Abstract

The Stockpol model was designed for decision support on New Zealand sheep and beef cattle farms. It indicates the biological feasibility of a livestock system and provides the opportunity for users to test alternative livestock policies. An evaluation of the Stockpol model was made using 44 data sets from farmlet trials at Gore, Ballantrae, and Whatawhata Research Centres and data from Massey University's Riverside farmin the Waiarapa. These data covered a range of sheep and beef finishing systems. The main test compared the measured pasture growth rate with that required by Stockpol to predict the monthly pasture cover recorded in the data. Measured and required pasture growth rates differed significantly (<0.01) in July, August and September. The model was conservative and required an average of 5.1 kgDM/ha/day more growth over these months than was measured. In other months the differences were small and non significant. When animal demand was adjusted to mimic pasture cover, adjustments of minus 22% from July to September were required. Overall it was considered tha Stockpol can be used with confidence for analysing options for farmer decision making. Areas for further improvement of Stockpol have been defined as the pasture growth model in early spring. Aspects of the animal demand in early spring also need to be tested more vigorously. The evaluation emphasised the importance of consistent input data for use in models. For example, the pasture measurement technique was different at each of the four sites studied. This may explain some of the variation between predicted and measured data.

CA, Morris, NG Cullen, and GC Upreti

Proceedings of the New Zealand Society of Animal Production, Volume 55, , 101-103, 1995