6 December 1996

Mapping out a yield pattern is easy with GPS

AFTER three years of using global positioning systems for yield mapping, Suffolk farmer Stuart Scarff has accumulated a large amount of data and has increased his confidence in the potential of precision farming.

He has already moved on to the second stage by using GPS for soil sampling and fertiliser application.

Mr Scarff farms 260ha (650 acres) of medium to heavy clay land at Fenns Farm, Combs, Stowmarket. All the farm is down to arable combineable cropping – 90% of crops are grown on contract for seed production. Wheat is the main crop with rape, peas, linseed and herbage seed included in the rotation.

Yield mapping covers the whole farm, with the data collected by a Massey Ferguson combine linked to the Focus FM broadcast network. Data is transferred by smart card to the office computer and down-loaded using Windows software.

Although it takes several years of yield mapping to iron out the variations caused by seasonal factors, Mr Scarff is confident that the data he has now built up will be increasingly useful as a management aid.

"The maps we made during the first year of yield mapping showed yield variations of 4t/ha, and even more in some fields," he says. "I realised there must be real possibilities for using this information to improve efficiency. We now have yield maps covering three years and we have reached the stage when we can identify consistent yield patterns in some of the fields."

The patterns show up in a comparison of the three years of maps for each field – another approach is to merge the three years of results into a single yield map.

As the rotation means different crops may be grown in each field during the three years, a direct comparison of yield in different years would be misleading.

To overcome this problem, the yield figures for each crop are expressed as a percentage of the top yield for each field and the percentages are entered on the map to average yields for different crops covering the three years.

The objectives as the yield patterns become clearer are to bring some of the lower yielding areas closer to the maximum in each field, and to reduce input levels and costs on areas identified as naturally low yielding due to soil type.

First stage was to take additional measures to deal with the headland compaction identified as a reason for reduced yields. A 12m (39.4ft) wide strip is subsoiled around each field as part of the post-harvest routine.

The next stage in developing precision farming techniques at Fenns Farm is to use GPS on other equipment and Mr Scarff has already made a start on this. He has equipped an ATV with a GPS unit to enable him to perform his own soil sampling.

Sampling achieved to date has been based on a 50m (164ft) grid – which he admits is time consuming and expensive, but it should be more accurate than the usual interval of 100m (328ft) or more. Mr Scarff believes the extra cost can be justified on the assumption that the sampling will not need to be repeated for at least five years.

Significant variations in soil analysis have been identified – particularly in terms of phosphate levels. To make use of the soil maps, Mr Scarff has started using an Amazone spreader linked to the Fieldstar GPS control on an MF tractor. This can target specific areas and he reports that he has made savings on phosphate and potash applications this autumn.

A GPS-linked seed drill is also included in the precision farming plans for the future. This will allow Mr Scarff to increase the plant population on some of the lower yielding areas in a bid to improve yields.

"We have learnt a lot in three years," he says. "We are beginning to get a clear pattern from the yield maps and can identify the areas needing special attention.

"However, one of the problems we now have is to decide when we should increase or decrease inputs on low yielding areas. In some cases we can only decide that by collecting more information and probably by some trial and error as well."n

Three years of yield mapping has enabled Stuart Scarff to build up a significant amount of data. Headlands are now subsoiled annually.

One of Stuart Scarffs yield maps. Note the yield variation from below 7.6t/ha to as high as 12.4t/ha.