2010, Darfield earthquake damages Canterbury, christchurch-earthquake.html, 16 September, retrieved 16 September New Zealand Historic Places Trust, 2010, Manchester Courts, RegisterSearch/RegisterResults.aspx?RID=5307; retrieved 9 September New Zealand Parliament, 2004, Building Act 2004, Department of Building and Housing Te Tari Kaupapa Whare, Ministry of Economic Development, New Zealand Government, Wellington, New Zealand, date of assent 24 August NewstalkZB, 2010, Most damage claims over $10k, asp?storyid=182291, 13 September, retrieved 13 September NZ Herald, 2010, Christchurch earthquake victim: I was pinned to the ground, nz/news/article.cfm?c_id=1&objectid= ; 14 September, retrieved 20 September NZSEE, 2006, Assessment and Improvement of the Structural Performance of Buildings in Earthquakes, Recommendations of a NZSEE Study Group on Earthquake Risk Buildings, New Zealand Society for Earthquake Engineering. Page, A. 1991, The Newcastle earthquake behaviour of masonry structures, Masonry International, Journal of the British Masonry Society, Vol. 5, No. 1, pp Restrepo-Velez, L. F. & Magenes, G. 2009, Static tests on dry stone masonry and evaluation of static collapse multipliers, Research Report No. ROSE-2009/02, IUSS Press, Fondazione EUCENTRE, Pavia, 72 pp. Russell, A. P. & Ingham, J. M. 2008, Architectural Trends in the Characterisation of Unreinforced Masonry in New Zealand, 14 th International Brick and Block Masonry Conference (14IBMAC), Sydney, Australia, February. Russell, A. P. & Ingham, J. M. 2010, Prevalence of New Zealand s Unreinforced Masonry Buildings, Bulletin of the New Zealand Society for Earthquake Engineering, Vol. 43, No. 3, pp Standards New Zealand, 2004, NZS :2004, Structural Design Actions Part 5:Earthquake actions New Zealand, Wellington, New Zealand. Wells, S. 2010, Council fast tracks earthquake policy, christchurch-earthquake/ /council-fasttracks-earthquake-policy, 15 September, retrieved 15 September 2010.

18 18 Performance of unreinforced masonry buildings during the 2010 Darfield... Ingham & Griffith JASON INGHAM Jason Ingham is Associate Professor in the Department of Civil and Environmental Engineering at the University of Auckland. Jason obtained his PhD from the University of California at San Diego in 1995, having previously completed his BE (Hons) and ME (Dist) from the University of Auckland. Jason also has an MBA from the University of Auckland. Jason is currently on the management committees of the New Zealand Society of Earthquake Engineering and the Structural Engineering Society of New Zealand, and is currently vice-president of the New Zealand Concrete Society. Jason s primary research interests are associated with seismic assessment and retrofit of concrete and masonry structures, and sustainable concrete technology. Over the last 6 years Jason has led a project considering the seismic assessment and retrofit of unreinforced masonry buildings, with further details available at www. retrofitsolutions.org.nz. MICHAEL GRIFFITH Michael Griffith is Professor in the School of Civil, Environmental and Mining Engineering at the University of Adelaide. He obtained his PhD in Structural Engineering from the University of California at Berkeley (1988) after completing his BSc (Eng) and MSc (Eng) degrees in Civil Engineering at Washington State University in the US. Michael is a member of the Standards Australia Australian Earthquake Loading Code committee, and he is also a member of Engineers Australia, state committee member of the SA Structural College and Past-President of the Australian Earthquake Engineering Society. His main professional and research interests are in the field of earthquake engineering and structural dynamics. He has co-authored two book chapters and over 100 research papers in the field of structural engineering.