Community forest governance and synergies among carbon, biodiversity and livelihoods

The urgency of the twin crises of climate change and biodiversity loss has intensified policy focus on forest conservation and restoration, including nature-based solutions for mitigation and expanded protected areas. Rural tropical landscapes—home to high biodiversity and major carbon sequestration potential—often have substantial human presence, with approximately 1.8 billion people living on lands needed to meet global biodiversity goals. These human-dominated forest landscapes are integral to rural livelihoods and adaptive capacity. Understanding relationships among carbon sequestration, biodiversity conservation, and rural livelihoods can help identify interventions that simultaneously support well-being and environmental objectives. Researchers and policy makers increasingly emphasize advancing multiple human and environmental objectives together and accounting for their interactions, trade-offs, and synergies. Yet many global analyses target single outcome domains, risking unintended trade-offs and missed co-benefits. Links between livelihood benefits and other outcomes are less well understood than carbon–biodiversity relationships. Although global opportunities for conservation and restoration are being mapped, careful analyses of institutional mechanisms that support co-benefits—especially at subnational scales—remain rare. Informed choices across competing priorities require knowledge of how interventions targeting one outcome affect others. Identifying factors associated with different combinations of forest outcomes is therefore central to calibrating interventions to minimize trade-offs and enhance co-benefits.

The paper situates its contribution within several strands of literature: (1) nature-based climate solutions and expansion of protected areas; (2) empirical work on trade-offs and synergies among multiple ecosystem services; (3) extensive research on carbon–biodiversity relationships in tropical forests; (4) emerging analyses of forest landscape restoration emphasizing local communities, equity, and governance; and (5) studies of decentralized and community-based forest management showing potential benefits for both conservation and poverty reduction. The authors highlight that global and single-outcome analyses can obscure multi-benefit interactions and trade-offs, that livelihood co-benefits remain understudied relative to carbon and biodiversity, and that there is a paucity of careful, subnational, institutional analyses linking governance mechanisms (formal community associations, participation in rule-making) to multiple outcomes. This work responds by examining institutional predictors of multi-dimensional benefits using a large, cross-country dataset of forest commons.

Study design and data: The analysis uses data on 314 forest commons in human-dominated tropical landscapes across 15 countries (Africa, Asia, Latin America) from the International Forestry Resources and Institutions (IFRI) program (compiled October 2018). IFRI sites represent smaller, fragmented forest patches embedded in agricultural matrices with relatively dense, lower-income populations and are selected to reflect a range of forest management regimes, not forest condition. For cases with multiple visits, the most recent plot data were used. Forests outside the low-income tropics, those <5 ha, and those lacking vegetation plot data were excluded. Biophysical sampling: Approximately 30 randomly located circular plots per forest (10 m radius; ~314 m²) were inventoried. For each tree (girth >31.4 cm at breast height ≈ DBH >10 cm), local/botanical name and girth were recorded. Benefits measured:

• Biomass (proxy for above-ground carbon stocks): basal area (m²/ha) computed by summing basal area of trees per plot and averaging across plots within each forest.
• Biodiversity: tree species richness estimated via the non-parametric Chao1 estimator using EstimateS software. Species abundance data per plot were used, with 100 randomized runs to produce 95% CIs. Observations with Chao1 >140 were excluded.
• Livelihoods: a livelihoods index was constructed via principal component analysis (PCA) using settlement-level dependence proportions for fodder, fuelwood, and timber supplied by the forest (aggregated to forest level). The first factor (Eigenvalue 1.77) explained 59.17% of variance (LR test χ²(3)=237.11, p<0.0001). Reliability: Cronbach’s alpha=0.613; the livelihoods index correlates strongly with alpha (Spearman’s rho=0.9669, p<0.0001). Governance and intervention variables (policy-relevant predictors):
• Association: binary indicator that the forest is owned/managed by a formal association of users under national law (formal community forest management association).
• Rule-making: indicator of meaningful local participation in formal rule-making (authority to make rules lies with a local association/government/NGO). Association and rule-making are orthogonal, capturing distinct governance aspects.
• Tree plantation: binary indicator that tree planting occurred in the past 10 years. Analytical approach:
• Bivariate associations among benefits: Spearman rank correlations among biomass, tree species richness, and livelihoods; log-transformations applied to biomass and tree richness where appropriate.
• Cluster analysis: Ward’s-linkage hierarchical clustering on standardized, log-transformed benefit measures (biomass, tree species richness, livelihoods) identified groups of forests with similar multi-benefit profiles. Five clusters were retained based on clustering structure (Supplementary Fig. 3; Supplementary Table 3): sustainable, carbon, conservation, subsistence, degraded.
• Multivariate test of cluster separation: MANOVA of benefits across clusters (Wilks’ λ=0.1088, F=156.44, p<0.0001; Lawley–Hotelling trace=4.8029, F=184.31, p<0.0001).
• Regression modeling: Multinomial logistic regressions (base category: degraded forests) examined the association of each predictor (association, rule-making, tree plantation) with cluster membership. Marginal effects on cluster probabilities and relative risk ratios (RRR) relative to degraded forests were computed. Robust (Huber-White) SEs with and without clustering by country produced similar results; Hosmer–Lemeshow GOF for multinomial models indicated good fit; LR and Wald tests did not indicate violations. Ethics and data availability: Data were collected by IFRI centers following local laws/ethics; latest IRB: University of Michigan HUM00092191. The dataset is publicly available (Mendeley Data).

Interrelationships among benefits:

• Biomass vs. tree species richness: weak positive correlation (Spearman’s rho=0.1989; p=0.0004).
• Livelihoods vs. tree species richness: weak positive correlation (Spearman’s rho=0.2268; p=0.0001).
• Biomass vs. livelihoods: no significant relationship (Spearman’s rho=−0.0246; p=0.6641). Clusters of multi-benefit outcomes (n=314):
• Sustainable forests (n=119): highest overall levels across benefits; above-average livelihoods and tree richness; biomass ranges from below to above average.
• Carbon forests (n=23): above-average biomass; average livelihoods; below-average tree richness.
• Conservation forests (n=88): average to above-average biomass and tree richness; below-average livelihoods.
• Subsistence forests (n=57): above-average livelihoods; below-average biomass and tree richness.
• Degraded forests (n=27): below-average biomass, tree richness, and livelihoods. Governance and interventions—marginal effects on cluster probabilities (direction and magnitude):
• Formal community management association: increases probability of sustainable (+0.19, p<0.001), carbon (+0.06, p=0.036), and subsistence (+0.11, p=0.007) clusters; decreases probability of conservation (−0.28, p<0.001) and degraded (−0.08, p=0.034).
• Local participation in rule-making: increases probability of carbon (+0.08, p=0.007) and subsistence (+0.16, p<0.001) clusters; decreases probability of conservation (−0.29, p<0.001).
• Tree plantation (past 10 years): increases probability of sustainable (+0.19, p<0.001) and subsistence (+0.24, p<0.001) clusters; decreases probability of carbon (−0.12, p<0.001) and conservation (−0.37, p<0.001) clusters. Avoiding negative outcomes—relative to degraded forests (RRR for one-unit change):
• Association: higher odds of sustainable (RRR=4.06, p=0.002), carbon (RRR=5.94, p=0.008), and subsistence (RRR=4.69, p=0.002) clusters.
• Rule-making: higher odds of carbon (RRR=5.24, p=0.01) and subsistence (RRR=4.07, p=0.004) clusters.
• Tree plantation: lower odds of carbon (RRR=0.05, p<0.001) and conservation (RRR=0.09, p<0.001); no significant association with sustainable or subsistence relative to degraded. Additional bivariate context (Extended Data Table 1): governance variables tend to have weak or negative associations with biomass and tree richness when examined singly, and positive associations with livelihoods; tree planting shows a strong negative bivariate association with biomass (rho=−0.4385, p<0.001), underscoring trade-offs when single outcomes are analyzed in isolation.

The cross-country analysis shows that forest commons managed and used by indigenous and rural communities can contribute to global environmental objectives (carbon storage and biodiversity) while also supporting local livelihoods. However, trade-offs are common across human-dominated landscapes; not all forest patches yield simultaneous maxima across benefits. Recognizing multifunctionality and heterogeneity is essential for designing policy interventions that enhance aggregate landscape outcomes rather than pursuing uniform “win–win” expectations. Institutional conditions—specifically, the presence of a formal community forest management association and meaningful local participation in rule-making—are consistently associated with more favorable multi-benefit clusters and with avoiding degraded outcomes. This supports theories that local actors have comparative advantages in context-specific governance, rule design, and monitoring. In contrast, tree planting, a widely promoted restoration strategy, displays heterogeneous and in some cases adverse associations with conservation and carbon-oriented clusters relative to degraded forests, suggesting that prevailing tree-planting practices may not, by themselves, deliver multi-objective restoration. These findings address the core research question by identifying governance features that are linked to synergies among carbon, biodiversity, and livelihoods across diverse contexts. The results underscore a policy shift from narrow tree-planting targets toward empowering local institutions with formal authority and participation mechanisms to achieve durable, multi-dimensional benefits. The analysis also cautions that the observed associations are not necessarily causal, motivating more context-sensitive, causal inference at national and subnational scales.

This study advances understanding of how community forest governance relates to combined outcomes in carbon storage, biodiversity, and livelihoods across tropical forest commons. Five distinct outcome clusters characterize prevalent synergies and trade-offs, with the largest group (sustainable forests) exhibiting jointly positive outcomes. Empowered local governance—formal community associations and meaningful participation in rule-making—is a robust predictor of favorable multi-benefit outcomes and reduced odds of degraded states, whereas tree planting shows mixed and sometimes negative associations with conservation and carbon-focused clusters. Policy implications include prioritizing institutional reforms that legally recognize and strengthen community management authority, support local rule-making, and ensure accountability, while implementing tree planting in locally responsive, socially inclusive ways. Future research should investigate which specific governance features and enabling conditions most effectively foster multi-benefit outcomes, how these evolve over time with decentralized management experience, and how governance interacts with socioecological factors to shape landscape-level multifunctionality.

• Sampling and selection: IFRI sites are not a random sample of all tropical forests; they represent human-dominated, fragmented forest commons. Results should be generalized cautiously beyond similar contexts.
• Causality: The study identifies associations, not causal effects. Unobserved confounders and context-specific dynamics may influence both governance arrangements and outcomes.
• Measurement proxies: Biomass (basal area) and biodiversity (tree species richness via Chao1) are proxies and may not capture all dimensions of carbon storage or biodiversity; livelihoods index is based on self-reported dependence and factor construction with moderate reliability (alpha=0.613).
• Temporal dynamics: Cross-sectional analysis limits inference about temporal trajectories and lagged effects of interventions such as tree planting or institutional change.