1 Introduction

Climate change (CC), alongside with rapid population growth and natural and manmade disasters, require modern approaches to be addressed; forward-looking design can have a profound impact on safety and prosperity of individuals and communities as a whole (Rus et al. 2018). In everyday life, the term “vulnerability” is widely used. Comte et al. (2019) claim that a consistent definition of vulnerability is not available. It is difficult for a definition to include every assessment context, as it is used in different policies and systems exposed to diverse hazards (Nguyen et al. 2016). However, related to CC, the Intergovernmental Panel on Climate Change (IPCC 2014) definition of vulnerability has been acknowledged by researchers (e.g., Li et al. 2019; Gargiulo et al. 2020). It is defined as “the propensity or predisposition to be adversely affected. Vulnerability encompasses a variety of concepts and elements including sensitivity or susceptibility to harm and lack of capacity to cope and adapt”.

According to Gargiulo et al. (2020), coastal vulnerability can be considered as the integration of sensitivity and social vulnerability of coastal areas. For this purpose, physical, socio-environmental and socio-economic aspects of coastal areas are taken into consideration. The analysis of vulnerability is occasionally carried out using complex indices (Gargiulo et al. 2020). Vulnerability is also strictly linked with adaptation, as the adaptive capacity is described as the ability of systems, institutions, humans and other organisms to adjust and respond to potential damage, for example due to weather extremes (IPCC 2001, 2014).

In terms of social vulnerability, characteristics of communities as well as inequities indicate an important factor in the ability of groups to respond to natural hazards and climate challenges (Cutter et al. 2003). Especially in coastal areas, social inequities, income, race, ethnicity, and demographic data are key factors in people’s vulnerability to environmental hazards (Li and Li 2011). Economic losses might be direct, affecting buildings, crucial infrastructure and their operations (Asariotis et al. 2017), or indirect, affecting local economy, causing issues in the supply chain, unemployment, investment decisions. The use of susceptibility functions (or damage functions) is the most frequently applied method (Meyer et al. 2013) to assess the direct costs related to natural hazards (e.g., Pistrika et al. 2014). Also, to assess indirect costs, economic models at local, regional or national level are used (Meyer et al. 2013). Financial costs vary from one place to another, as developed countries have higher property values and disaster insurance is affordable (Christian Aid 2021). However, it is rather difficult to define disaster cost, as it highly depends on the purpose of the assessment and the consequential uncertainties regarding the indirect cost (Hallegate and Przyluski 2010).

One question that needs to be addressed is “Why to use index-based approaches?” Indices are “measurable, observable quantities that serve as proxies for an aspect of a system that cannot itself be directly adequately measured” (McIntosh and Becker 2017). Abstract concepts like vulnerability and resilience can be visualized in spatial scales, using observable variables (McIntosh and Becker 2019). Within this approach, vague factors like “social factors” could be specified, and thus, quantified using multiple variables (e.g., age, education, unemployment) (Malone and Engle 2011).

In this study, a systematic review of assorted approaches applied globally is presented. In Sect. 2, the methodology followed in the literature review is analyzed. Section 3 discusses climate change vulnerability and resilience indices for coastal areas found in the literature. Section 4 briefly presents aggregation tools for the evaluation of the abovementioned indices. Finally, Sect. 5 presents the main conclusions and highlights gaps in the literature.

2 Review Methodology

The goal of the review was to identify and analyze relevant research of index-based methods for the evaluation of climate change vulnerability and resilience of coastal areas. Understanding and quantifying the risk is the main challenge that occurs, and is the first step in the development of an effective framework to address real-world cases. The PRISMA 2020 methodology by Page et al. (2021) was followed, as it includes both systematic reviews and meta-analyses. We searched, retrieved, classified and reviewed papers on climate-change vulnerability and resilience indices related to coastal water systems and relevant infrastructure. For this, Scopus, Thompson-Reuters Web of Science, PubMed, ScienceDirect, Google Scholar and other relevant databases were used. The search strategy included free text terms such as “climate change vulnerability index”, “coastal vulnerability indices”, “resilience of coastal areas to climate change”, “climate change resilience index”, “climate change and socio-economic indicators”, “coastal flood vulnerability”, “climate change and coastal flooding”. A summary of 1397 titles was revealed and screened by a single reviewer. The majority of the literature was in English. One hundred thirty-six (136) records were assessed for eligibility to be included in the study and 46 were finally included. The PRISMA flowchart is presented in Fig. 1.

Fig. 1
figure 1

PRISMA flowchart – Study methodology of excluded and included articles

Full size image

3 Addressing Vulnerability and Resilience in a Changing Climate

It is widely believed that coastal zones are exposed to diverse natural hazards under a changing climate. Coastal areas and infrastructure should be designed to address these challenges (Malliouri et al. 2017). The exposure level is affected by the land use of coastal areas, the population and social factors, the shoreline type and the existing (and planned) infrastructure. Current and future socio-environmental and socio-economic data are used to develop different scenarios (Fig. 2). Also, there is a significant variety of presumed socio-economic impacts of climate change among key coastal sectors. The environment is described with hydrologic, atmospheric and geologic variables. As a result, vulnerability components could be identified in spatial scale (Bevacqua et al. 2019). Using different projections for future scenarios could also add a temporal scale to the vulnerability assessment.

Fig. 2
figure 2

Flowchart for the definition of exposure and socio-environmental scenarios (Modified from Toimil et al. 2017)

Full size image

Important components affecting coastal areas are the astronomical tide, the storm surge, the wave run-up, the weather extremes and the sea-level rise. CC affects the frequency of storm surge events and the magnitude of wave heights, having crucial impact on coastal and port infrastructure (Tsoukala et al. 2016; Afentoulis et al. 2017). The main impacts of these phenomena are coastal erosion and coastal flooding. The majority of the literature analyzes risk caused by SLR and storm surges, while wave data have only been used by a few researchers (e.g., Gornitz et al. 1994; Dawson et al. 2009; Toimil et al. 2017). The use of historical weather data, as well as climate projections downscaled at the regional level have been used extensively in the literature to model the impacts of CC in various case studies (Fig. 3). However, climate projections and bias correction methods still have significant uncertainties (Kourtis and Tsihrintzis 2021).

Fig. 3
figure 3

Flowchart for the creation of hazard maps (Modified from Toimil et al. 2017)

Full size image

Social variables include population, nationality, economic status, age, gender, household type, housing quality, and health. Coastal areas are particularly exposed to climate change and extreme weather events, and the dependence of many regions on touristic and recreation activities intensifies the vulnerability level (Bevacqua et al. 2019). This also creates an uncertain environment for touristic activities (Uyarra et al. 2005).

3.1 Tools for Aggregation/Evaluation of Variables

The majority of the approaches in the literature promote a static understanding of human–environment interactions. Ford et al. (2018) raised similar concerns. The recent crisis caused by the COVID-19 pandemic is an example of the importance of dynamic approaches when social indicators are used. Hinkel (2011) also supported that, as useful as they can get, vulnerability and resilience indices solely cannot direct decision makers and stakeholders to raise public awareness, propose adaptation measures and monitor adaptation policy and funds.

A holistic approach is necessary to overcome the abovementioned concerns. This creates the urge to aggregate the divergent aspects of geophysical and social vulnerability into a sole index, or a limited group of indices. Integrating data from diverse sources is frequently necessary to understand potential relationships among them (Preston and Jones 2008). Aggregating variables reduces the amount of complexity, and the indices, often described as a “vulnerability score,” visualize the climate change vulnerability (Nguyen et al. 2016).

Flood risk management and defense projects interact with various sectors (socio-economic, environmental, etc.) of an area wider than the affected one. Different stakeholder groups are involved and expected to have diverse opinions about the projects' magnitude and the alternatives which seem viable. Therefore, policy and decision-making on coastal flood defense and risk management at strategic level depend on many criteria and stakeholders (Maragoudaki and Tsakiris 2005).

Weighting and ranking methods are used to express the relative importance of different individual variables in a system. With Multi-Criteria Decision Analysis (MCDA), it is rather possible to overwhelm many of the disadvantages of conventional evaluation methods. It allows for factors that cannot easily be expressed in monetary terms or quantified, but nevertheless play a decisive role in shaping policies, such as environmental, spatial and social impacts of a project, social justice etc., which should always be taken into consideration (Roukounis et al. 2020; Roukouni 2016).

Especially the Analytic Hierarchy Process (AHP) by Saaty (1987) is frequently met in the literature. Ellen et al. (2016) suggested that in climate planning, multi-criteria analysis has the ability to perform more efficiently than cost–benefit analysis and working group approaches during the decision-making process, as they are more comprehensive and scenario-driven.

GIS can be used to produce different kinds of hazard and vulnerability maps, as researchers integrate indicators regarding various aspects of vulnerability. To locate ‘hot spots’ of vulnerability, measures of exposure, sensitivity and adaptive capacity are combined and mapped. On the other hand, uncertainties in the model could also be identified (Malone and Engle 2011).

3.2 Different Approaches to Coastal Vulnerability and Resilience Indexation

In order to establish new approaches in vulnerability and risk analysis, the development of solidly identified and quantified variables could be effective (Brooks 2003). Index-based methods offer a simple way for the classification of alternative options during decision making procedures (Giannakidou et al. 2020).

One major challenge in climate vulnerability research is to define the subject of interest. Initially, research was focused on the physical profile of climate change vulnerability (e.g., Gornitz 1991; Thieler and Hammar-Klose 1999), developing robust indices that estimate the exposure of coastal areas. However, the understanding that climate change is a concern of humankind globally (UN 2015) had raised the question “who is vulnerable” rather than “what”. The assumption that people lacking wealth and resources are more susceptible to be damaged by climate change impacts has been common among researchers (Malone and Engle 2011). Indeed, Islam and Winkel (2017) illustrated how social inequality affects the ability of disadvantaged groups to respond to climate change hazards.

In order to address the abovementioned challenges, indicators should be consistent and coherent, based on robust data, using multiple criteria in the selection process. A holistic approach is crucial in order to assess coastal vulnerability, as excluding features may lead to unrealistic results (Li and Li 2011). Furthermore, high spatial resolution is crucial for the effectiveness of vulnerability indices (Gargiulo et al. 2020), taking into consideration factors that cannot be evaluated at large scales (Torresan et al. 2008).

The index-based approaches have been the most widely accepted methods to assess vulnerability (Debortoli et al. 2019; Blasiak et al. 2017; Nguyen et al. 2016; Balica 2012; Jun et al. 2011; Baker 2009). In complex human and other natural systems, more than one factors are present. The exposure and vulnerability of these environments is estimated by integrating multiple factors of every part of the system. As a result, these approaches are adopted more frequently than single-factor methodologies (Giri et al. 2020). In the present research various vulnerability indices were identified. The literature review showed that the majority of vulnerability indices were developed considering the geomorphological and socio-economic factors as main features.

In order to evaluate the geophysical characteristics of coastal areas and to allocate the future impacts of climate change among divergent coastal sections, the Coastal Vulnerability Index (CVI) was developed (Gornitz 1991). Different versions of the CVI were applied first on the coasts of the USA, Canada and Mexico (Gornitz et al. 1994; Thieler and Hammar-Klose 1999). This index was also applied, after modifications matching local preferences and data limitations in various other locations (Shaw et al. 1998; Doukakis 2005; Abuodha and Woodroffe 2006; Mendoza and Jiménez 2009; Gaki-Papanastassiou et al. 2010; Karymbalis et al. 2012; Yin et al. 2012; Kokkinos et al. 2014; Satta et al. 2015; Pantusa et al. 2018; Tragaki et al. 2018; Hawchar et al. 2020).

Pendleton et al. (2004) developed a weighted CVI method, based on Multi-Criteria Analysis, and more specifically, the AHP. Similar approaches to the CVI include those by: Bagdanavičiute et al. (2015); Devoy (2008); Mani Murali et al. (2013); Pendleton et al. (2004); Rao et al. (2008); Tibbetts and van Proosdij (2013); Addo (2013); Mani Murali et al. (2013); Bonetti et al. (2013). A summary of the studies implementing indices with geophysical characteristics is presented in Table 1.

Table 1 Geophysical Coastal Vulnerability Indices found in the literature
Full size table

A variety of physical characteristics have been included to these indices, in order to evaluate vulnerability levels of coastal areas to different climate change-related hazards. Geomorphological (slope, elevation, shoreline type, rock type, solid and drift geology, coastal erosion/accretion rate), sea (SLR, tidal range, storm surge), wave (wave height, wave run-up) and weather data, alongside with information regarding land cover, existing coastal flood protection measures and other environmental agents are used for this purpose. A summary of the physical variables used in the literature is presented in Table 2.

Table 2 Geophysical Variables used in the literature
Full size table

Climate change and socio-economic development are profoundly connected, as human activities have been the main driver of environmental harm. Socio-economic variables are used to determine the impacts of climate change on the society. These variables broaden physical indices to an inherent cultural bias (McLaughlin and Cooper 2010). About 10% of people who reside at low elevation coastal zones, located up to 10 m elevation of present-day mean sea level (MSL) (Wahl et al. 2018), face a great threat due to sea level rise (SLR) (McGranahan et al. 2007). Thus, identifying social vulnerability of coastal communities is crucial for planning resilience strategies (Tragaki et al. 2018). Social vulnerability may include characteristics of the communities and the built environment, such as the urbanization level, the growth rates, and the economic stability and development (Cutter et al. 2003). As it is believed that vulnerability depends on the sensitivity of both the natural and social components of a system (Heinz Center 2002), social inequities are key factors in natural hazard vulnerability assessment (Li and Li 2011). Characteristics of communities as well as social inequalities have also been taken into consideration. Socio-economic parameters such as demographic data, education level, population density and land use are constantly used in the literature to estimate social vulnerability. Cutter et al. (2003) used qualitative and quantitative data related to 42 variables (that were then reduced to 11 independent factors) to estimate the level of social vulnerability within American metropolitan coastal areas, creating the Social Vulnerability Index (SoVI). This index has been the baseline for the vast majority of the indices in the literature that take social factors into consideration. Alexandrakis et al. (2014) used the term socCVI, in an effort to combine the physical parameters of CVI with socio-economic factors. Guillard-Goncąlves et al. (2015) readapted SoVI to suit the social context of their case study in Lisbon, Portugal. Kleinosky et al. (2007) used a variety of socio-economic parameters within scenarios of sea-level rise to assess storm-surge flooding vulnerability level. In the approach of Tate et al. (2010), frequent hazards define a higher level of risk, as monetary losses are depicted via crucial infrastructure damage. A summary of the social coastal vulnerability indices used in the literature is presented in Table 3.

Table 3 Social Coastal Vulnerability Indices found in the literature
Full size table

Composite indices composed of various index types, such as geophysical, social, economic (i.e., socio-economic, socio-enviromental) are also frequently used. From simpler (Wu et al. 2002; Boruff et al. 2005; Chakraborty et al. 2005; Kantamaneni 2016) to more complex indices (Li and Li 2011; Satta 2014; Zanetti et al. 2016), recent researches tend to consider multiple factors. More and more variables describing climate data, wave and wind, geomorphology, land use/land cover, education, unemployment, and crucial infrastructure in coastal areas are used, with the proper adjustments for each case study (Briguglio and Galea 2003; McLaughlin and Cooper 2010; Mackey and Russell 2011; Balica et al. 2012; Mani Murali et al. 2013; Satta 2014; Ahsan and Warner 2014; Kantamaneni 2016; Zanetti et al. 2016; Calil et al. 2017; Toimil et al. 2017; Debortoli et al. 2019; Gargiulo et al. 2020; Edmonds et al. 2020; El-Zein et al. 2021). Especially in urban coastal environments (e.g., Zanetti et al. 2016; Gargiulo et al. 2020), the combination of different variable types seems to provide a more dynamic approach to human–environment interactions. As index-based methods are believed by some authors (e.g., Ford et al. 2018) to consist a static approach, the use of combined indices is a step to fill this gap, since human–environment interactions in coastal urban environments are highly dynamic. Especially in coastal cities, natural systems are distorted by human actions and co-exist with the built environment (Malvarez et al. 2021) in a complex definition of resilience (Masselink and Lazarus 2019). However, some authors suggest that one single complex index is not representative and prefer the use of multiple single indices instead (e.g., Tragaki et al. 2018). A summary of composite indices found in the literature is presented in Table 4, and the socio-economic variables used for social and composite indices are presented in Table 5.

Table 4 Composite Coastal Vulnerability Indices found in the literature
Full size table
Table 5 Socio-Economic Variables used in the literature
Full size table

Questionnaire surveys have also been used to evaluate vulnerability and resilience, as in the case of the Port Resilience Index (Morris and Sempier 2016). Global indices such as the World Risk Index (World Risk Report 2019), the Index for Risk Management (INFORM; Messina et al. 2019), the Climate Risk Index (CRI; Eckstein et al. 2020) and the University of Notre Dame’s Global Adaptation Index (ND-GAIN; Chen et al. 2015) have also been frequently used in order to validate custom indices, as they are designed to improve the common data basis for risk management so that the majority of governments and policy makers and other stakeholders can work consistently (Poljansek et al. 2020).

The CVI (Gornitz 1991) has been the ground for many indices with a geophysical approach developed later in the literature, and the SoVI (Cutter et al. 2003) for indices with a socio-economic approach. Past researches included simpler indices and variables were one-dimensional. As years advanced, more and more papers combining physical and socio-economic approaches have been published. Despite having similarities in variable selection, it was not possible to find consistency between different approaches, especially for complex indices. Also, even though the majority of the variables are quantitative, the evaluation process highly depends on the personal opinion of the researchers. Variable ranking and categorization may differ, taking into consideration specific aspects of each study area. There are even different approaches for the same variable. For example, some researchers suggest that low values for the variable “tidal range” indicate high vulnerability (e.g., Tragaki et al. 2018; Zampazas et al. 2022), while others suggest the opposite (e.g., Pantusa et al. 2018). Furthermore, vulnerability assessment using indices seems to be a static approach while human–environment interactions are highly dynamic. The approaches of Zanetti et al. (2016), Toimil et al. (2017) and Gargiulo et al. (2020) tried to fill this gap with indices matching diverse socio-economic issues; however, there is still room for improvement in this direction. The MHCRI by Satta (2014) included diverse variables that are implemented in both urban and rural environments.

4 Summary and Conclusions

A total of 46 studies that generate a vulnerability index for climate change related hazards in coastal areas has been reviewed. There are studies that focus on physical variables and others that focus on both physical and social factors. An interesting fact is the diversity of the approaches that have been developed. Authors have used different scales of analysis, criteria for variable selection and natural phenomena as main hazards for their research. Especially among social indicators, it is possible that limitations in data availability have led to many different approaches worldwide. Furthermore, the lack of consistency in the selection of variables and indicators (especially socio-economic) is observed, whereas the selection of geophysical indicators is rather standardized. Outcomes of models based on geophysical approaches seem to be more consistent as well. This research demonstrates that a broad range of variables have been used in this context, considering different types of time and space scales for coastal process operation, categorizing into different intervals and ranking each variable result in a large spectrum of vulnerability levels. However, in terms of suitability for vulnerability classification, variable ranking is often controversial, as the evaluation process highly depends on the personal opinion of the researchers.

The objective of this review paper was to provide a valid description of the key dimensions of heterogeneity within research on climate change vulnerability. Through this review, researchers and policy-makers can determine the gaps in knowledge on climate change vulnerability of coastal areas, and the challenges that need to be addressed in future research.

The absence of a specific framework for the selection of indicators was identified, especially regarding social vulnerability. In most of the case studies, indicators have been selected taking into consideration data availability and researchers’ aspirations. Moreover, researchers do not acknowledge concerns regarding static understanding of the human environment in index-based vulnerability research. A holistic approach in the selection of variables would be able to overcome these concerns, in order to identify not only areas in danger, but also the level of vulnerability of people and communities at local scales, as damage on crucial infrastructure will affect directly their livelihood.