6.5 Summary and conclusion on vulnerability and resilience The manner in which people may respond to flooding and their capacity to recover may be affected by their subjective severity of the flood impacts. Of the three surveys included in this reanalysis, only the Intangibles Survey sought to measure the subjective impacts of flooding and magnitudes of impacts in detail on its flooded sample by looking at the subjective severity of flooding upon households. Getting the house back to normal (the disruption to life and all the problems and discomfort during recovery) was rated as the most serious of the effects, followed by the stress of the flood event itself, having to leave home and worry about flooding in the future. These first three intangible effects were rated as markedly more serious than the tangible damages to the contents and structure of the property. There were striking and significant differences in the rating of the effects between men and women, with women giving a higher rating than men to almost all the effects and also rating the flood overall as having a more serious effect on their household than did the men.
The most highly rated impact, disruption, and all the problems of getting the home back to normal, was most closely associated with the stress of the flood event itself, also a highly rated impact. The stress rating was associated not only with disruption but also with having to leave home, worry about future flooding and with health effects. This rating therefore appeared to capture many of the most severe impacts of flooding on the lives of households. The overall rating of the seriousness of the effects of flooding was highly correlated with the stress rating. Other effects that were closely associated with the overall rating were disruption, having to leave home, health effects and damage to the house.
The reanalysis in this study is based on a model (Figure 6.1) that considers that vulnerability and resilience to flooding depend on a series of factors: flood event characteristics, social characteristics including prior health, dwelling characteristics and post-flood factors or intervening factors. In order to see what factors may affect vulnerability and resilience three variables were chosen as dependent variables: the GHQ12, overall subjective severity, and subjective stress on the household.
A number of key points emerged from the analyses on vulnerability and resilience. The four vulnerability measures examined appear to be measuring somewhat different aspects of vulnerability. The correlations between the variables are only moderately strong and there are only four predictor variables that are shared across more than two of the measures in the regression models.
The research on the GHQ12 measures shows that flooding has impact on the mental health of flood victims not only in the short term (at the worst time) but also in the long term as reflected in the current scores registered at the time of the interview, in most cases at least a year after the flooding. The current scores were higher for the flooded sample than for both the ‘at risk’ sample and the average for England in the Health Survey for England 2003. While there is recovery and resilience, as evidenced by the differences in the worst time and current scores, the flooding has long lasting impacts on the mental health of flood victims.
Vulnerability as measured here remains difficult to explain. Although the levels of explanation offered in the regressions analyses are not high, such levels of explanation are common in social science. Nevertheless, for two of the measures more than three quarters of the variance in the vulnerability remains unexplained when all the potential explanatory variables available within the study were included in the analysis. Community and social variables and psychological measures that were not included in the Intangibles Survey might offer further explanation.
The basic flood characteristics of depth and extent of flooding were not as prominent as explanatory factors as might be expected although they did play a part for some measures. All the respondents in the Intangibles Survey had flood waters inside their home and the results of the regression indicated that the actual depth of flooding is not such a salient factor. The contamination of the floodwaters was surprisingly important featuring as a predictor for all the measures and models, not only for the GHQ12 scores which focuses on mental health but also for the stress of the event and the overall severity.
Social variables that we might expect to be associated with vulnerability, such as old age, ill health and disability in the household, living alone, living alone in old age, having children or young children in the home did not feature as prominently as predictors as we might have expected. Indeed old age and or living alone were included in some models with an effect in the opposite direction to the expected one. The models were consistent with the bivariate analyses which showed that the middle aged tended to be more vulnerable than older people. Prior health was a predictor of the GHQ12 scores and featured in six of the eight models that included social variables indicating that health status contributed to many forms of vulnerability and both in the short term and long term. Gender was also a common factor in the models apart from those for the current GHQ12 The area house price ratings which reflect the wealth of the areas where people lived were significant factor in six of the models although the effect was not consistently in one direction. So too was tenure for some measures.
When post-flood events and responses were introduced, having to leave home and the time spent in getting the home back to normal were important explanatory variables. Institutional responses in the aftermath of flooding by insurers and loss adjustors were a very important explanatory factor common to all the vulnerability measures. This shows that how these institutions and the individuals within them deal with insurance claims can have a very significant role in mitigating or exacerbating the impacts of flooding on households. This emerged very strongly in FHRC’s earlier qualitative work (Tapsell et al., 1999; Tapsell and Tunstall 2001). Having uninsured losses was a predictor in two models indicating that where insurance cover was adequate, people were less vulnerable. Social and other institutional responses in terms of help from outside the household were factors that did not emerge as significant predictors of vulnerability. Indeed, the bivariate analyses showed that those helped tended to have higher scores on the vulnerability variables, probably because those who attracted help from outside the home were more seriously affected by the flooding. Thus such help did not emerge as a mitigating factor in vulnerability.
7 Summary and Conclusions The major objectives of FLOODsite Task 11 that this research aimed to address were:
to characterise types of communities with regard to their preparedness, vulnerability and resilience related to flood events;
to understand the driving forces of human behaviour before, during, and after floods;
In this report we addressed these questions through the reanalysis of three existing sets of data originally collected for other purposes between 2002 and 2005:
‘Intangibles’ data set
‘Warnings’ data set
‘Lower Thames’ data set
The first two surveys covered mainly those affected by flooding in a wide range of flood events and local communities in England and Wales and the third focused mainly on those at risk in a particular allocation in the Thames Valley, in the South of England.
Our key hypotheses were that individuals or households are vulnerable or resilient to flooding in the context of particular situations, especially their risk environments. Every flood therefore presents a combination of factors and the outcome in terms of vulnerability or resilience will be a combination of:
the characteristics and resources of the population affected;
their dwelling characteristics; and
the organisational and institutional responses to a particular event.
Therefore, the social and dwelling characteristics of the respondents in the three data sets were described in Chapter 3 and the flood events covered in the survey and the evidence on risk perceptions and constructions were presented in Chapter 4.
7.1 Social vulnerability and the drivers of human behaviour, before, during and after floods In chapter 5 we explored human behaviour before, during and after flooding in relation to the key factors outlined above, which may influence the levels of social vulnerability and the ability to cope with and recover from a flood. Table 7.1 summarises some key findings on preparedness prior to flooding.
Almost all the preparedness actions were found to vary according to the specific location surveyed, thus highlighting the importance of the combination of factors unique to each flood event.
According to the literature on social vulnerability, one might expect that many of the respondents and households in the surveys e.g. those with young children, older residents, long term ill or disabled, and those on lower incomes or in lower social grade groups would be particularly vulnerable during flood events. This was evidenced in the data analyses, however, the situation is complex and different groups were not necessarily vulnerable across all situations.
Flood awareness is a difficult concept to define and measure and it needs to be time-bounded. However, prior awareness of flood risk or the lack of it is important: it can be taken as indicative of vulnerability and may be a factor affecting response to flooding. The surveys show that awareness prior to actually experiencing flooding or on moving to the address was low (at between 24% to 30%). In some of the surveys, prior flood experience, being a property owner rather than renting property and a longer term residence were associated with prior awareness.
Table 7.1 Social vulnerability to flooding and preparedness actions: Intangibles and Warnings Surveys