Sulfur for Soybean in Kentucky: An Update

By Dr. John Grove, Professor of Agronomy
Soils Research and Extension

Some years have gone by since I summarized UK research about soybean yield response to sulfur (S) here in Kentucky. The objective of this article is to update my previous summary, a meta-analysis which included research work done between 2007 and 2012, reported in 2013. The intent of these irregularly timed summaries is to determine whether there is evidence of: a) a need for more S nutrition research to develop an S fertilizer recommendation; or b) continued vigilance via crop plant tissue surveys. I thank, at the outset, Drs. Edwin Ritchey, Chad Lee, and Carrie Knott for their contributions to this new data set; and Dr. Eugenia M. Pena-Yewtukhiw for assistance with the meta-analysis. Financial support from the Kentucky Soybean Board and Mosaic Fertilizer is also acknowledged. The interpretation of the results/meta-analysis is entirely mine.

In the 2013 report, there were 23 valid (three or more replications of each treatment) comparisons involving a no-sulfur (-S) treatment and a sulfur-added (+S) treatment (first data column in Table 1), and for the 2013 to 2020 growing seasons, there were 25 such comparisons (third data column in Table 1). In the earlier report, individual studies were sited in three counties (Russell-11, Caldwell/UKREC-11, and Fayette/Spindletop-1). For this report (Table 1), the trials were in two counties (Caldwell/UKREC-21 and Fayette/Spindletop-4). I also developed a separate set of the data from 2007-2012 with the 12 comparisons done in Caldwell/UKREC and Fayette/Spindletop (second data column in Table 1).

The later, 2013-2020, trials did not always have S nutrition as the primary, or only, study objective, but particular treatment combinations permitted an evaluation of the benefit of added S to soybean yield. Added S rates ranged from 10 to 800 lb S/acre, with the high rate resulting from a trial where the impact of high rates of gypsum on crop use of subsoil water was being evaluated. The soybean variety was appropriate to the area and all sites were planted without prior tillage (NT). Sulfur addition was accomplished with gypsum; ammonium sulfate; potassium sulfate; or sulpo-mag (K-Mag; potassium magnesium sulfate). Other fertilizer nutrients were applied according to experimental protocols. Grain yield was determined by small plot combine harvest. The mean -S treatment yield ranged from nearly 45 to nearly 85 bushels/acre.

After grouping, the meta-analyses of the populations of soybean yield responses to S addition (+S) was summarized both numerically and graphically. The numerical parameter summaries for the three populations (2007-2012/all; 2007-2012/research farm only; and 2013-2020/all) in soybean yield response are found in Table 1. The graphical demonstration of the cumulative frequency distribution in the two larger data sets is shown in Figure 1.

In the numerical summary shown in Table 1, the 2007-2012 data sets, the means were not different from zero, indicating that, on average, there was no benefit to S addition to soybean. That is not the case with the 2013-2020 data set, whose mean is both positive and significantly different from zero. Interestingly, there has been a large shift in the average response seen on the two research farms, from -1.95% to +2.60%, a shift of +4.55%.

The cumulative frequency distributions in Figure 1 are distinctly different. The 2013-2020 data exhibit less ‘spread’ that the 2007-2012 data. The newer data are shifted to the right, indicating a shift to a greater probability of a positive soybean yield response to added S. In the 2007-2012 data set there were only 3 of 23 comparisons where the yield response to S addition was greater than 5% (my personal criterion for a likely net economic benefit). However, in the 2013-2020 data set there were 6 out of 25 such outcomes with a yield response above +5% and 13 of 23 comparisons where the response fell between 0 and +5%. There were fewer negative yield responses in the later data, and most of those fell between -5% and 0.

The shift in the pattern of soybean yield response to added S may have several connected causes. First, the 2013-2020 data come from a more spatially limited set of study locations – locations that do not routinely receive S fertilizer but are experiencing less S deposition from the atmosphere. The fields at the research station farms may have suffered less erosion, resulting in greater topsoil thickness and soil organic matter stocks that reduce the yield response to added S. In the 2007-2012 data set, sulfur fertilizer comparisons from trials conducted on research station farms were much less positive that those conducted elsewhere in Kentucky.

To conclude, I note that many of our yield responses to S addition have shifted to the positive but remain rather small and are likely uneconomical. However, that general shift is based in research applicable to a small portion of Kentucky’s soybean field space. A greater portion of that space needs to be evaluated as regards the need for S fertilization of soybean.