In a publication brought to the public’s attention by news release from the University of Illinois several weeks ago, S.A. Khan, Richard Mulvaney and a colleague in the Department of Natural Resource and Environmental Sciences at the University of Illinois challenged a number of basic tenets of soil fertility, especially practices related to use of potassium (K) fertilizer. Citing hundreds of references and thousands of reported studies, they asserted that: K fertilizer is generally unnecessary in soils like most of those in Illinois; the soil test is not a reliable way to know how much K the soil will supply to a crop; K used as fertilizer can cause crops to have lower nutritional value; using K fertilizer can damage soil structure; and potassium chloride (KCl), which is the most commonly-used (and lowest-cost) K fertilizer material, is harmful to crops.
I won’t try to assess these assertions in detail, but will make some comments here to keep us focused on the larger issues of supplying nutrients to crops to prevent deficiencies and crop yield loss.
Can soils provide enough K without fertilizer?
Clay minerals in soils contain huge amounts of K, measured in tons per acre. Plant-available K – that found in soil solution or bound to exchange sites on soil particles – is typically measured in (low) hundreds of pounds per acre. The K in minerals is part of their chemical structure, and becomes plant-available very slowly – a few pounds per acre per year in many soils – as these minerals break down with weathering. Very low crop yields and K removal, or no crop removal (as was the case before Illinois soils were farmed) will result in plant-available K levels either not dropping or slowly increasing. Crop roots can pick up K from deeper soil layers, and over time will bring K up to the soil surface, where it will be released when plants die.
One of the items reported in the publication is that soil test K level in continuous corn in the Morrow Plots increased by about 67 percent over 50 years without any fertilizer application. They based this on a soil sample taken in 1955 and another in 2005. Soil test K levels in samples taken frequently from 1967 through 2008 in the Morrow Plots did not show such trends, however (Figure 1). Levels of K increased in both fertilized and unfertilized plots after deep tillage to relieve compaction in the late 1990s, but yields (and K removal) rose after that, and K levels came back down as removal rates rose.
The key to whether soils can supply enough K to meet crop needs is whether the crop removes K faster than the soil can free up K through the weathering process. At the low yield levels common 75 to 100 years ago, when return of nutrients to the soil as manure was common, K levels probably dropped slowly if at all even without fertilizer K. Today, a corn-soybean rotation with good yields will remove as much as 100 lb of K over a 2-year period. Most soils in Illinois can supply nowhere near these amounts, and so K levels will drop if no K fertilizer is added. How long it will take for a deficiency to appear will depend on how much plant-available K is present. But let’s not fool ourselves – K deficiency will appear at some point if removal continuously exceeds replacement. The only reasonable way to replace removed K is with K fertilizer.
There are soils in the world, including in parts of the western Corn Belt, where plant-available K levels are naturally high, due to the mineralogy of the parent materials from which the soils developed and the age of the soils. There are also places where soils have been weathered and leached for so long that K levels are very low. The soils in Illinois, and in most of the eastern Corn Belt US Corn Belt, fall into neither of these categories. Here, soils free up K more slowly than crops remove it, so if we are going to produce crops that remove K, at some point we are going to have to add some K back.
Is the K soil test useful?
There is no question that measuring plant-available K in soils is difficult, and that soil-test K levels vary over time, often not very predictably. Part of this is a sampling issue – soil test K levels often change quickly over short distances, so samples show variability. Soil moisture affects K tieup in clay minerals, and even the way soils are dried before testing can affect the soil test level. Despite all this, low soil test K values are often predictive of crop deficiencies, and crop K deficiencies are rare (though possible, especially when soils around the roots are dry) when soil test levels are high. Yield increases from adding K fertilizer are much more common when soil test K levels are low than when they are high. So while it’s easy to find fault with the soil test for K, the test (if samples truly represent the soils in a field) does tell us whether soil K levels are high enough to support full yields or whether deficiency and yield reductions are likely without adding fertilizer K. That’s usually all we need to know, especially when fertilizer K is routinely added back to replace that removed by crops.
Is K fertilizer harmful?
This is the area in which the authors take a rather odd turn, claiming that adding fertilizer K reduces crop nutritional value, damages soil structure, and (as KCl) can be toxic to plants. Evidence given on all of these counts is rather flimsy, and there’s little to say other than that any such effects are, based simply on quantities, minor or so small as to be non-detectable, or at worst, detectable but unimportant.
Claims that adding fertilizer K will lower “quality” of corn or soybean grain are without merit, and can be ignored. Corn and soybean grain K content can increase with increasing K rate in studies, but there is some recent evidence that grain K levels in both crops might be even lower than the “book values” used for years. Levels, and any problems claimed to be associated with such levels, are certainly not increasing.
The next claim has to do with the idea that K “makes the soil hard” and damages structure. As a monovalent cation, K can, if added in large quantities and given time, displace some of the other cations on soil exchange sites. If most of these exchange sites carried monovalent cations such as K or sodium (Na), soils would tend to “puddle” and be difficult to manage. The number of exchange sites is measured as the CEC, which range from single digits to the 40s or higher, depending on soil texture and organic matter. Each CEC unit occupied by K translates to 780 lb of K in the top 7 inches of soil. The great majority of exchange sites carry divalent cations such as calcium and magnesium, and this will continue to be the case even if we add hundreds (and in many soils thousands) of pounds of fertilizer K per acre.
There have been a number of efforts over years to show that the chloride in KCl can lower crop yield. How this would happen in Corn Belt soils is not clear. Like its cousin NaCl (table salt), KCl is a salt, and roots that encounter high salt concentrations in the soil can be damaged. But KCl can usually be banded with no negative effect, though there have been some reports of lower yields with banded K. Reasons for this are often not clear, but are more likely to be due to salt effects than to the chloride itself. As a negatively-charged anion, chloride moves through the soil quickly as water moves. Chlorine gas (Cl2) is harmful to all life, but its formation in the soil is chemically unlikely, and only tiny quantities would ever form. Chlorine is actually an essential plant nutrient, though some crops (like wheat) respond more to it than do corn and soybeans. It’s also known to reduce damage from some plant diseases. Plants, like people can deal with unneeded Cl by simply excluding it or getting rid of it once it’s been taken up.
The authors mentioned, without offering much evidence, that potassium sulfate would be a better K source than KCl. There’s nothing wrong with K2SO4 as a K source, but it is considerably more expensive than KCl, and there is no evidence that it’s a better source of K for corn and soybeans than is KCl.
Can we get better at fertilizing with K?
Even though the authors asserted that K fertilization is unnecessary in most Illinois soils, they (rather inconsistently) also suggested that producers do their own on-farm strip trials to see if adding K will increase crop yields. This is not a bad idea, as long as we keep in mind that soil test K levels in the medium or higher ranges – say above 250 lb per acre (125 ppm) usually means low chances of seeing a response. We can add to this the usual warning that looking at a few strips might lead us to wrong conclusions just by random chance of where the strips are located. But we have proposed a project to organize just such an effort in Illinois under the Nutrient Research & Education Council (NREC), and will be giving further information about this if the proposal is funded.