The promise of precision agriculture is intriguing to many growers. But the cost of related technology and time invested in global positioning systems, yield monitors, variable-rate seeders and fertiliser applicators strike a proper note of caution.
At paddock scale, one of the most obvious benefits is ascertaining which parts are likely to be profitable and which are not. There are other factors, too. Where crop performance is low, water is not taken up and is likely to drain below the root zone, contributing to dryland salinity.
Precision agriculture might encourage changing paddock boundaries better to reflect economic and environmental priorities. The discussion of benefits and costs is ongoing.
Below, Simon Cook and Robert Corner, Precision Agriculture Research Group, CSIRO Land and Water, discuss how precision agriculture may help farmers design their own on-farm experiments.
Q: What do the yield maps show?
Yield varies widely
The first thing a grower gets from precision agriculture is a detailed map of crop yields. The second thing is a surprise — "I knew it varied, but not that much!" This leads directly to the third stage, "What's causing variation?" Sometimes the cause is obvious.
For example, Doug Maitland, on whose Wyalkatchem (WA) property the paddock (shown in map below) lies, knows that the low yield in some parts of this map is due to soil type. Other areas are associated with weed-control problems, and the low yield on the extreme right-hand side was due to competition from trees along the paddock edge.
Enter basic information about prices and costs into the computer and it's possible to convert the yield map into a map of gross margin (below). This contains even more surprises, and may show that while some parts of a paddock are performing well, others are losing money.
Effectively, the better areas are subsidising the poor. If these poor areas continue to underperform, why not simply remove them from cultivation?
A 'gross margin' map
This grain-protein map (above right) was produced from grab samples at harvest, but technology is being developed in the USA to measure protein content in the header. The map shows that quality also varies widely within the paddock. The grain delivered to the bin will be of average quality.
Perhaps a grower would do better to segregate the grain harvest to be sure of making the required grades.
Map of variation in protein content
Q What do the maps mean?
These maps are very interesting, but they really start paying their way only when they help a grower answer the following questions:
- Do I know what is causing yield variation?
- Will the same thing happen next year?
- Can I manage the paddock better as a result?
Sometimes, a map will have clear meaning, and serve merely to clarify a grower's impression. For example, Mr Maitland knew that the low-yielding strip down the side of the paddock was caused by competition from trees, and another patch by waterlogging.
These problems occur in most years. He doesn't need numerical analysis for these, just keen observation, common sense and experience. Often a grower can't be sure what causes crop variation and can only make an educated guess. Incorrect interpretation on the basis of one or two maps could be quite risky—especially since Doug Abrecht at Agriculture WA has shown that it takes a long time to get an accurate representation of seasons.
Yield reversals (that is, areas which do well one year and poorly the next) are quite common.
All this means that there is no 'quick fix' to getting the most out of these maps. What growers require is a progressive build-up of experience to increase their certainty about what the maps mean and what they want to do about them.
Q Why analyse the yield maps?
In most cases where the cause of variation is not obvious, it's worth considering a more analytical approach. For example, analysis of the yield variation in Mr Maitland's paddock (together with soil-test information) showed that the problem was caused not by poor nutrition, or soil water alone, but by poor rooting conditions associated with acidic soils.
It looked like a problem of water unavailability, but was caused by low pH. This illustrates the other advantage of analysis, namely that many of the yield variations are subtle and cannot be detected by merely looking at the maps. You have to get behind the maps and look at the numbers that go into them.
This is not trivial, but methods are being developed by researchers at CSIRO Land and Water and at the Australian Centre for Precision Agriculture (at the University of Sydney) that could be packaged in user-friendly decision support. Comparable products are already marketed in the USA.
Q How to become more certain?
There are a few simple methods of using these yield maps. None are 'instant' solutions, nor are they alternatives to sound management. But in the hands of a skilled manager they can help a steady improvement in profitability. Here are a few suggestions:
Our experience is that crop patterns fluctuate from year to year so that one year's observations may be an unsound basis for change. The proportion of paddocks consistently 'good' or 'bad' is fairly small, compared to the proportion which is good in some years, bad in others. This doesn't mean: "Don't do anything for five years", but rather "tread carefully".
Difference in wheat yield for two consecutive years
Experiment with your own paddocks
A simple way of finding out how the paddock is likely to respond to changes in management is to experiment — simply vary fertiliser, seed or spray and measure the effects. The experiment can be as simple as splitting the paddock into blocks or strips. If variable-rate technology is available, the experimental design can be more complex, such as 'checkerboard' or 'egg-box' experiments. The GRDC now has available a manual describing the nuts and bolts of experimentation (see Resources p23).
US experience suggests that improvement is stepwise, and that neither growers nor advisers have all the answers. There, a network of commercial advisers and extension agents is springing up to provide a forum for discussion that steadily increases growers' confidence.
Q What's the pay-off?
No change is without risk and the grower is entitled to ask "What's in it for me?" This is impossible to answer categorically. In some years, the opportunity to benefit is likely to be better than others. However, we know that the potential profit can be substantial.
In Mr Maitland's paddock we analysed an experiment which compared the effect of adjacent varying fertiliser application and produced a map of 'optimal' fertiliser rate, seen in hindsight, which showed that even in this poor year, the potential benefit of variable application was about $25/ha, more than enough to pay for itself.
Map showing optimal fertiliser application rate
Q What does it cost?
The equipment needed to start yield mapping is a yield monitor and a differential GPS. Most new headers have a yield-monitor option. To buy one existing header would cost around $8,000. ential GPS costs about the same and will ur an ongoing signal charge of about $2,500 per year. Yield monitors are generally supplied with software for generating simple yield maps.
More complex software such as a Geographic Information System are needed for serious data analysis. Access is often available through advisers and yield monitor dealers who offer bureau services.
Q Will it work for me?
No-one can promise certain fortune. Like all new information, it will take time to learn how to use precision agriculture most effectively. This will require commitment to change, and a plan of where the industry wants to go.
These can be offered only if growers feel confident that precision agriculture will enhance long-term profitability. The solution lies in a concerted effort from growers, researchers and commercial providers in the face of stiff competition for time, money and patience.
Program 3.5.3 Contact: Dr Simon Cook 08 9333 6200