Timing of P application for corn and soybean production
Start Date: 2019-2021
Principal Investigators: Daniel Kaiser
Organization: University of Minnesota
- Both Bray-P1 and Olsen soil P in June was impacted consistently by P application rate and timing. Soil test P was greater in June following spring P application as indicated by significant P timing main effects and significant rate by timing interaction in sites with high carbonates.
- Corn leaf P concentration were consistently impacted by P application rate in 2019 and 2021 and impacted by timing in 2020.
- Soybean trifoliate P concentration was not consistently impacted by P application rate or timing.
- Corn and soybean yield were impacted by P application rate at 5 of 9 corn and 3 of 9 soybean locations.
- Corn grain yield was greater when P was applied in Spring and soybean grain yield was greater when P was applied in fall. Grain yield was increased when up to 60 lb P2O5 were applied for corn when applied in the spring, and 90 lbs P2O5 applied in the fall produced similar yield as 45 lbs P2O5 applied in spring.
- Soybean grain yield across the nine location was linearly increased when up to 90 lbs P2O5 were applied.
- Corn grain yield was at maximum when June Olsen soil test P was 6 ppm while soybean grain yield was at maximum when June Olsen soil test P was 20 ppm.
- Grain P concentration and P removed in the harvested grain were inconsistently impacted by P application rate and timing.
- Soybean grain quality (protein and oil concentration) were not consistently impacted by P application rate and timing.
- Corn grain yield was at maximum when leaf P concentration was 0.30% at V10 and 0.23% at R1.
- Soybean grain yield was at maximum when R1 trifoliate P concentration was 0.38%.
Phosphorus is a plant nutrient which, if deficient, can significantly limit crop growth and development. Phosphorus is considered a primary macronutrient. Primary macronutrients are elements that are essential for plant growth which more commonly require fertilizers to be applied to satisfy crop requirements. Orthophosphate is the form of phosphorus taken up by plants. The concentration of orthophosphate in the soil solution is low as orthophosphate is highly reactive with metal elements. Iron, aluminum, and calcium all can react with orthophosphate creating compounds varying in solubility. The ion which reacts with orthophosphate depends greatly on the pH of the soil.
Soils in major cropping regions in Minnesota predominantly formed under calcareous parent materials. Carbonates deposited in the material left following glaciation are still present near the soil surface in areas of the state. These soils with greater carbonate and calcium contents present challenges when managing phosphorus as it is difficult to increase available soil test phosphorus of calcareous soils. Soils with high calcium contents can fix phosphorus. Fixation is a process where orthophosphate reacts with calcium forming compounds like di- and tri-calcium phosphate. While fixed phosphorus is not technically lost from the soil it is rendered unavailable for plant uptake.
The relative rate of phosphorus fixation is not known in soils. Short term P sorption tests can be run to determine the amount of P which a soil will sorb, which can be substantial for some calcareous soils. Management of fertilizer P is common in the fall which gives more time for P to react and potentially bind phosphorus. Studies have been conducted focused on timing of P application, but many were conducted in soils with a neutral to slightly alkaline pH which did not contain appreciable amounts of calcium carbonate.
In a previous study funded by AFREC, on-farm strip trials established to determine corn and soybean response to a single rate of phosphorous fertilizer showed that a high rate of P applied one year can have multiple years’ benefits for crops in a two-year rotation. The exception was one location with a calcium carbonate equivalency of 20% where there was a yield benefit to P applied both years for a two-year corn-soybean rotation and P applied the previous year had not impact on the crop grown. Fall application provides more flexibility for farmers but there are questions as to whether spring is better under some circumstances.
We collected approximately 1440 GHG samples, 260 soil samples, and 240 water samples during the 2021 growing season. These samples were analyzed for nitrogen, carbon, and phosphorus.
The objective of this study is to establish whether there is a difference between fall and spring application of P fertilizer for corn or soybean production and whether potential differences may be tied to calcium carbonate content of the soil.
materials and methods
Field trials were established in farmer fields and at ag experiment stations in Minnesota (Table 1). Locations were targeted to test Low by either the Bray-P1 or Olsen tests (< 10 ppm Bray-P1 or <8 ppm Olsen P). Sites with a calcium carbonate equivalency of >5% CCE were given preference due to a greater capacity for P fixation but the exact CCE was not known until after trial establishment.
A split plot design was used where main plots consisted of four P rates and sub-plots consisted of timing (Fall or Spring). The four P rates were 0, 30, 60, and 90 lbs P2O5 per acre applied as MAP (11-52-0). All treatments were replicated four times. Nitrogen supplied with MAP was balanced with AMS. Gypsum was used to balance sulfur applied by AMS such that all plots received similar rates of N and S when P treatments are applied. Calcium supplied by gypsum was not expected to impact corn or soybean yield due to excessive levels of calcium already in the soil.
Corn and soybean will be the two crops utilized for this study. Additional crop species are not used as corn and soybean should provide sufficient information on potential differences in responses based on fertilizer timing which can be translated to additional crops. A total of three trials were established for each crop each year (6 trials total per year).
Soil samples (0-6”) were collected from each main block prior to fall treatment application, were air dried, ground, and analyzed for P by Bray-P1, Olsen, and Mehlich-3 P tests. Samples were additionally analyzed for calcium carbonate equivalency (modified pressure calcimeter method) and pH (1:1 soil:water). Additional 0-6” soil samples were collected in June from all plots to assess change in Bray-P1 and Olsen soil test P after treatment application.
Leaf samples were collected from each corn plot at V8-V10 (uppermost fully developed leaf) and R1 (leaf opposite and below the ear), and for soybean at the R1-R2 growth stage (uppermost fully developed trifoliate). All soybean plots will be harvested with a small plot combine. Corn will either be harvested with a plot combine or by hand. Soybean grain yield is reported at 13% moisture and corn grain yield is reported at 15.5%. Grain samples were collected from each location and analyzed for total P concentration for both crops while soybean was analyzed for protein and oil concentration in grain by NIR.
results and discussion
Location data are summarized in Table 1. Sites were selected to have low initial soil test P (STP) and measurable carbonate levels. The exception to this was Lamberton (both years) which was included due to a very acidic pH which is a good comparison to the remaining sites to determine the impact of free iron and aluminum on the retention of soil test P. The only location which tested above the medium STP class was Benson. To establish the location, we targeted high pH zones which tend to have low soil P, but that was not the case for Benson, Stewart 2019, and Danvers. The remaining sites were all within targeted parameters. Calcium carbonate equivalency (CCE) was highest at Crookston and Stewart, both in 2019. At the remaining four locations there was measurable CCE (except for Lamberton) but the levels were lower than anticipated.
2019 Data Summary
Table 2 and 3 summarizes main effects and main effect interactions for the ANOVA for the measured variables for the 2019 corn and soybean trials, respectively. Tables 4 and 5 summarize the phosphorous (P) rate and timing main effects for corn and soybean, respectively. Interactions were generally not significant, and the interaction data are not summarized for the majority of the measured variables. A lack of a significant interaction is an indication that there is no impact of P timing on fertilizer use by either crop. Exceptions however will be noted.
Soil samples were collected in June to assess potential loss of P availability following the fall application. In general, main effect significance was similar when P was analyzed by either the Bray P1 or Olsen P tests. The two exceptions were the corn location at Crookston and soybean location at Benson where the high level of carbonates neutralized the Bray solution and resulted in Very Low soil P tests which did not change with application rate. Benson was the only location where neither main effect was significant which could be due to greater variability in soil P due to the higher initial P soil tests. For Lamberton and both Morris sites, both main effects were significant for both soil tests while both main effects were also significant for the Olsen P test only at Crookston and Stewart. Since most sites were high pH where the Olsen P test is typically used, the Data in Figures 3 and 4 will be discussed for the corn and soybean locations, respectively.
The interaction between P source and rate was significant for the change in Olsen soil test P at the Morris corn and soybean locations. This indicates a difference in slope in the relationship between soil test P change and P application rate for the fall and spring applications. In all cases, Olsen soil test P was greater in June following spring application which is not surprising considering the greater time the P had to react with the soil from the fall application. From the relationships in Figures 3 and 4 it is surprising that more of the interactions were not significant. However, since most of the timing main effects were significant we have strong evidence that P is being tied up following fall application including at Lamberton which had an acidic pH. Sites like the soybean trial at Morris exhibited very little change in soil test P when up to 60 lbs P2O5 were applied.
Corn data are summarized in Table 5. Phosphorus application rate more consistently impacted measured variable compared to P application timing. Corn leaf P at V10 and R1 were generally increased linearly at most locations except for R1 leaf P at Morris which was no affected by P application. The only timing effect occurred for R1 leaf P concentration, which was greater for the fall application, but the difference was negligible. Expected concentration of P in corn leaves at R1 is between 0.2 and 0.4 %. All locations tested within that range with the lowest concentrations of P combing back at 0.23 %.
Corn grain harvest moisture was inconsistently affected by P application. Yield, however, was increased by P application at Lamberton and was impacted by timing at two locations, Lamberton and Morris. Corn grain yield favored spring application at both locations which were significant. The P rate by timing interaction was also significant at Morris indicating a significant impact of P rate that varied based on timing. The interaction is not shown in figures, but the analysis of data showed no yield difference between timing at the 0 and 90 lb application rates while spring application resulted in greater yield when 30 or 60 lbs P2O5 were applied. A regression was not run on the data but the ANOVA would indicate that application of 90 lbs would result in similar yield and that a small reduction in P could be taken if the P were applied at a rate greater than 60 lbs P2O5. The lack of a response to P at Crookston was likely due to high levels of Goss’ Wilt which reduced yield potential in the trial. Overall yield potential was significantly lower at Crookston compared to the remaining sites. Corn grain P concentration and P removed in the harvested grain were also inconsistently impacted by P rate and timing effects.
Soybean data are summarized in Table 6. Phosphorus rate and timing did not affect the concentration in the uppermost fully developed trifoliate at R1. In addition, P concentration in the harvested grain, soybean protein and oil concentration were inconsistently impacted by treatments.
Soybean grain yield was affected by P rate at Morris and Stewart and was not impacted by timing. Soybean yield was increased when up to 60 lbs P2O5 were applied at both locations. Removal of P in the harvested grain was increased by P application rate at all three locations, even Benson where yield was not affected which contrasts with corn where P removal was seldom impacted. Yield increase to P was greater at Morris where yield was nearly doubled with the 60 lb P2O5 application rate compare to the control, and grain yield was increased 2 bu/ac at Stewart.
2020 Data Summary
Table 7 and 8 summarizes main effects and main effect interactions for the ANOVA for the measured variables for the 2020 corn and soybean trials, respectively. Tables 9 and 10 summarize the phosphorous (P) rate and timing main effects for corn and soybean, respectively. Interactions were generally not significant for soybean and mostly for corn. The interaction data are not summarized for many of the measured variables. A lack of a significant interaction is an indication that there is no impact of P timing on fertilizer use by either crop. Exceptions however will be noted.
Soil samples were collected in June to assess potential loss of P availability following the fall application. In general, main effect significance was similar when P was analyzed by either the Bray P1 or Olsen P tests. There was only one location in 2020, Stewart, where carbonates were relatively high, and the Bray P test was not impacted. There were three significant interactions between time and rate for Bray P (Crookston, Lamberton, and Stewart) and there was a significant interaction for Olsen P only at Crookston and Stewart (Figures 3 and 4). The Bray-P interaction was a result of no change in June Bray-P when P for P applied in the fall. The same was true for the significant interaction for Olsen P at Crookston but for Stewart Olsen soil test change differed between timing for the 30 and 60 lb application rates but not for the 0 and 90 lb rate. Timing did differ but, in all cases, the two-way interaction was also significant as well.
Corn data are summarized in Table 9. Phosphorus application rate more consistently impacted measured variable compared to P application timing. Corn leaf P at V10 and R1 were impacted by timing but not by rate, which contrasts with 2019 where rate generally impacted leaf P concentration.
Corn grain harvest moisture was greater when P was applied in the Fall at Crookston and Lamberton but did not differ at Morris (Table 9). At Morris, grain moisture decreased with increasing P application rate. Grain yield as greater with spring application at both Crookston and Lamberton while P application rate increased yield at Morris with 60 lb P2O5 application rate. In all cases decreased moisture in the grain at harvest was accompanied by an increase in grain yield.
Soybean data are summarized in Table 10. Phosphorus timing did not affect the concentration in the uppermost fully developed trifoliate at R1, while trifoliate P concentration increased with increasing rate of P applied at Danvers. In addition, P concentration in the harvested grain, soybean protein and oil concentration were seldom impacted by treatments. The exception was grain P and protein concentration at Morris which was increased by the rates of P applied.
Soybean grain yield was affected by P rate at Danvers (Table 10). Rate and timing of P application did not affect soybean grain yield in all other cases. Yield was relatively low at Morris and Stewart. The low yield at Morris could not be explained while two separate hail events in July reduced soybean yield potential at Stewart. Grain P removal was increased by P application rate at both Danvers and Stewart. At Danvers, soybean grain yield and grain P removal were greater for the 90 lb and there was no difference among the remaining treatments. At Stewart, all rates of P increased P removal similarly compared to the no P control.
2021 Data Summary
Table 11 and 12 summarizes main effects and main effect interactions for the ANOVA for the measured variables for the 2021 corn and soybean trials, respectively. Tables 13 and 14 summarize the phosphorous (P) rate and timing main effects for corn and soybean, respectively. Interactions were generally not significant for soybean and mostly for corn. The interaction data are not summarized for many of the measured variables. A lack of a significant interaction is an indication that there is no impact of P timing on fertilizer use by either crop. Exceptions however will be noted.
Soil samples were collected in June to assess potential loss of P availability following the fall application. In general, main effect significance was similar when P was analyzed by either the Bray P1 or Olsen P tests. Carbonates were high at one corn location, Crookston, and at two soybean locations, Holloway, and Stewart. Carbonates in the soil impacted the Bray-P1 test more significantly at Crookston and Holloway and had less of an impact at Stewart. There were three significant interactions between time and rate for Bray P and Olsen P only at Morris. However, P application increased STP, but the increase was greater for spring applied P (Figures 5 and 6). Timing main effect was generally significant, and the two-way interaction wasn’t always significant when the timing main effect was significant. When timing was significant spring P application always resulted in a higher average STP concentration compared to fall.
Corn data are summarized in Table 12. Phosphorus application rate more consistently impacted measured variable compared to P application timing. Corn leaf P at V10 and R1 were impacted by timing and by rate at all locations except for the R1 leaf P concentration at Crookston like 2019 were rate generally impacted leaf P concentration. Moisture in the harvested corn grain was seldom impacted by P timing or rate. Phosphorus concentration in grain and P removed by harvested grain were impacted by timing and rate to Lamberton and Morris, but not at Crookston. Corn grain yield was impacted by P rate at Lamberton and Morris but timing only impacted corn grain yield at Lamberton. Unlike other years, corn grain yield was greater with the fall application at Lamberton. Grain yield was relatively low due to dry weather conditions at Lamberton in 2021 so the greater impact of fall P application may be a result of better mixing of P in the soil profile with more aggressive fall tillage. It’s likely that more of the P with the spring application may be closer to the soil surface in soils that are drier that may have limited P uptake. This cannot be proven with the data generated.
Soybean data are summarized in Table 13. Phosphorus timing did not affect the concentration in the uppermost fully developed trifoliate at R1, while trifoliate P concentration increased with increasing rate of P applied at Morris. Phosphorus concentration in the harvested grain was impacted by P rate at all soybean locations while timing only impacted grain P concentration at Morris and Stewart. Soybean protein and oil concentration were seldom impacted by treatments. Soybean grain yield was not impacted by P rate or timing at any location.
year 3 overall summary
A summary of corn and soybean P response across the six individual locations is given in Figure 7. In both cases P rate and timing significantly differed. In addition, the P rate by timing interaction was significant for corn only which was because timing differed but only for the two highest rates of P. For fall application, corn grain yield increased linearly to the 90 lb application rate which yield was increased up to 60 lbs of applied P for the spring application. The yield produced by the 90 lb fall rate was roughly equivalent to 45 lbs of P2O5 applied in the spring. The data indicates that for corn, P application in the spring should be favored with low P soils and that some of the P was likely tied up in the soil and rendered unavailable following application in the fall.
Soybean yield was greater for fall application compared to spring (Figure 7). Soybean grain yield was increased linearly up to the 90 lb P2O5 application rate across the sites. There was no interaction between rate and timing so the greater yield for fall application occurred regardless of the rate of P applied. This indicates that P in the fall is favored for soybean while P in the spring if favored for corn.
June Olsen soil test P (STP) was regressed with relative corn and soybean grain yield (Figure 8). There was a better relationship between June STP and corn yield as there was much greater variability in relative soybean grain yield. For corn, relative grain yield was at maximum when Olsen P was 6 ppm which is in the medium soil P classification. In contrast, soybean grain yield was at maximum when soybean grain yield was near 20 ppm which is the upper end of the high soil test classification. Responses were separated by fall and spring application but the soil test at maximum yield was similar for both corn and soybean. The STP level that maximized soybean yield was much greater than expected but with the amount of variability in the data it is likely that the actual STP concentration needed to maximize yield is much closer to that of corn.
Figure 9 summarizes the relationship between relative corn yield and V10 and R1 leaf P concentrations. There was a relationship between leaf P concentration and relative yield in both cases. For uppermost fully developed leaves at V10, grain yield was maximized when leaf P concentration was near 0.3%, while the concentration of P needed for maximin yield for the leaf opposite and below the ear was 0.23% at R1. The relationship between the concentration of P in the uppermost fully developed trifoliate in soybean at R1 was more poorly correlated with relative soybean yield. However, soybean grain yield was at maximum with R1 trifoliate P concentration was near 0.38% (Figure 10). When the study is completed, I will compile data with other trials upon completion of the study. The relationship between soybean grain yield and trifoliate P concentration has generally been poor compared to corn leaf P concentration.