The rapid classification of building damage states or placards after an earthquake is vital for enabling an efficient emergency response and informed decision-making for rehabilitation and recovery purposes. Traditional methods rely heavily on inspector-led on-site surveys, which are often time-consuming, resource-intensive, and susceptible to human error. This study introduces a machine learning-supported surrogate model designed to streamline the assessment of building damage, focusing on the automated assignment of damage placards within the context of New Zealand's post-earthquake evaluation frameworks. The study evaluates two key safety evaluation protocols—Rapid Building Assessment (RBA) and Detailed Damage Evaluation (DDE)—and integrates corresponding databases derived from the 2010–2011 Canterbury Earthquake Sequence (CES) in Christchurch. Six ML classifiers—Multilayer Perceptron (MLP), Random Forest (RF), Support Vector Machine (SVM), K-Nearest Neighbours (KNN), Gradient Boosting Classifier (GBC), and Gradient Bagging (GBag)—were rigorously tested across both databases. The results indicate that the RF-based surrogate model outperforms the other classifiers across both RBA and DDE protocols. Two distinct sets of critical predictors have been further identified for each protocol, allowing for the rapid retrieval of essential data for future on-site surveys, while retaining the RF model's predictive accuracy. The developed surrogate model provides a pragmatic tool for practising engineers to rapidly assign placards to damaged structures and for policymakers and building owners to make informed recovery decisions for earthquake-affected buildings
Quick and reliable assessment of the condition of bridges in a transportation network after an earthquake can greatly assist immediate post-disaster response and long-term recovery. However, experience shows that available resources, such as qualified inspectors and engineers, will typically be stretched for such tasks. Structural health monitoring (SHM) systems can therefore make a real difference in this context. SHM, however, needs to be deployed in a strategic manner and integrated into the overall disaster response plans and actions to maximize its benefits. This study presents, in its first part, a framework of how this can be achieved. Since it will not be feasible, or indeed necessary, to use SHM on every bridge, it is necessary to prioritize bridges within individual networks for SHM deployment. A methodology for such prioritization based on structural and geotechnical seismic risks affecting bridges and their importance within a network is proposed in the second part. An example using the methodology application to selected bridges in the medium-sized transportation network of Wellington, New Zealand is provided. The third part of the paper is concerned with using monitoring data for quick assessment of bridge condition and damage after an earthquake. Depending on the bridge risk profile, it is envisaged that data will be obtained from either local or national seismic monitoring arrays or SHM systems installed on bridges. A method using artificial neural networks is proposed for using data from a seismic array to infer key ground motion parameters at an arbitrary bridges site. The methodology is applied to seismic data collected in Christchurch, New Zealand. Finally, how such ground motion parameters can be used in bridge damage and condition assessment is outlined AM - Accepted manuscript
Territorial authorities in New Zealand are responding to regulatory and market forces in the wake of the 2011 Christchurch earthquake to assess and retrofit buildings determined to be particularly vulnerable to earthquakes. Pending legislation may shorten the permissible timeframes on such seismic improvement programmes, but Auckland Council’s Property Department is already engaging in a proactive effort to assess its portfolio of approximately 3500 buildings, prioritise these assets for retrofit, and forecast construction costs for improvements. Within the programme structure, the following varied and often competing factors must be accommodated: * The council’s legal, fiscal, and ethical obligations to the people of Auckland per building regulations, health and safety protocols, and economic growth and urban development planning strategies; * The council’s functional priorities for service delivery; * Varied and numerous stakeholders across the largest territorial region in New Zealand in both population and landmass; * Heritage preservation and community and cultural values; and * Auckland’s prominent economic role in New Zealand’s economy which requires Auckland’s continued economic production post-disaster. Identifying those buildings most at risk to an earthquake in such a large and varied portfolio has warranted a rapid field assessment programme supplemented by strategically chosen detailed assessments. Furthermore, Auckland Council will benefit greatly in time and resources by choosing retrofit solutions, techniques, and technologies applicable to a large number of buildings with similar configurations and materials. From a research perspective, the number and variety of buildings within the council’s property portfolio will provide valuable data for risk modellers on building typologies in Auckland, which are expected to be fairly representative of the New Zealand building stock as a whole
The devastating consequences of past events, such as the 2004 Indian Ocean and 2011 Tōhoku tsunamis, emphasise the need for continued improvement in resilience measures. Given that 80% of magnitude 8+ earthquakes occur on the Pacific Rim, New Zealand's tsunami risk is significant. This research develops a novel tsunami inundation model. The proposed model applies equations based on hydraulic principles, including energy conservation (friction loss). While it does not fully replicate hydrodynamic models, it maintains a two-dimensional approach and offers significant improvements over currently implemented simplified methods. It retains excellent computational efficiency (seconds to minutes) while achieving a significant increase in accuracy that is comparable to traditional hydrodynamic models, which typically take hours to days. Calibration of the roughness input variables to hydrodynamic modelling at Gisborne and Christchurch, New Zealand, optimised the model to achieve similarity index values of above 84% for inundation extent, while 77% of inundation depths were within ±1 m and over 93% within ±2 m. This research then produces the first nationally consistent tsunami exposure assessment for New Zealand using a physics-based modelling method. Using probabilistic shoreline wave amplitude data, the study generates high-resolution (10 m) inundation maps for seven return periods (50th and 84th percentiles). These maps are integrated with land cover and infrastructure data to quantify exposure and identify the most vulnerable locations. The results highlight exposure not only to the commonly studied cities but also to several provincial areas. The identification of exposure is the foremost step towards practical resilience efforts; however, understanding specific infrastructure impacts ensures that countermeasures and risk reduction practices are implemented. Therefore, a detailed evaluation of the NZTA Bridge Manual is conducted. Comparisons are made between the NZTA methodology and the rapid model developed in this research. The results reveal a significant overestimation of bridge and culvert exposure by NZTA methods. The study further highlights critical exposure locations for bridge and culvert assets. Flow depths calculated at bridge locations are significantly overestimated using the NZTA method compared to results derived from hydrodynamic modelling and the rapid model. This research then conducts component-level modelling of culvert assets, due to their identified vulnerability in the transportation network. At a 1:15 geometrical scale, laboratory experiments evaluated the response of different culvert set-ups to tsunami bores. The findings provide a detailed description into overtopping, flow regimes and pressure distributions and give laboratory experiments as validation studies for future numerical modelling and design improvements. Overall, this research performs a multi-modal tsunami inundation assessment, uniting macro-level exposure modelling with micro-level component responses by integrating modelling, exposure analysis, and experimental validation. The findings support refining current tsunami guidelines, improving infrastructure planning, and enhancing community preparedness. Overall, the study’s multi-model approach strengthens many elements of New Zealand’s ability to mitigate and respond to future tsunami events
The 2010 Darfield earthquake is the largest earthquake on record to have occurred within 40 km of a major city and not cause any fatalities. In this paper the authors have reflected on their experiences in Christchurch following the earthquake with a view to what worked, what didn’t, and what lessons can be learned from this for the benefit of Australian earthquake preparedness. Owing to the fact that most of the observed building damage occurred in Unreinforced Masonry (URM) construction, this paper focuses in particular on the authors’ experience conducting rapid building damage assessment during the first 72 hours following the earthquake and more detailed examination of the performance of unreinforced masonry buildings with and without seismic retrofit interventions
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
