Abstract Six fertiliser application scenarios were considered on a hill-country case study farm. The scenarios were developed for input into a decision tree model and were compared to a base line scenario that uses the farm’s current blanket application techniques. A further blanket application was considered along with several variable rate application scenarios. Throughout this study, variable rate application technology outperformed the fixed rate applications in terms of pasture production and fertiliser utilisation. One scenario, which used full variable rate application and a model optimised prescription map, produced the highest annual pasture yield across the 2518 ha Limestone Downs property, North Island, New Zealand, with 24 789 t DM, some 4854 t DM more than the property is currently modelled to produce. The blanket application techniques which are currently employed on this property produced the lowest annual pasture production with 19 935 t DM yr - 1. Variable rate techniques were predicted to increase annual production considerably, from 19 935 t DM yr - 1 to between 21 239 and 24 798 t DM yr - 1; in these cases, an increase in the production variability is also expected. Increased annual production from targeting additional fertiliser resources at more productive areas was identified through the decision tree modelling. Some areas were identified as having poor response to fertiliser, and fertiliser application to those areas was considered uneconomic.
Keywords agronomics; decision tree; fertiliser; GIS modelling; pasture production; topdressing; variable rate application
A06016; Online publication date 2 March 2007; Received 16 March 2006; accepted 17 October 2006
New Zealand Journal of Agricultural Research, 2007, Vol. 50:
53 - 63
0028 - 8233/07/5001 - 0053 © The Royal Society of New Zealand
2007
PDF file of entire paper: Print-quality (3814K) | screen-quality (494K)