Regenerative Soil Management: Restoring Soil Health, Biodiversity and Ecosystem Functions for Sustainable Agriculture

Climate-smart agriculture (CSA) has become a prominent framework for linking agricultural development with climate adaptation and, where feasible, mitigation. Yet its expansion in developing countries has often been treated as a technology-diffusion problem, even though persistent gaps in uptake, durability and distributional benefit point to deeper constraints in research systems, institutions, finance and policy coordination. This critical narrative review synthesises evidence published from 2010 to 18 July 2026 on how CSA science, innovation and policy can be integrated for responsible scaling in low- and middle-income settings. The literature indicates that many CSA practices can improve productivity, income, food security or resilience under particular conditions, but effects are heterogeneous and causal evidence remains uneven. Mitigation outcomes are less consistently measured than productivity and adaptation, while long-term resilience, dis-adoption, spillovers and equity are rarely evaluated. Adoption studies repeatedly identify extension, training, climate information, farmer organisations, tenure security and access to finance as enabling factors, yet their effects depend on the resource intensity and compatibility of specific practices. Evidence from scaling initiatives shows that reach is greatest when technical options are embedded in service delivery, value chains, climate information systems and public policy, but broad reach can weaken local fit and exclude resource-constrained groups. Science-policy platforms, participatory targeting and flexible financing can improve alignment, although institutional fragmentation and project dependence remain common. The synthesis therefore reframes CSA scaling as a portfolio-and-system alignment challenge rather than the replication of a standard package. Effective scale requires locally valid science, innovation systems that reduce risk and transaction costs, and policy architectures that coordinate incentives, public goods, safeguards and accountability across farm, landscape and national levels. Future research should prioritise longitudinal and experimental or quasi-experimental designs, comparable multi-pillar metrics, distributional analysis and evaluation of policy mixes rather than isolated technologies.