Property-level adaptation to pluvial flooding: An analysis of individual behaviour and risk communication material

In terms of behavioural insights, the risk communication materials were examined to determine how the concept of threat appraisal was employed (e.g., providing occurrence probabilities indications of potential severity) and whether the hazard was presented at a large scale (e.g., for the whole of Germany), or on a local level (e.g., neighbourhood level). While 78% of the materials communicated the hazard at a global level, i.e. without local context, only 43% of the risk communication materials communicated the hazard at the local level, i.e. with a local context (Fig. 6). It should be noted that a risk communication material can communicate a hazard on a larger scale and on a local level. Such a risk communication material was then counted in both categories. The threat was mentioned at the local and global levels twice as often in general terms as it was broken down into (future) probability and severity (Fig. 6). Flood maps are one way of presenting a hazard, the expected damage and the probability of occurrence at a local level. Within the analysed risk communication materials, flood and or/risk maps specifically for pluvial flooding were shown in approx. 15% of all risk communication materials analysed. The survey data reveal that only 32% (n = 1267) of residents recently affected by pluvial flooding believe it is likely or very likely that they could be hit again. Approx. 25% (n = 1237) of respondents said they considered the consequences of a future event to be bad or very bad for them.

To trigger coping appraisal in potentially affected households, flood adaptive measures can be presented in risk communication materials. In the material analyzed, measures are mentioned in 93% of risk communication materials. However, this does not mean that all of them described the measures in more detail, what hints to unused potential of risk communication: To trigger an adaptive behavior, one could additionaly describe the response efficacy of a measure and/or show the reader how to implement a measure (to enhance self-efficacy). In addition, implementation costs could be presented so that the reader can select a measure that fits their own budget (in terms of money or effort) and needs. In 52% of the materials, at least for one measure, more details are given on how this measure can be implemented in practice what can promote self-efficacy. Response efficacy is addressed in 53% of the materials, by describing how at least one of the measures reduces flood damage. Much less frequently (9%), the costs of at least one measure are mentioned.

The survey data reveals that most respondents (82%, n = 1206) see themselves in a position to implement a measure, hence have a high self-efficacy. However, response efficacity is much lower, as only 59% (n = 1255) of respondents believe that adaptive measures sufficiently reduce flood damage. In terms of cost, 34% (n = 1171) of the residents surveyed believe that adaptive measures are too expensive while just 13% (n = 905) of respondents believe that there are sufficient financial support programms.

Responsibility was communicated in about half of risk communication materials, and thus, less often than the threat or coping options (Fig. 6). In 41% of risk communication materials, the self-responsibility of potentially impacted residents was highlighted. The responsibility of public authorities was outlined in 28%. The survey data show that 33% (n = 1234) believe that provision against (pluvial) flooding is the responsibility of public agencies. 64% (n = 1244) of the respondents believe everyone must make their provisions. Both questions were not asked exclusively. Respondents could thus see the responsibility both with the public sector and themselves.

Making potentially affected people aware of their risks and coping options is one of the core ideas of PMT since it is a fear appeals strategy (Rogers 1975). Most of the respondents have a low perceived severity (68% of households surveyed) and a low perceived (future) probability (75% of households surveyed), leading to an overall low threat appraisal. However, those households had been flooded before being surveyed and thus are located in a potential flood zone. The fact that their flood experience did not result in a high appraisal of threat might be partly due to the perception that the flood event they experienced was significant (Wachinger et al. 2013). However, it may also be because households are not sufficiently informed about their hazard situation. It can be concluded that even after an event, municipalities cannot assume that the appraisal of threat is high among affected households and that those must be informed about flood hazards. From behavioural science, we know that in order to alert households to the threat, it is helpful to communicate both the severity and the probability of occurrence (Bubeck et al. 2012, Dillenardt et al. 2022, Grothmann and Reusswig 2006, Rogers 1975, Maddux and Rogers 1983). Therefore, future risk communication materials should be better aligned with behavioural science findings in this regard.

Additionally, presenting a pluvial flood threat on a general rather than a local level may result in potentially affected residents being generally aware of the threat of future pluvial flooding but unaware of their local and personal threat. The enhanced development of flood maps could provide the opportunity to convey localized information and be a great communication tool if understood correctly. The latter is essential since communicating an event's probability and magnitude can be challenging. Using return periods to communicate the probability of occurrence can lead to a "flood is due" effect (Grounds et al. 2017). This effect describes that lay people tend to misinterpret the probability of occurrence communicated via return periods. There are ways to overcome this issue, e.g. by (additional) communication of the likelihood in any given year (Grounds et al. 2017; Kellens et al. 2012). In any case, it is essential to ensure that communication via a map is well-designed and that sufficient explanations accompany the map to avoid misunderstandings.

A high coping appraisal has been identified as a clear and robust positive driver of adaptive behaviour in the context of pluvial and fluvial flooding (Dillenardt et al. 2022; Grothmann and Reusswig 2006). Hence, it is positive that almost all materials at least generally refer to adaptive measures and, therefore, trigger somehow the coping appraisal of households. However, coping appraisal could have been triggered more specifically by presenting measures of response efficacy and response costs, which has been done to a much lesser extent. This can not be due to medium used not providing enough space since flyers and checklists, whose nature is to briefly and compactly highlight a topic, together comprise only 14% of the risk communication materials examined. Segmenting coping appraisal into self-efficacy, response efficacy, and response cost offers the possibility of systematically promoting coping appraisal. It can be assumed that households will not implement a measure if it is considered unfeasible, ineffective or far too expensive. Hence, barriers to implementation are reduced by ideally addressing all three aspects of coping appraisal.

While over 80% of respondents feel able to implement at least one measure, attitudes toward the measures themselves could be better. Many respondents feel that measures are not effective and are too expensive. This emphasizes again the need to address response efficacy and response costs better. While response efficacy is considered in about half of the materials, the implementation costs were communicated in just 9%. Hence, there needs to be more information on the specific costs of adaptive measures. Although this is understandable because the cost of implementing a measure depends on many individual factors, e.g., the specific building structure and its location, it is essential to at least provide readers with further sources or an approximate range of costs.

Even if households view the efficiency of measures positively, the financial burden of implementing them can be a deterrent. Support programs that financially encourage the implementation of adaptive measures increase their affordability and the motivation of citizens to engage with these measures (Brockhoff et al. 2019). The need for financial support programs is evident among the surveyed households, where almost 90% see a need for support programs. This perception of households affected by pluvial flooding aligns with how reconstruction processes can be better managed in the sense of resilient flood recovery, also termed "build back better". This requires integrating and implementing adaptive measures, contrasts with rapid reconstruction and requires long-term support that provides time to plan and implement adaptive measures in private households (Slavíková et al. 2020). Programs that financially support the implementation of adaptive measures as part of a reconstruction process are rare, but such programs could be critical not only to enhance resilience but also to eliminate the criticism that government financial reconstruction assistance merely passes on the financial loss of households to taxpayers, allowing floodplains residents to remain unadapted (Slavíková et al. 2020).

In the context of integrated flood risk management, in which all stakeholders jointly manage flood risk, potentially affected households also bear part of this responsibility, particularly for the protection of their property (WHG (Wasserhaushaltsgesetz), 2009 § 5 Abs. 2). Studies have shown that households do not yet recognize this responsibility even in flood risk areas (Bamberg et al. 2017).

However, among the households surveyed, most respondents (two-thirds) showed a relatively high sense of (self) responsibility. In contrast, a smaller proportion of respondents (one-third) saw the responsibility instead of the public administration. This may be due to the lower severity of pluvial flooding, which will be considered in the next section. Anyways, in none of the surveys were these questions asked on a mutually exclusive scale, and as such, both seeing responsibility with public authorities and one’s responsibility can exist simultaneously. The extent to which the respondents correctly assess their responsibility and what personal responsibility should be assigned to private households cannot be clarified based on the data evaluated. However, responsibilities play a role in the decision-making process towards a protective response (Dillenardt et al. 2022; Lindell and Perry 2012; Maidl and Buchecker 2015). Naming them clearly can increase the effectiveness of flood management (Thistlethwaite 2017). Therefore, responsibilities should be communicated to advance pluvial flood management.

When discussing transferability, a distinction must be made between the methodology used, results and conclusions. Internationally transferable is the assignment of various adaptive measures to the super-groups used. However, it must be noted that the way insurance works is country-specific, as conditions differ (Hudson et al. 2019; Tesselaar et al. 2022). Whether and how information materials from other countries are based on PMT and PADM cannot be concluded from our results. Moreover, the attitudes of the surveyed households cannot be transferred to other countries.

The methodology applied is based on the behavioural models PMT and PADM. The relationships between threat appraisal, coping appraisal and protection motivation assumed by the PMT have been applied by the international scientific community to data from different countries in order to understand adaptive behaviour better (Bamberg et al. 2017; Grothmann and Reusswig 2006; Bubeck et al. 2013; Spekkers et al. 2017). Noll et al. (2022) compared the effects of PMT factors in the U.S., the Netherlands, Indonesia, and China. They found that, in essence, the positive effect of Threat and Coping Appraisal on protection motivation existed in all four countries. Internationally recognized frameworks such as the PADM and the Sendai Framework suggest that responsibilities need to be communicated and discussed (Lukasiewicz et al. 2017, Maly and Suppasri 2020, Wolff 2021, Lindell and Perry 2012). Therefore, the methodology presented here, based on PMT and PADM for the structured analysis of communication materials and survey data, can be applied to other countries.