| Sumario: | Inland valleys are among the most productive agricultural landscapes in West Africa, yet they are increasingly exposed to hydrological extremes, notably droughts and floods, which constitute the dominant climate-related stresses affecting rice-based systems. Importantly, these hazards do not occur independently: drought and flooding can occur within the same year and, in some cases, within the same growing season, creating compounded risks that severely constrain rice productivity, investment decisions, and livelihood resilience. Despite their growing importance, there remains limited understanding of the predictors and spatiotemporal dynamics of these concurrent hazards in inland valleys, hindering the development of effective early warning systems and targeted adaptation strategies.
This study analyzed the occurrence, drivers, and adaptation responses to drought and flood hazards across 68 inland valleys in Benin using a mixed-methods approach integrating 30-year climate data (1991–2021), geospatial analysis, machine learning, and socio-economic surveys of 1,124 farmers. Drought and wetness conditions were quantified using the Standardized Precipitation Evapotranspiration Index (SPEI), while Random Forest models were employed to identify key climatic and biophysical predictors. Farmers adaptation strategies and their determinants were analyzed using qualitative thematic analysis, principal component analysis, and Poisson regression.
Results revealed pronounced spatial and temporal heterogeneity in hazard exposure. Extreme drought events were observed in 2001 and 2016 across all sites, while extreme wet conditions occurred in 2008, 2009, and 2019. Cluster analysis classified inland valleys into three groups: highly drought-prone (35%), highly wet-prone (30%), and moderately exposed to both hazards (35%), closely matching farmers’ perceptions. Drought occurrence was primarily driven by soil organic carbon, SPEI, and sand content, whereas wetness was strongly associated with annual rainfall, soil organic carbon, and SPEI. Farmers adopted an average of four drought-related and six flood-related adaptation measures, with access to extension services, land ownership, and road infrastructure increasing the number of measures adopted.
Accurately predicting drought and flood occurrence in inland valleys is critical for optimizing cropping calendars, guiding investment in water management infrastructure, reducing yield losses, and enabling timely farmer-led adaptation. The findings provide a robust framework for predicting drought and flooding and developing inland valley–specific early warning systems, zoning hazard-prone landscapes, and scaling climate-smart solutions such as Smart-Valleys, improved drainage, and adaptive planting strategies.
Keywords: Lowland, climate extremes, hazard adaptation, West Africa.
|
|---|