The shock of Covid-19 has reminded us that disruptive events do not always materialise in the way that was expected. The UK Government’s 2016 Exercise Cygnus examined the scenario of a pandemic flu, which was profoundly different to a virus like Covid-19 that could transmit asymptomatically. Nonetheless, Cygnus revealed weaknesses in the country’s preparedness that were relevant to the Covid-19 pandemic but were not fully acted upon.
Critical infrastructure network failures in recent years have revealed the range of conditions under which infrastructure systems can fail: from floods and landslides to drones and cyber-attacks. All of those events could have been foreseen, though no doubt there are other failure conditions that would strike us as being more surprising.
In the face of the predictable and unpredictable, what can systems analysis tell us about the resilience of infrastructure systems? Even before we start considering the damaging threats to which an infrastructure network may be exposed, there is a lot that we can learn by examining the configuration of that network: how does it deliver services, what is it interconnected to, and how strong are the critical elements in the system?
When resources are scarce, governments need to know how to prioritise resilience investments in public assets. Retrofitting or replacing infrastructure assets can be extremely costly, so strengthening everything is unlikely to be affordable. Governments need rigorous methods of prioritisation which account for the uncertainties in future climate changes and in the condition and rate of deterioration of their infrastructure assets. Jim Hall, Professor of Climate and Environmental Risk at the University of Oxford
Though we can learn a lot from network analysis – from how different assets in the network are interconnected – to understand real networks we need to know about the capacities of each link in the network, the resources that flow along them and how resistant they are to extreme loading.
In many respects these are familiar engineering problems. The novelty of today’s system analysis is that we are now able to quantify the vulnerability of enormous networks. In an analysis for the National Infrastructure Commission (NIC) [1], my group analysed interdependent energy, transport, water and telecommunications networks for all of Great Britain. That inevitably involved some approximations, but we were still able to show how failures can propagate from one network to another… and back again.
That analysis for the NIC built upon previous analysis, inspired by officials at Infrastructure UK a few years previously, which identified what we referred to as ‘hotspots’ of critical infrastructure failure [2], where failure of one or more infrastructure could lead to a disproportionate amount of disruption.
These analyses provide a guide to infrastructure owners and operators for where they should be paying most attention, in surveillance, strengthening and protection. Actually, infrastructure operators usually know their own networks pretty well. They have asset management systems which record the location and condition of assets. Yet they seldom have the same amount of information about the networks upon which they depend – the power or telecommunications. This is where system-of-systems modelling is required, to enable sharing of information and management of systemic risks.
My research group’s analysis of critical national infrastructure in the UK has actually attracted even more attention around the world than it has in this country. There are ever-growing calls to enhance the resilience of infrastructure systems to the risks associated with climate change. The hazards from climate change to infrastructure include floods (river, coastal, pluvial), droughts (which can impact hydropower as well as water supplies), landslides, cyclones, heatwaves and wildfires. Damage from these extreme events is ever-growing, to a point that it can represent an unmanageable burden upon national balance sheets.
When resources are scarce, governments need to know how to prioritise resilience investments in public assets. Retrofitting or replacing infrastructure assets can be extremely costly, so strengthening everything is unlikely to be affordable. Governments need rigorous methods of prioritisation which account for the uncertainties in future climate changes and in the condition and rate of deterioration of their infrastructure assets.
Big new global datasets, notably OpenStreetMap and machine learning from satellite images and photos, are providing unprecedented amounts of information about infrastructure everywhere on Earth. Working with the World Bank, we analysed more than 50 million kilometres of road data to estimate global vulnerability to natural hazards [3]. That can be combined with large ensembles of simulated climatic extremes. None of this analysis is perfect – in fact there are big uncertainties – but it provides a hint of the global analysis of infrastructure resilience which may become possible in the future.
- The CSIC 2021 Distinguished Lecture by Professor Jim Hall, Professor of Climate and Environmental Risks in the University of Oxford, titled ‘The data revolution in global-scale analysis of climate risks to infrastructure systems’ takes place on Tuesday 13 July from 5pm to 6.30pm. This is a free-to-attend and online event. For more information see here.
- Please register here.
Footnotes
- https://wordpress-522998-1751448.cloudwaysapps.com/studies-reports/resilience/system-analysis-of-interdependent-network-vulnerabilities/
- Thacker, S., Barr, S., Pant, R., Hall, J.W., and Alderson, D. Geographic hotspots of critical national infrastructure. Risk Analysis, 11(1) (2018): 22-33. DOI: 10.1111/risa.12840.
- Koks, E.E., Rozenberg, J., Zorn, C., Tariverdi, M., Vousdoukas, M., Fraser, S.A., Hall, J.W., Hallegatte, S. A global multi-hazard risk analysis of road and railway infrastructure assets. Nature Communications, 10(1) (2019): 2677. DOI: 10.1038/s41467-019-10442-3.