Livelihood Vulnerability of Riverine-Island Dwellers in the Face of Natural Disasters in Bangladesh

Abstract

Bangladesh is one of the most disaster-prone countries in the world. In particular, its riverine-island (char) dwellers face continuous riverbank erosion, frequent flooding, and other adverse effects of climate change that increase their vulnerability. This paper aims to assess the livelihood vulnerability of riverine communities by applying the Intergovernmental Panel on Climate Change (IPCC) vulnerability framework and the livelihood vulnerability index (LVI). Results indicate substantial variation in the vulnerability of char dwellers based on mainland proximity. The main drivers of livelihood vulnerability are char-dweller adaptation strategies and access to food and health services. The study further reveals that riverbank erosion, frequent flood inundation, and lack of employment and access to basic public services are the major social and natural drivers of livelihood vulnerability. Char-based policy focusing on short- and long-term strategy is required to reduce livelihood vulnerability and enhance char-dweller resilience.

1. Introduction

Natural disasters are regular phenomena in Bangladesh due to the country’s topography, geographical position, and changes in climate over time [1,2]. These disasters often have dire impact on the social and economic activities of delta communities, the most vulnerable regions of the country. The greatest devastation often hits communities living in riverine-island regions (large sandbars that emerge from riverbeds due to silt and alluvium deposition), particularly in the form of dynamic riverbank erosion and accretion [1,3]. These regions, called chars, are known for their multiple natural hazards and social vulnerabilities. Each year, for instance, char dwellers lose considerable amounts of useable land due to continuous riverbank erosion. In addition to natural hazards, char life is hampered by poor communication structures that limit char dwellers’ equal access to the social and economic benefits enjoyed by mainland dwellers [4]. Such challenging conditions are common throughout the many char regions of Bangladesh, which constitute about 5% of both the nation’s total area (7200 sq. km) and population (6.5 million people) [1,5,6]. Equally challenging is the fact that people living in these regions are often unable to migrate and find employment on the mainland.

Due to the displacement caused by challenging conditions, however, char dwellers do frequently migrate across the char regions. According to CARE-Bangladesh [7], about 25% of char families migrated at least three times over the last ten years. According to model-based estimates regarding climate change, Bangladesh is expected to face average annual and seasonal temperature increases of up to 4.7 °C by the end of the century [4]. Rainy seasons appear to be intensifying, while winter seasons are becoming drier. Challenges such as these are harmful not only to human life but also to the landscape that serves as the basis for successful agricultural activity, e.g., cropping patterns, pest infestations, crop yields, and water availability. Char dwellers regularly lose their agricultural assets, crops, livestock, and poultry, as well as the fiscal and human capital needed to maintain economic success and overall survival.

Vulnerability is an emerging concept across disciplines, useful in understanding and assessing the status of people’s condition in the face of natural hazards. The major characteristics of vulnerability are dynamic and influence people’s social and biophysical processes and systems [2]. Significant mobilization is necessary from the government, nongovernmental organizations, researchers, and farmers to develop successful adaptation strategies [8,9]. The people of developing countries are a vulnerable community due to excessive dependency on agriculture and having low income [10]. However, these burdens may fuel the exploration of potential adaptive capacities of resource-poor communities [11,12]. The extent of people’s susceptibility is increased due to the increasing vulnerability to natural hazards of almost all spheres of life, like the social, physical, human, financial, and natural dimensions [13,14]. Though the effect of natural hazards may be occasional, seasonal, or year-round [15,16], the extent of exposure is not the same for all communities.

A context-specific approach is required for exploring and assessing vulnerability to draft proper policy and strategy at all administrative levels and reduce adverse effects on livelihoods [4,17,18]. The interaction between people and their biophysical and social environment is readily used to assess the development-policy framework by using specific indicators [19], representing context-specific adaptation strategies [20], to compare and monitor the extent of vulnerability over time, space, and resource allocation [4,21]. The main challenges of vulnerability assessment are to develop robust and sound measures [22].

This study focuses on riverine-island (char) areas in Bangladeshi deltas. Bangladesh is one of the world’s largest delta areas, with about 230 rivers, including the Ganges–Brahmaputra–Meghna (GBM) River. The adverse impact of natural hazards is generally seen in coastal and riverine islands, which makes dwellers an extremely susceptible community due to their geographical isolation [23,24]. The hazards of isolation of char dwellers are intensified due to the fact that Bangladesh faces heavy rainfall and flooding approximately four months each year [25,26]. Generally, these catastrophic floods cause huge riverbank erosion through morphological dynamism in the GBM river system [23]. Frequent flood inundation causes drastic riverbank erosion, accounting for the loss of about 150,000 square kilometers over the last ten years [27]. According to Center for Environmental and Geographic Information Services (CEGIS) [28], about 20 of 64 districts are prone to riverbank erosion, resulting in losses of 8700 ha of land and the displacement of around 200,000 people each year [23,29,30,31]. Despite such hazards and vulnerability, riverine islanders often choose char areas due to increasing population pressure (156.6 million in 2014 [32]), and cumulative pressure on limited areas of land. Char dwellers are considered the most vulnerable people to natural hazards and the poorest of the poor [33,34,35]. Char areas have no road communication with the mainland or even within char villages, which increases their vulnerability. They can only use local boats (normally used for carrying goods and catching fish in riverine Bangladesh) during the rainy season for their transportation. Chars also lack electricity, health-service, market, and financial-institution facilities, which reduces their resilience capacity [1,23].

Gathering accurate information and in-depth research findings is necessary for the government, nongovernmental organizations, and international donor organizations to develop any program, policy, and strategy for the economic, social, and environmental development for marginalized char dwellers [1,4]. Policy intervention cannot actually occur without understanding the actual situation of char-dweller vulnerability [4,36,37,38]. The government of Bangladesh [24] considers the issue of char-dweller vulnerability an urgent matter to address. This study intends to fill this important gap via employing the IPCC vulnerability framework [2] by developing a livelihood vulnerability index (LVI) and a climate vulnerability index (CVI). It also aims to explore the extent of vulnerability of char dwellers in terms of livelihood and climate change at a rural household level in the chars of Bangladesh.

2. Materials and Methods

A sustainable-livelihood framework was followed to guide vulnerability assessment. Vulnerability context is a major determinant of a sustainable-livelihood framework that is mainly based on 5 livelihood assets, namely, human, social, natural, physical, and financial capital, and directly influences the institutional process, and livelihood strategies and outcomes [39,40].

The study chose 2 local administrative units (Upazila) of Gaibandha district, namely, Saghata and Fulchhari Upazila. These areas are around 287 km from the capital of Dhaka and the northern part of Bangladesh (Figure 1). These areas comprise natural-hazard-prone and geographically isolated riverine areas. The study areas are riverine islands (chars) in Jamuna River, which faces huge riverbank loss every year. Frequent flood inundation and riverbank erosion are regular phenomena in these areas (Figure 2). The study areas were purposively selected considering natural-hazard severity, based on obtained information from literature reviews, expert opinions, available reports, and newspapers. The respondents for this study were selected randomly from the study areas.

The study mainly focused on 2 aspects of island char areas; first, char dwellers who live nearer to the mainland, within a 5 km distance from Saghata Upazila; second, those living more than 5 km away from the mainland, in the Fulchhari Upazila headquarters. Both of these areas regularly face the same extent of natural hazards. Each of them, however, has a unique identity with regard to the communication network in Upazila and the district headquarters, education facilities, health facilities, other basic public services, and livelihood assets. The studied villages in Saghata Upazila were Haldia, Patilbari, Garamara, Digalkandi, Guabari, Kanaipara, Kalurpara, Kumarpara, and Hatbari. The distant-island villages in Fulchhari Upazila were Deluabari, Jamira, Bajefulchhari, Kholabari, Pipulia, Tenrakandi, Gabgasi, and Baghbari.

2.1. Data Collection

The study used a questionnaire survey and focus-group discussions (FGDs) for data collection regarding livelihood assets, sociodemographic profiles, vulnerability indicators, and adaptation strategies. The questionnaire pilot was tested on 25 respondents to determine its suitability for the study and avoid any exaggeration in the questionnaire. The sample size was determined by the following formula, developed by Yamane [41]. This formula has been popularly used by researchers (see References [42,43,44,45,46,47]) for determining household sample size for livelihood research:

$$\mathrm{n}=\frac{\mathrm{N}}{1+{\mathrm{Ne}}^2}$$

where n = sample size, N = population, and e = confidence interval.

The total population in the study area was 5666. Therefore, sample size was 374 for this study. Data were collected from the head of every household by face-to-face interviews using a semistructured questionnaire. The questionnaire survey and FGDs for this study were conducted from January to August 2017. Oral consent was taken from the household head prior to the interview. The interviews were done in the local Bengali language and lasted an average of 50 min. One FGD was done comprising 10–12 household heads in every village to record opinions regarding socioeconomic and climate-related variables that were used to validate the obtained data from the questionnaire survey. Differences in vulnerability status between household living nearby villages (in Saghata Upazila), and household living distant villages from the mainland (in Fulchhari Upazila) were determined by chi-square and t-tests.

2.2. Vulnerability Analysis

Vulnerability is a condition of an individual or community to stresses due to changes in socioeconomic and environmental conditions disrupting livelihoods [18]. Vulnerability assessment can identify susceptible people and the context of natural hazards through exploring socioeconomic processes and natural outcomes [1,33,48,49]. According to the IPCC [50], vulnerability is a function of 3 dimensions: exposure, sensitivity, and adaptive capacity.

Vulnerability = ƒ (exposure, sensitivity, adaptive capacity)

Generally, vulnerability is positively related to a system’s exposure and sensitivity, but negatively related to adaptive capacity [1,14]. The livelihood vulnerability of char dwellers was measured by an LVI [4,38] and CVI [51], focusing on major determinants under the appropriate IPCC framework. The IPCC framework uses 3 major factors (exposure, sensitivity, and adaptive capacity) to measure vulnerability. This study used a composite index-oriented LVI, which comprises the human, natural, physical, social, and financial household capital of a sustainable-livelihood framework (SLF) to provide better integration with sensitivity and adaptive capacity. This kind of methodology has been used by other scholars [1,4,52,53,54,55]. The main limitation of SLF is its inability to integrate the indicators of sensitivity and adaptive capacity. In this study, the LVI approach deals with a group of 13 major components consisting of major indicators and subindicators under 5 categories of livelihood capital (human, natural, physical, social, and financial capital). It comprises health, food, water, knowledge, livelihood strategies, land, natural resources, natural disasters, climatic variability, social networks, housing and production means, and agricultural and nonagricultural assets. This context-specific LVI approach can properly explore the real circumstances of livelihood vulnerability caused by natural disasters [38].

Context-specific LVI and CVI were used with a weighted balance and integrated approach. These context-specific LVI and CVI adopted additional components after Hahn et al. [38] and indicators based on study-area context through literature review, expert consultation, and local circumstances. A scale ranging from 0 (least vulnerable) to 1 (most vulnerable) was used to show the vulnerability status of inter- and intragroups of respondents. Though each major indicator comprises some subindicators, each of them equally contributed to the index. Equal weight was given to all components. Since a specific scale was used for the specific component, standardization was done by Equation (1).

$${\mathrm{Index}}_{\mathrm{sv}}=\frac{{\mathrm{S}}_{\mathrm{v}}-{\mathrm{S}}_{\min }}{{\mathrm{S}}_{\max }-{\mathrm{S}}_{\min }}$$

(1)

where S_v is an original subcomponent value of area v; S_min and S_max are the minimum and maximum value of the subcomponent, respectively. The standardized index was developed by using these minimum and maximum values. A scale ranging from 0 to 100 was used to explore the percentage of some components.

An average of each subcomponent was calculated after standardization by using Equation (2).

$${\mathrm{M}}_{\mathrm{vj}}=\frac{{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{Index}}_{\mathrm{svi}}}{\mathrm{n}}$$

(2)

where M_vj is the value of major component j for area v; Index_svi denotes the subcomponent value indexed by i of major component M_j; n represents the number of subcomponents in major component M_j.

The values of 13 major components under the 5 major capitals of livelihood were directly used in Equation (3) or aggregated to 5 livelihood assets (H (human capital), N (natural capital), S (social capital), P (physical capital), and F (financial capital)) before being used in Equation (3) to obtain the weighted average of LVI:

$${\mathrm{LVI}}_{\mathrm{v}}=\frac{{{\displaystyle \sum }}_{\mathrm{i}=1}^{10}{\mathrm{W}}_{\mathrm{Mj}}{\mathrm{M}}_{\mathrm{vj}}}{{{\displaystyle \sum }}_{\mathrm{i}=1}^{10}{\mathrm{w}}_{\mathrm{mj}}}$$

(3)

Equation (3) above can also be expressed as Equation (4):

$${\mathrm{LVI}}_{\mathrm{V}}=\frac{{\mathrm{W}}_{\mathrm{H}}{\mathrm{H}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{N}}{\mathrm{N}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{S}}{\mathrm{S}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{P}}{\mathrm{P}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{F}}{\mathrm{F}}_{\mathrm{V}}}{{\mathrm{W}}_{\mathrm{H}}+{\mathrm{W}}_{\mathrm{N}}+{\mathrm{W}}_{\mathrm{S}}+{\mathrm{W}}_{\mathrm{P}}+{\mathrm{W}}_{\mathrm{F}}}$$

(4)

where LVIv is the livelihood-vulnerability index of area v; W_Mj is the weightage of component j; W_H, W_N, W_S, W_P, W_F are the weight value of human capital, natural capital, social capital, physical capital, and financial capital, respectively. Equation (4) can be expressed as:

$${\mathrm{LVI}}_{\mathrm{V}}=\frac{{\mathrm{W}}_{\mathrm{H}}{\mathrm{H}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{F}}{\mathrm{F}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{W}}{\mathrm{W}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{KS}}{\mathrm{KS}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{LS}}{\mathrm{LS}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{L}}{\mathrm{L}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{CC}}{\mathrm{CC}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{NDC}}{\mathrm{ND}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{SN}}{\mathrm{SN}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{HPM}}{\mathrm{HPM}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{AA}}{\mathrm{AA}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{NAA}}{\mathrm{NAA}}_{\mathrm{V}}+{\mathrm{W}}_{\mathrm{FI}}{\mathrm{FI}}_{\mathrm{V}}}{{\mathrm{W}}_{\mathrm{H}}+{\mathrm{W}}_{\mathrm{F}}+{\mathrm{W}}_{\mathrm{W}}+{\mathrm{W}}_{\mathrm{KS}}+{\mathrm{W}}_{\mathrm{LS}}+{\mathrm{W}}_{\mathrm{L}}+{\mathrm{W}}_{\mathrm{CC}}+{\mathrm{W}}_{\mathrm{ND}}+{\mathrm{W}}_{\mathrm{SN}}+{\mathrm{W}}_{\mathrm{HPM}}+{\mathrm{W}}_{\mathrm{AA}}+{\mathrm{W}}_{\mathrm{NAA}}+{\mathrm{W}}_{\mathrm{FI}}}$$

(5)

where W_H, W_F, W_W, W_KS, W_LS, W_L, W_CC, W_NDC, W_SN, W_HPM, W_NAA, W_AA, and W_FI are the weight of health, food, water, knowledge and skill, livelihood strategies, land, climatic variability, natural disasters and climate variability, social networks, housing and production means, agricultural assets, nonagricultural assets, and finance and income, respectively. Similarly, H_V, F_V, W_V, K_SV, L_SV, L_V, CC_V, NDC_V, SN_V, HPM_V, NAA_V, AA_V, and FI_V are the number of indicators under health, food, water, knowledge and skill, livelihood strategies, land, climatic variability, natural disasters and climate variability, social networks, housing and production means, nonagricultural assets, agricultural assets, and finance and income, respectively.

The exposure (Exp) index includes land (L), natural resources (NR), and natural disasters and climate variability (NDC); it was measured as follows (Equation (6)):

$$\mathrm{IndexExp}=\frac{{\mathrm{W}}_{\exp 1}\mathrm{L}+{\mathrm{W}}_{\exp 2}\mathrm{CC}+{\mathrm{W}}_{\exp 3}\mathrm{ND}}{{\mathrm{W}}_{\exp 1}+{\mathrm{W}}_{\exp 2}+{\mathrm{W}}_{\exp 3}}$$

(6)

where W_exp1, W_exp2, and W_exp3 represent the weight for land (L), climatic variability (CC), and natural disasters (ND), respectively.

The index of sensitivity (Sen) was calculated from health (H), Food (F), and water (W), as follows (Equation (7)):

$$\mathrm{IndexSen}=\frac{{\mathrm{W}}_{\mathrm{sen}1}\mathrm{H}+{\mathrm{W}}_{\mathrm{sen}2}\mathrm{F}+{\mathrm{W}}_{\mathrm{sen}3}\mathrm{W}}{{\mathrm{W}}_{\mathrm{sen}1}+{\mathrm{W}}_{\mathrm{sen}2}+{\mathrm{W}}_{\mathrm{sen}3}}$$

(7)

where W_sen1, W_sen2, and W_sen3 denote weight for health (H), Food (F) and water (W) respectively.

The index for adaptive capacity (Adacap) includes knowledge and skills (KS), livelihood strategies (LS), social networks (SN), household and production means (HPM), agricultural assets (AA), nonagricultural assets (NAA), and finance and income (FI), and was measured as follows (Equation (8)):

$$\mathrm{IndexAdaCap}=\frac{{\mathrm{W}}_{\mathrm{ad}1}\mathrm{KS}+{\mathrm{W}}_{\mathrm{ad}2}\mathrm{LS}+{\mathrm{W}}_{\mathrm{ad}3}\mathrm{SN}+{\mathrm{W}}_{\mathrm{ad}4}\mathrm{HPM}+{\mathrm{W}}_{\mathrm{ad}5}\mathrm{AA}+{\mathrm{W}}_{\mathrm{ad}6}\mathrm{NAA}+{\mathrm{W}}_{\mathrm{ad}7}\mathrm{FI}}{{\mathrm{W}}_{\mathrm{ad}1}+{\mathrm{W}}_{\mathrm{ad}2}+{\mathrm{W}}_{\mathrm{ad}3}+{\mathrm{W}}_{\mathrm{ad}4}+{\mathrm{W}}_{\mathrm{ad}5}+{\mathrm{W}}_{\mathrm{ad}6}+{\mathrm{W}}_{\mathrm{ad}7}}$$

(8)

where, W_ad1, W_ad2, W_ad3, W_ad4, W_ad5, W_ad6, and W_ad7 represent the weight for knowledge and skill (KS), livelihood strategies (LS), social networks (SN), household and production means (HPM), agricultural assets (AA), nonagricultural assets (NAA). and finance and income (FI), respectively.

The weighted average of CVI was calculated from the value of exposure, adaptive capacity, and sensitivity by the following formula (Equation (9)).

$$\mathrm{CVI}=1-\left|\left\{\frac{{\mathrm{N}}_1\mathrm{Exp}-{\mathrm{N}}_2\mathrm{Ada}.\mathrm{cap}}{\left({\mathrm{N}}_1+{\mathrm{N}}_2\right)}\right\}\right|\ast \left\{\frac{1}{\mathrm{Sen}}\right\}$$

(9)

where n_i is the number of major components in the i-th vulnerability dimensions. The value of each dimension ranged to a maximum value of 1 from a minimum of 0.

2.3. IPCC Framework Approach

The IPCC approach allows to integrate all 11 components into 3 dimensions: exposure, sensitivity, and adaptive capacity. The 3 contributing factors are accumulated in Equation (10).

$$\mathrm{LVI}-\mathrm{IPCCa}=\left(\mathrm{Exp}-\mathrm{AdaCap}\right)\times \mathrm{Sen}$$

(10)

where LVI – IPCCa is the LVI for a community with a minimum value of −1 (least vulnerable) and maximum value 1 (most vulnerable).

According to some scholars [4,56,57,58,59,60], it is very difficult to choose robust and relevant indicators to properly represent local communities. However, this limitation is addressed through an extensive literature review, direct observations, and expert opinions for obtaining representative and comprehensive results (Appendix A). Indicator-based studies are the best tools to simplify the telling of a complex story. However, indicator choices and weighting are always subjective arguments [1,4,23]. Scholars argued that nonweighted variables would not change the message conveyed through an index in comparison with weighted variables [4,49]. Most vulnerability indices are nonweighted averages of indicators and a weighted average of components [1,4,37,38,51]. Thus, in line with the existing literature, this study also applied equal weighting for all indicators.

3. Results and Discussion

The findings of LVI, CVI, and livelihood vulnerability are interpreted in this section. Figure 3 outlines the influencing factors of vulnerability. It also shows LVI and CVI values, highlighting the major and subcomponents that vary from indicator to indicator, and between Saghata Upazila (within 5 km from the mainland) and Fulchhari Upazila (more than 5 km away from the mainland).

3.1. Livelihood Vulnerability Index

The LVI value of char dwellers in Fulchhari Upazila (0.428) was higher than that of char dwellers in Saghata Upazila (0.417). These values indicate that char dwellers of the study area are vulnerable in terms of their livelihood assets. The char dwellers of the more-distant area were more deprived in terms of basic public services, with less access to education, health, and finances (Appendix B).

Sociodemographic characteristics between the two groups were similar but varied in some subindicators like knowledge and skill, livelihood strategies, health, and water. It was found that female-headed households were more vulnerable than male-headed households in both char areas. The values of knowledge and skill, livelihood strategies, and health of char dwellers of Saghata Upazila, meanwhile, was higher than Fulchhari Upazila.

The index values of land, natural resources, natural disasters, and climate variability of Fulchhari Upazila were slightly higher than those of Saghata Upazila char dwellers due to longtime settlement in their char areas. Similarly, the index value of social networks of Saghata Upazila char dwellers was higher than that of Fulchhari Upazila dwellers. On the other hand, the index values of housing and production means, agricultural assets, nonagricultural assets, of Fulchhari Upazila char dwellers was lower than Saghata Upazila char dwellers. Similarly, financial income index value was also higher in Saghata Upazila than in Fulchhari Upazila.

3.2. Climate Vulnerability Index

CVI values for Saghata Upazila and Fulchhari Upazila char dwellers was high, indicating that char dwellers of both near and distant areas were vulnerable to climatic variability and natural disasters. There was almost no significant difference between them (

Table 1. Major component dimension of char-dweller livelihood and climate vulnerability.

Major Dimensions	Saghata Upazila	Fulchhari Upazila
Exposure (land, climatic variability, and natural disasters)	0.498	0.562
Sensitivity (health, food, and water)	0.520	0.532
Adaptive capacity (knowledge and skill, livelihood strategies, social networks, housing and production means, agricultural assets, nonagricultural assets, and finance and income)	0.314	0.300
Climate vulnerability Index	0.838	0.958
LVI-IPCC	0.353	0.428

Source: field survey.

), but values were higher than riverbank and mainland dwellers [1].

The values of the major LVI dimensions are shown in Table 1. Significant difference exists between the values of major indicators of vulnerability among char-dweller groups. The value of exposure, sensitivity, and adaptive capacity of char dwellers of Saghata Upazila was less than Fulchhari Upazila (Table 1). The values indicate that Fulchhari Upazila char dwellers are more exposed and sensitive to natural hazards than Saghata Upazila char dwellers. Similarly, the adaptive capacity of Fulchhari Upazila char dwellers was less than that of Saghata Upazila dwellers. LVI-IPCC estimation findings indicate that Fulchhari Upazila char dwellers are more vulnerable, which is similar to previous findings [54,61,62].

3.3. Livelihood Vulnerability

The livelihood status of char dwellers was found to be highly vulnerable across the study areas. Findings show that both groups of char dwellers are vulnerable, but those living nearest to the mainland are less vulnerable than more distant dwellers. This is likely due to facilities being provided by public agencies and nongovernmental organizations, better communication and social networks, education facilities, and easy migration during extreme disasters [52,63]. Due to reduced access to education, knowledge, and skills, Fulchhari Upazila char dwellers are more at risk than Saghata Upazila dwellers. The number of educational institutions in Fulchhari Upazila is less than in Saghata Upazila, which also influences knowledge and skill level. Only primary schools are available in some villages, which causes school dropouts at the secondary level. Livelihood strategy is almost diversified in Saghata Upazila, but less diversified in Fulchhari Upazila.

Riverbank erosion is a common phenomenon in char areas, and Both study areas face it regularly. However, the extent of riverbank erosion in Saghata Upazila is greater than Fulchhari Upazila. Findings also indicate that Saghata Upazila char dwellers are more vulnerable than those of Fulchhari Upazila in terms of natural capital, including land, natural resources, natural disasters, and climate variability. The social network of char dwellers is not the same as mainland dwellers. The study reveals that the social capital of Saghata Upazila char dwellers is better than that of Fulchhari Upazila char dwellers. Like other types of capital, the physical capital, including housing and production means, agricultural assets, and nonagricultural assets, of Saghata Upazila char dwellers is better than that of Fulchhari Upazila char dwellers. This indicates that Fulchhari Upazila char dwellers are more vulnerable than Saghata Upazila dwellers in terms of physical capital. The financial capital of char dwellers is very low due to limited access to financial organizations like microfinance institutions, nongovernmental organizations (NGOs), commercial banks, and other voluntary organizations. The results also indicate that Fulchhari Upazila char dwellers are more financially vulnerable than Saghata Upazila char dwellers. Due to poor communication, nongovernmental microfinance institutions (MFIs) are not willing to work in distant char areas. Similar cases exist for public organizations [52,64]. The officials of various service-oriented organizations are not willing to work in char areas because of the lack of modern and health facilities, lack of electricity, almost no market, no communication means, sandy soil, long walking distances during the winter season, sandy wind storms, and frequent flood inundation. These reasons also cause food insecurity, poverty traps, and vulnerable livelihoods [4,26,44].

3.4. Policy Implications

Climate-resilience development may be considered a critical issue for Bangladesh. Though the nation has already taken some initiatives through formulated projects for the development of the riverbank dwellers, the need remains for strengthening char-dweller capacity to address recurrent disasters. Char dwellers face seasonal food insecurity and chronic poverty due to employment unavailability from September to November every year because of their dependency on agriculture. In addition, they face flood inundation and riverbank erosion every year. By losing almost all kinds of livelihood assets, they become highly vulnerable. Self-help is restricted due to a vicious cycle of poverty and the frequent attacks of natural disasters [4,23].

The findings of this study can help formulate a context-specific intervention program for the vulnerable communities of char areas. Particularly, targeted intervention is required to improve the livelihood of female-headed households as they are more vulnerable than male-headed households. The various social safety-net programs from GO and NGO efforts have been largely inefficient in securing char-dweller livelihoods [65]; new social safety-net programs should be implemented [66,67]. Similarly, a long-term development program should be implemented to develop char–mainland communication networks, season-oriented transportation, access to basic services, and markets for developing alternative livelihood strategies [68]. Many financial organizations are not willing to work in char areas due to geographical isolation and communication barriers. The government should, therefore, take initiative to control and monitor banking and nonbanking financial organizations so as to target their activities toward char areas and offer char dwellers greater access to financial capital. Since the professions of char dwellers are mainly related to agriculture, agricultural-research organizations should be encouraged to develop char-area-specific crop varieties and facilitate technology-transfer systems. The above means would help to develop resilience to natural disasters and maintain a sustainable livelihood throughout the country.

4. Conclusions

Due to its geographical position, Bangladesh is easily susceptible to natural disasters. Similarly, char areas are isolated from the mainland and exist throughout the country’s vast river-delta regions. This study sought to analyze the livelihood vulnerability of char dwellers, who face regular natural disasters like flood inundation, riverbank erosion, and drought. The major livelihood components were analyzed by developing a context-specific holistic approach. It was not easy to collect data from char areas due to poor accessibility. The researcher walked for miles, and sometimes used a local boat to visit char villages and conduct face-to-face interviews with the respondents. The study reveals that char dwellers are vulnerable in terms of livelihood assets, irrespective of areas. LVI and CVI results show that both char-dweller groups are vulnerable to natural disasters. They also report a difference in variability between major components and subcomponents, and with respect to mainland proximity. The main drivers of livelihood vulnerability are livelihood strategies, weak social networks, low access to food, water, and health facilities, and limited access to agricultural and nonagricultural assets, and finance. Interviews indicate the char-dweller perception that a long-term development plan, including road construction, social forestry, year-round employment, and capacity building would be helpful to build resilience against vulnerability. The adaptive capacity of char dwellers should be strengthened through the improvement of communication, transportation, livelihood diversification, and access to basic public services.