27 Jan 2014

Do we need to revisit potassium?

Paper presented at the Adelaide GRDC Advisor Update
Chickpea - K alone

Potassium in soils and plants

As management improves and yields increase, the extraction and relocation of nutrients also increases. Regional nutrient budgets show a progressive drawdown of potassium (K) as it is removed in plant products, and depending on cropping system and soils, there may be a need to consider the role of K. Use of K in Australia declined from 183 kt K in 2003 to 149 kt K in 2011 but did recover somewhat in 2012. At present around 75% of the K used is muriate of potash (MOP), also called sylvite (KCl) and about 1/3rd is used in WA, 1/3rd in Queensland and the balance across the other states.

Potassium (K) is a mineral nutrient essential to both plants and animals. Most of our annual crops contain about the same amounts of N and K, but K content of many high-yielding crops can demand more K than N. Unlike other nutrients, K does not form compounds in plants, but remains free in ionic form to 'regulate' many essential processes including enzyme activa­tion, photosynthesis, water use efficiency, starch formation and protein synthesis. Potassium plays a significant role in stomatal control so adequate K is often associated with some drought tolerance, and small grain can be a consequence of deficiency. Within general limits, neither grain protein (cereals) nor oil content (canola) are affected by K supply.

Table1. Potassium uptake by crops (Brennan pers. comm., Reuter pers. comm.)

Crop Use
K uptake
K content
K removal kg/ha
Wheat Growth
Wheat Grain
Wheat Straw
Wheat Hay
Canola Growth
Canola Grain
Canola Straw
Canola Hay

In South Australia about 10% of top 10 cm soil test values were less than a general lower limit of 100 mg/kg Colwell K, so the deficiency is not widespread. Low soil K values are predominately on the sandy grazing soils of the South East region (especially along the Coorong), parts of the Mount Lofty Ranges, Kangaroo Island and southern Mallee regions. Soils with marginal status are sandier soils on Eyre region. Consistently high values were observed in the northern and Yorke Peninsula cropping regions (Australian Agricultural Assessment 2001).

Table 1 shows K contents in wheat and canola crops cut for hay or left for grain. The K content of hay is much higher than the K content of grain, and as well the biomass in a hay crop will be more than the mass of grain removed so that around ten times more K will be removed in hay than grain.

Potassium concentration in the stubble of both canola and wheat is higher than in the seed, and if the stubble is burned, 30-40% of this K will be lost (Heard et al. 2006). If the stubble is retained, a lot of the K will be recycled through the topsoil, and unless the straw is spread, K can accumulate under the windrows, which is a reasonable field observation to diagnose K limitation.

As well as the amount of K demanded, the pattern of K uptake varies among crops. Some crops like maize takes up 50% of its K when it has accumulated only 30% of its peak biomass. In this crop, early K supply is important and early K stress cannot be remedied by tactical K applications. Rose et al. (2006) found that maximum K accumulation in wheat was around anthesis, but canola peaked a little later and had around 20% more K than wheat. Because K demand is quite high early, K should be applied early rather than later, and applications at booting in wheat or bud formation in canola are ineffective. Lupin is less responsive to applied K than either wheat or canola (Brennan 2012) and this is reflected in the lower critical Colwell K values (Table 2). Horticultural crops such as potatoes have very high demands.

Plants take up potassium actively and there are two mechanisms that operate depending on K concentration. Competition for uptake occurs between K and other monovalent ions such as sodium, and also between K and calcium and magnesium. The result can be low tissue Mg which is grazed pastures can induce grass tetany (if soil K/(Ca+Mg)<0.07, grass tetany is likely). The interactions among cations are quite complex and Na can replace part of the K demand of crops in mildly stressed cereals (Ma et al. 2011).

Potassium Deficiency Symptoms

One of the most common K defi­ciency symptoms is scorching or firing along leaf margins, usually appearing on older leaves first. Potassium deficient plants grow slowly and develop poor root sys­tems. Stalks are weak and lodging is common. Crops show lower resis­tance to disease and moisture stress.

Potassium in Soils

Most soils contain large amounts of K, which is present in clay minerals such as smectites, which are rich in K. However, probably less than 2% of the total soil K content is available to plants over the growing season, as most of it is a structural part of the clays. K exists in four forms.

    · Structural K – held in the lattice of soil minerals such as micas and feldspars. This is unavailable except over geologic time scales.
    · Fixed or interlayer K is trapped between layers of certain soil clays. This K is only very slowly available. The availability of this fraction is least reliably measured in soil tests, but it is the fraction that can be slowly depleted over time and this depletion will not show up in exchangeable or solution K soil tests.
    · Exchangeable K is present on the surface of clay and organic colloids and the size of this fraction depends on the CEC of the soil as well as its pH. K is displaced in acid soils.
    · Solution K is found in soil water and moves by diffusion to the plant root.

Plant available K is present as a cation in soil solution or in the clay complex, and can be accessed by roots through diffusion and this can be a slow process. Alternatively uptake can occur when crop roots contact soil colloids where K is held, but only a small proportion of the soil contacted. Figure 1 illustrates the way K diffuses to plant roots, and this transfer process can result in restricted supply to high demanding crops.

Figure 1. Potassium moves to plant roots either by slow diffusion or it taken up directly in exchange with soil colloids. Both can be slow processesBecause of the way K moves and is taken up, there are several things that cause problems when trying to predict K responsiveness using soil tests.

1. Because K is diffusion limited, wetted soil is critical for uptake, and if the topsoil or subsoil are dry, K will not be able to be accessed.

2. Under high-yielding condition, K diffusion can be slow and may not meet the rate at which K is demanded by the crop.
3. Rooting patterns differ among crops and tap rooted plants can be at a disadvantage in exploiting K compared to fibrous rooted plants. Row width can also change rooting patterns so wide seeding rows can reduce the volume of soil exploited.
4. Different species have different K demands, and wheat demand is a little lower than canola, and cotton tends be a lot higher. Data from India indicates that chickpea, peas, and lentils have a higher demand than wheat (Srinivasarao et al. 2003), while lupin has a lower demand or is more able to access soil K (Brennan and Bell 2013).
5. Other cations can affect the K demand but either competition or partial substitution, as well as through soil physical disruption.

Soil tests to predict crop response

As for other nutrients, soil tests seek to estimate the amount of K available to the plant using various extractants. Exchangeable K is estimated using ammonium acetate, while extractable K is estimated using bicarbonate (Colwell-K) or weak acid (Skene-K), and techniques using resins or exchange membranes are under development. There is a general relationship between Colwell K and ex-K, and for light soils the latter can be converted to the former by a factor of 391, but extractable can be more than exchangeable K on heavier soils. Mehlich-3 is a multi-element extractant used in the US but there is little data to support its use diagnostically in Australia. Tetraphenyl borate extractable K (TB-K) has been used to estimate the amount of K in the slowly available pools, and can help identify the amount of “back-up” available, and can be interpreted as a potential K buffering capacity. In the US there have been “new” tests proposed using moist rather than dried soils.

Predicting K response using soil tests is reasonably reliable for sandier soils, but on heavier soils the reliability declines. The search for more reliable soil K tests is a challenge and there have been concerns raised about the use of current tests (Khan et al. 2013). The Better Fertilizer Decisions for Crops (BFDC) project collated a large set of data on K responses and developed critical soil test ranges for a range of crops for Colwell-K (Table 2). The data were skewed to Western Australia and the values for heavy soils still seem relatively unreliable.

Table 2. Critical 0-10 cm Colwell-K soil test ranges (Brennan and Bell 2013) for a range of soil orders (values in mg/kg). Values are the 95% confidence range to achieve 90% of maximum yield.

All Soils
Ferrosols (Brown)
Tenosols 2-3 t/ha
Tenosols >3t/ha
Research in SA for sunflower by Lewis et al. (1991) proposed a critical value of 65 mg/kg which was similar to the broad range from the BFDC project. Research in the northern cropping zone has shown that K is becoming depleted, particularly in the subsoil, mainly because root growth in winter crops occurs below dry or drying topsoil. Deeper soil tests (to 30 cm) and estimates of buffering capacity and associated cations can help redefine critical limits (Table 3). Deeper sampling is also becoming used In Western Australia were K can leach on the coarse acid soils and the approach gave clearer critical values for wheat and canola than 0-10 cm sampling (Anderson et al. 2013).

Table 3 Tentative estimates of critical K levels for northern Vertosols (Guppy, pers.comm.). Values are those cited by Guppy as ex-K rather than Colwell K.

Topsoil (0-10 cm)
Subsoil (10-30 cm)
Ex K
If High Mg/Na*
If high Mg/Na*
<30 cmol/kg
30-60 cmol/kg
>60 cmol/kg
* High Mg >30% CEC, High Na >6% CEC

Critical Colwell K values for pastures are higher than the values in Table 2, and have been interpreted in terms of soil texture. Values range from 125 mg/kg in sands to 160 mg/kg in clay loams (Gourley et al. 2007).

Because of these interactions, it is difficult to reliably estimate K critical soil test values especially on Vertosols. The reason is that both rate of supply from less available K pools may be insufficient where rooting patterns or slow diffusion limit supply to a crop that experiences a period of high demand.

Plant Tissue Tests

As for some other nutrients, plant nutrient status can be assessed through tissues tests. The tests need to be timed to critical periods and the tissue selected needs to be responsive. The critical concentration for youngest emerged leaf blade K content at tillering in wheat is around 2% and about 1.8% youngest mature leaf during vegetative growth for canola (Reuter and Robinson, 1997). Concentration declines with later timings and with whole tops rather than blades. Grain K content is not a reliable indicator of paddock K status. Many plants show “luxury” accumulation of K so high values are not necessarily evidence of any toxicity.

Addressing K Deficiency

Timing, placement and rate of application all affect the response to K. Naturally it is critical to ensure that K is the limiting factor, and there are interactions among N, P, K and S as well as micronutrients. South Australia research showed that if N and P requirements were met, then added K gave good responses (Figure 2).

Figure 2. Grain yield response to the application of K at various N and P rates (Wilhelm, 2003). The soil Colwell K was 43 mg/kg in topsoil and 69 mg/kg in the subsoil.

Figure 3. Wheat grain yield in response to 50 kg K/ha with different application techniques (Wilhelm 2003). Soil Colwell K was 121 mg/kg in topsoil and 53 mg/kg in the subsoil.

The Right Time and Right Place to give the best responses are with K at seeding rather than topdressing. Muriate of potash is a salt and can cause damage to sensitive seeds when place together in the sowing row. The amount of damage will depend on row width, seeding points, soil texture and moisture. There is more information and access to the on-line damage tool at http://anz.ipni.net/article/ANZ-3076. Banding below the seed at planting has been shown to give much better results than topdressing or prespreading (Wilhelm 2003, Figure 3).

The Right Rate will need to be higher than replacement because K is relatively immobile. If the K buffering capacity is high, and the non-exchangeable K pool is strongly depleted, the competition between the soil and plant can mean minimum rates of 50 to 100 kg K/ha are needed to see responses. If using test-strips run out at seeding, then use a high rate to see if K supply is adequate. Another consequence of the low mobility – especially in alkaline soils – is that high rates can be used to cover two or three or even more crops. Work in Queensland using 200 kg K/ha deep (20 cm) banded is used before the most responsive crops and then cropped down over the seasons. So, it is better to use higher rates less frequently than lower rates every year.

The Right Source is usually MOP, mainly because it is significantly cheaper than sulfate of potash, potassium nitrate or potassium magnesium sulfate (langbenite). All commercially available K fertilizers are imported, although there is one current development to exploit greensands deposits of glauconite in WA. Some growers are concerned about adding extra chloride, but the amounts added are of little agronomic or environmental significance in adding to salt loads.


Potassium deficiency has been confirmed in South Australia in grain cropping regions. As K reserves are drawn down with higher yields, K replacement may need to be more widely addressed as it has been in Western Australia. The first evidence could be seen as good growth in windrows. Soil tests are reliable on sandy soils but less so on heavy soils. Responses to 50 to 100 kg K/ha banded below the seed at seeding where soil tests are below critical concentrations should give economic responses.


Anderson et al. 2013. WA Crop Updates. Paper at: http://www.giwa.org.au/2013-crop-updates

Australian Agricultural Assessment 2001. Volume 1, National Land and Water Resources Audit. Natural Heritage Trust.
Brennan and Bell 2013. Soil K-crop response calibration relationships. Crop and Pasture Science, 64, 514-522.
Brennan R. 2012. http://research.ipni.net/page/RANZ-2375
Brennan R. et al. http://www.australianoilseeds.com
Gourley et al. 2007. Better Fertilizer Decisions for Grazed Pastures. Victorian DPI. 16 pp.
Heard J et al. 2006. Nutrient loss with straw burning, Better Crops, 90, 10-11.
Khan et al. 2013. Renewable Agriculture and Food Systems. 29(1), 3-27.
Lewis et al. 1991. Fertilizer Research, 28, 185-190.
Ma Q, et al. 2011. Crop and Pasture Science 62, 972-981.
Reuter and Robinson, 1997. Plant Analysis, An interpretation manual. CSIRO Publishing.
Rose T. et al. 2006. Journal of Plant Nutrition and Soil Science 170, 404-411.
Srinivasarao et al. 2003. Better Crops. 17(1), 8-10.
Wilhelm N, (2003 date). Potassium responses observed in South Australian cereals. Available at: http://anz.ipni.net/article/ANZ-3197
Zubaidi et al. 1999. Australian Journal of Experimental Agriculture. 39, 721-732.

More about: Best Management Practices