Agroecological Food Systems Transition Architecture (AFSTA): From Evidence Synthesis to Modelling Transboundary Climate Risks and Food-System Transition Pathways in Africa

Technical Summary: Agroecological Food Systems Transition Architecture (AFSTA)

Problem Statement
Climate-related risks to food systems extend beyond the immediate location of a climatic hazard, propagating through production changes, resource availability, markets, and trade interdependencies. While existing analytical traditions offer robust methods for examining climate impacts, food-system transformation, agroecological transitions, resilience, and network transmission, these processes are frequently analyzed in isolation. A critical gap exists in connecting climate forcing with production responses, network-mediated transmission, transboundary effects, transition pathways, and multidimensional outcomes within a single analytical structure. Current models often fail to represent how transition pathways alter the very network configurations through which subsequent climate shocks are transmitted.

Methodology
The study employs a structured narrative evidence synthesis to develop the Agroecological Food Systems Transition Architecture (AFSTA). The synthesis is organized around six interconnected analytical domains:

1. Climate hazards and food-system shocks: Distinguishing between the initiating climatic disturbance and the resulting shock to production or system conditions.
2. Production and resource responses: Mediating mechanisms involving water, land, soil, inputs, management, and adaptive capacity.
3. Trade, markets, and network transmission: Analyzing how production shocks are buffered or transmitted through trade relationships and network structures.
4. Cross-border and cascading climate risks: Tracing direct, transmitted, and cascading effects across connected systems.
5. Sustainable and agroecological transition pathways: Representing transitions as changes in system configuration rather than fixed intervention coefficients.
6. Trade-offs, synergies, and resilience outcomes: Assessing multidimensional outcomes including food security, livelihoods, environmental performance, and resilience.

The methodology distinguishes between three evidence types: foundational (concepts and relationships), empirical (observed relationships, with emphasis on African systems), and methodological (modelling approaches). The synthesis utilizes structured narrative comparison rather than statistical aggregation, identifying established relationships and missing connections to define modelling requirements. The architecture is applied comparatively to Kenya, Mozambique, and Senegal, which serve as differentiated national configurations within the same analytical framework to illustrate how the common architecture operates under varying climatic, production, and institutional conditions. Parameterisation, calibration, and validation are identified as subsequent stages rather than components of the current study.

Key Contributions
The primary contribution of this work is the definition of AFSTA, an evidence-derived architecture designed to formalize the relationships between climate forcing, production responses, network transmission, transition pathways, and outcomes. Key conceptual contributions include:

• System Boundary Definition: The system boundary is functional rather than strictly geographical, allowing external countries, markets, or commodities to enter the model where they constitute material transmission pathways.
• Transition as Configuration Change: Transition pathways are defined as changes in food-system and network configuration that modify the conditions under which subsequent climate shocks are absorbed and transmitted, rather than as static scenarios.
• Distinction of Risk Types: The architecture explicitly traces the sequence from initiating hazard to direct effect, transmission pathway, receiving-system effect, secondary response, and cascading outcome.
• Modular Integration: AFSTA does not replace existing computational methods (e.g., climate-crop models, economic equilibrium models, network models) but provides the analytical basis to connect them. It specifies the relationships required for formalisation, parameterisation, calibration, and validation.

Results
The synthesis identifies that the principal modelling problem is one of connection rather than replacement. The evidence confirms that:

• Climate hazards generate differentiated food-system shocks depending on exposure and system conditions.
• Production responses are mediated by resource conditions and adaptive capacity, acting as an intermediate mechanism between disturbance and shock.
• Markets and trade function as a transmission layer whose structure can change with production and transition conditions, rather than remaining a fixed background network.
• Transition pathways can alter production diversity, resource efficiency, and market dependence, thereby changing the system's susceptibility to future shocks.
• Outcomes such as food security, resilience, and environmental performance often involve trade-offs and synergies that must be assessed separately before higher-order interpretation.

The comparative application demonstrates that the same AFSTA architecture can accommodate differentiated national settings (e.g., Kenya's devolved governance and diverse systems, Mozambique's smallholder focus and cyclone exposure, Senegal's mixed crop-livestock systems) without assuming identical responses to climate forcing or transition pathways.

Significance and Limitations
The paper positions AFSTA as a necessary analytical foundation for subsequent formal modelling of transboundary climate risks and food-system transitions in Africa. Its significance lies in providing a common architecture that links climate risk, production, networks, transition, and multidimensional outcomes, addressing the "connection problem" where existing models operate in silos.

The authors explicitly state that AFSTA is not an executable model and does not present model-generated pathway results. It lacks empirical parameterisation, calibration, and validation, which are identified as subsequent tasks. The study acknowledges limitations regarding the heterogeneity of evidence in spatial and temporal resolution and the uneven availability of African data across countries and food-system components. Consequently, the architecture serves as a framework for future reproducible variable definitions, functional relationships, and scenario assessments, rather than a predictive tool in its current form.