10 November 2001

Why so variable?

Can we do enough about in-field variability in crop yield to justify precision farming? And is there a cheaper way to measure it than grid soil sampling? Sarah Henly finds out

THE soil auger could have had its day for precision agriculture. Within 10 years it might have been replaced by information from sensors, believes Dr Margaret Oliver of the University of Reading.

The tool that once offered the only solution to finding out about soil properties is being superseded by sensor equipment in the drive to improve crop yields. Aerial photography, satellite imagery, EMI scanning and yield mapping are now providing cheaper information about within-field variation, enabling more targeted soil sampling and management, she explains.

"Precision farming requires an understanding of how the soil varies within a field, but one of the stumbling blocks to this technology is its cost. Sampling the soil to determine the nutrient content is extremely expensive, and is wasteful if the variability is insignificant, or there are other underlying causes that cannot be ameliorated. What we need is a low cost method of assessing variation to back up growers generally good knowledge of their land."

Contour maps based on the conventional sample of one core a hectare can be misleading because they are based on too little information, stresses Dr Oliver. In addition, the cores sample just the topsoil and we have no idea of the soil conditions at depths where crop roots penetrate.

She believes aerial photography offers a good starting point in the search for reasons why crops perform better in certain areas of the field.

The technology is not new. Aerial photographs taken up to 15 years ago both of bare and cropped land are being used to find permanent patterns of variability in the trial. These are being compared with slope, soil depth, stoniness, texture, organic matter, moisture or other factors to assess what factors other that nutrients are responsible for the variation.

"Its short-sighted to assume that the soil nutrient status is to blame for all variations in cereal yield. We usually find that high yielding areas have low nutrient levels, due to greater uptake, and these should have greater applications rather than the low yielding areas. The nutrient status if often deceptive and can fluctuate. Aerial photography looks at permanent patterns."


Factors such as soil texture and moisture availability are very important to yield. For example, clay soils often support better yields in dry seasons than wet ones. In wet years, the top of a sloping field may produce better yields than the bottom. Soil texture and moisture, often based on landscape characteristics, cannot be altered easily, therefore it is futile to pay to have a detailed grid assessment of the soil carried out, stresses Dr Oliver.

Techniques that measure the moisture holding capacity of soils, such as electromagnetic inductance scanning (EMI), give long-term insight into the soils ability to support a particular yield. This non-intrusive sensor, which can be towed behind a quad bike, is used widely in the US and Australia to identify salinity problems. It identifies the patterns of variation within a field, most likely in relation to the clay content and moisture content. The different patches in the pattern might have areas with low yield one year and high another – related to weather conditions.

If we can find out why a particular soil characteristic limits yield, through low cost sensor information, we can use that information to improve crop yields and minimise the environmental impact of fertiliser use."

Last but not least, yield mapping offers some very useful pointers to soil problems, according to a complementary project at SRI. Project leader Dr Murray Lark recognises it has limitations, but believes it can identify enough about soil variation and its effect on crop yield to save on whole farm soil sampling by targeting it more efficiently

"We are analysing yield maps to devise rules for identifying fields where the variation is large enough to justify further investigation. Then we are grouping together field zones which perform similarly and differently. Once we have located such zones, targeted soil sampling will help identify the soil factors which are limiting crop yield."

Yield mapping isnt precise enough to pick up changes within a distance of about 20m, but it could be used to highlight where further investigation is needed, perhaps by targeted grid sampling.

The project findings will help ADAS assess the scope for precision management. For example, when the problem looks to be low pH, growers need to know whether or not it is cost-effective to apply lime to selective areas only.

"Our work shows there is value in the information we can get relating to yield and soil type or moisture capacity, which could save growers money by allowing them to target inputs more precisely. My advice is to use yield maps to gain a long-term picture of variability, even if you dont intend to use a variable rate spreader or sprayer."