In this paper, the characteristics of near-fault ground motions recorded during the Mw7.1 Darfield and Mw 6.2 Christchurch earthquakes are examined and compared with existing empirical models. The characteristics of forward-directivity effects are first examined using a wavelet-based pulse-classification algorithm. This is followed by an assessment of the adequacy of empirical models which aim to capture the effect of directivity effects on amplifying the acceleration response spectra; and the period and peak velocity of the forward-directivity pulse. It is illustrated that broadband directivity models developed by Somerville et al. (1997) and Abrahamson (2000) generally under-predict the observed amplification of response spectral ordinates at longer vibration periods. In contrast, a recently developed narrowband model by Shahi and Baker (2011) provides significantly improved predictions by amplifying the response spectra within a small range of periods surrounding the directivity pulse period. Although the empirical predictions of the pulse period are generally favourable for the Christchurch earthquake, the observations from the Darfield earthquake are significantly under-predicted. The elongation in observed pulse periods is inferred as being a result of the soft sedimentary soils of the Canterbury basin. However, empirical predictions of the observed peak velocity associated with the directivity pulse are generally adequate for both events.
Liquefaction-induced lateral spreading during the 2011 Christchurch earthquake in New Zealand was severe and extensive, and data regarding the displacements associated with the lateral spreading provides an excellent opportunity to better understand the factors that influence these movements. Horizontal displacements measured from optical satellite imagery and subsurface data from the New Zealand Geotechnical Database (NZGD) were used to investigate four distinct lateral spread areas along the Avon River in Christchurch. These areas experienced displacements between 0.5 and 2 m, with the inland extent of displacement ranging from 100 m to over 600 m. Existing empirical and semi-empirical displacement models tend to under estimate displacements at some sites and over estimate at others. The integrated datasets indicate that the areas with more severe and spatially extensive displacements are associated with thicker and more laterally continuous deposits of liquefiable soil. In some areas, the inland extent of displacements is constrained by geologic boundaries and geomorphic features, as expressed by distinct topographic breaks. In other areas the extent of displacement is influenced by the continuity of liquefiable strata or by the presence of layers that may act as vertical seepage barriers. These observations demonstrate the need to integrate geologic/geomorphic analyses with geotechnical analyses when assessing the potential for lateral spreading movements.
The 2010 Darfield and 2011 Christchurch Earthquakes triggered extensive liquefaction-induced lateral spreading proximate to streams and rivers in the Christchurch area, causing significant damage to structures and lifelines. A case study in central Christchurch is presented and compares field observations with predicted displacements from the widely adopted empirical model of Youd et al. (2002). Cone penetration testing (CPT), with measured soil gradation indices (fines content and median grain size) on typical fluvial deposits along the Avon River were used to determine the required geotechnical parameters for the model input. The method presented attempts to enable the adoption of the extensive post-quake CPT test records in place of the lower quality and less available Standard Penetration Test (SPT) data required by the original Youd model. The results indicate some agreement between the Youd model predictions and the field observations, while the majority of computed displacements error on the side of over-prediction by more than a factor of two. A sensitivity analysis was performed with respect to the uncertainties used as model input, illustrating the model’s high sensitivity to the input parameters, with median grain size and fines content among the most influential, and suggesting that the use of CPT data to quantify these parameters may lead to variable results.
Liquefaction-induced lateral spreading in Christchurch and surrounding suburbs during the recent Canterbury Earthquake Sequence (2010-2011) caused significant damage to structures and lifelines located in close proximity to streams and rivers. Simplified methods used in current engineering practice for predicting lateral ground displacements exhibit a high degree of epistemic uncertainty, but provide ‘order of magnitude’ estimates to appraise the hazard. We wish to compare model predictions to field measurements in order to assess the model’s capabilities and limitations with respect to Christchurch conditions. The analysis presented focuses on the widely-used empirical model of Youd et al. (2002), developed based on multi-linear regression (MLR) of case history data from lateral spreading occurrence in Japan and the US. Two issues arising from the application of this model to Christchurch were considered: • Small data set of Standard Penetration Test (SPT) and soil gradation indices (fines content FC, and mean grain size, D50) required for input. We attempt to use widely available CPT data with site specific correlations to FC and D50. • Uncertainty associated with the model input parameters and their influence on predicted displacements. This has been investigated for a specific location through a sensitivity analysis.
This poster provides a summary of the development of a 3D shallow (z<40m) shear wave velocity (Vs) model for the urban Christchurch, New Zealand region. The model is based on a recently developed Christchurch-specific empirical correlation between Vs and cone penetration test (CPT) data (McGann et al. 2014a,b) and the large high-density database of CPT logs in the greater Christchurch urban area (> 15,000 logs as of 01/01/2014). In particular, the 3D model provides shear wave velocities for the surficial Springston Formation, Christchurch Formation, and Riccarton gravel layers which generally comprise the upper 40m in the Christchurch urban area. Point-estimates are provided on a 200m-by- 200m grid from which interpolation to other locations can be performed. This model has applications for future site characterization and numerical modeling efforts via maps of timeaveraged Vs over specific depths (e.g. Vs30, Vs10) and via the identification of typical Vs profiles for different regions and soil behaviour types within Christchurch. In addition, the Vs model can be used to constrain the near-surface velocities for the 3D seismic velocity model of the Canterbury basin (Lee et al. 2014) currently being developed for the purpose of broadband ground motion simulation.
The Canterbury Earthquake Sequence (CES), induced extensive damage in residential buildings and led to over NZ$40 billion in total economic losses. Due to the unique insurance setting in New Zealand, up to 80% of the financial losses were insured. Over the CES, the Earthquake Commission (EQC) received more than 412,000 insurance claims for residential buildings. The 4 September 2010 earthquake is the event for which most of the claims have been lodged with more than 138,000 residential claims for this event only. This research project uses EQC claim database to develop a seismic loss prediction model for residential buildings in Christchurch. It uses machine learning to create a procedure capable of highlighting critical features that affected the most buildings loss. A future study of those features enables the generation of insights that can be used by various stakeholders, for example, to better understand the influence of a structural system on the building loss or to select appropriate risk mitigation measures. Previous to the training of the machine learning model, the claim dataset was supplemented with additional data sourced from private and open access databases giving complementary information related to the building characteristics, seismic demand, liquefaction occurrence and soil conditions. This poster presents results of a machine learning model trained on a merged dataset using residential claims from the 4 September 2010.
Recent earthquakes have shown that liquefaction and associated ground deformations are major geotechnical hazards to civil engineering infrastructures, such as pipelines. In particular, sewer pipes have been damaged in many areas in Christchurch as a result of liquefaction-induced lateral spreading near waterways and ground oscillation induced by seismic shaking. In this paper, the addition of a flexible AM liner as a potential countermeasure to increase sewer pipe capacity was investigated. Physical testing through 4-point loading test was undertaken to characterise material properties and the response of both unlined pipe and its lined counterpart. Next, numerical models were created using SAP2000 and ABAQUS to analyse buried pipeline response to transverse permanent ground displacement and to quantify, over a range of pipe segment lengths and soil parameters, the effectiveness of the AM liner in increasing displacement capacity. The numerical results suggest that the addition of the AM liner increases the deformation capacity of the unlined sewer pipe by as much as 50 times. The results confirmed that AM liner is an effective countermeasure for sewer pipes in liquefied ground not only in terms of increased deformation capacity but also the fact that AM-Liner can prevent influx of sand and water through broken pipes, making sewer pipes with liner remaining serviceable even under severe liquefaction condition.
Recent earthquakes in New Zealand proved that a shift is necessary in the current design practice of structures to achieve better seismic performance. Following such events, the number of new buildings using innovative technical solutions (e.g. base isolation, controlled rocking systems, damping devices, etc.), has increased, especially in Christchurch. However, the application of these innovative technologies is often restricted to medium-high rise buildings due to the maximum benefit to cost ratio. In this context, to address this issue, a multi-disciplinary geo-structural-environmental engineering project funded by the Ministry of Business Innovation and Employment (MBIE) is being carried out at the University of Canterbury. The project aims at developing a foundation system which will improve the seismic performance of medium-density low-rise buildings. Such foundation is characterized by two main elements: 1) granulated tyre rubber mixed with gravelly soils to be placed beneath the structure, with the goal of damping part of the seismic energy before it reaches the superstructure; and 2) a basement raft made of steel-fibre rubberised concrete to enhance the flexibility of the foundation under differential displacement demand. In the first part of this paper, the overarching objectives, scope and methodology of the project will be briefly described. Then, preliminary findings on the materials characterization, i.e., the gravel-rubber mixtures and steel-fibre rubberised concrete mixes, will be presented and discussed with focus on the mechanical behaviour.
The city of Christchurch and its surrounds experienced widespread damage due to soil liquefaction induced by seismic shaking during the Canterbury earthquake sequence that began in September 2010 with the Mw7.1 Darfield earthquake. Prior to the start of this sequence, the city had a large network of strong motion stations (SMSs) installed, which were able to record a vast database of strong ground motions. This paper uses this database of strong ground motion recordings, observations of liquefaction manifestation at the ground surface, and data from a recently completed extensive geotechnical site investigation program at each SMS to assess a range of liquefaction evaluation procedures at the four SMSs in the Christchurch Central Business District (CBD). In general, the characteristics of the accelerograms recorded at each SMS correlated well with the liquefaction evaluation procedures, with low liquefaction factors of safety predicted at sites with clear liquefaction identifiers in the ground motions. However, at sites that likely liquefied at depth (as indicated by evaluation procedures and/or inferred from the characteristics of the recorded surface accelerograms), the presence of a non-liquefiable crust layer at many of the SMS locations prevented the manifestation of any surface effects. Because of this, there was not a good correlation between surface manifestation and two surface manifestation indices, the Liquefaction Potential Index (LPI) and the Liquefaction Severity Number (LSN).
The objective of this study is to examine the influence of near-fault motions on liquefaction triggering in Christchurch and neighboring towns during the 2010-2011 Canterbury earthquake sequence (CES). The CES began with the 4 September 2010, Mw7.1 Darfield earthquake and included up to ten events that triggered liquefaction. However, most notably, widespread liquefaction was induced by the Darfield earthquake and the Mw6.2, 22 February 2011 Christchurch earthquake. Of particular relevance to this study is the forward directivity effects that were prevalent in the motions recorded during the Darfield earthquake, and to a much lesser extent, during the Christchurch earthquake. A 2D variant of the Richart-Newmark fatigue theory was used to compute the equivalent number of cycles (neq) for the ground motions, where volumetric strain was used as the damage metric. This study is unique because it considers the contribution and phasing of both the fault-normal and fault-parallel components of motion on neq and the magnitude scaling factor (MSF). It was found that when the fault-normal and fault-parallel motions were treated individually, the former yielded a lower neq than the latter. Additionally, when the combined effects of fault-normal and fault-parallel components were considered, it was found that the MSF were higher than those commonly used. This implies that motions containing near-fault effects are less demanding on the soil than motions that do not. This may be one of several factors that resulted in less severe liquefaction occurring during the Darfield earthquake than the Christchurch earthquake.
In the period between September 2010 and December 2011, Christchurch was shaken by a series of strong earthquakes including the MW7.1 4 September 2010, Mw 6.2 22 February 2011, MW6.2 13 June 2011 and MW6.0 23 December 2011 earthquakes. These earthquakes produced very strong ground motions throughout the city and surrounding areas that resulted in soil liquefaction and lateral spreading causing substantial damage to buildings, infrastructure and the community. The stopbank network along the Kaiapoi and Avon River suffered extensive damage with repairs projected to take several years to complete. This presented an opportunity to undertake a case-study on a regional scale of the effects of liquefaction on a stopbank system. Ultimately, this information can be used to determine simple performance-based concepts that can be applied in practice to improve the resilience of river protection works. The research presented in this thesis draws from data collected following the 4th September 2010 and 22nd February 2011 earthquakes. The stopbank damage is categorised into seven key deformation modes that were interpreted from aerial photographs, consultant reports, damage photographs and site visits. Each deformation mode provides an assessment of the observed mechanism of failure behind liquefaction-induced stopbank damage and the factors that influence a particular style of deformation. The deformation modes have been used to create a severity classification for the whole stopbank system, being ‘no or low damage’ and ‘major or severe damage’, in order to discriminate the indicators and factors that contribute to ‘major to severe damage’ from the factors that contribute to all levels of damage a number of calculated, land damage, stopbank damage and geomorphological parameters were analysed and compared at 178 locations along the Kaiapoi and Avon River stopbank systems. A critical liquefiable layer was present at every location with relatively consistent geotechnical parameters (cone resistance (qc), soil behaviour type (Ic) and Factor of Safety (FoS)) across the study site. In 95% of the cases the critical layer occurred within two times the Height of the Free Face (HFF,). A statistical analysis of the geotechnical factors relating to the critical layer was undertaken in order to find correlations between specific deformation modes and geotechnical factors. It was found that each individual deformation mode involves a complex interplay of factors that are difficult to represent through correlative analysis. There was, however, sufficient data to derive the key factors that have affected the severity of deformation. It was concluded that stopbank damage is directly related to the presence of liquefaction in the ground materials beneath the stopbanks, but is not critical in determining the type or severity of damage, instead it is merely the triggering mechanism. Once liquefaction is triggered it is the gravity-induced deformation that causes the damage rather than the shaking duration. Lateral spreading and specifically the depositional setting was found to be the key aspect in determining the severity and type of deformation along the stopbank system. The presence or absence of abandoned or old river channels and point bar deposits was found to significantly influence the severity and type of deformation. A review of digital elevation models and old maps along the Kaiapoi River found that all of the ‘major to severe’ damage observed occurred within or directly adjacent to an abandoned river channel. Whilst a review of the geomorphology along the Avon River showed that every location within a point bar deposit suffered some form of damage, due to the depositional environment creating a deposit highly susceptible to liquefaction.
Study region: Christchurch, New Zealand. Study focus: Low-lying coastal cities worldwide are vulnerable to shallow groundwater salinization caused by saltwater intrusion and anthropogenic activities. Shallow groundwater salinization can have cascading negative impacts on municipal assets, but this is rarely considered compared to impacts of salinization on water supply. Here, shallow groundwater salinity was sampled at high spatial resolution (1.3 piezometer/km2 ), then mapped and spatially interpolated. This was possible due to a uniquely extensive set of shallow piezometers installed in response to the 2010–11 Canterbury Earthquake Sequence to assess liquefaction risk. The municipal assets located within the brackish groundwater areas were highlighted. New hydrological insights for the region: Brackish groundwater areas were centred on a spit of coastal sand dunes and inside the meander of a tidal river with poorly drained soils. The municipal assets located within these areas include: (i) wastewater and stormwater pipes constructed from steel-reinforced concrete, which, if damaged, are vulnerable to premature failure when exposed to chloride underwater, and (ii) 41 parks and reserves totalling 236 ha, within which salt-intolerant groundwater-dependent species are at risk. This research highlights the importance of determining areas of saline shallow groundwater in low-lying coastal urban settings and the co-located municipal assets to allow the prioritisation of sites for future monitoring and management.
This thesis is concerned with springs that appeared in the Hillsborough, Christchurch during the 2010-2011 Canterbury Earthquake Sequence, and which have continued to discharge groundwater to the surface to the present time. Investigations have evolved, measurements of discharge at selected sites, limited chemical data on anions and isotope analysis. The springs are associated with earthquake generated fissures (extensional) and compression zones, mostly in loess-colluvium soils of the valley floor and lower slopes. Extensive peat swamps are present in the Hillsborough valley, with a groundwater table at ~1m below ground. The first appearance of the ‘new’ springs took place following the Mw 7.1 Darfield Earthquake on 4 September 2010, and discharges increased both in volume and extent of the Christchurch Mw 6.3 Earthquake of 22 February 2011. Five monitored sites show flow rates in the range of 4.2-14.4L/min, which have remained effectively constant for the duration of the study (2014-2015). Water chemistry analysis shows that the groundwater discharges are sourced primarily from volcanic bedrocks which underlies the valley at depths ≤50m below ground level. Isotope values confirm similarities with bedrock-sourced groundwater, and the short term (hours-days) influence of extreme rainfall events. Cyclone Lusi (2013-2014) affects were monitored and showed recovery of the bedrock derived water signature within 72 hours. Close to the mouth of the valley sediments interfinger with Waimakiriri River derived alluvium bearing a distinct and different isotope signature. Some mixing is evident at certain locations, but it is not clear if there is any influence from the Huntsbury reservoir which failed in the Port Hills Earthquake (22 February 2011) and stored groundwater from the Christchurch artesian aquifer system (Riccarton Gravel).
The 2015 New Zealand strong-motion database provides a wealth of new strong motion data for engineering applications. An important component of this database is the compilation of new site metadata, describing the soil conditions and site response at GeoNet strong motion stations. We have assessed and compiled four key site parameters for the ~460 GeoNet stations that recorded significant historical ground motions. Parameters include: site classification (NZS1170.5), Vs30, fundamental site period (Tsite) and depth to bedrock (Z1.0, i.e. depth to material with Vs > 1000 m/s). In addition, we have assigned a quality estimate (Quality 1 – 3) to these parameters to provide a qualitative estimate of the uncertainty. New highquality Tsite estimates have largely been obtained from newly available HVSR amplification curves and spectral ratios from inversion of regional strong motion data that has been reconciled with available geological information. Good quality Vs30 estimates, typically in urban centres, have also been incorporated following recent studies. Where site-specific measurements of Vs30 are not available, Vs30 is estimated based on surface geology following national Vs30 maps. New Z1.0 values have been provided from 3D subsurface models for Canterbury and Wellington. This database will be used in efforts to guide development and testing of new and existing ground motion prediction models in New Zealand. In particular, it will allow reexamination of the most important site parameters that control and predict site response in a New Zealand setting. Furthermore, it can be used to provide information about suitable rock reference sites for seismological research, and as a guide to site-specific references in the literature. We discuss compilation of the database, preliminary insights so far, and future directions.
The magnitude Mw7.8 ‘Kaikōura’ earthquake occurred shortly after midnight on 14 November 2016. This paper presents an overview of the geotechnical impacts on the South Island of New Zealand recorded during the postevent reconnaissance. Despite the large moment magnitude of this earthquake, relatively little liquefaction was observed across the South Island, with the only severe manifestation occurring in the young, loose alluvial deposits in the floodplains of the Wairau and Opaoa Rivers near Blenheim. The spatial extent and volume of liquefaction ejecta across South Island is significantly less than that observed in Christchurch during the 2010-2011 Canterbury Earthquake Sequence, and the impact of its occurrence to the built environment was largely negligible on account of the severe manifestations occurring away from the areas of major development. Large localised lateral displacements occurred in Kaikōura around Lyell Creek. The soft fine-grained material in the upper portions of the soil profile and the free face at the creek channel were responsible for the accumulation of displacement during the ground shaking. These movements had severely impacted the houses which were built close (within the zone of large displacement) to Lyell Creek. The wastewater treatment facility located just north of Kaikōura also suffered tears in the liners of the oxidation ponds and distortions in the aeration system due to ground movements. Ground failures on the Amuri and Emu Plains (within the Waiau Valley) were small considering the large peak accelerations (in excess of 1g) experienced in the area. Minor to moderate lateral spreading and ejecta was observed at some bridge crossings in the area. However, most of the structural damage sustained by the bridges was a result of the inertial loading, and the damage resulting from geotechnical issues were secondary.
Semi-empirical models based on in-situ geotechnical tests have become the standard of practice for predicting soil liquefaction. Since the inception of the “simplified” cyclic-stress model in 1971, variants based on various in-situ tests have been developed, including the Cone Penetration Test (CPT). More recently, prediction models based soley on remotely-sensed data were developed. Similar to systems that provide automated content on earthquake impacts, these “geospatial” models aim to predict liquefaction for rapid response and loss estimation using readily-available data. This data includes (i) common ground-motion intensity measures (e.g., PGA), which can either be provided in near-real-time following an earthquake, or predicted for a future event; and (ii) geospatial parameters derived from digital elevation models, which are used to infer characteristics of the subsurface relevent to liquefaction. However, the predictive capabilities of geospatial and geotechnical models have not been directly compared, which could elucidate techniques for improving the geospatial models, and which would provide a baseline for measuring improvements. Accordingly, this study assesses the realtive efficacy of liquefaction models based on geospatial vs. CPT data using 9,908 case-studies from the 2010-2016 Canterbury earthquakes. While the top-performing models are CPT-based, the geospatial models perform relatively well given their simplicity and low cost. Although further research is needed (e.g., to improve upon the performance of current models), the findings of this study suggest that geospatial models have the potential to provide valuable first-order predictions of liquefaction occurence and consequence. Towards this end, performance assessments of geospatial vs. geotechnical models are ongoing for more than 20 additional global earthquakes.
The objective of the study presented herein is to assess three commonly used CPT-based liquefaction evaluation procedures and three liquefaction severity index frameworks using data from the 2010–2011 Canterbury earthquake sequence. Specifically, post-event field observations, ground motion recordings, and results from a recently completed extensive geotechnical site investigation programme at selected strong motion stations (SMSs) in the city of Christchurch and surrounding towns are used herein. Unlike similar studies that used data from free-field sites, accelerogram characteristics at the SMS locations can be used to assess the performance of liquefaction evaluation procedures prior to their use in the computation of surficial manifestation severity indices. Results from this study indicate that for cases with evidence of liquefaction triggering in the accelerograms, the majority of liquefaction evaluation procedures yielded correct predictions, regardless of whether surficial manifestation of liquefaction was evident or not. For cases with no evidence of liquefaction in the accelerograms (and no observed surficial evidence of liquefaction triggering), the majority of liquefaction evaluation procedures predicted liquefaction was triggered. When all cases are used to assess the performance of liquefaction severity index frameworks, a poor correlation is shown between the observed severity of liquefaction surface manifestation and the calculated severity indices. However, only using those cases where the liquefaction evaluation procedures yielded correct predictions, there is an improvement in the correlation, with the Liquefaction Severity Number (LSN) being the best performing of the frameworks investigated herein. However scatter in the relationship between the observed and calculated surficial manifestation still remains for all liquefaction severity index frameworks.
This thesis presents an assessment of historic seismic performance of the New Zealand stopbank network from the 1968 Inangahua earthquake through to the 2016 Kaikōura earthquake. An overview of the types of stopbanks and the main aspects of the design and construction of earthen stopbanks was presented. Stopbanks are structures that are widely used on the banks of rivers and other water bodies to protect against the impact of flood events. Earthen stopbanks are found to be the most used for such protection measures. Different stopbank damage or failure modes that may occur due to flooding or earthquake excitation were assessed with a focus on past earthquakes internationally, and examples of these damage and failure modes were presented. Stopbank damage and assessment reports were collated from available reconnaissance literature to develop the first geospatial database of stopbank damage observed in past earthquakes in New Zealand. Damage was observed in four earthquakes over the past 50 years, with a number of earthquakes resulting in no stopbank damage. The damage database therefore focussed on the Edgecumbe, Darfield, Christchurch and Kaikōura earthquakes. Cracking of the crest and liquefaction-induced settlement were the most common forms of damage observed. To understand the seismic demand on the stopbank network in past earthquakes, geospatial analyses were undertaken to approximate the peak ground acceleration (PGA) across the stopbank network for ten large earthquakes that have occurred in New Zealand over the past 50 years. The relationship between the demand, represented by the peak ground acceleration (PGA) and damage is discussed and key trends identified. Comparison of the seismic demand and the distribution of damage suggested that the seismic performance of the New Zealand stopbank network has been generally good across all events considered. Although a significant length of the stopbank networks were exposed to high levels of shaking in past events, the overall damage length was a small percentage of this. The key aspect controlling performance was the performance of the underlying foundation soils and the effect of this on the stopbank structure and stability.
The collapse of Redcliffs’ cliff in the 22 February 2011 and 13 June 2011 earthquakes were the first times ever a major failure incident occurred at Redcliffs in approximately 6000 years. This master’s thesis is a multidisciplinary engineering geological investigation sought to study these particular failure incidents, focusing on collecting the data necessary to explain the cause and effect of the cliff collapsing in the event of two major earthquakes. This study provides quantitative and qualitative data about the geotechnical attributes and engineering geological nature of the sea-cut cliff located at Redcliffs. Results from surveying the geology of Redcliffs show that the exposed lithology of the cliff face is a variably jointed rock body of welded and (relatively intact) unwelded ignimbrite, a predominantly massive unit of brecciated tuff, and a covering of wind-blown loess and soil deposit (commonly found throughout Canterbury) on top of the cliff. Moreover, detailing the external component of the slope profile shows that Redcliffs’ cliff is a 40 – 80 m cliff with two intersecting (NE and SE facing) slope aspects. The (remotely) measured geometry of the cliff face comprises of multiple outstanding gradients, averaging a slope angle of ~67 degrees (post-13 June 2011), where the steepest components are ~80 degrees, whereas the gentle sloping sections are ~44 degrees. The physical structure of Redcliffs’ cliff drastically changed after each collapse, whereby seismically induced alterations to the slope geometry resulted in material deposited on the talus at the base of the cliff. Prior to the first collapse, the variance of the gradient down the slope was minimal, with the SE Face being the most variable with up to three major gradients on one cross section. However, after each major collapse, the variability increased with more parts of the cliff face having more than one major gradient that is steeper or gentler than the remainder of the slope. The estimated volume of material lost as a result of the gradient changes was 28,267 m³ in February and 11,360 m³ in June 2011. In addition, surveys of the cliff top after the failure incidents revealed the development of fissures along the cliff edge. Monitoring 10 fissures over three months indicated that fissured by the cliff edge respond to intense seismicity (generally ≥ Mw 4) by widening. Redcliffs’ cliff collapsed on two separate occasions as a result of an accumulated amount of damage of the rock masses in the cliff (caused by weathering and erosion over time), and two Mw 6.2 trigger earthquakes which shook the Redcliffs and the surrounding area at a Peak Ground Acceleration (PGA) estimated to be around 2 g. The results of the theoretical study suggests that PGA levels felt on-site during both instances of failure are the result of three major factors: source of the quake and the site affected; topographic amplification of the ground movement; the short distance between the source and the cliff for both fault ruptures; the focus of seismic energy in the direction of thrust faulting along a path that intercepts Redcliffs (and the Port Hills). Ultimately, failure on the NE and SE Faces of Redcliffs’ cliff was concluded to be global as every part of the exposed cliff face deposited a significant volume of material on the talus at the base of the cliff, with the exception of one section on the NE Face. The cliff collapses was a concurrent process that is a single (non-monotonic) event that operated as a complex series of (primarily) toppling rock falls, some sliding of blocks, and slumping of the soil mantle on top of the cliff. The first collapse had a mixture of equivalent continua slope movement of the heavily weathered / damaged surface of the cliff face, and discontinuous slope movement of the jointed inner slope (behind the heavily weathered surface); whereas the second collapse resulted in only discontinuous slope movement on account of the freshly exposed cliff face that had damage to the rock masses, in the form of old and (relatively) new discontinuous fractures, induced by earthquakes and aftershocks leading up to the point of failure.
The purpose of this thesis is to evaluate the seismic response of the UC Physics Building based on recorded ground motions during the Canterbury earthquakes, and to use the recorded response to evaluate the efficacy of various conventional structural analysis modelling assumptions. The recorded instrument data is examined and analysed to determine how the UC Physics Building performed during the earthquake-induced ground motions. Ten of the largest earthquake events from the 2010-11 Canterbury earthquake sequence are selected in order to understand the seismic response under various levels of demand. Peak response amplitude values are found which characterise the demand from each event. Spectral analysis techniques are utilised to find the natural periods of the structure in each orthogonal direction. Significant torsional and rocking responses are also identified from the recorded ground motions. In addition, the observed building response is used to scrutinise the adequacy of NZ design code prescriptions for fundamental period, response spectra, floor acceleration and effective member stiffness. The efficacy of conventional numerical modelling assumptions for representing the UC Physics Building are examined using the observed building response. The numerical models comprise of the following: a one dimensional multi degree of freedom model, a two dimensional model along each axis of the building and a three dimensional model. Both moderate and strong ground motion records are used to examine the response and subsequently clarify the importance of linear and non-linear responses and the inclusion of base flexibility. The effects of soil-structure interaction are found to be significant in the transverse direction but not the longitudinal direction. Non-linear models predict minor in-elastic behaviour in both directions during the 4 September 2010 Mw 7.1 Darfield earthquake. The observed torsional response is found to be accurately captured by the three dimensional model by considering the interaction between the UC Physics Building and the adjacent structure. With the inclusion of adequate numerical modelling assumptions, the structural response is able to be predicted to within 10% for the majority of the earthquake events considered.
This thesis presents the application of data science techniques, especially machine learning, for the development of seismic damage and loss prediction models for residential buildings. Current post-earthquake building damage evaluation forms are developed for a particular country in mind. The lack of consistency hinders the comparison of building damage between different regions. A new paper form has been developed to address the need for a global universal methodology for post-earthquake building damage assessment. The form was successfully trialled in the street ‘La Morena’ in Mexico City following the 2017 Puebla earthquake. Aside from developing a framework for better input data for performance based earthquake engineering, this project also extended current techniques to derive insights from post-earthquake observations. Machine learning (ML) was applied to seismic damage data of residential buildings in Mexico City following the 2017 Puebla earthquake and in Christchurch following the 2010-2011 Canterbury earthquake sequence (CES). The experience showcased that it is readily possible to develop empirical data only driven models that can successfully identify key damage drivers and hidden underlying correlations without prior engineering knowledge. With adequate maintenance, such models have the potential to be rapidly and easily updated to allow improved damage and loss prediction accuracy and greater ability for models to be generalised. For ML models developed for the key events of the CES, the model trained using data from the 22 February 2011 event generalised the best for loss prediction. This is thought to be because of the large number of instances available for this event and the relatively limited class imbalance between the categories of the target attribute. For the CES, ML highlighted the importance of peak ground acceleration (PGA), building age, building size, liquefaction occurrence, and soil conditions as main factors which affected the losses in residential buildings in Christchurch. ML also highlighted the influence of liquefaction on the buildings losses related to the 22 February 2011 event. Further to the ML model development, the application of post-hoc methodologies was shown to be an effective way to derive insights for ML algorithms that are not intrinsically interpretable. Overall, these provide a basis for the development of ‘greybox’ ML models.
Liquefaction-induced lateral spreading in large seismic events often results in pervasive and costly damage to engineering structures and lifelines, making it a critical component of engineering design. However, the complex nature of this phenomenon leads to designing for such a hazard extremely challenging and there is a clear for an improved understanding and predicting liquefaction-induced lateral spreading. The 2010-2011 Canterbury (New Zealand) Earthquakes triggered severe liquefaction-induced lateral spreading along the streams and rivers of the Christchurch region, causing extensive damage to roads, bridges, lifelines, and structures in the vicinity. The unfortunate devastation induced from lateral spreading in these events also rendered the rare opportunity to gain an improved understanding of lateral spreading displacements specific to the Christchurch region. As part of this thesis, the method of ground surveying was employed following the 4 September 2010 Darfield (Mw 7.1) and 22 February 2011 Christchurch (Mw 6.2) earthquakes at 126 locations (19 repeated) throughout Christchurch and surrounding suburbs. The method involved measurements and then summation of crack widths along a specific alignment (transect) running approximately perpendicular to the waterway to indicate typically a maximum lateral displacement at the bank and reduction of the magnitude of displacements with distance from the river. Rigorous data processing and comparisons with alternative measurements of lateral spreading were performed to verify results from field observations and validate the method of ground surveying employed, as well as highlight the complex nature of lateral spreading displacements. The welldocumented field data was scrutinized to gain an understanding of typical magnitudes and distribution patterns (distribution of displacement with distance) of lateral spreading observed in the Christchurch area. Maximum displacements ranging from less than 10 cm to over 3.5 m were encountered at the sites surveyed and the area affected by spreading ranged from less than 20 m to over 200 m from the river. Despite the highly non-uniform displacements, four characteristic distribution patterns including large, distributed ground displacements, block-type movements, large and localized ground displacements, and areas of little to no displacements were identified. Available geotechnical, seismic, and topographic data were collated at the ground surveying sites for subsequent analysis of field measurements. Two widely-used empirical models (Zhang et al. (2004), Youd et al. (2002)) were scrutinized and applied to locations in the vicinity of field measurements for comparison with model predictions. The results indicated generally poor correlation (outside a factor of two) with empirical predictions at most locations and further validated the need for an improved, analysis- based method of predicting lateral displacements that considers the many factors involved on a site-specific basis. In addition, the development of appropriate model input parameters for the Youd et al. (2002) model led to a site-specific correlation of soil behavior type index, Ic, and fines content, FC, for sites along the Avon River in Christchurch that matched up well with existing Ic – FC relationships commonly used in current practice. Lastly, a rigorous analysis was performed for 25 selected locations of ground surveying measurements along the Avon River where ground slope conditions are mild (-1 to 2%) and channel heights range from about 2 – 4.5 m. The field data was divided into categories based on the observed distribution pattern of ground displacements including: large and distributed, moderate and distributed, small to negligible, and large and localized. A systematic approach was applied to determine potential critical layers contributing to the observed displacement patterns which led to the development of characteristic profiles for each category considered. The results of these analyses outline an alternative approach to the evaluation of lateral spreading in which a detailed geotechnical analysis is used to identify the potential for large spreading displacements and likely spatial distribution patterns of spreading. Key factors affecting the observed magnitude and distribution of spreading included the thickness of the critical layer, relative density, soil type and layer continuity. It was found that the large and distributed ground displacements were associated with a thick (1.5 – 2.5 m) deposit of loose, fine to silty sand (qc1 ~4-7 MPa, Ic 1.9-2.1, qc1n_cs ~50-70) that was continuous along the bank and with distance from the river. In contrast, small to negligible displacements were characterized by an absence of or relatively thin (< 1 m), discontinuous critical layer. Characteristic features of the moderate and distributed displacements were found to be somewhere between these two extremes. The localized and large displacements showed a characteristic critical layer similar to that observed in the large and distributed sites but that was not continuous and hence leading to the localized zone of displacement. The findings presented in this thesis illustrate the highly complex nature of lateral displacements that cannot be captured in simplified models but require a robust geotechnical analysis similar to that performed for this research.
The Mw 6.2 February 22nd 2011 Christchurch earthquake (and others in the 2010-2011 Canterbury sequence) provided a unique opportunity to study the devastating effects of earthquakes first-hand and learn from them for future engineering applications. All major events in the Canterbury earthquake sequence caused widespread liquefaction throughout Christchurch’s eastern suburbs, particularly extensive and severe during the February 22nd event. Along large stretches of the Avon River banks (and to a lesser extent along the Heathcote) significant lateral spreading occurred, affecting bridges and the infrastructure they support. The first stage of this research involved conducting detailed field reconnaissance to document liquefaction and lateral spreading-induced damage to several case study bridges along the Avon River. The case study bridges cover a range of ages and construction types but all are reinforced concrete structures which have relatively short, stiff decks. These factors combined led to a characteristic deformation mechanism involving deck-pinning and abutment back-rotation with consequent damage to the abutment piles and slumping of the approaches. The second stage of the research involved using pseudo-static analysis, a simplified seismic modelling tool, to analyse two of the bridges. An advantage of pseudo-static analysis over more complicated modelling methods is that it uses conventional geotechnical data in its inputs, such as SPT blowcount and CPT cone resistance and local friction. Pseudo-static analysis can also be applied without excessive computational power or specialised knowledge, yet it has been shown to capture the basic mechanisms of pile behaviour. Single pile and whole bridge models were constructed for each bridge, and both cyclic and lateral spreading phases of loading were investigated. Parametric studies were carried out which varied the values of key parameters to identify their influence on pile response, and computed displacements and damages were compared with observations made in the field. It was shown that pseudo-static analysis was able to capture the characteristic damage mechanisms observed in the field, however the treatment of key parameters affecting pile response is of primary importance. Recommendations were made concerning the treatment of these governing parameters controlling pile response. In this way the future application of pseudo-static analysis as a tool for analysing and designing bridge pile foundations in liquefying and laterally spreading soils is enhanced.
The Acheron rock avalanche is located in the Red Hill valley almost 80 km west of Christchurch and is one of 42 greywacke-derived rock avalanches identified in the central Southern Alps. It overlies the Holocene active Porters Pass Fault; a component of the Porters Pass-Amberley Fault Zone which extends from the Rakaia River to beyond the Waimakariri River. The Porters Pass Fault is a dextral strike-slip fault system viewed as a series of discontinuous fault scarps. The location of the fault trace beneath the deposit suggests it may represent a possible source of seismic shaking resulting in the formation of the Acheron rock avalanche. The rock mass composition of the rock avalanche source scar is Torlesse Supergroup greywacke consisting of massive sandstone and thinly bedded mudstone sequences dipping steeply north into the centre of the source basin. A stability analysis identified potential instability along shallow north dipping planar defects, and steep south dipping toppling failure planes. The interaction of the defects with bedding is considered to have formed conditions for potential instability most likely triggered by a seismic event. The dTositional area of the rock avalanche covers 7.2 x 105 m2 with an estimated volume of 9 x 10 m3 The mobilised rock mass volume was calculated at 7.5 x 106 m3• Run out of the debris from the top of the source scar to the distal limit reached 3500m, descending over a vertical fall of almost 700m with an estimated Fahrboschung of 0.2. The run out of the rock avalanche displayed moderate to high mobility, travelling at an estimated maximum velocity of 140-160 km/hour. The rapid emplacement of the deposit is confirmed by highly fragmented internal composition and burial of forest vegetation New radiocarbon ages from buried wood retrieved from the base of Acheron rock avalanche deposit represents an emplacement age closely post-dating (Wk 12094) 1152 ± 51 years B.P. This differs significantly from a previous radiocarbon age of (NZ547) 500 ± 69 years B.P. and modal lichenometry and weathering-rind thickness ages of approximately 460 ± 10 yrs and 490 ± 50 years B.P. The new age shows no resemblance to an earthquake event around 700- 500 years B.P. on the Porters Pass-Amberley Fault Zone. The DAN run out simulation using a friction model rheology successfully replicated the long run out and velocity of the Acheron rock avalanche using a frictron angle of 27° and high earth pressure coefficients of 5.5, 5.2, and 5.9. The elevated earth pressure coefficients represent dispersive pressures derived from dynamic fragmentation of the debris within the mobile rock avalanche, supporting the hypothesis of Davies and McSaveney (2002). The DAN model has potential applications for areas prone to large-scale instability in the elevated slopes and steep waterways of the Southern Alps. A paleoseismic investigation of a newly identified scarp of the Porters Pass Fault partially buried by the rock avalanche was conducted to identify any evidence of a coseismic relationship to the Acheron rock avalanche. This identified three-four fault traces striking at 078°, and a sag pond displaying a sequence of overbank deposits containing two buried soils representing an earthquake event horizon. A 40cm vertical offset of the ponded sediment and lower buried soil horizqn was recorded, which was dated to (Wk 13112 charcoal in palosol) 653 ± 54 years B.P. and (Wk 13034 palosol) 661 ± 34 years B.P. The evidence indicates a fault rupture occurred along the Porters Pass Fault, west of Porters Pass most likely extending to the Red Lakes terraces, post-dating 700 years B.P., resulting in 40cm of vertical displacement and an unknown component of dextral strike slip movement. This event post dates the event one (1000 ± 100 years B.P) at Porters Pass previously considered to represent the most recent rupture along the fault line. This points to a probable source for resetting of the modal weathering-rind thicknesses and lichen size populations in the Red Hill valley and possibly the Red Lakes terraces. These results suggest careful consideration must be given to the geomorphic and paleoseismic history of a specific site when applying surface dating techniques and furthermore the origin of dates used in literature and their useful range should be verified. An event at 700-500 years B.P did not trigger the Acheron rock avalanche as previously assumed supporting Howard's conclusions. The lack of similar aged rupture evidence in either of the Porters Pass and Coleridge trenches supports Howard's hypothesis of segmentation of the Porters Pass Fault; where rupture occurs along one fault segment but not along another. The new rock avalanche age closely post-dating 1200-1100 years B.P. resembles the poorly constrained event one rupture age of 1700-800 years B.P for the Porters Pass Fault and the tighter constrained Round Top event of 1010 ± 50 years B.P. on the Alpine Fault. Eight other rock avalanche deposits spread across the central Southern Alps also resemble the new ages however are unable to be assigned specific earthquake events due to the large associated error bars of± 270 years. This clustering of ages does represent compelling lines of evidence for large magnitude earthquake events occurring over the central Southern Alps. The presence of a rock avalanche deposit does not signify an earthquake based on the historical evidence in the Southern Alps however clustering of ages does suggest that large Mw >7 earthquakes occurred across the Southern Alps between 1200-900 years BP.
A number of reverse and strike-slip faults are distributed throughout mid-Canterbury, South Island, New Zealand, due to oblique continental collision. There is limited knowledge on fault interaction in the region, despite historical multi-fault earthquakes involving both reverse and strike-slip faults. The surface expression and paleoseismicity of these faults can provide insights into fault interaction and seismic hazards in the region. In this thesis, I studied the Lake Heron and Torlesse faults to better understand the key differences between these two adjacent faults located within different ‘tectonic domains’. Recent activity and surface expression of the Lake Heron fault was mapped and analysed using drone survey, Structure-from-Motion (SfM) derived Digital Surface Model (DSM), aerial image, 5 m-Digital Elevation Model (DEM), luminescence dating technique, and fold modelling. The results show a direct relationship between deformation zone width and the thickness of the gravel deposits in the area. Fold modelling using fault dip, net slip and gravel thickness produces a deformation zone comparable to the field, indicating that the fault geometry is sound and corroborating the results. This result Is consistent with global studies that demonstrate deposit (or soil thickness) correlates to fault deformation zone width, and therefore is important to consider for fault displacement hazard. A geomorphological study on the Torlesse fault was conducted using SfM-DSM, DEM and aerial images Ground Penetrating Radar (GPR) survey, trenching, and radiocarbon and luminescence dating. The results indicate that the Torlesse fault is primarily strike-slip with some dip slip component. In many places, the bedding-parallel Torlesse fault offsets post-glacial deposits, with some evidence of flexural slip faulting due to folding. Absolute dating of offset terraces using radiocarbon dating and slip on fault determined from lateral displacement calculating tool demonstrates the fault has a slip rate of around 0.5 mm/year to 1.0 mm/year. The likelihood of multi-fault rupture in the Torlesse Range has been characterised using paleoseismic trenching, a new structural model, and evaluation of existing paleoseismic data on the Porters Pass fault. Identification of overlapping of paleoseismic events in main Torlesse fault, flexural-slip faults and the Porters Pass fault in the Torlesse Range shows the possibility of distinct or multi-fault rupture on the Torlesse fault. The structural connectivity of the faults in the Torlesse zone forming a ‘flower structure’ supports the potential of multi-fault rupture. Multi-fault rupture modelling carried out in the area shows a high probability of rupture in the Porters Pass fault and Esk fault which also supports the co-rupture probability of faults in the region. This study offers a new understanding of the chronology, slip distribution, rupture characteristics and possible structural and kinematic relationship of Lake Heron fault and Torlesse fault in the South Island, New Zealand.
Documenting earthquake-induced ground deformation is significant to assess the characteristics of past and contemporary earthquakes and provide insight into seismic hazard. This study uses airborne light detection and ranging (LiDAR) and conducts multi-disciplinary field techniques to document the surface rupture morphology and evaluate the paleoseismicity and seismic hazard parameters of the Hurunui segment of the Hope Fault in the northern South Island of New Zealand. It also documents and evaluates seismically induced features and ground motion characteristics of the 2010 Darfield and 2011 Christchurch earthquakes in the Port Hills, south of Christchurch. These two studies are linked in that they investigate the near-field coseismic features of large (Mw ~7.1) earthquakes in New Zealand and produce data for evaluating seismic hazards of future earthquakes. In the northern South Island of New Zealand, the Australian-Pacific plate boundary is characterised by strike-slip deformation across the Marlborough Fault System (MFS). The ENE-striking Hope Fault (length: ~230 km) is the youngest and southernmost fault in the MFS, and the second fastest slipping fault in New Zealand. The Hope Fault is a major source of seismic hazard in New Zealand and has ruptured (in-part) historically in the Mw 7.1 1888 Amuri earthquake. In the west, the Hurunui segment of the Hope Fault is covered by beech forest. Hence, its seismic hazard parameters and paleoearthquake chronology were poorly constrained and it was unknown whether the 1888 earthquake ruptured this segment or not and if so, to what extent. Utilising LiDAR and field data, a 29 km-long section of the Hurunui segment of the Hope Fault is mapped. LiDAR-mapping clearly reveals the principal slip zone (PSZ) of the fault and a suite of previously unrecognised structures that form the fault deformation zone (FDZ). FDZ width measurements from 415 locations reveal a spatially-variable, active FDZ up to ~500 m wide with an average width of 200 m. Kinematic analysis of the fault structures shows that the Hurunui segment strikes between 070° and 075° and is optimally oriented for dextral strike-slip within the regional stress field. This implies that the wide FDZ observed is unlikely to result from large-scale fault mis-orientation with respect to regional stresses. The analysis of FDZ width indicates that it increases with increased hanging wall topography and increased topographic relief suggesting that along-strike topographic perturbations to fault geometry and stress states increase fault zone complexity and width. FDZ width also increases where the tips of adjacent PSZ strands locally vary in strike, and where the thickness of alluvial deposits overlying bedrock increases. LiDAR- and photogrammetrically-derived topographic mapping indicates that the boundary between the Hurunui and Hope River segments is characterised by a ~850-m-wide right stepover and a 9º-14° fault bend. Paleoseismic trenching at Hope Shelter site reveals that 6 earthquakes occurred at A.D. 1888, 1740-1840, 1479-1623, 819-1092, 439-551, and 373- 419. These rupture events have a mean recurrence interval of ~298 ± 88 yr and inter-event times ranging from 98 to 595 yrs. The variation in the inter-event times is explained by (1) coalescing rupture overlap from the adjacent Hope River segment on to the Hurunui segment at the study site, (2) temporal clustering of large earthquakes on the Hurunui segment, and/or (3) ‘missing’ rupture events. It appears that the first two options are more plausible to explain the earthquake chronologies and rupture behaviour on the Hurunui segment, given the detailed nature of the geologic and chronologic investigations. This study provides first evidence for coseismic multi-segment ruptures on the Hope Fault by identifying a rupture length of 44-70 km for the 1888 earthquake, which was not confined to the Hope River segment (primary source for the 1888 earthquake). LiDAR data is also used to identify and measure dextral displacements and scarp heights from the PSZ and structures within the FDZ along the Hurunui segment. Reconstruction of large dextrally-offset geomorphic features shows that the vertical component of slip accounts for only ~1% of the horizontal displacements and confirms that the fault is predominantly strike-slip. A strong correlation exists between the dextral displacements and elevations of geomorphic features suggesting the possibility of age correlation between the geomorphic features. A mean single event displacement (SED) of 3.6 ± 0.7 m is determined from interpretation of sets of dextral displacements of ≤ 25 m. Using the available surface age data and the cumulative dextral displacements from Matagouri Flat, McKenzie Fan, Macs Knob and Hope River sites, and the mean SED, a mean slip rate of 12.2 ± 2.4 mm/yr, and a mean recurrence interval of ~320 ± 120 yr, and a potential earthquake magnitude of Mw 7.2 are determined for the Hurunui segment. This study suggests that the fault slip rate has been constant over the last ~15000 yr. Strong ground motions from the 2010 Darfield (Canterbury) earthquake displaced boulders and caused ground damage on some ridge crests in the Port Hills. However, the 2011 Christchurch earthquake neither displaced boulders nor caused ground damage at the same ridge crests. Documentation of locations (~400 m a.s.l.), lateral displacements (8-970 cm), displacement direction (250° ± 20°) of displaced boulders, in addition to their hosting socket geometries (< 1 cm to 50 cm depth), the orientation of the ridges (000°-015°) indicate that boulders have been displaced in the direction of instrumentally recorded transient peak ground horizontal displacements nearby and that the seismic waves have been amplified at the study sites. The co-existence of displaced and non-displaced boulders at proximal sites suggests small-scale ground motion variability and/or varying boulder-ground dynamic interactions relating to shallow phenomena such as variability in soil depth, bedrock fracture density and/or microtopography on the bedrock-soil interface. Shorter shaking duration of the 2011 Christchurch event, differing frequency contents and different source characteristics were all factors that may have contributed to generating circumstances less favourable to boulder displacement in this earthquake. Investigating seismically induced features, fault behaviour, site effects on the rupture behaviour, and site response to the seismic waves provides insights into fault rupture hazards.
In the period between September 2010 and December 2011, Christchurch (New Zealand) and its surroundings were hit by a series of strong earthquakes including six significant events, all generated by local faults in proximity to the city: 4 September 2010 (Mw=7.1), 22 February 2011 (Mw=6.2), 13 June 2011 (Mw=5.3 and Mw=6.0) and 23 December 2011 (M=5.8 and (M=5.9) earthquakes. As shown in Figure 1, the causative faults of the earthquakes were very close to or within the city boundaries thus generating very strong ground motions and causing tremendous damage throughout the city. Christchurch is shown as a lighter colour area, and its Central Business District (CBD) is marked with a white square area in the figure. Note that the sequence of earthquakes started to the west of the city and then propagated to the south, south-east and east of the city through a set of separate but apparently interacting faults. Because of their strength and proximity to the city, the earthquakes caused tremendous physical damage and impacts on the people, natural and built environments of Christchurch. The 22 February 2011 earthquake was particularly devastating. The ground motions generated by this earthquake were intense and in many parts of Christchurch substantially above the ground motions used to design the buildings in Christchurch. The earthquake caused 182 fatalities, collapse of two multi-storey reinforced concrete buildings, collapse or partial collapse of many unreinforced masonry structures including the historic Christchurch Cathedral. The Central Business District (CBD) of Christchurch, which is the central heart of the city just east of Hagley Park, was practically lost with majority of its 3,000 buildings being damaged beyond repair. Widespread liquefaction in the suburbs of Christchurch, as well as rock falls and slope/cliff instabilities in the Port Hills affected tens of thousands of residential buildings and properties, and shattered the lifelines and infrastructure over approximately one third of the city area. The total economic loss caused by the 2010-2011 Christchurch earthquakes is currently estimated to be in the range between 25 and 30 billion NZ dollars (or 15% to 18% of New Zealand’s GDP). After each major earthquake, comprehensive field investigations and inspections were conducted to document the liquefaction-induced land damage, lateral spreading displacements and their impacts on buildings and infrastructure. In addition, the ground motions produced by the earthquakes were recorded by approximately 15 strong motion stations within (close to) the city boundaries providing and impressive wealth of data, records and observations of the performance of ground and various types of structures during this unusual sequence of strong local earthquakes affecting a city. This paper discusses the liquefaction in residential areas and focuses on its impacts on dwellings (residential houses) and potable water system in the Christchurch suburbs. The ground conditions of Christchurch including the depositional history of soils, their composition, age and groundwater regime are first discussed. Detailed liquefaction maps illustrating the extent and severity of liquefaction across Christchurch triggered by the sequence of earthquakes including multiple episodes of severe re-liquefaction are next presented. Characteristic liquefaction-induced damage to residential houses is then described focussing on the performance of typical house foundations in areas affected by liquefaction. Liquefaction impacts on the potable water system of Christchurch is also briefly summarized including correlation between the damage to the system, liquefaction severity, and the performance of different pipe materials. Finally, the characteristics of Christchurch liquefaction and its impacts on built environment are discussed in relation to the liquefaction-induced damage in Japan during the 11 March 2011 Great East Japan Earthquake.
Spatial variations in river facies exerted a strong influence on the distribution of liquefaction features observed in Christchurch during the 2010-11 Canterbury Earthquake Sequence (CES). Liquefaction and liquefaction-induced ground deformation was primarily concentrated near modern waterways and areas underlain by Holocene fluvial deposits with shallow water tables (< 1 to 2 m). In southern Christchurch, spatial variations of liquefaction and subsidence were documented in the suburbs within inner meander loops of the Heathcote River. Newly acquired geospatial data, geotechnical reports and eye-witness discussions are compiled to provide a detailed account of the surficial effects of CES liquefaction and ground deformation adjacent to the Heathcote River. LiDAR data and aerial photography are used to produce a new series of original figures which reveal the locations of recurrent liquefaction and subsidence. To investigate why variable liquefaction patterns occurred, the distribution of surface ejecta and associated ground damage is compared with near-surface sedimentologic, topographic, and geomorphic variability to seek relationships between the near-surface properties and observed ground damages. The most severe liquefaction was concentrated within a topographic low in the suburb of St Martins, an inner meander loop of the Heathcote River, with liquefaction only minor or absent in the surrounding areas. Subsurface investigations at two sites in St Martins enable documentation of fluvial stratigraphy, the expressions of liquefaction, and identification of pre-CES liquefaction features. Excavation to water table depths (~1.5 m below the surface) across sand boils reveals multiple generations of CES liquefaction dikes and sills that cross-cut Holocene fluvial and anthropogenic stratigraphy. Based on in situ geotechnical tests (CPT) indicating sediment with a factor of safety < 1, the majority of surface ejecta was sourced from well-sorted fine to medium sand at < 5 m depth, with the most damaging liquefaction corresponding with the location of a low-lying sandy paleochannel, a remnant river channel from the Holocene migration of the meander in St Martins. In the adjacent suburb of Beckenham, where migration of the Heathcote River has been laterally confined by topography associated with the volcanic lithologies of Banks Peninsula, severe liquefaction was absent with only minor sand boils occurring closest to the modern river channel. Auger sampling across the suburb revealed thick (>1 m) clay-rich overbank and back swamp sediments that produced a stratigraphy which likely confined the units susceptible to liquefaction and prevented widespread ejection of liquefied material. This analysis suggests river migration promotes the formation and preservation of fluvial deposits prone to liquefaction. Trenching revealed the strongest CES earthquakes with large vertical accelerations favoured sill formation and severe subsidence at highly susceptible locations corresponding with an abandoned channel. Less vulnerable sites containing deeper and thinner sand bodies only liquefied in the strongest and most proximal earthquakes forming minor localised liquefaction features. Liquefaction was less prominent and severe subsidence was absent where lateral confinement of a Heathcote meander has promoted the formation of fluvial stratum resistant to liquefaction. Correlating CES liquefaction with geomorphic interpretations of Christchurch’s Heathcote River highlights methods in which the performance of liquefaction susceptibility models can be improved. These include developing a reliable proxy for estimating soil conditions in meandering fluvial systems by interpreting the geology and geomorphology, derived from LiDAR data and modern river morphology, to improve the methods of accounting for the susceptibility of an area. Combining geomorphic interpretations with geotechnical data can be applied elsewhere to identify regional liquefaction susceptibilities, improve existing liquefaction susceptibility datasets, and predict future earthquake damage.
The purpose of this thesis is to conduct a detailed examination of the forward-directivity characteristics of near-fault ground motions produced in the 2010-11 Canterbury earthquakes, including evaluating the efficacy of several existing empirical models which form the basis of frameworks for considering directivity in seismic hazard assessment. A wavelet-based pulse classification algorithm developed by Baker (2007) is firstly used to identify and characterise ground motions which demonstrate evidence of forward-directivity effects from significant events in the Canterbury earthquake sequence. The algorithm fails to classify a large number of ground motions which clearly exhibit an early-arriving directivity pulse due to: (i) incorrect pulse extraction resulting from the presence of pulse-like features caused by other physical phenomena; and (ii) inadequacy of the pulse indicator score used to carry out binary pulse-like/non-pulse-like classification. An alternative ‘manual’ approach is proposed to ensure 'correct' pulse extraction and the classification process is also guided by examination of the horizontal velocity trajectory plots and source-to-site geometry. Based on the above analysis, 59 pulse-like ground motions are identified from the Canterbury earthquakes , which in the author's opinion, are caused by forward-directivity effects. The pulses are also characterised in terms of their period and amplitude. A revised version of the B07 algorithm developed by Shahi (2013) is also subsequently utilised but without observing any notable improvement in the pulse classification results. A series of three chapters are dedicated to assess the predictive capabilities of empirical models to predict the: (i) probability of pulse occurrence; (ii) response spectrum amplification caused by the directivity pulse; (iii) period and amplitude (peak ground velocity, PGV) of the directivity pulse using observations from four significant events in the Canterbury earthquakes. Based on the results of logistic regression analysis, it is found that the pulse probability model of Shahi (2013) provides the most improved predictions in comparison to its predecessors. Pulse probability contour maps are developed to scrutinise observations of pulses/non-pulses with predicted probabilities. A direct comparison of the observed and predicted directivity amplification of acceleration response spectra reveals the inadequacy of broadband directivity models, which form the basis of the near-fault factor in the New Zealand loadings standard, NZS1170.5:2004. In contrast, a recently developed narrowband model by Shahi & Baker (2011) provides significantly improved predictions by amplifying the response spectra within a small range of periods. The significant positive bias demonstrated by the residuals associated with all models at longer vibration periods (in the Mw7.1 Darfield and Mw6.2 Christchurch earthquakes) is likely due to the influence of basin-induced surface waves and non-linear soil response. Empirical models for the pulse period notably under-predict observations from the Darfield and Christchurch earthquakes, inferred as being a result of both the effect of nonlinear site response and influence of the Canterbury basin. In contrast, observed pulse periods from the smaller magnitude June (Mw6.0) and December (Mw5.9) 2011 earthquakes are in good agreement with predictions. Models for the pulse amplitude generally provide accurate estimates of the observations at source-to-site distances between 1 km and 10 km. At longer distances, observed PGVs are significantly under-predicted due to their slower apparent attenuation. Mixed-effects regression is employed to develop revised models for both parameters using the latest NGA-West2 pulse-like ground motion database. A pulse period relationship which accounts for the effect of faulting mechanism using rake angle as a continuous predictor variable is developed. The use of a larger database in model development, however does not result in improved predictions of pulse period for the Darfield and Christchurch earthquakes. In contrast, the revised model for PGV provides a more appropriate attenuation of the pulse amplitude with distance, and does not exhibit the bias associated with previous models. Finally, the effects of near-fault directivity are explicitly included in NZ-specific probabilistic seismic hazard analysis (PSHA) using the narrowband directivity model of Shahi & Baker (2011). Seismic hazard analyses are conducted with and without considering directivity for typical sites in Christchurch and Otira. The inadequacy of the near-fault factor in the NZS1170.5: 2004 is apparent based on a comparison with the directivity amplification obtained from PSHA.
Geologic phenomena produced by earthquake shaking, including rockfalls and liquefaction features, provide important information on the intensity and spatiotemporal distribution of earthquake ground motions. The study of rockfall and liquefaction features produced in contemporary well- instrumented earthquakes increases our knowledge of how natural and anthropogenic environments respond to earthquakes and improves our ability to deduce seismologic information from analogous pre-contemporary (paleo-) geologic features. The study of contemporary and paleo- rockfall and liquefaction features enables improved forecasting of environmental responses to future earthquakes. In this thesis I utilize a combination of field and imagery-based mapping, trenching, stratigraphy, and numerical dating techniques to understand the nature and timing of rockfalls (and hillslope sedimentation) and liquefaction in the eastern South Island of New Zealand, and to examine the influence that anthropogenic activity has had on the geologic expressions of earthquake phenomena. At Rapaki (Banks Peninsula, NZ), field and imagery-based mapping, statistical analysis and numerical modeling was conducted on rockfall boulders triggered by the fatal 2011 Christchurch earthquakes (n=285) and compared with newly identified prehistoric (Holocene and Pleistocene) boulders (n=1049) deposited on the same hillslope. A significant population of modern boulders (n=26) travelled farther downslope (>150 m) than their most-travelled prehistoric counterparts, causing extensive damage to residential dwellings at the foot of the hillslope. Replication of prehistoric boulder distributions using 3-dimensional rigid body numerical models requires the application of a drag-coefficient, attributed to moderate to dense slope vegetation, to account for their spatial distribution. Radiocarbon dating provides evidence for 17th to early 20th century deforestation at the study site during Polynesian and European colonization and after emplacement of prehistoric rockfalls. Anthropocene deforestation enabled modern rockfalls to exceed the limits of their prehistoric predecessors, highlighting a shift in the geologic expression of rockfalls due to anthropogenic activity. Optical and radiocarbon dating of loessic hillslope sediments in New Zealand’s South Island is used to constrain the timing of prehistoric rockfalls and associated seismic events, and quantify spatial and temporal patterns of hillslope sedimentation including responses to seismic and anthropogenic forcing. Luminescence ages from loessic sediments constrain timing of boulder emplacement to between ~3.0 and ~12.5 ka, well before the arrival of Polynesians (ca AD 1280) and Europeans (ca AD 1800) in New Zealand, and suggest loess accumulation was continuing at the study site until 12-13 ka. Large (>5 m3) prehistoric rockfall boulders preserve an important record of Holocene hillslope sedimentation by creating local traps for sediment aggradation and upbuilding soil formation. Sediment accumulation rates increased considerably (>~10 factor increase) following human arrival and associated anthropogenic burning of hillslope vegetation. New numerical ages are presented to place the evolution of loess-mantled hillslopes in New Zealand’s South Island into a longer temporal framework and highlight the roles of earthquakes and humans on hillslope surface process. Extensive field mapping and characterization for 1733 individual prehistoric rockfall boulders was conducted at Rapaki and another Banks Peninsula site, Purau, to understand their origin, frequency, and spatial and volumetric distributions. Boulder characteristics and distributions were compared to 421 boulders deposited at the same sites during the 2010-2011 Canterbury earthquake sequence. Prehistoric boulders at Rapaki and Purau are comprised of two dominant lithofacies types: volcanic breccia and massive (coherent) lava basalt. Volcanic breccia boulders are found in greatest abundance (64-73% of total mapped rockfall) and volume (~90-96% of total rockfall) at both locations and exclusively comprise the largest boulders with the longest runout distances that pose the greatest hazard to life and property. This study highlights the primary influence that volcanic lithofacies architecture has on rockfall hazard. The influence of anthropogenic modifications on the surface and subsurface geologic expression of contemporary liquefaction created during the 2010-2011 Canterbury earthquake sequence (CES) in eastern Christchurch is examined. Trench observations indicate that anthropogenic fill layer boundaries and the composition/texture of discretely placed fill layers play an important role in absorbing fluidized sand/silt and controlling the subsurface architecture of preserved liquefaction features. Surface liquefaction morphologies (i.e. sand blows and linear sand blow arrays) display alignment with existing utility lines and utility excavations (and perforated pipes) provided conduits for liquefaction ejecta during the CES. No evidence of pre-CES liquefaction was identified within the anthropogenic fill layers or underlying native sediment. Radiocarbon dating of charcoal within the youngest native sediment suggests liquefaction has not occurred at the study site for at least the past 750-800 years. The importance of systematically examining the impact of buried infrastructure on channelizing and influencing surface and subsurface liquefaction morphologies is demonstrated. This thesis highlights the importance of using a multi-technique approach for understanding prehistoric and contemporary earthquake phenomena and emphasizes the critical role that humans play in shaping the geologic record and Earth’s surface processes.
