Simulation and quantification of pasture condition and animal performance on boreal grasslands in Alberta

Abstract We developed a computer model (PASTURE) to simulate both condition of pastures grazed by farmed wildlife and animal performance as dual measures of system performance. For representative habitat types (e.g., aspen forest and open grasslands), the model simulates herbage growth, maturation and senescence, habitat and diet selection, and nutrient intake of several wild ruminant species. The model generated seasonal patterns of pasture productivity and animal behavior expected for Bromus-Poa pastures in aspen boreal ecosystems. Model calibration was conducted using data collected from studies conducted within the aspen boreal ecosystems at the Ministik Research Station in 1997 and 1998, and Elk Island National Park and the University of Alberta greenhouse in 1999 and parameters adjusted until the predicted and observed results were visibly close. Evaluation of the model was conducted using statistical criteria of calculating average error, root mean square, root mean square error, coefficient of residual mass and modeling efficiency and comparing these statistics against optimal values. Statistical tests indicated a good fit between predicted and observed values for cumulative phytomass, herbage use and animal performance.