A comparative study of farming and fishing households’ livelihood vulnerability in the Niger Delta, Nigeria

Multiple stressors such as climate shocks, environmental degradation and resource conflicts may pose a great challenge for African communities struggling to adapt to climate change. Yet, there is limited empirical evidence on the differential vulnerabilities of farming and fishing households to these triple stressors. Using a household survey in the Niger Delta region of Nigeria, we compare the vulnerability of farming and fishing households. We find that the farming households were more exposed to the triple stressors while the fishing households were more sensitive to the triple stressors owing to their poor physical and natural asset base. In addition, we find that the two groups share similar adaptive capacity and composite vulnerability scores. Overall, interventions such as credit schemes can enable fishing households to address their livelihood disruptions and build their asset endowment, and climate-responsive interventions such as early warning systems can partly reduce the climate exposure of farming households.


Introduction
Globally, climate change is expected to affect all countries, sectors, and livelihoods, albeit differently.It has reduced global agricultural total factor productivity by about 21% since 1961, and the effect is substantially more severe in warmer regions such as Africa and Latin America (Ortiz-Bobea et al. 2021).The number of people at risk of hunger due to climate shocks will increase by 33% by 2050 in the absence of adequate adaptation measures and sub-Saharan Africa is among the regions that face the most severe hunger impacts from climate change (Janssens et al. 2020).Thus, Africa is the most vulnerable continent to climate change (Intergovernmental Panel on Climate Change [IPCC] 2007; Asafu-Adjaye 2014).While all sectors of human endeavor in Africa are vulnerable to climate change, the agricultural sector is more adversely affected by it (Diffenbaugh and Burke 2019;Baarsch et al. 2020).Africa's substantial dependence on rain-fed agriculture and fisheries in coastal areas, widespread poverty, as well as its limited adaptive capacity exacerbate its vulnerability (Azzarri and Signorelli 2020;Asfaw et al. 2021).Sub-Saharan Africa, and in particular coastal regions, are highly susceptible to climate disturbances because of exposure to extreme weather events, and high dependence on climate-sensitive sectors and activities (Cline 2007;Zewdie 2014;Connolly-Boutin and Smit 2016).This is aggravated by the prevalence of weak support systems and the slow pace of economic development (IPCC 2007;Preston et al. 2008 andIPCC 2014).More worrisome is that Africa is exposed to other stressors, including conflicts, environmental degradation and population pressure, amongst others (Babatunde 2020;Call and Gray 2020;Adhvaryu, Kala, and Nyshadham 2022).
The combination of multiple stressors such as environmental degradation, resource conflicts and lack of resources for adaptation may pose a greater challenge for African communities struggling to adapt to climate change.This is particularly the case in the oil-rich region of Nigeria, where the twin burden of environmental degradation associated with oil spillage and conflict over resources have persistently affected the livelihoods of rural communities over the years (Ejiba, Onya, and Adams 2016;Osuagwu and Olaifa 2018;Babatunde 2020).There is a growing realization that climate-change studies must take into account the effects of other stressors (O'Brien et al. 2004;McCubbin, Smit, and Pearce 2015;Bennett et al. 2016).Understanding the complex interactions between climate change and other stressors will be essential in designing effective strategies for adapting to climate change (Tschakert 2007;O'Brien, Quinlan, and Ziervogel 2009;Staudt et al. 2013;R€ as€ anen et al. 2016;Lede et al. 2021).In this regard, while climate change is widely considered a threat to human health, food and water security and socio-economic development in Africa, and may undermine its ability to achieve the UN Sustainable Development Goals, an in-depth understanding of livelihood vulnerabilities to inform policy interventions should not be limited to climate change.
Previous theoretical and empirical studies in Nigeria and other African countries have often focused on climate change, environmental degradation, and conflict as standalone stressors, which provide partial insights on vulnerability to stressors.For instance, a large body of studies in different parts of Africa have focused on issues relating only to climate shocks (e.g.Azzarri and Signorelli 2020;Asfaw et al. 2021), environmental degradation (e.g.Kirui, Mirzabaev, and von Braun 2021;Usman et al. 2020) and conflict (e.g.George, Adelaja, and Weatherspoon 2020;Adhvaryu, Kala, and Nyshadham 2022).Only a few studies have looked at climate shocks and conflict jointly as stressors affecting vulnerability (e.g.Okpara, Stringer, and Dougill 2017;Mbaye 2020).Similarly, few studies have focused on climate shocks and environmental degradation jointly (e.g.Ahmed, Khan, and Augustine 2018; Call and Gray 2020).In addition, very few studies have focused on conflict and environmental degradation jointly (e.g.Onuoha 2008;Babatunde 2020).To our knowledge, no empirical study in a developing country setting has combined the three stressors in a single livelihood vulnerability assessment study at the local scale, which presents a knowledge gap that our study intends to address using a composite index approach.The composite index integrates two frameworks, including the sustainable livelihood framework (Chambers and Conway 1992) and the IPCC vulnerability framework (Hahn, Riederer, and Foster 2009;Busby, Smith, and Krishnan 2014;Koomson, Davies-Vollum, and Raha 2020;Asfaw et al. 2021), and has been noted to comprehensively capture the multi-dimensionality of vulnerability (Leichenko and O'Brien 2002).
By drawing on insights from the two frameworks, we are better able to assess the livelihood vulnerability of farming and fishing households.Specifically, we draw on how best to assess available capital or assets (natural, human, social, physical and financial) at the disposal of the households, which is at the core of the sustainable livelihood framework (Scoones 1998;Scoones 2009).This informs how the households make a living and how vulnerability can be reduced by making use of capital assets at the disposal of the households.To operationalize vulnerability to the triple stressors, we draw on the IPCC's vulnerability framework, which posits vulnerability as a function of exposure, sensitivity, and adaptive capacity (IPCC 2007).Thus, we are able to explain the vulnerability of the households' livelihoods in terms of their level of exposure to stressors, how sensitive they are to this exposure and their ability to resist, recover from and adapt to the impact of the exposure.
In this paper, we analyze the livelihood vulnerabilities of agricultural households to triple stressors of climate shocks, environmental degradation, and resource conflicts in the Niger Delta region of Nigeria.Specifically, we apply a composite climate-conflictenvironmental degradation vulnerability index to identify and compare differential vulnerabilities to the triple stressors faced by farming and fishing households, who are the predominant livelihood groups in the study region.In this way, we present a method for mapping vulnerability to the three stressors at the household level.Given that households are often faced with multiple stressors, a growing literature on vulnerability in different regions of the world increasingly recommends joint analysis of multiple stressors over a focus on a single stressor (O'Brien et al. 2004;Eakin and Wehbe 2009;Casale et al. 2010;McCubbin, Smit, and Pearce 2015;R€ as€ anen et al. 2016;Lede et al. 2021;Venus et al. 2022).This is because jointly analyzing multiple stressors accounts for the fact that the stressors often interact to alter livelihoods, and there could be some complementarities or tradeoffs in responses to the stressors, which cannot be addressed in studies that only consider a single stressor.In this regard, our study addresses the need to understand how the three stressors drive vulnerability amongst different livelihood groups in the region, and the ways in which adaptive capacities might be built to spur resilience or reduce vulnerability.
We make three contributions to different strands of literature on micro-level vulnerability assessment.First, we contribute to the large empirical evidence on livelihood vulnerabilities to multiple stressors in a developing country setting (e.g.Casale et al. 2010;Etwire et al. 2013;Bennett et al. 2016;Shackleton and Cobban 2016).Our focus on the Niger Delta region of Nigeria, a predominantly coastal area, is particularly important because of the growing manifestation of climate shocks (e.g.flooding) in the region, and its long-standing history of environmental degradation (e.g. oil spillage on land and water bodies) and resource conflicts (e.g.conflict over land) over the years.Second, we examine heterogeneity in livelihood vulnerabilities by exploring how different categories of agricultural households (e.g.farming and fishing households in our study region) respond to multiple stressors.This provides useful insights that can inform the proper targeting of policy interventions.The latter helps to allay concerns about promoting one-size-fits-all policy interventions in developing countries.Third, from a methodological perspective, we combine different indicators to construct a composite climate-environmental degradation-conflict vulnerability index drawing on the IPCC's concept of vulnerability in terms of exposure, sensitivity, and adaptive capacity (IPCC 2007).The indicators were selected deductively through a review of relevant literature and validated in the field, as applied in previous studies (e.g.Busby, Smith, and Krishnan 2014;Okpara, Stringer, and Dougill 2017;Koomson, Davies-Vollum, and Raha 2020;Asfaw et al. 2021).Our approach addresses some shortcomings of the approaches adopted by previous studiese.g.use of hypothetical adaptation in crop model for vulnerability assessment (Kandlikar and Risbey 2000), selection of indicators normatively (Vincent 2007), limited objectivity of experts in selection and aggregation of indicators (Brooks, Adger, and Kelly 2005;Alberini, Chiabai, and Muehlenbachs 2006).

Theoretical background
Livelihood has been defined as the means through which people make a living and this consists of their capabilities, assets at their disposal and activities they engage in to make a living (Chambers and Conway 1992).This closely relates to vulnerability, which is the degree to which socio-ecological systems are affected by some forms of hazard or, simply put, the capacity to be wounded.Vulnerability is a function of three components, namely exposure, sensitivity and adaptive capacity (IPCC 2001;McCarthy et al. 2001).The first component is concerned with the external side of risks, shocks and stress to which a system is subjected, while the last two components capture the internal side, which refers to the response and means of coping or adapting to the stress (Chambers 1989;F€ ussel and Klein 2006).Exposure is the nature and degree to which a system experiences stress (Adger 2006).The important characteristics of these stresses are magnitude, frequency and duration (Burton 1993).Sensitivity is the degree to which a system is affected by perturbations or stressors (Adger 2006).Adaptive capacity is the ability of a system to adjust to accommodate or cope with stress (F€ ussel and Klein 2006).This is a prerequisite for adaptation to occur and it involves the ability to harness a set of available assets to cope with stress, and the assets are a set of livelihood resources that individuals harness to build their livelihood adaptation strategies (Scoones 1998).
There is generally a consensus that a number of interacting factors or stressors (biophysical and socio-economic factors) shape vulnerability (Casale et al. 2010;O'Brien and Leichenko 2000;Reed et al. 2013).The biophysical drivers are factors related to biology and physical environment such as climate variability and change, land and water degradation, etc. while the socio-economic drivers are factors such as demographics, economics, institutions, policies, culture and conflicts.Vulnerability increases when exposure and sensitivity to hazards increase beyond the adaptive capacity of a socio-ecological system or region.This applies to our study region, a predominantly coastal area prone to flooding and coastal erosion.The region is vulnerable as it is faced with the menace of a degraded environment and exposure to resource conflict.
Several integrated frameworks are currently being used for the analysis of multiple stressors.One such framework is the concept of "double exposure" (Leichenko and O'Brien 2008), used to investigate the impacts of climate change in the context of economic globalization.Various case studies have applied the "double exposure" framework to show how climate change interacts with economic globalization to deepen the vulnerability of smallholder farmers (Silva, Eriksen, and Ombe 2010).Both Okpara, Stringer, and Dougill (2017) and Busby, Smith, and Krishnan (2014) adopted the double exposure framework to show how climate change interacts with conflict to deepen the vulnerability of countries and households.Conversations in this field typically draw upon the environmental security thesis (Homer-Dixon and Blitt 1998;Le Billon 2001) as the basis for a theoretical understanding of the role of environmental resources in conflict events.We build on insights from the "double exposure" framework as applied by Okpara, Stringer, and Dougill (2017) and Busby, Smith, and Krishnan (2014) and further extend it by incorporating a third exposure, which is environmental degradation.We also draw insights from other studies in the selection of indicators for construction of a livelihood vulnerability index (Koomson, Davies-Vollum, and Raha 2020;Asfaw et al. 2021).The focus of Asfaw et al. (2021) on only crop-farming households in Ethiopia and Koomson, Davies-Vollum, and Raha (2020) on only fishing households in a coastal area of Ghana provided a similar setting to our study, as we focused on both farming and fishing households.

Description of the study area
The study area is the Niger Delta region.It is situated on the Atlantic Coast of southern Nigeria where the river Niger divides into many branches (Uyigue and Agho 2007).It is the second biggest delta in the world, having a coastline covering around 450 kilometers, which ends at the mouth of the Imo river (Awosika 1995).The region extends to over 20,000 square kilometers and it is the largest wetland in Africa and is one of the three largest and richest wetlands in the world (CLO 2002).The region accounts for all the oil and gas exported from Nigeria, which represents 80% of the country's revenue (Obi 2009).Yet, it is among Nigeria's least developed regions with poverty and unemployment levels higher than the national average and lacking basic infrastructures such as electricity, healthcare facilities, roads, tap water, etc. (NDDC 2004).
The region is divided into four ecological zones, namely: coastal inland zone, mangrove swamp zone, freshwater zone, and lowland rainforest zone.The Niger Delta region officially comprises nine states, namely: Abia, Akwa Ibom, Bayelsa, Cross River, Delta, Edo, Imo, Ondo, and River States.It has 185 local government areas (LGAs) and over 40 ethnic groups in an estimated 3,000 communities (Idemudia 2009).The region has an estimated population of about 36 million (World meter 2020), and the majority depend on fishing and farming for their livelihoods.Artisanal fishing is mostly carried out in the region on a small-scale basis by self-employed fishermen and women using wooden or motorized canoes, rather than by commercial enterprises, but commercial trawlers do operate offshore.However, in recent times, aquaculture has begun to gain popularity because of the dwindling fish catch from the capture fisheries (Ibemere and Ezeano 2014).The major type of farming practiced in the region is crop farming.The main food crops grown include plantain, maize, yam, cassava, and vegetables.
Resource conflict in the Niger Delta includes struggles over agricultural lands and land with oil deposits.Conflicts over lands with oil deposits usually occur between the communities and the federal government, or between communities and multinational oil companies.Struggles over agricultural lands are usually between communities or individuals.The root cause of conflict in the region has been largely attributed to oil extraction, which isolates the locals from their land and livelihoods, (Obi 2009, Ikelegbe 2010;Obi and Rustad 2011).The activities of the multinational oil companies in the region spur oil spillage and gas flaring.Spilt oil on farmlands and water bodies destroys fish ecosystems, vegetation, and natural habitat.This, in turn, undermines rural livelihoods and spurs local grievances.For instance, between 1976 and 1996, about 4,600-7,000 oil spills were recorded with a total volume of 2.4 À 3.6 million barrels of oil wasted (Agbola and Olurin 2003;Iyayi 2004).We conducted the field survey activities in two states within the Niger Delta region (See Figure 1).

Sampling procedure
We used a multi-stage sampling technique to select the households used in the study.In the first stage, two states (Rivers and Bayelsa) were purposively selected out of the nine states due to their high dependence on cropping and fishing activities, notable pollution activities of oil companies which degrade the soil and water bodies, the prevalence of conflict, and the coastal nature of the states, which predisposes it to frequent flooding and coastal erosion.In the second stage, local government areas (LGAs) that are predominantly agrarian were selected, 13 LGAs out of 23 were selected from Rivers and 4 LGAs out of 8 were selected from Bayelsa.In the third stage, proportional random sampling was used to select 18 and 8 communities from the selected LGAs in Rivers and Bayelsa States, respectively.Finally, proportional random sampling was used to select the number of farming and fishing households in the selected communities, resulting in a sample size of 503 agricultural household heads (251 farming households and 252 fishing households).The United Nations (2008,(44)(45) sample size formula (see Equation ( 1)) was used to determine the number of households selected for the study.Using a confidence interval (Z) of 95%, 50% default value for prevalence of indicators (r), a sample size of 430 households was required.However, to account for possible missing values and outliers, the sample size was increased to 503.
where: N ¼ sample size, Z ¼ confidence interval (95% level is 1.96), r ¼ estimate of key indicators being measured (default value is 0.5), f ¼ sample design effect (has a default value of 2), k ¼ multiplier accounting for non-response (1.1), p ¼ proportion of the total population accounted for by the target population (0.4), n ¼ mean of household size ( 5), e ¼ precision level (10% precision level equals 0.01r).

Data collection
We employed cross-sectional data from a household survey, conducted between March and April 2018.The survey instrument used was a structured questionnaire administered to the respondents via face-to-face interviews by a well-trained team of enumerators and supervisors, including an open-ended questionnaire for focus group discussions with local experts (community leaders who are considered very knowledgeable about the prevalence of the triple shocks over the years in their respective communities) in each of the local government areas.The questionnaire was first pre-tested on a randomly selected sample of 20 respondents (10 farming and 10 fishing households) in February 2018, and some modifications were made to the questionnaire before the final data collection.We had two sets of questionnaires, one for the farming households and the other for the fishing households.The questionnaires had sections on household socio-demographic and institutional characteristics, livelihood income strategies, social and political networks, household income sources and expenditure, household assets, perceptions on climate shocks and impact, conflict, and environmental degradation.For ethical consideration, we designed a brief informed consent form and included it as part of the introductory note on the purpose of the survey and the survey team used it to obtain verbal consent for each respondent's willingness to participate in the survey.The dataset for this study is available in Harvard Dataverse (https://doi.org/10.7910/DVN/GLWYSZ).

Analytical framework
We used the composite index approach to calculate the vulnerability index for the farming and fishing households, as well as the pooled sample.The composite index has been noted to comprehensively capture the multi-dimensionality of vulnerability (Leichenko and O'Brien 2002).The index integrates two frameworks: the sustainable livelihood framework and the IPCC vulnerability framework (Busby, Smith, and Krishnan 2014;Koomson, Davies-Vollum, and Raha 2020;Asfaw et al. 2021).The sustainable livelihood framework provides a holistic approach to understanding how people make a living (Scoones 1998;Scoones 2009).At its core is the assessment of the available capital or assets (natural, human, social, physical and financial) at the disposal of people from which they make a living and an evaluation of the vulnerability context (shocks, stresses, trends and seasonality) in which this capital exists.We adopted the IPCC's definition of vulnerability as a function of exposure, sensitivity, and adaptive capacity as a starting point in operationalizing vulnerability to the triple stressors (IPCC 2007).This has been applied in some empirical vulnerability studies (e.g.Islam et al. 2014;Okpara, Stringer, and Dougill 2017;Senapati and Gupta 2017;Koomson, Davies-Vollum, and Raha 2020;Asfaw et al. 2021).
To derive the three major componentsexposure, sensitivity, and adaptive capacity, eight sub-components were used.For the exposure component, we captured three sub-components: exposure to climate shocks, resource conflict and environmental degradation.Sensitivity was measured by considering two sub-components: current state of food, water and health status, and physical/natural assets.Adaptive capacity was Overall, the analytical framework helps to give a comprehensive view of the vulnerability of livelihoods to the three stressorsclimate shocks, environmental degradation, and resource conflicts in our study setting.In particular, the framework shows the links between livelihood vulnerability and the three stressors, the major and subcomponents used in calculating the composite vulnerability index and the indicators that make up each sub-component used in operationalizing vulnerability (see Figure 2 for a summary of the framework).A detailed description of all indicators associated with each of the sub-components, their units of measurement and the basis for selecting the indicators are presented in Table 1A in the Appendix (online supplemental data).

Data analysis
We analyzed the collected data starting with the pooled sample (both farming and fishing households) and then the disaggregated samples (first sample: farming households and second sample: fishing households) using Stata.The first step in processing the raw data to determine the vulnerability index was to address the different units of measurement associated with the different indicators of the sub-components.This has to do with standardization that entails the transformation of each indicator into a uniform scale to allow for comparison and aggregation into a single index (Vincent 2007).We adopted the maximum-minimum standardization technique used by Hahn, Riederer, and Foster (2009) in standardizing the indicators.The formula is stated as: where Index S ¼ standardized indicators for each livelihood group, S ¼ raw data for the indicators associated with each livelihood group, S min ¼ minimum value of the indicator, S max ¼ maximum value of the indicator.The next step involved the assignment of weights.We used the equal weighting method employed in previous empirical studies (Hahn, Riederer, and Foster 2009;Okpara, Stringer, and Dougill 2017;Adu et al. 2018), which assumes that all sub-components contribute equally to the overall index.The standardized indicators were averaged to derive the value for each sub-component using the formula in Equation (3): where M i ¼ one of the eight sub-components for each livelihood group, index s i ¼ the standardized indicators that make up each sub-component, n ¼ number of indicators in each sub-component.
Drawing on the IPCC's framework of vulnerability, the next step involved aggregating the sub-components to derive the major componentsexposure, sensitivity, and adaptive capacity.We derived the major components using the formula in Equation (4): where M j ¼ one of the three major components for each livelihood group, index s j ¼ the standardized sub-components that make up each major component, n ¼ number of sub-components in each major component.Finally, the major components were averaged using the formula in Equation ( 5) to derive the composite vulnerability index.
The adaptive capacity was subtracted from one because it reduces vulnerability.The CVI was scaled from 0 (least vulnerable) to 1 (most vulnerable).

Socio-economic characteristics of households
The socio-economic characteristics considered in this study include gender of household head, marital status of household head, age of household head, level of education of household head, household size, marital status, membership of an association, access to health care and extension services, number of years of experience in farming/fishing and participation in off-farm work.A summary of the socioeconomic characteristics is shown in Table 1.About 62% of the sampled households were maleheaded households, the majority (77.3%) of them were married and only a few (3%) of them had no formal education.Most (94%) of the households had no access to extension services, no access to credit (about 88%) and do not belong to any farmer/ fisher-based association (89%).About 79% of the households had access to health care and 51% of the households were engaged in off-farm work.On average, the sampled household heads were aged 48 years, had 9 years of schooling, a household size of 7, a farming/fishing experience of 25 years, a farm size (for farming households) of 0.3 hectares and a gross annual income of N698,955.8($1,839).In terms of access to infrastructure, the average distances to a healthcare facility and water source were 2.8 km and 5.2 km, respectively.

Livelihood vulnerability indices
The computed values for the various indicators used in computing the composite vulnerability index are presented in Table 2.In the exposure component, the farming households had higher exposure to climate shocks with an index of 0.60, while the fishing households had an index of 0.45, and the observed difference between the two groups is statistically significant at the 1% significance level.This is expected given the higher value of the indicating variables -shifts in temperature, rainfall, climaterelated losses, and the number of flooding events reported by the farming households.
The resource conflict index for the farming households (0.35) was significantly higher than the index for the fishing households (0.22) at the 1% significance level.The farming households were more involved in conflicts than their fishing counterparts, especially conflicts over land.The fishing households had a lower likelihood of conflict because most of the water bodies where they fish are open to anyone who wants to fish.Hence, there is less struggle for fishing space.Also, the farming households had higher reports of land/water conflict, as most of the farmers reported witnessing other people in the community being involved in conflicts over land.The fishing households had a higher feeling of insecurity than the farming households, probably because of the much larger volatility (in terms of resource conflict) of some of the fishing-dominated communities in the study area (e.g.Degema, Buguma and Abonima communities).For conflict-related losses, the farming households reported more losses than the fishing households.The losses incurred by the farming households include destruction of their crops, property, injury, money spent in the treatment of injuries and even death in some cases.
The environmental degradation index for the farming households (0.56) was higher than the index for the fishing households (0.51) but the observed difference is not significant at the conventional significance levels.The plausible reason for the higher score on the environmental degradation sub-component by farming households was because they reported a higher incidence of land pollution and pollution-related losses than the fishing

À1.31
Note: HHs ¼ Households, maximum and minimum values of the indicators were 100 and 0 respectively except for the average number of floods, average number of days household head was ill, index of house quality and diversification which had maximum values of 7, 15, 0.17 and 4 respectively and minimum values of 0, 0, 0.34 and 1 respectively.
households.Even though the incidence of land pollution reported by fishing households was low, they reported higher incidence of water pollution than the farming households.
In the sensitivity component, the farming households were better off in terms of health, food, and water status, with an index of 0.31, while the fishing households had an index of 0.24, and the observed difference between the two groups is statistically significant at the 1% significance level.For the indicators of health, food and water status, the farming households reportedly travel lower distances to health care facilities than the fishing households.However, the average number of times the farming households were sick and unable to carry out their livelihood activities were the same as the fishing households.A larger share of the farming households depend solely on their own farm as their main source of food while a larger share of the fishing households depend on natural water sources such as stream/rivers and lakes as their only source of water.A larger share of the farming households have other water sources such as pipe-borne water which could be private pumps or community pumps, boreholes, and wells.
The index of the physical and natural assets sub-component for the fishing households was higher (0.43) than that of the farming households (0.11) and the observed difference is statistically significant at the 1% level.This suggests that the fishing households were more vulnerable with respect to this sub-component compared to the farming households.The physical and natural assets sub-component is comprised of two indicators -house quality index and land tenure/access.The farming households had a higher house quality index of 0.34 than the fishing households which had an index of 0.21.A plausible reason for this could be because the majority of the farming households live in relatively better houses, made with cement and iron sheets as roofing material and, on average, had two adults sleeping in a room.Only a very small share (1.2%) of the farming households reported not having access to land for their farming, while about 66% of the fishing households reported not having access to land for farming.Most of the lands operated by farming households are owned by the households.
In the adaptive capacity component, the index of the socio-demographic profile subcomponent for the farming households (0.70) was lower than the index for the fishing households (0.80) and the observed difference is statistically significant at the 1% significance level.This suggests that the farming households were more vulnerable with respect to this sub-component.In terms of the indicators of this sub-component, the fishing households had a larger adult workforce, larger male-headed households while the farming households had a larger share of household heads who are educated at least to secondary school level compared with the fishing household heads.In terms of farming/ fishing experience, both groups have a similar number of years of experience.
The livelihood income strategies index for the farming households (0.33) was significantly higher than the index for the fishing households (0.23) at the 1% significance level, which suggests that the fishing households were more vulnerable with respect to this sub-component than the farming households.In terms of the indicators of this sub-component, a larger share of the fishing households did not receive remittances from family members or friends living and working outside the community and had no access to credit for their livelihood activities compared with the farming households.However, a larger share of the farming households indicated that their income was enough to cover important expenses such as food, water, shelter, education, and health compared with the fishing households.In addition, the farming households, on average, engaged in more livelihood activities than the fishing households.
The index of socio-political network for the farming households (0.35) was significantly higher than the index for the fishing households (0.25) at the 1% significance level, which suggests that the fishing households were more vulnerable.This is expected because socio-political networks reduce vulnerability.In terms of the indicators for this sub-component, a larger share of the fishing households does not belong to any association with the farming households.A share of the fishing households received external support in difficult times and reported local cooperation in the communities where they lived during difficult times compared with the fishing households.However, both groups had similar access to climate information.
Figure 3 shows that the farming households were more vulnerable in relation to two of the exposure components (climate shocks and resource conflict), to one of the sensitivity components (health, food and water status) and to one of the adaptive capacity components (socio-demographic profile).The fishing households were more vulnerable in relation to one of the sensitivity components (physical and natural assets) and two of the adaptive capacity components (livelihood income strategies and sociopolitical network).Figure 4 shows some considerable differences in exposure and sensitivity components between the farming and fishing households.Overall, the farming and fishing households share a similar vulnerability score of 0.42 and 0.43 respectively, which indicates moderate vulnerability to climate shocks, environmental degradation, and resource conflicts across the two groups.
To gain better insights on vulnerability, we present the vulnerability level for the farming and fishing households in Table 3.The majority (81.9%) of the surveyed households fell into the category of moderately vulnerable and only a few (2%) of them fell into the highly vulnerable group.This stems from the values of the indicators associated with the components of exposure, sensitivity and adaptive capacity of the households, as shown in Table 2.A larger share of the farming households fell into the category of moderately vulnerable compared with the fishing households.While none of the farming households fell into the highly vulnerable category, about 4% of the fishing households fell into the highly vulnerable category.

Discussion
Climate shocks which are reflected in variability in rainfall and temperature, resourceuse conflict in terms of water and land, and environmental degradation in terms of water and land degradation are prevalent livelihood stresses in different parts of Africa (Asfaw et al. 2021;Azzarri and Signorelli 2020;Mbaye 2020;Usman et al. 2020).The triple stressors may interact to alter the livelihoods of natural resource-dependent households, such as the farming and fishing households in the Niger Delta region of Nigeria.Using a relevant set of indicators, we developed and applied a composite vulnerability index to generate useful insights about the factors that contribute to the vulnerability of the farming and fishing households to the triple stressors.Such insights can inform policy initiatives on the indicators to properly target vulnerability reduction.
Our study provides empirical evidence to demonstrate that climate and conflictrelated losses are key stressors that impact livelihoods, especially in the case of farming households in our study context.The over-dependence on own farm for food could  explain the reason why farmers reported climate-related losses more than the fishing households.Though the farming households were more exposed to climate-related shocks than the fishing households, the former had higher adaptive capacity.This stems from their relatively higher socio-political networks (membership of associations, access to external assistance and local cooperation) and livelihood income strategies (remittances, access to credit, diversification and having an income sufficient to cover expenses).This finding is contrary to Okpara, Stringer, and Dougill (2017) who reported that the fishermen were more vulnerable to climate-related losses, but had more adaptive capacity in their study setting.We find that fishing households were more vulnerable to environmental degradation of their water bodies, resulting in low fish catch.This finding is consistent with Koomson, Davies-Vollum, and Raha (2020) who reported that household incomes reduced significantly due to the interaction of climatic and non-climatic factors in Ghana.Despite their exposure to environmental degradation, they have a higher sociodemographic profile than the farmers, such as education, experience and adult workforce, which builds into their adaptive capacity.Relatedly, the farming households also reported that their lands were being polluted and are experiencing losses resulting from this pollution.
Our findings show that although both farming and fishing households had a low livelihood income strategies index and socio-political network, the fishing households were worst off.This is likely because the majority of them depend only on fishing with limited livelihood diversification.The farming households were more diversified, engaging mainly in trading, artisanship, and food processing.Diversification is a viable means through which people reduce risk or vulnerability and it can improve household welfare, such as a reduction in poverty, as shown in previous studies (Coulthard 2008).Enhancing the adaptive capacity of the fishermen by supporting their existing livelihood strategies and facilitating a more diverse and flexible fishery system will reduce their vulnerability to these stressors.The fishing households had a lower index for socio-political networks because only a small fraction of them belonged to any formal association and received external assistance during difficult times.This accords with Iwasaki, Razafindrabe, and Shaw (2009) who opined that the lack of social cohesion and local cooperation could lead to loss of assets such as boats and buildings by fishing households during extreme events.However, our result is not consistent with the findings of Okpara, Stringer, and Dougill (2017), which showed that fishermen were better off in terms of adaptive capacity through socio-political networks and livelihood income strategies than farmers.
Our findings suggest that climate shocks contributed more to the local exposure challenges faced in the area than resource conflict and environmental degradation.This is contrary to the findings of Okpara, Stringer, and Dougill (2017), which showed that water conflict contributes more to the exposure element than climate variability.While a higher percentage of farming households reported more shifts in temperature, rainfall, and the absolute number of flooding events, this should be interpreted with caution, as it may not be the case, since the farming and fishing households are both located in the same region.The self-reported data on shifts in temperature and rainfall in our study are subject to recall bias because the households that are affected by these shifts are likely to report more shifts in temperature and rainfall than those less affected.We acknowledge that a more appropriate method would have been to use spatially explicit rainfall and temperature data, but, unfortunately, we do not have access to reliable weather data for our study area.This calls for more empirical studies to integrate household survey data and weather data to better understand vulnerability to climate shocks.
Overall, although the farming households were more exposed to the triple stressors, the fishing households were more sensitive to the stressors due to their poor physical and natural asset base relative to the farming households, which makes them more vulnerable.On the other hand, the farming households were less sensitive compared to the fishing households because they were better off in terms of health, food, and water status.These findings show that fishing households were more vulnerable than farming households which contradicts the findings of Okpara, Stringer, and Dougill (2017) that farming households were more vulnerable.
While our study provides relevant insights into the differential vulnerability of farming and fishing households to triple stressors, we acknowledge that our study has some limitations.First, there is an inherent limitation associated with the use of any index approach, due to the likelihood of masking underlying multidimensional realities that are shaping vulnerability and the relatively subjective process that is involved in selecting the indicators and the weighting schemes.Second, while we considered resource conflict largely arising from conflict over land in the case of farming households and conflict over water in the case of fishing households, a more in-depth consideration of resource-specific conflict in our study design would have generated more insights from our study.Third, we rely on cross-sectional data, which makes it impossible to explore temporal heterogeneity in livelihood vulnerability.The use of panel data would have provided better insights into time-varying aspects of livelihood vulnerability.Finally, our study only considered household-level analysis.Our study could be further refined to focus on the regional context to assess other factors operating beyond the household level, which shape local vulnerability.

Conclusions and recommendations
In this study, we assess the livelihood vulnerabilities of agricultural households to triple stressors of climate shocks, environmental degradation, and resource conflicts in the Niger Delta region of Nigeria.In particular, we focus on the differential vulnerability of two livelihood groups -farming and fishing households.We employed household survey data elicited from 503 agricultural households to develop a composite vulnerability index for assessing the vulnerability of the households to the triple stressors.The results of the study showed that the farming households were more exposed to the triple stressors of climate shocks, resource conflict and environmental degradation.The high exposure of the farmers to climate shocks and environmental degradation is not surprising given the rising sea level and floods often recorded in the region, as well as oil spillage by multinational companies.Reducing exposure to climate shocks may be difficult at the micro level, but adaptive capacity and sensitivity can be improved to enhance resilience to climate shocks, especially for the fishing households who showed greater sensitivity and lower adaptive capacity compared to the farming households.Given that the fishing households were worst off in terms of health, food and water status and had a relatively poorer physical and natural asset base, which made them more sensitive to the triple stressors, there should be well-concerted policy interventions tailored to the fishing households.This is because policy interventions in Nigeria have often focused on farmers with limited attention to fishermen.
Policy interventions to improve the adaptive capacity of both groups should strongly consider livelihood diversification strategies, among others.While there are credit schemes for farmers to expand their production and diversify their means of livelihood, little attention has been paid to developing such schemes for fishermen.Examples of the schemes that have been tailored to farmers in recent years include the Agricultural Credit Guarantee scheme (ACGS), the Growth Enhance Scheme (GES) and the Anchor Borrowers Programme (ABP), among others.These schemes mainly support farmers through the provision of loans to boost the production of some selected agricultural commodities.In addition, some of these schemes provide crop farmers with agricultural inputs, such as improved seeds and fertilizers.With a much higher sensitivity among the fishing households largely due to poor asset endowment, government and development partners should promote credit schemes to enable them to overcome challenges associated with their livelihood disruptions.Such credit schemes should be designed to provide credit in cash and in kind to allow the fishing households to have timely access to improved fishing equipment, processing facilities, infrastructure and essential services to enhance their fishing operations.Having such credit schemes can help them re-build their asset base in the event of climate shock, conflict or pollution resulting from oil spillage.Given the fact that the farming households were more exposed to the triple shocks, especially in terms of climate shocks, climate-responsive interventions such as early warning systems should be put in place to reduce their exposure to climate shocks.Finally, it is important that policy actors and other stakeholders promote policies and locally tailored programmes aimed at reducing vulnerability to the triple stressors, especially as the stressors often interact to alter livelihoods.Such policies and programmes are better designed as a package of interventions to simultaneously address the triple stressors.These could include a mix of some of the following interventions that have been applied in other developing country settings, namely livelihood diversification, climate-resilient production technologies, information transfer, index-based insurance, payment for environmental protection, credit support, and community peace clubs.In addition, proper monitoring of the activities of multi-national oil companies can curtail the menace of oil spillage and gas flaring that degrade the environment.Overall, bundling some of these interventions may likely be more effective than the promotion of a single intervention.

Figure 1 .
Figure 1.Location of the study area.
measured by considering three sub-components: socio-demographic profile, livelihood income strategies, and social/political networks.The eight sub-components are composed of different indicators, which were selected deductively from a review of relevant literature (e.g.Islam et al. 2014; Okpara, Stringer, and Dougill 2017; Senapati and Gupta 2017; Amevenku et al. 2019).The indicators were validated through a focus group discussion with local experts (community leaders) who are deeply knowledgeable about the triple stressors that we considered.

Figure 3 .
Figure 3. Vulnerability radar chart of sub-components of the composite vulnerability index for fishing and farming households.Note: The composite vulnerability index in the chart ranges from zero to one, indicating from least vulnerable to most vulnerable.

Figure 4 .
Figure 4. Indexed major components and overall composite vulnerability scores for farming and fishing households.

Table 1 .
Socio-economic characteristics of households.

Table 2 .
Computed values of livelihood vulnerability indices.

Table 3 .
Vulnerability levels of farming and fishing households.