From Knowledge Transfer to Operational Competence: A Co-Produced Training Model for Early Warning in West Africa

Operational Numerical Weather Prediction (NWP) remains a central yet fragile component of early warning systems in many National Meteorological Services (NMSs) of Sub-Saharan Africa. Although access to global models, regional configurations and cloud-based computing has expanded, the effective use of these tools in routine operations is often inconsistent. This situation highlights a critical gap between acquiring knowledge and technical skills in NWP, and developing operational competence in configuring and running models, parameterising model physics, and interpreting, verifying and integrating forecasts into warning processes. While knowledge and skills can be transferred through short technical courses, competence—according to the World Meteorological Organization competency framework—requires demonstrated performance in real operational settings, under institutional constraints and accountability conditions.

This contribution presents and critically analyses a competency-oriented training model in operational NWP for flood early warning developed by the WMO Regional Training Centre in Italy, held by the Institute for Bioeconomy of the National Research Council, in partnership with the NMSs of Niger and Burkina Faso. The initiative reinterprets competency-based education not as a sequence of discrete courses, but as a co-produced institutional transformation process in which bottlenecks along the forecast chain were jointly analysed and learning priorities collectively defined. Model configuration challenges, verification procedures, and forecast interpretation and comparison practices were iteratively refined through shared experimentation. Training content therefore emerged directly from real service-delivery constraints.

The long-term embedding of West African forecasters within the operational team of the Tuscany Regional Meteorological Service represented the core mechanism of this approach. By participating in daily operations—model runs, post-processing, verification analysis and warning formulation—trainees developed competence through accountable practice rather than simulated exercises. A training-of-trainers logic enabled returning forecasters to adapt and replicate these co-developed practices within their home institutions, strengthening institutional autonomy and internal mentoring capacity.

Co-production also reshaped the collaborative dynamic. Continuous exchanges supported the emergence of a transnational community of practice in operational forecasting, within which complementary expertise was recognised and mobilised: the Italian service contributed experience in model configuration and verification, while West African forecasters brought essential knowledge of tropical dynamics, local observational constraints and impact contexts. This mutual recognition reduced dependency patterns and fostered shared ownership of operational solutions. Monitoring and stakeholder feedback indicate improved forecasting autonomy and greater stability in NWP-based early warning services.

The study argues that sustainable NWP implementation in resource-constrained environments depends on systemic and relational processes rather than on isolated individual learning or tool transfer. Co-production—through shared problem framing, iterative feedback and mutual accountability—can bridge the gap between knowledge and skill acquisition and the long-term transformation of operational early warning services.