Developing variable rate application technology: Modelling annual pasture production on hill country

A great deal of modelling work has been completed in New Zealand and overseas on factors contributing to variations in pasture production. Although there are many factors that affect this, some have far greater influence than others. This paper discusses some of the important governing factors that previous researchers have identified, and also comments on the modelling techniques used by others in attempting to model annual pasture production. One of the more adaptable techniques, commonly called decision tree modelling, has been selected and applied to a large sheep and beef, hill‐country farm. The resulting annual pasture production was compared with an animal intake model. Both approaches were in very close agreement. This indicates that the decision tree method could be a very useful tool in the management of hill country properties when linked with a geographical information system. Embedding a pasture production model within a geographical information system is useful as it helps to identify strategic opportunities for either increased production or greater efficiency of inputs, particularly if within‐field variation is scrutinised.