Measuring Social Vulnerability to Climate Change at the Coast: Embracing Complexity and Context for More Accurate and Equitable Analysis

Abstract

Social vulnerability indices are often used to quantify differential vulnerability to the impacts of climate change within coastal communities. In this review, we examine how “tried and tested” methodologies for analysing social vulnerability to climate hazards at the coast are being challenged by a new wave of indices that offer more nuanced conclusions about who is vulnerable, how, and why. Instead of producing high-level, generalised, and static conclusions about vulnerability, this new wave of indices engages more deeply with the interlinked socioeconomic, cultural, political, and economic specificities of place, as well as the multi-scalar and temporal dynamics, incongruities, and inconsistencies that are inherent to peoples’ lived, felt experiences of social vulnerability. By integrating these complex observations into an output that is still readily accessible to decision- and policy-makers, the new wave of indices supports the pursuit of more tailored, context-appropriate, and equitable climate adaptation. We suggest one way that these more nuanced forms of vulnerability analyses might be operationalised, by reflecting on an experimental research project that uses personas or fictional characters to examine social vulnerability to climate change in coastal Aotearoa New Zealand.

1. Introduction

Residents of cities and settlements in coastal locations are at risk from a range of climate change-related hazards and processes [1]. These include sea-level rise, severe storms, flooding and inundation, modified geomorphology (erosion, accretion, etc), marine heatwaves, increased water tables, groundwater salinity, pests and diseases, reduced sea ice in polar regions, ocean acidification, and associated transformations in species behaviour and abundance [1,2,3]. Biophysical changes to coastal environments and ecosystems are already impacting the social, cultural, economic, and political dimensions of life in coastal communities [4,5,6,7,8]; however, not all individuals, households, or social groups within these communities are equally affected or at risk [9,10,11]. Social vulnerability is a concept that is applied to understand the differential impacts of climate change on people, both in coastal and inland locations [12,13]. Social vulnerability examines how peoples’ positioning in a particular social context mediates their experiences of climate change, making them more or less likely to experience harm or adverse consequences associated with biophysical climate-related processes [14,15,16,17]. As the International Institute for Environment and Development (IIED) and the UN Office for Disaster Risk Reduction (UNDRR) observe (p. 11), “vulnerability is structured by social, economic, and political factors, particularly access to resources, social capital, and decision-making power. Vulnerability is highly differentiated by gender, sex, age, ability, ethnicity, locality, wealth, Indigenous group, and marginalisation” [18].

Scholars, organisations, and government authorities often make use of a small pool of well-established, “tried and tested” social vulnerability indices in order to identify social groups or neighbourhoods most at risk from climate change in both coastal and non-coastal areas [19,20,21,22]. Social vulnerability indices consist of a collection of characteristics (known as ”indicators”) that have a demonstrated link or causal association with a particular condition or outcome; in this case, vulnerability to climate change [23,24]. Social vulnerability indices produce quantitative, actionable assessments of vulnerability to climate change that can help decision-makers use time, money, and resources effectively to reduce harm and adapt to future change [25,26]. Although these indices have been in use for a significant time period in many diverse locations and settings, they do not necessarily provide an accurate reflection of social vulnerability as it is lived on the ground [27,28]. There is a tendency to transpose indicators from one location to another regardless of whether they align with the local social context, and many extant indices produce static ”snapshots” of social vulnerability that are fixed in time, space, and onto particular groups of people such as BIPOC (Black, Indigenous, and People of Colour) or those of lower socio-economic status [29,30,31,32].

We use this review to examine the theoretical roots of social vulnerability analysis and indices and to engage with a new wave of scholarship and praxis that seek to nuance existing approaches to quantifying climate vulnerability. This new wave consists of novel social vulnerability indices and frameworks that embrace the local social context, community-informed, participatory design, and the complexities and incongruities inherent to social vulnerability [33,34,35,36,37,38]. A significant portion of these novel approaches are applied in coastal locations [19,21], and we argue that the greater uptake of these complex, context-specific forms of analysis is crucial if climate adaptation is to be robust, successful, and equitable for coastal communities.

This review emerges from research into methods for measuring social vulnerability to climate change in the coastal locations within and beyond the shores of Aotearoa New Zealand. The research was conducted under the Resilience to Nature’s Challenges National Science Challenge, a research programme funded by the New Zealand Government to enhance the nation’s understanding and resilience to natural hazards [39]. The material presented in this paper was collated between September 2022 and January 2023. In total, we reviewed 70 indices and indicator-based assessments of social vulnerability to climate change and the hazards contained in peer-reviewed journal articles, international and Aotearoa-based policy documents, and the grey literature including reports produced by organisations such as the Intergovernmental Panel on Climate Change (IPCC), the World Health Organisation (WHO), and the United Nations Children’s Fund (UNICEF). We conducted searches using combinations of key terms such as “social vulnerability index” and “coastal” and “climate change”, which were entered into the University of Auckland’s library catalogue and Google Scholar’s search engine. Rather than conducting a systematic review, we present a narrative or thematic review of key theoretical concepts and critiques structuring the development of social vulnerability analysis and indices.

The review proceeds as follows. We begin by detailing three major “intellectual lineages” [40] that characterise international climate and hazard-related social vulnerability scholarship (Section 2). After this, we connect these vulnerability lineages to four ”foundational” indices or frameworks that are most widely and frequently used to measure social vulnerability to climate change both coastally and elsewhere (Section 3). In Section 4, we identify the new wave of scholarship and praxis that is pushing the boundaries of traditional social vulnerability indices. Section 5 concludes the review by summarising the state of the field and suggesting how these nuanced approaches to measuring social vulnerability might be operationalised via novel research focused on the creation of personas (fictional people consisting of composite characteristics) to support more robust, context-sensitive decision-making for climate adaptation.

2. Conceptualising Social Vulnerability to (Climate Change) Hazards: Lineages and Approaches

The Intergovernmental Panel on Climate Change (IPCC) defines climate change vulnerability as “the propensity or predisposition to be adversely affected. Vulnerability encompasses a wide variety of concepts and elements including sensitivity or susceptibility to harm and lack of capacity to cope and adapt” [41] (p. 826).

Sensitivity or susceptibility describes factors or phenomena that influence the degree and severity with which a group or individual is affected by climate change, while adaptive capacity or capacity to cope and adapt refers to the ability of individuals or groups to respond to and reduce harm from climate change [14]. In the last fifty years, the study of vulnerability to climate change has moved far beyond its original, limited focus on exposure to biophysical hazards, largely as a result of research highlighting how nations and social groups with fewer economic resources and socio-political privileges experienced greater sensitivity or lesser capacity to cope with and adapt to natural hazards like hurricanes or drought [12,42,43]. Three distinctive lineages of hazard scholarships (risk hazard, political economy/ecology, and human–environment/resilience) characterise this transformation, each with divergent understandings of the mechanics of vulnerability [44,45]. These lineages (and combinations of them) are the conceptual foundations upon which indices for social vulnerability to hazards and climate change are based.

2.1. Risk Hazard

The risk-hazard lineage evolved from discussions within engineering, economics, and epidemiology [44]. From a risk-hazard perspective, vulnerability arises through exposure to biophysical hazards like sea-level rise or cyclones. Although sensitivity is considered within this lineage, there is very little engagement with social context and its’ implications for vulnerability [40,44]. Risk-hazard approaches to vulnerability generally support adaptation based on technological or engineering interventions (such as sea walls or irrigation). Although these adaptations may reduce the severity of impacts associated with hazards, they often fail to target the social, political, and economic drivers of social vulnerability amongst those most at risk [46].

2.2. Political Economy/Ecology

As opposed to the risk-hazard approach, the political economy/ecology lineage explicitly seeks to identify the social dynamics and drivers that are implicated in differential vulnerability to the hazards within a population [44]. The political economy/ecology lineage has roots in the development, poverty, food security, and geography literature [44]. As such, the lineage is attentive to how combinations of biophysical processes and local, regional, and global socio-political, cultural, and economic factors increase exposure and sensitivity and/or constrain adaptive capacity for some people within a community or group [12,40,47,48,49]. As Eakin and Lynd Luers [40] observe (p. 370), “in this literature, vulnerability is not an outcome but rather a state or condition of being—and a very dynamic one at that—moderated by existing inequities in resource distribution and access, the control individuals can exert over choices and opportunities, and historical patterns of social domination and marginalisation”.

2.3. Human–Environment/Resilience

The human–environment/resilience lineage is closely linked with the field of ecology and therefore examines the vulnerability of coupled human–environment systems [40,50]. Here, vulnerability arises through the interplay of multiple stressors impacting on social and ecological systems, and it is counter-balanced by factors that enhance the resilience of or enable the system to recover, rebound, and adapt to shocks and perturbations [44,45].

2.4. Integrative Approaches

In addition to the three approaches above, there are also “integrative” methods for assessing vulnerability that combines the characteristics of the risk hazard, political ecology/economy, and/or human–environment/resilience lineages [40,44,45,51]. Integrative approaches typically examine biophysical and social systems concurrently and are widely used in climate risk assessments to measure and map vulnerable locations and populations such that adaptation resources can be channelled to those most in need [40].

3. Quantifying Social Vulnerability: Foundational Indices and Frameworks

As Mustafa et al. [25] note, theoretical analyses of vulnerability (including some of the lineages above) are highly effective at producing nuanced and qualitative data that uncovers the complex social processes driving vulnerability; however, policy-makers require simplified and often quantitative versions of these analyses to support robust decision-making. A wide range of indices have been developed to provide such quantitative assessments of social vulnerability to climate hazards in both coastal and non-coastal locations. Despite the breadth of locations and hazards these indices are applied to, the majority are based conceptually and/or methodologically upon four “tried and tested” foundational indices, which we detail below: the SoVI, Expanded Vulnerability Analysis, the SVI, and the SeVI/BeVI. The foundational indices draw on one or more of the vulnerability lineages identified above and have either been developed in or applied to coastal locations. Appendix A provides an overview of the statistical methodologies employed to formulate two of the most widely used social vulnerability indices (the SoVI and SVI).

3.1. The SoVI

The Social Vulnerability Index (or SoVI) represents one of the longest standing efforts to quantify and compare social vulnerability to hazards. Although the SoVI was created for a North American context by United States-based geographer Susan Cutter and colleagues and targets hazards broadly defined [52], the SoVI serves as the basis for many indices measuring social vulnerability to climate change globally [53,54,55], including in coastal locations [35,56,57,58]. Figure 1 demonstrates the foundational role the SoVI has played in the development of climate vulnerability indices. The SoVI has gone through a number of iterations over the years and comprises 29–32 indicators designed to capture the socio-economic, demographic, and built environment characteristics of a place, which have a bearing on vulnerability to hazards [59].

When created in the 1990s, the SoVI presented a novel approach to vulnerability. It accounted for factors that had largely been overlooked in vulnerability analyses (including income, age, and race), especially in the global north [43,60,61,62]. For instance, Cutter et al. recognised that Americans on a low income faced challenges with making savings that could otherwise be used to prepare for and recover losses related to disasters. Similarly, they acknowledged that race-based marginalisation of Black, Indigenous, and other people of colour (BIPOC) can reduce access to resources (time, money, vehicles, quality housing, etc.) that facilitate the ability to cope and respond to hazards.

3.2. Expanded Vulnerability Analysis

The expanded vulnerability analysis (EVA) framework was developed by Turner et al. [63] and exemplifies the human–environment/resilience lineage (see Figure 1). The EVA is focussed on the vulnerability of coupled human–environment systems and seeks to capture the dynamic interactions and feedback loops between systems that mediate vulnerability. The EVA was created to counter the political economy/ecology approaches to vulnerability which Turner et al. critique for neglecting both the interconnections between vulnerability in human and biophysical systems, and the agency that even the most marginalised groups exhibit when responding and adapting to change [63].

The EVA assesses vulnerability by examining exposure, sensitivity, and adaptive capacity. Exposure pertains to whom or what is exposed within a system (households, states, and ecosystems) and how often, for how long, and at what intensity. Sensitivity pertains to the systems under consideration (that is, factors such as social capital, institutions, or even soil quality that affect how deeply impacts are felt). Adaptive capacity pertains to the capacity of systems to cope, respond, adjust, and adapt to challenges or changes (via policy, decentralised actions, etc.). The EVA has been used to evaluate vulnerability in diverse locations, including Mexico’s Yucatan Peninsula and Yaqui Valley, and regions of the Greenland and Norwegian Arctic coastline [50], and it has also influenced a range of subsequent indices [38,64].

3.3. The Social Vulnerability Index (SVI)

The SVI was created by the United States (US) government Centers for Disease Control to allow the identification and ranking of census tracts that are vulnerable to hazards [65]. Although the SVI draws extensively on Cutter et al.’s [52] work, it is widely acknowledged in its own right in the hazard vulnerability literature [66] and has been employed extensively in the US to identify communities most at risk from natural disasters and climate hazards including sea-level rise, coastal flooding or inundation, tornadoes, and volcanic activity [67,68,69,70].

The SVI comprises 14 indicators clustered into four themes or domains. These are socioeconomic status, household composition and disability, minority status and language, and housing and transportation [66]. Despite providing a quantitative measure of vulnerability, the SVI has been developed through extensive validation to ”field-test” and verify the relevance of each indicator, and it is updated every two years to reflect the social and demographic trends [66,71,72,73,74].

3.4. The SeVI and BeVI

The Socioeconomic Vulnerability Index (SeVI) and the Built Environment Index (BeVI) were developed by Holand et al. in 2011 in order to evaluate social vulnerability to natural hazards in Norwegian municipalities (where many inhabitants live in coastal locations) and to direct hazard mitigation and climate adaptation [75]. The SeVI and BeVI take essential components of the SoVI (socio-economic and build environment characteristics) and measure these separately in an attempt to provide a more holistic analysis of vulnerability in a place [75]. As Holand et al. note, this approach aims to allows the user to pinpoint where crucial variables involved in the production of vulnerability lie (be it socioeconomic or built environment systems) as opposed to measuring these variables together within one index, which can result in artificially high overall measures of vulnerability when in fact one or the other system is exerting a disproportionate effect on vulnerability.

The 25 indicators for socioeconomic vulnerability in the SeVI focus on variables such as income, employment, age, population change, and gender. The eight built environment indicators in the BeVI include housing and population density and measures of infrastructure and public services. Although developed for a Norwegian context, the SeVI and BeVI have been applied internationally and the UNDP draws on the SeVI/BeVI model in official guidance on how to construct a social vulnerability index [15].

4. Nuance and Novelty in Social Vulnerability Indices

Despite the ongoing and widespread influence that foundational indices exert in the field, scholars, applied researchers, development, emergency management, and climate adaptation practitioners have developed alternative approaches to quantifying social vulnerability, with many targeting coastal locations. A number of key critiques have been applied to foundational indices which have spurred methodological and conceptual innovation such that indices generate more nuanced readings of (climate) vulnerability and enhance the efficacy of related adaptation decision- and policy-making. Below, we review three dimensions of this innovation: greater engagement with social context, capturing dynamism in vulnerability, and accounting for adaptive capacities and strengths.

4.1. Greater Engagement with Social Context

Indices like the SoVI have been critiqued for spurning a trend towards generalisability in vulnerability analysis, whereby a standard suite of indicators are readily applied to a wide range of different locations around the world. Indeed, Susan Cutter’s home institution, the Hazards and Vulnerability Research Institute (University of South Carolina), hosts a “SoVI Recipe” on its website [76] that gives instructions on how to replicate the SoVI via statistical methodologies like the Principle Components Analysis (PCA)—see Appendix A for further details. Whilst standardised approaches may allow for rapid assessments of vulnerability and enable comparison between regions, scholars like Nguyen et al. [33] and Działek et al. [30] state that standardisation precludes the consideration of the specific variables that influence social vulnerability in different contexts, including coastal locations. The same problem arises when scholars/practitioners produce vulnerability assessments deductively [29] by selecting indicators of social vulnerability because they are frequently used in major indices like the SoVI or SVI [26,77,78,79] or because the literature demonstrates a correlation between the indicator and social vulnerability [67,80,81,82] as opposed to the indicators being drawn from engagement with the community/ies under consideration.

The use of inductive, qualitative methodologies in index design [29] is one way that scholars and/or practitioners have sought to overcome the limitations of generalisability and deductive vulnerability analyses. Household surveys, interviews, and workshops with affected community members, key stakeholders, and decision-makers are used to identify locally relevant drivers of vulnerability which serve as the basis for index development [24,26,34,79,83,84,85]. For example, Nguyen et al. [33] used “expert judgement” to develop a Social Vulnerability Index in coastal Vietnam that is based upon stakeholders’ views of the most salient indicators. Stakeholders included local environmental professionals, decision-makers, and local village leaders who participated in in-depth interviews and focus groups where they discussed and evaluated the most pertinent causal factors associated with local vulnerability.

Collaborative validation is another strategy used to enhance the local specificity of vulnerability indices. Often, scholar/practitioners identify a potential suite of indicators using existing research or other social vulnerability indices and then work with community members, local stakeholders, or other experts to “field-test”, “ground truth”, or validate the indicators [23,73,86,87,88]. Oulahen et al. [55] used this approach to develop a social vulnerability index for coastal climate hazards in Vancouver. A baseline set of indicators were created using Cutter’s hazards of place model [89], the SoVI [52], and Fussel’s framework for social vulnerability [44]. The indicators were then reviewed using focus groups and surveys with local hazard-planning practitioners and was amended as necessary. Some indices also use validation to assign “weight” or particular importance to some indicators within an index [19].

Grounded, qualitative, and participatory methodologies are now internationally recognised as best practice for social vulnerability index design [15,90]. For instance, in a guide to creating social vulnerability indices (aimed at disaster risk reduction, climate risk management, and climate adaptation practitioners) the United Nations Development Programme (UNDP) recommends testing and validating the index [15] after the initial indicator selection. Ideally, the UNDP notes, validation should be both quantitative (for example, using historic disasters as test cases), and qualitative (for example, using interviews, focus groups, and the Delphi method). The Delphi method is aniterative, participatory technique to reach a decision that is based upon deliberation and consensus between experts. The technique is often used in social research to explore the implications of policy and alternative scenarios or solutions [91]). The UNDP also recommends the final index should be based on qualitative data (like household surveys) and quantitative formation (like the census). Engaging participatory approaches to index design is one way to enhance the likelihood that the measures of vulnerability will align with local peoples’ lived realities and successfully identify some of the major driving forces of social vulnerability specific to a particular locale. In turn, this may support adaptation decision-making to target households, groups, or neighbourhoods that are most at risk, as opposed to developing standardised, ”one-size-fits-all” plans.

4.2. Capturing Dynamism in Vulnerability

Foundational social vulnerability indices (and indices based upon them) have also been critiqued for producing static “snapshots” that overlook or simplify the dynamic nature of vulnerability [20,21,27,28,31,44]. Critics note that indicators are preoccupied with analysing variability in social vulnerability across space (for example, between different communities) whilst overlooking shifts in socio-economic, demographic, and institutional dynamics over time (including policy interventions, population, or behaviour change) [20,27,28,57]. Neglect for the temporal dynamics of vulnerability is often associated with methodology, since the majority of indicators are quantified through recourse to government statistics like the census [52,54,75,79,92]. Such databases present a snapshot of population demographics and behaviours at a particular point in time and can quickly become outdated [77].

In addition, most major indicators target a single spatial scale (be it a neighbourhood, city, county, or region) and a single pressure (like climate change) [27]. They neglect cross-scale linkages between the local/regional and global, as well as the interaction of multiple pressures on a system (such as pollution, pandemics, or policies), which together influence how social vulnerability is experienced at any given scale and how this experience changes over time [19,22,24,85,93,94].

As a result of these oversights, a range of indices approach social vulnerability with dynamism in mind and account for both the interacting scales [34,35,36] and social change over time [54,56,57,87,95,96]. A number of these indices have been developed in coastal locations [34,57,87,95]. For example, Kashem et al. [35] paired theories of neighbourhood change with a social vulnerability index to evaluate the evolving nature of social vulnerability to hazards amongst neighbourhoods in three US coastal cities. Kashem et al. demonstrate how the demographic composition of particular neighbourhoods has changed over time and become home to socially vulnerable populations such as elders and migrants. Interacting local, regional, and global influences—including policies promoting facilities and services designed to attract retirees from elsewhere, and the development of manufacturing premises providing goods for global markets that are dependent upon cheap, migrant labour—coalesce to form and re-form socially vulnerable landscapes at the neighbourhood scale [35].

Indices and frameworks developed by intergovernmental organisations also engage with dynamism in climate-related social vulnerability [97,98,99]. For instance, the Children’s Climate Risk Index, developed by UNICEF [100], has a strong focus on the policy/institutional dimension of vulnerability, reflecting elements of the political economy/ecology and human–environment/resilience lineages. The Children’s Climate Risk Index is the first set of indicators dedicated to identifying children’s vulnerability to climate change at a global level. A significant proportion of indicators target government structures and policies (including domestic health and education expenditure or the provision of Water, Sanitation, and Hygiene services) that are influenced by international agendas, such as the Millennium Development Goals, and change over time as a result of the flows of international aid, national development, and other interacting pressures like conflict and migration.

The International Union for Conservation of Nature and Natural Resources takes a different approach to social vulnerability that also engages with dynamism. The Rapid Assessment of Climate Change Vulnerability and Adaptation Planning [101] is a participatory tool intended to assist communities in the Indo-Burma region to evaluate their social vulnerability to climate change. The framework sets out a methodology for villagers whose livelihoods depend on wetlands to score their exposure, sensitivity, and adaptive capacity in a way that accounts for interactions between scales and changes over time. Drawing on the human–environment/resilience lineage, the framework guides villagers to think about how external stressors (like the encroachment of agriculture into wetlands or upstream dams) are impacting their livelihoods and may interact with climate change and how existing and planned forms of adaptation might alter their vulnerability to climate change into the future [101].

Accounting for dynamism can aid the development of adaptation strategies that reflect current demographics and, moreover, are flexible and holistic. For instance, considering how local vulnerability is influenced by shifting regional, national, and international social, political, economic, or environmental phenomena, it may enable the identification of a wider variety of households or groups in need of adaption support than would have been identified by simply examining the local social dynamics. Attending to the potential for changing vulnerability over time also builds in opportunities to restructure and re-prioritise adaptation resources as needed and may help to avoid common pitfalls associated with adaptation such as path dependency (where adaptation is committed to a particular course of action that could have negative consequences over time) [102] or maladaptation (where adaptation increases rather than alleviating vulnerability) [103,104].

4.3. Accounting for Heterogeneity and Adaptive Capacities

Social vulnerability indices have also been critiqued for overlooking internal heterogeneity within vulnerable social groups and downplaying or ignoring adaptive capacities, strengths, or assets they may possess [26,37]. Most foundational indices (and the indices they inspire) score, rank, or otherwise quantitatively compare social vulnerability by adding together characteristics known to enhance vulnerability (such as poverty, limited education, older age, or limited fluency in the local language). By equating vulnerability with broad social characteristics, these indices treat all members of a social group as vulnerable and fail to acknowledge evidence that vulnerability is far more nuanced in reality [105,106]. Empirical research demonstrates that social vulnerability is fluid and multi-faceted, whereby people in the same locale who share a common identifier such as age, indigeneity, or gender have diverse experiences of the same climate hazard owing to other aspects of their personhood such as marital, health, or employment status [107,108,109]. The different facets of personhood can increase the exposure and sensitivity to climate hazards, but also enhance the capacity to cope and adapt, and it is these intersections of personhood that produce unique (and changing) experiences of hazards within broad demographic or social groups [110,111].

The relative lack of consideration for adaptive capacities within most social vulnerability indices compounds the issue [32] and could mean indices produce erroneous conclusions and misdirected policy [13]. Despite a wealth of scholarship attesting to the capacity of otherwise vulnerable groups to cope and adapt to climate change [112,113,114], most social vulnerability indices do not explicitly measure life experiences, values, norms, behaviours, skills, social networks, and other characteristics that groups who are considered vulnerable engage and enact to respond to and overcome climate hazards.

In a special report on climate change vulnerability, the IPCC [90] states (p. 80) that, “static checklists of vulnerable groups do not reflect the diversity or dynamics of people’s changing conditions”. On page 81 of the report [90], the authors continue that, “the vulnerability ‘label’ can reinforce notions of passivity and helplessness, which obscure the very significant, active contributions that socially marginalised groups make in coping with and adapting to extremes”.

Scholars and practitioners have developed a range of social vulnerability indices that engage with adaptive capacities, strengths, and heterogeneity [37,38,97,101]. For example, the Integrated Vulnerability Assessment [99], co-developed by SPREP, national and regional experts, and agencies within the Pacific, is guided via a sustainable livelihoods approach that emphasises the capacity of low-income, island communities to improve their circumstances, wellbeing, and reduce vulnerability using diverse, locally relevant means [115,116,117]. This assessment parallels other social vulnerability indices that specifically emphasises strength or resilience amongst less privileged social groups, communities, or regions [37,118].

A small pool of social vulnerability indices engage internal heterogeneity through the use of composite characters that combine a variety of social characteristics to produce more nuanced measures of social vulnerability [26,36]. For example, Atyia Martin [26] developed an index for social vulnerability to coastal hazards in Boston that conveyed how intersecting social categories elevated vulnerability. Atyia Martin identified that gender, housing tenure, transportation access, and health status clustered together, leading to greater vulnerability amongst female renters with no car and a health condition, rather than vulnerability simply arising from being a renter. Despite this progress, these indices still focus primarily on deficits or characteristics that enhance vulnerability, leaving the adaptive capacities under examined.

5. Reflection: State of the Field and Applying New Directions for Vulnerability Analysis in Coastal Locations

Since the notion of social vulnerability emerged in the 1970s, scholars, applied researchers, and hazard and climate adaptation practitioners have sought to conceptualise and quantify who and where is most at risk and why. Three major vulnerability lineages (risk hazard, political economy/ecology, and human–environment/resilience) provide the conceptual basis upon which most social vulnerability indices and assessments have been built [44,47,63]. Many indices worldwide take an integrative approach, combining insights from two (or more) of these three lineages and often integrate measures of biophysical exposure to (climate) hazards alongside social and demographic-focussed indicators [40,51]. Several foundational indices (the SoVI, expanded vulnerability analysis, SVI, SeVI, and BeVI) have been instrumental in the field, guiding the content, focus, and methodologies of subsequent generations of social vulnerability indices [52,63,66,72,75] both at the coast and inland. However, the field is advancing and maturing, and it is possible to detect a new “wave” of indices that respond to the critiques of foundational indices like the SoVI.

This new wave seeks to overcome a number of pitfalls within foundational indices, including issues with generalisability, the repetition of standardised indicators for social vulnerability regardless of social context, static portrayals of vulnerability, limited engagement with adaptive capacities, and internal diversity within vulnerable groups [21,28,29,30,32,33,37]. Indices and assessments developed by university institutions, governments, and intergovernmental organisations engage participatory methodologies to ensure the indicators are grounded in local social context, and they integrate indicators designed to capture the multi-scalar and temporal dynamics of social vulnerability [15,35,55,100,101]. Indices and assessments also provide more holistic suites of indicators that balance the factors exacerbating vulnerability alongside those (knowledge, skills, social norms, and behaviours) that reduce risk [38,90,99,118]. Additionally, a small sample of this new “wave” have developed techniques to identify heterogenous experiences of vulnerability within social groups that offer a promising pathway to convey how climate risk is mediated through intersections of different social categories, rather than being fixed onto particular characteristics like poverty [26,36]. A significant proportion of the new wave of indices are developed for and applied to coastal locations and present opportunities for adaptation to better support those at risk in coastal communities by generating context-rich and accurate observations of vulnerability that are packaged into a (semi) quantitative format that policy- and decision-makers are familiar with.

One way to further operationalise this more nuanced approach to social vulnerability found within the new wave is through the creation of personas. As Pruitt [119] observes (p. 1), “personas are fictitious, specific, concrete representations of target users” and are typically engaged in market research to think through consumer needs, demands, and experiences with new products. Despite their relative lack of use in climate research and practice internationally [120], personas offer an alternative framework for understanding social vulnerability that aligns with the new wave of indicator scholarship and strikes a balance between the high-level, generalisable observations of social vulnerability produced by indices like the SoVI and SVI and the deeply grounded, highly specific accounts of vulnerability emerging from research within human geography, anthropology, and so on [60,121,122,123,124]. Below, we briefly reflect on an emergent and experimental research project that utilises personas to better understand social vulnerability in coastal communities in Aotearoa New Zealand. Whilst this work is rooted in the socio-political, cultural, economic, and environmental context of coastal Aotearoa, the concept and methodology are applicable internationally and provides a pathway for more refined vulnerability analyses that can support appropriate, inclusive, and equitable adaptation decision-making.

Despite the New Zealand government’s recent efforts to evaluate and plan for the impact of climate change on society [125,126,127], there are still only a handful of dedicated climate-focussed social vulnerability indices and assessments in existence for Aotearoa [78,86,128,129,130,131]. Since a high proportion of New Zealanders live in coastal locations, most extant indices and assessments apply to coastal communities which are at risk from a range of climate change impacts including sea-level rise, coastal inundation and erosion, storms, cyclones, extreme rain events, fluvial flooding, and the knock-on effects of ocean acidification, marine heatwaves, and increases to marine pests and diseases [126,132]. Aotearoa-based vulnerability indices and assessments vary in their level of nuance, with some exhibiting characteristics of the new wave, including participatory design and validation [86,129,133]. However, most draw heavily on indicators that comprise foundational indices like the SoVI [78,130,133] and reproduce other pitfalls associated with the “tried and tested” methods for assessing vulnerability, including the oversight of dynamism [78,128,130] and the tendency to ascribe vulnerability onto whole social groups such as Indigenous and other non-European groups, the elderly, or those of lower socio-economic status, without examining the internal diversity or adaptive capacities that may exist within these groups [78,86,126,129,130].

In order to facilitate more accurate understandings of social vulnerability in the coastal locations of Aotearoa, we are currently engaged in an experimental research project that pairs the use of personas with climate science and agent-based modelling. Our interdisciplinary research team is based at New Zealand’s National Institute of Water and Atmospheric research (NIWA) and comprises climate scientists and social scientists with expertise in modelling, climate risk and vulnerability assessment, serious games, participatory planning and climate adaptation, and community-based and ethnographic research. The work draws in part on the Adaptive Futures game developed by Blackett et al. [134,135], which has demonstrated the value of personas for enhancing inclusive and context-appropriate adaptation decision-making. Adaptive Futures is an interactive online serious game or simulation that is designed to build players’ capacity for responding to the complex challenges that climate change presents for the fictional coastal community of Seaview. One of the key elements of the game is to balance the various adaptation options against the diverse needs and interests of the “non-player characters” who populate Seaview. The non-player characters have differing experiences of climate change and adaptation preferences based on their physical location, identity, values, interests, and attitudes. Evaluation of the Adaptive Futures game shows that the presence of personas in the form of non-player characters challenges game players to consider how adaptation can accommodate the heterogeneous experiences of Seaview residents.

Our current research supports this type of complex, inclusive, and grounded adaptation decision-making. Drawing on the most frequently used indicators of social vulnerability to climate change that are found within the indices reviewed for the Resilience Challenge work described above and our combined 50 years of social and climate research in coastal communities [123,136,137,138,139,140,141,142,143,144,145,146,147,148,149], we developed a framework for creating personas. The framework (Figure 2 below) comprises a series of identity markers (such as socio-economic status, education, race, renter, or homeowner) that contribute to social vulnerability and that we call “positionality”, a range of adaptive capacities, skills, or assets that offset vulnerability, and a category for beliefs and values. These three elements denote the potential combinations of characteristics that personas may hold and which may influence the personas’ particular adaptation preferences. From the three spheres of the framework and our combined research experience with coastal communities, we have identified likely combinations of characteristics to form personas, each of which are more or less vulnerable to the impacts of climate change according to their positioning vis á vis several key variables (including social connections, net disposable income, mobility and health, willingness to change, personal agency, trust in mechanisms of the state, and experiences of discrimination). Our next step will involve agent-based modelling to determine how the personas will be differentially impacted and vulnerable to the impact of coastal climate hazards over a series of ten-year increments.

We acknowledge that the creation of these personas is based upon gross generalisations of the types of people found residing in coastal communities. However, by collating different characteristics that influence social vulnerability into a single persona, there is potential for more nuanced readings of climate vulnerability than are currently present in Aotearoa New Zealand. For instance, Māori, the Indigenous peoples of Aotearoa, are often identified as vulnerable to the impacts of climate change because of their close social, cultural, spiritual, and economic connections to the natural world, as well as their relative social disadvantage when compared with non-Māori [125,126,150,151,152]. One of our personas (a well-educated, high income, socially connected, and influential Māori person) presents an alternative narrative by showing that internal diversity exists within Māori society, and that strategies for vulnerability reduction need to be tailored to specific circumstances rather than treating entire groups as vulnerable.

Similarly, we approach the experiences of those involved with coastal horticulture and agriculture in a range of ways, rather than simply assuming that all farmers and farm-workers are equally vulnerable to the loss of income from climate change (as is often posited by vulnerability assessments) [125,126]. We have developed three separate personas that work on farms, who are more or less vulnerable or at risk from different hazards, depending on their positionality, adaptive skills, and outlook or worldview. For instance, one farmer has a long history of residence in a particular landscape, good equity in property, a diversified farm, and connection to the market, which endow them with a high level of practical skills, a large social network, access to savings, and a quality home. These facets of their identity together enhance their ability to continue production through adverse weather events, whilst insulating them from climate hazards like storm damage, flooding, or inundation. A younger farmer with higher debt, a shorter history within a landscape, a less diversified portfolio, and an older or lower quality home may be more vulnerable to these same climate hazards. Yet, the younger farmer may also possess fewer fixed habits or engrained processes than the more senior farmer, which may in turn enhance the ability for the younger farmer to change direction quickly in response to climate hazards and even take advantage of new opportunities presented by climate change. And, whilst both farmers may possess large social networks through professional and personal connections, the younger farmer may have a higher propensity to seek emotional support through their network than the more senior farmer, and therefore be less at risk of mental health burdens that are increasingly associated with farming in a rapidly changing climate [153,154,155,156].

The use of personas offers a novel approach to social vulnerability assessment that strikes a balance between the detail of deeply grounded, place-specific accounts of vulnerability and the high-level observations of vulnerability generated via foundational indices, indicators, and their progeny. In combination with agent-based modelling, our use of personas offers an accessible way for policy- and decision-makers to grapple with and gain insight into the incongruities and inconsistencies of social vulnerability to climate change that are not otherwise conceptualised via social vulnerability indices, including the ability of otherwise vulnerable groups to cope with, respond, and adapt to hazards, and for vulnerable people to become more or less vulnerable over time. The project presents opportunities for adaptation strategies to become more responsive to and inclusive of heterogeneous needs, interests, and perspectives within coastal communities, and to ultimately pursue actions that are not only more effective but also likely to be more equitable.