Preparing for complex interdependent risks: A System of Systems approach to building disaster resilience

https://doi.org/10.1016/j.ijdrr.2014.05.001Get rights and content

Abstract

The uncertainty posed by natural and human-made disasters arises from both known risks and a range of unforeseeable risks, some of which may be novel, not having been observed before. These interconnected risks may evolve over short periods of time and may feed into one another. In a network of multiple causes and effects, such risks may not be foreseeable at the disaster preparedness level, and may only be observed at the time of disaster response. This creates a higher level of complexity and requires new approaches with individual organizations and members needing to make decisions outside predefined frameworks and hierarchical command-control structures while still operating in the ethos of their organizations.

This study advocates the need for disaster preparedness strategies to go beyond linear approaches to risk management. This is necessary in order to better address complex interdependent risks where such risks may be novel or unforeseen and which may connect in a cascading manner. The resulting causal network needs to be addressed with a networked approach to enrich existing linear approaches by recognizing the need for an interconnected holistic approach to deal appropriately with interconnected risk factors.

This paper takes an interpretive perspective rather than the more typical positivist one. System of Systems (SoS) and complex systems thinking were used to inform a sense-making framework to distinguish between approaches to known/knowable and unknown risks.

Finally, the paper reports on how this framework was used in South Australia on three different scales of the SoS: community, NGOs and government.

Introduction

The Hyogo Framework for Action (HFA, [57]) and subsequent national strategies such as the Australian National Strategy for Disaster Resilience (NSDR, [42]), highlight the importance of assessing risks and preparing for them [16].

This study is based on a research project, which was funded by the Australian Natural Disaster Resilience Grant Scheme, an Australian government initiative to enhance disaster resilience building. The Australian National Strategy for Disaster Resilience (NSDR) in particular has encouraged research beyond the borders of individual organizations, to focus on the empowerment of communities to increase their resilience in response to disasters. This strategic goal was formulated in response to bushfires, cyclones and floods that have recently intensified in Australia due to climate change [54]. The intent of the NSDR is to promote collaboration between disaster management practitioners and local communities to build disaster resilience.

Disaster preparedness strategies are deployed to mitigate disaster risks and to build community resilience. Therefore, there is a strong connection between disaster resilience, disaster preparedness and risk management.

Current disaster preparedness strategies often focus on building resilience for known disaster risks. However, disasters are characterized by interdependent and systemic risks that can trigger cascading effects [36] which are hard to predict. The ‘unexpected’ is already part of the life of many communities. For this reason, there is an urgent need to investigate ways to prepare for what we are not able to predict or to communicate.

Drawing on complexity theories, we provide a sense-making framework on preparing for the unexpected by creating networks capable of building general resilience, that is, resilience to unknown risks.

As we discuss later, these topics are often addressed from a positivist empiricist perspective. While acknowledging the importance of these approaches to prepare for known and knowable disaster risks, this study argues the need for a complementary way of thinking to develop capability to prepare for unknown risks by building general resilience in the wider community.

The majority of risk assessments take the approach of ISO 310001 [33], which defines risk as ‘the effect of uncertainty on objectives’. This standard is based on the assumption that risk identification is the first step in managing risks. Therefore, many people use the ISO 31000 process of identifying a series of individual risks, assessing their individual likelihoods and consequences, and then comparing these. They then calculate the individual risks, either by qualitative or quantitative measures, evaluate their potential impact (as a product of likelihood and consequence) and finally sort them by severity of impact. Normally only the highest rated risks are addressed Fig. 1.

This approach is valuable in the case of traditional projects where problems, or, in this case, risks, can be broken down into their components and strategies can be created to deal with them. Unfortunately, linear risk assessments fail to take into account the impact of unpredicted risks (Homer-Dixon 1995 in [47]). This has become particularly significant as we confront climate change and the uncertainty of its effects on environment and populations [58].

ISO 31000 has clarified the steps to manage risks and has thus opened the way for risk management to be implemented in and by public organizations and governments. This has led to better informed decisions and more robust policies in terms of disaster preparedness. This approach has been used, for example, in areas of Australia that are prone to bushfires, by targeting community preparedness campaigns to raise awareness of risks and consequences; and in the Philippines, where local governments are seeking to address multiple disaster risks (e.g. typhoons, earthquakes and volcanic eruptions) through existing plans and government regulatory processes [5].

While there seems to be a robust basis for programs and research around how best to prepare the community for identified risks, disasters show increasingly unprecedented consequences. For example, typhoon Yolanda struck with unprecedented wind strength and storm surge in the Philippines in November 2013 [21]. A further example is the Fukushima disaster in Japan in March 2011. An unprecedented chain reaction involved an earthquake and a tsunami triggering a nuclear disaster2: the blackout of cooling systems to reactors of the Daiichi nuclear power plant led to meltdowns and the release of radioactivity [44].

Often when disasters cause heavy losses, a lack of preparation and information is blamed. However, modern disasters have shown that more information does not always correspond to higher safety or preparedness levels, especially where unanticipated events occur. For example, in the case of the previously discussed Fukushima disaster, the Japanese government, the Nuclear Safety Commission (NSC) and the company at the centre of the nuclear disaster, the Tokyo Electric Power Company (TEPCO) were aware of seismic, tsunami and nuclear risks [24], [56]. Despite this, the Nuclear Safety Commission (NSC), which oversees the Nuclear and Industrial Safety Agency (NISA) did not put in place adequate regulations and these in turn were not imposed on TEPCO. For example, outages were addressed as temporary risks, because they could have been quickly solved as they occurred. Consequently, TEPCO addressed the crisis inadequately and even contributed to its worsening [26] Fig. 2.

The underlying assumption of this analysis may be to suggest that stricter well implemented regulations would benefit the disaster severity. However, appropriate international regulations existed already at the time of the disaster, but TEPCO did not abide by them. This leads to a question on the effectiveness of international control and monitoring institutions and regulations, since government, nuclear authorities and TEPCO knew about them, but they did not put them into place [14].

An increased level of knowledge does not necessarily mean better disaster management. Disasters are often emergences. The word ‘emergence’, which is not to be confused with ‘emergency’, refers to the formation of new system properties as the result of evolutionary behavior [11]. In this sense, disasters trigger unique events that have never occurred before and that may never happen again with the same characteristics.

Statistical and modeling approaches to the complexity of systemic resilience have increased our understanding of past events, because they are based on probability theory [18]. These are extremely important to prepare for risks on which some information is available. However, modern disasters have shown a number of unanticipated emergent and cascading effects. For this reason, there is a strong need to investigate new ways to prepare for unexpected interdependent risk events, which trigger cascading effects [28], [47].

The prediction of cascading failures can be particularly difficult for two reasons. First, information is often scarce or insufficient to use approaches such as robust Bayesian methods [49], which are normally drawn on when the degree of uncertainty is considered very high and unmanageable without modeling support [18]. Second, public awareness of risks as identified by disaster prevention agencies is limited, leaving the rest of the population unprepared. Consequently, the challenge for regulators and practitioners is to design policies and disaster preparedness strategies that align with the capacity of, for example, individual communities and businesses.

A number of researchers acknowledge the need to go beyond current approaches by reviewing the underlying assumptions of today׳s risk assessment methods. For example, Aven [6] suggests an integrated risk perspective based on probability, knowledge strength and consequence as an alternative to the product of probability and consequence [51]. Furthermore, Aven and Krohn [7] identify the introduction of quality management and mindfulness related tools as a possibility to extend the risk perspective in a complementary way to positivist empiricist approaches that often do not reflect the observer׳s perception of uncertainty [55].

Research around disaster preparedness has been mainly linear in approach, focusing on known risks [18], [27]. However, this approach is limited, with sometimes inadvertently serious consequences, when for example:

  • The interactions between different risks are not considered, e.g. earthquake, tsunami and nuclear disaster in Japan [31], [58].

  • The gradual worsening of environmental conditions is neglected, e.g. increased number of natural disasters or sea level rise [45]

  • After disasters, lessons learned focus on actions specified in existing regulations rather than on an analysis of the disaster׳s development and changed conditions. As a result, lessons learned do not provide appropriate interconnected guidance for future risk and disaster preparedness [14].

  • ‘Extreme weather’ has been planned for as an exceptional condition, something that happens rarely (e.g. [39]); however, natural disasters have become so frequent (e.g. [17]) that they are increasingly part of a community׳s routine [8].

  • Information fails to reach communities and/or it is not specific to the community׳s context [1].

Even though many disaster risks cannot be predicted, there is potential to prepare for them [25]. Unknown risks can be managed.

Section snippets

Why system thinking in disaster preparedness

Risks are often described as linear cause-effect relationships [8], [13]. They are compiled on a risk register, evaluated and addressed individually [32], [48]. This process has brought important outcomes such as information sharing with the wider community and awareness raising campaigns.

However, in some cases such as those already discussed, risk assessments of this kind fail to reflect the actual risk situation [38]. This is the case when relatively unrelated systems affect each other.

Complex and systemic risks

Linear causal relationships constitute only a part of the risk network that emerges when disasters strike. We argue that, from a cognitive point of view, in a disaster, there are at least two types of risks. Complicated risks are characterized by cause–effect relationships that can be understood in advance, before the risk event occurs. Bush fires can be considered as an example. Stochastic approaches are appropriate for many complicated risks, where information is available from past disasters.

Resilience thinking

Resilience. This ancient word – Cicero used it in his Orations [3] – has been used to mean different things over the years and across a plethora of disciplines such as engineering, social sciences and, more recently, disaster risk reduction. It is also something of a buzzword requiring its usage here to be made clear.

Characteristics

Resilience is often defined through the characteristics of a resilient system. For example, ‘Your community will prove resilient in the event of a severe emergency

Specified and general resilience

Table 1 illustrates the two types of approaches to resilience building that we have discussed. It should be noted that the following is a sense-making table, not an operational one [52], that can support disaster risk reduction practitioners and researchers in explaining what many intuitively already know.

Specified resilience refers to known risks, whose consequences have already been observed in the world. This is the case with many natural disasters. Action plans are normally reductionist,

Building general resilience starting from the community

For the purpose of this paper, we describe how we framed this research project, rather than discussing findings that are specific of the SoS studied Fig. 7.

In the following section, we describe how we overlapped different scale views to arrive at a holistic view for interdependent and systemic preparedness to unexpected risks.

Disaster management involves a range of very different systems: governmental, political, historical, social, economic, financial, environmental, etc. For this reason,

Conclusion

Natural and human-made disasters are increasingly posing a threat to many communities. Areas that have not previously been affected by disasters are increasingly at risk. Disaster risks have been principally assessed following linear methods that neglected to consider risks in their causal networks. Recent disasters have drawn attention to the cascading effects of disaster risks that emerge in every disaster. Cascading effects of interdependent risks are challenging to model and predict.

This

Acknowledgments

A version of this paper has been developed for the HFA Thematic Review and as an input to the Global Assessment Report on Disaster Risk Reduction 2015 (GAR15). The authors wish to acknowledge the financial support of the Australian Government, the South Australian Fire and Emergency Services Commission (SAFECOM) (Grant no. NDRP-1213-33) and the Entrepreneurship, Commercialisation and Innovation Centre (ECIC) (Grant no. 56111309) at the University of Adelaide, South Australia. Moreover, they

References (64)

  • U. Beck

    Cosmopolitanism as imagined communities of global risk

    Am Behav Scientist

    (2011)
  • R. Bhaskar

    The possibility of naturalism. A philosophical critique of the contemporary human sciences

    (1998)
  • J. Boardman et al.

    Systems thinking. Coping with 21st century problems

    (2008)
  • F.E. Boteler

    Building disaster-resilient families, communities, and businesses

    J Extension

    (2007)
  • A. Cavallo

    Risk management in complex projects. An exploratory study to managing unknown unknowns in uncertain environments

    (2010)
  • Cavallo A, Ireland V. SoS in disasters: why following the manual can be a mistake. In: Proceedings of the IEEE...
  • P. Checkland et al.

    Learning for action: a short definitive account of soft systems methodology, and its use practitioners, teachers and students

    (2006)
  • Childs D, Gordy M., Gordon M. Implementation of the Hyogo Framework for Action. Summary of Reports 2007–2013, The...
  • Climate Council of Australia. Angry Summer 2013/2014; 2014....
  • Comes T, Cavallo A. Designing decision support systems at the interface between complex and complicated domains. In:...
  • Crotty M. The Foundations of Social Research: Meaning and Perspective in the Research Process Allen & Unwin, Crows Nest...
  • A.M. Cruz et al.

    Vulnerability of the oil and gas sector to climate change and extreme weather events

    Clim Change

    (2013)
  • J. Daniell et al.

    Super Typhoon Haiyan/Yolanda

    (2013)
  • De Rosa JK, Grisogono A-M, Ryan AJ, Norman D. A Research Agenda for the Engineering of Complex Systems. SysCon 2008 –...
  • J.F. Flach

    Complexity: learning to muddle through

    Cogn, Technol Work

    (2012)
  • Y. Funabashi et al.

    Fukushima in review: a complex disaster, a disastrous response

    Bull Atom Scientists

    (2012)
  • D.R. Gilpin et al.

    Crisis Management in a Complex World

    (2008)
  • Hatamura Y., Oike K., Kakinuma S., Takasu Y., Takano T., Tanaka Y., Hayashi Y., Furukawa M., Yanagida K., Yoshioka H.,...
  • C.E. Haque et al.

    People and community as constituent parts of hazards: the significance of societal dimensions in hazards analysis

    Nat Hazards

    (2007)
  • D. Helbing

    Globally networked risks and how to respond

    Nature

    (2013)
  • E. Hollnagel

    Safety-I and Safety-II: the past and future of safety management

    (2014)
  • M.E. Ibarraran et al.

    Climate change and natural disasters: macroeconomic performance and distributional impacts

    Environ, Dev Sustainability

    (2009)
  • Cited by (0)

    Prepared for the Hyogo Framework for Action (HFA) Thematic Review and as an input to the United Nations Global Assessment Report on Disaster Risk Reduction 2015 (GAR15).

    View full text