key: cord-0710231-zq0p2oen
authors: Walkling, Ben; Haworth, Billy Tusker
title: Flood Risk Perceptions and Coping Capacities among the Retired Population, with implications for risk communication: a study of residents in a north Wales coastal town, UK
date: 2020-08-07
journal: Int J Disaster Risk Reduct
DOI: 10.1016/j.ijdrr.2020.101793
sha: ba120091f8b422f97f204fcc41b102a4f0ae51b0
doc_id: 710231
cord_uid: zq0p2oen

Flood risk communication strategies have been ineffective for older adults as they have failed to accommodate diversity, viewing retired populations as homogenous. There have been calls from academics and NGOs to develop more detailed understandings of older adults’ risk experiences to inform disaster risk reduction (DRR) and communication approaches. We conducted in-depth interviews with twelve members of the retired population, of which the majority happened to be members of a local church, in a flood risk area of north Wales, UK, in 2018 to ascertain risk perceptions, coping capacities, and risk communication preferences to inform more age-centred approaches. Results present retired population are a diverse group with varying perceptions and capacities. While personal risk perceptions were low overall, coping capacities varied and were primarily social in nature, which can be sustained despite mobility or other limitations typical of older age. Participants expressed preference for traditional/interpersonal risk communication methods, such as telephone calls or home visits. A key recommendation from this study is that risk communication and DRR practices should adopt people-centric approaches that are co-produced and respect the differentiated vulnerabilities, capacities and needs of at-risk populations. This study and its findings are important is providing a more nuanced picture of the vulnerabilities and capacities of the particularly at-risk population of older adults. We must ensure that future DRR research, policies and practices focus on all experiences of at-risk populations, not only the dominant narratives or extremes of groups, to capture differences within groups’ abilities to support more effective community DRR.

1. Introduction 1. Introduction 1. Introduction 1. Introduction 27 Natural hazard and climate-related disasters are on the rise (Dominey-Howes 2015), with wide-28 ranging impacts on societies. Hazard events, such as cyclones, floods, or bushfires, do not become a 29 disaster until they intersect with a vulnerable population (Blaikie et al. 2005) . Vulnerability to 30 disaster is widely understood as a social construct, and is closely linked to systemic inequalities and 31 marginalisation (Sword-Daniels et al. 2018; Blaikie et al. 2005) . Consequently, socially-marginalised 32 groups are also some of the most vulnerable to, and disproportionately impacted by, disasters, such 33 as the economically poor ( To effectively reduce losses from disasters, recent scholarship and international policies inform that 37 we need strategies centred on understanding vulnerabilities, reducing risk, and building community 38 resilience and coping capacities. International disaster management has shifted away from a sole 39 focus on emergency response towards reducing vulnerabilities, preparing communities, and disaster 40 risk reduction (DRR) (e.g. UNDRR 2015). Emphasis is placed on resilience-i.e. the processes and 41 outcomes related to increasing communities' abilities to cope with disruption (UNDRR 2015) . 42

Building resilience involves understanding risk and uncertainty, effective communication, social 43 cohesion, empowerment and capacity building (Haworth et al. 2018) . 44

However, research shows policies and practice have been until recently mainly focused on 45 predominant groups in society (Eriksen 2014), inadequately capturing the needs of some groups 46 (e.g. Lakhina et al. 2019 ). Invisibility of vulnerable groups is associated with inadequate DRR (Walters 47 age alongside characteristics such as living with family or a long residency provide indicators of 121 preparedness, but general levels remain relatively low (Loke et al. 2012) . 122

The vulnerabilities of older adults were clearly demonstrated in the disproportionate mortality rates 124 of 49% for older adults in Louisiana following Hurricane Katrina, considering that older adults only 125 represented 6% of the total population (Kawasaki 2014 ). This high rate has been attributed to 126 evacuation challenges and invisibility of older adults to emergency services (Kawasaki 2014 ). This 127 high rate is a typical pattern, especially in the US (Ashley & Ashley 2008) . 128

The root causes of older adults' vulnerability differ across geriatric literature. As many older adults are not engaged in day-to-day economic activities they may be able to invest 158 more time on DRR activities in their community (Harris & Mihonvits 2015) . 159

However, as with vulnerability, capacities vary among the older adult population. For instance, those 160 more connected to society or family are often seen as better-able to cope with disruption compared 161 to older adults without such support networks (Kessel 2013) . Additionally, many older adults provide 162 support to others in their community. Over 2 million older adults in the UK are carers, and over half 163 of these have a disability or health condition themselves, and therefore many need additional 164 support during evacuation or preparedness activities (AgeUK 2018). 165

But older adults' capacities do not negate their specific vulnerabilities or limitations. While 166 vulnerability and resilience are related concepts, they are not simple opposites (Whittaker 2008) . 167

While resilience has its own complex history and critiques, these are not the focus of our article (for 168 more detailed discussions on these topics see Alexander 2013; Cretney 2014). Here we want instead 169 to highlight conceptual differences of importance for understanding our study on older adults and 170 their disaster vulnerability and capacities. Vulnerability informs that everyday social and economic 171 circumstances construct differing experiences of hazard exposure and disaster impacts (Whittaker 172 2008) . Reducing vulnerability, therefore, involves mitigating short term conditions, such as poor 173 mobility or evacuation access for older adults, as well as addressing historical, political, and socio-174 economical root causes (Blaikie et al. 2005) . Resilience, on the other hand, refers to capacities and 175 processes to enable fast and effective response and recovery from disruption, such as disasters 176 (Zhou et al. 2010 ). Thus, one can be simultaneously vulnerable and resilient in different ways; 177 reducing vulnerability does not necessarily increase resilience, and vice versa. This is particularly 178 important in the context of older adults. While capacities do need to be given more attention and 179 value, the conditions that create vulnerability -i.e. part of the reason why older adults need to be 180 more resilient -remain, and can be continually reinforced unless the root causes of vulnerability are 181 addressed. Working towards recognising, valuing, and building resilience capacities is vital for 182 effective DRR, but does not mean we can overlook the complexities of vulnerability. 183

When dissecting the complexities of vulnerability and resilience capacities, we should also consider 184 the notion of intersectionality (Vickery 2018 ). Intersectionality originates from feminist studies theorising the relationship between different social categories such as gender, race, and sexuality, 186 among others (Crenshaw 1990; Valentine 2017) . Intersectionality is often viewed as the way that any 187 particular individual stands at a crossroad of multiple groups with a combination of identities 188 influencing the individual's experiences and relative advantage/disadvantage (Minow 1997) . Age 189 alone doesn't define the experiences of older adults. Hopkins and Pain (2007) argue that some 190 markers intersect with age in meaningful ways, such as gender, as gender role expectations often 191 differ between generations. Thus older adults are not only older but also belong to a specific gender 192 or previous job/home role. A significant intersection for older adults is disability (Gibson & Hayunga 193 2015 (Orr et al. 2015) . This failure can occur due to the patchiness in geographical coverage of official 248 warning systems; inability to adequately forecast a hazard; or the inability of individuals to receive or 249 personalise warnings (Parker et al. 2009 ). The latter is important when considering older adults as 250 impairments may limit message comprehension (Mayhorn 2005) . Current best practice to address 251 some of these issues is to use multiple channels to ensure messages are disseminated as widely as 252 possible and to reinforce them through consistent exposure (Riley 2014) . 

Participants were recruited through a purposive sampling method, a typical method in disaster 299 research (Stallings 2007) . One author utilised personal networks to identify participants with the aid 300 of key informants. Although this may have led to a bias of including engaged participants and a form 301 of gatekeeping, this was a necessary trade-off with ensuring that a sufficient number of individuals 302 were interviewed. The participation criteria included: at least 65 years of age (minimum pensionable 303 age in the UK), not in full-time employment (as it was felt those in employment would have different 304 support networks to rely upon), able to communicate, and a study area resident (to ensure 305 participants were at risk of flooding to some degree and had local experience). A relatively flexible 306 participation criteria allowed for capturing unexpected demographics of the study area, reflecting 307 the intention to demonstrate heterogeneity among the age group. 308

In total 12 individuals were interviewed in June 2018, ranging in age from 65 to 88 with a median of 309 78.5. There was an uneven gender distribution, with three male and nine female participants. 310

However, this reflects the general demographic composition of the older population in the UK (ONS 311 2017). Five participants reported direct flooding experience flooding, four indirect experience, and 312 three had no flood experience. Two thirds of the participants happened to be members of a local 313 church in the study area, which may have introduced an unintended bias into the findings as social 314 connectedness may be higher in church-attending older adults compared to others. 315

While we acknowledge the sample size and study area do not allow us to make broad claims 316 representative of all retired people and flood prone areas, we consider the participants key 317 informants and experts in their own experience of flood, vulnerability, and risk communication. 318

Thus, the participant sample is appropriate for the contribution our study seeks to make and does 319 not detract from the value of discussing specific issues faced and risk communication preferences 320 directly with individuals of an at-risk group. 321

Semi-structured interviews were adopted over other qualitative research methods, such as surveys 323 or focus groups, to ensure participants were not influenced by others' responses, to allow 324 opportunities for participants to raise unexpected topics of interest, and for the interviewer to 325 clarify responses. Interviews were conducted in semi-public space (to balance ethical safety 326 concerns and ensuring participants felt comfortable to discuss freely) in two church halls in the study 327 area, and lasted up to 45 minutes (balancing the need for in-depth discussion and avoiding 328 participant fatigue). The fieldwork period during the summer meant that the flood risk in the area 329 was low, meaning participants' views may represent the pre-disaster or post-disaster stage and 330 access to participants was easier than if a flood had recently occurred. All participants were asked 331 the same base questions to ensure consistency and enable comparisons between participants, 332 broadly covering themes of risk perceptions, coping capacities, and experience with risk 333 communication. Interview questions were pilot-tested prior to data collection. To mitigate social 334 desirability bias (Krosnick 1999 ) multiple questions on the same topic were asked in different ways 335

and key topics were broken down into sub-questions to observe if different answers were given in a 336 different context (Grimm 2010). Interviews were recorded and transcribed verbatim by one author 337 to ensure to ensure consistency, comprehensiveness and data familiarisation (Braun & Clarke 2006) , 338 with results verified by the second author. Interviews were analysed using thematic content analysis, 339 as guided by Braun and Clarke's (2006) methodology, allowing for elucidation of meaningful themes 340 directly from the transcripts. Insights and conclusions were drawn directly from the interview data 341 without influence from pre-determined codes. The steps suggested by Braun and Clarke (2006) 342 prompted regular comparisons and validation of codes and themes to avoid anecdotal conclusions. 343

The findings are presented and discussed in the next section. All names used are pseudonyms and 344 ages are correct at the time of the interviews. 345 4. Results and discussion 4. Results and discussion 4. Results and discussion 4. Results and discussion 346

The following section will present the interview results and accompanying discussion. First, the 347 diversity of challenges and vulnerabilities described by participants will be presented. Then, varied 348 risk perceptions are explored, followed by coping capacities with a particular focus on older adults' 349 ability to support others, and subsequently a discussion of preferred risk communication methods. 350

Finally, we present recommendations for the field of disaster management. A considerable challenge for some participants was their need to rely on others to undertake DRR 358 activities. Transport and vehicle access was important, with many participants being without their 359 own vehicle. This reliance on others could result in inadequate and inefficient flood response and 360 evacuation for older adults. These issues can be compounded if those people the older adults rely on 361 are also impacted by floods, or also vulnerable, such as other older adults in a neighbourhood or 362 friendship network. Provisions should be made to ensure that older residents are adequately 363 considered and informed regarding evacuation protocols, ensuring that older residents and their 364 support networks can pre-plan access to transport and other vital resources. Additionally, those 365 without such informal support networks, who are furthermore vulnerable, will require support 366 through other means, such as volunteer organisations. 367

Poor health is not necessarily a feature of the experience of a retired person, as there are many 369 older adults who are healthy and continue to move with relative ease. However, 'biological' aging 370 will eventually affect everyone, and thus poor health is a problem for many within this demographic 371 (Mayhorn 2005) . For the participants who experienced this, reduction in mobility was of greatest concern. Two participants relied on rollators and others felt that reduction in mobility affected their 373 ability to respond quickly to potential flooding. They were unable to move items higher as they were 374 too heavy, or they did not have upper floors due to living in a bungalow. In this way, the 375 recommended flood preparations were not achievable for many members of the retired population 376 (EA 2010; NRW 2010; NRW 2018b). Addressing these limitations for older adults would prove 377 beneficial for other groups who face similar challenges, such as those with disabilities. 378 Some participants with limitations had made adaptions which could also prepare them for a crisis 379 event, such as carrying a panic-button, keeping a torch close-to-hand at night to aid declining 380 eyesight, or having a packed bag ready for unexpected hospital visits, potentially speeding up 381 evacuation. Therefore, adaptions in one part of an older adult's life may increase flood preparedness 382 in ways not seen for other demographics. adults, for some of the participants, health concerns, such as overcoming illness and reduction of 413 mobility following falls, were of greater concern than flood, as seen in Table 1 . 414 Table 1 . Reasons given for low/no flood concern. 415 Other concerns / No concern for others Alice (88)

"…I don't…really worry too much about that because, err, I think my only concern is that I don't fall over again…" High risk perceptions were less prevalent with just two participants expressing personal concern 418 about flooding. The first resides in very close proximity to the sea. The close proximity meant that, 419 for their household, they had to accept the risk of flooding or they would have to leave their home. 420

This acceptance of risk had led them to prepare and mitigate as far as they could within limitations 421 of cost, likelihood and political systems. The second participant had previously experienced flood in 422 1990. They had not been concerned prior to that event, but concerns have risen since. However, 423 concern has not transferred into preparedness, as they also displayed avoidance tactics in their 424 responses; they did not wish to receive flood warnings and said they would respond with panic and 425 alarm. For some older adults, previous experience may increase concern about the hazard and 426 event, but not translate into action due to a lack of awareness of how to adequately respond within 427 their limitations. 428

In contrast to the infrequency of floods described in the previous section, the inevitability of flood 429 was raised as an issue by some participants (Table 2) . They felt that nature and water specifically 430 were uncontrollable and if there was a flood it would occur with or without preventative measures. 431

They were unlikely to prepare as they felt powerless against its effects. 432 "Flooding is something that is, is quite devastating because there is nothing that you can stop it. Because you can't say 'oh well be careful don't start a fire or whatever'. Flooding when it comes is something that you can't do anything about isn't it"

"…if a flood is going to come a flood is going to come."

"…It's very sad but you can't do nowt about the weather can we."

"…I would sacrifice, I wouldn't attempt to -I know the power of water there is [no] trying to stop it. You can slow it down but if it wants to come in, it will come in."

"…it would probably affect me less actually. Affect us less because we accept, you know, that we are close to the sea, and if we can't accept that, err we can't live there." Many participants showed greater concern for their wider community than for themselves (Table 3) . 445 This is indicative of unrealistic optimism bias; as they felt that others in their community were at 446 higher risk of flooding ( J o u r n a l P r e -p r o o f hold for those who are older, as well as causing them to devalue the contribution they can make 450 towards flood preparedness and DRR. 451 Table 3 . Participants describing concern for others in their community. 452

"…A lot of people in that part of the world are disabled as well because it's very flat area and I do think they feel frightened. I know from talking to people who live over there when is a big and really bad storm. They don't go to bed."

"The concerns are that there are a lot of people in our neighbourhood that are elderly, but are infirm also. So the priority is to get them out, or make sure that someone knows where they are"

"…I would concern myself with people around me because some of them are…they have disabilities of different types. So I would want to be sure they were safe. Whereas I'm lucky enough to be able to get around." "There are people of our age that are more vulnerable than we are because they haven't got vehicles or they are on their own and would just that they were aware of the situation."

"…I would be concerned for people's immediate welfare."

"…because I care about the community (Pause). And I really, I [am] really distressed for people who have suffered in this way because, because that's who I am." Those without direct experience of flooding relied on others' experiences to estimate how they 462 would respond. Some participants expressed sympathy for people experiencing flooding in other areas of the UK during the interview period (summertime 2018), demonstrating that although 464 personal risk perceptions were low, awareness and worry for others may be higher. This could be 465 used to prompt personal awareness through linking others' experiences to their own circumstances. 466

Past employment experience also played a role in shaping risk perceptions and preparednes. A 467 former solicitor ensured that all important documents were stored in secure box and that deeds 468

were not kept on site as a protective measure. Another participant had worked in the Police 469

Headquarters during the 1990 flooding event and had heard many of the issues for those who 470 experienced the flood. The authoritative manner of their work made them feel that response efforts 471 should be left to emergency services. Influences that pre-date retirement and current circumstances 472 also need to be considered when understanding and planning for DRR among older adults. 473

The main capacity participants highlighted was an ability to support each other through social and 475 practical means, as illustrated in Table 4 . This mutual-aid was considered possible regardless of any 476 health conditions, for example, that may limit physical mobility. Others were or had been directly 477 responsible for the care of others, either in or outside their household, demonstrating willingness to 478 work around limitations and reduced capacities to provide care and remain active in social networks, 479 which may have been stronger for those participants that were members of the church. Utilising 480 these capacities among already established networks could provide opportunities to foster flood 481 preparedness more widely in communities. 482 

"Yes, communication and I do try and do that and I try and…ring people or write to them and things like that because that is my limitation."

"I've got to say there are between 8 and 10 [flood] wardens and we're all given a plan of who we look after. But we cover for other people, like I'm only supposed to do 5 'cos I said I don't want to do really but they said 'come on'. So I also help along the coast road. Ah, both sides because I've got friends there. So I pass it on to them."

"…probably check the neighbours were okay first thing."

"…the neighbours know if they need a lift or, they want something they can shout me."

"Just other members of the Church who are on their own. We help each other out, yeah…well The participants expressed desire to leave their home to minimise risk during floods, but overall 502 participant responses reflected a lack of awareness of practical steps that could be taken before 503 evacuating to minimise losses. Participants described different modes of and triggers for evacuation 504 (Table 5 ). Many were unaware of evacuation destinations. Risk communication needs to address 505 these differing perceptions to ensure resources can be flexibly directed where necessary to assist 506 those who cannot leave without support. Addressing the appropriate means, timing and location of 507 evacuation within a personal flood response plan, which are encouraged by emergency agencies, 508 that is embedded in the local context, would reduce uncertainty. 509 

Self-reported preparedness levels were low amongst participants. Lack of preparedness can be 513 linked to participants' low risk perception. However, other barriers such as health constraints also 514 influence preparedness. 515

Various sources and methods are utilised to communicate flood preparedness information for 516 flooding in the UK. These include traditional media, online messaging, and flood packs delivered to 517 homes (EA 2012). The latter is the way that eight participants received information. This form of 518 dissemination appeared ineffective to participants as it was often discarded as junk mail, as 519 described by Rhiannon (84), "Oh trustworthy, erm, well I suspect the pamphlet that you get from 520 your local authority. And if you bother to read them because they come in the junk mail, don't they?" 521

Although those that did recall reading the material felt it was straightforward and simple, they could 522 not remember if they had implemented any of the advice due to it being received many years ago 523 (Table 6) . 524 For participants that had not received preparedness information, there was a general feeling that 527 information was distributed unevenly, preferencing those who had previously experienced a flood. 528

This was highlighted by Martha (67), "I think we should all be told, yeah. Even if we're not, our own 529 property is affected, we've all got family and friends who could be and that. I think they should really 530 be giving more information to everybody in the community not just certain streets, which is how it 531 appears to be." Some participants lived outside the areas covered by flood warden programmes, 532 and thus were concerned they were missing out on important information. However, the lack of 533 actions by those who had received information suggests availability is not the only barrier to 534 engagement ( Table 7) . As previously highlighted, risk perceptions and capacities to act upon 535 information are important. 536 Table 7 . Participants' lack of action after receiving preparedness information 537 (Table 8) . 542 

For those that experienced flooding in 2013, a common complaint was that no warning was given 545 because flooding was expected at another area, and therefore people did not have adequate time to 546 respond (NRW 2014). The significance of time reflects the additional time the retired population require in preparing for floods. As reduced mobility and physical incapacity lead to slower response 548 times, or inability to complete actions themselves, retirees may require assistance in evacuation 549 preparation. 550

Echoing others in the literature, participants of this study felt that flood warnings should include the 551 magnitude and location(s) of events (e.g. Sheppard et al. 2012) , although this is currently challenging 552 due to the complex processes that lead to flooding. The use of a colour coding and scale system was 553 also discussed with participants. Within the UK there are three levels of flood warnings (NRW 554 2018c). Most participants were unaware of the scaling system but felt that any system would be 555 useful as it would provide context for the warning given. Some would change their response based 556 on the level of warning given, such as Mary (69) (Table 9 ). The added benefit of an SMS for some was that if 568 they weren't home when the warning was issued they could still respond appropriately. 569 communication should address organisations' roles to ensure that the public are aware of who will 595 be available and responsible during crises. Interviews also suggested that older adults are more 596 trusting of locally-focused information and may find messages that are related to their specific area 597 more actionable. 598

In light of the above, an age-centred flood risk communication strategy should be underpinned by 599 actionable messages that are primarily delivered by interpersonal means. The actionable messages 600 should be personalised to older adults with various needs and embedded in local contexts (Wood et 601 al. 2012) . Messages applicable to bungalow residents or with alternatives to lifting items would help 602 some older adults to better-prepare. Additionally, messages could use the retired population's 603 perceptions of community risk to help them relate risks back to themselves, increasing relatability 604 and potentially prompting risk reduction behaviours. 605

The findings of this study suggest a number of potential improvements for risk communication and 607 DRR concerning retired populations, especially in the context of coastal flooding, and societies 608 broadly. In this section we first provide recommendations for DRR policy and practice, before 609 suggesting areas for further research. Fourth, risk communicators should promote awareness and preparedness activities at times when 629 concern is at its greatest to capitalise on already-heightened interest. This could be during winter or 630 when floods occurring elsewhere are being covered by the media, for example. Older adults that are 631 socially connected often have groups that are ready made for such interventions, for example local 632 groups connected to Churches, community or age centres often hold weekly informative talks which 633 could be communicators to share practical preparedness advice (AgeUK 2018). interesting to explore whether there is now a generation of older adults who know how to use 658 technology or whether uneven use of digital technology persists. We hypothesise there might be an 659 increase in comfort for using communication technologies among older adults but divides between 660 those more and less connected will remain. Similarly, exploring the long term effect of the pandemic 661 on community cohesion and its benefits for communication could be an interesting area of future 662 research. As other technologies (e.g. geographic information systems for spatial analysis) continue to 663 change, these may also become increasingly relevant for assessing and mapping older adults' 664 vulnerabilities and communication needs (Szewranski et al. 2018) . 665

Finally, the sampling method drawing on the author's social network resulted in the majority of the 666 participants being members of a church, and thus studies exploring a more diverse retired 667 population should be undertaken. In this study, the church group sample may have presented a bias 668 of religious or spiritual individuals into the study, seen sometimes as a source of disaster resilience 669 (Almazan et al. 2018) , although faith was only mentioned by two of the participants as a source of 670 support. Regardless, there is the potential omission of social connectedness as an issue in the retired 671 population as a church community can provide this. Other older adults may not have similar 672 communities or support structures. The relationship between social connectedness, vulnerability, 673 and DRR needs to be considered outside of a group of predominately religious older adults. Despite 674 their church attendance, the capacities of the participants were similar, suggesting they were 675 generational rather than influenced by church membership; those participants unaffiliated with the 676 church were some of the most active people in the community, including by acting as flood wardens. 677

We hypothesise that research into older adults who experience mobility issues or are less socially 678 connected, such as those who are not members of churches, will highlight different issues to issues 679 to address for more effective community DRR. 680

Conclusions need to be considered carefully within the parameters of this study, being based on 682 interviews with 12 participants, many of whom were church members, in a coastal flooding context 683 in the UK. Some of the findings presented above may not relate to a wider at-risk retired population 684 -either with less social connections or in a context dissimilar to Rhyl. However, the insights provided 685 by the participants are useful in providing more nuanced discussions of older adults' disaster 686 preparation experiences and risk communication needs. Through this analysis of older adults and 687 flood risk in Rhyl, we have added to the to-date limited data and scholarship on disaster risk 688 perceptions, coping capacities, and risk communication needs for older adults. 689

Our findings demonstrate that vulnerabilities, capacities, and risk perceptions of older adults in 690 relation to disasters are diverse and dynamic. The coping capacities of participants varied, with some 691 able to socially support others, and others exhibiting limited means to prepare themselves for 692 floods. Some were highly concerned and aware of the risk they faced, while others showed no 693 concern and/or limited awareness. There was more in common in expectations and requirements of 694 risk communication, such as preferring the information and warnings to be communicated by 695 interpersonal methods. 696

A 'one-size-fits-all' approach to risk communication is ineffective for heterogeneous populations (Orr 697 & Twigger-Ross 2009). This includes the retired population, who may be demographically similar, but 698 have diverse vulnerabilities and capacities. Collectively grouping individuals as vulnerable without 699 also investigating their divergent experiences and capacities may lead to overlooking their 700 contributions to DRR (Cornell et al. 2012) . Risk communication should be sensitive to, and address, 701 differential capacities through actionable messages that allow individuals to better prepare within 702 their individual limitations (Wood et al. 2012) . This requires consultation with those who have 703 differing abilities to discover the challenges they face in implementing preparedness actions. 704

Methodologically, the findings of this study support ensuring at-risk populations and groups are 705 consulted to expound vulnerabilities, capacities and risk communication needs. Future research will 706 benefit from a larger, diverse participant sample, and consultation with medical professionals, social 707 workers or home-carers that are aware of the conditions that affect at-risk subgroups or minorities. 708 This is especially important for those who lack the agency to speak for themselves, such as those 709 with advanced dementia or speech-affecting disabilities. A combination of approaches will produce a 710 more nuanced and complete picture of the experience of vulnerable groups. 711

A key recommendation from this study is that future risk communication and DRR practices should 712 adopt people-centric approaches that are co-produced and respect the differentiated vulnerabilities, 713 capacities and needs of at-risk populations. One such means for contributing to this is developing an 714 aged-centred risk communication strategy in areas where there is a high proportion of retired 715 population. 716

The insights presented in this article also have relevance beyond older adults. For instance, the need 717 to discuss individuals' capacities and vulnerabilities in their local contexts with them directly to 718 elucidate specific challenges and adaptions for risk reduction, especially when not understood well 719 by those who are designing communication approaches, applies to any minority group for whom 720 common risk communication approaches may not be effective. 721

This study and its findings are important as they provide a more nuanced picture of the 722 vulnerabilities and capacities of a particularly at-risk population of older adults. We must ensure that 723 future DRR research, policies and practices focus on all experiences of at-risk populations, not only 724 the dominant narratives or extremes of groups, to capture differences within groups' abilities to 725 support more effective community DRR. 726

Changes to State Pension Age: The age at which you can get your State Pension is 731 changing

Later Life in the United Kingdom

Flood Fatalities in the United States

Resilience and disaster risk reduction: an etymological journey. Natural 739 hazards and earth system sciences

Predicting patterns of disaster-related resiliency among Typhoon Haiyan survivors

Disaster-related resiliency theory among older adults who survived Typhoon Haiyan

Preparedness for Natural Disasters Among older 747 US adults: A nationwide survey

Older 749 people's needs following major disasters: a qualitative study of Iranian elders' experiences of 750 the Bam earthquake

Population Aging and the Voluntary Sector: Key Figures and Projected Trends

Commission on the Voluntary Sector and Aging

At risk: natural hazards, people's vulnerability 755 and disasters

Investigating rural community communication for flood and bushfire preparedness

Risk perception -Issues for flood management 760 in Europe

The effectiveness of disaster risk communication: a 762 systematic review of intervention studies

Using thematic analysis in psychology

Older adults' disaster lifecycle experience of the

Queensland floods

Chapter 17: Literacy and Disaster Preparedness for Seniors

Re-defining community and vulnerability in the context of emergency 771 management

Effects of Air Temperature on Climate-Sensitive Mortality and Morbidity Outcomes in 774 the Elderly: a Systematic Review and Meta-analysis of Epidemiological Evidence

Disaster recovery among older adults: Exploring the intersection of vulnerability 777 and resilience

Vulnerability, "innocent" disasters and the imperative of cultural understanding

Disaster Prevention and Management

Flood risk perceptions and the UK media: Moving beyond 782

Older people and disaster preparedness: a literature 784 review

The Covid-19 crisis reveals how much we value old age. LSE USCentre

Risk Communication: Evolutions and Revolutions

Mapping the margins: Intersectionality, identity politics, and violence against 792 women of color

Resilience for whom? Emerging critical geographies of socio-ecological resilience

Rhyl town and area plan: looking forward together

Explainer: are natural disasters on the rise? The Conversation

The Impact of Intergenerational Oral History on

Perceived Generativity in Older Adults

What to do before, during and after a flood: practical advice on 803 what to do to protect yourself and your property

Flood Warnings: A Guide to the Environment Agency's flood 807 warning codes

Mapping the vulnerability of older persons to disasters

Older adults in disaster and 813 emergency management: What are the priority research in Australia

Disaster Risk Reduction

Online Government Services and the Offline Older Generation -Executive

Sexual and gender minorities in disaster

We Can Do Better: Lessons Learned for Protecting Older Persons 821 in Disasters

Position Paper: Climate Change in an Ageing World

Geographies of age: thinking relationally. Area

Local perceptions of infrequent, extreme upland flash flooding: 857 prisoners of experience?

Risk Communication and Natural Hazards

Chapter 5: Disasters and the frail elderly

Preparedness for Seniors: A Comprehensive Guide for Healthcare Professionals

Perception and communication of flood risks: A 866 systematic review of empirical research

The ability of older people to overcome adversity: A review of the resilience 868 concept

Coastal climate change, vulnerability and 870 age friendly communities: Linking planning for climate change to the age friendly communities 871 agenda

Maximising measurement quality: principles of good questionnaire design

Measures of Political Attitudes

Flood risk and climate change: global and regional 878 perspectives

Case study: 880 People from refugee backgrounds contribute to a disaster-resilient Illawarra

Social vulnerabilities and Hurricane Katrina: an unnatural disaster in

The concept of vulnerability in disaster research

Editorials: Bearing the brunt of covid-887 19: older people in low and middle income countries

At-home disaster preparedness of elderly people in Hong Kong

Vulnerability of older adults in disasters: emergency department utilization by geriatric patients 892 after Hurricane Sandy

Cognitive aging and the processing of hazard information and disaster 894 warnings

Warning the world of extreme events: A global perspective 896 on risk communication for natural and technological disaster

Designing effective risk communications for older adults

Elderly perceptions of social capital and age-related disaster vulnerability

Disaster Medicine and Public Health Preparedness

The social psychology of public response to warnings of a nuclear power 902 plant accident

The disaster flood experience: Older people's poetic voices of 904 resilience

Not only for myself: Identity, politics and the law

The resilient community and communication practice

What to do before, during and after a flood: practical advice 910 on what to do to protect yourself and your property

Wales Coastal Flooding Review Phase 1 Report -Assessment 914 of Impacts, A National Overview by Natural Resources Wales for Alun Davies AM, Minister for 915 Natural Resources and Food

Flood zones

Residential: How to prepare for flooding

Flood Warnings: Flood Codes

Enhancing flood resilience through improved risk communications

Rise of the Social Seniors revealed

Overview of the UK Population

Delivering benefits through science: Communicating risk and 937 uncertainty in flood warnings: a review of Defra/Environment Agency FCERM literature: Science

Project SC070060/SR2, Flood and Coastal Erosion Risk Management Research and Development 939

Delivering benefits through evidence: public 943 dialogues on flood risk communication -literature review: SC120010/R3, Flood and Coastal 944

Erosion Risk Management Research and Development Programme

Understanding and enhancing the public's behavioural 946 response to flood warning information

Communication, sense of community, and disaster recovery: A 948 Facebook case study

The benefits of flood warnings: Real 950 but elusive, and politically significant

Social media usage

Aged citizens in the warning phase of disasters: re-examining the 956 evidence

Happy to help? Exploring the factors associated with variations 958 in rates of volunteering across Europe

A comparison of direct 960 versus self-reported measures for assessing physical activity in adults: a systematic review

ISO 31000:2009-Setting a New Standard for Risk Management

Flood risk perceptions and spatial multi-965 criteria analysis: An exploratory research for hazard mitigation

The Experience of Flooding in the UK: A Research Study. London

Crisis and Emergency Risk Communication

Control and Prevention

Mental models in warnings message design: A review and two case studies

Influence on climate in the 2014 Southern England winter floods and their impacts

Adjustments to physical-social environment of 982 the elderly to climate change: proposals from environmental gerontology

Gerontology in Europe and Latin America

The missing link between flood risk awareness and 986 preparedness: findings from case studies in an Alpine Region

Gender roles: an incomplete revolution?

British Social Attitude: the 30 Report

Assessing adaptive capacity to tropical 993 cyclones in the East Coast of India: a pilot study if public response to cyclone warning 994 information

Understanding Risk Communication Theory: A Guide for

Improving Disaster 998 Resilience among Older adults: Insights from public health departments and Aging-in-place 999 efforts

Disaster Exposure in Older Adults

Comprehensive Guide for Healthcare Professionals

Methodological Issues

People with Disabilities and Disasters

Handbook of Disaster Research

Embodied Uncertainty: Living with complexity and natural hazards

Environmental Vulnerability Mapping for Environmental and Flood Resilience Assessment: The

Improving disaster 1015 preparedness of older adults living in the community

Older adults' narratives about a flood disaster: resilience, 1017 coherence, and personal identity

Older adults' disaster preparedness in the context of 1019 the

Disaster Risk Reduction

Flood risk and coastal change guidance

United Nations International Strategy for Disaster Risk Reduction (UNDRR) (2015)

Ageing and disability

Theorizing and researching intersectionality: A challenge for feminist 1032 geography

Using an intersectional approach to advance understanding of homeless persons' 1034 vulnerability to disaster

Viewpoint -The challenges of urban ageing: 1036 making cities age-friendly in Europe

People-centred early warning

Disaster Management. International 1040 lessons in risk reduction, response and recovery. Routledge: London

Disaster risk at the margins: homelessness, vulnerability and 1042 hazards

Resistance of personal risk perceptions to debiasing interventions

Research in social gerontology: social exclusion of 1046 aging adults

Vulnerability to Bushfires in South-Eastern Australia: a Case Study from East 1049

Resilience: thoughts on the value of the concept for critical 1051 gerontology

Global drivers of future river flood risk, letters 1054 nature climate change

Communicating 1056 actionable risk for terrorism and other hazards

Resilience to natural hazards: a geographic perspective

Natural Hazards

We thank the editor, Dr Sébastien Penmellen Boret, three anonymous reviewers and the study 728 participants. 729

Contributions of volunteered geographic information to community disaster resilience. 851Frontiers in Earth Science, 6: 183. 852