Hybrid and Uncertainty-Aware Modelling for Climate-Smart Agronomic Systems: From Soil and Water Processes to Pest and Crop Dynamics

Climate change is fundamentally reshaping agronomic systems through interacting pressures such as rising temperatures, altered rainfall patterns, soil degradation, and increasing frequency of extreme events. These changes affect not only crop productivity but also soil processes, water availability, irrigation efficiency, and the dynamics of pests and vectors. In smallholder systems across the Global South, these impacts are particularly pronounced, as limited access to forecasting tools and adaptive technologies constrains timely and effective responses. Despite rapid advances in modelling, many current approaches remain fragmented. Correlative models often capture statistical relationships but may fail under non-stationary climate conditions, while mechanistic models better represent biological and physical processes but are often difficult to scale or generalize. There is therefore a growing need for hybrid modelling frameworks that integrate mechanistic understanding (e.g., crop growth, pest life cycles, soil–water interactions) with data-driven approaches, including machine learning and remote sensing. This Research Topic seeks contributions that advance robust and reproducible modelling practices for climate-smart agronomy, with particular emphasis on integrating multiple components of agricultural systems. We welcome studies addressing soil–water processes, irrigation and water-use efficiency, crop productivity, pest and vector dynamics, and their interactions under changing climatic conditions. Contributions that bridge disciplines, linking agronomy, ecology, hydrology, and data science, are especially encouraged. A central focus of this topic is the explicit treatment of uncertainty and bias in agronomic modelling. Climate projections, model structure, parameter estimation, and spatial bias all contribute to uncertainty that can propagate through modelling workflows. We therefore encourage approaches that incorporate ensemble modelling strategies, including both multi-algorithm ensembles (e.g., combining statistical and process-based models) and multi-GCM ensembles to capture variability across future climate scenarios. Methods that quantify agreement, identify uncertainty hotspots, and improve the interpretability of projections are particularly relevant. We also highlight the importance of climate attribution and scenario-based analysis, enabling the separation of climate-driven effects from other environmental and anthropogenic influences. Such approaches are critical for moving beyond prediction toward actionable insights that support adaptation planning, policy development, and investment prioritization. Finally, this Research Topic promotes transparent and scalable modelling workflows, including improved validation strategies, bias correction methods, and the integration of Earth observation data and decision-support tools. By bringing together advances in hybrid modelling, uncertainty quantification, and system-level integration, this collection aims to contribute to the development of next-generation agronomic modelling frameworks that are both scientifically rigorous and operationally relevant. Topic editors: Komi Mensah Agboka, Siyabusa Mkuhlani, Luca Rossini Source: https://www.frontiersin.org/research-topics/80686/hybrid-and-uncertainty-aware-modelling-for-climate-smart-agronomic-systems-from-soil-and-water-processes-to-pest-and-crop-dynamics