Rethinking climate vulnerability in livestock systems: evidence from exposure, adaptation, and within-system heterogeneity in West Africa

Abstract

Climate vulnerability in livestock systems is often assumed to be intrinsically linked to production system characteristics. This study challenges this assumption by examining the relative roles of exposure, adaptive capacity, and production systems in shaping vulnerability outcomes in West Africa. Using household-level data from 1,126 livestock producers in Benin, Togo, and Côte d’Ivoire, we construct a Climate Vulnerability Index (CVI) combining exposure, sensitivity, and adaptive capacity. We employ a stepwise econometric approach based on Ordinary Least Squares (OLS) models, complemented by variance decomposition analysis, to assess the determinants of vulnerability. The results show that while semi-intensive and transhumant systems initially appear more vulnerable than sedentary systems, these differences diminish once exposure and adaptive capacity are accounted for. Exposure significantly increases vulnerability, whereas adaptive capacity exerts a strong mitigating effect. Contrary to common assumptions, no significant interaction or non-linear effects are found, suggesting that vulnerability is driven by additive and linear processes. Furthermore, the analysis reveals substantial heterogeneity within production systems, with more than 97% of the variation in vulnerability occurring within systems rather than between them. These findings provide strong evidence that climate vulnerability is not system-driven but process-driven, shaped primarily by the balance between environmental stress and adaptive capacity. The study contributes to the literature by highlighting the limits of system-based classifications and emphasizing the importance of household-level conditions. From a policy perspective, the results suggest that reducing vulnerability requires strengthening adaptive capacity and addressing exposure, rather than promoting specific livestock production systems. This calls for more context-sensitive and targeted adaptation strategies in West African livestock systems.

1 Introduction

Climate change presents profound challenges for agricultural systems globally, with livestock-dependent livelihoods in sub-Saharan Africa facing acute risks due to increasing climate variability, erratic rainfall, prolonged droughts, and rising temperatures (Aboye et al., 2023; Ahmad and Afzal, 2021; Bashiru and Oseni, 2025; Ngongolo and Gayo, 2025; Wongnaa et al., 2024; Zougmoré et al., 2016). These changes are altering the availability of water and pasture, directly affecting livestock productivity, herd dynamics, and the economic stability of pastoral and agropastoral households (Bashiru and Oseni, 2025; Ngongolo and Gayo, 2025; Wongnaa et al., 2024; Zougmoré et al., 2016). In contexts where livestock systems rely heavily on climate-sensitive ecosystems and face limited institutional support, these impacts are especially pronounced (Bashiru and Oseni, 2025; Ngongolo and Gayo, 2025).

Livestock production systems in the region are highly diverse, including sedentary, semi-intensive and transhumant forms (Azalou et al., 2023; Devendra et al., 2005; Koutchoro et al., 2025). This diversity is often assumed to explain differences in vulnerability, with mobile systems typically considered more exposed and intensive systems viewed as more resilient (Montcho et al., 2022; Zougmoré et al., 2016). However, this system-based perspective has been increasingly challenged. Emerging evidence suggests that vulnerability is not intrinsic to production systems but results from the interaction between environmental exposure, socio-economic conditions, and access to key resources (Madaki et al., 2025; Tofu et al., 2025). This raises an important question: to what extent do differences in climate vulnerability reflect the type of production system, and to what extent do they reflect differences among households operating within the same system?

The vulnerability framework conceptualizes vulnerability as a function of exposure, sensitivity, and adaptive capacity (Adger, 2006; Assani et al., 2024; Descheemaeker et al., 2016; Edmonds et al., 2020; Tofu et al., 2025; Zougmoré et al., 2016). Empirical studies show that access to resources—such as water, veterinary services, financial capital, and information—plays a decisive role in shaping vulnerability outcomes (Gautier et al., 2016; Madaki et al., 2025; Montcho et al., 2022; Munyaka et al., 2025; Tofu et al., 2025). Studies of pastoral and agro-pastoral households also show that vulnerability can differ substantially among households exposed to similar climatic conditions, reflecting differences in wealth, resource access, social networks, household characteristics, and adaptive capacity (Ayantunde et al., 2015; Opiyo et al., 2014; Qin et al., 2022; Segnon et al., 2021). Yet, much of the empirical literature continues to treat production systems as homogeneous categories, overlooking substantial variation within systems.

Recent research highlights that vulnerability can vary significantly within the same production system due to differences in resource access, adaptation strategies, and socio-economic conditions (Montcho et al., 2022; Munyaka et al., 2025; Wongnaa et al., 2024). Other studies have documented differences in vulnerability across livestock production systems (Berhe et al., 2017; de Glanville et al., 2020; Klinck et al., 2022), while also pointing to differences in resources and adaptive capacity within these systems. However, these two dimensions of variation have rarely been considered together. There remains limited empirical evidence on the relative importance of these factors and the extent to which vulnerability is driven by system characteristics versus underlying household conditions, particularly in West African livestock systems. More specifically, it remains unclear how much of the observed variation in household climate vulnerability is associated with differences between production systems and how much occurs among households within the same system.

This study addresses this gap by moving beyond system-based classifications to examine the mechanisms and patterns driving vulnerability. Specifically, it investigates the relative roles of livestock production systems, exposure to climatic stress, and adaptive capacity in shaping climate vulnerability among livestock producers in West Africa. It also examines whether adaptive capacity conditions the association between exposure and vulnerability and whether vulnerability varies substantially among households within the same production system. The central research question is: What drives climate vulnerability in livestock systems in West Africa?

Using micro-level data from 1,126 livestock-producing households across Benin, Togo, and Côte d’Ivoire, this study adopts a process-based approach to vulnerability. The results show that while production systems are initially associated with vulnerability differences, these effects weaken substantially once exposure and adaptive capacity are taken into account. Exposure significantly increases vulnerability, whereas adaptive capacity plays a strong mitigating role. Importantly, the findings provide no evidence of interaction or non-linear effects, suggesting that vulnerability is shaped by additive and linear processes. Furthermore, the analysis reveals substantial heterogeneity within production systems, with the vast majority of variation in vulnerability occurring within systems rather than between them.

These results demonstrate that climate vulnerability is not system-driven but primarily shaped by the balance between environmental stress and adaptive capacity. They also show that production-system categories alone provide limited information about the distribution of vulnerability among households. By highlighting the importance of intra-system heterogeneity, this study contributes to a more nuanced understanding of vulnerability dynamics and challenges approaches that rely on production system classifications alone.

2 Conceptual framework and hypothesis

2.1 Empirical literature on climate vulnerability in livestock systems

Previous studies have examined climate vulnerability in livestock systems from different perspectives. Many of them assess vulnerability through exposure, sensitivity, and adaptive capacity. These dimensions are often combined into composite indices, such as the Livelihood Vulnerability Index and related household-level measures (Bahta et al., 2025; Bahta and Nyaki, 2024; Gatew and Guyo, 2024; Opiyo et al., 2014; Qin et al., 2022). The findings consistently show that climate exposure is only one part of the vulnerability process. Household characteristics, access to resources and services, livelihood diversification, social networks, and access to climate information can also shape how households cope with climatic stress (Ayantunde et al., 2015; Opiyo et al., 2014; Qin et al., 2022).

Other studies have compared vulnerability across different livestock and livelihood systems. Evidence from pastoral, agro-pastoral, semi-pastoral, smallholder, and semi-commercial systems shows that production arrangements differ in their access to water, forage, productive assets, markets, and other resources needed to respond to climate stress (Berhe et al., 2017; de Glanville et al., 2020; Kapruwan et al., 2024; Klinck et al., 2022; Koutchoro et al., 2025). These differences can translate into different levels of vulnerability. However, households belonging to the same production system may also face different levels of exposure and have different capacities to respond. For example, Ayantunde et al. (2015) found differences in drought vulnerability within agro-pastoral communities according to wealth, gender, location, social networks, and previous drought exposure. Similarly, Segnon et al. (2021) identified distinct household vulnerability profiles within and across communities in the semi-arid areas of Mali. Qin et al. (2022) also showed that vulnerability varied according to household life-cycle characteristics, while Guye et al. (2023) found important spatial differences associated with exposure and adaptive capacity.

This literature shows that climate vulnerability cannot be understood solely from the type of livestock production system. It also depends on the conditions faced by individual households and their ability to respond to climate-related stress. However, previous studies have generally focused either on measuring household vulnerability and its determinants or on comparing different livelihood and production systems. Less attention has been given to the relative importance of between-system differences and within-system differences in explaining observed variation in vulnerability. This issue is particularly relevant in West Africa, where livestock production systems are diverse and households operating within the same system may face different environmental and socioeconomic conditions.

The present study addresses this issue by examining climate vulnerability at the household level across livestock production systems in Benin, Togo, and Côte d’Ivoire. The main characteristics and findings of the studies reviewed, together with their analytical approaches and relevance to the present study, are summarized in Table 1. We assess whether vulnerability differs across production systems and, at the same time, examine how much variation remains among households within each system. This approach allows us to move beyond the assumption that a production system represents a homogeneous level of vulnerability.

2.2 Theoretical and conceptual framework

This study adopts a process-based and widely used vulnerability framework in which climate vulnerability results from the interaction of exposure, sensitivity, and adaptive capacity, rather than from any fixed, inherent property of livestock systems (Asfaw et al., 2021; Opiyo et al., 2014; Tofu et al., 2025; Weng et al., 2023). In this view, vulnerability is an outcome of the way socio-economic and biophysical conditions combine under climatic stress and is best captured through composite indices that integrate these three components (Asfaw et al., 2021; Karimi et al., 2018; Opiyo et al., 2014; Tofu et al., 2025). In this study, this framework provides the basis for examining how differences in production systems, exposure, and adaptive capacity are associated with differences in household climate vulnerability.

Exposure denotes the degree to which livestock systems are subjected to climatic hazards such as recurrent droughts, rainfall variability, heat stress, and water scarcity (Habte et al., 2022; Opiyo et al., 2014; Weng et al., 2023). Sensitivity reflects how strongly production and livelihoods respond to these stressors, for example through feed shortages, livestock disease, mortality, and degradation of rangelands or water resources (Habte et al., 2022; Tofu et al., 2025, 2023). Adaptive capacity is the ability of households and systems to anticipate, cope with, and adjust to such shocks, shaped by access to assets, services, information, institutions, and diversification options (Karimi et al., 2018; Ntombela et al., 2024; Ogoudou et al., 2025; Tofu et al., 2023). These three dimensions are therefore central to the construction of the Climate Vulnerability Index used as the outcome measure in this study.

Consistent with index-based applications in pastoral and agro-pastoral areas, vulnerability can be expressed as the net effect of these components, where exposure and sensitivity jointly form the potential impact, which is offset by adaptive capacity (Belay and Lebeza, 2024; Opiyo et al., 2014; Tofu et al., 2025; Weng et al., 2023). Empirical studies suggest that higher adaptive capacity is generally associated with lower vulnerability, although the nature of this relationship may vary across contexts (Belay and Lebeza, 2024; Opiyo et al., 2014; Tofu et al., 2025, 2023; Weng et al., 2023). Beyond these core dimensions, the empirical framework considers production system and country as contextual factors that may be associated with differences in vulnerability, while household characteristics are included to account for differences in socioeconomic conditions.

The framework also provides a basis for examining whether these associations vary across households and production systems. In particular, the study considers whether adaptive capacity changes the association between exposure and vulnerability, whether the association between adaptive capacity and vulnerability is non-linear, and whether the association between production system and vulnerability varies according to adaptive capacity. These relationships correspond to the interaction and non-linear terms examined in the empirical models.

2.3 Hypotheses

Building on this conceptual framework and the empirical evidence reviewed above, the study formulates six hypotheses to examine the associations between production systems, country context, exposure, adaptive capacity, and household climate vulnerability.

First, H1 posits higher vulnerability in semi-intensive and transhumant systems relative to sedentary systems. Evidence from pastoral and semi-arid rangelands suggests that mobile and extensively managed herds are often located in areas with higher climate variability and weaker infrastructure, potentially leading to higher exposure and sensitivity (Karimi et al., 2018; Weng et al., 2023).

Second, H2 builds on comparative studies showing that vulnerability may differ across countries and regions due to variations in institutions, markets, public services, and policy environments, even under similar climatic conditions (Karimi et al., 2018; Ntombela et al., 2024; Tofu et al., 2023).

Third, H3 examines whether adaptive capacity moderates the association between exposure and vulnerability. While several studies suggest that adaptive capacity can mitigate the effects of exposure, the extent to which it modifies or conditions this relationship remains an empirical question (Ntombela et al., 2024; Opiyo et al., 2014; Weng et al., 2023). This study therefore tests whether the association between exposure and vulnerability varies with adaptive capacity.

Fourth, H4 examines whether the association between production system and vulnerability varies according to adaptive capacity. This hypothesis builds on evidence that production systems differ in their exposure, resource access, and adaptive capacity, while households within these systems may also have different capacities to respond to climatic stress (Berhe et al., 2017; de Glanville et al., 2020; Gatew and Guyo, 2024; Klinck et al., 2022). The study therefore examines whether differences in adaptive capacity are associated with different vulnerability patterns across production systems.

Fifth, H5 examines whether adaptive capacity has non-linear a non-linear association with vulnerability. Previous research suggests that threshold effects may exist, whereby improvements in assets or institutional access lead to disproportionate reductions in vulnerability, although such patterns are not consistently observed across contexts (Azine et al., 2025; Belay and Lebeza, 2024).

Finally, H6 is motivated by evidence that vulnerability may vary substantially within production systems due to differences in household characteristics, resource access, and adaptation strategies, potentially exceeding variation observed between systems (Alary et al., 2022; Descheemaeker et al., 2016).

On this basis, the study focuses on three dimensions: (i) how livestock production systems are associated with vulnerability patterns, (ii) the respective roles of exposure and adaptive capacity in shaping vulnerability, and (iii) the extent of intra-system heterogeneity in vulnerability outcomes. This conceptualization guides the construction of the Climate Vulnerability Index and provides a framework for testing hypotheses H1–H6.

3 Methodology

3.1 Study area and source of data

This study is based on primary data collected from livestock producers in three West African countries: Benin, Togo, and Côte d’Ivoire (Figure 1). Within each country, the survey was conducted in key livestock-producing regions. In Benin, data were collected in the Borgou (Tchaourou, N’Dali and Bembereke) and Alibori (Gogounou) departments; in Côte d’Ivoire, in the Bounkani and Tchologo regions; and in Togo, in the Oti-Sud, Kpendjal-Ouest, and Kpendjal areas. These regions were selected due to their strong engagement in livestock production and their exposure to climatic variability (Comoé and Siegrist, 2015; Djakpa et al., 2024).

Figure 1

The three countries were selected to capture differences in livestock production systems and in the climatic and institutional contexts in which livestock-producing households operate. The selected areas also provide a relevant setting for comparing sedentary, semi-intensive, and transhumant production systems. The unit of analysis is the livestock-producing household, with each observation corresponding to a livestock-producing household. Data were collected through a structured survey administered to household heads.

3.2 Sampling strategy and sample size

The sampling procedure was conducted within the selected study areas. The sampling frame was established using available agricultural census data for the respective study areas. In Benin, the reference population comprised 196,723 livestock producers in the Borgou and Alibori departments, according to the Recensement National de l’Agriculture (RNA) in 2021. In Côte d’Ivoire, the reference population comprised 55,324 agricultural households in the selected areas, according to the Recensement des Exploitants et Exploitations Agricoles (REEA) in 2015. In Togo, the reference population comprised 37,522 households in the selected prefectures, according to the Recensement National de l’Agriculture (RNA) in 2012. The sample size was estimated separately for each country using the finite population correction formula:

Where n is the estimated sample size, N is the reference population size, z is the standard normal deviate corresponding to a 95% confidence level, e is the margin of error set at 5%, and p is the estimated population proportion, set at 0.50 in the absence of a prior population estimate. The sample size was calculated using the SurveyMonkey sample size calculator. Based on these parameters and the respective population sizes, the estimated sample sizes were 384 households for Benin, 383 for Côte d’Ivoire, and 381 for Togo.

Within the selected areas, households were randomly selected from the identified livestock-producing populations, while ensuring coverage of the main livestock production systems and geographical areas included in the study. The final dataset comprised 1,126 livestock-producing households: 388 in Benin, 355 in Togo, and 382 in Côte d’Ivoire.

Table 2 presents the distribution of the sample by country and gender. The sample is relatively balanced across countries, with 34.5% of respondents in Benin, 31.5% in Togo, and 34.0% in Côte d’Ivoire. Male-headed households dominate the sample (87.9%), although the proportion of female-headed households varies significantly across countries, reaching 23.0% in Côte d’Ivoire compared to 8.0% in Benin and 4.8% in Togo.

Two observations had missing values for the experience variable and were therefore omitted from the multivariate regression models, resulting in an analytical sample of 1,124 observations for these models.

3.3 Construction of the climate vulnerability index (CVI)

3.3.1 Main indicators used for constructing CVI

To empirically assess climate vulnerability, we constructed a composite Climate Vulnerability Index (CVI) based on the conceptual framework described above. The index integrates three core dimensions: exposure, sensitivity, and adaptive capacity (Adger, 2006; Edmonds et al., 2020; Opiyo et al., 2014).

Each component is derived from a set of indicators capturing different aspects of livestock production systems and household conditions (Table 3).

The exposure index captures the extent to which households are subject to climatic and environmental stress (Edmonds et al., 2020; Füssel, 2010; Said et al., 2024). It is constructed using indicators related to resource pressure, climatic shocks, and water stress. These indicators are aggregated into a composite exposure index, where higher values indicate greater exposure to climatic stress.

The sensitivity reflects the degree to which livestock systems are affected by climatic stress (Füssel, 2010; Mekonen and Berlie, 2021; Said et al., 2024). It is primarily based on livestock mortality over the past 12 months. Higher levels of mortality increase the sensitivity of households, thereby contributing to higher vulnerability.

The adaptive capacity index captures the ability of households to cope with and respond to climatic shocks (Asfaw et al., 2021; Choden et al., 2020; Tessema and Simane, 2019). It is constructed from a set of variables reflecting access to key resources and strategies, including financial capital, feed supplementation, fodder production, preventive animal health practices, access to veterinary services, human capital, and social capital. These variables are aggregated into a composite index, where higher values indicate greater adaptive capacity.

3.3.2 Aggregation and standardization

The Climate Vulnerability Index is computed by combining the three components, such that vulnerability increases with exposure and sensitivity and decreases with adaptive capacity. Formally, the index is expressed as:

To facilitate interpretation and comparability across observations, the CVI is standardized using a z-score transformation:

This standardization ensures that the index has a mean of zero and a standard deviation of one, allowing for consistent interpretation in the econometric analysis.

3.4 Variables description

3.4.1 Dependent variable

The dependent variable is the standardized Climate Vulnerability Index (ZCVI), constructed as described in Section 3.3.2. This index captures the overall vulnerability of livestock-producing households by combining exposure, sensitivity, and adaptive capacity. Higher values of ZCVI indicate higher levels of vulnerability.

3.4.2 Key explanatory variables

The main explanatory variables relate to livestock production systems and the core components of vulnerability.

Livestock production systems are captured through dummy variables representing sedentary, semi-intensive, and transhumant systems, with sedentary systems used as the reference category.

Exposure and adaptive capacity are measured using the composite indices constructed in Section 3.3. These variables capture the extent of climatic stress and the ability of households to respond to such stress. To test for interaction effects and non-linear relationships, the model also includes an interaction term between exposure and adaptive capacity, as well as the squared term of adaptive capacity.

3.4.3 Control variables

We included controls variables to account for household socio-economic characteristics. These include age and experience of the household head, household size, gender, and formal education level.

Education is represented through dummy variables for primary, secondary, and higher education, with no formal education as the reference category. Country-specific effects are controlled for using dummy variables for Togo and Côte d’Ivoire, with Benin as the reference category.

Table 4 provides a summary of all variables used in the analysis, including their definitions and measurement.

3.5 Econometric strategy

To examine the determinants of climate vulnerability, this study employs Ordinary Least Squares (OLS) regression models. This choice is appropriate given that the dependent variable, the standardized Climate Vulnerability Index (ZCVI), is continuous. The analysis follows a stepwise modeling approach to test the hypotheses and to assess the contribution of different sets of variables.

3.5.1 Baseline model

The baseline model (Equation 1) examines the association between livestock production systems and climate vulnerability, while controlling for household characteristics and country effects:

where X_i represents household-level control variables (age, experience, household size, gender, and education), and C_i denotes country fixed effects.

This model is used to test Hypotheses H1 and H2.

3.5.2 Extended model

To assess the role of exposure and adaptive capacity, the model (Equation 2) is extended as follows:

This specification allows testing Hypothesis H3, focusing on the independent effects of exposure and adaptive capacity.

3.5.3 Full model

To examine interaction and non-linear effects, the full model (Equation 3) includes an interaction term and a quadratic term:

This model is used to test Hypotheses H4 and H5.

3.5.4 Variance analysis

To assess whether vulnerability varies more within production systems than between them, an analysis of variance (ANOVA) is conducted. This analysis is used to test Hypothesis H6 (Equation 4).

where captures differences in mean climate vulnerability across production systems.

In addition, robust standard errors are used in all regressions to account for potential heteroskedasticity. This provides consistent inference when the assumption of constant error variance is violated.

3.6 Diagnostic tests

To assess the validity of the OLS regression models, a series of diagnostic tests were conducted, the results of which are reported in Table 5.

Overall, the results confirm that the main assumptions underlying OLS estimation are reasonably satisfied. Multicollinearity is not a concern, as indicated by low Variance Inflation Factor (VIF) values (mean VIF = 1.38). Similarly, the Breusch–Pagan test does not provide strong evidence of heteroskedasticity (p = 0.063), as the null hypothesis of homoskedasticity is not rejected at the 5% significance level. However, the result is marginal at the 10% level, suggesting that some degree of non-constant variance cannot be ruled out. Robust standard errors are therefore used in all estimations to ensure consistent inference. The normality test suggests that residuals deviate from a normal distribution (p < 0.01). However, given the relatively large sample size (N = 1,126), this does not affect the consistency and asymptotic properties of the OLS estimators (Newey and McFadden, 1994; Wooldridge, 2016). Thus, the significant normality test result is interpreted as a departure from the normality assumption rather than as evidence that the OLS estimates are invalid.

In addition, the analysis of Cook’s distance indicates the absence of influential outliers, as all values remain well below conventional thresholds. Finally, the Ramsey RESET test does not reject the null hypothesis of correct model specification at the 5% level (p = 0.085). However, the result is marginal at the 10% level. To assess theoretically plausible non-linear and conditional relationships, the full model includes the exposure × adaptive capacity interaction and the squared adaptive capacity term. Neither term is statistically significant (Table 6).

Taken together, these diagnostic tests support the reliability of the econometric results presented in this study.

4 Findings

4.1 Effect of livestock production systems on climate vulnerability

Table 7 presents the results of the OLS regression estimating the effect of livestock production systems and country context on climate vulnerability, while Table 8 reports the pairwise comparison between production systems.

The results indicate that livestock production systems are significantly associated with climate vulnerability. Compared to sedentary systems (reference category), both transhumant and semi-intensive systems exhibit higher vulnerability levels. Specifically, transhumant systems increase vulnerability by 0.366 units (p < 0.01), while semi-intensive systems show a larger effect (β = 0.471, p < 0.01), supporting Hypothesis H1.

However, the pairwise comparison between semi-intensive and transhumant systems (Table 5) reveals no statistically significant difference (β = 0.105, p = 0.323), suggesting that these systems face comparable levels of vulnerability despite their structural differences.

Country-level effects are also significant. Households in Togo and Côte d’Ivoire exhibit higher vulnerability compared to those in Benin, with coefficients of 0.319 and 0.470, respectively, (p < 0.01). This supports Hypothesis H2, highlighting the role of contextual factors.

Among control variables, only household size is significant (β = −0.032, p < 0.01), indicating that larger households are less vulnerable. Other socio-demographic variables do not show significant effects, suggesting that structural and contextual factors dominate individual characteristics.

4.2 Role of exposure and adaptive capacity in climate vulnerability

Table 6 presents the results of the full model including exposure, adaptive capacity, their interaction, and the quadratic term capturing potential non-linear effects.

4.2.1 Independent effects of exposure and adaptive capacity

Exposure and adaptive capacity emerge as the main determinants of vulnerability. Exposure significantly increases vulnerability (β = 2.646, p < 0.01), while adaptive capacity has a strong negative effect (β = −2.968, p < 0.01) (Table 8).

The inclusion of these variables substantially improves the explanatory power of the model, with R² increasing from 0.099 in the baseline model to 0.557. This indicates that vulnerability is primarily driven by the balance between environmental pressure and adaptive responses rather than by structural characteristics alone.

These findings support Hypothesis H3, showing that vulnerability is not solely determined by exposure, as adaptive capacity plays a compensatory role.

4.2.2 Interaction between exposure and adaptive capacity

The interaction term between exposure and adaptive capacity is not statistically significant (β = 0.817, p = 0.203) (Table 8). This provides no empirical support for Hypothesis H4, indicating that adaptive capacity does not significantly modify the effect of exposure on vulnerability.

Both variables remain significant with expected signs, suggesting that their effects are additive rather than interactive.

4.2.3 Non-linear effects of adaptive capacity

As reported in Table 8, the quadratic term of adaptive capacity (ADAPT²) is not statistically significant (β = −0.073, p = 0.878), indicating no evidence of non-linear effects. This leads to the rejection of Hypothesis H5.

This result suggests that the effect of adaptive capacity on vulnerability is linear and constant across different levels of adaptation. In other words, each additional increase in adaptive capacity contributes proportionally to reducing vulnerability, without evidence of threshold or diminishing returns.

4.3 Heterogeneity within livestock production systems

The results provide strong evidence supporting Hypothesis H6, which posits that vulnerability varies more within production systems than between them.

As shown in Table 6, the coefficients associated with livestock production systems are either not statistically significant or relatively small in magnitude. In particular, the coefficient for transhumant systems is not significant (β = −0.056, p = 0.442), while the effect of semi-intensive systems, although significant, remains modest (β = 0.124, p = 0.020).

To further examine this hypothesis, an analysis of variance (ANOVA) was conducted to compare the distribution of vulnerability across production systems (Table 9). The results indicate that although differences between systems are statistically significant (F = 12.60, p < 0.01), the variance within systems is substantially larger than the variance between systems. Specifically, the within-group sum of squares (SS = 1100.31) far exceeds the between-group sum of squares (SS = 24.69), indicating that most of the variation in vulnerability occurs within production systems rather than across them.

This result is further supported by the relatively small differences in mean vulnerability across systems, with average ZCVI values of −0.118 for sedentary systems, 0.193 for semi-intensive systems, and 0.164 for transhumant systems.

Our findings demonstrate that vulnerability is primarily driven by intra-system heterogeneity rather than by structural differences between production systems. Households operating within the same system can experience markedly different levels of vulnerability depending on their exposure to climatic stress and their capacity to adapt.

These results highlight the importance of moving beyond system-based classifications and focusing on the underlying socio-economic and environmental conditions shaping vulnerability. They also suggest that policy interventions should prioritize strengthening adaptive capacity and reducing exposure, rather than promoting specific production systems.

The results consistently show that climate vulnerability is not an intrinsic property of livestock production systems. Rather, it emerges from the combined effects of exposure and adaptive capacity, with substantial heterogeneity observed within systems.

5 Discussion

This study examined whether climate vulnerability in livestock systems is primarily shaped by production system type or by broader socio-economic and environmental conditions. The results suggest that vulnerability is not adequately characterized by production system type alone. It is more closely associated with differences in exposure and adaptive capacity, consistent with process-based vulnerability frameworks (Asfaw et al., 2021; Pandey et al., 2017; Tessema and Simane, 2019; Tofu et al., 2023). This finding shifts attention from broad production-system categories toward household-level conditions.

First, the weak and non-robust effect of production system type suggests that system classifications alone are insufficient to explain vulnerability outcomes. Although transhumant and semi-intensive systems initially appear more vulnerable, these differences vanish once exposure and adaptive capacity are controlled for. Similar patterns have been observed in pastoral and mixed systems, where system labels provide contextual information but do not correspond systematically to vulnerability levels (Alary et al., 2022; Belay and Lebeza, 2024; Hammami and Ferchichi, 2023). This finding questions the assumption that mobility or intensification necessarily correspond to different levels of resilience. It instead highlights the importance of underlying household and environmental conditions (Alary et al., 2022; Belay and Lebeza, 2024; McDermott et al., 2010). Production-system categories may partly reflect differences in exposure and resource access (Madaki et al., 2025; Montcho et al., 2022; Tofu et al., 2025). This interpretation is consistent with evidence showing that access to productive resources and other forms of household capital can shape vulnerability independently of livelihood or production-system characteristics (Qin et al., 2022). Once these conditions are controlled for, the observed differences between systems become substantially weaker.

Second, the strong and opposing effects of exposure and adaptive capacity, and the large rise in explanatory power when they are included, reinforce process-based views of vulnerability as a function of exposure, sensitivity and adaptive capacity (Asfaw et al., 2021; Pandey et al., 2017; Tessema and Simane, 2019; Tofu et al., 2023). Studies in pastoral, mixed crop–livestock and smallholder systems similarly show that higher exposure and sensitivity are associated with greater vulnerability, whereas stronger adaptive capacity (assets, infrastructure, social networks, information) is associated with lower vulnerability (Karimi et al., 2018; Liu et al., 2024; Tofu et al., 2025; Weng et al., 2023; Xu et al., 2020). Evidence from Ethiopia, China and semi-arid pastoral regions indicates that areas or groups with similar production systems can exhibit very different vulnerability levels depending on access to water, markets, services and livelihood diversification (Asfaw et al., 2021; Liu et al., 2024; Qin et al., 2022; Tessema and Simane, 2019; Tofu et al., 2025; Weng et al., 2023; Xu et al., 2020). This result is further supported by recent empirical evidence showing that access to resources, institutional support, and socio-economic conditions are associated with differences in adaptation and resilience outcomes in agricultural systems (Adéchian and Baco, 2025; Akowedaho et al., 2022; Liu et al., 2024; Makate et al., 2024). These findings suggest that households within the same production system may experience different levels of vulnerability when their access to resources and services differs.

Third, the absence of a significant interaction between exposure and adaptive capacity, and the lack of non-linear effects, suggest that in this case adaptive capacity and exposure are associated with vulnerability in a largely additive pattern. While some studies report threshold-like gains from certain assets or services (Asfaw et al., 2021; Belay and Lebeza, 2024; Tessema and Simane, 2019), many operationalize vulnerability with approximately linear contributions of exposure, sensitivity and adaptive capacity. These studies also find that higher adaptive capacity is associated with lower vulnerability without clear evidence of saturation points (Asfaw et al., 2021; Pandey et al., 2017). The present results are therefore consistent with a broad body of index-based work in which vulnerability is positively associated with climatic stress and negatively associated with adaptive capacity (Asfaw et al., 2021; Pandey et al., 2017; Tofu et al., 2025). The absence of significant interaction and non-linear terms is also informative. It suggests that no specific threshold or conditional effect is evident within the range observed in this study.

Fourth, the very high share of variance explained within, rather than between, production systems (over 97%), underscores the centrality of household- and locality-specific factors. Comparable micro-level studies show substantial heterogeneity in vulnerability among households facing similar agro-ecological conditions, driven by differences in livelihood strategies, human and financial capital, social networks, and access to infrastructure and institutions (Tessema and Simane, 2019; Xu et al., 2020). This supports arguments that vulnerability is highly context-specific and cautions against relying on broad production system typologies alone to identify at-risk groups (Alary et al., 2022; Belay and Lebeza, 2024; McDermott et al., 2010). This strong intra-system heterogeneity aligns with evidence emphasizing the multidimensional nature of resilience, including the role of social capital, knowledge, and local practices (Dado Tossou et al., 2026; Houanyé et al., 2025; Xu et al., 2020). The magnitude of this within-system variation is an important contribution of the study. It shows that substantial differences in vulnerability remain among households sharing the same production-system category.

These results contribute to ongoing debates on climate vulnerability in agricultural and pastoral systems by empirically reinforcing approaches that emphasize access to resources, institutions and livelihood capitals as factors associated with differences in vulnerability (Asfaw et al., 2021; Pandey et al., 2017; Tessema and Simane, 2019; Tofu et al., 2025). The finding that semi-intensive systems can remain highly vulnerable aligns with evidence that intensification may introduce new forms of sensitivity—through dependence on external inputs, markets and credit—if not matched by investments in adaptive capacity and risk management (Demem, 2023; McDermott et al., 2010). However, the cross-sectional design does not allow us to conclude that intensification itself increases vulnerability.

From a policy perspective, the findings suggest the relevance of moving beyond system-focused prescriptions. Interventions could instead strengthen adaptive capacity and reduce exposure. This is consistent with recommendations from pastoral and smallholder contexts, which highlight the importance of improving access to water, veterinary and extension services, financial instruments, early warning systems, and livelihood diversification options tailored to local conditions (Asfaw et al., 2021; Karimi et al., 2018; Pandey et al., 2017; Tofu et al., 2025). The strong within-system variation further suggests that interventions should consider differences in household resources, exposure and adaptive capacity. Policies that promote particular production systems without simultaneously addressing underlying socio-economic and environmental drivers of vulnerability are unlikely to yield durable reductions in climate risk.

This study has some limitations that should be considered when interpreting the findings. First, the cross-sectional design does not allow us to establish temporal relationships between exposure, adaptive capacity, and vulnerability. Reverse causality may therefore be possible, and the observed associations should not be interpreted as causal effects. Second, some information refers to the previous 12 months. This reference period may create some overlap between reported exposure, adaptive responses, and current vulnerability conditions. This should be considered when interpreting the relationships between these variables. Third, although the models include a range of household, production-system, exposure, adaptive-capacity, and country-level variables, some relevant factors may not be fully captured in the available data. The possibility of omitted variable bias therefore remains. Future studies using longitudinal data and repeated observations could provide further evidence on the temporal relationships between climate exposure, adaptive capacity, and vulnerability.

6 Conclusion

This study set out to examine whether climate vulnerability in livestock systems is primarily determined by production system characteristics or by underlying socio-economic and environmental conditions. The results provide a consistent and nuanced answer: vulnerability is not an intrinsic feature of livestock production systems, but rather the outcome of interacting processes related to exposure and adaptive capacity.

First, the findings show that livestock production systems are associated with vulnerability outcomes, but this relationship remains limited when considered in isolation. While transhumant and semi-intensive systems initially appear more vulnerable than sedentary systems, the absence of significant differences between these systems suggests that vulnerability is not strictly hierarchical across production models. This result challenges conventional assumptions that mobility or intensification alone determine resilience, and instead points to the importance of underlying mechanisms shaping vulnerability within each system.

Second, the analysis highlights the central role of exposure and adaptive capacity. The strong and opposite effects of these variables indicate that vulnerability emerges from the balance between environmental pressures and households’ ability to respond. The substantial increase in explanatory power when these variables are introduced (R² rising from 0.099 to 0.557) reinforces the idea that vulnerability is primarily driven by dynamic processes rather than by structural characteristics. This finding aligns with conceptual frameworks that view vulnerability as a function of exposure, sensitivity, and adaptive capacity, rather than as a fixed attribute of production systems.

Importantly, the results do not support the hypothesis that adaptive capacity moderates the effect of exposure. The absence of a significant interaction effect suggests that adaptation operates through direct and independent channels, rather than by conditioning the impact of climatic stress. Similarly, the lack of evidence for non-linear effects indicates that adaptive capacity contributes to reducing vulnerability in a consistent and proportional manner, without threshold or saturation effects. Together, these findings point to a relatively simple but robust mechanism in which exposure and adaptation act as additive and opposing forces.

Third, the analysis reveals substantial heterogeneity within production systems. The ANOVA results show that more than 97% of the variation in vulnerability occurs within systems rather than between them. This finding provides strong empirical support for the argument that vulnerability is context-specific and shaped by household-level conditions. It also reinforces the limitations of system-based classifications, which may obscure important variations in access to resources, institutional support, and adaptive strategies.

Taken together, these results contribute to ongoing debates on climate vulnerability in agricultural systems. They provide empirical support for approaches that emphasize the role of access to resources and institutional conditions (Eakin and Luers, 2006; Ribot, 2013), while challenging perspectives that attribute vulnerability primarily to production system types. In particular, the finding that semi-intensive systems remain vulnerable despite their higher level of intensification suggests that dependence on external inputs and market conditions may introduce new forms of sensitivity.

From a policy perspective, these findings have important implications. Interventions aimed at reducing vulnerability should prioritize strengthening adaptive capacity—such as access to financial resources, veterinary services, and information—while also addressing exposure to climatic stress. Policies that focus solely on promoting specific production systems are unlikely to be effective if they do not simultaneously address the underlying drivers of vulnerability.

References

• 1

  AboyeA. B.KinsellaJ.MegaT. L. (2023). Farm households’ adaptive strategies in response to climate change in lowlands of southern Ethiopia. Int. J. Clim. Chang. Strateg. Manag.15, 579–598. doi: 10.1108/ijccsm-05-2023-0064

  • CrossRef
  • Google Scholar

• 2

  AdéchianS. A.BacoM. N. (2025). How does land tenure security contribute to the sustainable development of rural households? Evidence from Benin (West Africa). Front. Sustain. Food Syst.9:1621537. doi: 10.3389/fsufs.2025.1621537

  • CrossRef
  • Google Scholar

• 3

  AdgerW. N. (2006). Vulnerability. Glob. Environ. Change16, 268–281. doi: 10.1016/j.gloenvcha.2006.02.006

  • CrossRef
  • Google Scholar

• 4

  AhmadD.AfzalM. (2021). Impact of climate change on pastoralists’ resilience and sustainable mitigation in Punjab, Pakistan. Environ. Dev. Sustain.23, 11406–11426. doi: 10.1007/s10668-020-01119-9

  • CrossRef
  • Google Scholar

• 5

  AkowedahoB. D.Guinin AssoI.O’heixB. C. P.AdéchianS. A.BacoM. N. (2022). Access to land for agricultural entrepreneurial activities in the context of sustainable food production in Borgou, according to land law in Benin. Land11:1381. doi: 10.3390/land11091381

  • CrossRef
  • Google Scholar

• 6

  AlaryV.CaulfieldM. E.AmsidderL.JuanesX.BoujenaneI.SraïriT. M.et al. (2022). Heterogeneity of resilience of livelihood strategies in pastoral and agropastoral farming systems of rural semi-arid to arid areas in Morocco. Front. Sustain. Food Syst.5:723994. doi: 10.3389/fsufs.2021.723994

  • CrossRef
  • Google Scholar

• 7

  AsfawA.BantiderA.SimaneB.HassenA. (2021). Smallholder farmers’ livelihood vulnerability to climate change-induced hazards: agroecology-based comparative analysis in northcentral Ethiopia (Woleka sub-basin). Heliyon7:e06761. doi: 10.1016/j.heliyon.2021.e06761,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 8

  AssaniA. S.YarouA. K.DedehouN. V. F. G.WorogoH. S.BacoM. N.HouinatoM.et al. (2024). Towards indigenous community-based adaptation to climate change: a typological analysis of tree-livestock integration in smallholding systems in dryland areas of Benin (West-Africa). Agrofor. Syst.98, 197–211. doi: 10.1007/s10457-023-00899-z

  • CrossRef
  • Google Scholar

• 9

  AyantundeA. A.TurnerM. D.KalilouA. (2015). Participatory analysis of vulnerability to drought in three agro-pastoral communities in the west African Sahel. Pastoralism5:13. doi: 10.1186/s13570-015-0033-x

  • CrossRef
  • Google Scholar

• 10

  AzalouM.AssaniS. A.AssogbaB. G. C.IdrissouY.AlabiC. D.YabiJ. A.et al. (2023). Dynamics of transhumant livestock systems in west African coastal countries: a review. J. Livestock Sci.14. doi: 10.33259/JLivestSci.2023.109-121

  • CrossRef
  • Google Scholar

• 11

  AzineP. C.MugumaarhahamaY.MutweduV. B.MondoJ. M.ChumaG. B.BuchekabiriA.et al. (2025). Assessing smallholder farmers’ vulnerability to climate change and coping strategies in south Kivu Province, eastern Democratic Republic of Congo. Environ. Syst. Res.14:2. doi: 10.1186/s40068-025-00393-8

  • CrossRef
  • Google Scholar

• 12

  BahtaY. T.NyakiS. A. (2024). Livelihood vulnerability from drought among smallholder livestock farmers in South Africa. Hydrology11:11090137. doi: 10.3390/hydrology11090137

  • CrossRef
  • Google Scholar

• 13

  BahtaY.NyakiS.MaréF. (2025). South African commercial livestock farmers’ adaptation and coping strategies for agricultural drought. Open Agric.10:454. doi: 10.1515/opag-2025-0454

  • CrossRef
  • Google Scholar

• 14

  BashiruH.OseniS. (2025). Simplified climate change adaptation strategies for livestock development in low-and middle-income countries. Front. Sustain. Food Syst.9:1566194. doi: 10.3389/fsufs.2025.1566194

  • CrossRef
  • Google Scholar

• 15

  BelayT.LebezaT. M. (2024). Livelihood vulnerability of rural households to climate variability and change: an agroecological system-based approach in northwestern Ethiopia. Heliyon10:e40570. doi: 10.1016/j.heliyon.2024.e40570,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 16

  BerheM.HoagD. L.TesfayG.TadesseT.OnikiS.KagatsumeM.et al. (2017). The effects of adaptation to climate change on income of households in rural Ethiopia. Pastoralism7:12. doi: 10.1186/s13570-017-0084-2

  • CrossRef
  • Google Scholar

• 17

  BlackW.BabinB. J. (2019). “Multivariate data analysis: its approach, evolution, and impact,” in The Great Facilitator, eds. BabinB. J.SarstedtM. (Cham: Springer International Publishing), 121–130.

  • Google Scholar

• 18

  ChodenK.KeenanR. J.NitschkeC. R. (2020). An approach for assessing adaptive capacity to climate change in resource dependent communities in the Nikachu watershed, Bhutan. Ecol. Indic.114:106293. doi: 10.1016/j.ecolind.2020.106293

  • CrossRef
  • Google Scholar

• 19

  ComoéH.SiegristM. (2015). Relevant drivers of farmers’ decision behavior regarding their adaptation to climate change: a case study of two regions in Côte d’Ivoire. Mitig. Adapt. Strateg. Glob. Change20, 179–199. doi: 10.1007/s11027-013-9486-7

  • CrossRef
  • Google Scholar

• 20

  Dado TossouB.AdéchianS. A.EgahJ.Moumouni-MoussaI.BacoM. N. (2026). Local actors’ perceptions and adaptation strategies in response to agricultural policies on neglected and underutilized species in Benin. Front. Sustain. Food Syst.10:1760706. doi: 10.3389/fsufs.2026.1760706

  • CrossRef
  • Google Scholar

• 21

  de GlanvilleW. A.DavisA.AllanK. J.BuzaJ.ClaxtonJ. R.CrumpJ. A.et al. (2020). Classification and characterisation of livestock production systems in northern Tanzania. PLoS One15:e0229478. doi: 10.1371/journal.pone.0229478,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 22

  DememM. S. (2023). Impact and adaptation of climate variability and change on small-holders and agriculture in Ethiopia: a review. Heliyon9:e18972. doi: 10.1016/j.heliyon.2023.e18972,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 23

  DescheemaekerK.OostingS. J.Homann-Kee TuiS.MasikatiP.FalconnierG. N.GillerK. E. (2016). Climate change adaptation and mitigation in smallholder crop–livestock systems in sub-Saharan Africa: a call for integrated impact assessments. Reg. Environ. Chang.16, 2331–2343. doi: 10.1007/s10113-016-0957-8

  • CrossRef
  • Google Scholar

• 24

  DevendraC.MortonJ.RischkowskyB.ThomasD. (2005). “Livestock systems,” in Livestock and Wealth Creation: Improving the Husbandry of Animals Kept by Resource-Poor People in Developing Countries, eds. OwenE.KitalyiA.JayasuriyaN.SmithT. (Nottingham: Nottingham University Press), 29–52.

  • Google Scholar

• 25

  DjakpaM. G.EgahJ.DeglaP. (2024). Adaptation to climate change among transhumant herders: evidences from northern Benin (West Africa). Front. Sustain. Food Syst.8:1426771. doi: 10.3389/fsufs.2024.1426771

  • CrossRef
  • Google Scholar

• 26

  EakinH.LuersA. L. (2006). Assessing the vulnerability of social-environmental systems. Annu. Rev. Environ. Resour.31, 365–394. doi: 10.1146/annurev.energy.30.050504.144352

  • CrossRef
  • Google Scholar

• 27

  EdmondsH. K.LovellJ. E.LovellC. A. K. (2020). A new composite climate change vulnerability index. Ecol. Indic.117:106529. doi: 10.1016/j.ecolind.2020.106529

  • CrossRef
  • Google Scholar

• 28

  FieldA. (2024). Discovering Statistics using IBM SPSS Statistics. London: Sage publications limited.

  • Google Scholar

• 29

  FüsselH.-M. (2010). How inequitable is the global distribution of responsibility, capability, and vulnerability to climate change: a comprehensive indicator-based assessment. Glob. Environ. Change20, 597–611. doi: 10.1016/j.gloenvcha.2010.07.009

  • CrossRef
  • Google Scholar

• 30

  GatewS.GuyoN. (2024). Livelihood vulnerability of Borana pastoralists to climate change and variability in southern Ethiopia. Int. J. Clim. Chang. Strateg. Manag.16, 157–176. doi: 10.1108/IJCCSM-06-2023-0077

  • CrossRef
  • Google Scholar

• 31

  GautierD.LocatelliB.CorniauxC.AlaryV. (2016). Global changes, livestock and vulnerability: the social construction of markets as an adaptive strategy. Geogr. J.182, 153–164. doi: 10.1111/geoj.12115

  • CrossRef
  • Google Scholar

• 32

  GuyeM.LegesseA.MohammedY. (2023). Pastoralists under threat continuum: quantifying vulnerabilities of pastoralists to climate variability in southern Ethiopia. GeoJournal88, 1785–1806. doi: 10.1007/s10708-022-10710-8,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 33

  HabteM.EshetuM.MaryoM.AndualemD.LegesseA. (2022). Effects of climate variability on livestock productivity and pastoralists perception: the case of drought resilience in southeastern Ethiopia. Vet. Anim. Sci.16:100240. doi: 10.1016/j.vas.2022.100240,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 34

  HammamiO.FerchichiA. (2023). Vulnerability of pastoral ecosystems in northwestern Tunisia to climate change. Land Degrad. Dev.34, 4680–4689. doi: 10.1002/ldr.4801

  • CrossRef
  • Google Scholar

• 35

  HouanyéA. K.AdéchianS. A.BacoM. N.BadjanaH. M.AmoussouE. (2025). Integrating local knowledge and community practices for flood resilience in the Volta Basin. Sustainability17:8087. doi: 10.3390/su17178087

  • CrossRef
  • Google Scholar

• 36

  KapruwanR.SakshamA. K.BhadoriyaV. S.KumarC.GoyalY.PandeyR. (2024). Household livelihood resilience of pastoralists and smallholders to climate change in Western Himalaya, India. Heliyon.10:e24133. doi: 10.1016/j.heliyon.2024.e24133

  • CrossRef
  • Google Scholar

• 37

  KarimiV.KaramiE.KeshavarzM. (2018). Vulnerability and adaptation of livestock producers to climate variability and change. Rangel. Ecol. Manag.71, 175–184. doi: 10.1016/j.rama.2017.09.006

  • CrossRef
  • Google Scholar

• 38

  KlinckL.AyisiK. K.IsselsteinJ. (2022). Drought-induced challenges and different responses by smallholder and semicommercial livestock farmers in semiarid Limpopo, South Africa—an indicator-based assessment. Sustainability14:14148796. doi: 10.3390/su14148796

  • CrossRef
  • Google Scholar

• 39

  KoutchoroA. M.AdéchianS. A.HouessouL. G.YaoitchaA. S.ToureY.Moumouni-MoussaI. (2025). Pastoral transitions in West Africa: analysis of factors influencing herders’ sedentarization in northern Benin. Pastoralism15:15199. doi: 10.3389/past.2025.15199

  • CrossRef
  • Google Scholar

• 40

  LiuX.LiuS.WangR.TangH.ZhangF.JiaL.et al. (2024). Vulnerability of farmer households to climate change in rocky desertification areas—a case study of Guizhou Province. Land13:582. doi: 10.3390/land13050582

  • CrossRef
  • Google Scholar

• 41

  MadakiM. Y.BarnabasB.ShehuA.UllahA.BavorovaM. (2025). Institutional services towards climate action: a case of climate change adaptation of agro-pastoralists in the drylands of Nigeria. Clim. Pol.25, 1222–1236. doi: 10.1080/14693062.2024.2447487

  • CrossRef
  • Google Scholar

• 42

  MakateC.AngelsenA.HoldenS. T.WestengenO. T. (2024). Smallholder access to purchased seeds in the presence of pervasive market imperfections and rainfall shocks: panel data evidence from Malawi and Ethiopia. Agric. Econ.12:33. doi: 10.1186/s40100-024-00325-7

  • CrossRef
  • Google Scholar

• 43

  McDermottJ. J.StaalS. J.FreemanH. A.HerreroM.Van de SteegJ. A. (2010). Sustaining intensification of smallholder livestock systems in the tropics. Livest. Sci.130, 95–109. doi: 10.1016/j.livsci.2010.02.014

  • CrossRef
  • Google Scholar

• 44

  MekonenA. A.BerlieA. B. (2021). Rural households’ livelihood vulnerability to climate variability and extremes: a livelihood zone-based approach in the northeastern highlands of Ethiopia. Ecol. Process.10:55. doi: 10.1186/s13717-021-00313-5

  • CrossRef
  • Google Scholar

• 45

  MontchoM.PadonouE. A.MontchoM.MutuaM. N.SinsinB. (2022). Perception and adaptation strategies of dairy farmers towards climate variability and change in West Africa. Clim. Chang.170:38. doi: 10.1007/s10584-022-03311-4

  • CrossRef
  • Google Scholar

• 46

  MunyakaJ.AbdemS. A. E.GallayO.ChenalJ.LolemtumJ.AdierM. B.et al. (2025). Integrating socioeconomic and community-based strategies for drought resilience in west Pokot, Kenya. Climate.13:148. doi: 10.3390/cli13070148

  • CrossRef
  • Google Scholar

• 47

  NeweyW. K.McFaddenD. (1994). “Chapter 36 large sample estimation and hypothesis testing,” in Handbook of Econometrics, (Amsterdam: Elsevier), 2111–2245.

  • Google Scholar

• 48

  NgongoloK.GayoL. (2025). Climate change impacts and mitigation strategies in sub-Saharan Africa’s livestock production sector. A brief review. Pastoralism Res. Policy Pract.15, 1–7. doi: 10.3389/past.2025.14225

  • CrossRef
  • Google Scholar

• 49

  NtombelaK.AngulaM.SamuelsI.CupidoC.SwartsM.Menjono-KatjizeuE.et al. (2024). Pastoral coping and adaptive management strategies to climate change in communal areas in Namibia and South Africa. Pastoralism Res. Policy Pract.14:13548. doi: 10.3389/past.2024.13548

  • CrossRef
  • Google Scholar

• 50

  OgoudouC.AdéchianS. A.EgahJ.BacoM. N.Assani SeidouA.WorogoH. S. S.et al. (2025). Governing digital innovation in livestock systems: institutional gaps and coordination challenges in Benin. Pastoralism15:15108. doi: 10.3389/past.2025.15108

  • CrossRef
  • Google Scholar

• 51

  OpiyoF. E.WasongaO. V.NyangitoM. M. (2014). Measuring household vulnerability to climate-induced stresses in pastoral rangelands of Kenya: implications for resilience programming. Pastoralism4:10. doi: 10.1186/s13570-014-0010-9

  • CrossRef
  • Google Scholar

• 52

  PandeyR.JhaS. K.AlataloJ. M.ArchieK. M.GuptaA. K. (2017). Sustainable livelihood framework-based indicators for assessing climate change vulnerability and adaptation for Himalayan communities. Ecol. Indic.79, 338–346. doi: 10.1016/j.ecolind.2017.03.047

  • CrossRef
  • Google Scholar

• 53

  QinZ.HailiX.XiaoL.LuweiD.BojieW.CuiF.et al. (2022). Livelihood vulnerability of pastoral households in the semiarid grasslands of northern China: measurement and determinants. Ecol. Indic.140:109020. doi: 10.1016/j.ecolind.2022.109020

  • CrossRef
  • Google Scholar

• 54

  RibotJ. (2013). “Vulnerability does not just fall from the sky: toward multi-scale pro-poor climate policy,” in Handbook on Climate Change and Human Security, (Cheltenham: Edward Elgar Publishing), 164–199.

  • Google Scholar

• 55

  SaidF. F.KamisM. M.Abdul MauludK. N.JauhariA.MohammedN. (2024). Assessing climate change vulnerability in coastal communities: a composite vulnerability index approach in Kuala Gula, Malaysia. Environ. Dev. Sustain.26, 32409–32444. doi: 10.1007/s10668-024-05533-1

  • CrossRef
  • Google Scholar

• 56

  SegnonA. C.TotinE.ZougmoréR.LokossouJ.Thompson-HallM.OforiB. O.et al. (2021). Differential household vulnerability to climatic and non-climatic stressors in semi-arid areas of Mali, West Africa. Clim. Dev.13, 697–712. doi: 10.1080/17565529.2020.1855097

  • CrossRef
  • Google Scholar

• 57

  TessemaI.SimaneB. (2019). Vulnerability analysis of smallholder farmers to climate variability and change: an agro-ecological system-based approach in the fincha’a sub-basin of the upper Blue Nile Basin of Ethiopia. Ecol. Process.8:5. doi: 10.1186/s13717-019-0159-7

  • CrossRef
  • Google Scholar

• 58

  TofuD. A.DilbatoT.FanaC.DirbabaN. B.TessoG. (2025). Analysis of vulnerability, its drivers, and strategies applied towards reducing the pastoral and agro-pastoral livelihood vulnerability to climatic shocks. Sci. Rep.15:2567. doi: 10.1038/s41598-024-79165-w,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 59

  TofuD. A.FanaC.DilbatoT.DirbabaN. B.TessoG. (2023). Pastoralists’ and agro-pastoralists’ livelihood resilience to climate change-induced risks in the Borana zone, South Ethiopia: using resilience index measurement approach. Pastoralism13:4. doi: 10.1186/s13570-022-00263-3

  • CrossRef
  • Google Scholar

• 60

  WengC.BaiY.ChenB.HuY.ShuJ.ChenQ.et al. (2023). Assessing the vulnerability to climate change of a semi-arid pastoral social–ecological system: a case study in Hulunbuir, China. Ecol. Inform.76:102139. doi: 10.1016/j.ecoinf.2023.102139

  • CrossRef
  • Google Scholar

• 61

  WongnaaC. A.SeyramA. A.BabuS. (2024). A systematic review of climate change impacts, adaptation strategies, and policy development in West Africa. Reg. Sustain.5:100137. doi: 10.1016/j.regsus.2024.100137

  • CrossRef
  • Google Scholar

• 62

  WooldridgeJ. M. (2016). Introductory Econometrics a modern Approach. Cincinnati: South-Western cengage learning.

  • Google Scholar

• 63

  XuX.WangL.SunM.FuC.BaiY.LiC.et al. (2020). Climate change vulnerability assessment for smallholder farmers in China: an extended framework. J. Environ. Manag.276:111315. doi: 10.1016/j.jenvman.2020.111315,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 64

  ZougmoréR.ParteyS.OuédraogoM.OmitoyinB.ThomasT.AyantundeA.et al. (2016). Toward climate-smart agriculture in West Africa: a review of climate change impacts, adaptation strategies and policy developments for the livestock, fishery and crop production sectors. Agric. Food Secur.5, 1–16. doi: 10.1186/s40066-016-0075-3,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar