17 January 1997


Are you planning to use precision farming to pep up the accuracy of your fertiliser applications? If so, there may be more

to consider than you first thought. Charles Abel reports the views of a leading independent consultant

TO make the most of what precision farming technology has to offer fertiliser users, growers will need to map with care, think carefully about the results and take full account of crop off-take.

That was the core message Chris Dawson had for delegates at the Fertiliser Societys annual conference in Cambridge last month.

As an independent consultant and member of the Shuttleworth Precision Farming Alliance – a group formed to promote precision farming throughout Europe – Mr Dawson has spent the past few years studying the new technology in both Europe and North America.

His conclusion is that the technology, which allows farmers to achieve very accurate positioning within a field when taking measurements and when making applications, does have a role in crop nutrition, but is likely to be exploited for base nutrients first.

The need for more accurate use of nutrients was clearly shown by a survey of 76 arable fields. That showed 57% of their area fell outside the desirable P index, 31% was outside the desirable K index and 37% outside the optimum Mg index.

Nutrients differ

But the way precision farming technology is used to improve fertiliser applications will vary, depending upon the nutrient in question. Mr Dawson differentiated between P, K, Mg and Ca which provide for both the current crop and the soil nutrient reserve and N, S and micro-nutrients which have more to do with the current years crop production only.

"We have to separate the correction of deficiencies from the feeding of the crop," said Mr Dawson. Variations in levels of the first group of nutrients which show up when the field is mapped can be corrected with one off treatments which last several seasons. Mapping for N, S and micro-nutrients will require in-season assessments of soil reserves and crop need, posing more of a challenge.

Separating the two groups will make mapping and treatment far easier, suggested Mr Dawson.

But variations which seem substantial, may have no impact on crop production, he noted. Unless the variation goes beyond the acceptable limits for optimum crop production there should be no need to make adjustments. To do so would result in an endless cycle of chasing insignificant variations which make no difference to the final crop performance, he warned.

Instead, by using robust GPS-assisted sampling, application maps and yield mapping, meaningful variability of base nutrients could be eliminated in a few years, he said.

But before kicking the GPS-directed fertiliser spreader into action to apply variable rates of base nutrients according to soil reserves one further factor needs taking into account. That is the variable removal of nutrients by the crop, which is itself probably the main cause of variable nutrient levels within a field in the first place, Mr Dawson suggested.

This will occur where higher yielding areas remove more nutrient, so depressing the soils nutrient status. Over time reserves in those areas could dwindle and so limit crop growth. There is growing evidence that the areas within fields which give the highest yields are often associated with the lowest soil nutrient reserves, particularly of phosphate, he said.

A time bomb

"This is a time bomb waiting to go off. It really is very critical. Without a precision farming approach this serious situation will never become apparent. If you only take one result from a field you will never know. It is like having one foot in the oven and the other in the fridge, your average temperature is acceptable. And that is the situation you may have in the field. There is no doubt that in some well farmed fields the problem is already occurring."

This is where yield mapping can show what is being removed. Other mapping systems – like aerial photography and satellite imaging – may give a good assessment of crop vigour, but are of little use in determining final nutrient off-take in crop yield, he stressed.

"If you dont replace what the crop is removing, then the nutrient status is going to get lower and lower on the high yielding parts of the field, because they are only getting an average return of P and K fertilisers."

By adding variable crop off-take figures to an application map, inputs can be better tailored to crop need across the field. The total amount of fertiliser used will remain the same, because the yield of the field as a whole has not changed. But the way it is distributed will be different, offering scope for a general uplift in yield, he suggested. &#42

Precise mapping of soil nutrient levels, combined with an assessment of crop off-take using yield maps, could help growers improve the precision of fertiliser use, says independent consultant Chris Dawson (inset).


&#8226 Map P, K, Mg and Ca content by soil analysis.

&#8226 Take multiple samples at each sampling point.

&#8226 Take account of crop off-take using yield maps.

&#8226 Only adjust inputs where variation is likely to limit crop production.

&#8226 N, S and micro-nutrient mapping tricky, needing in-crop measurements.

&#8226 Ensure spreader can work accurately at single rate before trying variable rates, otherwise errors could be compounded.


Collecting 25 soil cores across a field to create a single representative sample for analysis is the established method of getting an average figure for nutrient content of a field. That can be confirmed by taking 25 samples from a grid pattern and analysing them separately, as done in a trial on three fields last year, explained Mr Dawson.

The results showed a great range in nutrient content, but when averaged gave a figure similar to the first method.

"The excitement is the variation. But should it be taken at face value?" Mr Dawson suggested not. He argued that it is not valid to consider each of the individual samples as being representative of the soil at that point. Up to 16 sub-samples may need collecting at each point to get a representative picture, he suggested.

"This needs more study to come up with an appropriate protocol. It is very important not to make the wrong assumptions. The protocol we use to take the samples must be robust."

See more