U of M Extension Home : U of M Home

Minnesota Crop News > 2001-2008 Archives

Soil test values for phosphorus (P) in many fields in Minnesota fall into the medium category. Using University of Minnesota guidelines, the medium category is defined at 11 to 15 ppm if the Bray or Mehlich III test is used. If the Olsen or bicarbonate test is used, the medium range is defined as 8 to 11 ppm.

Is this medium range adequate for optimum yields in corn and soybean production systems? Should soil test values for P be built up from the medium to high or very high categories? What guidelines should be used if soil test values for P are in the medium range? These are some of the many questions that are being asked by both crop producers and ag professionals as we work our way through the fall harvest of 2004.

Answers to these questions can be obtained from the results of trials conducted at the Southern Research and Outreach Center, Waseca and the West Central Research and Outreach Center, Morris. Phosphate fertilizer application at each site was based on two concepts.

The corn-soybean rotation was used and each crop was grown each year. In one concept, phosphate application rates were based on estimates of crop removal of P. The phosphate rates in the second concept were selected to match either the band or broadcast suggestions from the University of Minnesota. Crop yields were measured and soil samples were collected and analyzed for soil test P each year. The Bray P-1 procedure was used for soil samples collected from the Waseca location. The Olsen or bicarbonate procedure was used for soil samples collected from the Morris site.

The study was started in 1998 at Morris and 1999 at Waseca. For planning purposes, the expected yield at Morris was 140 bu. per acre for corn and 40 bu. per acre for soybeans. Expected yields were higher at the Waseca location (170 bu./acre for corn and 60 bu./acre for soybeans).

The corn and soybean grain yields at the Waseca location are summarized in Table 1. Rather than show yields for all years, yields from the first year (1999), final year (2003), and averages are listed.

Some explanation of the various treatments is appropriate. In the control, phosphate was not applied during the study. For the annual removal treatment, calculated removal for the anticipated yield of the corn and soybean crops was broadcast and incorporated before planting each year. For the biennial removal treatment, the calculated removal for both corn and soybean crops was broadcast and incorporated prior to planting the corn crop. The U of M broadcast suggestion for each crop to match the expected yields and a medium soil test for P was broadcast and incorporated each year before planting. Likewise, in the U of M banded treatment the banded suggestions for anticipated yield and medium soil test values for P were applied in a band near the seed at planting.

Table 1. Corn and soybean yields at the Southern Research and Outreach Center as affected by two concepts for fertilizer application.

The average yield for both crops for the 5 year period (1999 through 2003) were reasonably close to the expected yields. Statistical analysis of the data from the individual years as well as the yield averages shows that the treatment used had no significant effect on production. There are some differences in yields associated with the various treatments. These differences, however, are caused by natural variability in the research plot area and are not caused by the treatment applied.

Grain yield increases from phosphate application were not expected for the soybean crop. Small increases in corn yield might be expected at the medium soil test level; but, none were measured. All treatments received adequate rates of nitrogen and potash so that these two nutrients were not limiting production.

Soil samples (0 to 6 inches) were collected in the spring of 2004 and analyzed for P using the Bray procedure. The results of this analysis for the area where corn was planted in 2003 are provided in Table 2.

Table 2. Soil test P (Bray test) from samples collected (0 to 6 inches) in the spring of 2004 at Waseca.

Statistical analysis of the data in Table 2 shows that the differences in soil test P shown are the result of the treatment that was applied. The initial value was approximately 15.0 So, there was a small reduction when no phosphate fertilizer was applied. The reduction, however, was small and would not have substantially altered the suggested rate of phosphate over a period of five years.

Basing phosphate application on crop removal produced a substantial increase in the soil test for P. There was a less dramatic increase when the U of M broadcast suggestions were used; there was essentially no change in soil test P when U of M banded applications were used (16.5 ppm compared to 15.0 ppm).

In order to evaluate the impact of differences in recommendation concepts for calcareous soils, the study was also conducted at the West Central Research and Outreach Center at Morris (soil pH > 7.5). The corn and soybean yields are summarized in Table 3. Because of severe wind damage and subsequent lodging of stalks, corn grain yields were not measured in 2002. So, the average yields in Table 3 are for five years of corn production and 6 years of soybean production.

Table 3. Corn and soybean yields at the West Central Research and Outreach Center as affected by two concepts for fertilizer application.

The average grain yields of corn and soybeans over five years were higher than the original expected grain yields. The excellent growing conditions that were experienced were not anticipated.

Statistical analysis of the data shows that application of phosphate fertilizer had no effect on corn yield in 1998. The differences among treatments that were measured were random. Considering corn grain yields from 2003 and the average yields, application of any phosphate fertilizer increased grain yield when compared to the control. The concept used in determining the rate of phosphate to apply, however, had no significant effect on yield. Throughout the study, application of phosphate fertilizer had no significant effect on soybean yield. Results of past research have shown that this might be expected when soil test values for P are in the medium range.

Soil samples from this site were collected following harvest each year. The soil test P values for 1998 (the first year of the study) and 2003 (the last year) are summarized in Table 4.

Table 4. Soil test P (Olsen test) from soil samples (0 to 6 inches) collected in the fall of 1998 and spring of 2004 at Morris WROC.

Changes in soil test P were small when compared to the changes measured at Waseca. After six years of crop production, there was a small decline when phosphate was not applied (control). Except for the situation where the phosphate was applied in a band at planting, there was an increase in soil test P following the first year of phosphate application. These same small increases were measured for samples collected in the spring of 2004. In contrast to the Waseca location where soil pH is usually in the range of 6.5 to 7.0, changes in soil test P were less dramatic when the soil was calcareous. These lower soil test values illustrate the difficulty in building soil test values for P in calcareous soils using reasonably economic rates of phosphate.

Phosphate costs and returns to phosphate use are summarized in Table 5. Since the application of phosphate had no significant effect on yield at the Waseca location, there was no calculation of return to phosphate use at this site.

Table 5. Fertilizer recommendation concepts and the cost of phosphate application.

In calculating costs, the price of phosphate was set at $.25/lb. P₂O₅. To calculate returns, a value of $2.25/bu. of corn was used.

At Waseca, the fertilizer recommendations based on crop removal did produce substantial increases in soil test values for P. However, there were no added bushels of corn over a period of 5 years. So, for an additional cost of about $35 to $50 per acre, the only advantage was an increase in soil test values.

For the Morris location average corn yields were increased by the use of phosphate in all comparisons. The yield increases were small, but consistent. To calculate return, an average yield increase of 5 bu./acre/year was multiplied by $2.25/bu times 6 years. The cost of the phosphate was then subtracted from the total value of the yield increase. The smallest return was associated with crop removal applied on a biennial basis. The greatest return was produced by University of Minnesota recommendations applied in a band.

The results of this study conducted at two locations leads to the conclusion that it is more expensive to use crop removal as a basis for phosphate recommendations when soil test values for P are in the medium range. It’s not surprising that the crop removal concept is thought of by many when phosphate fertilizer rates are being considered. It appears, at first, to be a logical concept.

However, use of the crop removal concept for phosphate recommendations assumes that the soil is not able to supply and P to the growing crop. If this were true, soil test values for P should be 0.0 for some fields. Unless the Bray test is used on highly calcareous soils, there is seldom a report of 0.0 for soil test P.

The results of this study conducted at two locations using a corn/soybean rotation show that use of crop removal as a basis for making phosphate recommendations can be expensive.