Citation: Babanawo, D.; Mattah,

P.A.D.; Agblorti, S.K.M.; Brempong,

E.K.; Mattah, M.M.; Aheto, D.W.

Local Indicator-Based Flood

Vulnerability Indices and Predictors

of Relocation in the Ketu South

Municipal Area of Ghana.

Sustainability 2022, 14, 5698. https://

doi.org/10.3390/su14095698

Academic Editor: César

Vázquez-González

Received: 3 March 2022

Accepted: 29 April 2022

Published: 9 May 2022

Publisher’s Note: MDPI stays neutral

with regard to jurisdictional claims in

published maps and institutional affil-

iations.

Copyright: © 2022 by the authors.

Licensee MDPI, Basel, Switzerland.

This article is an open access article

distributed under the terms and

conditions of the Creative Commons

Attribution (CC BY) license (https://

creativecommons.org/licenses/by/

4.0/).

sustainability

Article

Local Indicator-Based Flood Vulnerability Indices andPredictors of Relocation in the Ketu South Municipal Areaof GhanaDaystar Babanawo 1,2,*, Precious Agbeko D. Mattah 1,2 , Samuel K. M. Agblorti 1,3,4 ,Emmanuel K. Brempong 1,2 , Memuna Mawusi Mattah 5 and Denis Worlanyo Aheto 1,2

1 Centre for Coastal Management—Africa Centre of Excellence in Coastal Resilience (ACECoR),University of Cape Coast, Cape Coast PMB TF0494, Ghana; pmattah@ucc.edu.gh (P.A.D.M.);sagblorti@ucc.edu.gh (S.K.M.A.); brempong5@gmail.com (E.K.B.); daheto@ucc.edu.gh (D.W.A.)

2 Department of Fisheries and Aquatic Sciences (DFAS), University of Cape Coast,Cape Coast PMB TF0494, Ghana

3 Department of Population and Health, University of Cape Coast, Cape Coast PMB TF0494, Ghana4 Centre for Mixed Migration and Diaspora Studies (CeMMiDS), University of Cape Coast,

Cape Coast PMB TF0494, Ghana5 Department of Environment and Development Studies, Central University, Tema PMB TF253, Ghana;

mmmattah@st.ug.edu.gh* Correspondence: daystarbabanawo@gmail.com

Abstract: Certain communities along the coast of Ketu South Municipality in south-eastern Ghana,remain vulnerable to coastal flood events from storm surges, high tidal waves, lagoon overflow,and heavy rainfall. However, the local conditions that make these communities vulnerable arepoorly understood and knowledge on which communities are most vulnerable is lacking. This studyimproves the conceptual understanding of different dimensions of vulnerability that exist acrossthe communities and the various levels of vulnerability that each exposed community exhibits. Thestudy surveyed 354 household heads from selected flood-prone communities including Blekusu,Agavedzi, Salakope, Amutsinu, and Adina. The survey collected data on demographic, social,economic, physical, exposure, and adaptive capacity to flood hazards. The data was then used toconstruct composite vulnerability indices at community levels. Results from the study demonstratethat the communities have different levels of vulnerability as a result of differences in their exposure,sensitivity, and adaptive capacity levels. The willingness to relocate as an adaptation strategy wasdetermined by household flood duration, number of livelihoods, and sea defence preference. Theseresults are relevant to flood disaster management programs and the adoption of effective adaptationmeasures that take into account local knowledge. The findings imply that interventions aimed atreducing vulnerability should take into account household characteristics, as well as flood exposure,and adaptive capacity factors.

Keywords: climate change; coastal flood; vulnerability indices; adaptation options; Ketu Southmunicipal; Ghana

1. Introduction

Coastal flooding induced by sea-level rise, heavy rainfall, and a flash flood is oneof the increasing coastal hazards threatening coastal communities [1–3]. According toMcGranahan et al. [4], 56 million people reside in the low elevation coastal zones (LECZ)of Africa. The LECZ are areas of 0–10 m above sea level and these zones are often exposedto the threats of sea-level rise.

Global sea level has risen by 1–2.5 mm/y in the past one hundred years [5] andpredictions of future sea-level rise range from 20 cm to 86 cm by the year 2100 [6]. Fromobservations, Ghana’s sea level rose at a rate of 2.1 mm/year in the Takoradi port between

Sustainability 2022, 14, 5698. https://doi.org/10.3390/su14095698 https://www.mdpi.com/journal/sustainability

Sustainability 2022, 14, 5698 2 of 26

1925 and 1970, which conformed with the global rate [7]. Sea level rise may be responsiblefor the frequent storm surges and tidal wave flooding being experienced along the coastof the Gulf of Guinea [8]. The phenomenon is having detrimental effects on many coastalsettlements in the West African sub-region and predictions are that the situation may beworsened in the 21st century and beyond [9].

The coast of Ghana, particularly the eastern coastline, has been noted to be at highrisk and prone to the impacts of these threats [10]. Hence, communities along this coast-line, including those in the Ketu South Municipality, remain vulnerable to flood hazards.According to Aboagye et al. [11], vulnerability to hazards is influenced by various factors,which are grouped into physical, environmental, institutional, social, and economic factors,and its interpretations depend on disciplinary perspective and context. For example, socialscientists tend to define vulnerability as a set of social, economic, and demographic factorsthat combine to determine people’s ability to cope with stressors [12].

In the literature, vulnerability to hazards such as sea-level rise is a multidimensionalconcept, encompassing biophysical, socioeconomic, and political factors [6,13]. Knowledgeon all these factors provides policy makers with a holistic approach to instil climate re-silience in vulnerable areas. According to Wu et al. [14], assessing the physical vulnerabilityalone does not capture the pattern of differentiated impacts as well as social factors amongthe populations that are exposed to the hazards. Thus, to provide more efficient andeffective adaptation solutions, especially for marginalized and poor communities, compos-ite vulnerability information encompassing the biophysical, political, and socioeconomicaspects of the society is required [15–17]. Investigating vulnerability at the local scale iscrucial for understanding its characteristics, which are necessary for developing site-specificand appropriate adaptation measures to match the level of exposure and sensitivity of theparticular area under study [3,6,18]. Understanding the various factors of vulnerabilitywith respect to exposure, sensitivity, and adaptive capacity of population groups at thelocal level also drives prioritization and efficient allocation of scarce resources to mitigate,prepare, respond, and recover from disasters. In this regard, this study assesses compositevulnerability levels of selected coastal communities in the Ketu South Municipal area ofGhana, using an indicator-based approach. Secondly, it assessed the predictors of relocationas a possible adaptation strategy in the study area.

1.1. The Concept of Vulnerability

The quest for vulnerability assessment dates back to the 1970s when it became morerelevant for reducing disaster risks [19–21]. In the research community of natural hazards,vulnerability was first defined as “the degree of loss to a given element, or set of elements,at-risk” and was often quantified in different indices. But the narrative started to changein the 1980s and 1990s when there was a recognition on the importance of environmental,economic, social, and political factors influencing the vulnerability of social systems [22,23].The idea was to consider different perspectives, including the assessment of susceptibilityto harmful impacts and the ability to adapt or moderate the impacts [22], rather than justan exposure assessment, which communicates only the likelihood of risk to policy makers.Recently, IPCC has refined the definition of vulnerability (in terms of climate change) toinclude exposure, susceptibility (sensitivity), and coping or adaptive capacity [20,24]. Thesefactors can be measured as biophysical and socioeconomic variables.

Aside from the mentioned evolutionary concepts, vulnerability assessment can beconceptualized in terms of disciplines and that can be construed from three main per-spectives [25]. The three perspectives are: (1) the risk-hazard concept, which assesses asystem’s physical exposure to hazards [26–28]. This perspective is common to the disasterrisk management discipline. (2) The social constructivist concept dominates the fieldsof Political Economy and Human Geography and regards vulnerability as a function ofsocioeconomic and political factors [29]. (3) The third concept is the integrated assessmentof biophysical, socioeconomic, and institutional factors to assess the vulnerability of anysystem to climate change events, which is interdisciplinary [3,11,16,30].

Sustainability 2022, 14, 5698 3 of 26

The conception of vulnerability can also depend on the level or scale of analysis, be itthe global, regional, country, or local level [20,31]. Some of the factors that are consideredimportant for vulnerability assessment at the global scale often include indicators suchas population distribution, relative mortality rate, and relative GDP losses [20,32]. At theregional scale, indicators are selected based on the characteristic of exposure, socioeconomicstatus, and resilience [30]. Local-scale or community-level assessments of vulnerabilitydiffer largely depending upon the scope of the assessment and data availability [16,30,33].

Many vulnerability frameworks and conceptual models have been developed toprovide a context-specific understanding of vulnerability at local scales [16,33,34]. Anexample is the “Hazard-of-Place Model” proposed by [16] and expounded by [3,27,30] tomeasure composite vulnerability at local levels, which was adapted for this study.

1.2. Developing a Vulnerability Index (Composite Index)

The concept of developing a composite index was introduced in the 1990s to capturethe complexity and multidimensionality of development issues [35]. Since then, somestudies and international organizations such as the World Bank, United Nations, andEuropean Commission have developed composite indices [35]. Examples include theEnvironmental Performance Index (EPI), the Human Development Index (HDI), GenderEmpowerment Index (GEM), Livelihood Vulnerability Index [36], and Flood VulnerabilityIndex [30,37], among others.

Developing an index involves the conceptualization of the phenomenon and opera-tionalization of the concept with identified measurable indicators [38]. The operational-ization stage mainly involves (i) normalization, (ii) weighting, (iii) aggregation of theindicators scores into an index, and (vi) classification of the results into either quantile,equal interval, natural breaks, or standard deviation, and (iv) uncertainty and sensitivityanalysis [39].

When all the indicators are measured with the same unit (e.g., percent or ratios), datacan be aggregated without being scaled. However, in many instances, the indicators to beaggregated have various units and measurements such as nominal, ordinal, interval, andratio scales. In this circumstance, normalization is the technique used to standardize theindicators on a common scale. The choice of a preferred normalization approach should bemade with consideration when constructing composite indices, considering the compositeindex’s aims as well as the data attributes [35]. This is because varied normalizationprocedures provide different results and can have significant implications for compositeindex scores [35].

The most widely used normalization techniques in the literature are ranking, Z-score,min-max, and distance to target, as presented in Table 1 below.

Table 1. Normalization methods.

Method Equation Description References

Ranking Nias = Rank(Xias)Uses on ordinal variablesthat can be converted toquantitative variables.

[34,36]

Z scores Z =x−µ

σ

Transforms all indicatorsvalues to a single scale witha mean of 0 and a standarddeviation of 1.

[36,37]

Min–max Isc = S−SminSmax−Smin

Rescales indicator valuesbetween 0 (worst rank) and1 (best rank).

[3,36]

Distance to target Nias = XiasTarget Xias

Rescales values between 0and 1. It is the ratio of thevalue of the indicator to itsmaximum value.

[34,36]

Sustainability 2022, 14, 5698 4 of 26

The final score and ranking of the composite indices depend on the weighting of thenormalized values of the indicators. Weighting reflects the importance of each indicatorrelative to the overall. Weighting can be very important because it modifies the sub-indicesvalues before aggregation to a composite index is done. However, Sajeva et al. [40] foundthat different weighting techniques do not influence the ranking of composite indices. Thereis no agreed-upon mechanism for weighing individual indicators. The two main typesof weighting in the literature include equal weighting, used in the Human DevelopmentIndex, HDI [41], and flood vulnerability indices, [42] or unequal weighting [3,41,43].

The aggregation method of a composite index can be compensatory (linear aggrega-tion) or non-compensatory (geometric aggregation). Compensability of indicators, which isdefined as compensating for any indicator’s dimension with an appropriate surplus in an-other indicator’s dimension, is the most fundamental issue in aggregation. The method ofaggregation used has a significant impact on the index scores. Moreira et al. [39] found thatwhen geometric aggregation was opposed to linear aggregation (additive), the geometricaggregation (multiplication) approach was more sensitive, as it provided less compens-ability for the indices, resulting in lower scores. Many vulnerability indices have beenaggregated from the IPCC’s three components, Exposure (E), Sensitivity (S) and AdaptiveCapacity (C) using any of the two aggregation techniques expressed as:

Linear aggregation (additive function)—V = α1E + α2S − α3C [3].Geometric aggregation (multiplication function)—V = E∗S

C [30,44,45].

1.3. Adaptation Options

Adaptation has to do with the reduction of risk and vulnerability through takingadvantage of opportunities, building or improving the capacity of nations, regions, cities,the private sector, communities, individuals, and natural systems. It is to moderate theimpacts of climate change-related disasters on the sectors and mobilize the capacity toimplement decisions and actions [46].

Adaptation to climate change hazards in any specific system requires adequate infor-mation to select robust options that could be helpful. There is therefore the need to engagepeople with different knowledge, experiences, and backgrounds. In the case of coastalareas, there are so many adaptation strategies including protection, accommodation, andretreats [47].

The protection option involves reducing the risk of a hazard by decreasing the prob-ability at which it occurs, and this is done by building systems that protect societies orpopulations and their assets through hard structural (sea defences) or soft structural in-terventions (e.g., beach nourishment, dune restoration and creation, wetland restorationand creation, etc.). The accommodation option, on the other hand, has to do with changingconditions to improve adaptive strategies or a society’s ability to withstand or cope withthe harm that a hazard may cause. This is done through hazard emergency planning,early warning and evacuation systems, hazard insurance, modification of land use andagricultural practices, and modification of building styles, among others. In a physicalvulnerability study by Boateng [26], policies that could allow settling in exposed coastalareas were recommended as temporary adaptation measures in certain parts of Ghana.Retreat options can be in the form of relocating away from the hazard through propertyacquisition, buyouts, or relocation programs to reduce the risk of the event and limit itspotential effects [26].

The preference for any of these options is dependent on the local context, such asthe type of support (political and public) available, technical and financial resources,institutional capacity, as well as socioeconomic characteristics [48]. It is also influenced bythe complexity of the hazard risk in terms of the magnitude of impacts as a result of localcharacteristics like topography, hydrology, ecological systems, tourism, and the presence ofother hazards [48].

Sustainability 2022, 14, 5698 5 of 26

1.4. Relocation

Some areas such as regions with low-lying coastal plains, islands, and deltas are highlyvulnerable [49]. According to Bukvic et al. [50], due to the complexity of hazards thatthese areas are exposed to, particularly, areas with interconnected waterbodies (deltas), therobust adaptation option may either be the combination of all the options or consideringa retreat, which is relocation. Relocation is an effective strategy for coastal flood hazardmitigation [48], yet its implementation can depend on the integration of governance orinstitutional frameworks [51]. The willingness of communities and households to considerrelocation is another important factor to consider [39,42].

1.5. Conceptual Base of the Study

The concept underpinning this study is the “Hazard-of-Place Model” proposed andexplained by Yankson et al. [3], Boateng et al. [27], and Sajeva et al. [40]. The conceptdiscusses the integration of exposure, sensitivity, and adaptive capacity factors in deter-mining the extent of vulnerability of societies to natural phenomena. The concept remainsthe standard approach used in most climate change vulnerability studies [25]. Accordingto Fussel [25], the conceptualization of vulnerability in a specific setting tends to containfactors of vulnerability that are considered targets for policy interventions. The hazard-of-place model captures vulnerability as a combined effect of physical (exposure), social(sensitivity), as well as wealth and institutional concepts (Adaptive capacity). Figure 1 is aconceptual framework adapted from the Hazard-of-Place Model for this study; it illustratesthe three vulnerability factors of exposure, sensitivity, and adaptive capacity with theirrespective indicators.

The exposure component is a measure of the character, including, the magnitude,depth, duration, and frequency of flood and the population that is exposed to the impact offloods [3]. The sensitivity, on the other hand, is a measure of the social characteristics of acommunity, including gender, age, disability, marital status, and household size, amongothers. Thirdly, adaptive capacity measures wealth and institutional capacity, a collectionof governance measures (local and external measures), which influence the adaptabilityof societies and communities that are exposed to extreme events that are characteristic ofclimate change [28,43,45,52].

Sustainability 2022, 14, 5698 6 of 26Sustainability 2022, 14, x FOR PEER REVIEW 6 of 26

Figure 1. Conceptual framework of the study. Source: Adapted from Yankson et al. [3], Marshall et al. [34], and Messner & Meyer [53].

Figure 1. Conceptual framework of the study. Source: Adapted from Yankson et al. [3], Marshallet al. [34], and Messner & Meyer [53].

2. Materials and Methods2.1. Study Area

Ketu South Municipality (Figure 2) is located between 6◦3′ N and 6◦10′ N latitude and1◦6′ E and 1◦11′ E longitude. It shares borders on the east with the Republic of Togo, onthe west with Keta Municipality, on the north by Ketu North District, and on the south bythe Gulf of Guinea. With a total population of 102,905 (2020 population projection, VoltaRegion), the Municipality covers around 779 km2, accounting for 3.8% of the Volta Regionland area [54].

Sustainability 2022, 14, 5698 7 of 26

Sustainability 2022, 14, x FOR PEER REVIEW 7 of 26

2. Materials and Methods 2.1. Study Area

Ketu South Municipality (Figure 2) is located between 6°3’ N and 6°10’ N latitude and 1°6’ E and 1°11′ E longitude. It shares borders on the east with the Republic of Togo, on the west with Keta Municipality, on the north by Ketu North District, and on the south by the Gulf of Guinea. With a total population of 102,905 (2020 population projection, Volta Region), the Municipality covers around 779 km2, accounting for 3.8% of the Volta Region land area [54].

The coast of Ketu South Municipality is located on Ghana’s eastern coast and contains valuable resources such as wetlands (particularly mangroves and marsh lands), lagoons, living marine resources (fisheries), minerals (salt and sand), and groundwater, which would have a significant impact on the municipality’s economy when properly managed. However, the coastline has been classified as being particularly vulnerable to sea-level rise in terms of physical vulnerability; with one meter of sea-level rise, half of the shore can be inundated [10]. In addition, the Ketu South municipal coastal area is a deltaic area with eight major communities wedged between the Keta Lagoon, salt marshes, and the sea, making it prone to sea flooding, lagoon overflows, and heavy rainfalls. Furthermore, the area is already subjected to frequent storm surges and high tidal wave flooding, which has a negative influence on the people and their economic activities [26,55].

Figure 2. A map of the study area.

2.2. Data Collection A mixed-method and GIS research approaches were deployed for this study. In line

with the mixed-method approach, exploratory sequential research design was adopted for the study. The first phase of the research was qualitative, using Focus Group Discus-sions (FGDs) as a Participatory Rural Appraisal (PRA) to explore and gain insight and identify local indicators that contribute to the vulnerability of the communities to coastal

Figure 2. A map of the study area.

The coast of Ketu South Municipality is located on Ghana’s eastern coast and containsvaluable resources such as wetlands (particularly mangroves and marsh lands), lagoons,living marine resources (fisheries), minerals (salt and sand), and groundwater, whichwould have a significant impact on the municipality’s economy when properly managed.However, the coastline has been classified as being particularly vulnerable to sea-level risein terms of physical vulnerability; with one meter of sea-level rise, half of the shore can beinundated [10]. In addition, the Ketu South municipal coastal area is a deltaic area witheight major communities wedged between the Keta Lagoon, salt marshes, and the sea,making it prone to sea flooding, lagoon overflows, and heavy rainfalls. Furthermore, thearea is already subjected to frequent storm surges and high tidal wave flooding, which hasa negative influence on the people and their economic activities [26,55].

2.2. Data Collection

A mixed-method and GIS research approaches were deployed for this study. In linewith the mixed-method approach, exploratory sequential research design was adopted forthe study. The first phase of the research was qualitative, using Focus Group Discussions(FGDs) as a Participatory Rural Appraisal (PRA) to explore and gain insight and identifylocal indicators that contribute to the vulnerability of the communities to coastal flooding,and also to map recent flood extents in the various communities. PRA is an approachthat allows local communities to share, evaluate, and improve their knowledge on thephenomenon, as well as plan and act [47,48,56,57]. The study was carried out from Januaryto March 2021. In all, nine FGDs were conducted, with a maximum of two in each of thestudy communities. The groups involved flood victims and community members whosupported in rescuing those affected by flood disasters. Following this was a householdsurvey using a structured interview guide, which was developed and pre-tested to collect

Sustainability 2022, 14, 5698 8 of 26

data on the identified indicators contributing to vulnerability during the PRA. An iterativetechnique was used to develop the structured interview guide, such that informationreceived from the PRA was used to enhance the structured interview instrument.

Based on the results from the qualitative studies, communities including Blekusu,Agavedzi, Amutsinu, Salakope, and Adina were identified as being regularly exposed toflood events, hence they were selected for the quantitative studies. Population data onthe selected communities were collected from the municipal office of the Ghana Statisticaloffice. The population sizes for Adina, Amutsinu, Agavedzi, Blekusu, and Salakope were1637, 217, 798, 1698, and 200, respectively. The total population size of the five communitieswas 4550: the target population size. Using Krejcie et al.’s [58] table (Table 1) for sample sizecalculation, a sample size of 357 was arrived at and was distributed among the communitiesusing the probability proportional to size (PPS) sampling procedure [59,60].

The sample size for each of the community is as follows:

Blekusu =16984550

× 357 = 133

Agavedzi =798

4550× 357 = 62

Adina =16374550

× 357 = 128

Amutsinu =217

4550× 357 = 17

Salakope =200

4550× 357 = 15

The growing demand for research required the development of an efficient methodfor calculating the sample size required to be representative of a specific target population.The National Education Association used the formula:

SX2NP(1− P)

d2(N − 1) + X2P (1− P)

for the sample size determination in Krejcie et al.’s [58] table, which is widely used; where:

→ S = Required Sample Size,→ X = Z value (e.g., 1.96 for 95% confidence level),→ N = Population size,→ P = Population proportion (expressed as decimal) (assumed to be 0.5 (50%) since this

would provide the maximum sample size), and→ d = Degree of accuracy (5%), expressed as a proportion (0.5): the margin of error.

Using the systematic sampling technique, every third housing structure was selectedand, in each structure, the first household that was met by the research team was inter-viewed if the household head was available and willing to be interviewed. The structuredinterviews were therefore administered to the household heads. Before the interviews, thehousehold heads were briefed on the objectives of the study. They were also providedthe option of agreeing or refusing to be interviewed at any point in time of the interview.The household heads were further assured of confidentiality and anonymity of the studysuch that their submissions during the interviews would not be disclosed to any thirdparty. Questions asked during the structured interviews bordered on household demo-graphics, socioeconomic factors, flood characteristics, as well as adaptation measures in thestudy area.

Sustainability 2022, 14, 5698 9 of 26

2.3. Data Analysis2.3.1. Determination of Indices for Vulnerability

In the determination of vulnerability indices, this study used variables from the surveydata as indicators (Table 2). Similar to Yankson et al. [3] and Chakraborty et al. [61], thevariables were summarised using percentages and averages (Tables 3–5).

Following Yankson et al. [3], the standardization method (min-max) used to developvulnerability indices at community levels was adopted. The method transforms eachindicator into scores that range between 0 and 1. The procedure used by Yankson et al. [3]was followed to standardize each indicator as follows:

Isc =S− Smin

Smax − Smin

where:

Isc is the standardized index for each community (c),S is observed value for each community, andSmax and Smin are the observed maximum and minimum values respectively.

The mean index, for each of the factors, sensitivity, exposure, or adaptive capacity, wasdetermined using the following formula:

f ISC,mean =1n ∑ ISC

where f refers to either exposure (E), sensitivity (S), or adaptive capacity (A) index, and nis the total number of indicators for the factor.

Following [30,44,45] composite community vulnerability indices were finally deter-mined using the geometric aggregation method (multiplication), as follows:

V =E ∗ S

C

where:

→ V = Composite community vulnerability index,→ E = Exposure sub-index,→ S = Sensitivity sub-index, and→ C = Adaptive capacity sub-index.

Following Cutter et al. [16], Balica et al. [30], and Bathi & Das [42], the values of thevulnerability factors were aggregated without weighting. Weighting is done to solve theproblem of aggregating indicators with different dimensions (units) and magnitude. How-ever, this problem can be addressed with standardization of the individual indicators [62],and this has been done in the min-max standardization steps above.

The min-max method used to develop the indices transforms the indicator scoresbetween zero (0) and one (1), with 0 being the worst score and 1 being the best score.Following Yankson et al. [3], exposure and sensitivity were ranked as high (≥0.30), medium(0.18–0.29), and low (<0.18); and potential impact was ranked as high (≥0.7) and medium(0.5–0.69). According to Weis et al. [63], since potential impact (exposure + sensitivity) andthe composite vulnerability indices are aggregated from the main vulnerability factors, noabsolute conclusion can be drawn on their scores; the correct conclusions are relative state-ments for the communities. Hence, the composite vulnerability scores were ranked usingrelative statements after [3]; a score of 0.67–1 is highly vulnerable, 0.34–0.66 vulnerable,and 0–0.33 least vulnerable.

2.3.2. Determination of Predictors of Relocation

Negative log-log regression was applied in STATA 16 to examine the predictors ofrelocation decisions. This type of regression was applied because the outcome of the

Sustainability 2022, 14, 5698 10 of 26

response variable, relocation, is dichotomous, “No” or “Yes” and more than 50% of theresponses were “No”, which is not affirmative. A negative log-log regression model issuitable for a dichotomous response variable that has 55% or more of responses that arenot affirmative [64].

The independent variables used for the analyses were selected based on practicalsignificance and theoretical relevance [39,42,54]. The variables included flood durationin houses, number of livelihoods of household heads, and sea defence preference. Theanalysis also controlled for theoretically relevant compositional factors and contextualfactors [64,65]. The compositional factors included gender of household head, age ofhousehold head, household size, education, and monthly income of household head,while the contextual factors included communities such as Adina, Agavedzi, and Blekusu.Amutsinu and Salakope were considered as part of Adina since their sample sizes weretoo small for the analysis. The compositional and contextual factors were controlled inthe model, taking into consideration that these factors might affect the responses on thepredictor variables [64,65].

In the analysis, a 95% confidence interval was employed, and the level of statisticalsignificance was set at 0.05. The results were reported as odd ratios (OR). An OR of 1 meansthat the predictor does not affect the odds of relocation, OR > 1 means that the predictor isassociated with higher odds of a relocation decision, and OR < 1 means that the predictoris associated with lower odds of a relocation decision.

2.4. Ethical Considerations

Ethical clearance for the study was provided by the University of Cape Coast’s In-stitutional Review Board (UCCIRB). The ethical approval identification number is UC-CIRB/CANS/2021/15.

Table 2. Indicators used for developing the vulnerability indices.

Components Indicators Description

ExposureFlood frequencyAverage flood frequency in communityAverage flood frequency in households

Flood frequency measures the return period of floodevents in the communities.

Average flood duration Flood duration is the number of days the flood takesto recede in the communities.

Flood depthPercentage of households with flood depth atwaist height

Flood depth determines the height of flood from theground level to the water surface, the higher the depththe greater the degree of damage [66].

Flood magnitudePercentage of households that reported floodmagnitude as more

Flood magnitude was measured based on theperception of the respondents, and this is classified asless, medium, or more.

Flood impactsPercentage of households who have experiencedhouse property damagePercentage of households who have experiencedlivelihood impactsPercentage of households who had experiencedimpacts on water sourcePercentage of households who experienced impacts onfood sourcePercentage of households who experiencedhealth impacts

Flood impacts were measured at household levels inthe dimension of house property damage, livelihoodloss, water and food source impacts, and healthimpacts, as identified in the qualitative studies.

Sustainability 2022, 14, 5698 11 of 26

Table 2. Cont.

Components Indicators Description

Sensitivity Percentage of female-headed households

Studies have demonstrated that the femalepopulations have lower chances of gaining access toresources and information during and after adisaster, and this had had a negative impact on theirphysical and mental health. It is also widelydocumented that women have higher mortality andpoverty rates in disaster occurrences, and studieshave found that the female population andfemale-headed households have positive andsignificant statistical effects or relation to the severityof social vulnerability of a locality [16,29,55,67,68].

Average household size In high-density areas, there is less probability ofevacuation and a higher risk of death [66].

Number of children <5 years

The young, that is, children under five (5) years ofage, are most often unable to respond to disasterswithout assistance [56,57], and they are moresusceptible to significant physical and psychologicalimpacts [69–71]. Children who have inadequatesupport from family are usually disadvantagedwhen they have to respond to a disaster [72].

Number of elderlies >65 years

Elderly groups, even if they are not poor orphysically weak, are more likely to lack the physicaland economic resources necessary to respond to adisaster efficiently and effectively [72]. Besides thephysical challenges that evacuation and relocationbring, elderly people become depressed aboutleaving their own homes to stay in a group quarteror a rescue place.

Number of disables

The mentally or physically disabled have a lessercapability to respond to a disaster effectively, as theyrequire additional assistance to prepare for andrecover from disasters. Disaster managers need totarget areas with more disabled people, for earlyevacuation and also for disaster preparationmeasures [59,60,73].

Number of women

Considering factors such as domesticresponsibilities, women are, in a way, less able torespond appropriately to a crisis. Their domesticresponsibilities and status may restrict their abilityto respond quickly in terms of evacuation to rescuegrounds or seeking relief on time in the advent of adisaster [71,72].

Adaptive Capacity Percentage of households that receive earlywarning information on flood

The availability of early warning systems in acommunity provides an opportunity for disasterpreparedness, early warnings, and emergencyinformation, which in extent substantially reduce thevulnerability of the exposed population to a hazard,including saving lives and minimizing potentialinjuries and property loss [74].

Percentage of households that were aware ofrecent flood before flooding Flood awareness reduces flood risk [3].

Sustainability 2022, 14, 5698 12 of 26

Table 2. Cont.

Components Indicators Description

Percentage of households that have communitysupport to address flood risk

Societal groups involved in flood disasters are critical tomanage the effects of the disaster in the absence of officialstate agencies. In comparison to communities withoutevidence of civil society flood mitigation/adaptation, acommunity having evidence of civil society floodmitigation/adaptation was judged as better equipped [3].

Percentage of households that receivegovernment intervention

Flood victims’ access to any type of support might be acrucial adaptation technique. Households that reportedreceiving support from their local government, friends, andfamily networks were considered to be more adaptable thanthose who did not [3].

RecoveryPercentage of households that are satisfied withgovernment interventionPercentage of households that recovers to theprevious efficient state after a floodPercentage of households that have floodinsurancePercentage of households with multiple sourcesof income

The need to recover after a disaster necessitates long-termrehabilitation efforts that are influenced by the underlyingsocioeconomic processes and structural limitations. Therecovery of an individual or a society is influenced bycapital re-accumulation processes and external interventions[75]. In [75], income, government interventions, andnumber of businesses (livelihoods), among others, areidentified as the determinants for recovery after a disaster.

Percentage of households withinformation assets

Ownership of household assets, information, andcommunication gadgets (e.g., televisions, radio and mobilephones) makes a household better off in receiving andprocessing information on imminent hazards, and also inpreparation for and evacuating from a hazard [17].Televisions, radios, and mobile phones are important inmediating socioeconomic vulnerability. They act as amedium of information access, and their usage does notnecessarily require a high literacy level or formaleducation [76].

Percentage of households withtransportation assets

Lack of transportation assets is an important aspect thatincreases the vulnerability of an individual or a social group.Empirically, we find evidence that the lack of transportationassets resulted in unnecessary suffering for persons living inpoverty or near poverty in the central region of NewOrleans, who did not have privately owned vehicles orother means of transportation to leave their homes to safergrounds [23].

Percentage of literate household heads

Households with limited education are usually lessproficient in reading and are therefore less likely to accessemergency information if they are not assisted. They arealso more subjected to income fluctuations due to unsecuredemployment and are less able to manage risk [77].

Average income

Low-income people are economically weak and are affectedby disasters disproportionately. It is identified that they areunable to afford assets or generate income that can helpthem prepare for a disaster or recover after a disaster [16,59].

Sustainability 2022, 14, 5698 13 of 26

Table 3. Indicators for exposure index.

Indicators Max Min Adina Amutsinu Salakope Agavedzi Blekusu

Average flood frequency in community(per annum) 5.5 3.65 3.65 5.18 5.5 4.23 4.69

Average flood frequency in households(per annum) 5.75 3.067 3.13 4.29 5.75 3.66 3.067

Average flood duration (days) 22.19 6.06 22.19 11.412 6.06 12.53 18.64Percentage of households with flood

depth at waist height 29.03 88.24 71.43 88.24 87.50 29.03 81.20

Percentage of households that reportedflood magnitude as more 81.95 67.74 74.60 76.47 75.00 67.74 81.95

Percentage of households who haveexperienced house property damage 88.71 64.66 69.84 76.47 75.00 88.71 64.66

Percentage of households who haveexperienced livelihood impacts 87.50 67.67 85.71 76.47 87.50 72.58 67.67

Percentage of households who hadexperienced impact on water source 42.86 8.06 42.86 23.53 25.00 8.06 11.28

Percentage of households whoexperienced impact on food source 94.12 81.25 92.06 94.12 81.25 88.71 74.44

Percentage of households whoexperienced health impacts 68.75 34.92 34.92 52.94 68.75 58.06 44.36

Table 4. Indicators for sensitivity index.

Indicators Max Min Adina Amutsinu Salakope Agavedzi Blekusu

Percentage of female headedhouseholds 82.35 56.35 56.35 82.35 81.25 59.68 61.65

Average household size 12.18 9.63 10.59 12.18 14.63 9.63 9.66Number of children <5 years 241 37 214 37 46 95 241

Number of elderlies >65 years 171 13 111 13 16 52 171Number of disables 92 3 43 3 5 39 92Number of women 569 79 459 79 97 240 569

Table 5. Indicators for adaptive capacity index.

Indicators Min Max Adina Amutsinu Salakope Agavedzi Blekusu

Percentage of households that receive earlywarning information on flood 18.80 53.97 53.97 23.53 31.25 27.42 18.80

Percentage of households that were awareof recent flood prior to flooding 0.00 22.58 11.11 17.65 0.00 22.58 12.78

Percentage of households that havecommunity support to address flood risk 0.00 6.35 6.35 0.00 0.00 0.00 6.02

Percentage of households that receivegovernment intervention 0.00 38.35 2.38 0.00 0.00 11.29 38.35

Percentage of households that are satisfiedwith government intervention 0.00 18.80 2.38 0.00 0.00 9.68 18.80

Percentage of households that recovers tothe previous efficient state after a flood 0.00 53.97 53.97 41.18 75.00 37.10 49.62

Percentage of households that have floodinsurance 0 3.76 2.38 0 0 0 3.759398

Percentage of households with multiplesources of income 0 73.68 71.43 52.94 56.25 67.74 73.68

Percentage of households with informationassets 0 92.06 92.06 88.24 100 87.10 87.22

Percentage of households withtransportation assets 35.29 51.88 40.48 35.29 37.5 51.61 51.88

Sustainability 2022, 14, 5698 14 of 26

3. Results3.1. Sociodemographic Characteristics of Survey Respondents (Household Heads)

Figure 3 illustrates the demographics of the respondents (household heads), revealingthat 38.7% of the 354 respondents were male household heads and 61.3 percent were femalehousehold heads. The most common age group, accounting for 37.6% of the total, wasover 60 years old, followed by 35% between 40 and 59 years old, and 27.4% between20 and 39 years old. This indicates that a large proportion of household heads exposedto flood events were of an economically inactive age group. In addition, the majority ofrespondents (71.8%) were married, with just a small percentage (3.1%) divorced, 4.2% single,and 20.9% widowed. The level of educational attainment among respondents was mainlybasic education, with 59% of the respondents attaining that, while 29.4% had no formaleducation, 7.9% had secondary education, and just a few (3.7%) had higher education.

Sustainability 2022, 14, x FOR PEER REVIEW 14 of 26

3. Results 3.1. Sociodemographic Characteristics of Survey Respondents (Household Heads)

Figure 3 illustrates the demographics of the respondents (household heads), reveal-ing that 38.7% of the 354 respondents were male household heads and 61.3 percent were female household heads. The most common age group, accounting for 37.6% of the total, was over 60 years old, followed by 35% between 40 and 59 years old, and 27.4% between 20 and 39 years old. This indicates that a large proportion of household heads exposed to flood events were of an economically inactive age group. In addition, the majority of re-spondents (71.8%) were married, with just a small percentage (3.1%) divorced, 4.2% sin-gle, and 20.9% widowed. The level of educational attainment among respondents was mainly basic education, with 59% of the respondents attaining that, while 29.4% had no formal education, 7.9% had secondary education, and just a few (3.7%) had higher educa-tion.

Figure 3. Sociodemographic characteristics of survey respondents (household heads).

3.2. Vulnerability Indices for the Exposed Communities Table 6 presents flood exposure and sensitivity indices for the communities as well

as the potential impact, adaptive capacity, and composite vulnerability indices. Adina, Amutsinu, Salakope, and Agavedzi scored high, 0.48, 0.51, 0.54, and 0.44, respectively, while Blekusu had a medium score, 0.23 of flood exposure. From the table, Adina, Sala-kope, Amutsinu, and Blekusu have high scores for sensitivity, 0.51, 0.51, 0.33, and 0.7, respectively, while Agavedzi recorded medium sensitivity: 0.23. In Table 2, results re-vealed that Blekusu had a higher number of children (241), the aged (171), and women (569). Adina followed suit with 214 children, 111 aged, and 459 women. These two com-munities, therefore, recorded higher sensitivity indices, as stated above.

Table 6. indices of sensitivity, exposure, potential impact, adaptive capacity, and composite vulner-ability.

Community Sensitivity Exposure Potential Impact

Adaptive Capacity

Community Vulnerabilit

y Adina 0.51 0.48 0.99 0.68 0.36

Amutsinu 0.33 0.51 0.85 0.37 0.45

Figure 3. Sociodemographic characteristics of survey respondents (household heads).

3.2. Vulnerability Indices for the Exposed Communities

Table 6 presents flood exposure and sensitivity indices for the communities as wellas the potential impact, adaptive capacity, and composite vulnerability indices. Adina,Amutsinu, Salakope, and Agavedzi scored high, 0.48, 0.51, 0.54, and 0.44, respectively,while Blekusu had a medium score, 0.23 of flood exposure. From the table, Adina, Salakope,Amutsinu, and Blekusu have high scores for sensitivity, 0.51, 0.51, 0.33, and 0.7, respectively,while Agavedzi recorded medium sensitivity: 0.23. In Table 2, results revealed that Blekusuhad a higher number of children (241), the aged (171), and women (569). Adina followedsuit with 214 children, 111 aged, and 459 women. These two communities, therefore,recorded higher sensitivity indices, as stated above.

Table 6. Indices of sensitivity, exposure, potential impact, adaptive capacity, and composite vulnerability.

Community Sensitivity Exposure PotentialImpact

AdaptiveCapacity

CommunityVulnerability

Adina 0.51 0.48 0.99 0.68 0.36Amutsinu 0.33 0.51 0.85 0.37 0.45Salakope 0.51 0.54 1.05 0.43 0.64Agavedzi 0.23 0.44 0.68 0.63 0.16Blekusu 0.70 0.23 0.93 0.95 0.1

Exposure and Sensitivity is ranked as high (≥0.30), medium (0.18–0.29) and low (<0.18), potential impact isranked as high (≥0.7) and medium (0.5–0.69), and composite vulnerability classifications, a score of 0.67; 1 ishighly vulnerable, 0.34–0.66 vulnerable, and 0–0.33 least vulnerable, after Yankson et al. [3] and Schmidt-Thomé& Greiving [78].

Sustainability 2022, 14, 5698 15 of 26

Table 6 also presents the potential impact indices of the communities. Potentialimpact is a combination of the flood exposure index and sensitivity to understand thelevel of impact a community would face from a flood hazard. From the table, Salakopehas the highest potential impact of 1.05, with the rest, Adina, Amutsinu, Agavedzi, andBlekusu, recording high potential impacts at 0.99, 0.85, 0.63, and 0.93, respectively. The highpotential impact recorded for the communities is an indication of a high level of exposureand sensitivity to flood hazards in the communities.

The adaptive capacity index ranges from 0.37 to 0.95, as presented in the table. Blekusu,Adina, and Agavedzi had higher adaptive capacity, 0.95, 0.68, and 0.65, respectively, whilethe other communities, Amutsinu and Salakope, had the lowest scores, 0.37 and 0.43,respectively. A higher percentage of responses on having institutional (government andcommunity institutional) supports, access to early warning information, formal education,information assets, transportation assets, recovery to a previous efficient state, and higheraverage income levels for Blekusu, Agavedzi, and Adina (Table 3) are the main contributoryfactors to their high score on adaptive capacity, as compared to the other communities.

The table also presents the composite vulnerability scores for the communities. Salakopeand Amutsinu scored 0.64 and 0.45, respectively, indicating vulnerability scores higherthan Adina, Agavedzi, and Blekusu, which had lesser vulnerability scores of 0.36, 0.16, and0.17, respectively.

To enhance the visualization and appreciation of the vulnerability levels across the com-munities, the vulnerability indices were presented on maps, as illustrated in Figures 4–8.Figures 4–8 show the exposure, sensitivity, potential impact, adaptive capacity, and com-posite vulnerability levels, respectively, for the various communities.

Sustainability 2022, 14, x FOR PEER REVIEW 16 of 26

Figure 4. Flood exposure levels for the communities.

Figure 5. Flood sensitivity levels for the communities.

Figure 4. Flood exposure levels for the communities.

Sustainability 2022, 14, 5698 16 of 26

Sustainability 2022, 14, x FOR PEER REVIEW 16 of 26

Figure 4. Flood exposure levels for the communities.

Figure 5. Flood sensitivity levels for the communities. Figure 5. Flood sensitivity levels for the communities.

Sustainability 2022, 14, x FOR PEER REVIEW 17 of 26

Figure 6. Potential Impact for the communities.

Figure 7. Composite vulnerability for the communities.

Figure 6. Potential Impact for the communities.

Sustainability 2022, 14, 5698 17 of 26

Sustainability 2022, 14, x FOR PEER REVIEW 17 of 26

Figure 6. Potential Impact for the communities.

Figure 7. Composite vulnerability for the communities.

Figure 7. Composite vulnerability for the communities.

Sustainability 2022, 14, x FOR PEER REVIEW 18 of 26

Figure 8. Adaptive capacity levels for the communities.

3.3. Determinants of Relocation Adaptation Option Table 7 presents the predictors for relocation, with their probability values, confi-

dence intervals (CI), odd ratios (OR), and probability parameters. The values of the prob-ability parameters presented in the table include: Akaike information criterion (AIC) = 0.653236, Bayesian information criterion (BIC) = −1784.577, Log pseudolikelihood = 99.622777052, Residual df = 338, (1/df) Deviance = 0.5894838, and (1/df) Pearson = 0.7939348.

Table 7. Negative log-log regression model, illustrating the relationship between explanatory and dependent variables (relocation adaptation option).

Variable Predictors + Compositional and Contextual Factors

OR SE P Value Confidence Interval Flood duration 1.009646 0.0033071 0.003 1.003185 1.016148

Livelihoods (ref: 1 livelihood) 2 livelihoods 0.5704749 0.1086301 0.003 0.3927809 0.8285577

Sea defence (ref: No) Yes 0.1879353 0.0349485 0.000 0.1305325 0.2705814

Age of household head (ref: 20–30) 40–59 1.067389 0.2157025 0.32 0.718304 1.586124 60+ 0.8981923 0.1996318 −0.48 0.5810074 1.388535

Gender of household head (ref: male)

Female 1.377811 0.2834103 1.56 0.9206629 2.061951 House size (ref: 1–4)

5–7 1.282795 0.3962804 0.81 0.7001713 2.350228 Above 8 1.104163 0.307661 0.36 0.6395235 1.906381

Education (ref: No education)

Figure 8. Adaptive capacity levels for the communities.

Sustainability 2022, 14, 5698 18 of 26

3.3. Determinants of Relocation Adaptation Option

Table 7 presents the predictors for relocation, with their probability values, confidenceintervals (CI), odd ratios (OR), and probability parameters. The values of the probabilityparameters presented in the table include: Akaike information criterion (AIC) = 0.653236,Bayesian information criterion (BIC) = −1784.577, Log pseudolikelihood = 99.622777052,Residual df = 338, (1/df) Deviance = 0.5894838, and (1/df) Pearson = 0.7939348.

Table 7. Negative log-log regression model, illustrating the relationship between explanatory anddependent variables (relocation adaptation option).

VariablePredictors + Compositional and Contextual Factors

OR SE p Value Confidence Interval

Flood duration 1.009646 0.0033071 0.003 1.003185 1.016148Livelihoods (ref: 1 livelihood)

2 livelihoods 0.5704749 0.1086301 0.003 0.3927809 0.8285577Sea defence (ref: No)

Yes 0.1879353 0.0349485 0.000 0.1305325 0.2705814Age of household head (ref: 20–30)

40–59 1.067389 0.2157025 0.32 0.718304 1.58612460+ 0.8981923 0.1996318 −0.48 0.5810074 1.388535

Gender of household head (ref: male)Female 1.377811 0.2834103 1.56 0.9206629 2.061951

House size (ref: 1–4)5–7 1.282795 0.3962804 0.81 0.7001713 2.350228

Above 8 1.104163 0.307661 0.36 0.6395235 1.906381Education (ref: No education)

Basic school 1.119896 0.2231549 0.57 0.7578159 1.654975Secondary school and above 1.544611 0.453259 0.138 0.8690359 2.745366Monthly Income (ref: <100)

100–400 1.024614 0.199909 0.12 0.6990134 1.501881500–900 1.691449 0.5515699 1.61 0.892664 3.205014

1000 and above 0.9576796 0.3090339 −0.13 0.5087984 1.802581Community (ref: Adina)

Agavedzi 1.089387 0.1560924 0.32 0.6476522 1.832408Blekusu 0.8819346 0.1560924 −0.71 0.6234237 1.24764

Probabilities ParametersAIC 0.653236 Residual df 338BIC −1784.577 (1/df) Deviaance 0.5894838

Log pseudolikelihood −99.622777052 (1/df) Pearson 0.7939348

In the table, the probability values for flood duration, livelihood, and sea defencepreference were significant at p = 0.003, p = 0.003, and p = 0.00 respectively, hence thesevariables were the predictors for relocation in the communities.

The table demonstrates that flood duration, a continuous variable (OR = 1.009646,p < 0.0001), in households is more likely to influence relocation decision. Livelihood(OR = 0.5704749, p < 0.0001), a categorical variable, where two (2) livelihoods are beingcompared with one (1) livelihood, a reference variable, indicates that respondents withtwo (2) livelihoods were 43% less likely to relocate as compared to respondents with one(1) livelihood. The third predictor, sea defence preference (OR = 0.18, p < 0.0001), is a binaryvariable (Yes or No) where “No” responses were treated as a reference variable indicatingthat respondents who preferred sea defence for protection were 82% less likely to make arelocation decision, as compared to respondents who did not prefer sea defence structures.

In the table, the compositional factors included gender of household head, age ofhousehold head, household size, education, and monthly income of household head, aswell as the contextual factors including communities such as Adina, Agavedzi, and Blekusu,which were controlled for in the model; it had probability values that were not significant(p > 0.0001).

Sustainability 2022, 14, 5698 19 of 26

4. Discussion4.1. Community Vulnerability Levels

Recognizing which spatial scales are more vulnerable to flooding, and where thisvulnerability may be reduced more easily, might help decision-makers prioritize floodprotection measures in local and regional areas [30]. In this study, community vulnera-bility indices were developed at community levels to identify more and less vulnerablecommunities, based on the IPCC vulnerability factors.

Although these particular communities have been previously identified as vulnera-ble [26], the specific type of actions to implement remained unclear. Using the IPCC vul-nerability factors, including exposure, sensitivity, potential impacts, and adaptive capacity,to develop vulnerability indices at community levels resulted in differential vulnerabilityscores across the studied communities. The variation in the vulnerability levels could beattributed to the inequality of socioeconomic characteristics of the studied communities andthe differences in their flood exposure levels. According to Wongbusarakum & Loper [79],no one threshold defines whether a community is vulnerable to climate change; as a result,social indicators can assist in determining where limited resources should be invested.

This study found that vulnerability levels in the dimension of sensitivity, exposure,and adaptive capacity varied across the studied communities. This finding is directly in linewith the findings of Yankson et al. [3], which demonstrated that flood-prone communitiesin the Greater Accra Metropolitan Area of Ghana exhibited different levels of vulnera-bility with respect to their exposure, sensitivity, and adaptive capacity. Similar findingswere also reported in a previous study of flood-prone rural municipalities of Bosnia andHerzegovina [80].

It was found in this study that, in communities such as Adina, Salakope, Amutsinu,and Blekusu, sensitivity to coastal floods was high, but that Agavedzi had medium sensi-tivity. A higher number of people with disability, female-headed households, dependentage groups, women, and larger household sizes were the contributory factors to the highsensitivity recorded in these four communities (Table 4). It is worth discussing that thecommunities with high sensitivity are demographically more vulnerable than Agavedzi,which recorded medium sensitivity. Demographically, vulnerable groups are individualswho are more vulnerable than others in a locality due to their specific demographic or so-cioeconomic traits [79]. For example, it is reported in Owusu (2016) that, in many countries,women’s societal roles restrict their ability to adapt to climate change, and their obligationsfor childcare, water collection, and cooking fuel collection often increase their sensitivity toclimate change.

In the case of exposure, Agavedzi, Adina, Salakope, and Amutsinu had high scores,with Blekusu recording a medium score. Higher frequency of flood, with higher responseson high flood depth, magnitude, and impacts (such as household property damage, liveli-hood impact, food source impact, water source impact, and health impacts) are the contrib-utory factors to the high scores for exposure in the communities (Table 1). In a previousstudy [81], similar local indicators were found as the contributory factors for the highexposure of flood-prone areas in West Bengal, India.

According to Scheuer & Haase [82], flood vulnerability is determined by the quantityand value of elements at risk, their susceptibility, and their level of exposure to the hazard.This implies that communities found with higher exposure scores have higher populationsand elements exposed to higher flood magnitude. These communities also suffer morenegative impacts from the flood events than the community recording medium exposure.Empirically, this finding can be explained by the qualitative study, which explains thatBlekusu is protected with sea defence structures, minimizing its exposure to flood impacts,while the other communities are left unprotected.

When the two vulnerability components (sensitivity and exposure) were combined,all the communities had a high potential impact, with Salakope recording the highest.Contrary to the findings of Sitaula [80], which illustrated that sensitivity indices were themajor determinant of the high scores of potential impact in Rural Municipalities of Bosnia

Sustainability 2022, 14, 5698 20 of 26

and Herzegovina, this study demonstrated that the determinant of high potential scoresalternate between sensitivity and exposure, depending on the community. For example,Salakope had a sensitivity score of 0.51 and an exposure score of 0.54, which indicatesthat exposure contributed slightly more to the high potential impact than sensitivity. Also,Blekusu had a sensitivity score of 0.70 and an exposure score of 0.23, indicating thatsensitivity is the major determinant of the high potential impact.

Despite the high scores of potential impacts recorded for all the communities, Adina,Agavedzi, and Blekusu had their overall vulnerability positions remediated with highadaptive capacity, unlike Salakope and Amutsinu, which had medium adaptive capacity.The differences in their potential impact and adaptive capacity resulted in varied compositevulnerability scores across the communities. This finding is consistent with Yanksonet al. [3], where all the communities studied exhibited different levels of vulnerability inthe Greater Accra Metropolis, Ghana.

Contributing to the literature, these findings illustrate how complex vulnerability is,and that there are various levels of exposure, sensitivity, and adaptive capacity that needto be explored in various locations to understand the overall vulnerability levels. Thisunderstanding is crucial to designing effective adaptation strategies that are robust andsuitable for affected communities.

4.2. Predictors of Relocation

The value of AIC = 0.653236 in the results demonstrates that the model’s goodness offit is less than 2, indicating there is substantial evidence to support the fact that the modelis almost as good as the best model. According to Fabozzi et al. [83], “if the AIC is lessthan 2, this indicates there is substantial evidence to support the candidate model (i.e., thecandidate model is almost as good as the best model); between 4 and 7, this indicates thatthe candidate model has considerably less support; greater than 10, this indicates that thereis essentially no support for the candidate model”.

Evacuation and relocation are the most effective measures for vulnerable communitiesto avoid disasters and can ensure the safety of life and properties; thus, it is necessaryto identify the factors that influence residents’ willingness to evacuate and relocate fromhazard zones [84]. The last objective of this study was to identify the predictors of relocationin flood-prone communities in Ketu South Municipal, Ghana.

The findings indicate that the residents’ relocation decisions were not simple. The pre-dictors that influence their decision interplay between physical and economic factors, whilecompositional and contextual factors (control variables) were not significant predictors.The result from modelling the determinants of relocation as an adaptation option revealedthat flood duration in households, number of livelihoods, and sea defence preference arethe significant predictors of relocation in the study area; their probability values weresignificant (p < 0.0001).

In line with the findings of Buchori et al. [85], the residents were willing to relocatedue to flood duration in the households. The odd ratio for flood duration recorded was1.009646, which implies that the longer the flood duration in a household, the higher thelikelihood for the respondent to make a relocation decision. This also agrees with thefindings of Xu et al. [86], which demonstrated that every one-unit increase in severity ofearthquake disaster in China corresponds to increase in the odds of willingness to evacuate.

In the case of livelihood, it was found that the respondents with two livelihoods were43% less likely to relocate as compared to respondents with one livelihood. This suggeststhat livelihood can be considered an essential asset that influences relocation adaptationmeasures in hazard zones. This is consistent with the findings of Addo et al. [28], whichdemonstrate that voluntary and permanent relocation were overlooked by most floodvictims due to fear of losing income-generating ventures that serve as sources of livelihoodin the Sekondi-Takoradi Metropolis in Ghana. In addition, it can be argued that respondentswith multiple livelihoods possess wealth that can help them accommodate flood disasters,as compared to respondents with one source of livelihood. Wealth improves one’s ability

Sustainability 2022, 14, 5698 21 of 26

to plan for and endure losses in the event of an emergency, and vulnerability is thoughtto be exacerbated by a lack of wealth [17]. The wealth of a household is characterized byincome sources, quality of the housing structure, and possession of household assets [17].

It was also found that sea defence preference plays a strong role in predicting willing-ness to relocate. The odd ratio recorded for this variable, 0.18, implies that respondentsthat prefer sea defence for protection are 81% less likely to make relocation decisions, ascompared to respondents that do not prefer sea defence.

The compositional and the contextual factors that were controlled in the model hadprobability values that were not significant. This implies that the likelihood of a respondentto agree to a relocation decision in the study area is not based on gender, age, householdsize, education, monthly income, or the community in which the respondent resides. Thisfinding is consistent with findings in other studies [42,54], where contextual factors suchas communities and social dimensions did not influence relocation decisions. Similarly,Seebauer & Winkler [87] found that a relocation decision is made exclusively withinhouseholds, regardless of their neighbours’ actions or influence. There has also beena similar outcome on the compositional factors (household characteristics), especiallygender and its association with willingness to relocate. Xu et al. [86] found that genderand other household characteristics that were used as control variables in their studywere not significantly related to willingness to relocate from a hazard zone in SichuanProvince, China.

However, other previous studies [72,74,88,89] have demonstrated contradictory out-comes to these findings. In these studies, compositional factors such as average income, andcontextual factors such as place dependence (community bond), were significant influentialfactors for relocation aside from hazard severity and types of livelihoods or occupations.

5. Conclusions

Community vulnerability indices were constructed based on the IPCC vulnerabilityfactors: exposure, sensitivity, potential impacts, and adaptive capacity. The communitieshad different scores, with some recording high and some medium for all the aspects ofvulnerability. The study found that Adina, Salakope, Amutsinu, and Blekusu had highsensitivity scores, while Agavedzi had a medium score. A higher number of female-headed households, elderlies above 65 years, children below five years, women, and largerhousehold sizes are the contributing factors to the high scores for the four communities(Table 4).

With the exposure component of vulnerability, Agavedzi, Adina, Salakope, andAmutsinu had high scores, with Blekusu recording a medium score. A higher frequencyof flood exposure, with higher responses on high flood depth, magnitude, and impactsin the dimension of household property damage, livelihood, food and water sources, andhealth impacts are the contributing factors to high scores for exposure in the communitiesrecording high exposure indices (Table 3).

The combination of sensitivity and exposure scores resulted in high potential impactscores across the communities, with Salakope recording the highest. However, with adap-tive capacity, Salakope and Amutsinu recorded medium scores, while Adina, Agavedzi,and Blekusu recorded high scores, with Blekusu recording the highest. Despite the highscores of potential impacts recorded for the communities, Adina, Agavedzi, and Blekusuhad their overall vulnerability positions remediated with high adaptive capacity, unlikeSalakope and Amutsinu, which had medium adaptive capacity. The findings on the overallcommunity vulnerability scores are as follows: Salakope had the highest composite vul-nerability score 0.62, followed by Amutsinu, 0.50, with the rest of the communities Adina,Agavedzi, and Blekusu recording the lowest composite vulnerability scores: 0.31, 0.04, and−0.01, respectively.

Lastly, the study examined the predictors of relocation decisions at the household level,and it was found that flood duration, number of livelihoods, and sea defence preferencewere the main predictors, while compositional factors (control variables), such as gender of

Sustainability 2022, 14, 5698 22 of 26

household head, age of household head, household size, education, and monthly income ofhousehold head, were not significant predictors of relocation decisions in the communities.Contextual factors (control variables), such as the communities, were also not significant inthe result. Households with a longer flood duration are more likely to agree to relocationdecisions than households that experience a shorter duration of a flood. Also, householdswith multiple sources of livelihood, as well as households that prefer sea defence structures,are less likely to relocate. However, the likelihood of a respondent to agree to a relocationdecision is not based on gender, age, household size, education, monthly income, orthe community in which the respondent resides, since these factors were not significantpredictors of relocation.

These findings suggest that communities that were exposed to coastal flooding wouldsuffer impacts disproportionately, based on the varied vulnerability levels they exhibit.Hence, it is recommended that stakeholders, such as National Disaster Management Orga-nization (NADMO) and concerned Non-Governmental Organizations (NGOs), strengthenthe adaptive capacity of the communities by providing early warning systems, increasingthe scope of interventions to the vulnerable communities that receive fewer interventions,and increasing flood risk awareness about sea-level rise effects on coastal communities.

To address the sensitivity of the communities to coastal flooding, flood disaster riskreduction programs organized by governmental institutions, NADMO, and NGOs shouldfocus on locating and giving priority to households with age-dependent groups, femalehousehold heads, and low levels of education, disability, and low wealth levels. The studyfound that these factors influence the sensitivity of households and communities to coastalflooding.

Nonetheless, if relocation policies are being implemented, NADMO and other stake-holders should first target households that experience longer flood duration and haveonly one livelihood, since they are more willing to relocate than the others that experienceshorter flood duration, have multiple livelihoods, and prefer sea defence structures forprotection.

It is also recommended that the severity of post-event impacts of coastal floods onthe communities and the economic value of facilities and elements that are exposed toflood hazards are assessed. This will help in adopting robust adaptation options for thecommunities.

Author Contributions: Conceptualization, D.B.; methodology, D.B. and P.A.D.M.; formal analysis,D.B., E.K.B. and P.A.D.M.; validation, P.A.D.M. and S.K.M.A.; writing—original draft preparation,D.B.; writing—review and editing, P.A.D.M., S.K.M.A., M.M.M. and D.W.A.; supervision, P.A.D.M.and S.K.M.A. All authors have read and agreed to the published version of the manuscript.

Funding: This study is part of an MPhil project at the University of Cape Coast, Ghana, funded bythe World Bank Africa Centre of Excellence in Coastal Resilience (ACECoR) Project (World Bank ACEGrant Number 6389-G).

Institutional Review Board Statement: Ethical clearance for the study was provided by the Uni-versity of Cape Coast’s Institutional Review Board (UCCIRB). The ethical approval identificationnumber is UCCIRB/CANS/2021/15.

Informed Consent Statement: Informed consent was acquired from all subjects who participated inthe study, following the UCCIRB’s guidelines.

Data Availability Statement: The policies of UCCIRB policies do not support the sharing of thesurvey data publicly.

Acknowledgments: The authors would like to express their gratitude to the World Bank for financingthis study and also to the field assistants; Breeze Babanawo, James Macho Akligo, Michael Shitor,Dan Etse Agbayiza, and Gad Atsu Agbayiza, who supported with the data collection for the study.

Conflicts of Interest: The authors declare no conflict of interest.

Sustainability 2022, 14, 5698 23 of 26

References1. Owusu, M. Gender Vulnerability to Climate Change and Livelihood Security in Urban Slum Communities in Accra, Ghana. Ph.D.

Thesis, The University of Adelaide, Adelaide, Australia, 2017.2. Osman, A.; Nyarko, B.K.; Mariwah, S. Vulnerability and risk levels of communities within Ankobra estuary of Ghana. Int. J.

Disaster Risk Reduct. 2016, 19, 133–144. [CrossRef]3. Yankson, P.W.K.; Owusu, A.B.; Owusu, G.; Boakye-Danquah, J.; Tetteh, J.D. Assessment of coastal communities’ vulnerability

to floods using indicator-based approach: A case study of Greater Accra Metropolitan Area, Ghana. Nat. Hazards 2017, 89,661–689. [CrossRef]

4. Mcgranahan, G.; Balk, D.; Mcgranahan, G.; Bartlett, S. The rising tide: Assessing the risks of climate change and humansettlements in low elevation coastal Zones. Environ. Urban. 2007, 19, 17–37. [CrossRef]

5. Sterr, H. Assessment of vulnerability and adaptation to sea-level rise for the coastal zone of Germany. J. Coast. Res. 2008, 24,380–393. [CrossRef]

6. IPCC. IPCC Fourth Assessment Report: Climate Change 2007 (AR4)—Synthesis Report; IPCC: Geneva, Szwitzerland, 2007; Volume2099, p. 104.

7. Woodworth, P.L.; White, N.J.; Jevrejeva, S.; Holgate, S.J.; Church, J.A. Evidence for the accelerations of sea level on multi-decade.Int. J. Climatol. J. R. Meteorol. Soc. 2009, 789, 777–789. [CrossRef]

8. Mensah, C.; Kabo-bah, A.T.; Mortey, E. Assessing The Effects of Climate Change on Sea Level Rise Along the Gulf Of Guinea. J.Energy Nat. Resour. Manag. 2017, 4, 15–22. [CrossRef]

9. Church, J.A.; White, N.J. A 20th century acceleration in global sea-level rise. Geophys. Res. Lett. 2006, 33, 94–97. [CrossRef]10. Boateng, I.; Wiafe, G.; Jayson-Quashigah, P.N. Mapping vulnerability and risk of Ghana’s coastline to sea level rise. Mar. Geod.

2017, 40, 23–39. [CrossRef]11. Aboagye, D.; Attakora-Amaniampong, E.; Owusu-Sekyere, E. Place-based assessment of intersection of biophysical and social

vulnerability to flooding in Accra, Ghana. Int. J. Appl. Geospat. Res. 2020, 11, 55–68. [CrossRef]12. Aksha, S.K.; Juran, L.; Resler, L.M.; Zhang, Y. An Analysis of Social Vulnerability to Natural Hazards in Nepal Using a Modified

Social Vulnerability Index. Int. J. Disaster Risk Sci. 2019, 10, 103–116. [CrossRef]13. Grinsted, A.; Christensen, J.H. The transient sensitivity of sea level rise. Ocean Sci. Discuss. 2020, 17, 1–5. [CrossRef]14. Wu, S.; Yarnal, B.; Fisher, A. Vulnerability of coastal communities to sea-level rise: A case study of Cape May County, New Jersey,

USA. Clim. Res. 2002, 22, 255–270. [CrossRef]15. Sendai Framework. Sendai framework for disaster risk reduction 2015–2030. Aust. J. Emerg. Manag. 2015, 30, 9–10.16. Cutter, S.L.; Boruff, B.J.; Shirley, W.L. Social vulnerability to environmental hazards. Soc. Sci. Q. 2003, 84, 242–261. [CrossRef]17. Felsenstein, D.; Lichter, M. Social and economic vulnerability of coastal communities to sea-level rise and extreme flooding. Nat.

Hazards 2014, 71, 463–491. [CrossRef]18. Munji, C.A.; Bele, M.Y.; Nkwatoh, A.F.; Idinoba, M.E.; Somorin, O.A.; Sonwa, D.J. Vulnerability to coastal flooding and response

strategies: The case of settlements in Cameroon mangrove forests. Environ. Dev. 2013, 5, 54–72. [CrossRef]19. Birkmann, J. Risk and vulnerability indicators at different scales: Applicability, usefulness and policy implications. Environ.

Hazards 2007, 7, 20–31. [CrossRef]20. Birkmann, J.; Wisner, B. Measuring the Unmeasurable: The Challenge of Vulnerability; UNU-EHS: Bonn, Germany, 2006; Volume 5,

Available online: http://www.ihdp.unu.edu/file/get/3962.pdf (accessed on 5 January 2022).21. Wisner, B. Vulnerability as Concept, Model, Metric, and Tool. In Oxford Research Encyclopedia of Natural Hazard Science; Oxford

University Press: Oxford, UK, 2016. [CrossRef]22. Cutter, S.L. Vulnerability to environmental hazards. Prog. Hum. Geogr. 1996, 20, 529–539. [CrossRef]23. Kelly, P.M.; Adger, W.N. Theory and practice in assessing vulnerability to climate change and Facilitating adaptation. Clim. Chang.

2000, 47, 325–352. [CrossRef]24. Schneider, S.H.; Semenov, S.; Patwardhan, A.; Burton, I.; Magadza, C.H.D.; Oppenheimer, M.; Yamin, F. Assessing key vulnerabili-

ties and the risk from climate change. In Climate Change 2007: Impacts, Adaptation and Vulnerability; Contribution of Working GroupII to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change; Parry, M.L., Canziani, O.F., Palutikof, J.P.,van der Linden, P.J., Hanson, C.E., Eds.; Cambridge University Press: Cambridge, UK, 2007; pp. 779–810.

25. Fuessel, H.M. Vulnerability in Climate Change Research: A Comprehensive Conceptual Framework. 2005. Available online:https://escholarship.org/content/qt8993z6nm/qt8993z6nm.pdf (accessed on 5 January 2022).

26. Boateng, I. An assessment of the physical impacts of sea-level rise and coastal adaptation: A case study of the eastern coast ofGhana. Clim. Chang. 2012, 114, 273–293. [CrossRef]

27. Balaganesh, G.; Malhotra, R.; Sendhil, R.; Sirohi, S.; Maiti, S.; Ponnusamy, K.; Sharma, A.K. Development of compositevulnerability index and district level mapping of climate change induced drought in Tamil Nadu, India. Ecol. Indic. 2020, 113,106197. [CrossRef]

28. Addo, K.A.; Nicholls, R.J.; Codjoe, S.N.A.; Abu, M. A biophysical and socioeconomic review of the Volta delta, Ghana. J. Coast.Res. 2018, 34, 1216–1226. [CrossRef]

29. Noy, I.; Yonson, R. Economic vulnerability and resilience to natural hazards: A survey of concepts and measurements. Sustainabil-ity 2018, 10, 2850. [CrossRef]

Sustainability 2022, 14, 5698 24 of 26

30. Balica, S.F.; Wright, N.G.; Van der Meulen, F. A flood vulnerability index for coastal cities and its use in assessing climate changeimpacts. Nat. Hazards 2012, 64, 73–105. [CrossRef]

31. Dolan, A.H.; Walker, I.J. Understanding vulnerability of coastal communities to climate change related risks. J. Coast. Res. 2006,1316–1323.

32. Brouwer, R.; Akter, S.; Brander, L.; Haque, E. Socioeconomic vulnerability and adaptation to environmental risk: A case study ofclimate change and flooding in Bangladesh. Risk Anal. Int. J. 2007, 27, 313–326. [CrossRef] [PubMed]

33. Balica, S.F.; Douben, N.; Wright, N.G. Flood vulnerability indices at varying spatial scales. Water Sci. Technol. 2009, 60,2571–2580. [CrossRef]

34. Marshall, N.A.; Marshall, P.A.; Tamelander, J.; Obura, D.; Malleret-King, D.; Cinner, J.E. A Framework for Social Adaptation toClimate Change: Sustaining Tropical Coastal Communitites [Sic] and Industries; Iucn: Grand, Switzerland, 2010.

35. Talukder, B.; Hipel, K.W.; van Loon, G.W. Developing composite indicators for agricultural sustainability assessment: Effect ofnormalization and aggregation techniques. Resources 2017, 6, 66. [CrossRef]

36. Hahn, M.B.; Riederer, A.M.; Foster, S.O. The Livelihood Vulnerability Index: A pragmatic approach to assessing risks from climatevariability and change—A case study in Mozambique. Glob. Environ. Chang. 2009, 19, 74–88. [CrossRef]

37. Dinh, Q.; Balica, S.; Popescu, I.; Jonoski, A. Climate change impact on flood hazard, vulnerability and risk of the Long XuyenQuadrangle in the Mekong Delta. Int. J. River Basin Manag. 2012, 10, 103–120. [CrossRef]

38. Hinkel, J. “Indicators of vulnerability and adaptive capacity”: Towards a clarification of the science–policy interface. Glob. Environ.Chang. 2011, 21, 198–208. [CrossRef]

39. Moreira, L.L.; de Brito, M.M.; Kobiyama, M. Effects of different normalization, aggregation, and classification methods on theconstruction of flood vulnerability indexes. Water 2021, 13, 98. [CrossRef]

40. Sajeva, M.; Gatelli, D.; Tarantola, S.; Hollanders, H. Methodology Report on European Innovation Scoreboard; European Comission:Brussels, Belgium, 2005.

41. Qasim, A.W. United Nations Development Programme (UNDP). Human Development Report 2013. Pak. Dev. Rev. 2013, 52,95–96. [CrossRef]

42. Bathi, J.R.; Das, H.S. Vulnerability of coastal communities from storm surge and flood disasters. Int. J. Environ. Res. Public Health2016, 13, 239. [CrossRef]

43. Frigerio, I.; Ventura, S.; Strigaro, D.; Mattavelli, M.; De Amicis, M.; Mugnano, S.; Boffi, M. A GIS-based approach to identify thespatial variability of social vulnerability to seismic hazard in Italy. Appl. Geogr. 2016, 74, 12–22. [CrossRef]

44. Cutter, S.L.; Barnes, L.; Berry, M.; Burton, C.; Evans, E.; Tate, E.; Webb, J. A place-based model for understanding communityresilience to natural disasters. Glob. Environ. Chang. 2008, 18, 598–606. [CrossRef]

45. Choi, H. Assessment of aggregation frameworks for composite indicators in measuring flood vulnerability to climate change. Sci.Rep. 2019, 9, 1–14.

46. Tompkins, E.L.; Vincent, K.; Nicholls, R.J.; Suckall, N. Documenting the state of adaptation for the global stocktake of the ParisAgreement. Wiley Interdiscip. Rev. Clim. Chang. 2018, 9, 1–9. [CrossRef]

47. Nicholls, R.J. Adaptation Options for Coastal Areas and Infrastructure: An Analysis for 2030. 2007. Available online: https://www.unfccc.int/files/cooperation_and_support/financial_mechanism/application/pdf/nicholls.pdf (accessed on 5 January 2022).

48. Bukvic, A.; Zhu, H.; Lavoie, R.; Becker, A. The role of proximity to waterfront in residents’ relocation decision-making post-Hurricane Sandy. Ocean Coast. Manag. 2018, 154, 8–19. [CrossRef]

49. Williams, S.J. Sea-level rise implications for coastal regions. J. Coast. Res. 2013, 63, 184–196. [CrossRef]50. Bukvic, A.; Smith, A.; Zhang, A. Evaluating drivers of coastal relocation in Hurricane Sandy affected communities. Int. J. Disaster

Risk Reduct. 2015, 13, 215–228. [CrossRef]51. Abel, N.; Gorddard, R.; Harman, B.; Leitch, A.; Langridge, J.; Ryan, A.; Heyenga, S. Sea level rise, coastal development and

planned retreat: Analytical framework, governance principles and an Australian case study. Environ. Sci. Policy 2011, 14,279–288. [CrossRef]

52. Browne, E.; Siegel, P. No Regrets Approach to Human Vulnerability to Climate Change GEQ Heltberg Siegel Jorgensen 2009Related Papers. 2010. Available online: https://www.academia.edu/5641848/No_Regrets_Approach_to_Human_Vulnerability_to_Climate_Change_GEQ_Heltberg_Siegel_Jorgensen_2009 (accessed on 5 January 2022).

53. Messner, F.; Meyer, V. Flood Damage, Vulnerability and Risk Perception—Challenges for Flood Damage Research; Springer:Berlin/Heidelberg, Germany, 2005.

54. Allan, A.; Hissen, N.F.; Ghosh, A.; Samling, C.L.; Tagoe, C.A.; Nelson, W.; Mensah, A.; Salehin, M.; Mondal, S.; Spray, C.Stakeholder Mapping for Adaptation in Deltas. 2015, p. 89. Available online: https://generic.wordpress.soton.ac.uk/deccma/wp-content/uploads/sites/181/2017/07/D1.1.2_Stakeholder-mapping-Fast-Track.pdf (accessed on 5 January 2022).

55. Cazcarro, I.; Arto, I.; Hazra, S.; Bhattacharya, R.N.; Osei-Wusu Adjei, P.; Ofori-Danson, P.K.; Asenso, J.K.; Amponsah, S.K.;Khondker, B.; Raihan, S.; et al. Biophysical and socioeconomic state and links of deltaic areas vulnerable to climate change: Volta(Ghana), Mahanadi (India) and Ganges-Brahmaputra-Meghna (India and Bangladesh). Sustainability 2018, 10, 893. [CrossRef]

56. Chambers, R. The Origins and Practice of Rural Appraisal. World Dev. 1994, 22, 953–969. Available online: http://www.ircwash.org/sites/default/files/125-94OR-16929.pdf (accessed on 5 January 2022). [CrossRef]

Sustainability 2022, 14, 5698 25 of 26

57. Ahsan, N.; Warner, J. The Socioeconomic Vulnerability Index: A Pragmatic Approach for Assessing Climate Change Led Risks–ACase Study in the South-Western Coastal Bangladesh; Elsevier: Amsterdam, The Netherlands, 2014. Available online: https://www.sciencedirect.com/science/article/pii/S221242091300071X (accessed on 5 January 2022).

58. Krejcie, R.V.; Morgan, D.W. “Determining sample Size for Research Activities”, Educational and Psychological Measurement. Int.J. Employ. Stud. 1996, 18, 89–123.

59. Yansaneh, I.S. An analysis of cost issues for surveys in developing and transition countries. In Household Sample Surveys inDeveloping and Transition Countries; 2005; pp. 253–266. Available online: http://unstats.un.org/unsd/hhsurveys/pdf/Household_surveys.pdf (accessed on 5 January 2022).

60. Antwi, E.K.; Boakye-Danquah, J.; Barima Owusu, A.; Loh, S.K.; Mensah, R.; Boafo, Y.A.; Apronti, P.T. Community vulnerabilityassessment index for flood prone savannah agro-ecological zone: A case study of Wa West District, Ghana. Weather. Clim. Extrem.2015, 10, 56–69. [CrossRef]

61. Chakraborty, L.; Rus, H.; Henstra, D.; Thistlethwaite, J.; Scott, D. A place-based socioeconomic status index: Measuring socialvulnerability to flood hazards in the context of environmental justice. Int. J. Disaster Risk Reduct. 2020, 43, 101394. [CrossRef]

62. Lee, Y.J. Social vulnerability indicators as a sustainable planning tool. Environ. Impact Assess. Rev. 2014, 44, 31–42. [CrossRef]63. Weis, S.W.M.; Agostini, V.N.; Roth, L.M.; Gilmer, B.; Schill, S.R.; Knowles, J.E.; Blyther, R. Assessing vulnerability: An integrated

approach for mapping adaptive capacity, sensitivity, and exposure. Clim. Chang. 2016, 136, 615–629. [CrossRef]64. Armah, F.A.; Ekumah, B.; Yawson, D.O.; Odoi, J.O.; Afitiri, A.R.; Nyieku, F.E. Predictive Probabilities of Access to Clean

Cooking: Evidence from the Demographic and Health Surveys in 31 Countries in Sub-Saharan Africa. Environ. Justice 2019, 12,118–131. [CrossRef]

65. Bukvic, A.; Owen, G. Attitudes towards relocation following Hurricane Sandy: Should we stay or should we go? Disasters 2017,41, 101–123. [CrossRef] [PubMed]

66. Hadipour, V.; Vafaie, F.; Kerle, N. An indicator-based approach to assess social vulnerability of coastal areas to sea-level rise andflooding: A case study of Bandar Abbas city, Iran. Ocean. Coast. Manag. 2020, 188, 105077. [CrossRef]

67. Wood, N.J.; Burton, C.G.; Cutter, S.L. Community variations in social vulnerability to Cascadia-related tsunamis in the U.S. PacificNorthwest. Nat. Hazards 2010, 52, 369–389. [CrossRef]

68. Zhang, X.; Yi, L.; Zhao, D. Community-based disaster management: A review of progress in China. Nat. Hazards 2013, 65,2215–2239. [CrossRef]

69. Clark, G.E.; Moser, S. Assessing the Vulnerability of Coastal Communities to Extreme Storms: The Case of Revere, MA, USA.Mitig. Adapt. Strateg. Glob. Chang. 1998, 3, 59–82. [CrossRef]

70. Chen, W.; Cutter, S.L.; Emrich, C.T.; Shi, P. Measuring social vulnerability to natural hazards in the Yangtze River Delta region,China. Int. J. Disaster Risk Sci. 2013, 4, 169–181. [CrossRef]

71. Enarson, E. Through women’s eyes: A gendered research agenda for disaster social science. Disasters 1998, 22, 157–173. [CrossRef]72. Morrow, B.H. Identifying and mapping community vulnerability. Disasters 1999, 23, 1–18. [CrossRef]73. IPCC. Climate Change 2014 Synthesis Report_AR5 FINAL; IPCC: Geneva, Switzerland, 2014.74. Sufri, S.; Dwirahmadi, F.; Phung, D.; Rutherford, S. A systematic review of Community Engagement (CE) in Disaster Early

Warning Systems (EWSs). Prog. Disaster Sci. 2020, 5, 100058. [CrossRef]75. Jordan, E.; Javernick-Will, A. Indicators of Community Recovery: Content Analysis and Delphi Approach. Nat. Hazards Rev. 2013,

14, 21–28. [CrossRef]76. Noble, I.R.; Huq, S.; Anokhin, Y.A.; Carmin, J.; Goudou, D.; Lansigan, F.P.; Osman-Elasha, B.A.V. Adaptation needs and options.

In Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects; Contribution of Working Group IIto the Fifth Assessment Report of the Intergovernmental Panel on Climate Change; Cambridge University Press: Cambridge, UK,2014; pp. 833–868.

77. World Bank. World Development Report 2000/2001: Attacking Poverty; The World Bank: Washington, DC, USA, 2000. Availableonline: https://www.ipcc.ch/site/assets/uploads/2018/02/WGIIAR5-Chap14_FINAL.pdf (accessed on 10 January 2022).

78. Schmidt-Thomé, P.; Greiving, S. European Climate Vulnerabilities and Adaptation: A Spatial Planning Perspective; John Wiley & Sons:Hoboken, NJ, USA, 2013.

79. Wongbusarakum, S.; Loper, C. Indicators to assess community—Level social vulnerability to climate change. In NOAA Document,April; 2011; pp. 1–41. Available online: https://reefresilience.org/pdf/SocMon_Climate_change_guidlelines_FINAL_april_2011.pdf (accessed on 10 January 2022).

80. Žurovec, O.; Cadro, S.; Sitaula, B.K. Quantitative Assessment of Vulnerability to Climate Change in Rural Municipalities of Bosniaand Herzegovina. Sustainability 2017, 9, 1208. [CrossRef]

81. Das, M.; Chattopadhyay, A.; Basu, R. Spatial flood potential mapping with flood probability and exposure indicators of floodvulnerability: A case study from West Bengal, India. Int. J. Georesources Environ. 2017, 3, 85–93.

82. Scheuer, S.; Haase, D. Exploring multicriteria flood vulnerability by integrating economic, social and ecological dimensions offlood risk and coping capacity: From a starting point view towards an end point view of vulnerability. Nat. Hazards 2011, 58,731–751. [CrossRef]

83. Fabozzi, F.J.; Focardi, S.M.; Rachev, S.T.; Arshanapalli, B.G. The Basics of Financial Econometrics: Tools, Concepts, and AssetManagement Applications; John Wiley & Sons: Hoboken, NJ, USA, 2014.

Sustainability 2022, 14, 5698 26 of 26

84. Zhou, W.; Ma, Z.; Guo, S.; Deng, X.; Xu, D. Livelihood capital, evacuation and relocation willingness of residents in earthquake-stricken areas of rural China. Saf. Sci. 2021, 141, 105350. [CrossRef]

85. Buchori, I.; Pramitasari, A.; Pangi, P.; Sugiri, A.; Maryono, M. International Journal of Disaster Risk Reduction Factors distin-guishing the decision to migrate from the flooded and inundated community of Sayung, Demak: A suburban area of SemarangCity, Indonesia. Int. J. Disaster Risk Reduct. 2021, 52, 101946. [CrossRef]

86. Xu, D.; Peng, L.; Liu, S.; Su, C.; Wang, X. Influences of Sense of Place on Farming Households’ Relocation Willingness in AreasThreatened by Geological Disasters: Evidence from China. Int. J. Disaster Risk Sci. 2017, 8, 16–32. [CrossRef]

87. Seebauer, S.; Winkler, C. Should I stay or should I go? Factors in household decisions for or against relocation from a flood riskarea. Glob. Environ. Chang. 2020, 60, 102018. [CrossRef]

88. Vincent, K.; Cull, T. A Household Social Vulnerability Index (HSVI) for Evaluating Adaptation Projects in Developing Countries.Available online: https://www.semanticscholar.org/paper/A-Household-Social-Vulnerability-Index-%28HSVI%29-for-Vincent-Cull/1166265e6df595c90fcca96be504bd0b68ce1802 (accessed on 10 January 2022).

89. Correll, R.M.; Lam, N.S.N.; Mihunov, V.V.; Zou, L.; Cai, H. Economics over Risk: Flooding Is Not Considerations on a VulnerableCoast. Ann. Am. Assoc. Geogr. 2021, 111, 300–315. [CrossRef]