IEEE Systems, Man and Cybernetics Magazine - July 2019 - 17

modeling can be used to inform investment prioritization in
a country whose requirements and economic development
are very different from those of the United Kingdom [26].
The study considered Curaçao's options for energy,
water, solid waste, wastewater, and transport infrastructure, highlighting "quick wins" that could deliver benefits
in the next few years (like construction of waste-to-energy
facilities and addressing leakage and waste in water supplies) as well as setting out longer-term options. It
assessed the risks of sea-level rise and storm flooding to
the road network and critical buildings, such as those
related to health and education. Minister of Traffic, Transportation, and Urban Planning Zita Jesus-Leito stated,
"Our infrastructure is vital for the functioning of Curaçao
today and its future success. Therefore, it should be optimized, efficient, and resilient. In that context, cross-sectoral long-term planning is essential for maximizing the
full potential of our island for the benefit of all its people"
[26]. Conducting an analysis in relatively data-scarce situations on a short time scale was challenging, but global data
sets from remote sensing and crowdsourcing, combined
with good stakeholder support within countries, are making this type of study increasingly feasible.
Modeling the Risks Associated With
Infrastructure Interdependencies
In the preceding section, I explained how interrelationships can be incorporated in the long-term planning of
infrastructure systems. In that context, interdependencies
are significant if there are major flows of resources (e.g.,
electricity and water) from one sector to another, on a
scale whereby omitting these fluxes would result in inaccurate planning assumptions. I have also considered interdependencies in final demand; so, for example, an increase
in population would be expected to put upward pressure
on both energy and water usage.
Some of the greatest concerns about interreliance
derive from its potentially disruptive effects during natural
catastrophes or security breaches. Understanding the
nature of these interdependencies tends to require much
more detailed information than for long-term planning
purposes. In this context, the interrelationships are often
categorized as follows:
◆ Geographic: Two or more systems are colocated in
physical space.
◆ Physical: A physical output from one system is a necessary input to another.
◆ Cyber: Information produced by a system affects the
operation of another.
◆ Human: There are shared dependencies on people,
e.g., workers or organizational or social systems.
Considering geographical interdependencies, it is necessary to know the location of individual assets in relation to
the hazards to which they are exposed (e.g., are the assets
in a floodplain?). An analysis of the likelihood of the failure
of a given asset, given a particular hazard scenario, such as

a flood, involves understanding the geometry of the asset
(e.g., its elevation relative to the flood level), its condition,
and its function. Though comprehensive analysis at a
national scale is still some way off, significant elements of
such analysis are already in place. For example, by interrogating network data from electricity transmission and distribution network operators in the United Kingdom and
using statistical techniques to fill in missing data, it has
been possible to develop the full electricity network hierarchy for Great Britain and the dependence of most other
infrastructure assets on that network (Figure 3).
In addition, the analysis of rail bridge failures due to
foundation scour at river crossings used a unique data set
of 100 bridge failures over the period 1830-2003 to estimate the fragility of bridges in river floods [27]. The number of customer s who a re dependent upon a ny
infrastructure asset was also estimated using hours of
customer disruption as a common metric across infrastructure sectors. Using source-sink relationships
between supply points and users and modeling the
amount of disruption during a very large number of scenarios of network degradation, taking account of the
potential for flow rerouting, it is possible to estimate how
many customers might be disrupted in hazard scenarios
of an array of different severities.
An analysis of network vulnerabilities and customer
disruptions helps to identify hot spots of vulnerability
in the network [28]. The next step is to use this information to prioritize interventions that increase resilience.
The business case for intervention to increase resilience rests on the comparison of the cost of the intervention with the value of the losses that are expected to
be avoided. We have, for example, undertaken a version
of this cost-benefit calculation for electricity substations, exploring a variety of possible interventions,
such as building flood protection around substations,
raising plants above the ground, and relocating facilities [29]. This results in a prioritization of substations
for investment in protective measures, though that prioritization is sensitive to estimates of the scale of econom ic i mpa ct . T hu s, a robu st nes s a na lysis wa s
implemented to illustrate over what range of economic
impacts we expect investments in protection to continue to be cost beneficial.
An analysis of cyberattacks on infrastructure networks
is even more challenging because the consequences of the
attack need not be geographically coherent. Faults and/or
hacks in software systems can simultaneously impact
many different assets. In an analysis of the potential consequences of cyberattacks for electricity substation infrastructure [30], several different threat scenarios, including
simultaneous cyberattacks on the same type of hardware
located at several dispersed locations across eastern England, were analyzed to identify the potential scale of disruption and hence help to inform the prioritization of
protective measures.
Ju ly 2019

IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE

IEEE Systems, Man and Cybernetics Magazine - July 2019

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - July 2019