Agent-based model to analyse human-drought feedbacks in dryland regions

The Horn of Africa Drylands (HAD) region is vulnerable to climate-related risks, including droughts. Drought is a persistent and costly hazard; it negatively impacts the environment, people’s livelihoods, access to education, and food security. The rural livelihoods of countries in the HAD region are highly dependent on rain-fed agriculture and livestock herding (agro-pastoralism), which is extremely sensitive to weather and climate variability. Communities can take adaptation decisions that influence the impacts of drought. In turn, adaptation behaviour by agro-pastoralists influences the drought hazard. However, only a few drought models incorporate these feedbacks in drought analysis. This research presents a dynamic drought adaptation modelling framework that combines socio-hydrological and agent-based modelling approaches. By coupling the spatially explicit hydrological Dryland Water Partitioning model (DRYP) with a behavioural model capable of simulating (bounded rational) behavioural theories, the model enables simulating the feedbacks between the environment and agro-pastoralists and amongst agro-pastoralists (both through ‘communication structures’ and indirectly via their water and land use). Qualitative and quantitative data inform the model parameterisation on adaptive decision behaviour of agro-pastoralists, applied to the upper Ewaso N’giro river basin in Kenya. We demonstrate the possibility of capturing adaptive human-drought feedbacks, laying the foundation for in-depth analysis of human-drought feedbacks and future climate and policy scenarios.