Ground Cover Supplement : GC Supplement - Precision Agriculture
GROUND COVER PRECISION AGRICULTURE 6 ADOPTION Applying a few simple rules ensures results from precision agriculture trials are of value By Roger Lawes PRECISION AGRICULTURE (PA) technology can be used to run on-farm paddock-scale trials. In effect, one new technology can be used to test another. However, trials are only useful if they are correctly designed and implemented. In 2008 CSIRO, the Department of Food and Agriculture, Western Australia (DAFWA) and the Liebe Group ran a series of trials using quite complex designs. They ended in disaster because the trial design failed to account for the difficulties created by a 15-metre seeder bar, a 30-metre boomspray and a harvester with a 10-metre front. The research team set about simplifying the on-farm trial process. Our objective was to create a design that was simple and worked, so new technologies could be tested with farmers' equipment. The approach outlined here was successfully implemented by growers in 2009. 1Fewer rather than more treatments It is extremely important to ask an appropriate question and put it to the test. Conventional experiments can be complex. For example, scientists may evaluate multiple rates of fertiliser on multiple crop species. In on-farm experiments, where treatments must fit in with a commercial operation, we recommend restricting the number of treatments to just one or two at most. The remainder of the paddock should be thought of as a control. A simple trial that generates a definitive outcome is better than a very complex and time-consuming trial that confounds the issue. Trials not tribulations the measurement scales at which the majority of spatial variability in crop and soil properties occurs. New methods of analysis are being used to break the data down into components of variability that range between 'noise' at one extreme and large- scale trends at the other. If most of the variability occurs closer to the noise end of the scale then field measurements need to be made at a finer scale. Alternatively, where much of the variability is captured at the trend level, field measurements can be made further apart and still be useful. As the focus of PA management begins to move from broad within-field management classes to finer scales, a better understanding of the minimum detail that should be measured for different crop and soil properties will improve our ability to build accurate agronomic links at this scale. While this work should help us determine which of the data layers that we can now gather at fine scale are the most useful to help predict yield and nutrient requirements, it will also provide estimates of the maximum measurement resolutions. A spin-off from these methods of analysis is that we can provide a map of data that shows variability above a certain scale. Such maps would provide growers with a level of detail in variability to match the operation specifications of application implements, or remove noise to a defined level of variability for directing nutrient replacement operations. Gathering quality PA data is the starting point; our work is aiming to help growers use this data to make robust management decisions that improve profitability and sustainability. □ GRDC Research Code US00044 More information: Dr Brett Whelan, senior research fellow, Australian Centre for Precision Agriculture, 02 9351 2947, email@example.com www.grdc.com.au/US00044 Researchers Brett Whelan (right) and James Taylor (left) discuss the integration of data layers with South Australian grain grower Malcolm Sargent. On-farm trials using commercial equipment and PA technology need to be simple, with one or two treatments each running the whole length of a paddock for two seeder-bar widths.
GC Supplement - Biosecurity
GC Supplement - Climate