Developing countries can adapt to climate change effectively using nature-based solutions

The study investigates the effectiveness of climate change adaptation interventions in low- and middle-income countries (LMICs), focusing on both risk reduction and development-related outcomes. Motivated by Sustainable Development Goals linking adaptation with poverty reduction, economic stability, and public health, the paper addresses the fragmented understanding of intervention effectiveness across sectors and outcomes. The core research question is: to what extent are different types of adaptation interventions in the agricultural and coastal sectors effective in achieving risk-reduction and development outcomes in LMICs? The work contributes to debates on whether targeting climate protection (hazards/exposure) versus broader sustainable development (vulnerability/adaptive capacity) is more effective, and on the relative effectiveness of hard (technological/infrastructure) versus soft (behavioural/institutional) interventions, including the growing role of Nature-based Solutions (NbS).

Evidence gap maps and systematic reviews have increasingly synthesized adaptation effectiveness, especially in agriculture and coastal sectors of LMICs and SIDS. Agricultural evidence spans case studies, experiments, quasi-experiments, and reviews targeting farmer behaviour, productivity, and livelihood resilience. Coastal evidence includes case studies and experiments on NbS and other measures to reduce economic damages and livelihood vulnerability. Prior syntheses emphasized state-of-the-art, metrics, planning, financing, response types, or specific interventions/outcomes, often not distinguishing LMICs vs. industrialized contexts or sectors, and frequently focusing only on effective interventions. Literature underscores interactions between adaptation, risk reduction, and development priorities. There is a notable interest in NbS for cost-effectiveness and multiple co-benefits, but past evidence was biased towards Global North contexts and lacked socio-economic outcome assessments. This review addresses those gaps by integrating sectoral distinctions, LMIC focus, and both risk- and development-related outcomes.

Design: Systematic review drawing on an existing Evidence Gap Map (EGM) of adaptation in LMICs. The final dataset includes 103 studies and 363 empirical observations (coastal: 19 studies, 97 observations; agricultural: 84 studies, 266 observations). An observation is defined as one intervention-outcome effect within a study. Inclusion criteria (PICOS): Population: individuals, groups, institutions, systems, and communities in agricultural and coastal sectors in LMICs. Interventions: actions to adjust to, reduce, stop, or leverage benefits from climate change/climatic hazards. Comparators: no intervention, different levels, or between interventions. Outcomes: reduction of climate hazard/exposure, reduction of social/economic vulnerability (including adaptive capacity), or contribution to the enabling environment. Study types: quantitative (correlation, impact, or review methods) in peer-reviewed/grey literature, 2007–2018, English or with English abstract (for French, Spanish, German). Conceptual framework: Climate risk conceptualized as interaction of hazards with vulnerability and exposure. Outcomes categorized as (1) climate hazard and exposure reduction, (2) social/economic vulnerability reduction (including adaptive capacity), and (3) enabling environment improvements. Interventions classified into seven types: nature-based solutions (NbS); built infrastructure; technological; informational/educational; institutional (policies, plans, regulations); financial (credit, markets, transfers, insurance); social/behavioural (social support, behavioural change). Screening and coding: PRISMA-informed process using the Doswald et al. (2020) EGM database. Studies coded for author, year, country, design, sector, intervention type, outcome category, and effect direction/size. A qualitative consensus approach ensured coding reliability: initial joint coding of six studies until agreement saturation, followed by independent coding with collaborative resolution of questions. Hierarchical dataset allowed multiple observations per study; all observations weighted equally. Effect measures: Effect direction coded as positive, neutral (non-significant or explicitly no effect), or negative. Effect size coded on an ordinal scale (small, medium, large) for non-neutral effects, translating reported quantitative metrics; defaulted to medium unless authors indicated very small effects or metrics were below specified thresholds. Direction and size recoded to a 7-point scale: -3 (large negative), -2 (small/medium negative), -1 (very small negative/negligible), 0 (neutral), 1 (very small positive/negligible), 2 (small/medium positive), 3 (large positive). Geographic coverage: Studies spanned multiple regions with concentrations in China, multi-country analyses, India, Pakistan, Bangladesh, Ethiopia, Kenya; coastal studies concentrated in South Asia; agricultural studies prominent in South Asia, East Asia & Pacific, and Sub-Saharan Africa. Analysis: Descriptive synthesis of effect directions and mean effect strengths by intervention type, sector, and outcome category; assessment of single vs. combined interventions’ effectiveness; contextual covariates noted where multivariate analyses were present.

• Overall effectiveness: Across sectors and interventions, positive effects outnumber negative effects. Positive effects were more frequent in the agricultural sector than in the coastal sector.
• Sectoral focus of outcomes: Coastal sector evidence concentrates on risk reduction; agricultural sector evidence more often on development-related outcomes (social/economic vulnerability and enabling environment). In agriculture, only 23% of observations addressed risk reduction.
• Nature-based Solutions (NbS): Only intervention type with positive effects across all outcome categories in both sectors. Coastal sector NbS: 29 positive vs 9 neutral/negative observations; mean effects by outcome: hazard/exposure 1.2; social/economic vulnerability 0.8; enabling environment 0.67. Agricultural sector NbS mean effects: hazard/exposure 1.8; social/economic vulnerability 1.3; enabling environment 1.1.
• Coastal sector patterns:
  • Social/behavioural interventions associated more with neutral/negative effects (17 neutral/negative vs 13 positive). Mean effect for social/economic vulnerability is negative (-0.7), but positive for hazard/exposure (0.6). A large share of negative development-related effects stems from one relocation study in Bangladesh finding worsened living conditions and risk-related outcomes.
  • Technological interventions showed mostly negative development-related effects (3 of 4 observations; mean -2.0 for social/economic vulnerability; -0.5 for enabling environment).
  • Built infrastructure and informational and institutional interventions had positive mean effects where evidence exists (e.g., institutional mean 2.0 across categories, though based on few observations).
  • Overall coastal means (ALL): hazard/exposure 1.1; social/economic vulnerability negative (-0.7); enabling environment 0.3.
• Agricultural sector patterns:
  • Majority of observations positive (~69%), with ~19% neutral and ~12% negative.
  • Technological and informational interventions were mostly positive (~75% positive within each). Informational mean effects: hazard/exposure 1.3; social/economic vulnerability 1.5; enabling environment 1.3.
  • Built infrastructure and NbS showed relatively high positive mean effects on social/economic vulnerability (built infrastructure 1.3; NbS 1.3). Examples include irrigation investments improving yields and intercropping increasing yield stability.
  • Financial and social/behavioural interventions had notable shares of negative effects (19% and 22% within-group, respectively). Examples include migration reducing milk production/income in Pakistan and market/credit access decreasing production in parts of Brazil.
  • Overall agricultural means (ALL): hazard/exposure 1.2; social/economic vulnerability 1.2; enabling environment 0.8.
• Distribution of interventions and outcomes (Table 1 highlights): In coastal sector, most observations were NbS (25 hazard/exposure; 10 social/economic vulnerability; 3 enabling) and social/behavioural (14 hazard/exposure; 14 social/economic vulnerability; 2 enabling). Agricultural sector covered all types: informational (15 hazard/exposure; 42 social/economic vulnerability; 9 enabling), financial (11; 28; 9), built infrastructure (14; 28; 1), social/behavioural (7; 28; 6), institutional (8; 14; 10), NbS (5; 16; 7), technological (1; 7; 0).
• Combinations of interventions: 61% of studies used a single intervention. Coastal studies predominantly single (74%); agricultural studies more often combined 3+ interventions. Combining three or more interventions yielded only slightly higher proportions of significant positive effects versus single or double interventions.
• Contextual robustness: Several studies’ positive effects hold after controlling for contextual covariates (e.g., mangroves’ flood protection controlling for population density, elevation, etc.; extension services’ resilience effects controlling for socio-economic and farm characteristics).

Findings affirm that adaptation interventions in LMICs can achieve both risk reduction and development-related outcomes, addressing the study’s central research question. Sectoral differences emerged: coastal interventions more consistently reduce immediate climate risks, whereas agricultural interventions more often improve development-related dimensions, highlighting the interlinkages between adaptation and development in agriculture. NbS stand out as broadly effective across sectors and outcomes, supporting their promotion for co-benefits and cost-effectiveness. However, evidence in coastal contexts suggests potential trade-offs where some interventions (e.g., social/behavioural measures including relocations) may reduce hazards yet harm socio-economic outcomes, underscoring the need for careful design and monitoring. Informational interventions in agriculture show consistently positive effects on both risk reduction and development outcomes, emphasizing the role of perceptions, learning, and decision support under climate stress. Technological and infrastructure measures also contribute meaningfully to reducing vulnerability, particularly through improved water management. Policymakers should align interventions with outcome domains where evidence is strongest, integrate assessments with contextual covariates to identify scope conditions, and remain vigilant for negative effects to enable corrective action. Combining interventions may offer synergies, but current evidence shows only modest gains over single interventions, suggesting more targeted research on optimal combinations is needed.

This systematic review synthesizes evidence from 363 observations across 103 studies in LMICs to evaluate adaptation effectiveness in agricultural and coastal sectors. Key contributions include demonstrating that: (1) adaptation interventions can positively affect both risk reduction and development outcomes; (2) NbS are uniquely positive across all outcome categories and sectors; (3) informational interventions in agriculture are reliably effective for both risk and development outcomes; and (4) sectoral patterns and potential trade-offs necessitate tailored intervention design and evaluation. Policy and practice implications are to prioritize interventions with demonstrated co-benefits (especially NbS and informational measures), assess interventions alongside contextual factors, and proactively monitor for negative impacts. Future research should expand primary studies in coastal contexts and on long-term development outcomes, investigate synergies among combined interventions (particularly including NbS), and strengthen integrated monitoring, evaluation, and learning systems, incorporating qualitative data to capture mechanisms and equity impacts.

The review acknowledges: (1) potential language bias (English-only search terms) and underrepresentation of non-English literature; (2) predominance of quantitative studies, likely underrepresenting interventions suited to qualitative evaluation (e.g., informational/institutional, enabling environment) and limiting insight into definitions of effectiveness and causal mechanisms; (3) publication bias toward positive findings, with coastal sector evidence clustered in NbS and social/behavioural interventions and underrepresentation of infrastructure interventions due to fewer ex-post evaluations; (4) difficulty identifying adaptation interventions when not explicitly labeled and lack of specificity about climate threats in some studies; (5) limited use of multivariate analyses in many studies, leaving influential covariates underexplored; (6) heterogeneity in adaptation contexts and outcomes (dependent variable problem) precluding rigorous meta-analysis and limiting systematic coding of socio-economic, institutional, and biophysical contexts; and (7) small sample sizes in certain sector-intervention-outcome cells (e.g., coastal technological/built infrastructure), reducing generalizability of those estimates.