Coupled biophysical and decision-making processes in grassland systems in East African savannahs – A modelling framework

Increasing livestock densities on limited grazing areas in African savannahs lead to resource degradation through overgrazing, aggravated by drought. Assessing herd management strategies over longer periods at landscape scale is important to propose options for sustainable land use. This requires an understanding of processes related to hydrology, nutrient cycling, herd movement, pasture degradation, and animal resilience that involve dynamic soil-plant-animal interactions and human decisions about stocking rates, livestock purchases and sales.

We present the coupled model system MPMAS-LUCIA-LIVSIM (MLL), the combination of a spatially explicit agent-based model for human decision-making (MPMAS), a spatially distributed landscape model for water flows, nutrient cycles and plant growth (LUCIA), and a herd model (LIVSIM) representing grazing, body weight, nutrition and excreta of individual animals. MLL represents daily vegetation growth in response to grazing and organic inputs, monthly animal performance influenced by forage availability and quality, and herders' management in response to resource status. New modules for selective grazing, resprouting of pasture, herd movement and model coupling were developed for MLL.

The test case of a pastoral system in the Ethiopian Borana region demonstrates the capabilities of MLL to simulate key soil-plant-animal-human interactions under climate-related management scenarios with varying access to grazing land, changing cattle prices and different spending / saving behaviour of herders. 20-year simulations showed the negative impact of consecutive drought years on vegetation biomass, on herd development and movement and how reserving grazing areas for dry seasons could mitigate overgrazing and improve income. Seasonality and drought response of vegetation growth, selective grazing of different plant parts, resprouting after grazing, calving intervals, milk yields and lactation in response to forage supply and quality as well as herder reactions to shocks were plausibly represented.

Building upon this successful proof-of-concept, MLL can be used to identify robust management options for improved grazing systems in savannahs in follow-up research.