Improving the Reliability of Soil Potassium Testing and Recommendations

There is a long history of potassium (K) soil testing and management in Iowa and the North-Central Region. A period of especially extensive and diverse research extended from the late 1950s to the middle 1970s, especially with work in Illinois, Indiana, and Iowa but also in other states. During these years there was a great deal of basic K chemistry and mineralogy studies and also field applied research focusing on soil K testing and crop response to K fertilization. This research was published in dozens of journal papers of all kinds, and the basic and applied research results were the basis for most current soil-test K (STK) interpretations and K fertilizer recommendations in states of the region. With major issues apparently resolved, and several key investigators retiring, much less research was conducted until the early 1990s in Iowa and other states.

During the late 1980s and early 1990s, Iowa crop consultants and producers began observing clear symptoms of K deficiency in corn and soybean in some soils with assumed adequate K supply or even high STK levels where a deficiency should not be expected. Field observations indicated that these deficiencies were not deficiencies induced by known conditions such as soil compaction, dry soil, or incidence of root diseases and pests. Therefore, in 1994 an extensive applied K research effort began in Iowa. A brief review of this effort is warranted to better understand the reasons for the proposed work.

More than 400 short-term and long-term response trials with corn-soybean rotations or continuous corn have been established across Iowa since 1994. For some trials we used a conventional plot methodology, whereas for many (mainly those focusing on study of spatial variability of STK and crop response to K) the methodology involved replicated, on-farm strip trials using precision agriculture technologies. Eight long-term experiments established at university research farms continue to be evaluated today.

Soil-test K field calibration and placement methods research conducted from the middle 1990s until 2001 resulted in a major update of Iowa STK interpretations and K recommendations in 2002, which are current today. This research impacted four K management issues: (1) It revealed the inadequacy of existing STK interpretations using the ammonium-acetate K extraction method (the suggested optimum STK levels were too low). (2) Updated suggested average K concentrations in corn and soybean grain useful to better estimate K removal and maintenance fertilization rates. (3) It showed that deep K placement should be used with the ridge-till tillage system in all conditions and with no-till or strip-till with dry weather, but not with chisel-plow/disk tillage. (4) Improved soil testing for K by establishing the first field-based calibrations for the Mehlich-3 K test in the North-Central Region. Therefore, there were major updates of Iowa interpretations and recommendations, first in 1996 and last in 2002. The last update included the most radical change, in which all STK interpretations categories for the ammonium- acetate and Mehlich-3 K tests were increased. For example, the old Optimum category, for which maintenance fertilization based on removal was recommended, was increased from 90-130 ppm to 130-170 ppm (6-inch sampling depth). These changes had a significant impact in K management in Iowa, since a major extension effort and collaboration with coops and influential crop consultants resulted in widespread adoption of the recommendations.

Although this research solved some serious problems, at the same revealed other problems and suggested new areas of research. Several projects conducted since 2001 had significant results to improve K management but are not directly relevant to the proposed research, so will not be described here. These include characterization of the within-field spatial variability of STK and crop response to K (Clover and Mallarino, 2008), interactions between corn and soybean diseases or pests and K nutrition (Fixen et al., 2008, Mallarino and Clover, 2009), use of starter K with or without broadcast K for corn (Mallarino et al., 2010; Mallarino et al., 2011), and relationships among corn and soybean grain yield, tissue K concentration, and K fertilization (Clover and Mallarino, 2013).

However, results of two recent research projects are very relevant to the idea and justification of the proposed research.

Soil cation exchange capacity (CEC), K saturation, and moist soil testing.

In spite of the extensive new STK field correlation and calibration effort, the currently used STK methods are much less reliable to predict crop K sufficiency and K fertilizer needs than are soil tests for P, for example. The performance of these tests is so poor that their main value is to show that no K fertilizer is needed above a certain level, but predictions of response and K needs below such a level are not at all reliable. Therefore, extension efforts since 2002 until this year began to emphasize build-up of STK levels to minimize risk of under-fertilization and lower economic benefits. As a consequence, research from 2001 until 2006 studied the effects of soil CEC and K saturation on field correlations of ammonium- acetate and Mehlich-3 tests with corn and soybean and at the same time evaluated a soil test for K based on field-moist samples, instead of the common method of drying of samples before analysis. The research involved 200 site-years for corn, 162 site-years for soybean, and 32 Iowa soils series, some of which also are found in neighboring states (Barbagelata and Mallarino, 2013). The results showed that consideration of soil CEC, K saturation, or ratios between Ca, Mg, and K did not improve the value of soil testing for K in Iowa. Research in other regions has shown that critical STK levels tend to be higher for soils with high CEC than for soils with very low CEC. Perhaps the variation across Iowa soils was not wide enough, although, for example, ranges were 9.7 to 34.1% clay, 12.0 to 36.5 cmol_c kg^-1 CEC, and 18 to 94 (Ca + Mg)/K ratio (Barbagelata and Mallarino, 2013). This result agreed with unpublished Iowa survey-type of research by the late Dr. Alfred Blackmer, who found no relationship between soybean yield and CEC or soil cation ratios (A.M. Blackmer, 2007, unpublished).

A field-moist K test was developed and adopted by the Iowa State University (ISU) soil testing laboratory from 1963 until 1988 since regional research (mainly in the greenhouse) showed it was better for soils of the region. It was among procedures recommended by the North-Central Regional Committee for Soil Testing and Plant Analysis (NCERA-13) committee during the 1980s (Eik et al., 1980; Eik and Gelderman, 1988). No other laboratory adopted the moist test, citing impractical handling procedures, however. So in 1988 the ISU laboratory discontinued use of the moist K test, and in 1998 the NCERA-13 committee also dropped this procedure from its methods publication (Gelderman and Mallarino, 1998). Recent research has confirmed that testing undried soil samples was a better predictor of K sufficiency and K fertilizer needs for crops (Barbagelata and Mallarino, 2013), and again showed that suggested levels for the dried tests still are too low. In 2012, the moist sample handling procedure was updated and re-introduced in the NCERA-13 publication with recommended soil testing methods (Gelderman and Mallarino, 2012). Since fall 2012, this test is offered as the only K test or as an option by two large private laboratories that operate mainly in Iowa, Minnesota, and northern Missouri. The ISU Soil and Plant Analysis Laboratory will implement it as an option for routine testing this year. Therefore, new interpretations for the commonly used dried samples (to increase the suggested STK optimum levels) and for this "revived" moist test will be included in Iowa recommendations during 2013.

Potassium recycling to soil and temporal STK variability:

We in Iowa theorized that the poor reliability of soil testing for K could partly be explained by variable K recycling from residue to soil during the fall and spring soil sampling seasons. Potassium is soluble in plant tissue, and both rainfall intensity and distributions could affect K leaching to the soil and STK. Therefore, research was conducted over four years and involving 33 corn site-years and 14 soybean site- years (Oltmans and Mallarino, 2012). The main findings were that K recycling to soil from standing soybean plants at physiological maturity in the fall and from residue since harvest until the next spring was much more complete and faster than for corn. About 65 to 70 % of soybean plant K (except grain) was recycled by early December, and by early spring 80 to 90% was recycled to soil. In corn, the K release from plant tissue and residue is more gradual and about 45% of the K remained in the residue by early spring. Much less rainfall was needed to recycle a certain amount of K in soybean than in corn. There was a significant positive linear relationship between K loss from residue and increased STK in the spring compared with the fall. These results have great value to better understand the very high temporal variability and uncertainty of soil testing for K.

The results of the recent Iowa research on the field-moist test and on impacts of rainfall and type of crop residue on K recycling to the soil no doubt have improved soil testing for K or understanding of some failures, although some things are being implemented at this time. However, this very same research also showed that much temporal STK variability remained unexplained, and that the difference between soil K extracted by moist and dry testing methods varies greatly across soil series, soil moisture levels, and years. These types of variation could not be explained by soil texture, CEC, or K saturation. In recent work, soil clay mineralogy and moisture were not measured at any site due to budget limitations.

We believe that one key "missing piece" of information relates to poor understanding and absolute lack of consideration of the equilibrium between exchangeable and non-exchangeable soil K pools in soil testing for crop production. Extensive research has shown that this equilibrium is affected by many factors, such as soil mineralogy (mainly the clay and fine silt fractions), the concentration of these soil K pools, soil moisture and reduction/oxidation conditions, K additions and removal, and interactions with ammonium- N in the soil, among others. The general view in published research and among many soil scientists has been that this equilibrium is relevant mainly at decadal or longer periods, and this is the main reason that the measurement of non-exchangeable K is not used in soil-testing for crop production. Another very important practical reason is that the measurement of non-exchangeable K is cumbersome and expensive.

In soils, K occurs as a monovalent ion with low hydration energy. Positively charged K ions are held in soil by the attraction of negative charges that are associated with clay-size, layer silicate minerals and soil organic matter. Plant-available K may be thought of as being held "exchangeable" at two kinds of clay mineral sites: (1) on exterior particle surfaces and (2) between aluminosilicate layers at the “frayed” edges of weathered particles. Potassium ions in the “deep” interlayer region of micas or within crystals of feldspars are released too slowly to the soil solution to be important in plant nutrition during a growing season. In this simplified model, K ions that are held at exterior surfaces (or in the interlayer zones of low-charge minerals such as montmorillonite) can be readily displaced (“exchanged”) by other cations in the soil solution and therefore are readily available for plant uptake. On the other hand, K ions that are held at the frayed edges of minerals (such as clay mica, i.e., illite) are more slowly released to the soil solution by either exchange processes or by a concentration deficit in solution (disequilibrium). If the soil solution has low K⁺ activity, K at the exterior surfaces of clay minerals may be released into solution by exchange with H⁺, Na⁺, Mg²⁺, or Ca²⁺; and this is a primary source of K for plant nutrition. Potassium in interlayer positions near the edges of high-charge layer silicates has been termed "non-exchangeable" or “fixed” because it is not readily subject to exchange with other cations. The affinity of high-charge layer- silicate surfaces for K is influenced greatly by the ion’s weak hydration energy and size, which is about the same as that of the ditrigonal cavity in the basal plane of oxygen atoms of layer silicates.

The frayed edges of clay minerals such as muscovite, biotite, illite, and vermiculite possess high levels of negative charge and provide locations where K ions are not easily exchanged with other ions. In contrast, montmorillonite is a smectite mineral that does not have sufficiently high negative charge to fix K tightly under most conditions, although drying may promote temporary K fixation when montmorillonite layers collapse around the easily dehydrated K ions. Other smectites, such as beidellite, may have high enough inherent charge to retain K more strongly than montmorillonite under both moist and dry conditions.

Estimating the capacity of silicate clay minerals to supply K to the soil solution (that is, the K buffer capacity) should be important for predictions of plant nutrition and growth. The K buffer capacity of a given soil is not a constant value but depends on soil pH and on concentrations of other cations, especially Ca, in the soil solution. Fertility recommendations for crops are normally based not on K buffer capacity but on the amount of K released by extraction of a soil sample with ammonium acetate (as discussed above), and that value is not always well correlated with plant uptake of K. Drying soil samples before the extraction may significantly alter quantities of ammonium-exchangeable K compared with extraction from field-moist samples, and research has suggested this happens by the collapse of silicate layers around K ions when water molecules that normally hydrate the K are lost by evaporation (Luebs et al., 1956, among others).

Another reason that STK may not reflect plant-available K is related to variations in the negative charge of clay minerals that contain Fe in their octahedral sheets (e.g., beidellite). Reduction and oxidation of Fe within such layer silicate minerals can lead to cycling in the fixation and release of K⁺ (Fig. 1). This process has been extensively documented in laboratory settings by Stucki and co-workers (Stucki et al., 1987; Kostka et al., 1996, 1999; Gates et al., 1998). For example, temporary fixation may occur during the spring when soils become water saturated during periods of heavy rainfall. Saturation reduces the redox potential of the soil solution. Low redox potential results in reduction of Fe³⁺ to Fe²⁺ -- thereby increasing the effective negative charge that holds cations in mineral interlayer regions. When the soil dries, oxidizing conditions return, Fe²⁺ is oxidized to Fe3⁺, the effective negative charge holding K⁺ and NH₄⁺ ions decreases, and K⁺ and NH₄⁺ exchangeability increases. Because K⁺ and NH₄⁺ ions are similar in size and hydration energy, the same mineralogical and environmental factors that control the availability of K to plants also affect the availability of ammonium. In some soils, added ammonium fertilizer has been shown to inhibit the release of fixed K, presumably because when they are adsorbed at the weathered edges of micaceous minerals, NH₄⁺ ions block diffusion of K from interlayer positions (Welch and Scott, 1961). Thus, another source of variation in STK measurements may be applications of fertilizers such a monoammonium and diammonium phosphate to the soil before sampling for STK.



Fig. 1. Temporary fixation of K and NH₄ by vermiculite and smectite (courtesy of Darrel Schulze, Purdue University).


Research has shown that a portion part of the non-exchangeable K held in the interlayers of expandable or partially expandable clay minerals can be released relatively easily to re-supply a fraction of the K taken up by plants (Singh et al., 1983; Meyer and Jungk, 1993; Mengel and Uhlenbecker, 1993). Cox and Joern (1996) showed that the ammonium-acetate STK method predicted plant-available K poorly in soils where non-exchangeable K contributed significantly to K nutrition in winter wheat and that often ammonium- acetate extractable K underestimated plant available K in some soils with high K retention capacity (Cox et al., 1999). Therefore an ideal soil-test for K perhaps should quantify the proportion of non- exchangeable K that may potentially become available to plants early during the growing season. The sodium tetraphenylboron (NaBPh₄) method for measuring non-exchangeable (Scott et al., 1960; Smith and Scott, 1966) is a less drastic treatment of the mineral structures than extraction with boiling 1 M HNO₃ (McLean and Watson, 1985), but Scott and Reed (1960 and 1962) demonstrated it is an effective extractant for non-exchangeable K from mica structures. The BPh₄ anion facilitates the release of non- exchangeable K by combining with K in solution and precipitating it as KBPh₄, while Na exchanges with interlayer K.

Cox et al. (1996) modified the NaBPh₄ method by using Cu instead of Hg to destroy the BPh₄ anion and recover precipitated K. Cox et al. (1999) simplified the method further, mainly with the objective of making more feasible for routine testing, by decreasing extraction time. They showed that K measured by the modified NaBPh₄ test was a better predictor of K uptake by winter wheat than with ammonium- acetate because the latter does not measure potentially available non-exchangeable K. Schindler et al. (2002), working in South Dakota under field conditions, found no advantage of using the modified NaBPh₄ method over the commonly used ammonium-acetate STK method at predicting plant K concentration in corn. They postulated that one reason NaBPh₄ was not better was that montmorillonite dominated the clay fraction of the soils used, which would not be expected to contribute much plant- available non-exchangeable K. Unpublished Iowa work by Mallarino research group developed field correlations of the modified NaBPh₄ method and both the ammonium-acetate and Mehlich-3 STK methods, all based on dried samples, for corn (63 site-years) and soybean (54 site-years) from 2001 until 2004 (Barbagelata and Mallarino, 2006). The amount of NaBPH₄-extractable K always was significantly higher than amounts extracted by the STK methods, and the difference increased with increasing soil K and decreased as the Ca and Mg to K ratio increased. The found that NaBPH₄ did not have a consistently superior capacity to predict corn and soybean response to K fertilizer than the STK methods, and it was very laborious and expensive. Soil texture, CEC, or K saturation did not help explain results by any test. However, soil clay mineralogy and soil moisture were not measured at any site due to budget limitations.

This research and additional, more recent, unpublished research by Mallarino and students (M.W. Clover and A.P. Mallarino, unpublished; Villavicencio and Mallarino, 2011) provided evidence for a faster than assumed equilibrium between exchangeable K (estimated by STK methods) and the more reactive fraction of non-exchangeable K as measured by the modified NaBPh₄ method. They showed that K additions and removal changed this non-exchangeable K fraction as much as exchangeable K not only after many years of different K additions and removal by cropping but also within one year. Moreover, this non-exchangeable fraction may sometimes, but not always, be plant-available within a short period. For example, they found that a pre-plant K application several times higher than removal greatly increased post-harvest STK levels over initially low levels at some sites. In contrast, at other sites with similar K application rate and removal, the post-harvest STK increased very little or did not increase whereas post-harvest NaBPH₄-extractable K did increase by a very large amount. The most interesting result was, however, that the following crop did not respond to additional K application in any case. Soil texture, CEC, or K saturation did not help explain results by any test. Soil clay mineralogy and soil moisture were not measured at any site due to budget limitations.

Therefore, the long history of K research in the North-Central region, but mainly Iowa research during the last decade, clearly suggests a need for an in-depth study of short-term changes in the exchangeable and non-exchangeable soil K pools to better understand short-term K dynamics in soils. This should include a combination of incubation and field studies that explores the interactions of K addition and removal, clay mineralogy, and soil moisture changes in the context of reduction/oxidation conditions.

The general goal of the project is to take a significant step forward at improving soil testing for K and the assessment of crop K fertilization needs by building on recent research that partially solved some problems but revealed information gaps that need to be addressed. We hypothesize that the so-called non- exchangeable soil K pool has a much more important and short-time impact on STK levels and its temporal variability over a period of a few weeks or months and also on crop-available K and crop response to K fertilization.

This project will combine basic research to characterize fundamental chemical processes with field-scale research that is directly applicable to production agriculture via improved K testing and management. This kind of study will increase understanding of K retention and release in soils of the North Central Region and should result in better recommendations and more profitable K management practices.

The specific objectives are the following:

1. Study under controlled conditions the relationships among STK measured by the ammonium-acetate and Mehlich-3 extractants, non-exchangeable K, and relevant properties of soils of the western humid Corn-Belt with large acreages dedicated to row-crop production.
2. Study at the field scale how K additions and crop K removal influence changes in the ratio between STK and exchangeable K as assessed in samples collected in the fall (post harvest) and spring (shortly before crop planting).
3. Determine if knowledge of both STK and non-exchangeable K for specific soil types and/or conditions is useful to improve the interpretation of routine soil tests for K and K fertilizer recommendations for crops.

The main study for achieving Objective 1 will consist of mineralogical analyses in conjunction with investigation of STK / non-exchangeable K relationships based on an indoor incubation technique. Soil samples will be collected from soils of Iowa and the western humid Corn-Belt region that are extensively dedicated to corn and soybean production. Soils of approximately 25 soil series will be used for this objective. At least one-half of the soils will be from Iowa because this is where field trials for the project will be established, and some Iowa soil series also are found in areas of neighboring states. The soils will be selected based on contrasting properties which previous knowledge suggests may determine differences in the dynamics of the ratio of non-exchangeable K to STK. These properties include clay mineralogy, texture, extractable (Ca+Mg) / K ratio, drainage class, and history of K application and removal. Samples will be collected from the surface 15-cm layer, preferably from K research sites having a good history of soil testing, management, and crop response to K fertilization.

The soils will be characterized for particle size distribution and several routine soil tests, such as P, K, organic matter, extractable cations, pH, calcium carbonate, and others. For STK, soil will be analyzed by the ammonium-acetate and Mehlich-3 methods by using both the dried sample handling procedure and field-moist procedure (Gelderman and Mallarino, 2012). The chemical analyses for both sample handling procedures will use equivalent ratios of dry soil to extracting solution and other procedures suggested by the NCERA-13 regional committee (Warncke and Brown, 1998). Also, all samples will be analyzed for non-exchangeable K by the two versions of the NaBPh₄ method mentioned in the background section. One version is the original method with a long digestion period as proposed by Scott et al. (1960) and Smith and Scott (1966), modified to use Cu²⁺ instead of Hg²⁺ to destroy the BPh^4- anion and recover precipitated K (Cox et al., 1996). The other version measures the most weakly retained fraction of the so called soil “non-exchangeable” K pool by the short-digestion method of Cox et al. (1999). Total soil K will be measured after digestion using aqua regia and HF (Helmke and Sparks, 1996).

The mineralogical composition of the clay (< 2 ìm equivalent settling diameter) and fine silt (2 – 20 ìm) fractions of each soil will be determined. Semi-quantitative determination of layer silicates will be made by a combination of x-ray diffraction (using both oriented samples and powder mounts), Cu Ká radiation, and an x-ray diffractometer (Whittig and Allardice, 1986; Harris and White, 2008) and thermal analysis (Karathanasis, 2008). The proportion of vermiculite and smectite in the clay fractions of the soil will be quantitatively determined by using the Ca / Mg exchange method for smectite and the K/ NH₄ exchange method for vermiculite (Jackson, 1979). The concentration of discrete illite in the clay fraction will be determined by measuring non-exchangeable K after digestion of the clay fraction with aqua regia and HF (Jackson, 1979).

The soil characterization data will be used to select soil series for an indoor incubation experiment to study non-exchangeable K / STK relationships in which treatments will be soil series, two K application rates, and two hydrological regimes over a 6-month period. We suspect that some of the soils will have similar properties that are relevant to K release, and eliminating potential duplication by selecting contrasting soils will allow us to adjust the work and costs to the budget available. The basic methodology (except for the hydrological regime treatment) has been developed in previous work in our laboratory (Ruiz Diaz et al., 2008). The soil samples will be air-dried, ground to pass a 2-mm sieve, and stored at 4°C until the incubation is initiated. The treatments will be replicated three times. The K application treatments will consist of no K application or mixing soil and KCl fertilizer at an equivalent rate of 100 kg K ha^-1. This rate is intermediate between rates currently suggested in Iowa for corn and soybean in soils testing Very Low and the suggested maintenance K fertilization rates that are based on K removal with grain harvest. These two K treatments will be applied to three portions (for each replication) of each of the 25 soils. After thorough mixing, 200 g of soil will be poured into a 300-mL plastic cup in which two simulated hydrological regimes and two incubation periods can be applied. One hydrological regime will consist of maintaining the soil at near field capacity and aerobic conditions at constant 25°C. Each cup will have a plastic lid with two small holes drilled at the center to allow air exchange. De- ionized water will be added periodically as needed to maintain approximately 80% of the soil water- holding capacity. For the other hydrological regime, the soil will be maintained at waterlogged condition for one week by adding excess water, opening a hole at the bottom of each cup to let water drain and incubate for one week (with no water additions), create again one week of waterlogged conditions, and let water drain. Once soil in these cups reaches about 80% of the water-holding capacity, all cups will be maintained at this water level. The incubation will be ended after three months for a set of cups and after six months for another set (two incubation periods).

At the end of each incubation period, soil will be analyzed (1) for crop-available K (by the ammonium- acetate and Mehlich-3 methods, using both the dried and field-moist sample handling procedures mentioned above), (2) for water-exactable K, and (3) for non-exchangeable K by the modified (short digestion time) NaBPh₄ method mentioned above. The results of these tests will be studied in relation to the treatments and the soil properties.

Seventeen existing Iowa K response trials with corn-soybean rotations that were last evaluated in 2012 will be continued or modified as needed to study how K fertilization and crop K removal influence changes in the ratio between non-exchangeable K and STK over time. The benefit of using existing trials with different K application histories include lower cost and that represent prevailing conditions in production agriculture, since the vast majority of fields managed with corn-soybean rotations have long and varied histories of K application. Two trials have been conducted for 36 years at sites with the soil series Kenyon (mixed, superactive, Typic Hapludolls) and Webster (mixed, superactive, Typic Endoaquolls) (developed in glacial till of different ages). Five trials have been conducted for 20 years at sites with soil series Kenyon, Webster, Galva (mixed, superactive, Typic Hapludolls), Mahaska (smectitic Aquertic Argiudolls), and Marshall (mixed, superactive, Typic Hapludolls) (the last three soils developed in loess of different thickness and age). Also, ten trials have been conducted for four to eight years at ten sites with some of the previously mentioned soils and also the series Canisteo (mixed, superactive, calcareous, Typic Endoaquolls), Clarion (mixed, superactive, Typic Hapludolls), Floyd (mixed, superactive, Aquic Pachic Hapludolls) (these three soils developed in glacial till of different age), Exira (mixed, superactive, Typic Hapludolls), Grundy (smectitic Aquertic Argiudolls), Haig (smectitic Vertic Argiaquolls), and Taintor (smectitic Vertic Argiaquolls) (the last four soils developed in loess of different thickness and age).

Some trials are managed with chisel-plow/disk tillage and others with no-till. Crop grain yield and K removal with grain harvest will be evaluated for three crop years (2013, 2014, and 2015), although soil sampling will extend until spring 2016. Notice of approval or not of the proposal will be received by June 1st. With the expectation of approval, however, we plan to take this spring all the soil samples needed to use the 2013 crop year and advance faster at obtaining results.

Broadcast K fertilizer treatments (KCl fertilizer) have been applied at all trials. Treatments at some trials include a control receiving no K and several K rates, whereas at others only a control and one high K rate had been applied. To adjust the work to the budget available, we will use all field replications (three to five) of a control treatment that has not received K and of a high K rate, which has not been the same at all trials. Having different high rates at some trials should not be a serious limitation since these high rates have maximized crop yield over time and the main objective is to study variations in the ratio of non- exchangeable K to STK from fall until spring in relation to K removal and soil properties.

Soil samples (12 cores, 15-cm depth) will be collected for each crop year from the same plots at three times. The first samples will be collected in spring, once soils have thawed and moisture is appropriate for good soil sampling and before planting the crops (probably during the first two weeks of April). The other two sampling times will be after crop harvest in the fall. These samples will be collected immediately after harvest (late September to the middle of October depending on the year and crop maturity) and again later in the fall right before snow or soils freeze (usually middle to late November). Comparison of soil K measured by various methods between spring and fall will be useful to study possible relationships with K removal. Comparisons of soil K measured for the two sampling dates in the fall and in the following spring will be useful to study time effects interacting with weather and soil properties across sites and years.

A composite sample for each soil series included in the field trials will be prepared from soil collected for the first time from control plots that have not received K fertilizer. These samples, one per site, will be used for the incubation study described above for Objective 1, so the soil properties will be characterized as described for Objective 1. The samples collected from each plot at each of the three sampling times will be analyzed for soil K using the same methods described for the two incubation periods of the study described for Objective 1. These will include (1) crop-available K by the ammonium-acetate and Mehlich-3 methods, using both the dried and field-moist sample handling procedures and (2) for non- exchangeable K by the by the modified (short digestion time) NaBPh₄ method. In addition, soil moisture at each sampling date will be determined by a gravimetric method.

Rainfall, temperature, crop hybrids or varieties, and the results of the crop and soil evaluations will be studied to interpret relationships between K treatments, grain yield, K removal, and the measured soil properties with emphasis on the ratio between non-exchangeable K and STK.

Objective three will be achieved by using the results of evaluations described for Objective 1 and 2 to see if the new knowledge of STK / non-exchangeable K relationships for specific soil series and soil moisture conditions is useful to improve the interpretation of routine soil testing for K. Our hypothesis is that this will indeed be the case, and we believe the study will prove it. If so, we should be able to suggest ways in which this knowledge can be implemented in practice for improving crop K fertilizer needs. No clear procedures can be detailed at this time, because the results of the crop and soil test evaluations will indicate what can be done. However, there is no doubt that procedures will include extensive simple and multiple correlation and regression analyses across and for each soil series included in the incubation study, and across and for each site-year of the field experiments. This portion of the project will include input by colleagues from other states who will be invited to collaborate by providing soil samples. The specific soil series and collaborators have not been identified at this time.

The soil series characterization and incubation study will be completed by the end of second funding year (May 2015). The crop harvests will be in the fall of 2013, 2014, and 2015. Therefore, the chemical analyses of the last grain samples to estimate K removal will be completed by February 2016. Analyses of the last soil samples, which will be collected in April 2016, will be completed by August 2016. Data management and statistical analysis for the entire project will be completed by January 2017. A final comprehensive report and articles for publication in peer-reviewed scientific journals will be completed by May 2017.

Barbagelata and Mallarino, 2006. Field calibration of ammonium acetate, Mehlich-3, and sodium tetraphenylboron soil potassium tests for corn and soybean. In: P.A. Barbagelata, Ph.D. Dissertation. Iowa State University. Ames.
Barbagelata, P.A., and A.P. Mallarino. 2013. Field correlation of potassium soil test methods based on dried and field-moist soil samples for corn and soybean. Soil Sci. Soc. Am. J. 77:318-327.
Clover, M.W., and A.P. Mallarino. 2008. Using precision agricultural technologies to assess within-field variation of corn and soybean grain yield responses to potassium fertilization. In: M.W. Clover. Ph.D. Dissertation. Iowa State University, Ames.
Cox, A.E., and B.C. Joern. 1996. Predicting plant available potassium using a modified sodium tetraphenylboron method. p.70-81. In: Proceedings North Central Extension-Industry Soil Fertility Conference. Vol. 12. St. Louis, Missouri.
Cox, A.E., B.C. Joern, C.B. Roth. 1996. Non-exchangeable ammonium and potassium determination in soils with a modified sodium tetraphenylboron method. Soil Sci. Soc. Am. J. 60:114-120.
Cox, A.E., B.C. Joern, S.M. Brouder, and D. Gao. 1999. Plant-available potassium assessment with a modified sodium tetraphenylboron method. Soil Sci. Soc. Am. J. 63:902-911.
Eik, K., and R.H. Gelderman. 1988. Soil sample preparation. p. 2-4. In: Recommended chemical soil test procedures for the North Central region. North Central Regional Publ. 221 (rev.). Bull. 499 (rev.). North Dakota Agric. Exp. Stn., Fargo.
Eik, K., E.L. Hood, and J.J. Hanway. 1980. Soil sample preparation. p. 2-3. In: Recommended chemical soil test procedures for the North Central region. North Central Regional Publ. 221 (rev.). Bull. 499 (rev.). North Dakota Agric. Exp. Stn., Fargo.
Fixen, P. E.; R.W. Schneider, D.L. Wright, A.P. Mallarino, K.A. Nelson, S.A. Ebelhar, and N.A. Slaton. 2008. Implications of Asian soybean rust in nutrient management - research update. Better Crops 92(1):26-29.
Gates, W.P., A.-M. Jaunet, D. Tessier, M.A. Cole, H.T. Wilkinson, and J.W. Stucki. 1998. Swelling and texture of iron-bearing smectites reduced by bacteria. Clays Clay Miner. 46:487-497.
Gelderman, R.H. and A.P. Mallarino. 2012. Soil Sample Preparation. Chapter 1, Revised October 2012. p. 1.1-1.4. In: Recommended chemical soil test procedures for the North Central region. North Central Regional Publ. No. 221 (Rev.). Missouri Exp. Stn. Publ. SB 1001. Univ. of Missouri. Columbia. Published online at http://ncera-13.missouri.edu/pdf/sb1001_updatedOct2012.pdf.
Gelderman, R.H., and A.P. Mallarino. 1998. Soil sample preparation. In J. Brown (ed.) Recommended chemical soil test procedures for the North-Central Region. North-Central Regional Publ. No. 221 (Rev.). Missouri Agric. Exp. Station. Columbia. MO.
Harris, W., and G.N. White. 2008. X-ray Diffraction Techniques for Soil Mineral Identification. SSSA Book Series, Methods of Soil Analysis, Part 5 - Mineralogical Methods 5.5:81-115.
Helmke, P.A., and D.L. Sparks. 1996. Lithium, Sodium, Potassium, Rubidium, and Cesium. SSSA Book Series, Methods of Soil Analysis. Part 3 - Chemical Methods 5.3:551-574.
Jackson, M.L. 1979. Soil chemical analysis: Advanced course, 2nd edition. Department of Soil Science, University of Wisconsin, Madison.
Karathanasis, A. D. 2008. Thermal Analysis of Soil Minerals. SSSA Book Series, Methods of Soil Analysis, Part 5 -Mineralogical Methods 5.5:117-160.
Kostka, J.E., E. Haefele, R. Viehweger, and J.W. Stucki. 1999. Respiration and dissolution of Fe(III)- containing clay minerals by bacteria. Environ. Sci. Technol. 33:3127-3133.
Kostka, J.E., K.H. Nealson, J. Wu, and J.W. Stucki. 1996. Reduction of structural Fe(III) in smectite by a pure culture of the Fe-reducing bacterium Shewanella putrefaciens strain MR-1. Clays Clay Miner. 44:522-529.
Luebs, R.E., G. Stanford, and A.D. Scott. 1956. Relations of available potassium to soil moisture. Soil Sci. Soc. Am. Proc. 25:349-352.
Mallarino, A.P., and M.W. Clover. 2009. Do new corn hybrids and yield levels influence potassium fertilizer management? p. 125-131. In: The Integrated Crop Management Conf. Proceedings. Dec. 2-3, 2009. Ames, IA. Iowa State Univ. Extension.
Mallarino, A.P., M.W. Clover, and R.R. Oltmans. 2011. Identification of reasons for high temporal soil- test potassium variation. p. 65-73. In: North-Central Extension-Industry Soil Fertility Conf. Proceedings. Nov. 16-15. Vol. 27. Des Moines, IA.
Mallarino, A.P., N. Bergmann, and D.E. Kaiser. 2010. A look at effects of in-furrow applications of potassium on corn. Fluid J. 18: (70) 12-15.
Mallarino, A.P., N. Bergmann, and D.E. Kaiser. 2011. Corn responses to in-furrow phosphorus and potassium starter fertilizer applications. Agron. J. 103:685-694.
McLean, E.O., and M.E. Watson. 1985. Soil measurements of plant-available potassium. p 277-308. In: R.D. Munson (ed.) Potassium in agriculture. ASA, CSSA, and SSSA, Madison, WI.
Mengel, K., and K. Uhlenbecker. 1993. Determination of available interlayer potassium and its uptake by ryegrass. Soil Sci. Soc. Am. J. 57:761-766.
Meyer, D. and A. Jungk. 1993. A new approach to quantify the utilization of non-exchangeable soil potassium by plants. Plant and Soil 149:235-243.
Oltmans, R.R., A.P. Mallarino. 2012. Potassium uptake, recycling, and soil-test temporal variability in soybean and corn. In: R.R. Oltmans, M.S. Thesis. Iowa State University, Ames.
Ruiz Diaz, D.A., J.E. Sawyer, and A.P. Mallarino. 2008. Poultry manure supply of potentially available nitrogen with soil incubation. Agron. J. 100:1310-1317.
Schindler, F.V., H.W. Woodard, and J.J. Doolittle. 2002. Plant available potassium assessment trough chemical prediction methods. Commun. Soil Sci. Plant Anal. 33:1473-1484.
Scott A.D., R.R. Hunziker, and J.J. Hanway. 1960. Chemical extraction of potassium from soils and micaceous minerals with solutions containing sodium tetraphenylboron. I. Preliminary experiments. Soil Sci. Soc. Amer. Proc. 24:191-194.
Scott, A.D., and M.G. Reed. 1960. Determination of the precipitated potassium in sodium tetraphenylboron-micaceous mineral systems. Soil Sci. Soc. Am. Proc. 24:326-327.
Scott, A.D., and M.G. Reed. 1962. Chemical extraction of potassium from soils and micaceous minerals with solutions containing sodium tetraphenylboron: II. Bioite. Soil Sci. Soc Am. Proc. 26:41-45.
Singh, K.D., K.W.T. Goulding, and A.H. Sinclair. 1983. Assessment of potassium in soils. Commun. Soil Sci. Plant Anal. 14:1015-1033.
Smith, S.J., and A.D. Scott. 1966. Extractable potassium in grundite illite: I. Method of extraction. Soil Sci. 102:115-122.
Stucki, J.W., P. Komadel, and H.T. Wilkinson. 1987. Microbial reduction of structural iron(III) in smectites. Soil Sci. Soc. Am. J. 51:1663-1665.
Villavicencio, C.X., and A.P, Mallarino. 2011. Long‐term relationships between potassium fertilization, removal with crop harvest, and soil‐test potassium. In: C.X. Villavicencio, M.S. Thesis. Iowa State University, Ames.
Welch, L.F. and A.D. Scott. 1961. Availability of nonexchangeable soil potassium to plants as affected by added potassium and ammonium. Soil Sci. Soc. Am. Proc. 25:102-104.
Whittig, L.D., and W.R. Allardice. 1986. X-ray diffraction techniques. p. 331-362. In: A. Klute (Ed.) Methods of Soil Analysis, Part 1. SSSA, Madison, WI.