Combining participatory and modeling approaches to investigate factors and drivers of soil erosion risk in mixed crop-livestock farms

Soil erosion threatens mixed farms in marginal areas, endangering their cultural and economic role in territories where pastoralist systems are already under pressure for climatic, socioeconomic, and generational factors. The rise in extreme rainfall events worsens soil loss on farmland, underscoring the need to co-develop practices that boost climate resilience in agriculture. This study helps fill the gap in understanding how the integration of farmers’ perceptions with spatial modeling can inform land management strategies. We combined farmers’ perceptions, model predictions, and farm management to provide an integrated assessment of the soil erosion. We represented the geographical distribution of soil erosion risk through geographical information systems-based RUSLE modeling. Farmers’ perceptions on soil erosion were assessed through surveys and fuzzy cognitive mapping conducted across 25 sheep farms. Our model shows that 37% of cropland is at risk, mainly due to land topography and soil cover. Fuzzy cognitive maps reveal that farmers are aware of the main environmental and human-linked soil erosion drivers. Farmers recognize cropping system design, especially using perennial forage instead of annual crops, as key to reducing soil erosion, and also see temporary ditches, reduced tillage, and agroforestry as effective measures. Utilizing a multivariate ordinal logistic regression, we showed that sheep farmers with a higher education level tend to perceive higher soil erosion risk. The number of conservation measures adopted increases when farmers are more aware of soil erosion issues, when they identify a higher number of fuzzy cognitive map connections, and when the predicted soil erosion risk is higher. Farmers’ perceptions of erosion risks and soil conservation measures aligned with model predictions on soil erosion, highlighting the importance of systematically involving farmers in research and policy design. Their detailed mental models enhance environmental models and should be considered in the European Common Agricultural Policy for sustainable rural development.