On-farm field experiments for precision agriculture

Establishing simple, practical and meaningful on-farm field experiments is important for the future of PA. The process, at present, requires some form of stratification into potential management zones which partition soil and crop variation. The treatment levels, replication number and plot location for each zone can then be determined based on the economic impact on production and the spatial constraints of a minimum plot size and proximity to zone boundaries. A first tentative attempt at such a design is presented. The objective function is divided into two parts. Firstly, designs are selected that meet the economic criterion of x% penalty on expected profit. From those candidate designs the position of the plots are optimised in a biometric sense. An example for a 70 ha field in Eastern Australia is presented and the need for repeating the experiments with the same treatments in the same locations for a number of seasons is discussed.