The climate-biodiversity-livestock nexus in Mediterranean mixed agroecosystems: Modeling crop yields adaptation using a system-GMM approach

Abstract

Across the semi-arid landscapes of Mediterranean Agroecosystems, crop productivity is increasingly threatened by rising temperatures, erratic rainfall, and evolving land-use pressures. While crop diversification and livestock integration are widely promoted as adaptive strategies, their combined effectiveness remains underexplored in data-scarce contexts like North Africa. This study investigates the dynamic interactions between climate stressors, crop diversity, and livestock integration, with a specific focus on understanding how these components jointly shape crop productivity across diverse agroecological settings. Using a System-GMM estimation approach, the analysis addresses endogeneity, persistence, and interactions effects. By endogenizing livestock integration, the model empirically accounting for feedback loops between livestock dynamics and cropping systems diversification, particularly under variable climate conditions. Machine-learning techniques are used for optimal model selection. Interaction and marginal effects analysis are conducted to uncover synergistic effects. Findings reveal that crop yield responds non-monotonically to temperature anomalies, rainfall deficits, and diversification indices. The impact of crop diversification is strongly conditioned by livestock presence and type. Sheep grazing, when combined with diversified cropping systems, enhances yield resilience to heat stress but exacerbates vulnerability to rainfall deficits. Several double and triple interaction terms are statistically significant, confirming the presence of context-specific synergies and trade-offs. This study contributes by endogenizing livestock integration within a dynamic GMM framework, and by generating stress-specific insights that can guide more targeted and resilient adaptation strategies for North African agroecosystems. The results offer generalizable implications for mixed farming systems operating under climate uncertainty.