Summary

Accurate fertilizer N recommendations are a primary responsibility of the land grant university and are essential for optimizing profitability for the farmer and minimizing loss of N to the environment. Twenty nine small-plot and field-size strip experiments were conducted on farmer's fields in southern Minnesota from 1989 through 2001 to validate University of Minnesota N recommendations for corn after soybeans. Corn responded to fertilizer N at all but two sites. The N rate needed to optimize fertilizer N in the 14 small-plot studies ranged from 0 to 150 lb N/A and averaged 71 lb/A across all studies when using the least significant difference method (LSD). The economically optimum N rate (EONR) was 71 and 85 lb N/A for the linear response plateau (LRP) and quadratic response plateau (QRP) models, respectively. In the field-size strips, yields were optimized at N rates ranging from 60 to 120 lb N/A with an optimum N rate of 87 lb/A using the LSD method. The EONR was 84 and 100 lb N/A when using the LRP and QRP models, respectively. Although different experimental procedures were used, optimum N rates determined by the small-plot and field-strip studies were very similar. These results suggest  the 120-lb N rate presently recommended by the University of Minnesota is sufficient for expected corn yields ranging from 150 to 174 bu/A.

Introduction

Nitrogen (N) is an essential plant nutrient that is applied for corn in greater quantity than any other fertilizer in Minnesota. Applying the proper rate of N to achieve the optimum corn yield is essential if producers are to maximize economic return and minimize N losses to the environment. Profitability per acre can be sacrificed if inadequate rates of N are applied. On the other hand, using too much N increases the potential for losses of excessive nitrate to ground and surface water while also reducing economic return to the grower.

Hundreds of field studies to determine the optimum rate of fertilizer N for corn following soybeans have been conducted by University soil scientists in Minnesota since the early 1960's. These studies have led to a set of N rate recommendations that are updated periodically as N calibration studies are continually conducted using recently developed corn hybrids with greater yield potential.

Determining the optimum rate of fertilizer N for a crop involves assessing the amount of N becoming available from the soil and then determining the amount of fertilizer N needed to supplement the soil N to meet the crop's total N requirement. Because uncontrollable factors such as precipitation and temperature affect the release of N from soil organic matter and the amount of N needed by the crop in any particular year or location, the optimum rate of fertilizer N will likely vary from location to location and from year to year. Thus, one quickly sees the need for continually conducting studies to assess the magnitude and consistency of corn yield response to fertilizer N given the variations in weather and changing genetic potential.

In recent years, University of Minnesota N recommendations for corn have frequently been questioned by the fertilizer industry and some farmers and advisors as being too low for optimum yields and profitability. The fact that many of the studies were conducted on University experiment station land was often cited as a contributing factor to the recommendations being lower than deemed appropriate by the dealer/advisor/farmer. Thus, N rate calibration studies located on farmer's fields were strongly encouraged. The following reports the results from these on-farm studies conducted for the period 1989 through 2001.

Experimental Procedures

Fourteen small-plot studies were conducted in nine southern Minnesota counties from 1989 through 1999. These small plots measured from 10 to 15' wide and 40 to 60' long and were replicated four to six times at each site. The year, county, soil information, method and time of N application, and N source for each site are shown in Table 1. Tillage, planting, pesticide application, cultivation, and hybrid and planting rate selection were all conducted by the farmers. University scientists applied the fertilizer N, hand-harvested the yields, and collected other appropriate field data (weather, past cropping and nutrient history, etc.). Best management practices including spring application and incorporation within 1 day were used to minimize N losses.

Because concern was being expressed by some dealers, advisors, and farmers regarding the relevance of yield responses from small-plot studies, 15 field-size strip studies were conducted in five counties in southern Minnesota from 1997-2001. All sites were chosen by the farmer in conjunction with either local crop advisors, dealers, or state agency personnel. Site characteristics, method and time of N application, and N source used are shown in Table 2. The concentration of sites in Nicollet Co. was due to increasing nitrate concerns within the St. Peter Wellhead area and the presence of a multi-agency project to examine N use for corn in that area. Spring or sidedress application was used at 10 sites and fall anhydrous ammonia plus N-Serve was used at 4 sites. Fertilizer N was applied by the dealer or farmer in strips matching the applicator width (30 to 60'). Strip length ranged from about 400' to more than 1200'. All fertilizer N rates were replicated three times except for the two earliest sites (AA and BB) where only two replications were used. In addition, no zero N (0 lb N/A) strips were included at sites AA and BB and only one control strip was included at sites GG and KK. Because of these limitations, no statistical analyses were performed for these sites.

At each field-strip site, tillage, planting, pesticide application, and hybrid and planting rate selection were all conducted by the farmers. The strips were combine-harvested using yield monitors. All yield data were collected by the farmer and/or consultant. The strip-average yields were then provided to the authors of this publication for statistical and economic analyses.

Statistical analysis to determine the least significant difference (LSD) at the 90% probability level was conducted on yield data from all sites that had three or more control (0 lb N/A) plots/strips. Analysis to determine the economic optimum N rate (EONR) was performed on the yield data from all sites except I (only 4 N rates), AA, and BB. Based on a N price of $0.15/lb and a corn price of $2.00/bu, linear response plateau (LRP) and quadratic response plateau (QRP) models were used to calculate the EONR for each of the 26 sites.

Results and Discussion

Small Plot Studies

Corn yields for each of the N rates at all 14 sites are shown in Table 3. The optimum N rate for each site was determined using three statistical methods also shown in Table 3. The LSD value gives the smallest yield difference among treatments that is statistically significant at the 90% probability level. The optimum yield using the LSD is shown in bold print for each site. A response to fertilizer N was not obtained at two sites (A and J) even though yields at these sites ranged between 150 and 180 bu/A. This can be explained by the release of N from the soil organic matter. Greater availability of soil N usually occurs following dry years and long-term applications of N (fertilizer and/or manure) in excess of crop removal. Using the LSD approach corn yields were optimized at the 30-lb N rate at 1 site (L), at the 60-lb N rate at 4 sites (C, D, H, and K), at the 90-lb N rate at 5 sites, at the 120-lb rate at 1 site (F), and at the 150-lb rate at 1 site (E). Averaged across all 14 sites, the optimum N rate, using the LSD approach, was 71 lb N/A (82 lb N/A if using only the 12 responding sites) to produce a yield of 169 bu/A. The EONR averaged across all 12 sites that fit the LRP model was 71 lb N/A but ranged from 0 to 107 lb N/A. The QRP model fit 13 sites giving an average EONR of 85 lb N/A with a range from 0 to 140 lb N/A.

^1/   EONR = Economic Optimum N Rate based on 15¢/lb N and $2.00/bu corn. 

The yields were averaged across the 13 sites for each of the N rates and are plotted in Fig. 1. The EONR using the QRP model for this 13-site average is 105 lb N/A, which gave a yield of 173 bu/A. This 19-lb increase over the average of the EONR's in Table 3 is largely due to the two non-responding sites and the omission of the 30-lb N rate in calculations based on Fig. 1.

These data clearly show the variability encountered among site-years when determining an optimum N rate for corn. University scientists assess this variability and consider it in when making fertilizer N recommendations. Usually the recommendations tend to be slightly greater than a very strict mathematical interpretation of the response data would suggest. This slight cushion protects the farmer from risk of yield and profitability loss under unforeseen conditions. In this case, a N recommendation of 120 lb N/A was more than needed for optimum yields at 13 of 14 sites when using the LSD approach, at all 13 sites when using the LRP approach, and at 10 of 13 sites when using the more liberal QRP approach.  Based on these yield responses to N in small-plot studies, the 120-lb N rate, presently being recommended by the University of Minnesota for expected corn yields of 150 to 174 bu/A on these soils, was sufficient to optimize yield and profitability at 13 of 14 sites. Moreover, this rate of N could be considered excessive at one-half of the sites where a yield response to N did not occur at rates greater than 60 lb N/A.

Field-size strip studies

Corn yields for each of the 15 field-size strip studies are shown in Table 4. Least significant differences were calculated for 11 of the sites while EONR's were calculated for the 13 sites with 0-lb control plots. All 13 sites responded to fertilizer N. The yield shown in bold type for each site represents the N rate that optimized corn yield using the LSD approach. Yields were optimized at the 60-lb N rate at 3 sites, at the 90-lb rate at 6 sites, and at the 120-lb rate at 2 sites. Averaging these optimum N rates and associated yields across all 11 sites resulted in an optimum N rate of 87 lb/A and an average yield of 150 bu/A.

^1/ EONR = Economic Optimum N Rate based on 15¢/ lb N and $2.00/bu corn.

^2/  Only two replications, no statistical analysis conducted. Actual N rates used shown by superscript numbers.

^3/ The 0-lb control strip was not replicated at these sites. Thus, a statistical analysis to determine the LSD was not conducted.

The EONR for the 13 sites using the LRP method ranged from 50 to 137 lb N/A with an average of 84 lb N/A. Twelve of the 13 sites had EONR's less than 120 lb N/A. When using the more liberal QRP method, EONR's ranged from 55 to 169 lb N/A and averaged 100 lb N/A  for the 13 sites. Eleven of the 13 sites had EONR's less than 120 lb N/A.    When the yields were averaged across the 13 sites for each N rate (Fig. 1), the EONR using the QRP method for this 13-site average was 99 lb N/A, which gave a yield of 152 bu/A.

Similar to the small-plot studies, these field-size experiments also demonstrate the site-to-site variability associated with arriving at an optimum N rate for corn. But in the aggregate, a 120-lb N rate was sufficient to optimize corn yield at all sites, when using the LSD approach. Using the LRP and QRP approaches, the 120-lb N rate was sufficient to optimize economic return at 12 of 13 sites and 11 of 13 sites, respectively.

Small-plot vs Field-size strip studies

Although the experimental procedures used were much different for these two types of field studies, the yield responses to fertilizer N were remarkably similar between the small-plot and field-size strips (Fig. 1). When averaged across all sites, the QRP derived EONR's for the small plots and field-size strips were 105 and 99 lb N/A, respectively. These results suggest that fertilizer N rate recommendations are equally good, regardless whether the yield response data come from small plot or field-size strip studies.

Acknowledgement

The authors extend grateful appreciation to the Minnesota Corn Research and Promotion Council and to the St. Peter Wellhead Protection Project for their financial assistance in the conduct of this project. Sincere thanks is given to the many farmer-cooperators who participated in the project and to the consultants/local advisors (Blue Earth Agronomics - Karyn Wassman, GMA Agronomics - Kevin Bitterman, McPherson Crop Management - Bernie Paulson, and United AgTech - Steve Sodeman) whose collection of field strip data was vital to this project. We also greatly appreciate the technical assistance provided by Jeff Vetsch, Andy Scobbie, Brian Anderson, and David Groh in the collection and analyses of the data.