Introduction
Agriculture plays a pivotal role in the global economy, particularly for approximately 75% of the world’s poor population living in rural areas and relying primarily on this sector (World Bank, 2024). Agricultural growth is 2–3 times more effective in reducing poverty than equivalent growth in other sectors, with the greatest impacts on the poorest segments of society (World Bank, 2024). In Vietnam, agriculture remains a cornerstone of the Mekong Delta’s economy, contributing significantly to national food security and the livelihoods of millions of rural households (FAO, 2022). Amid escalating climate change and socioeconomic development, agricultural productivity depends not only on natural conditions but also on the effective mobilization and suastainable utilization of livelihood capitals (IPCC, 2023); (Abebaw, 2025).
The Sustainable Livelihoods Framework (SLF), developed by the UK Department for International Development (DFID), has become a key analytical tool in academic research and policy development for rural areas (Scoones, 1998; Carney, 1998). Although the SLF has been widely adopted for assessing rural vulnerabilities, including in climate change contexts, it has faced several criticisms. These include its limited attention to power dynamics, social inequalities (e.g., gender, class, and ethnicity), historical and political processes, and the dynamism of capital transformations over time (Hickey and du Toit, 2007; Morse and McNamara, 2013). In the Mekong Delta, where livelihoods are shaped by complex interactions between state policies, upstream dam developments, and market forces, the SLF may overlook adverse incorporation and structural exclusions that exacerbate vulnerability (Hickey and Bracking, 2005; Small, 2007). Despite these limitations, the SLF remains valuable for its people-centered, multidimensional approach to capitals, which aligns well with the need to disaggregate climate impacts on interconnected assets in this study. By applying MANOVA to quantify interrelationships among capitals, this research addresses some methodological gaps in prior SLF applications, which often rely on qualitative or index-based assessments without robust multivariate testing. This approach is widely applied to assess rural communities’ production capacity, adaptation, and resilience to shocks, including climate change (Ellis, 2000). Recent international studies highlight how livelihood opportunities, assets, and strategies shape pathways for sustainable agricultural transformation, particularly when innovations maximize access to natural capital and create new opportunities for farm households (Abebaw, 2025).
Thanh Phu commune (formerly in Ben Tre Province) is located in the coastal Mekong Delta region, between the Ham Luong and Co Chien rivers within the Tien river system. Covering a natural area of 425.7 km2 with a population density of about 300 people/km2, the commune features an intricate network of canals, lush coconut groves, and diverse farming systems adapted to three ecological sub-zones: freshwater areas (combining coconut cultivation with cacao, and freshwater shrimp and fish farming); brackish water zones (shifting from traditional rice-shrimp to large-scale industrial shrimp farming); and saline areas (integrating mangrove forests with shrimp, clam, and blood cockle aquaculture). This diversity underscores the need to examine the roles and contributions of various livelihood capitals in production activities. As a coastal commune, Thanh Phu has high rates of poor and near-poor households (many communes exceeding 15–20%) (Ben Tre DOST, 2023). Local livelihoods depend mainly on agriculture and aquaculture but are highly vulnerable to climate change, salinity intrusion, and agricultural price fluctuations. Most poor and near-poor households engage in mixed farming, including rice, vegetables, fruit trees (e.g., coconut and green-skinned pomelo), livestock (cattle, goats, pigs, poultry), and aquaculture (e.g., giant freshwater shrimp intercropped with rice). Specialized models such as model coconut gardens, beekeeping, ornamental leaf plants, and mushroom cultivation also enhance livelihood diversification. However, major constraints include limited or no access to farmland, unstable employment, inadequate production skills, and high vulnerability to natural disasters, diseases, and market volatility. Many households also lack basic social services, such as sturdy housing, clean water, sanitation, and health insurance. Poverty alleviation programs in the commune focus on agricultural and fisheries extension with climate-adaptive training, value chain development (e.g., coconut, pomelo, livestock), preferential loans for production expansion, farmer group formation, cooperatives, and scaling successful models to boost incomes and achieve sustainable poverty reduction.
Climate change severely impacts livelihood capitals in the coastal Mekong Delta, including Thanh Phu commune, reducing agricultural productivity and increasing community vulnerability (IPCC, 2023; Abebaw, 2025; Le and Nguyen, 2025; Renaud et al., 2015). Salinity intrusion and sea-level rise lead to land loss, water pollution, and declining soil productivity, directly affecting aquaculture and fruit cultivation (e.g., coconut and pomelo), with salinity levels of 4% penetrating 60–70 km inland in Tra Vinh (Le and Nguyen, 2025). Human capital is weakened by health declines from extreme weather and limited adaptive training (Huong et al., 2019). Physical capital suffers damage from erosion, flooding, and infrastructure collapse. Financial capital diminishes due to crop failures, reduced incomes, and rising recovery costs. Social capital erodes from inadequate community support and training, hindering collective adaptation. These impacts highlight the urgent need for adaptive strategies, such as livelihood diversification and investments in human and physical capitals (Tran et al., 2022; Smajgl et al., 2015).
According to the SLF, household agricultural production relies on five key capitals: natural (land, water, air, biological resources); human (knowledge, skills, health, information access); physical (infrastructure, equipment, housing); financial (savings, credit, income); and social (networks, organizations, institutions) (DFID, 1999). Each capital contributes uniquely to production processes, from providing inputs and technical support to facilitating market access and risk management.
International studies demonstrate the value of analysing livelihood capitals in agriculture. In China, livelihood resilience is measured across three dimensions: buffering capacity, self-organization, and learning ability (Zhang et al., 2023). In Ethiopia, factors such as land size, education, livestock holdings, gender, age, market distance, credit access, annual income, training, and household size drive diversification strategies (Asante-Addo et al., 2021). Although numerous studies provide broad overviews of rural livelihoods, detailed analyses of the availability, quality, and specific roles of individual capitals in agricultural production remain limited (Reardon et al., 2007), particularly regarding interactions with land use and climate adaptation (Ellis, 2000).
To analysis differences in the impacts of climatic factors (such as rainfall variability, flooding, drought, and salinity intrusion) on interrelated livelihood capitals, MANOVA is considered an appropriate method. According to Hair et al. (2019), MANOVA enables testing for mean differences across multiple correlated dependent variables between independent groups, thereby reducing the risk of Type I error that arises from conducting separate univariate ANOVAs. In the context of climate change impacts on farmers’ livelihoods, MANOVA is particularly effective because the five livelihood capitals (natural, human, social, financial, and physical) are closely interconnected and simultaneously influenced by climatic stressors.
Globally, MANOVA and similar multivariate techniques have been widely employed to evaluate the multidimensional effects of climate change on rural livelihoods and livelihood capitals, particularly within the Sustainable Livelihoods Framework. In Africa (Ethiopia and the northeastern highlands), multivariate approaches have been used to assess livelihood vulnerability based on the five capitals, highlighting concurrent declines in natural and financial capitals due to drought and rainfall variability (Simane et al., 2016; Dendir and Simane, 2021). In Asia (Pakistan, the Indo Gangetic Plains of India, and the Himalayan region), multivariate analyses, including the Livelihood Vulnerability Index (LVI) and multivariate regression models, have elucidated differences among farming households according to their exposure to extreme climate events, with notable effects on human and social capitals (Pandey and Jha, 2012; Venus et al., 2022). In Latin America (the Ecuadorian Andes and other rural areas), studies combining livelihood capital theory with multivariate analysis have examined income inequality and climate change perceptions among small scale livestock farmers (Torres et al., 2022). These studies demonstrate the effectiveness of MANOVA and related multivariate techniques in handling intercorrelations among livelihood capitals, similar to applications of the LVI or other multivariate models (Hahn et al., 2009; Reed et al., 2013).
Although few studies have directly applied MANOVA in Thanh Phu commune, related research in the Mekong Delta (e.g., vulnerability assessments in An Giang, Can Tho, and Tra Vinh provinces) has used similar multivariate approaches, such as the LVI or multivariate Probit models, to analyse the multidimensional impacts of climate change. The application of MANOVA in the present study will help clarify differences in impacts across the five livelihood capitals according to the degree of exposure to climate change, thereby informing suitable adaptation strategies, enhancing resilience in coastal communities of the Mekong Delta, and contributing to the global body of research on livelihood vulnerability under climate change.
Research Methods
Data Collection
Data were collected in January 2024 (updated as of November 2025) through face-to-face interviews with household heads using a structured questionnaire consisting of 33 questions divided into four main sections (i) demographic characteristics of the household head (ii) household livelihood attributes (iii) perceptions of climate change impacts and changes in natural resources (iv) livelihood adaptation strategies.
The section evaluating the five types of livelihood capital (88 indicators questions 16–29) comprised
- Human capital (31 indicators) assessment of 12 channels for accessing information/knowledge (across three dimensions production market and general information) and 7 core capability factors (health education farming techniques information acquisition etc.).
- Natural capital (22 indicators) quality of three main resource types (water soil air) 7 climate change phenomena degree of impact on 4 resource types and 8 local causes.
- Physical capital (10 indicators) current status of housing land ownership farming obstacles and support needs for five asset categories (infrastructure machinery tools transportation other).
- Financial capital (10 indicators) sources of capital investment scale and impacts of four extreme climate events on productivity costs and income (evaluated in percentages).
- Social capital (15 indicators) effectiveness of support and activities provided by five organizations/groups (cooperatives women’s unions farmers’ associations agricultural extension services other) in the context of climate change.
The questionnaire utilized a mixed measurement scale combining qualitative responses (5 levels from very poor to very good) and quantitative measures (specific quantities percentages) to ensure data accuracy and comparability.
Each capital index was calculated as the arithmetic mean of its constituent indicators:
where n is the number of indicators for that capital type, and Xij represents the response value for indicator j of household i.
The arithmetic mean approach was chosen over weighted averages because: (1) no theoretical or empirical basis exists for assigning differential weights to indicators within each capital type (Hahn et al., 2009); (2) equal weighting treats all aspects of each capital as equally important to household livelihoods, consistent with the holistic perspective of the Sustainable Livelihood Framework (Ellis, 2000); and (3) this approach enhances transparency and replicability (OECD, 2008).
Sample size was calculated using Yamane ‘s formula (1967) formula:
with N = 25,000 ( total agricultural households) e = 0.07 (margin of error ± 7%) yielding n ≈ 202. Therefore surveying 200 households was considered reasonable and sufficient to ensure representativeness (Krejcie and Morgan, 1970; Bartlett et al., 2001).
The sample was selected using simple random sampling (Yamane, 1967; Israel, 1992; Singh and Masuku, 2014) combined with geographic stratification across three ecological zones. Five hamlets were randomly selected Giao Thanh, My Hung, Dai Dien, An Nhon, An Quy. Within each hamlet households were randomly chosen based on the criterion of at least 5 years of agricultural experience (Table 1).
The sample achieved spatial balance (approximately 40 households per commune) and relative gender balance (74% male 26% female) ensuring objectivity and representativeness of local livelihood types.
Multivariate Analysis of Variance (MANOVA)
MANOVA is a multivariate statistical method used to test differences in the means of multiple correlated dependent variables across independent groups (Tabachnick and Fidell, 2019). In the context of examining the impacts of climate change on farmers’ livelihoods, MANOVA is particularly appropriate because the livelihood capitals comprise five interrelated components: natural capital, human capital, social capital, financial capital, and physical capital. Unlike conducting separate univariate ANOVAs for each capital type, which risks inflating Type I error rates, MANOVA simultaneously evaluates the combined effects of climate phenomena (heavy rainfall, flooding, drought, and salinity intrusion) on the entire livelihood capital system.
MANOVA decomposes the total variance matrix (T) into between-group (H) and within-group (E) components:
where T is the total matrix, H is the hypothesis (between-group) matrix, and E is the error (within-group) matrix.
Key test statistics include Wilks’ Lambda (
Λ = ∣ E ∣ ∣ T ∣
Prior to conducting MANOVA, the necessary assumptions were examined. The Box’s M test (Table 2) revealed a violation of the homogeneity of covariance matrices assumption across climate impact groups (Box’s M = 80.271, F = 4.178, df1 = 15, df2 = 1029.389, p < .001) (Box, 1949). Despite this violation, MANOVA proceeded with Pillai’s Trace as the primary test statistic due to its greater robustness to such assumption violations (Tabachnick and Fidell, 2019). While Pillai’s Trace is widely regarded as the most robust statistic when homogeneity of covariance matrices is violated ( as confirmed here by Box’s M test, p < .001), it is not immune to all assumption breaches, particularly in cases of severe non-normality or unbalanced designs (Warne, 2014). In livelihood studies, where capital indicators are often ordinal or skewed due to household heterogeneity, multivariate normality violations are common, potentially inflating Type I errors in less robust tests like Wilks’ Lambda. The choice of Pillai’s Trace here mitigates this risk, but future studies could complement MANOVA with non-parametric alternatives (e.g., permutation-based MANOVA) or structural equation modeling to further validate inter-capital relationships under violated assumptions. Additionally, the Roy’s Largest Root results, showing dominant effects (e.g., partial η2 up to 0.382 for flooding), suggest concentrated impacts along principal dimensions, which Pillai’s Trace captures conservatively but may understate in highly collinear data.
Further examination confirmed violations of the homogeneity of variance assumption at the univariate level for all dependent variables: human capital (F = 1.530, p = .017), financial capital (F = 2.335, p < .001), physical capital (F = 2.713, p < .001), social capital (F = 1.452, p = .031), and natural capital (F = 2.335, p < .001) (Table 3). These results reinforced the decision to interpret MANOVA outcomes using Pillai’s Trace rather than Wilks’ Lambda (Wilks, 1932).
The overall MANOVA results indicate statistically significant differences among climate-change impact groups with respect to livelihood capital when considered jointly. Following the recommendation of Tabachnick and Fidell (2019), because the assumption of homogeneity of covariance matrices was violated (Box’s M test, p < .001), Pillai’s Trace was employed as the primary test statistic due to its greater robustness to violations of this assumption.
Results and Discussion
Results of the Multivariate Analysis of Variance (MANOVA)
The MANOVA results (Table 4) demonstrate that flooding exerts a very strong and statistically significant multivariate impact on the livelihood system across all tests. Specifically, Pillai’s Trace, Wilks’ Lambda, and Hotelling’s Trace all report significance levels of p = .000, confirming a clear and consistent overall multivariate effect of flooding on livelihood structures. The Partial Eta Squared values range from 0.127 to 0.155, indicating moderate to moderately strong effect sizes and suggesting that flooding simultaneously affects multiple forms of livelihood capital, including human, natural, physical, financial, and social capital.
The Roy’s Largest Root test further reveals an exceptionally strong dominant effect, with a value of 0.618, F (6, 107) = 11.025, p = 0.000, and a Partial Eta Squared of 0.382. This very large effect size indicates that flooding has a highly concentrated and intense impact along the principal dimension of the multivariate livelihood space, reflecting the abrupt, direct, and highly destructive nature of flood hazards. Such impacts generate simultaneous shocks across the five livelihood capitals, leading to profound disruptions in livelihood structures as well as ecosystem functioning.
The MANOVA results (Table 5) demonstrate that drought exerts a statistically significant multivariate impact on the five livelihood capital dimensions, including human, natural, financial, physical, and social capital. All multivariate tests (Pillai’s Trace, Wilks’ Lambda, and Hotelling’s Trace) report p values of .000, confirming a robust overall effect of drought on the livelihood capital system. The Partial Eta Squared values, ranging from 0.122 to 0.133, indicate moderate effect sizes, suggesting that drought generates cumulative and widespread pressures across multiple forms of livelihood capital rather than affecting a single dimension in isolation.
The Roy’s Largest Root statistic further reveals a highly concentrated dominant effect, with a value of 0.382, F (6, 107) = 6.815, p = .000, and a Partial Eta Squared of 0.276. This large effect size indicates that drought strongly affects a principal dimension of the livelihood capital system, reflecting the critical role of water scarcity in simultaneously undermining natural resources, constraining financial capacity, weakening physical assets, increasing stress on human capital, and eroding social support mechanisms. Together, these results highlight the structurally embedded and systemic nature of drought impacts on livelihood sustainability.
The analysis results (Table 6) show that salinity intrusion exerts a very strong and statistically significant multivariate impact on the five livelihood capital systems, including natural capital, human capital, physical capital, financial capital, and social capital. Specifically, Pillai’s Trace, Wilks’ Lambda, and Hotelling’s Trace all reach significance at p = 0.000, confirming the presence of a comprehensive, pervasive, and severe overall effect of salinity intrusion on the livelihood structure. The Partial Eta Squared values range from 0.121 to 0.141, indicating moderate to moderately strong effect sizes and suggesting that salinity intrusion does not act in isolation but simultaneously degrades multiple forms of livelihood capital, particularly natural and physical capital, thereby generating indirect adverse consequences for financial and human capital.
The Roy’s Largest Root test further reveals a very strong dominant effect, with a value of 0.490, F (6, 107) = 8.741, p = .000, and a Partial Eta Squared of 0.329, indicating an extremely concentrated impact along a principal dimension of the multivariate livelihood space. This large effect size reflects the cumulative, persistent, and difficult to reverse nature of the salinization process, in which the degradation of natural capital plays a central role, undermining productive foundations, increasing livelihood risks, and constraining the long term adaptive capacity and resilience of households.
The MANOVA results (Table 7) indicate that unseasonal rainfall exerts a statistically significant multivariate impact on the livelihood variable system across all tests. Specifically, Pillai’s Trace, Wilks’ Lambda, and Hotelling’s Trace yield significance levels ranging from p = .002 to p = .001, confirming the presence of an overall multivariate effect of unseasonal rainfall on livelihood structures. The Partial Eta Squared values range from 0.098 to 0.108, indicating small to moderate effect sizes and suggesting that unseasonal rainfall causes noticeable disruptions but does not exert a fully dominant influence across all livelihood capital components.
The Roy’s Largest Root test shows a highly significant result with F (6, 107) = 5.063, p = .000, and a Partial Eta Squared of 0.221, indicating the presence of a dominant impact dimension within the multivariate livelihood space. This relatively large effect size suggests that unseasonal rainfall strongly and selectively affects a specific livelihood dimension, reflecting the high sensitivity of ecosystems, agricultural production cycles, and season dependent livelihood strategies to abnormal climatic variability.
Table 8 shows that the interaction between flooding and drought produces the strongest statistically significant effect, with Pillai’s Trace = 0.422, F (25, 535) = 1.970, p = .004, and partial η2 = 0.084. The alternating cycle of floods and droughts fundamentally disrupts the natural resilience of ecosystems, leading to continuous environmental degradation. In addition, Roy’s Largest Root = 0.358 with a partial η2 of 0.264 indicates a large effect size along the dominant dimension, providing strong evidence of a pronounced synergistic interaction effect.
The MANOVA results (Table 9) indicate that the combined effects of unseasonal rainfall and salinity intrusion on the dependent variables are statistically significant at the 5% level across all multivariate tests. Specifically, Pillai’s Trace, Wilks’ Lambda, and Hotelling’s Trace report significance values of 0.034, 0.029, and 0.025, respectively, all below the 0.05 threshold, providing clear statistical evidence of an overall multivariate effect. The corresponding Partial Eta Squared values range from 0.079 to 0.084, indicating small to moderate effect sizes and suggesting that these combined climatic stressors exert a meaningful influence on the set of dependent variables.
In addition, the Roy’s Largest Root test yields a statistically significant result (Sig. = 0.001) with a Partial Eta Squared of 0.176, indicating the presence of a dominant effect dimension. This result suggests that unseasonal rainfall and salinity intrusion strongly affect at least one dependent variable or a subset of dependent variables, highlighting heterogeneity in the magnitude of impacts across variables.
The MANOVA results (Table 10) indicate that the combined effects of drought and salinity intrusion on the dependent variables are not clearly significant at the 5% level for most multivariate tests. Specifically, Pillai’s Trace, Wilks’ Lambda, and Hotelling’s Trace yield significance values of 0.088, 0.079, and 0.071, respectively, all exceeding the 0.05 threshold, suggesting insufficient statistical evidence to confirm a significant overall effect. Nevertheless, the Partial Eta Squared values range from 0.108 to 0.116, indicating a moderate effect size and implying that practical effects may still be present despite the lack of strong statistical significance.
In contrast, the Roy’s Largest Root test shows a statistically significant result (Sig. = 0.001) with a Partial Eta Squared of 0.23, indicating that at least one dominant dimension of drought and salinity intrusion exerts a strong influence on a specific dependent variable or a subset of dependent variables. This result suggests heterogeneity in effects across variables and underscores the need for follow-up univariate ANOVA or post hoc analyses to identify the variables most strongly affected.
The divergence between Pillai’s Trace (p = .088, non-significant) and Roy’s Largest Root (p = .001, significant) reflects their fundamental difference: Pillai’s Trace evaluates effects across ALL dimensions, while Roy’s Largest Root examines only the FIRST dominant dimension (Tabachnick and Fidell, 2019).The significant Roy’s Largest Root indicates that drought combined with salinity strongly affects one specific capital—natural capital—rather than all five capitals uniformly. Given the violation of homogeneity of covariance matrices (Box’s M, p < .001), Pillai’s Trace is prioritized for conservative interpretation. Thus, this interaction is interpreted as having a targeted effect on natural capital rather than a systemic multivariate effect.
Following the multivariate analysis, the Tests of Between-Subjects Effects table was examined to assess both the individual and interaction effects of climate change variables on various livelihood capitals. The results are presented below.
Table 11 presents the multivariate test results examining the effects of both single and interactive climate change impacts on livelihood capitals. Overall, the findings indicate that different climate stressors exert heterogeneous and statistically significant influences across specific livelihood capital dimensions.
For single climate impacts, floods show statistically significant effects on natural, human, social, and physical capital (p < .05), while no significant association is observed with financial capital (p = .258). Unseasonal rainfall significantly affects natural capital (p = .001) and financial capital (p = .012), but does not exert statistically significant effects on human, social, or physical capital. Drought is found to significantly influence human, social, and physical capital (p < .05), whereas its effects on natural and financial capital remain statistically insignificant. Salinity intrusion demonstrates significant impacts on physical and financial capital (p < .01), while no significant effects are detected for natural, human, or social capital.
Regarding interactive climate impacts, the combined effects generally exhibit fewer statistically significant relationships compared to single stressors. The interaction between floods and unseasonal rainfall shows a significant effect only on social capital (p = .031). The combined impact of floods and drought significantly affects natural capital (p = .021) and financial capital (p = .004), highlighting compounding stress on both environmental and economic resources. Similarly, the interaction between floods and salinity intrusion significantly influences financial capital (p = .038). The interaction between unseasonal rainfall and drought does not yield any statistically significant effects across all livelihood capitals (p > .05). In contrast, unseasonal rainfall combined with salinity intrusion significantly affects human capital (p = .005), while the interaction between drought and salinity intrusion shows a significant association with natural capital (p = .012).
Taken together, these results suggest that single climate stressors tend to exert broader and more consistent impacts on livelihood capitals, whereas interactive effects are more selective and capital-specific. This pattern underscores the importance of disaggregating both individual and combined climate hazards when assessing livelihood vulnerability and designing targeted adaptation strategies. These findings must be interpreted cautiously given the SLF’s acknowledged limitations in capturing power relations and structural drivers beyond household-level capitals (e.g., upstream hydropower dams exacerbating drought and salinity in the Mekong Delta; Smajgl et al., 2015). The stronger single-stressor effects compared to interactions may partly reflect the framework’s static view of capitals, potentially underestimating cumulative, long-term synergies from compounded hazards (e.g., repeated drought-salinity cycles leading to irreversible land degradation; recent 2024–2025 DSI events in VMD). Moreover, the sample’s focus on agricultural households with ≥ 5 years’ experience may bias toward more resilient groups, excluding landless or newly vulnerable populations often hit hardest by climate shocks (IPCC, 2023). Despite these constraints, the MANOVA approach advances SLF applications by providing quantitative evidence of capital interdependencies, supporting more precise, capital-targeted policies.
Conclusion
This study applied SLF and MANOVA to examine the differential impacts of climate change stressors, including flooding, drought, salinity intrusion, and unseasonal rainfall, on the five livelihood capitals (natural, human, physical, financial, and social) of agricultural households in the coastal Thanh Phu commune, Mekong Delta, Vietnam. The findings reveal that climate change exerts significant and multifaceted effects on rural livelihoods, with single climatic stressors generally producing stronger and broader impacts than their interactions.
Flooding, drought, and salinity intrusion emerged as the most severe threats, each demonstrating robust multivariate effects (p = .000) with moderate to large effect sizes (partial η2 ranging from 0.121 to 0.382). Flooding exhibited the strongest dominant impact (Roy’s Largest Root partial η2 = 0.382), simultaneously affecting natural, human, social, and physical capitals, while sparing financial capital. Drought significantly influenced human, social, and physical capitals, reflecting cumulative stress on health, community networks, and infrastructure. Salinity intrusion, characteristic of coastal zones, strongly affected physical and financial capitals, underscoring persistent degradation of productive assets and income stability. Unseasonal rainfall, while statistically significant (p ≤ .002), displayed smaller effect sizes, primarily targeting natural and financial capitals.
Interaction effects were more selective and generally weaker. Significant synergistic impacts included the combination of flooding and drought on natural and financial capitals, flooding combined with salinity intrusion on financial capital, and unseasonal rainfall combined with salinity intrusion on human capital. These findings are consistent with prior research in the Mekong Delta and other vulnerable regions, confirming that compounded climatic hazards amplify vulnerability in specific capital dimensions rather than uniformly across the entire livelihood system (Tran et al., 2022; Le and Nguyen, 2025; Smajgl et al., 2015).
The results highlight the interconnected nature of livelihood capitals and the necessity of disaggregated analysis to identify capital-specific vulnerabilities. Natural and financial capitals appear particularly susceptible to direct environmental degradation and economic shocks, whereas human and social capitals are more affected by prolonged or interactive stressors. These insights contribute to the growing body of evidence on climate-induced livelihood vulnerability in deltaic environments and underscore the value of multivariate approaches such as MANOVA for handling intercorrelations among capitals (Hoa et al., 2025; Simane et al., 2016; Pandey and Jha, 2012).
From a policy perspective, adaptation strategies should prioritize integrated interventions that strengthen the most affected capitals. Investments in salinity tolerant crops, early warning systems, and climate resilient infrastructure are critical to mitigate impacts on natural and physical capitals. Enhancing human capital through targeted training and extension services, alongside strengthening social capital via farmer cooperatives and community-based networks, will improve adaptive capacity. Financial capital can be bolstered through accessible credit, crop insurance, and value-chain development for high value products such as coconut and pomelo. Livelihood diversification, promoting integrated aquaculture-agriculture models and off-farm opportunities, remains essential for reducing dependence on climate-sensitive activities. Policy recommendations should also address SLF critiques by incorporating power and equity considerations, such as gender-differentiated access to adaptation resources and support for marginalized groups (e.g., female-headed or land-poor households) facing higher vulnerability (IPCC, 2023). In the Mekong Delta context, where recent droughts and salinity intrusion (2024–2025) have intensified, integrated strategies must go beyond household-level interventions to tackle upstream governance and transboundary water issues for long-term resilience.
In conclusion, addressing climate change impacts in the Mekong Delta requires context-specific, capital targeted policies that move beyond generalized adaptation to foster resilient and sustainable rural livelihoods. Future research should explore longitudinal effects and evaluate the effectiveness of ongoing adaptation interventions to inform scalable solutions for coastal communities facing escalating climatic risks.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation
