Integrated study and modelling of the factors affecting small-scale subsistence farming in eSwatini (southern Africa) through the use of causal diagrams and agent-based models.

The dependence of small-scale subsistence farmers on natural resources highlighted the need of a transition towards a more sustainable management of resources. An extensive literature is available on the structural factors affecting small-scale agricultural production in developing countries, while few studies are available on the role of social dynamics on the capacity of the system to react to adverse conditions. What are then the social, economic and environmental factors affecting small-scale subsistence agriculture? and, what is the effect of social behaviours on the adoption of agricultural innovation? This study is based on the specific case of eSwatini (southern Africa). A causal loop diagram was adopted to show the cause–effect relationship between variables. The role played by social behaviors in the diffusion of innovation is explored with the use of agent-based modelling (ABM). The model is based on agents, environment, links and five processes: crop production, social learning, individual learning, decision making, and resource recovery. In the simulation three farming behaviours can be adopted: no cropping, traditional practices or conservation agriculture (CA). The ABM was implemented in the Netlogo 6.2.0 platform. The case study is supported by scientific literature and interviews with local stakeholders. The results highlight the role of gender equality in the diffusion of knowledge. The model was run with and without gender equality for 10, 25 and 50 cycles representing farming seasons. The results show an increase in the spread of CA in the scenarios with no gender discrimination for which the rate of innovation adoption is faster. Moreover, the study casts light on the role of social learning and its dependence on training and education centres for the diffusion of new behaviours. Also, results show how an increase in the number of learning centres leads to a higher rate of knowledge diffusion