Stochastic simulation model of Ankole pastoral production system: Model development and evaluation
MetadataShow full item record
Mulindwa, H., Galukande, E., Wurzinger, M., Ojango, J., Okeyo, A.M., Sölkner, J. 2011. Stochastic simulation model of Ankole pastoral production system: Model development and evaluation. Ecological Modelling 222(20-22): 3692-3700
Permanent link to cite or share this item: http://hdl.handle.net/10568/10485
In the Ankole pastoral production system animals are grazed on pasture all year round. The cattle are not supplemented with conserved pasture or commercial feed except minerals. The large number of factors that influence production makes it impractical and expensive to use field trials to explore all the farm system options. A model of a pastoral production system was developed to provide a tool for developing and testing the system; for example, drying off animals early and supplement them for quick return on heat, testing the economic and ecological viability of the different stocking rates. The model links climate information, on a monthly basis, with dynamic, stochastic component-models for pasture growth and animal production, as well as management policies. Some of the component models were developed and published by other authors but are modified to suit the Ankole pastoral conditions. The model outputs were compared with on-farm data collected over 3 years and data collected for other on-farm studies in the region. The relative prediction error (RPE) values for body weight after weaning across both breeds ranged from 3% to 12% which is below the acceptable 20% and means that the model predicts post weaning growth with an average error of 7.5%. The model predicted pasture production and milk yield across seasons with relative prediction errors of 17.6% and 3.33%, respectively. The graph shapes of actual and predicted average daily milk yield as influenced by season (month of the year) were similar. Because pasture growth and milk production predictions were acceptable, economic predictions can be made using the model to test different management options such as seasonal breeding, alterations in lactation length and determination of appropriate off-takes and evaluation of economic viability of various stocking rates.