Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, NATURAL HAZARD SCIENCE (naturalhazardscience.oxfordre.com). (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 16 December 2017

Flood Warning Systems and Their Performance

Summary and Keywords

Humankind is becoming increasingly dependent on timely flood warnings. Dependence is being driven by an increasing frequency and intensity of heavy rainfall events, a growing number of disruptive and damaging floods, and rising sea levels associated with climate change. At the same time, the population living in flood-risk areas and the value of urban and rural assets exposed to floods are growing rapidly. Flood warnings are an important means of adapting to growing flood risk and learning to live with it by avoiding damage, loss of life, and injury. Such warnings are increasingly being employed in combination with other flood-risk management measures, including large-scale mobile flood barriers and property-level protection measures.

Given that lives may well depend on effective flood warnings and appropriate warning responses, it is crucial that the warnings perform satisfactorily, particularly by being accurate, reliable, and timely. A sufficiently long warning lead time to allow precautions to be taken and property and people to be moved out of harm’s way is particularly important. However, flood warnings are heavily dependent on the other components of flood forecasting, warning, and response systems of which they are a central part. These other components—flood detection, flood forecasting, warning communication, and warning response—form a system that is characterized as a chain, each link of which depends on the other links for effective outcomes. Inherent weaknesses exist in chainlike processes and are often the basis of warning underperformance when it occurs.

A number of key issues confront those seeking to create and successfully operate flood warning systems, including (1) translating technical flood forecasts into warnings that are readily understandable by the public; (2) taking legal responsibility for warnings and their dissemination; (3) raising flood-risk awareness; (4) designing effective flood warning messages; (5) knowing how best and when to communicate warnings; and (6) addressing uncertainties surrounding flood warnings.

Flood warning science brings together a large body of research findings from a particularly wide range of disciplines ranging from hydrometeorological science to social psychology. In recent decades, major advances have been made in forecasting fluvial and coastal floods. Accurately forecasting pluvial events that cause surface-water floods is at the research frontier, with significant progress being made. Over the same time period, impressive advances in a variety of rapid, personalized communication means has transformed the process of flood warning dissemination. Much is now known about the factors that constrain and aid appropriate flood warning responses both at the individual and at organized, flood emergency response levels, and a range of innovations are being applied to improve response effectiveness. Although the uniqueness of each flood and the inherent unpredictability involved in flood events means that sometimes flood warnings may not perform as expected, flood warning science is helping to minimize these occurrences.

Keywords: flood warning, warning response, flood damage savings, flood forecasting, warning communication, warning chain, warning performance

Introduction

The simple idea of warning people of an impending flood is a beguiling one since actions may then be taken to avoid, or at least reduce, harmful effects. However, making flood warnings perform satisfactorily has proven to be far from simple. This article explores the nature and origins of flood warnings; the development and principal findings of research; key issues in research and practice including performance measurement; and the current state of flood warning research.

A flood warning is information in the form of a prediction about a flood that is likely to happen. This information is usually targeted at and communicated to people who are in the path of the flood in advance of the flood occurring, with the intention of enabling them to avoid harm. Such information may also be communicated to infrastructure providers (e.g., electrical power companies) to enable them to take actions to avoid the disruptive effects of power outages and also to those operating flood barriers. Many population centers (e.g., South-West Netherlands, London, Venice, St. Petersburg, New Orleans, and Providence, Rhode Island) are protected from flooding by flood forecasting and warning systems that operate in conjunction with large-scale flood barriers that are closed when forecast thresholds are reached. Warnings are also provided for professional emergency responders so that they are alerted, for example, to evacuate people from an area likely to be flooded.

Research into flood warnings is situated within a number of wider research areas from which it benefits and to which it contributes. The desire for personal safety and the avoidance of injury, death, and loss underpins all of this research. One of these wider research areas is natural hazards and disasters research because potential for early warnings is common to many natural hazards (Alexander, 1993; Tobin & Montz, 1997). Although beyond the scope of this article, even broader is the body of research on warnings of almost any type, including warnings related to roads, fire risks, hazardous substances, diseases, medications, consumer product safety, and even warning signals in intensive care units, cruise liners and cockpits (Edworthy & Adams, 1996; Wolgater, 2006). Research into flood warnings, and warnings more generally, benefits from a wide range of disciplinary perspectives. Each contributes to our understanding of the design of warning and their performance. In the flood warning field, these disciplines include climatology and hydrometeorological and hydraulic sciences that are primarily concerned with prediction and forecasting from which flood warnings are derived. They also include information and communication technology (ICT) and risk communication sciences that are concerned with the transmission of warnings (Wachter & Uslander, 2014). Decision making and the psychological/behavioral sciences contribute to our understanding of human information processing, individual differences, and behavioral responses (Fischhoff & Eggers, 2006; Smith-Jackson, 2006). Sociology and organizational science aid our understanding, for example, of group and institutional responses to warnings (Drabek, 1986; Rodriguez et al., 2007). Many other fields of research also contribute, including, for example, engineering and media studies.

Flood warnings are part of a larger system that includes flood detection and forecasting on which it is dependent. This article focuses mainly on flood warning communications, flood warning response, and consequent flood loss reduction. However, because developments in flood detection and forecasting have implications for flood warnings, on occasion it is necessary to refer to these developments.

Flood warnings in system context

Flood warnings are the central component of a larger system of flood detection, forecasting, warning, and response systems (FFWRS) which resemble a chain of components supported by a number of processes (Figure 1).

Flood Warning Systems and Their PerformanceClick to view larger

Figure 1. The flood forecasting, warning, and response system (FFWRS).

The conditions that may lead to a flood are first detected by monitoring meteorological, river, and tidal conditions and meteorological forecasting. Indeed, the first sign of a flood developing may be a severe weather warning rather than a flood warning. Hydrometeorological forecasting is a complex science that links numerical meteorological, hydrological, and hydraulic models (i.e., flood routing) in order to forecast the peak levels that a flood is expected to reach at particular locations and times. It is likely that flood risk will have already been assessed and flood risk zones mapped. One of the most advanced flood forecasting systems is the European Flood Awareness System (EFAS), which became operational in 2012 (Thielen et al., 2009; Demeritt et al., 2013). EFAS employs ensemble weather forecasts that are linked to hydrological models to generate twice daily river flow and flood forecasts. These forecasts produced by the European Centre for Medium-Range Weather Forecasting are provided for European Union national authorities and the European Commission up to ten days in advance of an anticipated flood. These forecasts are progressively corrected by real-time weather and river flow observations.

In the fluvial case, flood forecasters may use precipitation and river flow forecasts in rainfall–runoff and flow routing models in order to forecast river discharges and flood levels. In the coastal flood case, forecasts of atmospheric pressure, wind speed and direction, wave setup, and tidal conditions are used in flood forecast models. These forecasts may be coupled with tropical cyclone forecasts and warnings. Warnings are also required for (1) pluvial or surface-water floods, which are often caused by intense thunderstorms that can generate hazardous flash flooding; (2) groundwater floods; (3) dambreak floods; and (4) other types of floods, for example, tsunamis, glacial outburst flows, debris floods, and mud floods. In each case, forecasts are based on the most significant atmospheric and terrestrial mechanisms and parameters.

Once a flood forecast indicates that a flood warning should be issued, usually a predesigned flood warning message is communicated to a predetermined list of recipients. Typically, these recipients include flood defense officers who close flood barriers, professional emergency responders, infrastructure providers, the media, and the public, including householders and property owners who have registered to receive flood warnings from the flood warning agency. Designing effective flood warnings and communicating them in a timely manner in a compressed time frame so that they generate appropriate responses from recipients is complicated by a host of situational, cognitive, and behavioral variables. However, modern personalized ICT devices have dramatically improved the prospects of flood warnings rapidly reaching those who possess them. Flood risk and flood warning awareness is a basic prerequisite for an effective warning response and so providing pre-flood information and education is very important. Similarly, flood preparedness by individuals and by the emergency services is also crucial. Warning response is the penultimate phase in the FFWRS chain and mainly involves individual property occupants and emergency responders taking timely, appropriate actions to avoid harm.

Although it is not always the case, the FFWRS process should be an iterative one with information flows in both directions as a flood develops and passes. Some of the more advanced FFWRS use community sourcing of observed meteorological and hydrological conditions to adjust forecasts and warnings during flood events. The final component of FFWRS is the post-event review: an important opportunity to learn about the performance of the system so that it may be enhanced through lessons learned. The flood warning component of FFWRS may be perfectly designed and operated, but because flood warning performance depends heavily on the effectiveness of flood detection and forecasting on the one hand, and on warning response on the other, it is crucial to identify those factors in the FFWRS chain that impact flood warning performance.

The FFWRS described above and shown in Figure 1 represents a formal, scientific approach to flood warning that is normally introduced by flood risk management agencies. However, informal, “folk” flood warning systems may evolve where there is no formal system or where formal flood warnings are not well attuned to local needs (Parker & Handmer, 1998). Cases from West Bengal (Schware, 1982) and England (Haggett, 2000) demonstrate that such systems depend on residents measuring rainfall and observing river levels and simple social communication networks for disseminating warnings among neighbors. The lesson to be learned from research on informal flood warnings is that the performance of flood warnings is likely to be most effective when formal, “top-down” and informal, “bottom-up” systems are willingly integrated. This lesson is underpinned by other research demonstrating that the “co-production” of flood knowledge (i.e., using and integrating scientific and lay knowledge) is likely to be the most effective way to address flood hazards (Landstrom et al., 2011).

Flood warning lead time

Flood warning lead time is the amount of time between a warning being received and the onset of flooding or damage at any one location, and it differs from flood forecast lead time (Figure 2).

Flood Warning Systems and Their PerformanceClick to view larger

Figure 2. A simplified and idealized time sequence of flood forecasting, warning, and response illustrating lead times.

Floods may be slow or fast rising, providing different challenges to those who are providing and seeking to warn and respond to warnings. On large rivers such as the Rhine and Danube in Europe, the Mississippi in North America, and the Ganges in India, flood waves take weeks to travel the length of the river system. In these cases, a long forecast and warning lead time (i.e., weeks) are potentially available, allowing the flood and its effects to be tracked downstream, permitting those in its path plenty of time to act to avoid or reduce harm. Even so, because there are limits to damage and disruption avoidance, large floods on these types of rivers can still cause substantial harm. In contrast, in densely urbanized, metropolitan areas with impermeable surfaces, slow-moving or stationary thunderstorms can generate intense rainfall that in a matter of minutes can produce severe surface-water flooding disruptive and hazardous to life. It may only be feasible to provide a very short flood warning lead time (i.e., minutes) which may be insufficient to avoid harm. Fortunately, in these circumstances, severe weather warnings may precede flood warnings, allowing early warning of the threat. Flood warning lead time is a critical factor in flood warning effectiveness, which explains why much of the effort in improving flood warnings has been directed at lengthening lead times.

The origins of flood warnings

The origins of flood warnings lie in early, informal flood forecasting and warning practices on river systems and coasts around the world. Ever since floods have disrupted and damaged human endeavor and caused lives to be lost, there has been an interest in trying to anticipate the next flood to reduce harm and to take advantage of its benefits. More organized, formal attempts to forecast floods and to provide warnings often emerged in the 20th century as a natural extension of meteorological observations in meteorological agencies established in the 19th century. For example, the origin of the UK’s Met Office was an experimental government department established in 1854 to better understand weather conditions for shipping safety purposes. The establishment of Catchment Boards in 1930 stimulated a better understanding of river behavior in the UK. It also led to rudimentary river catchment-based flood forecasting and warning using recorded flow and flood levels and rainfall observations. Further impetus was given by the Second World War, and shortly afterward in 1953 the UK experienced the 1953 North Sea storm flood surge that killed thousands of people. This event led directly to the establishment of a Storm Tide Forecasting Service to provide warnings of similar threats.

In 2009, a Flood Forecasting Centre became operational for the UK. This is a joint venture by the Met Office, the Environment Agency, and its counterparts in Scotland, Wales, and Northern Ireland. The Centre’s flood forecasts underpin all flood warnings in the UK, providing Flood Guidance Statements for up to five days in advance, which represents a major advance in the quality of flood warnings (Environment Agency & Met Office, 2016). There is a long history of hydrometry on some European rivers such as the Rhine and Danube, but in the UK it was not until the 1960s that the river gauging network expanded sufficiently to underpin a credible flood warning service. A similar pattern of evolution occurred in India where the Indian Meteorological Department was established in 1875. Then in 1969 the Government of India created a Central Flood Forecasting Directorate and six interstate river basin flood forecasting divisions charged with flood forecasting and warning.

Responsibilities for public flood warnings emerged somewhat earlier in the United States. The Smithsonian Institution was in regular receipt of weather observations from over 150 volunteers, and in 1870 the National Weather Service was established. By 1891, the Weather Bureau was responsible for issuing flood warnings to the public followed by hurricane warnings in 1898. As in the UK and India, the impetus for flood warnings in the United States was driven by serious floods such as the 1889 earthen dam break and flooding near Johnstown, Pennsylvania.

The development of flood warnings research and key findings

Although they are interconnected, three main strands of knowledge contribute to the development of flood warnings research: (1) risk perception and the social psychology of behavioral response to crises and warnings; (2) the economic effectiveness of flood warnings; and (3) flood warning performance.

Risk perception and the social psychology of behavioral response to crises and warnings

Starting in the 1940s with Gilbert White’s research in the United States, geographers researched alternatives to structural, engineering approaches to floods (White, 1945). In seminal research, White set out a “theoretical range of adjustments” to floods including non-structural measures one of which was flood forecasting and warning (White, 1964). Together with Kates and Burton, White established the influential “Chicago School” of natural hazard geography (Kates, 1962) that led to similar studies in Canada (Burton, 1965), the UK (Harding & Parker, 1974), and elsewhere (White, 1974). The central idea was that people respond to risk through the prism of perception and experience. Although they make rational decisions in the face of hazards such as floods, they have “bounded rationality.” As long as science, engineering, and policy are sound, then society will be safer, especially when those at risk have their cognitive boundaries stretched by public risk education (Burton et al., 1978). Flood hazard perception and decision-making psychology therefore became an important research theme since it was acknowledged that there was no simple “risk-response” process at work in how people perceived and responded to the risk of flooding, including when they receive to flood warnings (Slovic et al., 1974).

The “natural hazards paradigm,” as it came to be known, became challenged by the “vulnerability paradigm” (Cutter, 1996). The hazard and perception focus did nothing to explain the plight of deprived, marginalized, and powerless people, especially in the developing world, and their vulnerability to hazards (Blaikie et al., 1994). For example, in the Philippines there is repeated high loss of life and population displacement in flood disasters. Here, the lack of flood warnings that meet the needs of the poor people of shanty towns has much more to do with a rigid political structure, land ownership that is in the hands of a few wealthy individuals, and the incredibly inequitable distribution of wealth than it has to do with flood risk perception (Mahmud, 2000). Ensuring that flood warnings meet the needs of those at risk of flooding is a very important issue. Even so, the research of White, Burton, and Kates provided an impetus to do research on flood risk awareness and perception which importantly underpins flood warnings.

From the 1970s onward, a wider range of disciplinary perspectives developed. Sociologists at the Disaster Research Center (DRC) at the University of Ohio and elsewhere in the United States focused on the organization of human behavior in emergencies and under stress (Dynes, 1970; Britton, 1988). With some exceptions (McLuckie, 1970, 1974), their initial focus was on processes closely related to warnings such as evacuation behavior but subsequently explicitly encompassed warnings (e.g., Quarantelli & Taylor, 1968; Drabek, 1986; Mileti, 1975; Perry & Mushkatel, 1984; Mileti & Sorenson, 1990; Hamilton et al., 2016). Sociological research into human behavior in crises contributed significantly to our understanding of human warning response. Disbelief in warnings is a common reaction that is most probably followed by behaviors that may confirm, or neutralize, the warning, if indeed there is any response at all (Perry et al., 1981). Research by U.S. geographers also contributed to this understanding. For example, Gruntfest’s research into people’s behavior in the 1976 Big Thompson flood led to a focus on flood warnings (e.g., Gruntfest, 1977; Gruntfest & Huber, 1989), including subsequent research (e.g., Hayden et al., 2007).

Psychologists and sociologists have made important contributions to risk and uncertainty perception and risk communication theory which is highly relevant to flood warnings. The overarching research finding was that although recent experience of being flooded usually leads to greater flood risk awareness and to a higher probability of an appropriate response to a flood warning, sociological and psychological factors condition and complicate human response to warnings. These factors help explain why warning response is often more constrained or limited than would otherwise might be the case. Douglas and Wildavsky (1983) and Beck’s (1992) classic panorama of “risk society” (1992) contributed to an understanding of the cultural framing of risk and risk communication. German researchers, Jungermann et al. (1988), focused on the reporting of risk information by the media and important subjects, including trust and credibility in risk communication, the right to know, and the public perception of risk. How the human mind deals with uncertainty was the subject of Tversky and Kahneman’s (1973, 1974) research on the availability heuristic and biases. The availability heuristic explained why systematic errors occur in people’s quantitative risk estimates and responses. Fischhoff et al. (1998) reported that people experience problems correcting their risk estimates for systematic biases arising from past exposure to risks or even thinking spontaneously about the possibility of such biases.

Morgan et al. (2002) and Renn (2008) are examples of subsequent and more recent psychological and social research on the communication of risk and uncertainty, whereas McCarthy et al. (2007), Faulkner et al. (2007), Parker et al. (2009), Kellens et al. (2013), and Demerrit & Norbert (2014) deal with some of the practical issues in communicating flood risk uncertainty. Italian sociologists have undertaken some particularly insightful case studies of people’s response to the risk and warning of floods (Parker et al., 2008, pp. 58–87). Research into flash fluvial and debris floods in northern Italian villages (De Marchi et al., 2007) revealed a progressive loss of a culture of self-protection among the people, so that they do not respond to flood warnings and consider risk mitigation the sole responsibility of the municipal authorities. Although these findings may apply to some extent in other locations in Europe, they are by no means typical, although lack of awareness of floods and flood warnings and limited warning response remains a widespread issue.

The economic effectiveness of flood warnings

Day’s seminal research on the effectiveness of flood warnings introduced what became known as the “Day curve,” which is the relationship between flood warning lead time measured in hours and the percentage reduction in flood damages for residences—one of the first measures of the performance of flood warnings (Day et al., 1969; Day, 1970) (Figure 3).

Flood Warning Systems and Their PerformanceClick to view larger

Figure 3. The Day Curve

(inferred from Day, 1970).

A civil engineer, Day explored the damage-saving potential of flood warnings coupled with temporary flood proofing and evacuation of residences in the Susquehanna River Basin in New York, Pennsylvania, and Maryland. The “Day curve” and the positive relationship between lead time and damage-savings is one of the foundations of modern FFWRS. One of the principal ways of improving FFWRS is to increase the forecast and warning lead times by scientific and technical improvements in flood forecasting and warning dissemination.

The economic benefits of flood warning systems became of interest to the UK River Authorities in the 1970s (Chatterton & Farrell, 1977; Chatterton et al., 1979). A long series of research projects followed on the economic and social performance of flood warnings funded by the UK’s flood management agencies (e.g., Penning-Rowsell et al., 1978, 1983, 2000; Parker, 1987, 1991; Parker & Neal, 1990). This customer-oriented research revealed that a significant proportion of flood warnings was received too late and after flooding had already commenced. Also, flood warnings often did not reach those targeted. Research on the economic savings generated by flood warnings focused on the warning recipients experienced, as revealed through questionnaire surveys in 50 locations in England and Wales flooded between 2001 and 2004 (Parker et al., 2007a, 2007b). This research also utilized standardized flood damage data (Penning-Rowsell et al., 2013). Further research on the damage-reducing effects of flood warnings crucially demonstrated the economic value of combining flood warnings with warning response pathways, including property-level protection measures and business continuity planning (Parker et al., 2008) (Table 1).

Table 1 Flood warning response pathways.

Warning response pathway

Description

Flood defense operations

Closure of warning-dependent flood barriers and floodgates in order to avoid flood damages

Watercourse capacity maintenance

Ensuring that watercourses are kept clear of debris, etc., so that flood levels can be suppressed, thereby reducing flood damage

Demountable flood defenses

Erecting warning-dependent temporary flood defenses to avoid damage

Property-level resistance and resilience measures

Activation of warning-dependent flood-proofing measures to avoid damage. Also activating business continuity plans.

Moving contents

Moving contents to higher levels or locations to avoid damage

Evacuation

Pre-flood evacuation of people in the path of a flood to save lives and reduce risk of injury

Other researchers developed different assessment methods (Carsell et al., 2004; Schroter et al., 2008; Ball et al., 2012; Wurster & Meissen, 2014). Some of the most insightful research has been undertaken at the GeoForschungsZentrum in Potsdam, based on extensive empirical surveys following the severe floods in the River Elbe and Danube catchments in 2002. This research focused on flood damages, flood preparedness, and damage savings generated by warnings (Kreibich et al., 2005; Thieken et al., 2005; Kreibich et al., 2007; Steinfuhrer & Kuhlicke, 2007; Thieken et al., 2007). Although many businesses and householders did not receive a formal flood warning, this research also demonstrates that flood warnings coupled with precautionary measures produce significant economic benefits.

Flood warning performance

The German research also illuminates ways in which flood warnings may be improved, and many of the insights into flood risk perception and behavioral response also provide valuable lessons. Sorenson (2000) reviewed progress and shortcomings in U.S. hazard warning systems, including flood warnings. The quality and effectiveness of warnings varied greatly between localities; new communication technologies increased the gap between those who could and could not afford them; and there was limited progress in improving integrative connections between the scientific, technical, and social components of flood warning systems. Emergency Management Australia (1999) emphasized the importance of integrated FFWRS by promoting the concept of the Total Flood Warning System, which prioritizes close coordination of all components. The EUROflood research project (1992–1994) provided a unique opportunity to compare the level of development of FFWRS in Germany, France, The Netherlands, Portugal, and the UK and to identify strengths and shortcomings influencing warning performance (Parker et al., 1994; Parker & Fordham, 1996). As a result of severe floods, there has been a drive in the UK to enhance flood warnings partly through government-funded research. Tapsell et al. (2004) evaluated the social performance of flood warning communication technologies, identifying those that performed best for particular vulnerable groups. These findings underscored the desirability of inclusive flood warning communication strategies in the sense of providing warnings by a variety of communication methods accessible to the full range of warning recipient needs. Shaw et al. (2005) recommended methods of enhancing flood risk awareness in low-probability, high-consequence flood risk areas. Fernandez-Bilbao and Twigger-Ross (2009) focused on improved ways of targeting flood warnings, and Kashefi et al. (2009) recommended careful use of probabilistic flood warnings.

Key issues arising from research and practice

Forecasts are not warnings

An issue that has beset many flood warnings is that the hydrometeorologists and hydraulic engineers who forecast floods sometimes not only confusingly refer to them also as flood warnings but genuinely believe that all that is necessary to avoid losses is to disseminate an accurate flood forecast. As a result of this narrow disciplinary thinking, the performance of flood forecasts or warnings has often been measured in purely technical terms: typically as the difference between forecast and actual peak flood levels. Also, as long as the forecast or warning has been passed onto to other official agencies (e.g., municipal or military authorities), it has often been assumed that the task of warning has been completed. In reality, unless forecasts are being passed between those who use the same technical language, it is vital for a technical flood forecast to be translated into a nontechnical flood warning in terms that are readily understandable by those directly at risk. Moreover, the warning must be communicated directly to those at risk in a timely and accessible manner, rather than to intermediaries who may or may not pass the warnings on. Within this process the design of warning messages, including their wording and presentation, is critically important. Flood warning performance may then be measured in social or economic terms: that is, (1) whether or not the warning reached those directly at risk in a timely manner and in comprehensible and useful terms; and (2) the extent to which loss of life and injury was avoided or damage reduced.

These issues hindered the UK’s flood warning systems until the mid-1980s. Flood engineers tended to measure the performance of flood warnings in technical terms and were surprised by the findings of social survey research which revealed that many warnings were not reaching those directly at risk, and often not in a timely manner (Penning-Rowsell et al., 1983). In Pakistan, where engineering approaches still dominate flood risk management, making sure that flood forecasts are translated into warnings that are subsequently issued to the many poor and powerless people who occupy the country’s floodplains remains a major issue (Webster et al., 2011).

How to measure performance

Some of the most common ways of measuring flood warning performance are shown in Table 2. Some are technical measures, and others are social ones requiring social survey responses. To promote public confidence and response, flood forecasts and warnings must be sufficiently accurate to maintain their credibility. The uncertainty (which involves accuracy, reliability, timing, etc.) associated with a flood warning increases as lead times increase. Enhancing forecast accuracy permits more accurate assessment of risk to life and property, leading to increased potential for flood damage avoidance. However, increasing accuracy may lead to flood warning lead times being reduced. Less common performance measurement criteria include geographical coverage (i.e., flood warnings may not be available in some areas) and the degree to which warnings are accessible to vulnerable groups (e.g., minorities whose language may not be the principal one).

Table 2 Common flood warning performance characteristics and measurement parameters.

Characteristic

Measurement parameter

Detection

Probability of detection

Accuracy

Forecast flood levels compared with actual flood levels; the proportion of those who received a warning who were subsequently flooded

Reliability

Flood hit, miss and false alarm rates; or false alarm ratio

Probability (i.e., uncertainty)

Amount or percentage of certainty/uncertainty associated with the forecast

Time range ahead of flood

How far ahead in time a forecast can be made

Timeliness

Forecast lead time; warning lead time; recipients’ assessments of adequacy of lead times

Spatial resolution

The smallest area for which a forecast can be made

Warning information

Recipients’ assessments of the degree to which the warning provided them with the flood information they needed

Satisfaction with flood warning service

Levels of satisfaction among those for whom flood warnings were/should have been provided

Damage reduction

The amount of flood damage saved by the warning

Protection of life and limb

The assessed number of lives and injuries avoided by the warning

Benefit-cost ratio

The ratio of the assessed benefits and costs of providing a flood warning

Legal responsibility for flood warnings

Legal responsibility for flood warnings is sometimes unclear, even when it is evident that legal responsibility may be compromised by insufficient training, inadequate communication skills or capacity, or conflicting roles. For example, although they had no statutory responsibility to do so, until 1996 in England and Wales, the police accepted responsibility for disseminating public flood warnings. However, this was not a police priority, and to broaden police officers’ experience, many authorities frequently moved police officers between duties, a practice that worked against these officers accumulating flood warning experience. Although the police disseminated flood warnings efficiently in some cases, in others the process was far from professional, with warnings being disseminated too late or not at all (Penning-Rowsell et al., 1983). With the situation becoming increasingly unsatisfactory, in 1996 the government gave the Environment Agency legal responsibility for communicating flood warnings in England and Wales to the public, leading to a much more professional approach. In France, mayors’ responsibilities are onerous. They are accountable to their local electorate and, in the case of public safety, to the national government as well. Mayors find that it difficult to discharge the two, sometimes conflicting, public roles, and they are legally responsible for ensuring public flood risk awareness and for efficiently disseminating public flood warnings. Successfully discharging these duties depends on mayors’ personal knowledge, skills, capabilities, and communication capacities. These qualities are known to vary greatly among the many mayors in France, and so, not surprisingly, the quality and effectiveness of flood warnings is most uneven (Parker et al., 2008, p. 105).

Integration of FFWRS

Researchers and practitioners agree that an integrated approach to FFWRS is necessary for satisfactory performance. Yet, all too often integration between the different components of FFWRs (Figure 1) and the principal actors/agencies is lacking (Sorenson, 2000), and so they function sub-optimally. Haggett (2000, p. 264) identified integration factors categorized as institutional, legal, technological, expertise and training, and informational. When successfully applied to FFWRS, these factors could produce an integrated system referred to as the “total flood warning system” in Australia. FFWRS require multidisciplinary and multiprofessional inputs, and it is sometimes challenging for disciplines spanning from meteorology to psychology and sociology to fully understand each other, let alone to work cooperatively. FFWRS also involve a range of organizations that must work cooperatively. For integration to occur, it must be afforded political credibility and legitimization through political commitment. Integration objectives must be established so that the role of each agency in meeting these objectives is defined. Where roles are ill defined, functions are divided between agencies, emergency communication technologies are incompatible, information-sharing is weak, and the organizational culture is competitive rather than cooperative, the potential for conflict or stalemate is great and FFWRS underperformance is highly likely.

Integration issues arise in transboundary river basins where the cooperation of FFWRS agencies is essential. Major intergovernmental and interagency cooperation difficulties have been experienced regarding upstream river flow forecasts for the Brahmaputra, Ganges, and Meghna rivers, which flow from India into Bangladesh. Until recently, this has limited flood warning lead times in Bangladesh to three days rather than the eight days thatis now achievable by the new American SERVIR satellite-based flood forecasting system. Haggett’s (2000) research revealed that few of these integration factors had been adopted in England and Wales prior to 1996 when the Environment Agency was created and given responsibility for flood risk management. Fortunately, since then, many of these integration issues have been acted upon, although until recently the police leadership role in flood emergency command structures has been ambiguous.

How best to raise and maintain flood risk awareness

Public awareness of flood risk and understanding of flood warnings is a prerequisite for adequate response to flood warnings. Even then, unless flood warning recipients know how to respond appropriately, response may be ineffective. The starting point is to raise and then maintain a high level of flood risk awareness. All too often, however, levels of awareness are low, as, for example, in a recent YouGov polling survey in the UK commissioned by the Know Your Flood Risk Campaign. In this survey, 67% of respondents had never checked their home’s level of flood risk. Flood risk awareness is likely to be low especially in locations where flooding is infrequent, people live out of sight of the source of flooding (e.g., a river), and few have experienced flooding.

A range of practical approaches to raising and maintaining flood risk awareness (Table 3) have been proposed. However, the most significant issue concerns the merits of different models of risk communication in which flood risk information and flood warnings may be embedded (Demeritt & Norbert, 2014). The first of these approaches, known as the “risk message model,” is based in transferring sound, unbiased flood risk information from scientists or government to citizens on the basis that they have an “information deficit,” which, when made good, leads to rational response. The “risk instrument model” is about using the lessons from social psychology about people’s risk perception and behavior to influence their risk awareness and response to warnings. So, for example, flood risk information and warning messages tap into the negative emotions associated with being flooded in order to raise awareness and motivate people to respond. Both of these models are one-way communication processes. A third model, the “risk dialogue model,” is based on two-way communication. Viewed as more human and trust-building, this model is more participatory and “democratic” and allows use of local knowledge of a flood risk. Arguably, using this model will be more successful in raising risk awareness and eliciting appropriate warning responses. Finally, the “risk governance model” views risk as a central part of modern life in which there is freedom of choice and responsibility for managing risks is transferred to individuals. This model relies on providing people with access to flood risk information (such as online flood risk maps) and “nudges” (e.g., further flood risk information delivered at local farmers’ markets) and transferring the responsibility of responding to warnings from the state to them. A two-way flood warning communication model seems preferable, whereas the risk governance approach is more theoretical and less practical.

Table 3 Some practical methods for raising and maintaining public flood risk awareness.

Method

Details and examples

Flood risk maps

High-quality maps accessible online for all principal sources of flooding. Users may click the location of their property to find out the flood risk level; e.g., UK—http://watermaps.environment-agency.gov.uk/wiyby/wiyby.aspx?&topic=floodmap#x=357865&y=355121&scale=11

South Australia—https://www.waterconnect.sa.gov.au/Hazard-Management/Flood-Awareness/SitePages/Home.aspx

USA—https://www.fema.gov/risk-map-flood-risk-project-lifecycle

Flood campaigns

Employing social marketing principles to raise flood risk awareness among residential communities, children, small and medium-size businesses etc.; e.g., New Zealand—http://www.flood-aware.com/topics/final_report_activity_2.pdf

Flood advisory and guidance services

Forums for providing users with information on flood risk, how to respond to a flood warning and other ways of reducing their flood risk; e.g.,

UK—http://www.knowyourfloodrisk.co.uk/our-campaign

Flood action groups and networks

Setting up a local flood action group that subsequently undertakes flood risk awareness activities; e.g., UK—https://www.buckscc.gov.uk/media/3321841/floodsmart_final-report.pdf

Participatory learning

Involves using language, stories, songs, and traditions to strengthen flood risk awareness and to develop a culture of responding to warnings, e.g., International— http://www.ifrc.org/Global/Publications/disasters/reducing_risks/302200-Public-awareness-DDR-guide-EN.pdf

Flood science fairs and flood fairs

Online flood science fair, with flood projects and experiments targeted at all school levels; e.g., USA—http://www.juliantrubin.com/fairprojects.html

Fairs to provide public information and advice from government agencies and self-help protection firms, e.g., UK—http://www.fadsdirectory.com/the-national-flood-forum

Flood museums

Flood museums help maintain the collective memory of floods and thereby raise and/or maintain flood risk awareness, e.g., The Netherlands—http://www.watersnoodmuseum.nl/NL/

How best to design flood warning messages

A great deal of research exists on the effects of warning message design on recipients, and numerous recommendations have been set forth on how to optimize their effectiveness (Edworthy & Adams, 1996; Emergency Management Australia, 1999; Bauer et al., 2013). An array of warning types is available, including verbal warnings, labels, danger signs, sirens, bells, and other sounds. Because people do not always comply with warnings, designers of warning messages use language, color, symbols, and the like in their efforts to overcome this reluctance. It is common for severe weather and flood warnings to be graded to indicate the level of risk. For example, the U.S. National Weather Service escalates its coastal flood warnings from a “flood advisory” (minor or nuisance flooding is imminently occurring) to a “flood watch” (moderate or major flooding is possible and could pose a threat to life) and then to “flood warning” (moderate or major flooding is imminent or occurring and will pose a threat to life). A similar tropical storm warning system is employed as such storms may cause coastal flooding. Somewhat controversially, whereas France, the Czech Republic, and most European countries have color-coded warnings (e.g., yellow, orange, and red denoting an escalating risk of flooding), because of inconsistent use, the UK replaced its color-coded flood warnings with symbolic ones. There are t now three levels of warning in England: flood watch, flood warning, and severe flood warning. Each is denoted by an orange or red warning triangle, within which the outline of a house is submerged in increasing levels of floodwater. A short verbal message accompanies each (e.g., Severe flood warning, Danger to life).

The style and wording of flood warning messages need careful consideration. Warning messages should be specific, accurate, consistent, and clear. They should specify the flooding source and say where, when and to what level flooding is expected. Messages are considered to have greater effect on warning response if they also indicate likely impact (e.g., lives may be lost) and provide simple guidance on how to respond. A wide variety of factors is taken into account in designing flood warning messages delivered by the broadcast media, including, for example, in the case of televised weather forecasts which include flood warning advice, the personality, appearance, and apparent trustworthiness of the presenter. A controversial issue is the extent to which flood warning messages should contain an element of fear. Harries (2012) cautions against flood risk communications designed to induce fear to motivate response. His research showed how people think of their homes as intrinsically safe places and suggests that warnings that threaten people’s security risk prompt denial of risk.

How best to disseminate flood warnings

One of the most transforming factors in flood warnings is the widespread application of rapid, personalized, information communication technologies (ICTs), which allow flood warnings to be transmitted directly to those at risk and flood warning lead time to be maximized. Table 4 shows conventional and more advanced communication methods for providing essentially the same warning information. Warnings need to be given in more than one language, and the social performance of each of these methods differs. For example, some warnings are inaccessible to those with hearing or sight difficulties, and it is therefore desirable for flood warnings to be communicated from a single, authoritative source via multiple channels and methods (Parker, 2004), including conventional, “low-tech,” and more advanced methods, especially where there is a risk of power supplies being disrupted by flooding. A key issue surrounding advanced ICTs is that disadvantaged, low-income groups may well not have access to them, and as flood warning communications develop, these groups may be excluded so that they fail to receive flood warnings. Even so, ICTs and social networking are advocated as ways of providing early warning of floods in low-income countries such as Mozambique (Mondlane et al., 2013).

Table 4 Flood warning communication methods.

Conventional

More advanced

Analogue telephone

Radio telephone/VHF

Radio

Facsimile

Sirens

Loudspeakers

Leaflets/letters

Local flood wardens

Water-level monitors

Digital telephone

Cell phone. SMS messaging

Social networking: Facebook, Twitter, Musmnet, etc.

Pagers

Automatic Voice Messaging (AVM)

Television/local radio broadcasts

TV crawlers and teletext

Recorded messages

Dial-and-listen services

E-mail

Electronic file transfer

Internet with real-time warnings

WAP telephones

Integrated multimedia warning services

Home computer/tablet links

Digital Radio Technology supported by internet to wireless devices, e.g., vibrating pillows or fire alarms

When best to issue a flood warning and false warnings

Decide to issue a warning too soon and the chances of certainty about flooding occurring are lower than issuing the warning later when there is greater certainty. However, the longer a warner waits, the shorter is the flood warning lead time. This erodes the ability of warning recipients to respond in time and to avoid loss, and so flood warners risk being criticized for not warning earlier. On the other hand, the earlier a warning is issued, the higher the probability of a “false warning” and possible public and media criticism. This is a classic dilemma—flood warners risk being damned if they do and damned if they don’t.

Providing an escalating scale of flood warnings reduces this dilemma to some extent, as long as those at risk fully appreciate the hierarchy of warnings (say from flood alert to severe flood warning) and the different responses required. Repeated flood warnings not followed by flooding are known in some cases to cause those living in flood-risk areas to lose confidence in them (Parker et al., 2011). However, the likelihood of people responding to a flood warning is not reduced by the “cry-wolf syndrome” (Dow & Cutter, 1998) if the quality of communication between warners and the warned is good and people understand the basis of false warnings (i.e., that they are not necessarily “false” and that because of uncertainty it is almost inevitable that some warnings will not be followed by flooding). Even so repetitive warnings may still reduce response.

How best to communicate uncertainty

With the advent of ensemble forecasting, the evaluation of uncertainty has advanced enormously in the meteorological and hydrological sciences (Demeritt et al., 2010). However, challenges still remain for scientists communicating uncertainty associated with flood warnings to professional responders and to the public. Faulkner et al. (2007) call for a “translation discourse” about such uncertainty to be developed between scientists and professional responders. Evaluations of the merits of issuing probabilistic flood warnings (i.e., warnings in which the percentage chance of flooding is expressed) to the public have been explored by Kashefi et al. (2009). However, Parker et al. (2011) found that, even among professional responders, the meaning of probabilistic flood warnings is sometimes misunderstood. Using probabilistic flood warnings may, however, allow flood warnings to be issued earlier, with some positive benefits.

Flood warnings: futile or effective?

Recent research into the flood warnings based on the forecasts of the continental-scale European Flood Awareness System (EFAS) indicates that these warnings are very effective in reducing monetary damage (Pappenberger et al., 2015). The benefits are estimated to be on the order of 400 euros for every 1 euro invested. Yet in 2000, Australian geographer John Handmer posed the question of whether flood warnings are futile (Handmer, 2000), and Parker and Priest (2012) questioned whether Europe’s flood warnings could ever be effective. As testified by many flood victims, there is little doubt that flood warnings save lives and that lack of warning contributes to deaths. Parker and Priest (2012) report that for the 1990–2010 period the number of recorded floods in Europe increased, whereas the number of recorded fatalities per flood remained stable or declined. This research also shows that, when combined with other damage-reducing measures, such as property-level protection, the damage savings from flood warnings are likely to be significant. However, the doubts expressed about the effectiveness of flood warnings are not so much a commentary on the losses that are undoubtedly avoided. These doubts are more to do with the inherent potential of flood warning chains to break down and to underperform unless they incorporate highly developed, continually tested, and well-maintained learning systems.

The current state of flood warning science

Since 2000, significant advances have been made in flood warning science. During this period, flood risk management, especially flood forecasting and warning, have climbed the political agenda and research funding has increased. Severe flood disasters and rising global flood losses have been key drivers. Exposure to flooding is growing rapidly (Hallegatte et al., 2012) and, with climate change, increased flooding is expected in many regions (Rojas, 2013; Intergovernmental Panel on Climate Change, 2013).

Flood forecasting science is developing rapidly, increasing flood warning lead times. Probabilistic forecasting models, such as hydrological ensemble prediction systems (HEPS), in which a suite of numerical predictions are employed in forecasting, are being integrated with conventional deterministic models (Thielen et al., 2009; Cloke & Pappenberger, 2009). By 2009, most European countries used traditional deterministic flood forecasting models, with just over half also using probabilistic models (World Meteorological Organisation, 2009), but this proportion is increasing year by year. The accuracy of deterministic flood forecasts has been improving about 3% to 5% per year. For example, in Slovakia and the Netherlands, with forecast levels to 0.2 meters accuracy, flood forecasts for medium-sized and large rivers can now be provided three to five days in advance.

The EFAS, which entered operational service in 2012, represents a major advance (Thielen et al., 2009) and not only for Europe. Through incorporating medium-range weather ensemble forecasts into flood predictions generating probabilistic information, EFAS produces twice daily river flow and flood warnings on a continental scale, with lengthier warning lead times. EFAS, which is based at the European Centre for Medium-Range Weather Forecasting (ECMWF) in England, generates information for flood aid preparation for the European Commission and national flood forecasting agencies. EFAS’s performance is measured in terms of (1) warning lead time and (2) rates of hits, false alarms, and misses (Table 1). Flood warning lead time is now up to 10 days. Furthermore, because EFAS is a learning system that accumulates data and experience of flood forecasting and warning, the system performance has improved about 10% to 30% every decade. In the United States, the National Weather Service’s (NWS) 13 regional weather and forecasting models have recently been integrated into a national-scale model using high-performance computing. This has led to hourly, nationwide flood forecasts with longer warning lead times, especially in rural areas. The larger U.S. urban areas often already possess their own flood forecasting capability. The NWS approach has now been adapted for global use, employing forecasting models from ECMWF building on an international project named Global Flood Awareness System (GloFAS), which will help Latin America. The system revealed its flood warning potential during severe floods in Pakistan and Sudan in 2013.

Intense tropical storms (i.e., hurricanes) that make landfall pose a major flood risk across the globe, and so warnings are crucial. Forecasting these storms has two main components: storm track and intensity. Satellite imagery, a move from static statistical models to more advanced dynamic models of storms, and the advent of supercomputers have greatly improved hurricane forecasts. Forecast hurricane track errors have been halved in the past 20 years (Masters, 2010). Twenty-five years ago, hurricane forecast lead times were three days; today they are five days and according to the U.S. Government’s Performance.Gov website, in the foreseeable future they will likely be extended even further. In recent years, a new generation of coastal surge forecasting and warning methods based on wind, wave, and water-level modeling has developed, which has enhanced flood warnings. For example, on Britain’s North Sea coastline, tidal surge warnings now begin to be available as an early warning five or six days ahead (Met Office, undated; Titley, Undated), leading to a useful flood warning lead time of 12 to 24 hours. Meteorological observations and numerical weather forecast models are linked to water level, wave, and surge ensemble models. The effects on beach movement, overtopping, and breaching are then predicted so that the probability of damage to people and property can be assessed prior to a decision to issue a warning.

Flash pluvial and surface-water flood forecasting and warning is more challenging (Pitt, 2008). Flooding from intense rainfall cannot yet be predicted with confidence. Through advances in radar detection and quantitative precipitation forecasting, heavy rainfall forecasting and warning has become almost routine. However, connecting these forecasts with digital terrain, hydrologic, and hydraulic models to produce surface-water flood forecasts and warnings is at the research frontier (Parker et al., 2011). The US Flash Flood Guidance method, based on the approximate threshold basin-average rainfall depth over a given duration that would cause a small stream to begin flooding, has been applied in Europe to enhance the accuracy of flash flood forecasts (Delrieu & Velasco, 2010). Ongoing research is investigating how improvements in radar observations and data assimilation can be exploited to improve precipitation forecasts in order to increase the lead time and accuracy of flood warnings. Researchers in the UK, Italy, and the Netherlands have been experimenting with community sourcing of real-time observation data and cell phone apps that can be used to calibrate forecasts as a flood event develops.

Dramatic advances in computer and communication technology have greatly improved opportunities to provide people with access to flood risk information and to rapidly disseminate flood warnings directly to large numbers of people via automated messaging and social network media. For example, whereas timely flood warning dissemination was patchy in England prior to the mid-1990s, it is no longer the research issue that it was, although communication systems may become overloaded in crises and power outages may cause communication systems to go down. Significant progress has also been made in estimating the flood damage saving potential of flood warnings in combination with warning response measures (Penning-Rowsell et al., 2013; Clarke et al., 2015).

Getting those at risk to respond appropriately to flood warnings has always been a key research theme, particularly in psychological and sociological research. However, it has received fresh interest because it is an area where further progress is needed to ensure that FFWRS perform effectively. The underpinning importance to warning response of flood risk awareness and understanding of flood warning codes, as well as flood preparedness, has been well demonstrated by research. The principles of effective warning message design are now well known, and good practice guides exist (e.g., Martini & De Roo, 2007). Psychological and sociological research has demonstrated the cognitive, social, and cultural processes that condition warning response and, for example, the importance of creating community memory of previous flood disasters through “flood museums” (Table 3). However, flood warnings often do not reach enough people at risk, especially those who are most vulnerable to flooding.

Two examples of recent research that seeks to break down this problem and to improve warning response illustrate the current state of the art. Research on the sociospatial behavior of motorists during flash floods in the Gard department of France has provided some important guidelines on how to reduce the number of fatalities and rescues among motorists (Ruin & Lutoff, 2004; Ruin et al., 2007). Severe weather warning lead times are short, and flood warnings are difficult to provide in this part of France; accordingly, and the number of flood fatalities and rescues among motorists has been high. The research employed cognitive mapping combined with Geographical Information Systems (GIS) data processing to evaluate drivers’ perceptions of flash flood risks on their daily journey itineraries. Drivers have a higher tendency to underestimate risk than to overestimate it, and there is a greater underestimation of risks on secondary roads than on primary ones. These kinds of research findings, coupled with car radio alerts, may in future provide motorists with life-saving warnings. UK research employing value-modes analysis, which identifies people’s social values according to a number of value groups, suggests how carefully constructing warning communications according to group value characteristics may alter their warning response behaviors (Fernandez-Bilbao & Twigger-Ross, 2009).

Research is aiding flood emergency response and management by emergency planners and responders. Innovations include agent-based flood evacuation simulations and IT-based, decision-support systems, incorporating geographical information systems. These innovations provide key visualization and other data during flood events, when time is compressed and circumstances are likely to be dynamic and complex (Mirfenderesk, 2009).

Conclusions

Flood warnings are part of wider FFWRS in which their performance substantially depends on the adequacy of flood forecasting at one end of the warning process and warning response at the other. For warnings to have optimum effect, each system component must work soundly and be connected successfully to others within a constrained time frame. A system with such enchainment and interconnectedness characteristics is inherently at risk of underperforming unless each component is based on sound science and there is a continual process of maintenance, rehearsal, and post-event learning.

Remarkable scientific and technical progress has been made in the last two decades in flood forecasting and flood warning communication. These advances have led to significant improvements in the accuracy of flood warnings and longer flood warning lead times, providing sufficient time for effective flood warning response. Research shows that the potential benefits of flood warnings in terms of avoidable flood losses are appreciable, particularly when flood warnings are employed in combination with a range of warning-dependent measures. Although the time taken to adopt new scientific and technical research innovations varies considerably from continent to continent and from country to country, major progress has been made in continental- and global-scale flood awareness, forecasting, and warning systems. At the same time, significant flood forecasting challenges remain, notably in providing accurate, reliable, and timely surface water flood forecasts and warnings.

Research reveals the potential complexity of the cognitive and social psychological process of public response to warnings. Even so, research has provided clear understandings about the human response to warnings. People behave in response to flood warnings in ways consistent with their situational perceptions of the flood risk. The factors that determine flood risk perception and behavior can readily be divided into sender and receiver factors and the characteristics of the flood warning communication process. Flood warnings that perform most successfully are those that utilize multiple communication methods and channels and where flood warnings can be confirmed by communicating with others and finding further confirmatory information. However, in modern societies in which people are aware of many, almost daily, risks (e.g., health, road traffic, financial) and there is a relatively high degree of residential mobility, floods are not among the most common, direct experiences of people. Therefore, getting them to respond appropriately to a warning of a flood, which they have never experienced before and may never experience again, remains challenging. Establishing FFWRS that meet the needs of millions of poor people at risk of flooding in low-income countries is equally, if not more, challenging.

Although the weight of recent advances in flood warning science suggests that the knowledge now exists to ensure the satisfactory performance of flood warnings, local circumstances are almost always changing. Each flood is characterized by a degree of unpredictability and uniqueness that continues to present significant challenges. For example, in the 2010–2011 floods in Victoria, Australia, flood warnings systems performed much less well than expected (Comrie, 2011). Even societies such as Japan which are well organized and disciplined; and where there is significant experience of major disasters and high levels of flood preparedness, people may not respond to warnings, with unfortunate consequences (Cyranoski, 2011). Flood warnings are always likely to be imperfect in some way, but the promise of flood warning science is to continually reduce the potential for underperformance and failure.

Suggested Readings

Ball, T., Black, A., Ellis, R., Hemsley, L., Hollebrandse, F., Lardet, P., et al. (2012). A new methodology to assess the benefits of flood warning. Journal of Flood Risk Management, 5, 188–202.Find this resource:

Comrie, N. (2011). Review of the 2010–11 flood warnings and response. Melbourne: State of Victoria.Find this resource:

Demeritt, D., & Norbert, S. (2014). Models of best practice in flood risk communication and management. Environmental Hazards, 3(4), 313–328.Find this resource:

Emergency Management Australia (1999). Flood warning (2d ed.) (Part III: Emergency Management Practice, Vol. 3, Guidelines, Guide 5). Dickson, ACT: EMA.Find this resource:

Environment Agency and Met Office (2016). Flood guidance statement user guide, version 3, August.

Fischhoff, B., Riley, D., Kovacs, D. C., & Small, M. (1998). What information belongs in a warning?. Psychology and Marketing, 15(7), 663–686.Find this resource:

Handmer, J. W. (2000). Are flood warnings futile? Risk communication in emergencies. The Australasian Journal of Disaster and Trauma Studies, 2.Find this resource:

Kellens, W., Terpstra, T., & De Maeyer, P. (2013). Perception and communication of flood risks: A systematic review of empirical research. Risk Analysis, 33(1), 24–49.Find this resource:

Parker, D. J. (2004). Designing flood forecasting, warning and response systems from a societal perspective. Meteorologische Zeitschrift, 13(1), 5–11.Find this resource:

Parker, D. J., & Priest, S. J. (2012). The fallibility of flood warning chains: Can Europe’s flood warnings be effective?. Water Resources Management, 26, 2927–2950.Find this resource:

Pitt, M. (2008). Learning lessons from the 2007 floods. (The Pitt Review). London: Cabinet Office.Find this resource:

Wolgater, M. S. (Ed.). (2006). Handbook of warnings. Mahwah, NJ: Lawrence Erlbaum Associates.Find this resource:

References

Alexander, D. (Ed.). (1993). Natural disasters. London: UCL Press.Find this resource:

Ball, T., Black, A., Ellis, R., Hemsley, L., Hollebrandse, F., Lardet, P., et al. (2012). A new methodology to assess the benefits of flood warning. Journal of Flood Risk Management, 5, 188–202.Find this resource:

Bauer, L., Bravo-Lillo, C., Cranor, L., & Fragkaki, E. (2013). Warning design guidelines, CMU-CyLab-13-002. Pittsburgh, PA: CyLab, Carnegie Mellon University.Find this resource:

Beck, U. (1992). Risk society: Towards a new modernity. London: SAGE.Find this resource:

Blaikie, P., Cannon, T., Davis, I., & Wisner, B. (1994). At risk: People’s vulnerability and disasters. London: Routledge.Find this resource:

Britton, N. (1988). Organized behaviour in a disaster: A review essay. International Journal of Mass Emergencies, 6(3), 363–395.Find this resource:

Burton, I. (1965). Flood damage reduction in Canada. Geographical Bulletin, 7, 161–185.Find this resource:

Burton, I., Kates, R. W., & White, G. F. (1978). The environment as hazard. New York: Oxford University Press.Find this resource:

Carsell, B., Pingell, N. D., & Ford, D. T. (2004, August). Quantifying the benefit of a flood warning system. Natural Hazards Review, 5(3), 131–140.Find this resource:

Chatterton, J. B., & Farrell, S. (1977). Nottingham flood warning system: Benefit assessment. Enfield, U.K.: Flood Hazard Research Centre, Middlesex Polytechnic.Find this resource:

Chatterton, J. B., Pirt, J., & Wood, T. R. (1979). The benefits of flood forecasting. Journal of the Institution of Water Engineers and Scientists, 33(3), 237–252.Find this resource:

Clarke, J., McConkey, A., Samuel, C., & Wicks, J. (2015). Quantifying the benefits of flood risk management actions and advice (Flood incident management and property-level responses. Report—SC090039/R Stage 3. Flood and Coastal Erosion Risk Management Research and Development). Bristol, EA.Find this resource:

Cloke, H., & Pappenberger, F. (2009). Ensemble flood forecasting: A review. Journal of Hydrology, 375, 613–626.Find this resource:

Comrie, N. (2011, June). Review of the 2010–11 flood warnings and response. Melbourne: State of Victoria.Find this resource:

Cutter, S. (1996).Vulnerability to environmental hazards. Progress in Human Geography, 20, 529–539.Find this resource:

Cyranoski, D. (2011). Japan faces up to failure of its earthquake preparations. Nature, 471, 556–557.Find this resource:

Day, H. J. (1970). Flood warning benefit evaluation: Susquehanna River Basin (Urban Residences). (ESSA Technical Memorandum WBTM Hydro 10. Silver Spring, MD). U.S. Department of Commerce, Environmental Science Services Administration, Weather Bureau.Find this resource:

Day, H. J., Bugliarello, G., Ho, H. P., & Houghton, V. T. (1969). Evaluation of benefits of a flood warning system. Water Resources Research, 5(5), 937–946.Find this resource:

De Marchi, B., Scolobig, A., Delli Zotti, G., & Del Zotto, M. (2007). Risk construction and social vulnerability in an Italian Alpine Region (FLOODsite Report No. T11-2006-08). Wallingford, U.K.: HR Wallingford.Find this resource:

Delrieu, G., & Velasco, D. (2010). Advancing flash flood forecasting in ungauged basins (Deliverable Report No. 8.1, HYDRATE project, GOCE-2006-03702). Padova, Italy: University of Padova.Find this resource:

Demeritt, D., & Nobert, S. (2014). Models of best practice in flood risk communication and management. Environmental Hazards, 3(4), 313–328.Find this resource:

Demeritt, D., Nobert, S., Cloke, H., & Pappenberger, F. (2010). Challenges in communicating and using ensembles in operational flood forecasting. Meteorological Applications, 17, 209–222.Find this resource:

Demeritt, D., Nobert, S., Cloke, H. L., & Pappenberger, F. (2013). The European Flood Alert System and the communication, perception, and use of ensemble predictions for operational flood risk management. Hydrological Processes, 27, 147–157.Find this resource:

Douglas, M., & Wildavsky, A. (1983). Risk and culture: An essay on the selection of technological and environmental Dangers. San Francisco: University of California Press.Find this resource:

Dow, K., & Cutter, S. (1998). Crying wolf: Repeat responses to hurricane evacuation orders. Coastal Management, 26, 237–252.Find this resource:

Drabek, T. E. (1986). Human system responses to disaster: An inventory of sociological findings. New York: Springer-Verlag.Find this resource:

Dynes, R. (1970). Organized behaviour in disaster. Lexington, MA: Heath Lexington Books.Find this resource:

Edworthy, J., & Adams, A. (1996). Warning design: A research perspective. London: Taylor and Francis.Find this resource:

Emergency Management Australia (1999). Flood warning (2d ed.) (Part III: Emergency Management Practice, Vol. 3, Guidelines, Guide 5). Dickson, ACT: EMA.Find this resource:

Environment Agency and Met Office (2016). Flood guidance statement user guide, version 3, August, http://www.ffc-environment-agency.metoffice.gov.uk/services/FGS_User_Guide.pdf.

Faulkner, H., Parker, D. J., Green, C., & Beven, K. (2007). Developing a translational discourse to communicate uncertainty in flood risk science and the practitioner. Ambio, 36(7), 692–702.Find this resource:

Fernandez-Bilbao, A., & Twigger-Ross, C. (2009). More targeted flood warnings: A review. Improving institutional and social responses to flooding (Science Report SC060019, Work Package 1b). London: Defra, EA.Find this resource:

Fischhoff, B., & Eggers, S. (2006). Mental models of warning decisions: Identifying and addressing information needs. In M. S. Wolgater (Ed.), Handbook of warnings (pp. 279–287). Mahwah, NJ: Lawrence Erlbaum Associates.Find this resource:

Fischhoff, B., Riley, D., Kovacs, D. C., & Small, M. (1998). What information belongs in a warning?. Psychology and Marketing, 15(7), 663–686.Find this resource:

Gruntfest, E. (1977). What people did during the Big Thompson Flood (Working Paper 32). Boulder: University of Colorado Institute of Behavioral Sciences.Find this resource:

Gruntfest, E., & Huber, C. (1989). Status report on flood warning systems in the United States. Environmental Management, 13(3), 279–286.Find this resource:

Haggett, C. M. (2000). An integrated approach to flood warning in England and Wales. PhD diss., Middlesex University.Find this resource:

Hallegatte, S., Green, C. H., Nicholls, R. J., & Corfee-Morlot, J. (2012). Future losses in major coastal cities. Nature Climate Change, 3, 802–806.Find this resource:

Hamilton, B., Liu, B. F, Madden, S., Sutton, J., Wood, M. W., & Mileti, D. S. (2016). Disaster warnings in your pocket: How audiences interpret mobile alerts for an unfamiliar hazard. Journal of Contingencies and Crisis Management, 24(3), 136–147.Find this resource:

Handmer, J. W. (2000). Are flood warnings futile? Risk communication in emergencies. The Australasian Journal of Disaster and Trauma Studies, 2.Find this resource:

Harding, D. M., & Parker, D. J. (1974). Flood hazard at Shrewsbury, UK. In G. F. White (Ed.), Natural hazards local national global (pp. 43–52). London: Oxford University Press.Find this resource:

Harries, T. (2012). The anticipated emotional consequences of adaptive behaviour—impacts on the take-up of household flood-protection measures. Environment and Planning A, 44(3), 649–668.Find this resource:

Hayden, M., Drobot, S., Gruntfest, E., Benight, C., Radil, S., & Barnes, L. (2007). Information sources for flash flood warnings in Denver, CO & Austin, TX. Environmental Hazards, 7, 211–219.Find this resource:

International Panel on Climate Change (2013). The physical science basis. Contribution of Working Group I to the Fifth Assessment Report of the IPCC. Cambridge: Cambridge University Press.Find this resource:

Jungermann, H., Kasperson, R., & Wiedemann, P. (Eds.). (1988). Risk communication. Jülich, Germany: Kernforschungsanlage Jülich.Find this resource:

Kashefi, E., Lumbroso, D., Orr, P., Twigger-Ross, C., Walker, G., & Watson, N. (2009). Communication and dissemination of probabilistic flood warnings (Science project SC070060/SR4. Flood and Coastal Erosion Risk Management Research and Development Programme). London: Defra, EA.Find this resource:

Kates, R. W. (1962). Hazard and choice perception in flood plain management (Research Paper No. 78). Chicago: Department of Geography, University of Chicago.Find this resource:

Kellens, W., Terpstra, T., & De Maeyer, P. (2013). Perception and communication of flood risks: A systematic review of empirical research. Risk Analysis, 33(1), 24–49.Find this resource:

Kreibich, H., Muller, M., Thieken, A. H., & Merz, B. (2007). Flood precaution of companies and their ability to cope with the flood in August 2002 in Saxony, Germany. Water Resources Research, 43(3), 1–15.Find this resource:

Kreibich, H., Thieken, A. H., Petrow, T. H., Muller, M., & Merz, B. (2005). Flood loss reduction of private households due to building precautionary measures—lessons learned from the Elbe flood in August 2002. Natural Hazards and Earth Systems Sciences, 5, 117–126.Find this resource:

Landström, C., Whatmore, S. J., Lane, S. N., Odoni, N. A., Ward, N., & Bradley, S. (2011). Coproducing flood risk knowledge: Redistributing expertise in critical “participatory modelling.” Environment and Planning A, 43(7), 1617–1633.Find this resource:

Mahmud, N. (2000). Flood disasters and population displacement: The case of Ormoc City, The Philippines, In D. J. Parker (Ed.), Floods (Vol. 1, pp. 400–411). London: Routledge.Find this resource:

Martini, F., & De Roo, A. (Eds.). (2007). Good practice for delivering flood-related information to the general public (EXCIFF project). Ispra, Italy: EU Joint Research Centre.Find this resource:

Masters, J. (2010). NHC increases hurricane forecast lead times.

McCarthy, S., Tunstall, S., Parker, D. J., Faulkner, H., & Howe, J. (2007). Risk communication in emergency response to a simulated extreme flood. Environmental Hazards, 7, 179–192.Find this resource:

McLuckie, B. F. (1970). The warning system in disaster situation: A selective analysis (Report Series No. 9). Newark, DE: Disaster Research Centre, University of Delaware.Find this resource:

McLuckie, B. F. (1974). Warning—A call to action. Warning and disaster response: A sociological background. Fort Worth, TX: NOAA, National Weather Service.Find this resource:

Met Office (Undated). 1953 East coast flood—60 years on.

Mileti, D. S., Drabek, T. E., & Haas, J. E. (1975). Human systems in extreme environments: A sociological perspective (Program on Technology, Environment and Man). Boulder, CO: Institute of Behavioral Science, University of Colorado.Find this resource:

Mileti, D. S., & Sorenson, J. S. (1990). Determinants of organizational effectiveness in responding to low probability, catastrophic events. The Columbia Journal of World Business, 22(1), 13–21.Find this resource:

Mirfenderesk, H. (2009). Flood emergency management decision support system on the Gold Coast, Australia. The Australian Journal of Emergency Management, 24(2), 48–58.Find this resource:

Mondlane, A., Hansson, K., Popov, O., & Muianga, X. (2013). ICT for social networking in flood risk and knowledge management strategies—An MCDA approach. International Journal of Computer, Electrical, Automation, Control and Information Engineering, 7(10), 1281–1287.Find this resource:

Morgan, M. G., Fischhoff, B., Bostrom, A., & Atman, C. J. (2002). Risk communication: A mental models approach. Cambridge: Cambridge University Press.Find this resource:

Pappenberger, F., Cloke, H., Parker, D. J., Wetherall, F., Richardson, D., & Thielen, J. (2015). The monetary benefit of early flood warnings in Europe. Environmental Science and Policy, 51, 278–291.Find this resource:

Parker, D. J. (1987). Flood warning dissemination: The British experience. In J. W. Handmer (Ed.), Flood hazard management: British and international perspectives (pp. 169–184). Norwich, U.K.: Geo Books.Find this resource:

Parker, D. J. (1991). The damage-reducing effects of flood warnings (Report prepared for Halcrow and the National Rivers Authority (Anglian Region) Regional Telemetry Scheme Appraisal). Enfield, U.K.: Flood Hazard Research Centre, Middlesex Polytechnic.Find this resource:

Parker, D. J. (2004). Designing flood forecasting, warning and response systems from a societal perspective. Meteorologische Zeitschrift, 13(1), 5–11.Find this resource:

Parker, D. J., Fordham, M., & Torterotot, J-P. (1994). Real-time hazard management: Flood forecasting, warning and reponse. In E. C. Penning-Rowsell & M. Fordham (Eds.), Floods across Europe: Flood hazard assessment, modelling and management (pp. 135–166). London: Middlesex University Press.Find this resource:

Parker, D. J., & Fordham, M. F. (1996). An evaluation of flood forecasting, warning and response systems in the European Union. Water Resources Management, 10, 279–302.Find this resource:

Parker, D. J., & Handmer, J. W. (1998). The role of unofficial flood warning systems. Journal of Contingencies and Crisis Management, 6(1). 45–60.Find this resource:

Parker, D. J., & Neal, J. (1990). Evaluating the performance of flood warning systems. In J. W. Handmer & E. C. Penning-Rowsell (Eds.), Hazards and the communication of risk (pp. 137–156). Aldershot, U.K.: Gower Technical.Find this resource:

Parker, D. J., & Priest, S. J. (2012). The fallibility of flood warning chains: Can Europe’s flood warnings be effective?. Water Resources Management, 26, 2927–2950.Find this resource:

Parker, D. J., Priest, S. J., & McCarthy, S. (2011). Surface water flood warnings requirements and potential in England and Wales. Applied Geography, 31, 891–900.Find this resource:

Parker, D. J., Priest, S. J., Schildt, A., & Handmer, J. W. (2008). Modelling the damage-reducing effects of flood warnings (FLOODsite Report No. T10-07-12). Wallingford, U.K.: HR Wallingford.Find this resource:

Parker, D. J., Priest, S. J., & Tapsell, S. M. (2009). Understanding and enhancing the public’s behavioural response to flood warning information. Meteorological Applications, 16, 103–114.Find this resource:

Parker, D. J., Tapsell, S. M., & McCarthy, S. (2007a). Enhancing the benefits of flood warnings. Natural Hazards, 43(3), 397–414.Find this resource:

Parker, D. J., Tunstall, S. M., & McCarthy, S. (2007b). New insights into the benefits of flood warnings: Results from a household survey in England and Wales. Environmental Hazards, 7(3), 193–210.Find this resource:

Penning-Rowsell, E. C., Chatterton, J. B., & Parker, D. J. (1978). The effect of flood warnings on flood damage reduction. London: Central Water Planning Unit and HMSO.Find this resource:

Penning-Rowsell, E. C., Parker, D. J., Crease, D. J., & Mattison, C. R. (1983). Flood warning dissemination: An evaluation of some current practices in the Severn-Trent Water Authority Area. Enfield, U.K.: Flood Hazard Research Centre, Middlesex Polytechnic.Find this resource:

Penning-Rowsell, E. C., Priest, S. J., Parker, D. J., Moris, J., Tunstall, S., Viavatenne, C., et al. (2013). Flood and coastal erosion management: A manual for economic appraisal. London: Routledge.Find this resource:

Penning-Rowsell, E. C., Tunstall, S. M., Tapsell, S. M., & Parker, D. J. (2000). The benefits of flood warnings: Real but elusive, and politically significant. Journal of the Chartered Institution of Water and Environmental Management, 14(1), 45–50.Find this resource:

Performance.Gov (Undated) Improve Forecast Accuracy and Lead Times for Severe Weather. An Official Website of the US Government.

Perry, R. W., Lindell, M. K., & Greene, M. R. (1981). Evacuation planning in emergency management. Lexington, MA: Lexington Books.Find this resource:

Perry, R. W., & Mushkatel, A. H. (1984). Disaster management: Warning response and community relocation. Westport, CT: Quorom Books.Find this resource:

Pitt, M. (2008). Learning lessons from the 2007 floods. (The Pitt Review). London: Cabinet Office.Find this resource:

Quarantelli, E. L. & Taylor, V. A. (1968). Some views on the warning problem in disasters suggested by sociological research. In Disaster Preparedness Report, Special Issue 1–4. Washington DC: US Department of Commerce.Find this resource:

Renn, O. (2008). Risk governance: Coping with uncertainty in a complex world. London: Earthscan.Find this resource:

Rodriguez, H., Quarantelli, E. L., & Dynes, R. (Eds.). (2007). Handbook of disaster research. New York: Springer.Find this resource:

Rojas, R., Feyen, L., & Watkiss, P. (2013). Climate change and river floods in the European Union: Socio-economic consequences and the costs and benefits of adaptation. Global Environmental Change, 23(6), 1737–1751.Find this resource:

Ruin, I., Gaillard, J-C., & Lutoff, C. (2007). How to get there? Assessing motorists’ flash flood risk perception on daily itineraries. Environmental Hazards, 7(3), 235–244.Find this resource:

Ruin, I., & Lutoff, C. (2004). Vulnerabilite face aux crues rapides et mobilites des population en temps de crise, La Houille Blanche, 6, 114–119.Find this resource:

Schroter, K., Ostrowski, M., Velasco, C., Torres, D. S., Nachtnabel, H. P., Kahl, B., et al. (2008). Effectiveness and efficiency of early warning systems for flash-floods (EWASE) (CRUE Research Report No I-5). London: ERA-NET CRUE.Find this resource:

Schware, R. (1982). Folk wireless: An example of indigenous technology for flood warnings (Food Systems and Society Project Monograph). Geneva: UN Research Institute for Social Development.Find this resource:

Shaw, J., Cudmore, S., Collier, D., Reed, D. W., Antonelli, A., Genna, R., et al. (2005). Improving flood warning awareness in low probability and medium—high consequence flood zones (R&D Technical Report W5-024). London: Defra/EA.Find this resource:

Slovic, P., Kunreuther, H., & White, G. F. (1974). Decision processes, rationality, and adjustment to natural hazards. In G. F. White (Ed.), Natural hazards local national global (pp. 187–205). London: Oxford University Press.Find this resource:

Smith-Jackson, T. L. (2006). Receiver characteristics. In M. S. Wolgater (Ed.), Handbook of warnings (pp. 335–344). Mahwah, NJ: Lawrence Erlbaum Associates.Find this resource:

Sorenson, J. H. (2000). Hazard warning systems: Review of 20 years of progress. Natural Hazards Review, 1(2) (May), 119–125.Find this resource:

Steinfuhrer, A., & Kuhlicke, C. (2007). Social vulnerability and the 2002 flood, Country Report Germany (Mulde River) (FLOODsite Report No. T11-07-08). Wallingford, U.K.: HR Wallingford.Find this resource:

Tapsell, S. M., Burton, R., Parker, D. J., & Oakes, S. (2004). The social performance of flood warning communications technologies (Technical Report W5C-016 to the Environment Agency). Enfield, U.K.: Flood Hazard Research Centre, Middlesex University.Find this resource:

Thieken, A. H., Kreibich, H., Muller, M., & Merz, B. (2007). Coping with floods: Preparedness, response and recovery of flood-affected residents in Germany in 2002. Hydrological Sciences, 52(5), 1016–1037.Find this resource:

Thieken, A. H., Muller, M., Kreibich, H., & Merz, B. (2005). Flood damage and influencing factors: New insights from the August 2002 flood in Germany. Water Resources Research, 41, 1–16.Find this resource:

Thielen, J., Bartholmes, J., Ramos, M.-H., & de Roo, A. (2009). The European Flood Alert System—Part 1: Concept and development. Hydrology and Earth Systems Sciences, 13, 125–140.Find this resource:

Titley, H. (Undated). Ensemble storm surge forecasting in the UK during the exceptional 2013/14 winter.

Tobin, G. A., & Montz, B. E. (1997). Natural hazards: Explanation and integration. New York: Guilford.Find this resource:

Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive Psychology, 5(2), 207–232.Find this resource:

Tversky, A., & Kahneman, D. (1974). Judgement under uncertainty: Heuristics and biases. Science (New Series), 185(4157), 1124–1131.Find this resource:

Wachter, J., & Uslander, T. (2014). The role of information communication technology in the development of early warning systems for geological disasters: The Tsunami showcase. In W. Friedmann & W. Zschau (Eds.), Early warning for geological disasters: Scientific methods and current practice (pp. 227–252). Heidelberg: Springer-Verlag.Find this resource:

Webster, P. J., Toma, V. E., & Kim, H-M. (2011). Were the 2010 Pakistan floods predictable?. Geophysical Research Letters, 38(4).Find this resource:

White, G. F. (1945). Human adjustment to floods. Research Paper No. 29. Chicago: Department of Geography, University of Chicago.Find this resource:

White, G. F. (1964). Choice of adjustment to floods. Research Paper No. 93. Chicago: Department of Geography, University of Chicago.Find this resource:

White, G. F. (1974). Natural hazards: Local, national, global. New York: Oxford University Press.Find this resource:

Wolgater, M. S. (Ed.). (2006). Handbook of warnings. Mahwah, NJ: Lawrence Erlbaum Associates.Find this resource:

World Meteorological Organisation (2009). WMO RAVI working group on hydrology subgroup on flood forecasting and warning (Activity Report 2005–2009. Final Report by Karro, I). Geneva: WMO.Find this resource:

Wurster, S., & Meissen, U. (2014). Towards an economic assessment approach for early warning systems: Improving cost-avoidance calculations with regard to private households. In M. S. Pfaff, L. Plotnick, & P. C. Shih (Eds.), Proceedings of the 11th International ISCRAM Conference. University Park, PA: S. R. Hiltz.Find this resource: