On the Proper Conceptualisation of the Warning, …...rainfall-derived riverine flooding which makes...

On the Proper Conceptualisation of the Warning, Evacuation and Community Education Tasks in theContext of Planning for Dam Failure

ANCOLD 2002 Conference on Dams

ON THE PROPER CONCEPTUALISATION OF THE WARNING,EVACUATION AND COMMUNITY EDUCATION TASKS IN THE

CONTEXT OF PLANNING FOR DAM FAILURE

Chas Keys1 and Steve Opper2

ABSTRACT

As the legislated ‘combat agency’ for dealing with floods, the NSW State Emergency Service has hadconsiderable experience in planning for flooding on the state’s rivers and in developing arrangements to helpkeep people safe when floods occur. This experience has been put to use over the past decade in the particularcontext of managing floods caused or exacerbated by dam failure. Some of the complexities of the dam-failureplanning problem are explored in this paper, specifically as they relate to warning and evacuation tasks and tothe issue of preparing communities for the extreme flooding which dam failure can be expected to cause. Thepoints are made that warning is not just about mechanical alerting devices, evacuation is not restricted tocommanding people to move, and public education requires a sensitive comprehension of the problems ofdisseminating information about rare and difficult-to-believe events.

1 Deputy Director General, NSW State Emergency Service2 State Planning Coordinator, NSW State Emergency Service.

1.0 INTRODUCTION

In New South Wales the State EmergencyService, as the ‘combat agency’ for themanagement of floods according to theState Disaster Plan (New South WalesGovernment, 2000, 27), has been activelyplanning with dam owners for the lastdecade to deal with the potential impact ofdam-failure flooding on downstreamcommunities. In that time a close workingrelationship has evolved between the SES,the state’s Dams Safety Committee andvarious dam owners, and arrangementshave been developed and refined to governthe notification of emerging problemswhich could lead to dam failure, to definethe needed inputs to the emergencyplanning process and to codify the scopeand content of the actual planning whichmust be carried out. These matters havebeen canvassed in a number ofcontributions to conferences of theAustralian National Committee on Large

Dams and published in the ANCOLDBulletin (see, for example, Haines, 1996;Keys, 1997a, 1997b). Moreover planningto warn and evacuate people below damswhich have been declared to be ‘deficient’and at some risk of failure has been carriedout for communities below more than 30such structures. This paper seeks to castlight on some of the complexities of theplanning work which the SES hasundertaken with dam owners and to build adeeper understanding in the damsmanagement community of the nature ofthe warning and evacuation tasks whichmust be carried out when dam failurebecomes possible or actually occurs. It alsodeals briefly with the question of engagingthe people who live in areas below damsand who would be affected by dam-failureflooding.

Much of what is reported here for thecontext of potential dam-failure floodsreflects the realities of flood management

On the Proper Conceptualisation of the Warning, Evacuation and Community Education Tasksin the Context of Planning for Dam Failure


more generally. Thus the SES’s experienceover many years in managing the tasks ofwarning, evacuation and public educationin relation to ‘normal’ riverine floods, oftenin environments in which very largepopulations are dangerously exposed to theflood threat as in the case of theHawkesbury-Nepean River on Sydney’snorth-western outskirts (see Gillespie et al.,2002), can inform the planning for floods inthe special and unusual circumstanceswhich dam failures would represent.Naturally, there must be a recognition inthis of the uniqueness of the dam failureflooding problem in terms of time framesof flood onset downstream, rates of riverrise, velocity of flow and severitygenerally. In all these contexts dam failurefloods are for obvious reasons particularlyserious and pose special problems.Nevertheless, the same principles apply asfor the management of floods in generaland the SES as combat agency must dealwith the range of flood threats – not justrainfall-derived riverine flooding whichmakes up the vast bulk of the floodexperience in NSW (see New South WalesGovernment, 2001, 2-3 for a discussion ofthe range of flood types which can occur).

2.0 WARNING OFPOTENTIAL DAM-FAILURE FLOODING

Warning of an impending emergency is acomplex matter which involves far morethan the mere imparting of informationabout what is or may be about to happen.There are many issues to consider,including what should actually be said andhow messages should be conveyed to thecommunity at risk. On the latter issue, adistinction is commonly made in theliterature of emergency warning between‘alerting’ and ‘notifying’ people about whatis likely to happen (Rogers and Sorenson,1988). Alerting refers in this context to thesudden bringing of a situation to people’sattention, usually by the use of noise to“break a routine acoustic environment to

cue people to seek additional information”(Molino et al, 2002, 55). Sirens arecommonly used to do this, sometimes inassociation with flashing lights. Noises andlights can perform only part of the warningtask, however, the remainder being madeup of the notification element whichexplains what the problem actually is andwhat people should do about it so as toremain safe or protect their property.

True warning of a pending emergencyrequires people to be both alerted andnotified. There is a tendency, however, forsome people who are required to developwarning systems to rely heavily on ‘bellsand whistles’ alerting devices to do all ormost of the job. It is easy to over-emphasise the technology-focussed,essentially engineering-based approach,and this is not uncommon. The problemswhich arise are several. They includefailure at a technical level of the equipmentinvolved because the machinery breaksdown, batteries are not charged or powerboards are not secure, or because sirenscannot be heard as a result of inappropriatecalibration (for example, the sound beingdrowned out by high winds or heavy rain).False alarms also occur as a consequence ofmechanical malfunction. Other problemsrelate to community members not acceptingthe presence of mechanical devices –perhaps because they do not accept thereality of the potential problem which theyaddress – and switching off or vandalisingthem. In addition there are the difficultieswhich inevitably arise when people, havingheard the noise, fail to understand what itmeans. In this circumstance the commonresponse is for people to do nothing exceptto become annoyed by theuncomprehended and unwelcomeinterruption: this is akin to the reaction ofmany to the sound of burglar alarms goingoff in cars when in the vast majority ofcases there is clearly no burglar involved.Complacency as well as anger isengendered in such situations.

Failure of ‘alerting-only’ systems can occurat both the mechanical and the humanlevels, then. Only with information-based



notification can these problems beovercome, which means that any alertingnoise must be followed by words. Thenature of the emergency must be explainedin the real time of its pending or actualoccurrence and people must be empoweredby advice to do what is necessary tomaximise their personal safety and protecttheir belongings. In other words,comprehension must be built, and thisrequires explanatory words which arecarefully crafted to ensure thatunderstanding is enhanced and people aremotivated to act appropriately. The inputshave to do with clear and completecommunication, which implies a need for apsychological dimension to be appreciatedin the context of persuading people to actand also reassuring them that the problemis being managed and that help is available.Relevant information must be incorporatedin the messages and expressed in wayswhich are designed to create particularresponses.

The key to effective warning is that peopleare alerted, are brought to an understandingof what is happening and are guided toappropriate actions in the appropriate timeframe. Not all of this is likely to beoptimally achieved solely as the emergencyis unfolding, of course: the ideal is to havesome prior understanding inculcated duringan educational phase. We return to thistheme later in the paper.

One approach which combines the alertingand notifying elements of warning whichhas been used during floods in NSW(though not yet in relation to potential damfailure) involves the Standard EmergencyWarning Signal (SEWS). This is a noisewhich was devised for use in northernAustralia to warn of approaching tropicalcyclones but is now available for use inimpending emergencies more generally: thenoise is played for several seconds overradio or television stations and isimmediately followed by words whichexplain the situation and advise appropriatebehaviour on the part of those who need toact. The use of the signal is incorporatedinto the dam-failure planning which is

being carried out in NSW. Radio stationshave the signal and arrangements have beendevised to guide its use when dam failurethreatens.

Considerable thought needs to be given towhat is actually said in notifying peopleabout potential dam-failure flooding andthe means by which the information is to betransmitted. In NSW the SES has come tothe conclusion that for normal riverineflooding it is difficult to ensure thatwarning messages can be made complete interms of necessary content and strike theappropriate notes of persuasion andreassurance if they are devised solely as anactual flood is rising. At such a time thereare simply too many tasks for emergencymanagers to deal with and messageconstruction is too complex a function toperform satisfactorily ‘on the run’ under thecircumstances which then apply. It is muchbetter for warning messages to be puttogether slowly and carefully in ‘planningtime’ – assuming, of course, that those whoprepare the messages understand in somedetail and beforehand what theconsequences of flooding of differingseverities will be and what sorts ofresponses will be required by people indifferent areas and with differing exposuresto the flood risk. In NSW the SES hasexpended much effort on the developmentof ‘flood intelligence’, and there is nowsufficient such intelligence on most of thestate’s rivers to sustain the development ofeffective warning messages for floods of allseverities up at least to flood-of-recordlevels. Naturally, the messages which areconstructed must be capable of real-timemodification so that content which couldnot be anticipated in planning time can bequickly incorporated as the flood unfolds.

The SES has embarked upon a program ofcreating pre-written flood warningmessages, designed to be broadcast overradio stations but also intended to providevehicles for other means of notification, forthe reference areas of all 160 gauges in thestate for which the Commonwealth Bureauof Meteorology provides flood heightpredictions when floods are developing.



These messages are being devised for arange of heights, beginning with thedesignated ‘minor flood’ level (which isintended to connote the onset of floodingcapable of causing significant communityimpact in terms of the inundation of low-lying land and the closure of local roads)and going up to heights at least equivalentto those reached in floods of record on therelevant river reaches. The messages arebased on a growing bank of intelligenceabout the consequences of flooding whichhas been compiled largely from real-timeobservations and from consultants’floodplain risk management reports. Itcovers impacts such as the closure of roads,the inundation of farm land, the flooding ofresidential, commercial and industrial areasand the overtopping of protective levees.Pre-written messages allow real meaning tobe incorporated in the communication andpermit appropriately arresting andpersuasive language to be devised: inessence their prior preparation encourages afine-tuning of content and tone and helpsensure that important information is notmissed out (a real problem when all of themessage formulation is done when floodsare actually occurring). All of this shouldimprove the quality of the communicationand enhance the likelihood that necessaryactions will be stimulated on the part ofthose who are likely to be affected by thecoming flood.

Dam-failure flooding is, obviously, muchrarer of occurrence than normal riverineflooding. Indeed it is virtually non-existentand there can be no expectation thatmanagers will be able to learn byexperience – let alone repetitive experience– how to construct appropriate warningmessages about it. That being the case itwill be no less necessary than with normalfloods for pre-written messages to beconstructed in anticipation of dam failure.Fortunately, the information which damowners provide on the likely ‘reach’ ofsuch floods in terms of envelopes ofinundation can be used to create thesemessages which can then be employed,with modification to incorporate on-the-day

information if necessary, when dam failurebecomes possible.

There remains the question of techniques ofmessage dissemination to consider. Thereare now many alerting, notifying and dualalert-notification techniques available (seeMolino et al, 2002), and the principlesrelating to selection and utilisation ofmethods are well known (EmergencyManagement Australia, 1999a, 43-53).Broadly these principles relate to the needfor methods to be chosen to suit the timeframes available, to ensure a variety ofmodes of transmission is used in a layeredfashion (for the purpose of creatingredundancy and to reach the maximumnumber of people within the relevant area),and to incorporate personal notification (forexample by doorknocking or personaltelephone call) if at all possible.

These latter methods are time-consuming toimplement on a time-taken-per-household-notified basis, and they are obviouslydifficult to implement when the timeavailable before possible dam failure isshort. They have the great advantage,however, of being able to deliver credibilitymerely because of the human dimensionthey incorporate. Their chief advantage isthis very characteristic, which reinforcesthe notion that warning strategies require ahuman element in addition to whatevermechanical devices are to be employed.

Within the compass of these principles, theactual choice of techniques to be used indifferent environments will depend on whatis practicable. The time available toevacuate is an important mediating factor inthe dam-failure context: the techniques tobe used to warn will not be the same forhouses just below a small council damwhich is structurally deficient and couldfail in a severe earthquake as for areas along way downstream of a dam whichcould fail only in a very large inflow floodwhich exceeded spillway capacity. In thefirst of these cases, in fact, real-timenotification as defined here might beimpossible in some instances: in thosesituations, a reliance on alerting devices



alone would be justifiable, but there wouldneed to be a special effort to ensure thatpeople’s comprehension of the devices waskept alive and current. The question of theuse of time in the context of evacuation isconsidered further below.

3.0 EVACUATING PEOPLEFROM HARM

Just as warning ideally is not a mattersimply of bells and whistles, evacuation isnot solely about ordering people to movefrom one place to another even though thelegal authority exists for emergencyofficers in NSW (including SESControllers) to force people to evacuatewhen they are at severe risk. Apart from theneed in practice to persuade people tomove, which relates to warning processes,there are complex questions about theactual movement process and itsmanagement. A vital element of planningto evacuate people in the face of impendingflooding is the effective utilisation of time.Careful consideration of the time variable isnecessary to ensure that the evacuation ofall those at risk can actually be effected andthat it can be done with as much quality aspossible – for example, by allowing peopleto save some belongings and to get to aplace of some comfort (a nearby town, forinstance, where accommodation and otherneeds can be met) rather than simplygetting to the nearest location which isabove the likely reach of the flood.

The SES has developed a method ofconceptualising the use of time in themanagement of warning and evacuationprocesses during periods of flooding(Emergency Management Australia, 1999b,33-34). Figure 1 illustrates the method bycompartmentalising the evacuation processinto its constituent phases and representingit in diagrammatic form. The horizontalaxis represents the time variable. Forsimplicity the various stages are depicted asbeing sequential, although in actuality thereis likely to be considerable overlap betweenadjacent stages. The model is describedbelow in relation to both riverine and dam-failure flooding and in relation to the

agencies responsible in NSW. In otherstates the allocation of responsibilitiesdiffers somewhat.

3.1 FLOOD PREDICTION

This element is represented by P in thediagram and is usually managed by theCommonwealth Bureau of Meteorology inrelation to riverine floods. In the dam-failure context the dam owner is normallyresponsible for forecasting the flooding.Factors that can influence P include:

• Data collection (methods, hardware,transmission);

• Flood modelling capability (data,software);

• Human resources (staff availability andactivation time);

• Requirements for the dissemination ofthe prediction;

• Weather forecasting capability (severeweather advice, Flood Watches).

3.2 RESPONSE INITIATION ANDMOBILISATION

This component is represented by R in thediagram and in the context of dam failure isnormally the responsibility of emergencymanagers downstream (in NSW, the SES asthe combat agency for flooding but with thepolice and other emergency services alsoinvolved). The time needed to initiateemergency response operations is acomplex mix of organisational,communications, procedural and legislativefactors. Factors that can influence Rinclude the effectiveness of:

• Emergency response planning;

• Training of personnel;

• Communications systems and methods;

• Exercising activation and the deliveryof specific response procedures.



3.3 WARNING DELIVERY

This component is represented by W in thediagram, and again is usually the

On the Proper Conceptualisation of the Warning, Evacuation and Community Education Tasks in the Context of Planning for Dam Failure




responsibility of the SES (although wherethe time for warning is very short the taskmay by agreement be managed by the damowner using alerting technologies).Warning preparation and delivery shouldbe expected to represent an element forwhich significant time is required. Themost reliable warning method is door-to-door delivery by emergency servicepersonnel which can be quantified inrelation to time taken. In floods many othermethods of warning may used includingradio, TV, sirens and telephones but thetime frames required for the application ofthese methods cannot easily be assessed.

Prior public education is fundamental to thesuccess of any warning strategy. A publiceducation program using a variety ofmethods must be implemented to develop acommunity which is ready for warnings aswell as for the floods they herald.

3.4 EVACUATION OPERATIONS

Again the primary responsibility in NSWbelongs to the SES when evacuations arerequired because of flooding. The keyelements of the evacuation operation arerepresented by the terms Ea, En, L and Sin the diagram. The evacuation operationresulting from a severe flood will usuallybe the most complex and time-consumingaspect of the overall flood operation,especially if the number of people to beevacuated is large.

En represents the total time required tocomplete the evacuation, assuming thereare no impediments to using all of thecalculated (theoretical) necessary time.

L represents the time lost (ie which has tobe subtracted from En) due to flooding orthe closure of evacuation routes or thefailure of transport systems.

Ea represents the actual (ie true) timeavailable within which evacuation eithermust be completed by adapting the processto the time constraints, or will have to beterminated before completion. The portionof evacuation not completed by time te will

result in the need to begin a rescueoperation.

3.5 SAFETY FACTOR

S represents the safety factor or time bufferwhich may be available if the interval Lcan be eliminated – that is, if there is noimpediment to using all of the totalnecessary evacuation time and there isadditional time available to deal withinterruption or uncertainty. Evacuationarrangements must maximise opportunitiesfor a timely response. Factors influencingthe response will include:

• Transport routes (capacity andresilience to damage, vehicle weightand size limits, vulnerability to localflood effects on creeks or mainstreams);

• Transport methods (private vehicles,public transport, special vehicles forexample for medical contingencies);

• Traffic control and coordination(priority setting of traffic lights, manualcontrol of intersections);

• Removal of personal effects, pets andlivestock from properties before orduring evacuation;

• Destination facilities for safe storage ofpersonal effects, pets and livestock;

• Marshalling and processing facilitiesfor evacuees;

• Accident-handling arrangements onevacuation routes.

3.6 SOME COMMENTS IN TERMSOF APPLICATION TO DAM-FAILURE FLOODING

Put in the context of the potential failure ofa dam, Prediction Time (P) refers to thetime from first concern that the dam couldfail in a developing event to the point atwhich the possibility is sufficiently credibleto require the evacuation of peopledownstream of it. During this period the



relevant emergency managers must benotified so that arrangements can be dustedoff and actions instituted as quickly aspossible should they become necessary:this will create benefits in terms ofshortening the response mobilisation phase.This phase (R) can only be very short if theplanning has already been done and theresponse managers are aware and trained,have appropriate communications systemsand methods ready and are able to activateand deliver specific response proceduresspeedily. It could be almost instantaneousif those who were assessing the potentialfailure were the same people who wereresponsible for mobilising these responseactions (in other words, employees of thedam owner). The warning phase (W),likewise, could only be speedily completedif the warnings were to be delivered solelyby mechanical means, presumablyactivated by the dam owner. As notedabove, this is not the preferred situation butthere will be occasions when time is veryshort (particularly if dam failure is alreadyoccurring) and when no additional methodscan realistically be applied.

In most cases, however, these three phaseswill take a considerable amount of time tocarry out. Just how time-consuming theprocess is likely to be can for someelements be quantified: the SES has, forexample, tested the doorknocking processand found that in urban areas an average offive minutes should be allowed perdwelling for a doorknocking team of twopeople. The figure would be higher in ruralareas because of the greater distancesbetween dwellings. Obviously the timetaken to doorknock all relevant dwellingscan be modified, if the time is short, byincreasing the number of doorknockers andthe SES routinely incorporates in itsplanning for this function the availability ofmembers of other organisations such as theRural Fire Service. Real instances ofevacuation operations in recent years haveshown how time-consuming and labour-intensive the task actually is: itsconsumption of these resources is easilyunderestimated. At Grafton in March 2001,when the SES sought to evacuate some

12,000 people on the rising limb of a floodwhich was predicted to overtop the town’sprotective levees, the time available (aboutsix hours) was not enough to allow all thedwellings to be doorknocked.

The remainder of Figure 1 deals with theactual movement of evacuees. The factorswhich are central to the process are the timeneeded to complete the evacuation, aquantity which can be estimated usingempirical data relating to numbers ofvehicles and roadway capacities (600vehicles per lane per hour is a commonlyused, well tested and appropriatelyconservative planning figure) and the timewhich is likely to be available before theevacuation can no longer proceed (forexample because low points on evacuationroutes are flooded or bottlenecks occurwhere routes converge). Clearly when Ea(time available) is less than En (timeneeded) there is a serious problem, and thisis likely to apply in the context of dam-failure flooding when flood levels will risevery quickly and the potential for laterrescue activities to be successfullyundertaken is likely to be minimal.

There remains in all this conceptualisationof the elements a great deal of uncertaintyabout the speed with which large-scaleevacuations can be carried out. Even withdata on the numbers of doorknockersrequired and the numbers of vehicles thatcan pass a given point on an evacuationroute there are many unmeasurables.Experience shows, however, that the earlyphases of an evacuation operation are likelyto proceed slowly given that people usuallyunder-react and delay their responses in thehope that the situation will improve andevacuation will not be necessary. This isespecially the case if there are no floodcues that people can actually see in theirown locations, or if there is an inherent lackof understanding of or trust in the warningsystems and processes that have beendeveloped. For dam failure, the likelihoodof such a lack of comprehension and trustcan be expected to be particularly highgiven the rarity of dam-failure flooding. Allthis militates, incidentally, against the



stricture that evacuation operations shouldbe initiated early if it is known that timewill be short. Great stress is also placed onthe need to ensure that people understandthe problem and the warning systembeforehand.

This time-line approach provides a valuablemeans of conceptualising the problem oflarge-scale evacuations (and, indeed,smaller-scale ones where time is short), andit will help the SES to better planevacuations necessitated by potential dam-failure events. The obvious cautionary noterelates to the fact that its application almostalways suggests that evacuation operationstake longer than one would intuitivelyexpect to be the case and it is not alwayseasy to ‘find’ extra time when it is needed.In the dam-failure context, for example,one cannot buy time simply by beginningthe evacuation before dam failure hasbecome a credible possibility. Especiallyfor small dams which could fail relativelyquickly in a severe flood event this mightlead to frequent evacuations which wouldturn out not to be necessary. The result islikely to be that a greater emphasis willhave to be placed on mechanical warningdevices, which in turn puts stress on theneed to ensure that people understand andaccept their purpose and will act when thedevices are activated.

4.0 EDUCATING AT-RISKCOMMUNITIESABOUT POSSIBLEDAM FAILURE

It is an axiom in emergency managementcircles that warning and evacuating people,while never simple, are at their leastdifficult if the people have a sound priorunderstanding of the threat which requiresthem to evacuate. Ideally thisunderstanding is best inculcated if there hasbeen recent experience of the threat.Clearly in Australia such comprehensioncannot be expected to exist in the dam-failure context since there have been few ifany cases of significant dam-failureflooding in living memory. The next best

thing is to create the understandingsynthetically – that is, by educating peopleabout the threat, about its management bydam owners and emergency responders,and about the actions which people cantake to respond in their own interests. Not agreat deal of this work has yet been done inNSW, and until it is done the situation isthat evacuation will be sought from peoplewho are unlikely to understand the need forit. This remains a significant challenge forthose who are responsible for themanagement of potential dam-failureflooding.

Numerous problems need to be faced here.Amongst them are the fact that large damsrarely fail, which means that dam failurelacks credibility in the community mind.When the topic is raised many people tendto react either scornfully, believing thatfailure will not happen and that somebodyis mischievously seeking to create ‘panic’,or angrily when they recognise that they,their families and their properties are atsome risk. These things are difficult tohandle in the public domain. So are actionsto protect dams in a safety context, asCowan (1998) has reported in the contextof the management of the deficiencies ofHume Dam by large-scale emergencyreleases of water (which were not expectedby farmers and others downstream) inOctober 1996. In this instance thecommunity reaction was overwhelminglyof disapproval and hostility.

It is important in any communityengagement program that delay ininforming people about the potential dam-failure problem be minimised. Otherwisesuspicion will grow about the reasons forthe lack of communication, the seeds ofmistrust will be sown and they will infectlater efforts to provide information aboutthe problem. Unfortunately it takes time fordam owners and emergency managers tocome to grips with the issue and to developstrategies for handling it and it would notbe wise to publicise the problem withoutbeing able to reassure people thatappropriate strategies are at least beingdevised. That said, doing things which are



visible (like installing warning systems)before there is any public consultation or anopportunity for people to come to gripswith the issue and have some input to itsmanagement will also create hostility.There was recently a case in NSW where asiren-based alarm system was switched off,probably by a community member who hadaccess to it and who was lobbied by othercommunity members angered by the factthat it had gone off for some hours andcreated a false-alarm situation. In this caseseveral people living nearby had beenhostile from the start about the sirensystem, and the public education that wascarried out in a series of public meetingsapparently did not fully defuse theirhostility. No real understanding andacceptance of the system’s purpose wasengendered.

This was a case, incidentally, in which thedam owner had taken the lead in themanagement of the issue of educating themembers of the community and theemergency management (SES)involvement had lagged behind. It is vitalthat the interests of both the dam owner andthe emergency manager are fullyincorporated and develop synchronously,just as it is vital that there is a genuineeffort to ensure that the wishes ofcommunity members are considered in thedevelopment of the management strategieswhich are adopted. In this day and age,consultative processes which are seen to betop-down in their nature are likely to beundermined in the community and will failas a consequence.

In NSW the SES and dam owners haveconducted public meetings in a number ofareas at risk to explain the local dam-failureissue, to indicate what is being done tomake the particular dam safe and to seekguidance on appropriate warning messagesand strategies. At one of these meetings, inthe Orange area, a specially-createdFloodSafe guide was distributed as it alsowas shortly afterwards to residents who hadnot attended the meeting (Figure 2). Theopportunity was also taken to encouragepeople to devise their own private written

plans for managing dam failure should itoccur: pro forma documents weredistributed and explained and individualswere helped to define appropriate actionsfor their own situations with regard toproperty protection and evacuation. Indoing this, the opportunity was taken toeducate people about warning processes sothey would be able to understandbeforehand the triggers to the actions theyhave planned.

These meetings need to be undertaken withconsiderable care and sensitivity. There aredelicate balances to be achieved betweenthe provision of advice to individuals andthe reception of information from them,and hostility is often close to the surfaceespecially if, for example, people havebought their land since the dam ownerbecame aware of the dam’s deficiency butbefore the deficiency was publicised.Equally, though, people can be angered onbeing advised that failure is extremelyunlikely: in this instance, some may arguethat the publicity is about events that willalmost certainly not happen but willnevertheless damage them in the prospectby reducing the value of their property.

There are no easy ways out of thesedifficulties. It is worth stressing publically,however, that the discovery of deficienciesshould not be taken as a simple matter offault on the part of the dam owner, sincethe problem is very widespread – naturallythis reassurance is best not given by anemployee of the owner – and that action isbeing taken to maximise community safety.What will be difficult, of course, is thesituation of having to keep people aware ofthe potential for dam failure over the longterm if engineering works are not to beundertaken to rectify the deficiency. So farthis issue has not been tackled, but it isobviously preferable from an emergencymanagement perspective to have warningsystems and evacuation arrangements asstrictly interim measures while the spillwayis being enlarged, the dam post-tensionedor other engineering fixes are beingcompleted.



Figure 2: An Example of a FloodSafe Guide (Orange)



5.0 CONCLUSION

The SES has had several decadesexperience of managing floods and,especially over the past ten years, growinginvolvement in formally planning for them.Improved warning systems and processeshave been developed for riverine flooding,experience has been gained in theconceptualisation of evacuation operationsand in their management in numerous partsof NSW, and there is an increasing focuson the engagement of flood pronecommunities to discuss the threat and itsmanagement. Progress in these areas istransferable to and informs the work whichis being done for communities below damswhich have been identified as having somerisk of failure. Although there is still a vastamount of work to do before all thenecessary warning and evacuationstrategies have been fully devised andimplemented and all the communitiesbelow the deficient dams have been (andcontinue to be) engaged, much has beendone.

There are two cautionary notes whichshould be sounded, however. The firstrelates to the critical importance ofcommunity education. Almost certainly thisis the most difficult area of all to manageand to date it is the one which has beengiven the least attention. Further effort isrequired here. The second is that it remainsvital that the efforts of dam owners andemergency managers are properlyintegrated so that the two sides of themanagement of potential dam failures arefully incorporated. It is not best practice forwarning systems to be instituted solely byemployees of dam owners and withoutemergency management expertise beingsought: when this happens, the likelihood isthat the warning processes will beincomplete and unbalanced. Similarly, damowners who have not had experience withcommunity attitudes to flood problems arenot in the best position to defuse thevarious tensions which almost inevitablycome to the fore during the publiceducation phase. It is equally the case thatthe communication with the people at risk

should not be left to the emergencymanagers alone since they are not wellplaced to discuss the technical issuesrelating to the deficiencies of dams or to theremedial works which are proposed.

The problems to be faced are multi-dimensional and multi-disciplinary, andthey will not be optimally handled withouta genuine partnership between the twogroups of interests responsible for theirsolution. This can create difficulties ofcoordination, of course, when thetimetables of relevant personnel in differentorganisations do not gel. The solution is tomake sure they do gel rather than for oneorganisation to become impatient and dothe work which should be shared by both.

The over-riding message is that it is not allengineering, and neither is it allpsychology. There must be a consciouscombining of the different types of relevantexpertise held in different cultures if thebest results are to be achieved.

6.0 REFERENCES

Cowan, M. (1998) Hume Dam: the DamOwner and the Community 1995-1997 – aRetrospective View, ANCOLD Bulletin,No. 110, 55-64.

Emergency Management Australia (1999a)Flood Warning, Australian EmergencyManuals Series, Part III, EmergencyManagement Practice, Vol. 3, Guide 5, 2nded.

Emergency Management Australia (1999b)Flood Preparedness, AustralianEmergency Manuals Series, Part III,Emergency Management Practice, Vol. 3,Guide 4.

Gillespie, C., Grech, P. and Bewsher, D.(2002) Reconciling development withflood risks: the Hawkesbury-Nepeandilemma, The Australian Journal ofEmergency Management, Vol. 17, No.1,27-32.



Haines, R.C. (1996) EmergencyManagement Plan and Deficient Dams:International Best Practice or a Home-Grown Approach? ANCOLD Bulletin,No. 102, 79-85.

Keys, C. (1997a) Planning for Cases ofPotential Dam Failure: an Audit ofProgress in New South Wales, ANCOLDBulletin, No. 106, 71-82.

Keys, C. (1997b) Planning to Warn andEvacuate Communities PotentiallyThreatened by Dam-Failure Flooding:some Problems Revealed by the Experiencein New South Wales, ANCOLD Bulletin,No. 107, 63-67.

Molino, S., Begg, G., Stewart, L andOpper, S. (2002) Bells and whistles, beltsand braces: designing on integrated floodwarning system for the HawkesburyNepean Valley (Part 1: bells and whistles,

available technologies), The AustralianJournal of Emergency Management, Vol.17, No. 1, 55-59.

New South Wales Government (2000) NewSouth Wales State Disaster Plan(DISPLAN), State EmergencyManagement Committee.

New South Wales Government (2001)New South Wales State Flood Plan, aSub-Plan of the New South Wales StateDisaster Plan (DISPLAN), StateEmergency Management Committee.

Rogers, G. and Sorenson, J. (1988)Diffusion of Emergency Warning:Comparing Empirical and SimulationResults, paper presented to a meeting of theSociety for Risk Analysis, Washington DC.

Presented at the ANCOLD 2002 Conference on Dams, Glenelg.

On the Proper Conceptualisation of the Warning, …...rainfall-derived riverine flooding which makes...

Documents

Transcript of On the Proper Conceptualisation of the Warning, …...rainfall-derived riverine flooding which makes...