Flood warning is a non-structural flood mitigation technique that aims to advise persons of an approaching flood. In the commercial sector context flood warnings aim to allow persons adequate time to evacuate staff and lift or remove stock and equipment.
The effectiveness of flood warnings nationally and internationally has come under increased debate over recent years. Literature originating from the Flood Hazard Research Centre of Middlesex University has criticized flood warnings on the basis that they “often don’t work well and too frequently fail completely” (Handmer, 2000, p. 1). The basis of criticism lies in the conclusions that response rates are much lower than previously estimated, few persons receive official warnings and warning recipients are often dissatisfied (Penning-Rowsell et al, 2000).
Flooding of the Kempsey central business district (CBD) in March of 2001, has provided a terrific opportunity to assess the effectiveness of flood warning systems in a commercial setting.
Kempsey is a small rural city located upon the Macleay River in the New South Wales Mid-North Coast region. Boasting a population of 10,000 residents, Kempsey’s CBD consists of approximately 200 businesses. Kempsey has been no stranger to flooding in the past, recording 21 major floods since 1838. The largest recorded in 1949 resulting in an estimated $200 million dollars damage and six deaths (Dutton, 2000). Structural flood protection is provided by a series of levees that protect the CBD up to an Average Recurrence Interval 10 year flood.
An ALERT1 radio-telemetered flood warning network is in place for Kempsey. The system aims to provide residents 24 hours notice of an approaching major flood, however, for floods produced by heavy rainfall in the lower half of the catchment, typically only 15 hours will be available. The network consists of 12 automatic rain gauges and 12 automatic stream level gauges placed across the 11,500 square kilometre catchment. Additional data are available from an extra 7 rain gauges and 11 river gauges located predominately in the Armidale area (McKay, 2002). The Bureau of Meteorology (BOM), local council and State Emergency Service (SES) monitor the system at all times. Flood prediction is the responsibility of the BOM and is performed through the use of hydrological models. The SES and local council interpret the data. Local warning broadcast and flood response is the responsibility of the SES.
1 ALERT is the acronym for Automatic Local Evaluation in Real Time
Kempsey 2001 Flood
Kempsey’s commercial district was flooded on Saturday the 10th of March 2001. Flooding was a result of an intense sub-tropical low-pressure system located offshore in the Tasman Sea. Heavy rainfall predominately over the lower catchment saw the river peak on Saturday night, recording a height of 6.9 metres at the Kempsey traffic bridge. The height was the 9th largest recorded with an Average Recurrence Interval of 12 years.
The most recent flood prior to 2001 occurred in 1963; as a result flood experience and preparedness in 2001 were low. Survey results indicated that 25% of business managers had previously experienced flooding, of which many described flooding encountered as minor.
Flood Warning Effectiveness
The effectiveness of flood warnings can be evaluated by the extent of warning coverage, avoided losses, warning system failures and recipient satisfaction (Parker and Neal, 1990). Data used to evaluate warning effectiveness was collected through surveys of 88 businesses damaged during the flood.
Extent of warning coverage
Warning coverage within the CBD was variable; some businesses received more than one day’s notice, whilst others received only a few hours (Table 1). The majority of businesses received greater than 20 hours warning. The critical warning of levees being overtopped was issued by the BOM 18 hours in advance.
Warnings were delivered by emergency services through door knocks and the media. Door knocks concentrated on warning businesses in Smith Street, many of which were not flooded. Personal official warnings were received by only 40% of businesses. Local radio was stated as the most effective source of information about the flooding as shown in Figure 1. Responses stating family, friends, neighbours and personal observations as effective sources of flood information indicate the existence of an informal warning system in Kempsey. The informal system involves the use of rules of thumb to predict flood heights using information from up river.
The Internet has recently emerged as a method of delivering warnings to the public. In Australia this information is presented on the BOM website, www.bom.gov.au. Numerous local councils including Kempsey are developing their own flood websites to deliver local flood information. During the 2001flood few businesses attempted to access the BOM site. Little awareness of the site was evident with the site receiving no mention in local flood education literature. Internet access was also limited with 48% of Australian businesses not having Internet access (ABS, 2001). Internet access was further reduced during the flood due to the loss of electricity and phone services.
Once warned of the flood, 80% of businesses attempted to validate warnings, either by listening to the radio, contacting emergency services, talking to neighbors, or observing river levels. Validation enabled managers to more appropriately assess the flood risk posed to their business.
Avoided losses are calculated by estimating the actual damage to potential damage ratio. Actual damage refers to losses that occurred during the flood event and potential damage the loss that may have resulted if no flood preparations were made.
The actual damage to potential damage ratio was calculated at 0.44, indicating that 56% of direct damage was averted through warning and emergency preparations. A maximum estimate of savings attributed to official warnings is estimated at $3.2 million. It is likely that the real amount is less than this figure since the majority of businesses maintain that they did not receive official warning. Savings may therefore be attributed to informal warnings as well as other measures such as pre-flood planning and flood education.
Savings tended to increase as warning time increased (Table 2). The initial four hour warning period was associated with the greatest savings, supporting Parker (1991) who stated that the majority of possible damage savings could be made within four hours. After four hours savings appeared to diminish as warning time increased, signifying diminishing marginal returns. The existence of diminishing marginal returns places doubt over the cost effectiveness of measures to increase potential warning time, since it is unlikely that increases will lead to substantial savings.
Warning system failures
Ten percent of businesses failed to make flood preparations even though in the majority of cases the warning time received was adequate. These businesses commonly lacked flood experience and awareness, resulting in poor flood preparedness. As a result such businesses were unable to identify the flood risk. A further 10% of businesses made ineffective flood preparations. Ineffective response was the result of underestimation of flood depths, limited space, short warning times, confusion and heavy or semi-fixed equipment which had to be lifted or removed.
Three river level gauges including Kempsey and one rainfall gauge failed. The Kempsey river gauge failed as a result of a bank slump, causing the gauge to register artificially high river levels (McKay, 2002). Predictions were forced upwards to reflect the recorded levels resulting in predicted heights much greater than actual heights.
Businesses expressed unhappiness about warnings, with 85% of respondents indicating dissatisfaction. Complaints are listed in Table 3. Businesses were most unsatisfied about not personally receiving warning from emergency services. Managers appeared to have an expectation that they should be individually warned by emergency services. Many businesses that received no official warning were located in the most hazardous areas.
Complaints about limited warning time were received from a variety of businesses ranging from those who only received several hours warning to others who received over 24 hours. Businesses tended to use only a fraction of the warning time available to them, on average taking only seven hours to prepare stores before evacuating. This meant that some businesses evacuated earlier than needed, whilst others conducted preparations after taking considerable time to evaluate the warnings received.
Confusion over warnings resulted from conflicting and inconsistent reports often communicated through informal channels. Confusion also occurred when persons were unable to relate gauge heights to their business. This problem is of significance since the underestimation of flood depth can lead to the escalation of tangible and intangible damages. Some businesses were able to overcome this problem by relating forecast depths to flood markers representing 1949 flood heights, while others contacted emergency services for further information.
Warnings delivered by the media were criticized for containing inappropriate information. Such warnings often contained information about flooding in Grafton and other areas, tending to neglect Kempsey.
Warnings were criticized for inaccuracy, with predicted flood levels perceived as being higher than actual levels. Inaccuracy was caused by the malfunctioning of the Kempsey gauge resulting in predictions up to 0.9 metres higher than actual levels.
Concern was issued about the lack of local input into flood predictions. Numerous businesses suggested that experienced local gauge readers should be used rather than automatic gauges due to their substantial knowledge of local flooding (McKay, 2002).
Where to now?
Despite wide public dissatisfaction the experience of Kempsey has shown that flood warning can be successful. Flood warnings allowed persons to safely evacuate resulting in no deaths and substantial avoided damage. Public dissatisfaction however must be addressed, as it is evident that more is demanded of the present system. Various areas must be given attention as a result.
Public dissatisfaction relates to the poor communication of flood warnings. The public expects that they will receive some form of personal warning from emergency services. Currently this expectation is difficult to fulfill, though emerging technology as discussed by Molino et al. (2001) may create improvements.
Presently radio is the main medium for flood warning dissemination. Increased public education is needed to inform the public of the importance of listening to the radio during times of emergency. The demise of local radio stations to the benefit of growing commercial regional radio networks covering numerous local government areas is a problem. The worry is that large commercial stations may reduce the airtime given to warning messages in favor of other programming.
Door knocks can be effective in delivering personal warnings though are commonly slow and inefficient. Their slow speed means that it is critical that they are focused in the most hazardous areas.
Flood warning response
can be improved through public awareness programs and pre-flood planning.
Individual “Flood Action Plans” combine both awareness and planning creating
a series of actions to be put in place in anticipation of a flood. These plans
allow households or businesses to develop knowledge of how floods will individually
affect them. Plans should be linked to flood warnings, with
It is unlikely that improvements in Kempsey’s potential flood warning time through predictions using forecast rainfall will lead to substantial increases in avoided commercial damage. Furthermore accuracy may be reduced resulting in further decreases in public satisfaction.
The NSW Flood Warning Consultative Committee has already implemented several programs in the aftermath of the 2001 floods including the introduction of river gauge redundancy upstream of key gauges to reduce problems associated with gauge failure. The BOM has also responded to public demands for real-time rainfall and river level information by presenting this information on their website.
In conclusion there are dangers in becoming over critical when evaluating flood mitigation. It is a rare event where everything goes to plan. The key message must be to identify any shortcomings within the system and improve on their failures. If experiences are ignored then disaster is likely to be our destiny.
ABS (2001). Characteristics of Small Business. Australian Bureau of Statistics 8127.0
Dutton, M. (2000). The 1949 Macleay River Flood and its 50th Anniversary Commemoration. Proceedings of the NSW Floodplain Management Authorities 40th Annual Conference, Parramatta NSW.
Handmer, J. (2000). Are Flood Warnings Futile? Risk Communication in Emergencies. The Australasian Journal of Disaster and Trauma Studies, Vol. 2000-2. [Available Online] www.massey.ac.nz/~trauma/issues/2000-2/handmer.htm
McKay, G. (2002). Personal Communication. Commonwealth Bureau of Metrology, Sydney.
McKay, G. (2002). Kempsey – 40 Years of Flood Forecasting. Proceedings of the NSW Floodplain Management Authorities 42nd Annual Conference, Kempsey NSW.
Molino, S., Begg, G. and Opper S. (2001). Bells and Whistles, Belts and Braces. A Preliminary Analysis of New and Emerging Warning Technologies. Proceedings of the NSW Floodplain Management Authorities 41st Annual Conference, Wentworth NSW.
Parker, D. J. and Neal, J. (1990). Evaluating The Performance of Flood Warning Systems. Penning-Rowsell, E. C. and Handmer, J. W. (eds), Hazard and the Communication of Risk, Gower Technical Press, Aldershot, pp. 137-156.
Parker, D. J. (1991). The damage reducing effects of flood warning. Middlesex University Flood Hazard Research Centre, London.
Penning-Rowsell, E. C., Tunstall S. M., Tapsell S. M. and Parker D. J. (2000). The Benefits of Flood Warnings: Real But Elusive, and Politically Significant. Water and Environmental Journal Vol. 14, October 2000.
For further information
please contact: Andrew Gissing