404

DAMAGE TO NON-STRUCTURAL COMPONENTS AND

CONTENTS IN 2010 DARFIELD EARTHQUAKE

Rajesh P. Dhakal1

SUMMARY
This paper describes the performance of (or damage to) non-structural components and contents in
buildings during the 4th September 2010 Darfield (Canterbury) earthquake and the subsequent
aftershocks. Even in buildings with little damage to their structural systems, non-structural and content
damages were significant; and these damages were reported to have increased during the aftershocks
(especially those of magnitude 5 and higher). Most commonly damaged non-structural components were
brick chimneys, parapets, ceilings, facades, internal walls and windows. The nature and extent of
damages in each of these components are discussed in this paper with the help of typical damage photos
taken after the earthquake. The extent of content damage in a building was dependent on its usage;
typically buildings using racks/shelves for displaying commodities (such as library, departmental stores,
liquor shops etc) suffered significantly greater loss from content damage than residential houses, office
buildings and other types of commercial buildings.

and type of building, whereas the extent of content damage

INTRODUCTION depended mainly on the usage of the building.
A significant portion of the estimated 4 billion dollars loss
incurred in the 4th September 2010 Darfield (Canterbury)
earthquake and the subsequent aftershocks can be attributed to
the losses from damage to non-structural components (also CHIMNEYS
termed as secondary structural elements) and contents.
Although noticeable structural damage occurred only in a Damage to chimneys was very common in all areas of
small proportion of the building stock, damage to non- Christchurch in this earthquake. In general, brick chimneys
structural components and contents was apparent in almost all damaged most severely (see Figure 1), whereas chimneys
buildings in this event. In many buildings, the extent of made of other materials (even brick chimneys that were
damage to non-structural components (such as chimneys, plastered outside) damaged to a much lesser extent and light
parapets, canopies, facades, partition walls, staircases, metal chimneys did not seem to have suffered any damage
windows) was more than that to the structural components; whatsoever. The damage to brick chimneys ranged from minor
except for old and unretrofitted unreinforced masonry (URM) cracking to the collapse of the whole chimney. In many cases,
buildings. This is in agreement with outcomes of previous falling down of the collapsed chimneys resulted in secondary
seismic loss estimation studies [1, 2] which have concluded damage to roofs; especially those made of tiles (see Figure 1).
that in several buildings non-structural and content damage A crude survey indicated that more than half of brick
contribute a major share of the total loss in an earthquake. chimneys damaged in residential buildings in Christchurch,
and the extent of damage did not show any specific correlation
Commonly observed damages to non-structural components with the age and type of the building.
and contents are described in this report with some typical
damage photos taken after the earthquake. However, it is not
to be misunderstood that the types of damage described herein
occurred in all buildings. At this stage, it is not possible to
PARAPETS
provide a concise figure on percentage of buildings
undergoing each type of damage. More information should Brick parapets (not retrofitted for earthquakes) are common
come to light as the insurance claim details come in. Since features in URM, confined masonry and concrete frame/wall
there is an excess in home insurance policies, damage of very buildings. The majority of parapets in buildings in
trivial nature is unlikely to be reported. However, owners of Christchurch were damaged to different extent in this
most surveyed houses have either lodged a building damage earthquake; parapet with no damage whatsoever was a rare
claim or are planning to do so after the aftershocks cease to sight. The damage included cracking, some bricks being
occur. As structural damage has been minimal except for dislodged and a segment of (in some cases the whole of)
liquefaction effected areas and unretrofitted URM buildings, parapet falling down (see Figure 2). In several cases, the
the majority of building/house damage claims are likely to be detached parapet blocks fell and damaged building parts in
based on damage to non-structural components. This indicates lower storeys and in a couple of locations squashed cars
that percentage of buildings that have undergone non-trivial parked in the roadside underneath. Collapsed parapets in the
damage to non-structural components is very high. The roadside buildings could easily have caused injury/casualty if
severity of non-structural damage was influenced by the age the earthquake had struck during the day.

1
Associate Professor, Department of Civil & Natural Resources Engineering , University of Canterbury, Christchurch (Member)

BULLETIN OF THE NEW ZEALAND SOCIETY FOR EARTHQUAKE ENGINEERING, Vol. 43, No. 4, December 2010
 405

Figure 1: Typical chimney damage.

Some URM buildings had parapet bracing measures prior to

the earthquake. In many cases, these parapets performed well
in clear contrast to severe damage of nearby unbraced parapets
in buildings with similar original construction. Even in
buildings with the walls anchored to the roof/floor, parapets
were found severely damaged (see Figure 2). This was
particularly common in buildings where the height of parapet
above the anchor line was significant. In a few cases, the
parapets on the front elevation of buildings were braced but
not on the side and back elevations which experienced severe
damage. Also in many cases, damaged parapets were typically
removed from tops of URM walls in the clean-up efforts, but
the integrity of the connections of the walls to roofs were still
compromised by the damage. As a result, tops of some
remaining damaged walls will need to be further stabilized
with reliable restraints before repairs commence. In some
cases, severely damaged parapets collapsed during the
aftershocks.

Figure 3: Typical canopy damage: (i) Supported canopy;

Figure 2: Typical parapet damage. (ii) Anchorage failure of suspended canopy; (iii)
Cracked wall due to canopy anchorage.
406

CANOPIES facades from storeys above. Even canopies supported by light

truss suffered damage in some cases (see Figure 3).
In several buildings, the canopy was damaged. A subjective
approximation is that between 10% and 20% canopies were
damaged. Some canopies which were secured to the building CEILINGS
by ties (anchored to the wall) caused stress concentration
around the anchorage, which resulted in cracks in the walls In low-rise residential houses, ceilings generally consist of
(see Figure 3). In some buildings, these anchors were unable plasterboard nailed and/or glued to a light timber frame. In
to resist the extra force generated by the shaking, which commercial buildings, ceilings consist of panels supported on
caused punching shear failure of the walls around the anchor a grid of aluminium beams that are hung though metal wires
(see Figure 3). Needless to mention, this resulted in complete anchored to the floor above. In both cases, the ceiling systems
collapse of the canopy. In many cases the canopy was are generally not engineered for seismic performance. Unlike
damaged because of the impact caused by falling parapets or in earthquake prone countries like USA and Japan, properly

Source: G MacRae

Missing ceiling panel Damaged ceiling Cracked plasterboard

Source: G MacRae

Source: Hush Interior Ltd

Failed grid connection Damaged ceiling components Collapsed plasterboard

Source: Hush Interior Ltd Source: Hush Interior Ltd

Ceiling damage caused by interaction with service

Source: G MacRae

Damaged ceiling being cleared Severely damaged ceiling system

Figure 4: Different types of damage to ceiling systems in residential and commercial buildings.
 407

designed seismic braces were not used in the ceilings in the FAÇADES, INTERNAL LINING & PARTITION WALL
inspected buildings. Hence, it was not surprising that the
ceilings were damaged in several buildings in this earthquake. In most URM buildings, brick walls collapsed in the out-of-
In a crude approximation, 10%-15% of commercial/industrial plane direction, but these are not included here as they are
buildings incurred ceiling damage to different extents. structural damage in URM structures. However, many other
types of buildings such as timber framed, infill masonry and
The proportion of residential houses with ceiling damage was concrete framed buildings had damage to infill walls, partition
significantly less than in commercial/industrial buildings. In walls and facades. Damage of masonry infill did occur in a
residential houses, the common form of ceiling damage was few buildings but was not very common. The worst observed
cracks (of varying length and width) on the plasterboard, damage was to the St. Elmo Courts (see Figure 5), a brick-
crushed plasterboard particles falling on the floor and masonry infilled RC frame building constructed in 1930s. The
plasterboards being detached from the frame (due to punching building exhibited large shear cracking of the infill between
though the nail or tearing off at the glue). On the other hand, windows. The cracks extended the full height of the building.
in commercial buildings, the observed ceiling damage Ceramic tiles attached to structural beams and columns were
included dislodging of the panels, breaking of the panels, also fractured, especially around beam-column joints. Note
failure of the ceiling grid members and connections, failure of that masonry infill in old buildings can easily be categorised
perimeter angles and damage of ceiling panels due to as structural components and the aforementioned damage in St
interaction with the services. Some photographs of typical Elmo Courts could also be argued as structural damage.
ceiling damage are presented in Figure 4.
Within the central city, the majority of facade damage was to
medium height buildings with infill and exterior lightweight
claddings. Damage to glass panes was visible from street
throughout greater Christchurch, but this has been included in

Damage to façades

Source: A Abu Source: G MacRae Source: R Diaferia

Fracture of ceramic tiles Detached

Infill masonry damagesignboard
in St Elmo Courts
Source: G MacRae

Source: G MacRae
Damage to walls and linings

Figure 5: Damage to non structural vertical elements.

408

Source: A Baird

Damage to old window panels Spider glazing damage Damage to window and wall
Figure 6: Typical damage to windows.
the next section under “windows damage”, which could also non-structural elements and suffered no interruption to their
have been categorized as facades. Many residential houses business services. One notable exception was damage to
exhibited warping of their joinery without any cracks visible storage rack systems and the subsequent loss of stock. The
in the glass. The other type of facade damage consisted damage to the rack systems varied. Complete collapse
predominantly of brick facades falling out due to poor occurred to heavily loaded, relatively light gauge racks. It was
connection with the structure. In the partition walls and reported by storeowners that more things fell from racks that
internal linings, it was very common to see cracks initiating were secured to the floor but not to the walls than the racks
from door and window corners (see Figure 5). Some cases of that were not secured to the floor. This might have been due to
observed façade, wall, lining, tile damage are shown in the lower forces in the unsecured racks as a result of rocking and
photos in Figure 5. As shown in the photos, ceramic tiles on sliding. In some cases, it has also been found that inadequate
the walls (and floor in some cases) also suffered damage. provision of bolts in one segment of racking system resulted in
twisting of the whole rack (see Figure 7). In some cases,
In many buildings, the aftershocks (especially; the 5.1 cracks were found in the infill wall panel where the racks were
magnitude crustal aftershock which originated in Lyttleton, anchored (see Figure 7). A sampling of the observed racking
less than 10 km from Christchurch, on the early morning of damage is shown in Figure 7.
Wednesday 8 September) caused additional damage to non-
structural components. It was reported that new cracks on In offices, unsecured book shelves fell to the floor. In libraries,
walls and internal linings appeared and the existing cracks even the shelves that were tied together collapsed (see Figure
widened and extended during the aftershocks. Being a near 7). It was found that the ties in this case were not strong
source and very shallow earthquake, this aftershock had a enough and failed to resist the tilting tendency of the shelves.
higher dominant frequency than the main event, which is Shelves tied to the wall and tied with each other using strong
closer to the natural frequency of the low rise residential ties were intact.
building stock. Understandably, this aftershock caused
noticeable damage to these buildings.

CONTENTS
The earthquake was strong enough to cause things fall down
WINDOWS
from their elevated position in almost every building in
Broken glass panels in windows were observed in several Christchurch. The extent of content damage varied greatly
buildings. The worst of the glass panel damage was focussed depending on the location of the building and the use of the
in the central city, where the majority of Christchurch’s taller building. In residential houses, the common contents that were
and historic buildings are located. Window panels cracked in damaged include racks, cutlery, vases, photo frames, arts,
the main shock were reported to have broken in the decorative pieces and aquarium. In a small number of houses;
aftershocks in some buildings. As the window framing system television sets were reported to have fallen off from the
used in the old buildings was rigid and did not allow relative cabinets. Contents were reported to have damaged also in the
glass displacement, most broken glass panes were observed in aftershocks; especially the Wednesday (8th September)
this type of window frames. On the other hand, modern morning’s 5.1 magnitude aftershock appeared to have caused
aluminium frame windows have deformable rubber sealing content damage amounting to more than half of that in the
which allow the glass panels to displace to some extent. main earthquake in some houses/shops. More than half of the
Hence, very few broken glass panels were observed in this surveyed residential households were not planning to lodge an
type of window. Although spider glazing is a modern system, insurance claim for content damage which means that the
damage to this type of glazing was observed; for example in content damage was less than the excess of their content
the Westpac building in the city centre (see Fig 6). In all cases, insurance policy (typically 250 dollars).
the window glass damage posed a falling hazard for
pedestrians. Some typical window damages are shown in Fig In industrial buildings, content damage was almost none
6. (except for damage to racks in some industrial buildings as
mentioned earlier) because the heavy machineries and
equipments expectedly did not fall down from their positions.
In offices, despite several things (including documents) fell
down, there was not much that could not be reused. Office
RACKS AND SHELVES equipments (such as printer, photocopiers) are generally not
Racks and shelves are common in industrial, commercial and secured to the floor with any seismic restraints; however, they
office buildings. Apart from liquefaction affected areas, mostly remained operational after the earthquake. On the other
industrial establishments typically suffered little damage to
 409

Source: SR Uma Source: SR Uma

Crack developed in the wall behind the rack at the anchor Twisted racks due to insufficient bolt in a segment

Source: M Hannah

Source: N Crannitch

Damaged racks at a storage facility Collapsed racks at a metal fabricator warehouse

Source: G MacRae

Damage to book shelves in university library

Figure 7: Typical damage to racks and shelves.

extreme, content damage was a major contributor to the total items in racks also reported extensive content damage. Some
loss in commercial buildings. typical photos of content damage are shown in Figure 8.
Depending on the type of business, the extent of content
damage varied greatly. In a poultry farm, it was reported that EXAMPLES OF GOOD PRACTICES
3,000 chickens were killed and thousands of eggs were
broken. In many shops, things put inside the freezer got spoilt The extent of damaged contents varied greatly depending on
because of power disruption, which lasted from a few hours in the type of display racks. For example, there was no content
most suburbs of Christchurch to a few days in some. As damage whatsoever in a shop selling fragile items such as
expected, there was little content lost in shops selling trophies, glassware, plaques, frames etc. A lot of these delicate
garments, shoes, beds, flowers, furniture and all other business things were hung on wall, displayed on racks secured to the
which did not display items in racks. Similarly, some wall and standalone racks, still nothing fell and broke. It was
businesses providing services also did not incur much content found that the racks secured to the walls had an angle at the
loss. Nevertheless, businesses selling everyday commodities front edge, which stopped things from falling down despite
suffered severe content damage. Flooding on the floor from being displaced from their original position. Even at other
broken bottles was a common sight in department stores, locations, lips and rods at the edge of shelves (see Figure 9)
liquor shops, bars and restaurants (see Figure 8). Pharmacies, performed quite well, reportedly preventing the sliding or
gift shops, and several other businesses which display fragile toppling of many contents. The standalone display racks were
410

Furniture and artifacts damage in a church. Liquor damage in a store.

Source: Greg MacRae

Cabinet and book damage in an office. Damaged contents in a warehouse.

Figure 8: Contents damage.
provided with rollers at the base, which acted like base interconnected by a stronger tie remained intact. Some typical
isolation and prevented the rack from rocking which would cases of good storage practice leading to reduced risk of
have caused the things to fall down. The racks on the roller content damage are shown in Figure 9.
moved a small distance, but the carpet on the floor restricted
the rack from rolling haphazardly. However, racks on rollers
may be very unstable in uncarpeted floors. LESSONS LEARNT

Similarly, there was generally no damage to racks that were  Brick chimneys are very vulnerable and should not be
anchored to walls. It was learnt from a liquor shop that a rack used in new constructions. Even in existing buildings,
with a bigger footprint had a much smaller likelihood of brick chimneys should ideally be properly braced or
bottles falling down than smaller racks. In the University replaced. Bracing is challenging as the chimneys tend
library, book shelves connected to each other by a small tie to be very heavy and any bracing will have to anchor to
toppled whereas shelves anchored to the wall and/or very light wood roof framing members. A viable option
may be to remove the chimneys down to the roof line,

Figure 9: Examples of good storage practice.

 411

provide a concrete confining cap at the roof line, anchor prescriptive installation and seismic bracing details are
to the roof at that point and replace with a properly utilized.
engineered light chimney above the roof line.  Analysis of insurance claims on non-structural and
 Unsecured parapets are highly susceptible to severe content damage.
damage, potentially leading to collapse in earthquakes.  Methods to improve seismic performance of racks and
They should be braced to the buildings to prevent shelves. In particular, the current NZ racking design
collapse. guidelines need to be compared with the observations
 Windows with tight frames and without any deformable from this earthquake, potentially leading to amendment
sealing (as in old windows) are very vulnerable to glass of the guidelines.
breaking in earthquakes. The modern windows with
aluminium frames and rubber sealing performed very ACKNOWLEDGMENTS
well.
 The members used in truss/grids/frames to support This paper is a modified and reformatted version of a report on
and/or brace non-structural components such as the same topic prepared for (and available in) the Darfield
canopies, ceilings, racks, shelves should be properly (Canterbury) Earthquake Clearinghouse. The author extends
designed to resist the effect of seismic actions. special thanks to Fred Turner and Greg MacRae for their
 Racks and shelves should be properly anchored to walls feedback on the first draft of the Clearinghouse report, which
wherever possible. In some cases, the racks with helped the author to refine the paper. The author also
insufficient anchorage performed poorly than acknowledges the information provided on ceilings and
unanchored racks. However, this should not be taken as facades by the Non-Structural Components Research Group at
a justification to put unanchored racks on the floor. A University of Canterbury (G. MacRae, R. Dhakal, A. Palermo,
properly designed racking system that is also adequately S.Pampanin and their students J. Hair, G. Paganotti, J. Singh,
anchored will provide the best performance in relation to A. Baird and R. Diaferia); by D. Mukai on industrial racks;
life safety and protection of contents. It is important for and by SR Uma of GNS on racks and shelves. The author has
new facilities to use seismically engineered rack systems based the facts stated in this paper on the observation and
which would consider the wall anchorage as a boundary information gathered from an extensive building survey
condition of the system. conducted in different parts of the earthquake affected area by
the author and G. MacRae, H. Gavin, D. Mukai, J. Crosier,
 Wherever feasible, wider racks should be used instead of
J.K. Min, M. Hannah, J. Byrne, C. Muir, P. Grange, Mustafa,
thin/slender racks.
V. Sadashiv, G. Cole, D. Gardiner, M. Newcombe, A. Lu. The
 Angles, lips or rods at the edges of racks are very
author thanks all these people and other people who took the
effective in preventing the contents from falling.
photographs used in this paper.

REFERENCES
RESEARCH NEEDS
1. Bradley, B., Dhakal, R. P., Cubrinovski, M., MacRae, G.
 Bracing scheme for existing brick chimneys and come and Lee, D., (2009), “Seismic Loss Estimation for
up with seismic resistant chimney system (probably Efficient Decision Making”, Bulletin of the New Zealand
already exists in the form of modern metal chimneys, Society for Earthquake Engineering, Vol. 42, No. 2, pp
but this needs to be verified/enforced). 97-110.
 Methods to stabilize facades, parapets and canopies in
existing buildings. 2. Dhakal, R.P., Mander, J.B. and Xu, L. (2010), “Seismic
 Seismic performance assessment and methods to Financial Loss Estimation of Steel Moment Frame
improve seismic performance of common ceiling Buildings”. Journal of the International Review of Civil
systems used in NZ. Engineering (IRECE), Vol. 1, No. 2, pp 130-142.
 Investigate suspended ceiling seismic bracing systems
currently in use (if any) and their relative performance.
Ceiling performance can improve if some simple and