Timothy Gulden, Chenna Reddy Cotla, Atesmachew Hailegiorgis and Andrew T. Crooks
This paper describes a preliminary, lightweight and robust method for simulating vegetation growth that has basic validity in the face of remotely sensed vegetation data while being simple enough to retain conceptual and computational tractability when it is incorporated into a large agent-based model of human subsistence, conflict, and displacement in East Africa. The sub-model predicts daily vegetation values for 2.5 million 1km2 land grid cells using remotely sensed monthly rainfall data. It has been informally validated against remotely sensed, bi-monthly normalized difference vegetation index (NDVI) data. We believe that the approach presented in this sub-model is uniquely well suited to representing dryland vegetation dynamics within a context of a large, human-environment interaction model such as the RiftLand model of which it is part.