Potassium myth, reality and decision support

Several combine harvesters working in concert.

Potassium (K), together with nitrogen, phosphorus and sulphur, is one of the four major nutrients required for healthy plant growth. But unlike the other nutrients, the grains industry relies almost entirely on soil reserves of potassium for production.

Large areas of Western Australia’s wheatbelt soils are naturally low in K and soil scientists have long realised that, in time, reserves would be depleted to the point where grain yield and grain quality might be affected.

New research by CSIRO and AGWEST and supported by growers and the Federal Government through the GRDC has shed new light on soil K and pointed the way to its profitable application.

Extracting potassium

The ancient sandy and duplex soils of the Western Australia wheatbelt were inherently low in K even when they were cleared for agriculture starting in the 1940s. There are exceptions where soils have formed on young rocks or the dolerite dykes that dissect the wheatbelt, but in general many soils have extractable K concentrations of 50 ppm or less and are potentially deficient.

Potassium levels in all cropped soils are decreasing for a number or reasons:

  • Inputs of K are very low — less than 1 kg/ha/year is added in rainfall and the store of slowly released K fixed in the clay minerals is very low compared to soils elsewhere.
  • Relatively large amounts of K are being removed in grain and biomass. One tonne of cereal grain contains 4 kg of K, compared to 8 kg for lupins and 9 kg for canola. The change to rotations including broadleaf crops has thus increased the rate at which K is being removed. Higher yields of all crops as a result of new varieties and improved management also contribute to greater loss of K. If crop biomass is also removed, the loss of K is much greater — about 12 kg of K/t.
  • Potassium is readily leached from coarse textured soils because it is not bound tightly to the negatively charged surfaces in the soil.

Better soil test, better yields

Economic use of fertiliser K depends on our ability to predict response to applied K. The CSIRO/AGWEST project measured grain yield responses at 19 sites on sandy duplex and deep sand soils. Colwell K, measured in the surface 10 cm, was shown to give a good prediction of wheat grain yield response to K fertiliser.

Grain yield was significantly increased by K application at 14 of the 19 sites and, from the fitted response curve, grain yield was 90 per cent of maximum when Colwell K was 40 mg/kg in the 0-10 cm layer and 60 per cent of maximum when Colwell K was 20 mg/kg.

Overall, a Colwell K of 35 mg /kg discriminates between responsive and non-responsive sites. Based on this value and the accumulation of information on soil levels, the researchers concluded that large areas of the wheatbelt are at risk of K deficiency.

Myths about subsoil K reserves

A ongstanding belief has been that K accumulation in the clayey subsoils of duplex profiles meant that these soils were unlikely to require K fertiliser. The research has shown that this K is only poorly accessible by the crop, probably because of low root density in the subsoil and the time delay until roots exploit this layer. The sandy topsoil is the major source of K and this is the reason the 0-10 cm soil test is effective even though K is readily leached down the profile.

Decision support

Mike Wong, CSIRO Land and Water, and his co-workers have developed a Decision Support System (DSS) for K application.

Fertiliser requirement depends on both the amount of nutrient available from the soil (before fertilisation) and the demand by the crop — larger crops will naturally require more nutrient uptake and place more demand on the finite soil supply. The DSS integrates all of these factors.

Using these relationships an estimate of K fertiliser requirement can be made for any level of yield given the soil test result. At present the DSS is available in the form of a ‘nomogram’ — a set of graphs from which the result can be read. Interestingly, it suggests that K requirement will differ between soft and ASW wheats. A computer model may be next.

Program 3.4.3

Contact: Dr Mike Wong 08 9333 6299 email m.wong@ccmar.csiro.au

Region North, South, West