The Impact of Precision Agriculture

Through the ages agricultural production systems have benefited from the incorporation of technological advances primarily developed for other industries. The industrial age brought mechanisation and synthesised fertilisers, the technological age offered genetic engineering and now the information age brings the potential for Precision Agriculture.

With the advent of tools such as the differential Global Positioning System (dGPS), Geographical Information Systems (GIS), and miniaturised computer components, agricultural enterprises are now capable of gathering more comprehensive data on production variability in both space and time. The desire (and ability) to monitor and respond to such variation on a fine-scale is the goal of Precision Agriculture.

This desire has both an economical and environmental basis. Matching inputs to crop and soil requirements as they vary within a field should improve the efficiency of resource use and minimise adverse environmental impact.

At present, monitoring and mapping the spatial variation in small-grain crop yields is receiving much publicity in Australia. Yield mapping is only one component of a Precision Agriculture system and small-grains is not the only enterprise to embrace the ideas. Crop yield monitors are also available for potato, peanut and forage harvesters and are under development for cotton, sugarcane and a range of horticultural crops.

The Precision Agriculture philosophy may be eventually applied to the spectrum of agricultural industries, for both quantity and quality control.

A Precision Agriculture System

There are 5 components to consider in the development of a Precision Agriculture system.

Spatial referencing

Gathering data on the pattern of variation in crop and soil parameters across a field requires an accurate knowledge of the position at which samples are taken. The dGPS network enables this information to be swiftly obtained with an accuracy here in Australia of approximately +/- 1 metre.

Crop & soil monitoring

Influential factors effecting crop yield, along with the crop yield itself, must be monitored at a fine-scale. Measuring soil factors such as texture, nutrient concentrations, pH etc. at present remains reliant on systematic manual soil sampling and analysis in the laboratory. Research is underway worldwide into real-time analytical soil sensors that will eventually automate the sampling and analysis procedures in the field.

Pest and disease dispersal along with crop growth indicators such as water stress can be successfully monitored using aerial or satellite photography in conjunction with crop scouting. In Australia two types of real-time small-grain yield sensor, measuring either volumetric or mass flow, are available from five manufacturers. This number will possibly double by 1998.

The total number of grain yield monitors operating in the country is below 200 at present. In the USA it is estimated by the manufacturers that between 5,000 and 10,000 units are operating, half with dGPS capability.

Spatial prediction & mapping

To produce a map of variation in soil, crop or disease factors that represents an entire field it is necessary to estimate values for unsampled locations. Various methods may be used for these predictions based on the values at the sampled locations. The most suitable methods for the various factors continues to be debated and the techniques refined.

Decision support

The degree of spatial variability found in a field will determine whether unique treatment is warranted in certain parts. Correlation analysis between the variation in crop yield and the measured factors influencing crop yield can be used to formulate agronomically suitable treatment strategies.

Differential action

To deal with spatial variability, operations such as fertiliser, lime and pesticide application, tillage, sowing rate etc. may be varied in real-time across a field. A treatment map can be constructed to guide rate control mechanisms in the field. Here in Australia there are presently three systems on the market that can integrate these operations and the number will continue to rise. The controller hardware is also available.

System Development

These components are at different stages of development and implementation. The technology required to gather detailed data leads the agricultural science of deciphering and applying the information it contains.

Technology

Ground positioning using dGPS receivers is well advanced and continues to increase in precision. Competition among an expanding number of GPS companies in Australia should also begin to reduce unit costs.

Crop yield monitors are considered very accurate at measuring the bulk yield of an entire field however less is known about the accuracy of the monitoring systems at the 1-2 metre level where individual yield measurements are matched with dGPS position. This contributes to uncertainty in the industry over the detail yield maps should attempt to display

Variable-rate controlling equipment is also well advanced with feed-forward times being reduced and rate changes becoming much smoother. Technological answers are less abundant in the search for information on what may be causing the observed yield variation. Data is required on the same scale as yield data (i.e. every 1-2m). This will eventually require sensors that either externally scan or invasively measure soil attributes as they pass in the field.

Agronomic Research

Here lies the greatest information gap. Scientists and commercial entities both in Australia and internationally are actively researching the causes of, and treatments for, the observed yield variation.

It is evident that grain yield can vary widely within a field and that the spatial pattern of this variation may change over time (Figure 1). This reflects interactions between influential field attributes and also between these attributes and the environment.

Figure 1. Wheat yield maps for 1995 and 1996, ‘Marinya’, Biniguy, NSW.

Figure 2 shows that spatial variability is also evident in a 1997 season cotton crop where irrigation usually mediates the significant environmental parameter of soil moisture.

Identifying a significantly yield limiting factor in one year may have limited bearing on the next growing season if its influence is considered singularly. Yield, soil, pest and environment variability data will have to be collected for a number of years (possibly up to 10 in highly variable environments) to adequately characterise and model this interaction.

In this manner a map of yield potential for a field may be constructed and then used each year in conjunction with early season environmental indicators and crop response models to guide differential actions.

Figure 2. Cotton yield for 1997, ‘Togo’, Narrabri.

Establishing a baseline understanding of the variability in yield potential within a field becomes essential if the most significant soil-based contributors to variability are shown to be difficult to manipulate.

Soil factors such as clay content and organic matter levels are known to contribute to nutrient availability and moisture storage capacity of the soil. They are also extremely difficult or impractical to amend in the short-term.

Our research has shown that the spatial variability in these two factors overwhelmingly affects the variation in sorghum yield in one northern NSW field. Intuitively, factors contributing to variability in the soil moisture regime will be important in the majority of cereal growing regions in Australia.

The more easily adjusted soil factors such as available nutrient levels and pH will also be important in many areas. However if the more rigid factors are going to limit yield then it would seem prudent to allow these to govern the application rates of any ameliorants in the field.

Precision Agriculture is not about treating a field to produce a uniform yield unless the potential is uniform. Its potential will be only be realised by acknowledging diversity in yield potential and environmental conditions when formulating field management operations.

Economics

The potential value of Precision Agriculture can best be displayed in a gross margin map (Figure 2). Uniform field treatment costs have been deducted from variable gross profit (yield x price). The 1996 wheat harvest produced a gross profit range between $A0/ha and $A560/ha at a mean of $A295/ha. Mean gross profit could have been increased with some form of differential treatment.

Figure 3. Gross margin map for 1996, ‘Marinya’, Biniguy, NSW.

Determining and attempting to manage variability in yield potentials will obviously raise the variable costs associated with sampling and amelioration. Estimates from the USA place this figure between $A12/ha and $A21/ha depending on the sampling detail. In Australia the projected cost would be between $A12/ha and $A63/ha due to greater unit sampling and analysis costs.

However, the economics of improved environmental stewardship does not easily fit the standard accounting paradigm. The allocation of monetary value to environmental gain is a fledgling science. Payments for positive actions or fines for deleterious actions could be accommodated, but at present Australia has no such remunerative or punitive legislation in place. It is apparent that Europe and the USA are moving in this direction.

Risk Assessment

The improved production information gathered using Precision Agriculture techniques also provides an ideal tool for risk assessment in potentially poor growing seasons. For example, well documented areas of low yield potential may be removed from production or have their inputs reduced to minimise potential financial losses. Such assessments would form part of the decision-support system, so that management actions may be used to disperse or lower production or capital risks across a whole farm.

Education

As with the introduction of all new approaches to crop production, education plays a pivotal role in its widespread adoption. Within the farming community, the main source of Precision Agriculture information has been the marketers of technology, and not agricultural systems managers or recognised educational bodies. The main reason for this being the as yet mimimal agronomic research being performed here in Australia. It is vital that the technology is utilised in an efficient systems approach that is suitable for the Australian environment.

This type of ‘high tech’ approach will probably see the advent of skilled consultants catering for a number of enterprises. Tertiary education will be required to train such people.

Politics

There is still not as yet a strong Precision Agriculture movement in Australia, driven by the economic-environmental imperative, as in the US and Europe. We anticipate legislation such as the 1996 US Farm Bill to expedite research and development.

Conclusions

Information is an economic necessity in any productive industry. The technology is now becoming available to monitor agricultural input/output at an increasingly detailed level. At present, it is necessary to gather data on output to characterise the variability that may be expected over space and time. Understanding the causes will be more difficult at this scale and require committed research from the agricultural industry and improvements in soil sampling and analysis technology. Ultimately these will be available but the impact of Precision Agriculture in Australia will depend on ensuring only suitable techniques are adopted within a fertile research, educational and political framework.

Acknowledgements

We wish to acknowledge the financial support of the Australian Research Council, and the Cotton Research and Development Corporation who provide scholarship and support for Brougton Boydell tenable at the CRC for Sustainable Cotton Production. We would also like to express our appreciation to National Mutual Cotton, ‘Togo’ and thank Craig and Judy Boydell, 'Romaka', Biniguy, NSW, for their considerable assistance .

THE UNIVERSITY OF SYDNEY
© 2008 - Australian Centre for Precision Agriculture