The Demise of the URM Building Stock in Christchurch During the 2010- 2011 Canterbury Earthquake Sequence

This may be the author’s version of a work that was submitted/accepted for publication in the following source:

Moon, Lisa, Dizhur, Dmytro, Senaldi, Ilaria, Derakhshan, Hossein, Grifﬁth, Michael, Magenes, Guido, & Ingham, Jason (2014) The demise of the URM building stock in Christchurch during the 2010- 2011 Canterbury earthquake sequence. Earthquake Spectra, 30(1), pp. 253-276.

This ﬁle was downloaded from: https://eprints.qut.edu.au/119501/

c Consult author(s) regarding copyright matters

This work is covered by copyright. Unless the document is being made available under a Creative Commons Licence, you must assume that re-use is limited to personal use and that permission from the copyright owner must be obtained for all other uses. If the document is available under a Creative Commons License (or other speciﬁed license) then refer to the Licence for details of permitted re-use. It is a condition of access that users recog- nise and abide by the legal requirements associated with these rights. If you believe that this work infringes copyright please provide details by email to [email protected]

Notice: Please note that this document may not be the Version of Record (i.e. published version) of the work. Author manuscript versions (as Sub- mitted for peer review or as Accepted for publication after peer review) can be identiﬁed by an absence of publisher branding and/or typeset appear- ance. If there is any doubt, please refer to the published source. https://doi.org/10.1193/022113EQS044M 1 The Demise of the URM Building Stock in 2 Christchurch during the 2010 –2011 3 Canterbury Earthquake Sequence

a) b) c) d) 4 Lisa Moon, Dmytro Dizhur, Ilaria Senaldi, Hossein Derakhshan, e) f ) g) 5 Michael Griffith, M.EERI, Guido Magenes, and Jason Ingham, M.EERI

6 The progressive damage and subsequent demolition of unreinforced masonry 7 (URM) buildings arising from the Canterbury earthquake sequence is reported. 8 A dataset was compiled of all URM buildings located within the Christchurch 9 CBD, including information on location, building characteristics, and damage 10 levels after each major earthquake in this sequence. A general description of 11 the overall damage and the hazard to both building occupants and to nearby 12 pedestrians due to debris falling from URM buildings is presented with several 13 case study buildings used to describe the accumulation of damage over the 14 earthquake sequence. The benefit of seismic improvement techniques that had 15 been installed to URM buildings is shown by the reduced damage ratios 16 reported for increased levels of retrofit. Demolition statistics for URM buildings 17 in the Christchurch CBD are also reported and discussed. [DOI: 10.1193/ 022113EQS044M]

18 INTRODUCTION

19 There have been over 11,000 earthquakes and aftershocks associated with what is 20 referred to here as the 2010 –2011 Canterbury earthquake sequence (Bradley et al. 2013). 21 The Christchurch Central Business District (CBD) was defined by the Canterbury Earth- 22 quake Royal Commission (2011) as the area bounded by the four avenues (Bealey, 23 Fitzgerald, Moorhouse and Deans) and Harper Avenue (Figure 1a), with Bradley et al. 24 (2013) reporting that the intensity of shaking recorded at the CCCC site (refer Figure 1a) 25 was representative of that experienced across most of the CBD. Ground motion in five

a) Doctoral researcher, School of Civil, Environmental and Mining Engineering, University of Adelaide, South Australia 5005, Australia b) Research Fellow, Department of Civil and Environmental Engineering, The University of Auckland, Private Bag 92019, Auckland 1142, New Zealand c) Doctoral researcher, ROSE Program, UME School, IUSS, Via Ferrata 1, Pavia, Italy d) Research Fellow, School of Civil, Environmental and Mining Engineering, University of Adelaide, South Australia 5005, Australia e) Professor, School of Civil, Environmental and Mining Engineering, University of Adelaide, South Australia 5005, Australia f) Associate Professor, Department of Civil Engineering and Architecture, University of Pavia and European Centre for Training and Research in Earthquake Engineering (EUCENTRE), Via Ferrata 3, Pavia, Italy g) Professor, Department of Civil and Environmental Engineering, The University of Auckland, Private Bag 92019, Auckland 1142, New Zealand

0.3 4 Sept 26 Dec 22 Feb 23 Dec 23 Dec 0.2 2010 2010 2011 2011 2011

0.1

-0.1

-0.2

-0.3 Ground Acceleration (g) -0.4 |<-- 127 seconds -->| 25 s |<-- 47 s -->|<-- 63s -->|<-- 64 -->| -0.5 0 50 100 150 200 250 300 Time (seconds) (a) Location of CCCC recording station (b) Horizontal component N64E

0.4 0.8 4 Sept 26 Dec 22 Feb 23 Dec 23 Dec 0.7 4 Sept 26 Dec 22 Feb 23 Dec 23 Dec 0.3 2010 2010 2011 2011 2011 2010 2010 2011 2011 2011 0.6 0.2 0.5 0.4 0.1 0.3 0.2 0 0.1 0 -0.1 -0.1 -0.2 -0.2 -0.3 |<-- 127 seconds -->| 25 s |<-- 47 s -->|<-- 63s -->|<-- 64 -->| Ground Acceleration (g)

Ground Acceleration (g) |<-- 127 seconds -->| 25 s |<-- 47 s -->|<-- 63s -->|<-- 64 -->| -0.3 -0.4 -0.5 -0.4 -0.6 0 50 100 150 200 250 300 0 50 100 150 200 250 300 Time (seconds) Time (seconds) (c) Horizontal component N26W (d) Vertical component

Figure 1. Details of combined earthquake recordings in the Christchurch CBD.

26 of these events, with a combined duration of approximately 325 seconds, caused horizontal 27 ground accelerations in excess of 0.15g (Figure 1b –d) at the CCCC site. 28 For completeness it is noted that while the Darfield earthquake (4/9/2010) had a greater 29 magnitude (Mw 7.1), its epicenter was located much further away (approximately 40 km) 30 from the Christchurch CBD than was the Mw 6.3 Christchurch earthquake (22/2/2011) 31 whose epicenter was only 10 km from the Christchurch CBD. Thus, from a ground accel- 32 eration perspective, the Darfield earthquake elastic spectra for sites in the CBD were roughly 33 consistent with the elastic design spectra for a 500-year return period event for Christchurch. 34 In contrast, the spectra for the same sites in the February 2011 event corresponded roughly to 35 the 1/2500 year design earthquake (Bradley and Cubrinovski 2011).

36 OBSERVED PERFORMANCE OF URM BUILDINGS 37 The results from two companion studies are presented below. The first study considered 38 all URM buildings in the Christchurch CBD and was undertaken at the request of the 39 Canterbury Earthquakes Royal Commission (Ingham and Griffith 2011c). The second 40 study specifically focused on stone URM buildings, due to both the disproportionately THE DEMISE OF THE URM BUILDING STOCK IN CHRISTCHURCH 3

41 large number of stone URM buildings in the Christchurch region when compared to all of 42 New Zealand and because of the particular heritage significance of many of these buildings 43 (Senaldi et al. 2013). A third companion study considering the seismic performance of 112 44 Christchurch churches (including 20 clay brick and 32 stone URM churches) has been 45 reported elsewhere (Leite et al. 2013).

46 MATERIAL PROPERTIES 47 The general observation from the debris of collapsed URM walls was that the kiln fired 48 clay bricks were of sound condition, but that the mortar was in poor condition. In most cases 49 the fallen debris had collapsed into piles of individual bricks, rather than as larger chunks of 50 masonry debris, and when rubbed the mortar readily crumbled when subjected to finger pres- 51 sure (refer Figure 2), suggesting that the mortar compression and shear strengths were very 52 low. Subsequent testing (Lumantarna 2012) of 293 mortar samples and 67 clay bricks col- 53 lected from 51 and 23 damaged URM building sites respectively located across the greater 54 Christchurch area resulted in average compressive strengths of 1.75 MPa and 23.6 MPa for 55 the mortar and bricks, confirming the initial observation that mortar was frequently in 56 poor condition (i.e., soft) but that the bricks were reasonably strong. These findings are con- 57 sistent with results obtained from testing of other URM buildings throughout New Zealand 58 (Lumantarna et al. 2013a, b). 59 When undertaking a post-earthquake review of archived files reporting the design of 60 earthquake strengthening interventions for URM buildings in the Christchurch CBD 61 there was little evidence found on comprehensive site investigations conducted by practicing 62 engineers in order to determine appropriate masonry material proprieties (i.e., mortar bed 63 joint shear tests and anchor pull-out tests) to be used in their designs. Instead, it appeared 64 that most conclusions regarding building condition and material properties were based on 65 visual observations. In contrast, extensive material investigation on existing buildings is

Figure 2. Condition of masonry rubble. 4 MOON ET AL.

66 routinely conducted in the U.S. in order to establish both the existing condition of the build- 67 ing and accurate and reliable material properties for use in structural designs.

68 TYPICAL FAILURE MODES AND DAMAGE PATTERNS 69 Comprehensive descriptions of the damage caused to URM buildings by the Canterbury 70 earthquake sequence have been previously reported (Dizhur et al. 2010, Dizhur et al. 2011,

Table 1. List of the most commonly observed URM building failure modes

Component Failure mode: Damage Chimney Frequent flexural failures, falling brickwork, damage to roof and/or adjacent structures (most commercial buildings had no chimneys, or chimneys that were not visible from the street, so numbers were not recorded) Gable end wall Out-of-plane failure due to missing or excessively spaced ties (63% of buildings with gables (120∕189) had partial or full collapses of gables) Parapets Frequent collapse of both unrestrained and restrained parapets. Unrestrained parapet failure usually occurred along the roofline if wall –roof connections were intact, otherwise in conjunction with out-of-plane wall failure. Restrained parapet failure was often due to missing or ineffective restraint either in adhesive anchor connection to masonry or punching shear failure in the masonry around the connection (66% of buildings with parapets (324 ∕491) had partial or full collapses of parapets) Awnings Falling parapets landed on awnings, causing failure of the awning canopy or awning support when tied back into the URM (42% of buildings with an awning (125∕299) had partial/full awning collapse) Out-of-plane walls Many examples of out-of-plane wall failures. Almost universally the masonry fell in the outward direction (70% of buildings were observed to have out-of- plane wall damage (438∕627)) Anchorages Anchorage failures at wall-to-floor and wall-to-roof connections were common, as well as anchorage failures for parapets and awnings (this information is not currently in the database) Corners Vertical connections between walls at building corners were observed to fail in some cases (66% of buildings were observed to have some corner failure (414∕627)) In-plane walls Relatively infrequent in-plane wall failures, typically at spandrel and pier locations. Failures were observed mostly in walls containing a large number of openings (56% of buildings showed signs of in-plane wall damage 352∕627–of these only 32 ∕352 (9%) showed extreme damage) Pounding Several cases were observed of damage caused by pounding of one URM building against its neighbor (see Cole et al. 2012) (this information is not currently in the database) THE DEMISE OF THE URM BUILDING STOCK IN CHRISTCHURCH 5

Figure 3. Wall-roof anchorage partial failure.

71 Ingham et al. 2011, Ingham and Griffith 2011a, 2011b, 2011c, 2011d). It should be noted that 72 the most commonly observed failure modes (listed in Table 1) sustained by URM buildings 73 were typical of those seen in other earthquake-affected cities around the world. Even before 74 the February 2011 earthquake, a significant number of wall –diaphragm connections had been 75 seriously damaged, with obvious out-of-plane bed joint sliding, although in some cases com- 76 plete wall failure had not yet developed (see Figure 3). 77 On the few occasions that building owners or occupants were in attendance it was 78 possible to gain access to the interior of URM buildings and often observe that some separation 79 had occurred between the floor and/or roof diaphragms and the masonry walls (in the out-of- 80 plane direction). This damage was not easy to detect from the outside of a building, so that the 81 damage reported from external building surveys in the first 72 hours is assumed to be a lower 82 bound estimate of structural damage to URM buildings (Ingham et al. 2011, Moon et al. 2012). 83 There were also many cases of buildings that were structurally sound but were yellow 84 or red-tagged owing to falling hazards from neighboring buildings. In some cases it was 85 clear that a parapet or chimney from a neighboring building had fallen onto or through 86 the roof, being the only damage to the structure. In other situations, a building abutting 87 a taller building with a damaged parapet or gable side walls or chimney was given a yellow 88 card (no public access) due only to the falling hazard posed by the adjacent structure. These 89 examples of collateral damage and risk and the associated business interruption costs ele- 90 vated the financial impact of this earthquake sequence beyond just the cost of rebuilding 91 (Cousins et al. 2012).

92 DAMAGE STATISTICS FOR CLAY BRICK URM BUILDINGS IN 93 CHRISTCHURCH CBD 94 The data reported in this section are for clay brick load bearing masonry buildings 95 located in the Christchurch CBD. These buildings were mostly (77%) of solid two or 6 MOON ET AL.

96 three wythe thick wall construction with timber floors and roofs, although some (23%) 97 brick cavity construction was also observed. It is clearly shown in Figure 4a that the 98 sizeable percentage (43%) of green-tagged URM buildings in September 2010 signifi- 99 cantly reduced (to just 1%) after the February 2011 earthquake. In the data that follows, 100 it was not possible to distinguish between two and three wythe walls because most mul- 101 tistory URM buildings were comprised of three wythe walls in the lower levels and of 102 two wythe walls in the upper levels. Nevertheless, in every case the out-of-plane damage 103 started in the upper levels where the accelerations were highest and the walls were their 104 thinnest (i.e., two wythe). 105 In addition to collating the placard data as reported in Figure 4, data were collected on 106 many attributes of the URM building stock in the Christchurch CBD, such as number of 107 stories, building footprint area, and building occupancy type (see Ingham and Griffith 108 2011d for full details). The observed extent of damage was correlated against both the 109 Wailes and Horner (1933) Damage Scale, which was specifically developed to describe 110 damage to URM buildings, and against the damage levels prescribed in ATC 13 (ATC 111 1985), because of the widespread use of this survey form in post-earthquake damage assess- 112 ments. Interestingly, no discernible trends regarding the extent of building damage were 113 apparent for any parameter, such as indicated in Figure 5 where it can be seen that an 114 approximately uniform extent of damage was observed independent of the number of 115 stories. 116 Based on the position of fallen debris, the risk to building occupants and to pedestrians 117 located outside but in close proximity to the 340 URM buildings in the CBD was assessed as 118 shown in Figure 6. The data are consistent with the documented location of fatalities (39) 119 associated with URM buildings (both clay brick and stone) within the CBD, where the

Unknown 15% Unknown Green 3% 1% Demolished 6% Yellow Red 17% 11% Green 43%

Yellow 31%

Red 73% Green Yellow Red Unknown Green Yellow Red Demolished Unknown

(a) 4 September 2010 earthquake (b) 22 February 2011 earthquake (361 (361 entries) entries including 21 (6%) demolished)

Figure 4. Placard assignments for URM buildings in Christchurch CBD (Note: “Demolished” refers to buildings dismantled by building contractors before the February earthquake). THE DEMISE OF THE URM BUILDING STOCK IN CHRISTCHURCH 7

35%

30% 100

25% 80 20% 60 15% 40 10%

5% Surveyed Buildings of Number Percentage of Buildings Surveyed Buildings of Percentage

0% 0 Insignificant Moderate Heavy Major Destroyed 1 - 10% 10 - 30% 30 - 60% 60 - 100% 100% Extent of Damage to Building 1 Storey High 2 Stories High 3+ Stories High

Figure 5. Correlation between extent of damage and story height for 340 URM buildings in CBD (22 February 2011).

100% 100%

80% 80%

60% 60%

40% 40%

20% 20% Percentage of Buildings of Percentage Percentage of Buildings of Percentage

0% 0% Insignificant Moderate Heav y Major Destroyed Insignificant Moderate Heav y Major Destroyed 1 - 10% 10 - 30% 30 - 60% 60 - 100% 100% 1 - 10% 10 - 30% 30 - 60% 60 - 100% 100% Extent of Damage to Building Extent of Damage to Building

unlikely likely near certain unlikely likely near certain

(a) Risk to building occupant (b) Risk to nearby pedestrians

Figure 6. Risk to building occupants and nearby pedestrians for different building damage levels.

120 majority of people killed by URM building damage (35) were located outside rather than 121 within the building (see Table 2).

122 DAMAGE STATISTICS FOR STONE URM BUILDINGS 123 The damage assessment inspections that were undertaken in September 2010 and 124 again in April and May 2011 identified 96 unreinforced stone masonry buildings in 125 the wider Canterbury region (approx. half being located in the Christchurch CBD and 126 the remainder located in surrounding suburbs), many of which are included on the 8 MOON ET AL.

Table 2. Cause of death for 39 people killed by URM buildings (extracted from Cooper et al. 2012)

Number Cause of death associated with URM building 35 (90%) Killed as a result of the facade or walls of URM buildings collapsing onto: 26 (67%) Pedestrians or persons in vehicles 6 (15%) People in a neighboring building 3 (8%) People who had run out of a building to escape 4 (10%)Killedinside a URM building

127 New Zealand Historic Places Trust register of heritage buildings. Many of these stone 128 masonry buildings were designed by the same architects or architectural firms and 129 most were constructed between 1850 and 1930 and hence have similar characteristics 130 both in terms of architectural features and in the details of their construction. These build- 131 ings are typically 2 or 3 stories in height, with two-story buildings being most common, 132 and either standalone or row buildings. The vast majority of stone masonry buildings, and 133 in particular those constructed in the Gothic Revival style, are characterized by structural 134 peripheral masonry walls that may be connected, depending on the size of the building, to 135 an internal frame structure constituted of cast iron or steel columns and timber beams or to 136 internal masonry walls that support flexible timber floor diaphragms and timber roof 137 trusses. A variety of wall cross section types were identified, as indicated in Table 3 138 and illustrated in Figure 7. 139 The seismic performance of stone masonry buildings was partially identified by consid- 140 ering the safety assessment data that were collected following the earthquakes that occurred 141 in September 2010 and February 2011. From Figure 8 it can be seen that there was a sig- 142 nificant escalation of damage due to the continuing earthquake activity in the Christchurch 143 region. Figure 9 gives a further breakdown of this data following the February 2011

Table 3. Stone wall cross section types and number of buildings for each type

Types of stone wall cross section No. of buildings Multi-leaf masonry walls (basalt or lava flow) 24 Three-leaf masonry walls with rubble lava flow façade, internal concrete core 13 and Oamaru stone facing Three-leaf masonry walls in Oamaru stone and concrete core 6 Ashlar stone facing (basalt or bluestone) 24 Ashlar stone facing and brickwork backing 17 Ashlar stone facing, concrete core and brickwork backing 3 Others 6 Undetermined 3 Total 96 THE DEMISE OF THE URM BUILDING STOCK IN CHRISTCHURCH 9

(a) Time Ball Station with (b) Cathedral of the Blessed (c) St. Luke’s Anglican multiple-wythe stone Sacrament in Oamaru stone Church characterized by masonry with poured concrete core stone facings with clay brick interior wythes

Figure 7. Representative examples of wall cross-sections for Christchurch stone masonry buildings.

green 18% unknown demolished 2% green 2% 16%

unknown 49% yellow yellow 26% 22% red demolished red 58% 0% 7% green yellow red demolished unknown green yellow red demolished unknown

(a) After September 2010 (b) Data updated 07 June 2011

Figure 8. Distribution of safety evaluation placards applied to 96 stone masonry buildings (Note: “Demolished” means buildings dismantled by building contractors).

144 earthquakes on the basis of building usage. From this data no clear trends in damage patterns 145 could be observed on the basis of building usage. 146 A comparison between the data reported in Figure 4b and in Figure 8b suggests that stone 147 URM buildings fared better than clay brick URM buildings. This finding is contrary to 148 expectation when accounting for the inherent vulnerability of the two forms of construction 149 and is attributed to a number of extensive earthquake strengthening interventions that had 150 been applied to iconic stone masonry buildings, such as the Christchurch Arts Centre (Bailey 151 et al. 2013). Further details on the performance of unreinforced stone masonry buildings are 152 reported by Senaldi et al. (2013). 10 MOON ET AL.

Number of buildings 5

l t l o e ty a o ic li ti ial ff ci n rc ch umen O a e S n F ide Church li c c b ti es Mo u is R omm P C Cultural/museum Tour Green Yellow Red

Figure 9. Distribution of safety evaluation placards applied to stone masonry buildings, differ- entiated by usage (data updated 7 June 2011).

153 OBSERVED PERFORMANCE OF EARTHQUAKE 154 STRENGTHENED URM BUILDINGS 155 Following the February 2011 earthquake an exercise was undertaken to identify the seis- 156 mic improvements to URM buildings in Christchurch existing prior to the September 2010 157 earthquake. These improvements were divided into three categories for ease of assessment 158 and the interpretation of data: Parapet restraints; Type A seismic improvements; and Type B 159 seismic improvements. Definition of these categories is provided below.

160 PARAPET RESTRAINTS 161 Parapet restraints (considered structural) were looked at separately (i.e., they were not 162 part of A or A þ B grouping which are discussed later). Recognizing the obvious hazard 163 that parapets pose to building occupants and nearby pedestrians, it follows that parapet 164 restraints are one of the most commonly encountered earthquake strengthening techniques 165 in URM buildings. Typically observed methods used to improve the earthquake response of 166 parapets include but are not limited to: (a) parapet removal; (b) bracing of the parapet back 167 into the roof structure via steel members connected to the parapet and secured into the 168 masonry using either adhesive anchors or through ties; and (c) addition of a concrete 169 ring beam (or cap beam) in order to tie a parapet to the end walls that run orthogonal to 170 the parapet. 171 A total of 435 records of parapets (predominantly solid two-wythe thick) were associated 172 with the surveyed buildings, with some buildings having multiple parapets, such as those on 173 street corners or for end-of-row or standalone buildings. As shown in Table 4, of these 435 174 parapets only 149 (34%) parapets could be positively identified as having parapet restraints 175 installed. Unfortunately, it was not possible to definitively identify a sufficient sample size of 176 specific types of parapet restraint systems from building inspections although steel bracing 177 was by far the most commonly observed form of restraint. THE DEMISE OF THE URM BUILDING STOCK IN CHRISTCHURCH 11

Table 4. Distribution of parapet restraints ( “cases” refers to a single line of wall)

No. of cases % of parapets Restrained 149 34% Unrestrained 89 21% Unknown 197 45% Total 435 100%

178 As expected (Figure 10), restrained parapets performed significantly better than parapets 179 having no restraint, with 84% ( 75 ∕89 ) of unrestrained parapets suffering full or partial col- 180 lapse while only 44% (65 ∕149) of restrained parapets suffered similar damage. Furthermore, 181 48% (71 ∕149) of restrained parapets suffered no or moderate damage while only 13% 182 (12 ∕89 ) of unrestrained parapets achieved such good response, such that 86% ( 71 ∕83 ) of 183 those parapets that performed satisfactorily were restrained. 184 Overall it may be concluded that unrestrained parapets were twice as likely to collapse 185 as were restrained parapets. It would seem that this is a disappointing finding as it would 186 have been assumed that the majority of restrained parapets would have performed satis- 187 factorily. However, it is emphasized that many of the restrained parapets had suffered 188 some damage in previous earthquakes and had not been repaired before the February 189 2011 earthquake. Nevertheless, only 13% of the unrestrained parapets performed satisfac- 190 torily, supporting the Royal Commission recommendation that all parapets and ornaments 191 require restraint. This finding suggests that further investigation is also required to better 192 understand why parapet restraints were not more uniformly successful. Ground motions 193 well in excess (six times higher in February 2011) of those assumed for the design of

60%

Unrestrained parapets 50% Restrained parapets 40%

30%

20%

10% Percentage of Buildings Surveyed Buildings of Percentage

0% Insignificant Moderate Heavy Major Destroyed 1 - 10% 10 - 30% 30 - 60% 60 - 100% 100% Extent of Damage to Building

Figure 10. Performance of unrestrained and restrained parapets. 12 MOON ET AL.

194 parapet restraints, low mortar strengths, the absence of superimposed load on parapets, 195 correspondingly low masonry shear and anchorage strengths, ineffective adhesive anchor 196 development and spacing, and damage from prior earthquakes, all clearly influenced the 197 performance of restrained parapets.

198 TYPE A SEISMIC IMPROVEMENT 199 Examples of Type A retrofits included restraints to gable-ended walls, the installation of 200 wall–diaphragm anchorages, and diaphragm improvements. As with parapets, because it is 201 widely known that gable end walls are prone to failure, it is not uncommon to find some form 202 of restraint provided to inhibit out-of-plane failure (see Figure 11). Where present, the type of 203 gable end wall restraint was identified and recorded. The Canterbury Earthquake Royal Com- 204 mission Report (2011) reports that for 185 gable end walls, 128 had some form of restraint 205 which were further broken down using three categories: (i) most (109∕128) used through ties 206 (Figure 11a); (ii) 5∕128 used adhesive anchors; and (iii) 14 ∕128 were restrained with some 207 form of tie at the time of original construction (Figure 11b). 208 In some cases connections between the walls and the roof and floor diaphragm systems of 209 the building had been installed to prevent walls responding as vertical cantilevers secured 210 only at their base. Connections were typically either older “rose” type through ties with exter- 211 nal plates (see Figure 12a), which were installed from 1940s onward, or more recent adhesive 212 anchors (see Figure 12b) which in many locations were seen to be ineffective. The out-of- 213 plane failure of URM walls was in many cases attributed to the low shear strength of 214 masonry, wide anchorage spacing and/or insufficient embedment depth of anchors. In 215 some cases the reasons for the adhesive anchor failures were apparent: anchor pull-out 216 due to insufficient embedment length or a lack of bonding between the anchor and the sub- 217 strate material. In other cases, the reasons for such failures were not evident from visual 218 observation. The construction quality of adhesive type anchorages was commonly observed 219 to be poor, due to insufficient anchorage depths and poor workmanship.

Figure 11. Examples of gable end wall restraints. THE DEMISE OF THE URM BUILDING STOCK IN CHRISTCHURCH 13

Figure 12. Wall-to-diaphragm connections.

220 Diaphragm improvement techniques typically consisted of an addition of: (i) steel bra- 221 cing (see Figure 13a); (ii) steel brace frames (see Figure 13b); (iii) plywood overlays (see 222 Figure 13c); (iv) concrete slab topping; or (v) combinations of these (see Figure 13d).

223 TYPE B SEISMIC IMPROVEMENT 224 Examples of Type B retrofits generally include: (i) out-of-plane strengthening of existing 225 masonry walls for response perpendicular to wall planes (i.e., addition of strong-backs);

Figure 13. Example of diaphragm improvement techniques. 14 MOON ET AL.

226 (ii) in-plane shear strengthening of masonry walls; and (iii) introduction of higher-level inter- 227 ventions to supplement or take over the seismic resisting role from the original unreinforced 228 masonry structure. Techniques to achieve these improvements include the use of: internal/ 229 external strong backs; steel and concrete (moment and braced) frames, additional cross walls, 230 and strengthening of existing masonry walls with shotcrete, surface bonded Fiber Reinforced 231 Polymer (FRP), or post-tensioning. In practice, Type B retrofits are seen only in conjunction 232 with Type A retrofits. Two case study buildings with Type A þ B retrofits are discussed later.

233 RETROFIT LEVEL (%NBS) 234 The level of seismic improvement was expressed as %NBS (Percentage of New Building 235 Standard), where a building having a score of %NBS ≤ 33 corresponds to the definition 236 given in Section 122 of the New Zealand Building Act (2004) (New Zealand Parliament 237 2004) as “earthquake-prone” and a building having a score of 33 ≤ %NBS ≤ 67 is defined 238 by NZSEE (2006) as “earthquake risk.” The value of %NBS was established either from 239 Christchurch City Council (CCC) records, personal communication with building owners, 240 engineers and heritage personnel, or in some cases where sufficient seismic details about the 241 building were known, was based upon estimation. 242 Accessed CCC records showed that most engineering reports provided the level of 243 strengthening in terms of design PGA values, which were correlated to the then-current 244 (2011) %NBS values using the seismic zone factor of Z ¼ 0.22 for Christchurch and the 245 procedures provided in NZS 1170.5 (Standards New Zealand 2004). Note that the seismic 246 zone factor for Christchurch was later elevated to Z ¼ 0.3.

247 COMPARISON BETWEEN %NBS AND ASSESSED DAMAGE LEVEL 248 Earthquake strengthening levels in terms of %NBS were identified for 94 URM buildings 249 using the pre-earthquake value for Z of 0.22. The distribution of %NBS data is reproduced in 250 Table 5 where it is shown that 61 buildings had been earthquake strengthened to at least 67% 251 NBS. With respect to Type A and B retrofit techniques, it can be generalized that Type A þ B 252 retrofit was needed to achieve a %NBS > 67% whereas Type A retrofit on its own was most 253 likely to correspond to a %NBS classification of between 33% and 67%. 254 The performance of these earthquake strengthened buildings was analyzed by determin- 255 ing the damage distribution for each category of %NBS. From Table 6 and Figure 14 it can be 256 seen that URM buildings having retrofits with less than 33%NBS did not perform appreci- 257 ably better than URM building having no retrofits at all. This finding is consistent with

Table 5. Distribution of %NBS classifications for 94 earthquake strengthened URM buildings

NBS Retrofit level No. of buildings % of buildings

%NBS < 33 15 16% 33 ≥ %NBS < 67 18 19% 67 ≥ %NBS < 100 50 53% %NBS ≥ 100 11 12% THE DEMISE OF THE URM BUILDING STOCK IN CHRISTCHURCH 15

Table 6. Damage levels for different %NBS categories (after February 2011)

33 ≥ %NBS 67 ≥ %NBS %NBS Damage level %NBS < 33 < 67 < 100 ≥ 100 Total Insignificant 1 7% 1 6% 10 20% 8 73% 20 1–10% Moderate 5 33% 4 22% 28 56% 3 27% 40 10 –30% Heavy 5 33% 9 50% 10 20% 0 0% 24 30 –60% Major 1 7% 4 22% 2 4% 0 0% 7 60 –100% Destroyed 3 20% 0 0% 0 0% 0 0% 3 100% Total 15 18 50 11 94

80%

60%

40%

20%

Percentage of buildings of Percentage 0%

Damage level Level of retrofit

%NBS ≥ 100 67 ≤ %NBS < 100 33 ≤ %NBS < 67 %NBS < 33 All buildings No retrofit

Figure 14. Damage levels for different levels of %NBS earthquake strengthening.

258 observations from past earthquakes (ATC-31 1992). In addition, incomplete retrofits can 259 only be expected to provide, at best, satisfactory seismic behavior of only the properly 260 strengthened parts of the building (Bruneau 1995). In contrast, there was a significant reduc- 261 tion in the “Major” and “Destroyed” levels of damage for URM buildings when strengthened 262 to 33% < %NBS < 67% and an even more dramatic reduction in “Heavy” damage when 263 URM buildings were retrofitted to a level of %NBS > 67%. Thus, it can be concluded 264 that the applications of seismic retrofits to URM buildings in the CBD were successful 265 in reducing building damage and improving life safety, especially when considering that 266 the intensity of the February 2011 earthquake was well in excess of the code level specified 267 for Christchurch. 16 MOON ET AL.

268 CASE STUDY BUILDINGS 269 Four case study buildings are considered here; two with Type A retrofit and two with 270 Type A þ B retrofit. The location of the four buildings is shown in Figure 15, where it can be 271 seen that two of the buildings, A1 and B2, lie within the Christchurch CBD. Buildings B1 and 272 A2 were included for illustrative purposes because they demonstrate excellent performance 273 of a seismic retrofit (B1) and the effect of cumulative damage (A2). A description of each 274 building and its performance is given below.

275 TYPE A RETROFIT

276 Example A1: Mayfair Building, 140–142 Lichfield Street 277 This four-story clay brick URM building had no evidence of parapet restraints but 278 did have through-bolts for wall-diaphragm (floor and roof) anchors which initially restricted 279 damage to just the top story. However, in subsequent earthquakes the out-of-plane wall 280 damage spread down and to the corners and eventually around into the side walls (see 281 Figure 16). This building has since been demolished.

282 Example A2: Sumner Borough Council Building, 16 Wakefield Street 283 This two-story clay brick URM building used through-bolts for wall-diaphragm (all 284 floors and roof) anchorage as well as having all parapets restrained (see Figure 17). The 285 September 2010 earthquake caused cracking in the side and front walls but no major cracking 286 was observed until after the February 2011 earthquake, and by June 2011 entire sections of 287 the side walls had fallen away. The building has since been demolished.

Figure 15. Locations of case study buildings and strong motion seismographs (active 22 February 2011). THE DEMISE OF THE URM BUILDING STOCK IN CHRISTCHURCH 17

Figure 16. Progressive damage to Mayfair building —Level A retrofit.

Figure 17. Progressive damage in Sumner Council building —Level A retrofit.

288 TYPE B RETROFIT 289 Example B1: The Smokehouse, 650 Ferry Road 290 The Smokehouse is a two-story freestanding clay brick URM building as shown in 291 Figure 18. The building’s exact construction date is unknown but can be confirmed as 292 pre-1930s, and the building has been categorized as a heritage building by the Christchurch 293 City Council. The building’s footprint has dimensions of 13 m along Ferry Road and 10 m

Figure 18. Progression of damage in Smokehouse —Level A þ B retrofit. 18 MOON ET AL.

294 along Catherine Street. The original mortar is a weak lime/cement mortar with large grain 295 size sand. In places the original mortar was re-pointed with strong cement mortar. 296 The Smokehouse was seismically retrofitted in 2007 with adhesive anchors, steel frame 297 stiffening of the roof, alterations to the internal layout which involved replacement of some 298 original external walls with moment resisting steel frames (see Figure 18b, 18d) and infilling 299 of one window at the second floor level. The retrofit design of the building won the 300 New Zealand Architectural Award in 2008 for initiative in retention, restoration and exten- 301 sion of a significant building and its adaption to new uses (Smokehouse Restaurant 2009). 302 There were no signs of damage from the September 2010 earthquake on the exterior face 303 of the building. However some minor vertical cracks could be seen in the interior at the wall 304 corners, with minor cracks around the perimeter of the in-filled window and one small hor- 305 izontal crack slightly above the base of one pier at the second floor level. The scaffolding 306 shown in Figure 18 was erected to repair damage prior to the February 2011 earthquake. The 307 February event caused only slight cracking as well as some minor damage to the exterior 308 brickwork, which required repointing with minimal interruption to business. Hence, the retro- 309 fit was seen by the owner to have been very successful and a good investment.

310 Example B2: Red Cross Shop, 223–227 High Street 311 This building was constructed in 1890 with timber truss roof and floors and was later 312 retrofitted (2006) with the addition of concrete floor diaphragms, internal steel strongbacks 313 and additional cross-walls. The building has survived all the earthquakes and has been 314 repaired with the expectation that the building will not be demolished. The progressive con- 315 dition of the building (see Figure 19) shows little damage to the façade, with the only major 316 damage caused by brickwork from the neighboring building falling through its roof and awn- 317 ing. However, minor visible damage occurred in a rear wall (not shown).

318 RESPONSE OF URM BUILDINGS TO CUMULATIVE GROUND SHAKING 319 Cabanas et al. (1997) presented the results of a comparative study aimed at identifying a 320 ground motion intensity index that best describes URM building damage based on post- 321 earthquake observations. Six scales of building damage were considered, ranging from

Figure 19. Progression of damage to Red Cross building. THE DEMISE OF THE URM BUILDING STOCK IN CHRISTCHURCH 19

322 0 (slight plaster cracking) to 5 (building collapse), and two indices were described, being the 323 Arias Intensity (Arias 1970) and the Cumulative Absolute Velocity with an acceleration 2 324 threshold of 20 cm ∕s , CAV 20 . A threshold of CAV 20 ¼ 1.5 g-s or of AI ¼ 1.5 m∕s was 325 suggested as appropriate for the onset of the “Destruction” damage level (Level 4, partial 326 building collapse, interior wall collapse), and it was suggested that the “Heavy” damage 327 level (Level 3, falling chimneys and deep and wide cracks in walls) could occur at intensities 328 as low as CAV 20 ¼ 0.6 g-s or AI ¼ 0.45 m∕s. 329 Ground accelerations at the Christchurch Cathedral College (CCCC) site (refer Figure 1a) 330 within the Christchurch CBD were investigated using the two earthquake intensity measures 331 described above, and the results were correlated with the observed building damage. Instead 332 of CAV20 , a similar but more recently proposed index, CAVDP , was calculated using the 333 procedure suggested by Campbell and Bozorgnia (2012a, b). The building damage incurred 334 in the 4 September 2010 earthquake was largely in the form of extensive wall cracking and 335 falling chimneys and parapets, which is best described as damage Level 3, whereas the exten- 336 sive URM building damage during the 22 February 2011 earthquake is consistent with 337 damage Levels 4 or 5 using the scale suggested by Cabanas et al. (1997). Figure 20a 338 shows that the cumulative AIs of both N64E and N26W components of the Canterbury 339 earthquake sequence first exceeded the 1.5 m∕s threshold for damage Level 4 during the 340 22 February 2011 earthquake.

341 Similar to the finding reported above, Figure 20b shows that the CAVDP value for the 342 4 September 2010 earthquake was between the proposed limits for damage Level 3 and 343 Level 4, but that the cumulative values exceeded the damage Level 4 threshold during 344 the 22 February 2011 earthquake. Consequently it is concluded that the correlation between 345 the CAVDP parameter and the observed building damage in the 4 September 2010 earthquake 346 is consistent with the recommendations of the Cabanas et al. (1997) study. The large cumu- 347 lative CAVDP values for the 22 February 2011 earthquake are also consistent with the wide- 348 spread observation of extensive damage to URM building during the 22 February 2011 349 earthquake.

6 4 04 26 22 23 23 04 26 22 23 23 Sep Dec Feb Dec Dec Sep Dec Feb Dec Dec 10 10 11 11 11 10 10 11 11 11 3 (g-s) 4 DP

2 N64E Lvl. 4 damage N26W N64E 2 Lvl. 4 damage N26W 1 Lvl. 3 damage Cumulative Cumulative CAV Lvl. 3 damage

Cumulative Arias Intensity, AI (m/s) 0 0

(a) Cumulative Arias intensity values (b) Cumulative CAV DP intensity values

Figure 20. Earthquake intensity of the main events of the Canterbury earthquake sequence based on recordings from CCCC site. 20 MOON ET AL.

350 In summary, both Arias Intensity and the CAVDP indices calculated for ground motions 351 in the Christchurch CBD greatly exceeded values previously proposed as resulting in 352 widespread and severe damage to URM buildings. Consequently, observations from the 353 Canterbury earthquake sequence support earlier recommendations that ground motion 354 metrics that account for both intensity and duration have the potential to effectively char- 355 acterize the damage to URM buildings associated with earthquake ground motions.

356 CHRISTCHURCH BUILDING DEMOLITION STATISTICS 357 The high percentages of red-tagged (73%) and yellow-tagged (17%) URM buildings after 358 the February 2011 earthquake has had a significant impact on the Christchurch CBD, leading 359 to the majority of these building having been subsequently demolished as indicated in 360 Figure 21. The large jump in the plots during August 2011 is due to a combination of (i) dis- 361 continuity in the data collection process; (ii) the fact that it took a long time for demolitions to 362 get started in earnest; and (iii) until the June 2011 earthquakes it was hoped that damaged 363 URM buildings could be saved. In hindsight, the difficult decision to demolish the seven- 364 story heritage listed Manchester Courts building, which was damaged in the September 2010 365 event probably saved lives in the February 2011 event.

100% 120 100%

120 100 80% 80% 100 80 60% 60% 80 60 60 40% 40% 40 40 20% 20 20% 20 Number of Buildings Demolished Buildings of Number Percentage of Buildings Demolished Buildings of Percentage Number of Buildings Demolished Buildings of Number 0% 0 Demolished Buildings of Percentage 0% 0 11 11 11 12 12 -12 -12 12 12 12 -11 -11 - Jul- earlier - Apr Mar Jul earlier May- -Jan-12 -Nov-11 -Aug- -Sep-12 Apr- Sep-11 Aug Mar 4-Oct- May- - 7-Feb- 5-Mar-12 8-Sep-11 3 27 - 9- 11- 4-Oct-11 11 3-Aug-11 8- 30-Aug-11 20-Sep-11 28 20- 7-Feb-12 5-Mar-12 27- 26 9 30-May-12 11 11-Jan-12 30- 20-Sep-11 28-Nov-11 20- 26-Sep-12 30-May-12 Date of Demolition Date of Demolition (a) No confirmed retrofit (b) Type A retrofit only

100% 70

80% 60

50 60% 40

40% 30

20 20% 10 Percentage of Buildings Demolished Buildings of Percentage Number of Buildings Demolished Buildings of Number 0% 0 11 12 12 12 12 -11 -11 - -12 - - Jul earlier Apr-12 Mar- May-12 - May - - 4-Oct 3-Aug-11 8-Sep-11 5-Mar 7-Feb 27- 9 11 11-Jan 20- 28-Nov-11 30-Aug-11 20-Sep- 26-Sep-12 30 Date of Demolition (c) Type A and B retrofit (d) Demolished URM buildings as at 27/06/12

Figure 21. Distribution of retrofit types for demolished buildings. THE DEMISE OF THE URM BUILDING STOCK IN CHRISTCHURCH 21

366 Despite the significantly reduced damage ratio for URM buildings strengthened using 367 Level A þ B retrofits to > 67%NBS, a relatively large number of these buildings were 368 demolished as indicated in Figure 21c. Indeed, Figure 21 shows that while more than 369 90% of the non-retrofitted URM buildings were demolished, roughly 80% of the Level 370 A retrofitted and 70% of the Level A þ B retrofitted URM buildings were also demolished. 371 The reasons for the relatively high demolition rate for strengthened URM buildings has not 372 yet been fully reported but it is thought that the cumulative effects of damage during the 373 sequence of earthquakes, the high level of insurance coverage, and uncertainties associated 374 with assessment of partly damaged buildings were all significant contributors. The location 375 of the demolished URM buildings is indicated on a map in Figure 21d, providing greater 376 detail of the former location of these buildings within the Christchurch CBD. The fact 377 that so many demolitions have been located in a relatively small area explains why businesses 378 in the Christchurch CBD have been so badly impacted (Cousins et al. 2012).

379 CLOSING REMARKS 380 381 1. Material properties for Christchurch clay brick URM buildings were consistent with 382 earlier findings from throughout New Zealand, characterized by good quality clay 383 bricks and weak mortar, such that failure modes tended to be associated with mortar 384 joint failures rather than brick splitting. 385 2. Observed damage patterns to URM buildings were consistent with failure modes 386 seen in previous New Zealand and Australian earthquakes, and documented in 387 many other earthquake prone cities worldwide. 388 3. The majority (90%) of the 39 fatalities caused by earthquake damage to URM build- 389 ings was to pedestrians, persons in vehicles or neighboring buildings, or people who 390 ran from URM buildings in an attempt to escape. 391 4. Earthquake strengthened buildings generally sustained less damage than buildings 392 that had not been seismically upgraded or had been upgraded to lower levels of 393 %NBS:

394 • 73% of URM buildings that had been earthquake strengthened to a level 395 greater than 100%NBS sustained insignificant levels of overall building 396 damage. 397 • 76% of URM buildings that had been earthquake strengthened to a level of 398 between 67% and 100%NBS sustained moderate or lower levels of overall 399 building damage. 400 • 72% of URM buildings that had been earthquake strengthened to a level of 401 between 33% and 67%NBS sustained heavy or major levels of overall building 402 damage. 403 • 60% of URM buildings that had not been retrofit or earthquake strengthened to 404 a level below 33%NBS sustained heavy or major damage or were destroyed.

405 5. A surprising number of strengthened parapets failed after multiple damaging earth- 406 quakes. Further investigation is required into why and, if possible, how these 407 strengthening techniques can be improved. 408 6. Standard earthquake strengthening procedures for URM buildings do not prevent 409 earthquake damage to these buildings. At best these procedures significantly reduce 22 MOON ET AL.

410 the chance of partial or total building collapse (Ingham and Griffith 2011d, EERI 411 1994). Many earthquake strengthened URM buildings required extensive repair, or 412 even demolition. 413 7. Buildings with Type A þ B earthquake strengthening techniques such as shotcrete 414 strengthened walls, additional cross-walls and reinforced concrete or steel strong- 415 backs generally performed reasonably well. 416 8. The five largest earthquakes in the Canterbury earthquake sequence had cumulative 417 absolute velocities and Arias intensities for recording sited in the Christchurch CBD 418 that were well in excess of previously recommended thresholds for partial building 419 collapse, and provide insight regarding the progressive damage to URM buildings 420 when subjected to an extended duration of moderate to severe intensity of ground 421 shaking. 422 REFERENCES 423 Applied Technology Council (ATC), 1985. ATC-13: Earthquake Damage Evaluation Data for 424 California, Redwood City, CA, 492. 425 Applied Technology Council (ATC), 1992. ATC-31: Evaluation of the Performance for Seismi- 426 cally retrofitted buildings, Redwood City, CA, 75. 427 Bailey, S., Dizhur, D., Trowsdale, J., Griffith, M. C., and Ingham, J. M., 2014. Performance of 428 historic posttensioned masonry seismic retrofits during the Canterbury earthquakes, Journal of 429 Performance of Constructed Facilities , under review. 430 Bradley, B. A., Quigley, M. C., Van Dissen, R. J., and Litchfield, N. J., 2013. Ground motion 431 and seismic source aspects of the Canterbury earthquake sequence, Earthquake Spectra 30 , 432 XXX–XXX. 433 Bradley, B. A., and Cubrinovski, M., 2011. Near source strong ground motions observed in the 434 22 February 2011 Christchurch earthquake, Bulletin of the NZSEE 44 , 181 –194. 435 Bruneau, M., 1995. Performance of masonry structures during the 1994 Northridge (Los Angeles) 436 earthquake, Canadian Journal of Civil Engineering 22 , 379 –402. 437 Cabañas, L., Benito, B., and Herráiz, M., 1997. An approach to the measurement of the potential 438 structural damage of earthquake ground motions, Earthq. Eng. Struct. Dyn. 26 , 79 –92. 439 Campbell, K. W., and Bozorgnia, Y., 2012a. Cumulative absolute velocity (CAV) and seismic 440 intensity based on the PEER-NGA database, Earthquake Spectra 28 , 457 –485. 441 Campbell, K. W., and Bozorgnia, Y., 2012b. A comparison of ground motion prediction equa- 442 tions for Arias intensity and cumulative absolute velocity developed using a consistent data- 443 base and functional form database, Earthquake Spectra 28 , 931 –941. 444 Canterbury Earthquakes Royal Commission, 2011. http://canterbury.royalcommission.govt.nz/. 445 Cole, G. L., Dhakal, R. P., and Turner, F. M., 2012. Building pounding damage observed in 446 the 2011 Christchurch earthquake, Earthquake Engineering & Structural Dynamics 41 , 447 893–913. 448 Cooper, M., Carter, R., and Fenwick, R., 2012. Canterbury Earthquake Royal Commission 449 Volume 4: Earthquake Prone Buildings , available at http://canterbury.royalcommission 450 .govt.nz/Final-Report —Part-Two. 451 Cousins, W. J., King, A. B., and Kanga, M., 2012. Accumulated losses from sequences of earth- 452 quakes: Implications for risk modeling, in Proceedings of the 15th World Conference on 453 Earthquake Engineering, 24 –28 September 2012, Lisbon, Portugal. THE DEMISE OF THE URM BUILDING STOCK IN CHRISTCHURCH 23

454 Dizhur, D., Ingham, J. M., Moon, L., Griffith, M., Schultz, A., Senaldi, I., Magenes, G., 455 Dickie, J., Lissel, S., Centeno, J., Ventura, C., Leiti, J., and Lourenco, P., 2011. Performance 456 of masonry buildings and churches in the 22 February 2011 Christchurch earthquake, Bulletin 457 of the NZSEE 44 , 279 –297. 458 Dizhur, D., Ismail, N., Knox, C., Lumantarna, R., and Ingham, J. M., 2010. Performance of 459 unreinforced and retrofitted masonry buildings during the 2010 Darfield earthquake, Bulletin 460 of the NZSEE 43 , 321 –339. 461 Earthquake Engineering Research Institute (EERI), 1994. Expected Seismic Performance of 462 Buildings , Ad Hoc Committee on Seismic Performance, Oakland, CA, 19 pp. 463 Ingham, J. M., Biggs, D. T., and Moon, L. M., 2011. How did unreinforced masonry buildings 464 perform in the February 2011 Christchurch earthquake?, The Structural Engineer 89 (6), 465 14 –18. 466 Ingham, J. M., and Griffith, M. C., 2011a. Damage to unreinforced masonry structures by seismic 467 activity, The Structural Engineer 89 (3), 14 –15. 468 Ingham, J. M., and Griffith, M. C., 2011b. Performance of unreinforced masonry buildings during 469 the 2010 Darfield (Christchurch, NZ) earthquake, Australian Journal of Structural Engineer- 470 ing 11 , 207 –224. 471 Ingham, J. M., and Griffith, M. C., 2011c. The Performance of Unreinforced Masonry 472 Buildings in the 2010 –2011 Canterbury Earthquake Swarm , Commissioned report to 473 the Royal Commission of Inquiry into Building Failure Caused by the Canterbury 474 Earthquake. 475 Ingham, J. M., and Griffith, M. C., 2011d. The Performance of Earthquake Strengthened URM 476 Buildings in the Christchurch CBD in the 22 February 2011 Earthquake , Addendum 477 report commissioned by Royal Commission of Inquiry into Building Failure Caused by 478 the Canterbury Earthquake. 479 Leite, J., Ingham, J. M., and Lourenco, P. B., 2013. Statistical assessment of damage to churches 480 affected by the 2010 –2011 by the Canterbury (New Zealand) earthquake sequence, Journal of 481 Earthquake Engineering 17 , 73 –97. 482 Lumantarna, R., 2012. Material characterisation of New Zealand ’s unreinforced masonry build- 483 ings, Doctoral dissertation, The University of Auckland, Auckland, NZ, 399 pp., available at 484 https://researchspace.auckland.ac.nz/handle/2292/18879. 485 Lumantarna, R., Biggs, D. T., and Ingham, J. M., 2013a. Uniaxial compressive strength and 486 stiffness of field extracted and laboratory constructed masonry prisms, ASCE Journal of Mate- 487 rials in Civil Engineering , available at http://dx.doi.org/10.1061/(ASCE)MT.1943-5533 488 .0000731. 489 Lumantarna, R., Biggs, D. T., and Ingham, J. M., 2013b. Compressive, flexural bond and shear 490 bond strengths of in-situ New Zealand unreinforced clay brick masonry constructed using lime 491 mortar between the 1880s and 1940s, ASCE Journal of Materials in Civil Engineering , avail- 492 able at http://dx.doi.org/10.1061/(ASCE)MT.1943-5533.0000685. 493 Moon, L. M., Griffith, M. C., Ingham, J. M., and Biggs, D. T., 2012. Review of transect of 494 Christchurch CBD following 22 February 2011 earthquake, in 15th World Conference on 495 Earthquake Engineering, Paper #4686, 24 –28 September, Lisbon, Portugal. 496 New Zealand Parliament, 2004. Building Act 2004, Department of Building and Housing — 497 Te Tari Kaupapa Whare, Ministry of Economic Development, New Zealand Government, 498 Wellington, NZ, Date of assent: 24 August 2004. 24 MOON ET AL.

499 New Zealand Society for Earthquake Engineering (NZSEE), 2006. Assessment and Improvement 500 of the Structural Performance of Buildings in Earthquakes , Recommendations of a NZSEE 501 Study Group on Earthquake Risk Buildings. 502 Senaldi, I., Magenes, G., and Ingham, J. M., 2013. Damage assessment of unreinforced stone 503 masonry buildings after the 2010–2011 Canterbury earthquakes, International Journal of 504 Architectural Heritage, under review. 505 Smokehouse Restaurant, 2009. Smokehouse, available at http://www.holysmoke.co.nz (retrieved 506 28 October 2010). 507 Standards New Zealand, 2004. NZS 1170.5:2004, Structural Design Actions Part 5: Earthquake 508 actions–New Zealand, Wellington, NZ. 509 Wailes, C. D., and Horner, A. C., 1933. Earthquake damage analyzed by Long Beach building 510 officials, Engineering News-Record, 25 May 1933. 511 (Received 21 February 2013; accepted 13 July 2013)

View publication stats