International Journal of Environment and Climate Change

Terraces, contour and graded bunds, check dams and farm ponds form the physical backbone of soil and water conservation in rainfed agricultural landscapes across Asia, Africa, the Mediterranean basin and Latin America. Their design conventions were established when the statistical properties of rainfall were assumed to be fixed, so that a single design storm derived from a historical record could be treated as a durable specification. That assumption is now under sustained challenge from evidence of intensifying short-duration rainfall, rising rainfall erosivity and changing storm structure, and from a parallel methodological debate about whether observed changes justify explicitly time-varying design statistics. This critical narrative review examines how far the engineering evidence base supports climate-proofing of these four structure classes, and where it does not. The review synthesises evidence on the changing design environment, the hydraulic and geotechnical behaviour of each structure class under high-intensity rainfall, catchment-scale interactions among structures, and the appraisal frameworks used to justify investment. Three findings recur. First, the performance of small conservation structures is governed less by mean annual rainfall than by the intensity, sequencing and antecedent-moisture context of a small number of extreme events, yet most reported effectiveness estimates derive from plot-scale or annual-aggregate observations that suppress precisely this variability. Second, the dominant failure pathways are hydraulic and geotechnical rather than agronomic, and they concentrate at conveyance and outlet components that are frequently omitted from both design guidance and monitoring. Third, structural resilience depreciates through siltation, deferred maintenance and abandonment at rates comparable to the rate of climatic change, so that institutional and financial arrangements determine realised safety as strongly as the design storm does. Confidence in current effectiveness estimates is limited by short observation periods, uneven geographical representation and scarce documentation of failure. Priorities are identified for failure-oriented monitoring, load-path-based design and robustness-oriented appraisal.