Understanding flood protection stopbanks across New Zealand

 

20/04/2018


Stopbanks (levees) protect many New Zealand towns and cities from floodwater damage.  At the moment, stopbanks are managed based on local priorities, and we don’t have a nation-wide understanding of comparative protection level, impact, and risk. A new collaborative research programme links researchers at the University of Canterbury and University of Auckland with river managers from regional and unitary councils, in order to better understand and compare the characteristics of stopbank networks across New Zealand.

 

 

Rangitikei River stopbank breach, April 2017
Rangitikei River stopbank breach, April 2017. © GNS Science

Stopbanks are a primary form of flood protection for many towns and cities in New Zealand.  Like many forms of protection, they help to shield us from certain types of harm during unusual events.  Just like seatbelts, stopbanks are rarely needed, but when they are, we want them to work properly to protect us. However, no form of protection is able to prevent damage in every situation. 

In New Zealand, we presently have a flood protection approach that amounts to ‘local solutions to local problems’.  This approach means that each region manages their own flood protection based on available resources and priorities.  However, every region differs in its available resources and exposure to floods.   Every region is different in terms of the number of stopbanks, likelihood of flood events, areas (and assets) that stopbanks protect, and ratepayer base (ability to pay for improved flood protection).  Is your seatbelt a piece of rope tied around your waist?  Or is it a 5-point racecar harness?  Is it frayed or brand new?

Given the regional nature of stopbank and flood protection management, direct comparison of stopbanks across the country is difficult. Do stopbanks in Palmerston North provide as much protection as those in Paeroa?  Unlike seatbelts, which are assessed based on known standards during a Warrant of Fitness, there are no standardised national data sets, indicators or methodologies to assess the reliability of stopbanks across New Zealand. 

Over the past 18 months, researchers at the University of Canterbury and The University of Auckland have been working with regional councils and unitary authorities to compile the first New Zealand Inventory of Stopbanks (NZIS).  The intent of this study is to understand the make-up of stopbank assets in New Zealand by creating a single, standardised, reliable and spatially-referenced inventory. Analysis of the NZIS is helping us to understand the properties and condition of the New Zealand stopbank network – for example: height, type, geometry, location, design and service levels of the embankments that protect New Zealand from floods.  By considering the environmental conditions in different areas, such as seismic risk and liquefaction potential, we are beginning to understand what kind of conditions our stopbanks need to withstand (will your seatbelt be used on a high-speed Formula 1 racetrack or in a golf-cart?).

 

 

 

To date, and with the help of council river managers, researchers have assembled a database of over 4,800 km of stopbanks located across all 16 regions in New Zealand.  These stopbanks are presently being classified in terms of national liquefaction susceptibility models.  This will help us understand which stopbanks should be prioritised for inspection or remediation after earthquake events.

In the coming months, the final outcomes of the NZIS project will be presented to New Zealand council river managers. One of the main findings of the study is how little we know about the engineering properties of stopbanks – especially those that were built many decades ago.  Attributes such as height, year of construction, and embankment material type are not widely reported for stopbanks in the NZIS.   It is expected that further work will be required to fill in data gaps so that we can eventually predict the likely performance of stopbanks across New Zealand.

Just like seatbelts, stopbanks constitute an important form of protection for New Zealanders.  However, we don’t currently have a “Warrant of Fitness” standard to compare, assess, or rank the likely level of protection afforded by our stopbanks.  Current research aims to assemble a nation-wide inventory of stopbanks (NZIS) so that councils, government departments, and the public can begin to understand the reliability of the structures that protect us in floods.

How do different infrastructure networks influence one another following a natural hazard event?

 

04/04/2018


We rely on infrastructure networks every day. They provide us with essential services like electricity and water as well as transportation and waste collection. We are reliant on infrastructure networks to function as a society, and these networks are reliant on each other too. Following a natural hazard event this interdependency becomes even more critical. It is important to understand what essential services will be available after an event, and to do this we need to investigate how infrastructure networks influence one another.

 

Map-based representation of a selection of national infrastructure networks in New Zealand and their link to regional population density in the bottom map
Map-based representation of a selection of national infrastructure networks in New Zealand and their link to regional population density in the bottom map

 

The National Interdependent Infrastructure Model

Through collaboration between the Resilience to Nature’s Challenges and University of Oxford members of the Infrastructure Transitions Research Consortium, a model has been developed to characterise and quantify how network disruption spreads, both within each network and across into other networks.

The model quantifies the direct disruptions and the indirect disruptions due to dependencies and interdependencies across infrastructure networks. Direct disruptions are those resulting from damage within a network. Indirect disruptions to one network caused by damage to another network are a result of dependencies. Indirect disruptions where two different networks affect one another are a result of interdependencies. This model has been used to assess the potential spread of disruption for different hazard scenarios, along with identifying locations across the country that are most vulnerable to disruption.

 

Example of damage to potable water pipes due to movement and damage of the bridge structure following the Kaikōura earthquake. Original freshwater pipelines have failed and been replaced with temporary pipes (in black)
Example of damage to potable water pipes due to movement and damage of the bridge structure following the Kaikōura earthquake. Original freshwater pipelines have failed and been replaced with temporary pipes (in black)

What is an example of infrastructure dependency?

If there is damage to the electricity network at a certain point, all areas beyond this point would lose mains power – this is the direct disruption. If there is a pump station for the water supply network within the area where power has been lost, the pump can no longer function unless there is a backup power supply – this represents an example of indirect disruptions caused by the dependence of the water supply network on electricity. With the pump station affected, all areas that depend on it for supplying water will be affected.

Why is this important?

Infrastructure networks are increasingly reliant on each other and are becoming more interconnected, especially with changes in technology. Recent natural hazard events have highlighted the importance of infrastructure networks for response and recovery, and the importance of dependencies across different networks. However, dependencies are often not modelled due to their number and complexity, or simplified approaches are taken. The National Interdependent Infrastructure Model has enabled the quantification of both direct and indirect disruption across New Zealand due to natural hazard events. By quantifying these effects, the influence of pre-event investment and post-event decision making across these networks can be better assessed.

 

Failure of powerpole in the Kaikōura earthquake and subsequent damage to a telecommunications cabinet
Failure of powerpole in the Kaikōura earthquake and subsequent damage to a telecommunications cabinet

What next?

The model is a fundamental project from which other work can be linked. This will provide an improved understanding of individual networks and components of these networks, and in turn enable future improvements to the National Interdependent Infrastructure Model. We are developing approaches to assess the detailed functionality of networks, such as the power flow within electricity networks. These approaches are scalable, so regional or city scale models can be developed and linked back to the national model.

 

How can we keep the lights on during and after a natural disaster?

 

15/03/2018


 

Switching on lights, charging phones, connecting to the internet and communicating with friends have become basic necessities of modern living. We assume that electricity required for these activities will be available when we want to use it. But what if an earthquake, storm or volcanic eruption hits and a blackout follows?

 

How do you get electricity?

Electricity is generated at power stations and is then ‘transported’ by transmission and distribution networks through overhead lines and cables to the end user. Substations are a key part of this system, as they have specialized equipment that establishes a consistent quality of electricity supply throughout the network. A simple overview of the structure of the power system is as shown below:

 

https://www.electricityforum.com/electrical-training/power-system-analysis-and-design

Before the event

Electricity is crucial in modern society and industry. Hazard events can impact electricity infrastructure and cause power outages, so it is important that we assess and build resilience in this area to reduce that impact where possible. This can be done through physical investment into pole/tower strengthening, substation elevation or line reconstruction and relocation, which typically involves millions of dollars of capital investment. Our existing electricity network providers currently benchmark their assets on reliability metrics, which focuses on the potential small scale events, but this does not effectively account for large-scale natural hazard events.  

Dr. Yang Liu is developing new metrics to assess the resilience of power systems during natural hazard events. The metrics will assess the likelihood of different levels of damage to affected areas under various hazards. This will enable electricity network providers to figure out how resilient their existing system is, as well as where they should invest to improve that resilience. This will enable more effective post-disaster system functionality, including restoration of power after a hazard event. In addition, Liu is investigating how the existing reliability metrics are linked with the new resilience metrics.

 

Immediately after the event

Even if there is investment in the physical side of the power system, it is difficult to pinpoint the areas that are going to be affected by natural hazard events. There can be damage to power station equipment, transmission/distribution overhead line or underground cables and substation equipment. As a recent example, the 2016 Kaikōura earthquake resulted in a range of equipment damage as shown in the images below. This damage and consequent power outages led to blackouts in multiple areas.

 

Top left: Damaged face of the HVDC Tower, Top right: Land crack between legs of the HVDC tower, Bottom left: Oaro 33kV pole damage, Bottom right: Cracked 66 kV two pole

Images obtained from paper: Y. Liu, N. Nair, A. Renton and S. Wilson, “Impact of the Kaikōura Earthquake on the Electrical Power System Infrastructure”, Bulletin of the New Zealand Society for Earthquake Engineering, vol. 50, no. 2, 2017.

 

Damage to different components and parts of the network means that blackouts may be unavoidable in some situations, and efforts should be geared towards the restoration of power as quickly as possible. As simple as this sounds, here are three scenarios that highlight some of the complex decisions that need to be made:

 

  1. A component in a transmission substation has been damaged and repairs are likely to span days or weeks. However, nearby there is a wind power plant or a small hydro power plant that cannot operate without the transmission grid being available because of regulatory requirements. Is it possible to use the generation nearby to temporarily supply electricity?
  2. There is a damaged pole that needs to be replaced before electricity can be supplied to your home but the access to your area has been hampered by a damaged road. Should this road be repaired before other roads in the area?
  3. A section of consumers are without electricity and the distribution company relies on the feedback from consumers to identify where repairs are needed. However, the communication mast and fibre network in that area have been severely damaged. Therefore, the consumers cannot relay information to the distribution company by their usual means.

 

Damages on cables after the Kaikōura earthquake: Top left and right: Instances of permanent ground deformation and fault ruptures at cable locations; Bottom left: example of stretched cable; Bottom right: example of broken cable (Photos courtesy Rob Ruiter, Chorus)

 

From these scenarios, Duncan Maina, a PhD researcher at the University of Auckland, is seeking to examine the best same day restoration practices to ensure that the end user can be supplied with electricity with limited time and resources. He will also take into consideration the interdependencies among power, transport, communication and other lifeline infrastructure.

 

A few days to weeks after the event

Depending on the scale and severity of natural hazard events, electricity outages could last from a few hours to weeks or more. To reduce these outage times new methods of managing and operating the power system in an affected region can be used while repair is ongoing. An example of this is a micro-grid, which is a self-sufficient region with distributed energy sources within the broader power system. Distributed energy sources include small hydropower turbines, solar generators, wind turbines, bio-diesel, diesel and also energy storage systems such as batteries and electric vehicles (EVs) that are becoming prevalent in power distribution networks.

 

Bidirectional restoration process

Figure obtained from: A. Gholami, F. Aminifar and M. Shahidehpour, “Front Lines Against the Darkness: Enhancing the Resilience of the Electricity Grid Through Microgrid Facilities,” in IEEE Electrification Magazine, vol. 4, no. 1, pp. 18-24, March 2016. doi: 10.1109/MELE.2015.2509879]

 

Infrastructure team PhD student Samad Shirzadi is exploring these new distributed energy processes, working closely with impacted distribution utilities.   

To implement these new processes a robust communication system will be required. This part of the research is being undertaken Master’s student Farrukh Latif. Farrukh willl investigate how communication networks are affected following large scale natural disasters, and based on this research, develop and propose strategies to improve resilience. In particular, he will focus on electric-power grid control/monitoring communication networks during earthquake sequences. 

 

Months after the event

Even after the power system has been brought back to normalcy and all levels of the network are operational, we need to monitor power network equipment that were in the regions affected by the natural hazard event but did not experience immediate damage. This is necessary as, for example, an underground cable might be functioning well after a disaster, but its rate of failure increases after the disaster. Ebad Rehman, a PhD researcher is analysing and investigating this phenomenon, which will help establish best practices of assessing long-term resilience of power infrastructure.

 

Conclusion

New Zealand electricity infrastructure has unique threats, and exploring these threats provides opportunities to reduce impacts from extreme natural hazards. In order to build resilience in this area we must maintain social awareness and consistent interdisciplinary cooperation between industry, academia and related governmental organizations.  

This Resilience to Nature’s Challenges Infrastructure Toolbox group of projects works alongside the Resilience to Nature’s Challenges Rural Toolbox, Project AF8, QuakeCoRE and electricity/communication industry stakeholders.  The research projects identified in this piece are being investigated and led by PI: Nirmal Nair and AI: Andrew Austin based at The University of Auckland.

How do different infrastructure networks influence one another following a natural hazard event?

 

14/03/2018


 

We rely on infrastructure networks every day. They provide us with essential services like electricity and water as well as transportation and waste collection. We are reliant on infrastructure networks to function as a society, and these networks are reliant on each other too. Following a natural hazard event this interdependency becomes even more critical. It is important to understand what essential services will be available after an event, and to do this we need to investigate how infrastructure networks influence one another.

 

Map-based representation of a selection of national infrastructure networks in New Zealand and their link to regional population density in the bottom map
Map-based representation of a selection of national infrastructure networks in New Zealand and their link to regional population density in the bottom map

 

The National Interdependent Infrastructure Model

Through collaboration between the Resilience to Nature’s Challenges and University of Oxford members of the Infrastructure Transitions Research Consortium, a model has been developed to characterise and quantify how network disruption spreads, both within each network and across into other networks.

The model quantifies the direct disruptions and the indirect disruptions due to dependencies and interdependencies across infrastructure networks. Direct disruptions are those resulting from damage within a network. Indirect disruptions to one network caused by damage to another network are a result of dependencies. Indirect disruptions where two different networks affect one another are a result of interdependencies. This model has been used to assess the potential spread of disruption for different hazard scenarios, along with identifying locations across the country that are most vulnerable to disruption.

 

Example of damage to potable water pipes due to movement and damage of the bridge structure following the Kaikōura earthquake. Original freshwater pipelines have failed and been replaced with temporary pipes (in black)
Example of damage to potable water pipes due to movement and damage of the bridge structure following the Kaikōura earthquake. Original freshwater pipelines have failed and been replaced with temporary pipes (in black)

What is an example of infrastructure dependency?

If there is damage to the electricity network at a certain point, all areas beyond this point would lose mains power – this is the direct disruption. If there is a pump station for the water supply network within the area where power has been lost, the pump can no longer function unless there is a backup power supply – this represents an example of indirect disruptions caused by the dependence of the water supply network on electricity. With the pump station affected, all areas that depend on it for supplying water will be affected.

 

Why is this important?

Infrastructure networks are increasingly reliant on each other and are becoming more interconnected, especially with changes in technology. Recent natural hazard events have highlighted the importance of infrastructure networks for response and recovery, and the importance of dependencies across different networks. However, dependencies are often not modelled due to their number and complexity, or simplified approaches are taken. The National Interdependent Infrastructure Model has enabled the quantification of both direct and indirect disruption across New Zealand due to natural hazard events. By quantifying these effects, the influence of pre-event investment and post-event decision making across these networks can be better assessed.

 

Failure of powerpole in the Kaikōura earthquake and subsequent damage to a telecommunications cabinet
Failure of powerpole in the Kaikōura earthquake and subsequent damage to a telecommunications cabinet

What next?

The model is a fundamental project from which other work can be linked. This will provide an improved understanding of individual networks and components of these networks, and in turn enable future improvements to the National Interdependent Infrastructure Model. We are developing approaches to assess the detailed functionality of networks, such as the power flow within electricity networks. These approaches are scalable, so regional or city scale models can be developed and linked back to the national model.

Can we evacuate Auckland before a volcano erupts?

 

12/03/2018


It is only a matter of time before another volcano erupts in Auckland’s Volcanic Field. The question that Resilience Challenge researchers are tasked with is whether we can evacuate everyone from the affected area in time when it happens.

 

Mount Eden (Maungawhau) volcanic cone, Auckland. Photo: Daniel Blake

 

The lack of data

As no evacuations of the Auckland city area have taken place to provide historical data, and an evacuation exercise of this magnitude is impractical, it is difficult to predict whether we can evacuate Auckland’s residents if monitoring of seismic activity suggests a volcanic eruption is imminent. In any case, the uncertainly is large, in terms of location, styles, sequence and advance warning of any volcanic eruption.

 

Driving under pressure

Traffic congestion in Auckland. Photo: russelsteet

 Auckland’s congested transport network has its own challenges. In particular, Auckland’s geographic location, between two isthmuses (narrow sections of land that connect larger landmasses), results in only a limited number of evacuation routes. In addition, all our knowledge of Auckland’s travel and driver behaviour is based on day-to-day activities, for example the daily commute, shopping trips and the school run. How we react and behave in an emergency adds another degree of complexity.

 

Our research

How to allow for such uncertainty, constraints and complexity is undoubtedly a challenge. To address this challenge two research projects are underway. The first is attempting to understand human and driver behaviour in emergency situations in Auckland. We are all human, and when faced with an emergency we behave differently. How this behaviour affects the ability to evacuate Auckland city is the ultimate aim of the research.

Maps of congestion hotspots during evacuation, generated using the AIMSUN software

The second research project underway is building a simulation model of Auckland’s transport network. Volcanic scenarios developed by other researchers in the Challenge and Determining Volcanic Risk in Auckland (DeVoRA) project will be used as case studies and evacuations will be simulated. Outputs from the evacuation models will include the length of time needed to vacate the evacuation zone at different times of the day, the number of people still left in the evacuation zone when the volcano erupts, and the bottlenecks in the road network. This information will then allow Civil Defence and Emergency Management, among others, to understand, prepare for and, if needed, amend the existing evacuation plan to maximise the number of people safely evacuated.

So, can we evacuate everyone in time? Unfortunately it is not a yes/no answer. However, given scenarios on the location, type, size and advance warning of the next volcanic eruption, the transport model being developed to simulate evacuation, will provide us with an answer. Where the answer isn’t the one we want to hear, mitigation measures can be proposed and simulated which will increase the number of people that can be evacuated and ultimately save lives.

 

Traffic congestion on Auckland’s SH16. Photo: Daniel Blake

 

Staff: Associate Professor Seosamh Costello, Dr Prakash Ranjitkar, Dr Subeh Chowdhury

Postdoc: Daniel Blake

PhD Students: Mujaddad Afzal and Snehalata Thakur

Student Profile: Xavier Bellagamba

 

09/03/2018


Perspectives on lifeline resilience

 

About Xavier

small

Xavier comes from the Canton of Valais at the heart of the Northern Alps in Switzerland. After completing an undergraduate program in Civil Engineering from EPFL, Lausanne, he joined the Master’s program of Civil Engineering at ETH, Zurich.

Interested to pursue his research on infrastructure resilience he started at ETH. As well as exploring new horizons, Xavier applied for a co-funded Resilience to Nature’s Challenges-QuakeCoRE scholarship and started his work in February 2016 at University of Canterbury under Professor Brendon Bradley. He is co-supervised by Dr. Liam Wotherspoon and Dr. Matthew Hughes. His work explores multiple aspects of underground lifeline seismic resilience and exploits the data collected on the water supply network of Christchurch after the Canterbury earthquake sequence events. Excited by the digital revolution, Xavier enthusiastically applies some machine learning and AI techniques to tackle problems related to the seismic resilience of urban water supply systems.

During his free time, Xavier enjoys reading (economics, sociology and psychology), jogging and cooking. If the weather is good, you better look for him in the mountains (on skis, in tramping or running shoes or harnessed at one end of a rope).

 

You’ve said resilience…

The resilience of a system (either if it is a building composed of several frames, an organization or spatially-distributed infrastructure) should be considered from three different perspectives: the loss assessment (what parts of the system are damaged and how does it affect its operability?), the risk mitigation (what parts of the system should be retrofitted, removed or strengthened to minimize future losses?) and the recovery (what parts of system should be repaired first to minimize disruption?).

 

Loss assessment

The loss assessment of an interconnected system is carried out in two steps: first, the damage estimation and second its impact on the system operability. As Christchurch is located in a liquefaction-prone area, new and easy-to-apply fragility functions for pipelines in such soils have been developed utilizing the aforementioned dataset.

These functions are then utilized in a connectivity-based network engine to estimate the systemic loss of service (e.g. disconnected population or dwellings). This engine has been used to estimate the number of people, businesses and global utility that have been affected by a water outage following the 2011 February Mw6.2 Christchurch earthquake as shown below.

 

Simulated water outage after the 2011 February Mw 6.2 Christchurch earthquake
Simulated water outage after the 2011 February Mw 6.2 Christchurch earthquake

 

Mitigation

Urban infrastructure are complex systems that evolve over time through the adoption of new standards or materials, growth, asset renewal and mitigation strategies. Hence, designing sound strategies accounting for multiple natural hazards is a very difficult task. In order to help horizontal infrastructure asset managers or urban planners to assess their options, a framework that could be applied across multiple systems and hazards is proposed. Based on the results of time-dependent event catalogues, it will rank the different strategies based on their relative performance, mitigating future losses for the considered hazards.

 

Recovery

The recovery of a system encapsulates two key factors: the recovery process optimization and the ability to rapidly estimate the damage extent. Based on the engine developed for the loss assessment, a resource-based recovery model is proposed. Its objective is to minimize the selected community metric (population, dwellings or utility) suffering from a water outage over the restoration period. Figure 2 provides an example of simulated water outage recovery.

Simulated recovery of water services following the 2011 February Mw6.2 Christchurch earthquake
Simulated recovery of water services following the 2011 February Mw6.2 Christchurch earthquake

Built on the CyberShake program and modern deep learning methods, a tool is being developed to rapidly estimate the most probable shake map after potential future earthquakes. Combining this approach with state-of-the-art damage assessment software will allow emergency response planners to fasten and sharpen their decisions within a chaotic environment.

 

Why is this useful?

The development of these tools will contribute to a better understanding of complex civil infrastructure systems subject to earthquakes. In particular, whereas parts of this work can easily be implemented into existing loss modelling software, other portions could be directly applied by consultants, urban planners or asset managers to assist them in the assessment of their mitigation, development or recovery strategies.