Simplifying disasters: developing a

model for complex non-linear events

Introduction The simplest disaster model is probably
All disasters are complex events. If this by Charles Kelly, the sequence Pre-event, Disaster, Post Event.
complexity can be reduced, a disaster Disaster Management Consultant More detailed variants of the linear
becomes less severe and thus less trau- Presented at the Disaster Management: model are provided below. The first is the
matic, disruptive and damaging. A first step Crisis and opportunity: Hazard Management standard process model of a disaster
toward reducing complexity is to better and Disaster Preparedness in Australasia and composed of stages involving Preparedness,
understand how the different facets of a the Pacific Region Conference, Response, Recovery, Mitigation (Neal,
disaster relate. November 1–4, 1998, Cairns, Queensland. 1997).
This understanding can be developed Currently, this model is presented as a
through the use of a model of the process continuum from disaster to development,
of a disaster event. The focus here is not on and clear model helps victims and assist- ‘. . . a linear progression from a state of crisis
models to predict specific types of events. ance providers play the same game in the through rehabilitation to development’
It is on a generic, process-oriented model same ballpark. (Macrea, 1997).
which applies to the unfolding of all types Finally, and in a similar manner, a The second variant of the linear disaster
of disasters, and which may be used in disaster process model can be useful to process model comes from Harrald and
monitoring and guiding disaster manage- disaster managers in explaining the course Stoddart (1998) and ‘ . . . characterises the
ment efforts. and possible future outcomes of a disaster phases of a disaster by the evolution of
to non-specialists. If the model is clear (a internal functions and tasks of the organ-
Why a model? major condition), then its presentation to isational structure’: Initiation/Mobilisation
There are a number of reasons why a dis- select or general audiences can facilitate (storming, forming), Integration (norming),
aster process model can be useful. First, a securing support for disaster management Production (performing), Demobilisation
model can simplify complex events by help- efforts. (transition).
ing to distinguish between critical elements Although using different terminology,
and noise. This is particularly useful in a Disaster process models both models cover the same pre-event,
high-pressure disaster response envir- An overview event, post-event sequence noted above.
onment, when little time is available to To date, the development of a disaster This linear sequencing of disaster stages
think about events or to identify critical process model has been based on ident- has been subject to criticism. Neal (1997)
issues. Second, contrasting actual con- ifying the stages, events, actions and time highlights three points, among others. First,
ditions with a theoretical model can lead frame which make up the course of a different stages occur at the same time for
to a better understanding of the current disaster (see Neal, Haas et al. and Frerks et different segments of a population. Second,
situation, and how a disaster may evolve. al. for different perspectives on the disaster some events are relevant to more than one
This facilitates planning and helps make process). Of the three elements, the time stage. Third, stage ‘. . . divisions are arbit-
plans more complete. element is easiest to define. It can be rary, and are only useful in distinguishing
Third, having available a model of the quantified down to seconds, but more often the major functional activities of a period’
disaster process is an essential element in it is measured in hours, days and then years (Haas et al, 1977). Other criticism has been
quantifying disaster events, itself key to (e.g. a log scale). that a view of the disaster process, moving
reducing the complexity of disasters. Disaster stages identify periods in the from pre-disaster to disaster and out-of-
Without models, efforts at quantificaiton unfolding of a disaster. Each stage serves disaster conditions, incorrectly supposes a
have no base from which to organise the to classify the nature of impacts that occur separation between disasters and non-
data collected. or actions which take place to address disaster (i.e. development) periods (see
Fourth, a written disaster process model impacts of a disaster. Events are interlinked, DHA, 1995).
helps establish a common base of under- as are actions, and events and actions. To address the complex relation between
standing for all involved. Whether con- These linkages can be tight or loose. They disasters and development a circular
sciously or not, everyone uses some kind are often the most difficult part of a model was proposed (see Figure 1).
of model as a guide for decisions and disaster to identify, define, and understand. The realisation that disaster phases and
actions. These models are developed from, Finally, the idea of defining stages in a development were often linked and could
and colored by, experience, formal learning disaster suggests some intrinsic order over exist at the same time led to the proposition
and prejudices. Some of the complexity and time. The reality is that the disorder within that disasters and development exist on a
confusion of a disaster can be attributed to and between the events and actions defines ‘Mobius strip’ (see Figure 2, Anderson,
the different cultures, languages and the existence of disastrous conditions (see 1985). The Mobius strip was proposed by
avocations, and thus the different mental Kiel). To be useful to disaster managers, a Cuny and Beaumont before the 1993 model
models, of those involved in a disaster event. disaster process model must go beyond a in Figure 1. Here it is presented as the stage
A written model serves as a common simple definition of disaster stages and after the simple circular model, which was
ground and makes possible better integ- shed light on (and provide a conceptual also put forward in an earlier form by Cuny
ration of indigenous relief and recovery organisation for) the basic events and and others in 1985 (see Anderson, 1985 and
efforts with outside assistance. An accurate actions which constitute a disaster. Cuny, 1985).

Autumn 1998–99 25
 Although the Mobius model clearly
presents the link and concurrent nature of
disaster and development, the model’s
circular form a basic question: Is disaster Warning
history doomed to continually repeat itself, erge
n cy Disaster Di
prepasaster
running like a dizzy rat around the tread Em ponse redn
ess
res
wheel of a Mobius strip?
A more optimistic approach is to see
disaster and development as linked in an
upward spiral or, pessimistically, in a Recon
structio revent
ion
downward spiral (Kelly, 1998). While a n Disaster p
Development
spiral has the advantage of being able to
show positive or negative change, it lacks Rehabilitation Disaster mitigation
the ability to portray the concurrent nature
Figure 1: Circular model of disaster
of disaster and development, so physically
demonstrated in the Mobius model. can be related to the status of other factor
Although the models presented above all Reconstruction relationships in other phase planes without
attempt to set out an organisational struc- Transition being influenced by intervening planes.
0ture for the disaster process, they are Response This is accomplished by making the x-y
broad statements. These models may help Development coordinates of two phase planes the x or y
Preparation
in understanding the general nature of the Mitigation factor of a new phase plane. Although only
disaster process, but provide little insight Prevention two factors can be compared in each phase
into actual unfolding of a disaster. What is plane, increasingly more complex situ-
needed is a disaster process model which ations can be created by progressively
Figure 2: Cuny-Beaumont Mobius strip model
starts with the basic events and actions of forming phase planes for increasingly
a disaster and serves as a functional tool lines represent two factors and each of the complex pairs of factors (i.e. the product
in guiding the disaster management effort. four quadrants a relationship between the of relating the results of two phase planes).
A model specifically for this purpose is two factors (see Figure 3, Priesmeyer and To create a generic model for disaster
presented in the following section. Cole, 1996). response, a disaster process phase plane
The center of the phase plane represents can be organised with Response/Input
Reconsidering the form of no change. Changes in the x and y factors along the y axis and Event/Impact along the
a disaster process model can be plotted over time and used to x axis (see Figure 4).
In reconsidering the form for a disaster monitor the status and progress of a The upper right quadrant represents
process model several points need to be disaster response. (Priesmeyer and Cole conditions when both inputs and impacts
taken into account. First, rapid change is discuss phase planes and their use as are increasing. The lower right quadrant
probably the single greatest contributing disaster management tools in detail.) represents conditions when inputs are
factor in events becoming disasters. Rapid The phase plane concept has three falling behind impacts. The upper left
change also adds to the demands on the advantages for disaster process modeling. quadrant indicates conditions when inputs
disaster manager. Simplifying how change First, each phase plane incorporates three exceed the needs for dealing with an event,
is present is critical to simplifying the factors, ‘x’, ‘y’ and time of the phase, and and the lower left quadrant shows con-
disaster process model. four areas of interaction (quadrants). Other ditions when both inputs and impacts are
Second, underlying the issue of change models generally cover only one area/ diminishing.
is the nonlinear nature—the chaos—of a quadrant and normally only two compon- The normal course of a disaster moves
disaster (Keil, 1996). This chaos ‘ . . . looks ents, ‘y’ and time (usually the ‘x’ axis.) from the lower right quadrant to the upper
like random behavior but is really unstable Second, phase planes can be sandwiched. right, then to the upper left and finally to
behavior over time that stays within clear Theoretically, an unlimited number of the lower left. An efficient response should
boundaries’. Importantly, chaos is probably planes, representing all basic events of a spend little time in the lower right and
a ‘ . . . necessary and desirable condition disaster, can be centered along a single axis, upper left, and move as quickly as possible
which accommodates adaptations, cross- representing no change in status (the from the upper right to the lower left
communications... and other such emer- juncture of the x and y). quadrants. A preemptive response (or
gent behavior essential to an efficient Third, phase planes are only represent- mitigation activity) would be indicated by
response’ (Priesmeyer and Cole, 1996). ations of reality. Thus, the position (x-y an increase in input in the upper right
Finally, as Haas et al. suggest, dividing coordinate) of two factors in a phase plane quadrant with little rightward movement
disasters into stages is only useful in a
general sense (see quote above). It is the Change in resource Y Resource/inputs Y
actual events of a disaster which are
Change in resource X

Quadrant 2 Quadrant 1
important and need to be the focus of
Event/impacts X

Exchange: Increasing Intensity: Decreasing impact Increasing impact

disaster management efforts: it does not More Y and less More of both Excessive inputs and inputs
X needed resources needed
matter what disaster stage you are in if you
don’t have a handle on what is happening. Subsidence: Exchange:
Less of both More X and less Decreasing impacts Increasing impact
One approach to disaster process model- resources needed Y needed and inputs Insufficient inputs
ing which takes these points into account Quadrant 3 Quadrant 4
is to use a two-dimensional Cartesian
plane. On this ‘phase plane,’ the ‘x’ and ‘y’ Figure 3: Resource needs presented on a phase plane Figure 4: Disaster process phase plane model

26 Australian Journal of Emergency Management

along the impact axis, followed by a shift • identifying all the relevant input/impact the basic intent to develop a model to make
into the lower left quadrant as the threat of associations it easier to deal with the non-linear com-
a event diminished. Movement to the left • establishing a system for collecting, plexity at the core of a disaster.
on the plane indicates a reduction in the quantifying and presenting data References
level of activity related to the specific • defining the complex interactions of Anderson M.B. 1985, ‘A reconsideration of
impact/input combination covered by the phase plane factors which exist in a the linkages between disasters and devel-
phase plane. disaster. opment’, Disasters, Harvard Supplement.
Going a step further, phase plane activ- The current capacities of relational data Cuny F.C. 1985, ‘What has to be done to
ities (i.e. the type of input in response to a base programs, as well as the possibility of increase the effectiveness of disaster inter-
specific type of event) can be organised managing phase planes through a geo- ventions’, Disasters, Harvard Supplement.
according to the general stages used to graphic information system approach, Cuny F.C. 1993, ‘Introduction to Disaster
describe a disaster process (see previous suggest that the development of an operat- Management: Lesson 2–Concepts and
sub-section). The locations of the input/ ional disaster process model is less of a Terms in Disaster Management’, in Pre-
impact relationship on the phase planes in technical than a sweat-of-the-brow prob- Hospital and Disaster Medicine, 8 (1).
these groupings (left or right of the y axis) lem. The fundamental organisation of the DHA (Department of Humanitarian
indicates whether an associated stage of a model is simple. The challenge is to detail Assistance) 1995, DHA News, 14, May–June.
disaster is of increasing (moving right on out the complexity of a disaster so it can Frerks G.E., Kliest T.J., Kirkby S.J., Emmel
the plane) or deceasing (moving to the left be represented in the simple x/y param- N.D., O’Keefe P. and Convery I. 1995,
of the plane) importance. This in turn helps eters of the model. Correspondence, Disasters, 19 (4).
the disaster manager identity and define Haas J.E., Kates R.W. and Bowden M.J.
what progress is being made in dealing Conclusion 1977, Reconstruction Following Disaster,
with the disaster. Reducing the complexity of disasters is key The MIT Press, Cambridge.
Phase plans can be reorganised as to reducing their negative impacts on Harrald J.R. and Stoddard L. 1998,‘Scen-
needed to present the range of activities society. This paper provides a generic ario Based Identification and Structuring
which are taking place at any time period model for the disaster process which can of Information Needs for the Response to
in a disaster. In this way, the phase plane help reduce the complexity of disaster and Complex International Crisis’, in Disaster
groupings enable a flexible definition of the also handle the non-linear nature of and Emergency Management: Internation-
stages of a disaster (Haas et al. 1977) disaster events. The model is more focused al Challenge for the Next Decade, proceed-
It is important to highlight that the centre on practical disaster management needs ings of the International Emergency Man-
point represents a static, but not necessarily than other disaster models. The strength agement Society Conference 1998, IEMS/
a positive, situation. Just because there is of the model lies in an ability to help the George Washington University, Washington.
no change does not mean there is no disaster manager or researcher define and Keil L.D. 1996, ‘Chaos Theory and Disas-
problem, only that it has not gotten worse understanding the relationship between ter Response Management: Lessons for
(or better). Since disasters are character- inputs and impacts starting from the Managing Periods of Extreme Instability’,
ised by constant change, this type of static comparison of two factors, rather than in G.A. Koehler (ed.), What Disaster Res-
situation is probably rare and only applic- beginning with the imprecise disaster stage ponse Management Can Learn from Chaos
able to a limited number of phase planes classification approach. Theory, California Research Bureau, Cali-
(i.e. input/impact relationships) at any one The immediate benefits of the disaster fornia State Library, Sacramento.
time. process model approach come from its Kelly C. 1996, Correspondence, Disasters,
Turning the phase plane-based disaster simplicity and ability to indicate trends in 20 (3).
process model into a practical tool for the disaster response process. Using the Kelly C. 1998, Correspondence, Disasters,
disaster managers is both simple and model as a formal decision support tool 22 (2).
complex. The simplicity comes from the requires developing a comprehensive data Macrea J., Badbury M., Jaspars S., John-
ease with which the plane-and-quadrant base of disaster impacts and input require- son D. and Duffield M. 1997, ‘Conflict, the
format presents the core challenge in ments, managed through a computer- Continuum and Chronic Emergencies: A
dealing with a disaster: the provision of based manipulation and analysis system. Critical Analysis of the Scope for Linking
sufficient inputs to reduce impact. The challenge in developing this system is Relief, Rehabilitation and Development
The phase plane model does this without in defining and organising the disaster Planning in Sudan’, Disasters, 21 (3).
regard for questions of response stage, the impact and input relationships. Manage- Neal D.N. 1997, ‘Reconsidering the
nature of specific inputs or perceived ment of the resulting data is feasible with Phases of Disaster’, International Journal
severity of an event. This simplicity and available computer programs. of Mass Emergencies and Disasters, 15 (2).
versatility in handling many different Moving from the simple phase plane Priesmeyer H.R. and Cole E. 1996, ‘Non-
aspects of a disaster has advantages in both approach of the basic model to a highly linear Analysis of Disaster Response Data’,
developing consensus among professionals automated process may not be totally in G.A. Koehler (ed.), What Disaster Res-
on the management of a disaster and in necessary. The basic model approach need ponse Management Can Learn from Chaos
explaining the evolution of a disaster to be implemented only for those events/ Theory, California Research Bureau, Cali-
non-professionals. impacts and related resource/input re- fornia State Library, Sacramento.
The complexity develops in trying to quirements which are identified as being Charles Kelly (MA, Geography) is a disaster management
build the phase plane model into a stand- most critical to the success of the disaster consultant who has worked on international disaster
alone system useful to disaster managers management operation. assistance in Africa, the Caribbean, Central Asia, Eastern
in developing specific plans and decisions. A minimalist approach to using the Europe and Indonesia for 18 years. In his free time he is
an independent scholar of disaster management and has
While the phase plane approach can easily model can provide useful information had numerous articles published in journals and
identify changes in disaster conditions, the without cumbersome, complex and time conference proceedings. He can be reached at email:
real complexity exists in three areas: consuming procedures. This fits well with 72734.2412@ Compuserve.com.

Autumn 1998–99 27