FHRC research on flood warning systems over the years has examined the issue of the extent to which those at risk from flooding rely upon official warning systems in making decisions at the time of flooding. For example, a 1986-7 study in the Upper Severn, Upper Trent and Avon Catchments (Neal and Parker, 1989) showed that farmers in all the catchments were more used to coping with flooding and were less reliant upon warnings systems than those in commercial property and that residents were most reliant on official warnings.
In a survey of residents affected by flooding along the River Thames in 1990, respondents were asked about their ‘need’ for an official warning (Tunstall, 1992). More than half (60%) felt that they needed an official warning, while a third (34%) were prepared to rely upon their own judgement, for the rest it depended upon the circumstances. Responses varied according to area, with residents in rural settlements reporting less need for an official warning than those in urban areas such as Guildford, Woking and Purley. Lack of experience of, and connection with, urban rivers, population mobility and thence lack of community memory of flooding may contribute to the greater reliance on official warning systems in urban areas. The perceived need for an official warning did not differ markedly according to the extent to which the residents in the survey were affected by the recent flood event nor according to whether or not residents had received an official warning although those who had been warned saw more need for an official warning than those not warned. These data are old and public expectations of warning services as well as their provision may have changed markedly in the intervening period.
In the Lower Thames Survey, respondents were asked “How much if at all, do you currently rely on the authorities or your own judgement of when the River Thames is going to flood?”. Only 10% reported that they ‘completely rely on the authorities’, 20% said that they ‘mainly rely on the authorities’, for the largest group (35%) the response was ‘about half and half’, a quarter (26%) ‘mainly rely on their own judgement’ and for very few (7%) the response was ‘completely rely on their own judgement (McCarthy et al., 2006). The Lower Thames area surveyed is an urban area but one with a fairly stable population in which a substantial minority live close by the river and may have knowledge of its behaviour. Furthermore, residents in focus groups pointed out that flood warnings for the area are issued in general terms for large reaches of the river and provide only a very general indication of what may happen on a particular section of the river, obliging residents to exercise some further judgement of their own.
4.5 Summary The data sets covered a very large number of floods with different characteristics (slow onset river floods, flash flooding in steep catchments, and extreme rainfall events), in different locations, and invoking different responses from the responsible organisations. It was hypothesised that specific flood characteristics and the extent of flooding might have an influence on the health effects of flooding and that these factors may be important for the analysis of vulnerability and resilience in Section 6. Those interviewed in the Intangibles Survey were more seriously affected in terms of flood depths than those in the Warnings Survey, with substantial minorities experiencing deep flood waters in their homes. In both surveys, a majority had flood waters in their homes for at least a day. It was also hypothesised that contaminants might add to the adverse impact of flooding on health, and three quarters of Intangibles flooded sample reported the flood waters being contaminated.
Awareness of flood risk may be a factor affecting response to flooding. Prior awareness was reported as highest among the Intangibles at risk sample (86%), followed by the Warnings sample (55%), the Lower Thames sample (30%) and finally the Intangibles flooded sample (24%). However, the questions asked about awareness were different in the different surveys and this may account for the variations. Prior flood experience, tenure and length of residence were significant factors affecting awareness in some of the data sets. Those who had experienced floods inside their homes prior to the most recent event were more likely to be aware of the risk. Those who owned their property outright (usually older and more long-term residents) were significantly more likely to be aware than those still buying their property, and they, in turn, were more likely to have prior awareness of the flood risk than those renting or in other forms of tenure. Awareness differed with length of residence with longer term residents most likely to have prior knowledge of flood risk. The key social variables (gender, age and social grade) which might be expected to have some influence on awareness do not appear to be important. However, there were significant and marked variations in prior awareness according to the specific location and context of where the interviews took place.
There is very little data on flood risk constructions across the three surveys. However, some perceptions of risk were a focus of the Lower Thames Survey and were analysed. Results showed that prior awareness of risk appeared higher among those in higher social grade groups. The risk of future flooding was perceived as higher among the flooded rather than non-flooded. It appears, therefore, that while residents in the area surveyed acknowledge a level of risk, for most of those without flood experience, the level was not associated with an immediate risk to their homes. It was striking that respondents drew on their own experience and judgement or on other informal sources such as neighbours to a large extent rather than on formal sources of information.
River bank location was also a significant factor in risk construction in the Lower Thames, with significant differences in the way those whose property was immediately on the river bank and those who lived further away constructed the risk. The river bank residents were more likely to view the risk of their property being flooded in the future as certain or very likely compared with non-river bank residents. Overall, 81% of those with some current awareness of the risk stated that they would have moved to the area even if they had known of the risk, illustrating the strong attraction of living in the Lower Thames area balanced against the limited knowledge of the risk of flooding and also limited experience and thence possibly poor understanding of the impacts of flooding.
A flood warning had been received at some time in the past by 53% of Lower Thames respondents Only 37% of Warnings respondents and 23% of flooded Intangibles respondents had received a warning in a recent flood event. Those more seriously affected were more likely to have received a flood warning than those with fewer main parts of their home affected, with an association between depth of flooding and receipt of a warning. Moreover, those owning their properties were more likely to have received a warning than those in other tenure groups. For Lower Thames respondents, reliance on warnings was more equally divided between formal and informal sources. In the Lower Thames area as in other parts of the country flood warnings are issued in general terms for large reaches of the river and provide only a very general indication of what may happen on a particular section of the river, thus obliging residents to exercise some further judgement. A key finding is that despite investment in improvements to warning systems, still only a minority of residents receive an official flood warning of any kind. Data from both the Intangibles and the Warnings Surveys also indicate that the receipt of a warning is very area and event specific.
The Environment Agency’s automatic voice messaging (AVM and now FWD) systems are becoming the main source of warnings, whilst unofficial warnings from neighbours, friends and relatives still feature significantly. The location and thence the nature of the flood event and the level and type of warning service available, are the key factors in variations in warning dissemination. Whether or not residents have registered on the automatic voice messaging system is very important and their prior awareness of the flood risk and their past experience of flooding are other factors. Lower social grades and those renting properties appear to be less likely to receive warnings, and vulnerable households appear at present to be no more likely than others to be warned.
Therefore, the data help to illustrate people’s experience of flooding and how their perceptions or constructions of flood risk may be influenced by a number of factors, such as flood event or dwelling type characteristics, organisation or institutional responses to the event and the characteristics and resources of the population affected. These factors may result in individuals and households being more or less vulnerable to the threat and extent of flooding and may affect their ability to recover.
5. Human behaviour before, during and after a flood People’s behaviour before, during and after a flood will be driven by a number of factors such as the flood event itself and individual or household characteristics. These factors will influence the levels of social vulnerability relating to capacity to cope with and recover from a flood event. Evidence of people’s behaviour examined in this section includes:
before the flood event: undertaking preventative measures including insurance;
during the event: actions undertaken to minimise the effects of the flood event on the household (e.g. move furniture upstairs); and
after the event: evacuation, undertaking preventative measures for future events.
In many cases, what happens after a flood can have a worse effect than the event itself. Other factors can also help reduce or increase the effects of a flood. These include:
resources: income, car ownership, insurance, education;
social support: help, social capital; and
intervening factors: problems with builders and insurers, evacuation.
In addition to the above factors, drivers or explanatory factors for people’s behaviour may include:
length of residence, prior flood experience, awareness of flood risk, the event characteristics including the receipt of a warning, people’s characteristics, community support
5.1 Drivers of human behaviour before, during and after a flood The actions that people take during a flood, such as moving furniture or valuable items, will depend to an extent on receiving a flood warning and also on the length of the warning. But with or without warnings, other variables such as having previous experience of flooding and how people construct flood risk (discussed in Section 4) may influence people’s behaviour. People who have prior experience or awareness of flooding will be more likely to know which actions to take or how to obtain help. Personal and household characteristics such as age, presence of children, number of people in the household, and disability are also likely to affect these actions.
Whether respondents or a member of their household left the home after the flood, the length of the evacuation, the time to get back to normal are examples of variables that will show aspects of people’s behaviour after the event. We can hypothesise that these actions will be determined by the characteristics and extent of the flood, i.e. large depths will cause more damages to the property and be the main cause of an evacuation. The type of property will also be key, as people living in ‘vulnerable’ housing will find living in their homes more difficult. The characteristics of the household members, i.e. age, health problems, presence of children, etc., are also likely to influence the decision to evacuate the home.
Preventative measures such as taking out insurance against flooding, keeping sandbags in the property, not buying expensive furniture for ground level rooms or avoiding keeping valuable or irreplaceable objects on the ground floor, can be undertaken before a flood occurs in which case they will be determined by the awareness of the risk or prior flooding experience. These measures can also be undertaken after an event in preparation for future flooding. Again, having experienced a flood will be the main driver for undertaking these sorts of measures as well as being worried about the possibility and consequences of a future event. These drivers of behaviour are discussed in the following sections.
5.2 Flood preparedness and preventative actions This section addresses the issue of the actions taken by residents in flood risk areas in advance of flood events such as registering on the AVM and obtaining insurance cover, and examines the drivers for taking preventative action. It draws mainly on the Warnings data set which included a number of questions on preparedness, some of which were only included in Phase 2 of the survey. The Intangibles data set only provides responses on insurance.
5.2.1 Automatic Voice Messaging system (AVM)
A key preparatory action that those in flood risk areas in England and Wales can take is to register with the Environment Agency to receive flood warning messages via the Agency’s national automatic voice messaging (AVM) system Floodline Warnings Direct. Prior to the national system, which has only recently become active, property owners in flood risk areas were recruited to the warning system by staff in the Agency’s various area offices and there were variations in the methods and success of local recruitment.
Questions about registering on the AVM were only asked in the Warnings Survey. In Phase 1 of the Warnings Survey, information was collected on those people on the AVM system at the time of the interview. In Phase 2 on the other hand, the questions differentiated between those on the system at the time of the last/worst flood and those who had registered since then. Only 28% (70) of the 278 residents interviewed in Phase 2 had taken the precaution of signing up to register on the AVM system before the recent/worst flood. Among this group, those who had experienced more than the one flood inside at their current address were more likely to have registered than those only affected once (39% compared with 21%). Those who reported that they were aware of the flood risk before the recent flooding were also significantly more likely be on the AVM system than those unaware (37% compared with 18%).
There were no differences in the AVM take up in relation to length of residence, age, tenure, social grade or income. Thus, flood risk awareness and experience were the only variables considered that accounted for registration before the last/worst flood in the Warnings Survey. Other reasons, which were not possible to examine in the research, have been suggested for the failure of those in flood risk areas to sign on to receive warnings:
a reluctance to mark property as ‘at risk’ for fear of an effect on the property’s value;
a belief that the flood risk at the property is low;
a reluctance to accept the disturbance of warning messages, some of which might be received at night, and may not always result in flooding; and
lack of knowledge that the system is available in an area.
Phase 2 of the Warnings Survey shows the importance of the experience of flooding in overcoming inertia or reluctance to sign onto the system. As many as 24% (60 households) signed up for the AVM after the flood and only three households that were on the system at the time of the last/worst flood left the system. Thus at the time of the interview, a majority (52%) of those interviewed were on the system. There were few other factors that appeared to explain why some residents decided to register whilst others did not. There were no differences in AVM uptake according to social grade, tenure, age, length of residence, and living in vulnerable housing. Although there were significant variations in uptake with household income, there was no pattern to the variation. There was no relationship between depth of flooding experienced inside the home and signing onto the AVM. The least affected, i.e. those who had not even had flood waters in their garden, were less likely to have signed onto the AVM than those who had had gardens flooded (9% compared with 26%). Whether or not residents had had other parts of their property flooded, however, did not make any difference to uptake.
Insurance and compensation (relief funds)
Residents in flood risk areas can prepare in advance for flooding by taking out insurance to cover damage to the building structure and to the contents of their property. Such insurance, rather than government or other compensation or relief schemes, has since the early 1960s provided the main mechanism by which flood victims in England and Wales are compensated for damage to their property (Arnell, Clark and Gurnell, 1984). Flood insurance has been provided routinely through the competitive insurance market as part of ordinary household insurance. Flooding is a standard feature in policies for properties in flood risk areas with annual probabilities of 1 in 75 or greater. However, since the flood events of 1998 and autumn 2000, the Association of British Insurers (ABI - the organisation representing insurers) has revised the basis on which insurers offer insurance in areas of significant flood risk in order to limit the liability.
The latest Statement of Principles from the ABI (ABI, 2005) commits its members to continue to provide flood insurance cover to existing customers in areas of significant risk where there are government plans to improve flood defences within five years. Moreover, if flood insurance is to remain widely available the government must make further progress in reducing flood risk. Should the government reduce investment in flood risk management measures in the long-term insurers will reconsider their flooding cover. New customers in flood risk areas are considered on a case by case basis. Cover is not normally refused but will be influenced by the level of existing or proposed flood defence measures and may result in higher ‘excess’ payments in the event of claims (ABI, pers. comm., 14.9.2006).
However, flood relief funds are also frequently set up following large-scale flood events. These receive donations from local authorities (where available), local communities and businesses and are usually administered by the local authorities, the local Mayor’s office or by voluntary organisations. However, previous experience shows that not everyone appears to receive information about these funds, e.g. those who were evacuated, and there is often resentment that those who were uninsured are able to benefit more from the funds than those with insurance (Tapsell and Tunstall, 2001). Voluntary donations of household goods are also often made available for flood affected households, particularly for those without insurance cover. Goods can include: clothing, furniture, decorating materials, and general household goods.
Insurance in the three Surveys
All three surveys in this study contained some information on insurance take-up that had not been previously analysed. Taking out insurance was a common form of preparedness measure taken by residents in the flood affected areas surveyed in England and Wales, although for many flood insurance may have come automatically as part of their general household cover (Figure 5.1). Both the respondents in the Warnings Survey and the flooded in the Intangibles Survey were asked about having insurance cover at the time of the recent /worst flood. The data do not cover those who may have sought insurance after the flood event.
In the Lower Thames Study, only 67% (138) of residents, most of whom had not experienced recent flooding inside their homes, reported currently having contents insurance that covered flooding. Respondents in the Lower Thames Survey were also asked “Have you ever experienced problems renewing or obtaining contents insurance because of the risk of flooding to your home?” From the responses, it appears that such problems only affected a minority (16%) in this flood risk area. The very small category (44) who had experienced some level of flooding at some time, not necessarily flooding inside their homes, reported more problems (25% compared with 14% for those not reporting themselves to have been flooded) but the differences were not statistically significant.
Figure 5.1: Household insurance: Intangible and Warnings Surveys
In both the Warnings Survey and the flooded households in the Intangibles Survey, when the two main forms of insurance (buildings and structure, and new for old) were considered, it emerged that taking out insurance against flooding did not vary significantly with the number of floods experienced, or length of residence. Those with prior awareness of the flood risk were no more likely to have insurance of any type than those not aware. Similarly, age was not a significant factor in having insurance cover in most cases. However, in the Warnings Survey, the very elderly (75 and over) appeared to be somewhat different from those younger. Those aged 75 and over were less likely to have the more modern ‘new for old’ insurance (68% compared with 82% for those under 75) and more likely to have other forms of contents insurance than the under 75s (22% compared with 12%). In the Intangibles Survey, the older age groups (those aged 65 and over) were also found to be less likely to have ‘new for old’ contents insurance and more likely to have other contents insurance. The key factor important in insurance take-up in both surveys was social grade, with those in the lowest social groups significantly less likely to have all kinds of insurance (Figures 5.2, 5.3 and 5.4).
Figure 5.2: Buildings and structure insurance by social grade: Intangibles Survey and Warnings Survey
Intangibles Survey: chi-square; p<0.001
Warnings Survey: chi-square; p <0.001
Household income was considered in the Warnings Survey and was also found to be a significant related factor (Table 5.1). This may have significant implications as it is estimated that half a metre of floodwater in a modern semi-detached house will result in an average cost of £15,000-£30,000 (22,000-44,000 Euros) to repair the building and £9,000 (over 13,000 Euros) to replace damaged belongings (ABI, pers. comm., 23.9.2006).
Figure 5.3: “New for old” contents insurance by social grade: Intangibles Survey and Warnings Survey
Intangibles Survey: chi-square; p<0.001
Figure 5.4: Other types of contents insurance by social grade: Intangibles Survey
and Warnings Survey
Intangibles Survey: chi-square; p<0.001
Table 5.1: Proportion of respondents with insurance cover according to gross monthly
Household income: Warnings Survey
Buildings and structure insurance
Gross household income (before tax) per month*
£4,000 or more
‘New for old’ contents insurance
Can you please indicate which one of the following represents your gross household income (before tax) per week, month, or year? **