System dynamics and hierarchical Bayesian inference to model Indigenous small-scale agri-food systems in Guatemala

Abstract:

System dynamics (SD) modelling are a useful tool for modelling complex interactions and feedbacks between human and environmental domains in the context of natural resource management. While a large number of studies have used SD models built with stakeholders as decision-support tools in the hydrology and natural resource management fields, their application to small-scale agriculture, especially in rural and Indigenous regions, is complicated by the lack of data pertaining to many environmental and especially socioeconomic variables key to the model's structure.

In this presentation, we present a novel method of combining stakeholder-built system dynamics models with a hierarchical Bayesian inference calibration algorithm for the calibration of uncertain empirical equations between socioeconomic variables of interest. Stakeholders contribute mostly to the determination of the causal loop and feedback structure of the model, while the inference algorithm uses spatially explicit databases (such as are often available from many governmental surveys) with low temporal resolution in order to substitute space for time during the calibration and validation phases of the model. Results from a case study of agricultural production and food security of different Indigenous agricultural systems in Tz'olöj Ya' and K'iche' (Guatemala) show that the model was successfully calibrated with this approach, while spatial validation indicates that the model does indeed show better performance in regions similar to the municipalities wtih whose stakeholders' help it was constructed. The system dynamics model was then coupled to an external cropping model and used in scenario analyses of community resilience to climate change and adverse socioeconomic conditions. The results show that socioeconomic interventions, such as universal education and minimum wages, are very effective at improving outcomes for both social (food security and poverty) and environmental (forest cover) indicators of system sustainability.