Illinois
Fertilizer Conference Proceedings
January 21-23, 2002
| Home | 2002 Index | Search |
| Home | 2002 Index | Search |
E.C. Varsa, S.A. Ebelhar, T.D. Wyciskalla, and C.D. Hart1
Corn producers collecting yield monitor data over a number of years usually observe patterns of productivity variability in fields that are somewhat repeatable year after year. That is, high- and low-yielding areas tend to assume similar patterns from year to year. It seems plausible, then, that areas of fields that vary in productivity should be fertilized differently, as crop removal of nutrients also would vary. Most producers, however, apply uniform rates of fertilizer to their fields, even though known productivity differences occur.
It would seem reasonable, then, that varying the nitrogen (N) rate according to the soils' productivity potential should improve N use efficiency. That is, higher-producing regions should receive more fertilizer N because of higher yields, and lower productivity portions of fields should receive less N because of lower yields. A number of previous reports (Carr et al., 1991; Redulla et al., 1996; Sawyer, 1994) have suggested this or similar approaches to whole-field fertilization.
With the advent of variable rate technology (VRT) and recent advances in fertilizer equipment, application rates can be tailored and varied as one traverses the field. It is the objective of this research to determine if agronomic, economic, and environmental benefits can be obtained by varying N application rates across the field as differing productivity areas are encountered. This would be compared with the standard practice of uniform N application based upon the average yield for the whole field. The research contained in this report is described in two parts. First, a whole-field comparison of two variable N application methods will be made with a fixed (uniform) rate of N application. A second phase of this research will evaluate response of corn to rates of N within selected historical low-, medium-, and high-productivity regions of the field as a small plot study.
Mr. Kelly Robertson, a farmer and crop consultant located in Franklin County, Illinois with at least five years of yield monitor data on the fields of his farm, has agreed to collaborate with us for a second year of this study. For the 2001 cropping season, he had a 39-acre field available for use in these studies that was soybean in 2000 and which has been in a cornsoybean rotation for 10 years. The dominant soil type in the field was a Cisne silt loam with lesser amounts of Hoyleton silt loam. The field's topography is quite flat, with a slope averaging from 0 to 2 percent. Adequate drainage is a major limitation in this field, and surface ditching is practiced where possible to remove excess water. Tile drainage is not practiced because of the restrictive claypan layer in the Cisne soil.
Figure 1 shows a map giving the outline of the field and the normalized yields broken down into three general productivity categories. The low-yield regions were identified as those with normalized yields that were 95 percent or less, medium-yield (or average-yield) regions of the field had normalized yields that ranged from 95 to 105 percent, and high-yield regions had normalized yields that exceeded 105 percent. The areas of the field that were in the low-, medium-, and high-yield categories were approximately 33, 32, and 34 percent, respectively. The average corn yield for this field (excluding drought years) over the past 10 years (five crops of corn) has been about 165 bushels per acre.
Twenty-one points for sampling were GPS-identified in the gridded map of the field (each grid being 60 ft. by 60 ft.). The normalized yield and the number of sampling points in each yield category was as follows: <60 percent = 1 point; 60-85 percent = 4 points; 85-100 percent = 6 points; 100-110 percent = 5 points; and 110-120 percent = 5 points (see Figure 1). Soil samples (0-8 inches) for fertility and crop parameter measurements were collected at each central, geo-referenced point itself and 30 feet to either side of the central point, perpendicular to crop row direction, for a total of three subsamples at each sampling point. This allowed sampling of the three fertilizer N rates being compared at the varying normalized yields occurring at each sampling point. Soil test averages for the field, based upon the 21 sampling points, were pH = 7.04 (range 6.55- 7.50) and exchangeable K = 194 lb/ac (range 147-395 lb/ac). Soil test P values were unavailable at the time of this report preparation. These geo-referenced sampling points were also the locations for other crop and soil evaluations, including ear leaf N sampling at silking, stalk nitrate at maturity, grain yield and moisture, and post-harvest soil nitrate analysis to a 3-foot depth.
Strip comparisons of variable rate N and uniform N application were based upon a field average yield of 165 bushels per acre. The uniform N application rate was calculated by multiplying yield x 1.2 less a soybean credit of 40 lb N per acre [(165 x 1.2) - 40 = 158 lb N per acre].
Two computational methods of variable rate N (VRN) were used as a comparison with the uniform rate of N. The variable rate application method used in 2000, referenced as VRN-Old, is the first method. It is based upon varying the N rate with the normalized, historical yields as they occur in the field. The formula for VRNOld is 1.2 x normalized (proven) yield within a cell less a soybean credit of 40 lb N/ac. This method essentially reduces N rates where proven, established yields are less than 100 percent and increases N rates when proven yields exceed 100 percent.
The second variable rate N method, referenced as VRN-New, essentially reverses the process of VRN-Old. That is, it increases N rates in lower productivity areas (less than 90 percent) and decreases N rates when normalized yields exceed 100 percent. The formula for VRN-New is 1.2 x normalized (proven) yield - 40 lb N/ac soybean credit for 90-110 percent yield levels; 1.0 x normalized (proven) yield - 40 lb N/ac soybean credit for proven yields greater than 110 percent; and 1.4 x normalized (proven) yield - 40 lb N/ac soybean credit for proven yields less than 90 percent.
Figure 2 shows a graph of the predicted rates of N application by the three methods. Variable application of N (as anhydrous ammonia) was accomplished with a controller on the applicator programmed in synchrony with a prescription map of soil productivity indexes (normalized yield map) on a computer in the tractor. Both uniform and variable rate strips were 30 feet wide (12 rows) for the entire length of the field, excluding head lands. The tool bar was equipped with shanks spaced between each row, and the anhydrous ammonia was applied as a side dressing to the corn at the five-leaf stage of development.
In a selected portion of the field, where normalized yields of <90 percent (low), 90-110 percent (average), and >110 percent (high) productivity were closely contiguous to each other, a small-plot, intensive N rate study was conducted within each productivity region. N rate treatments selected were equivalent to 0.8, 1.0, 1.2, 1.4, 1.6, and 1.8 lb N per bushel expected yield, plus a zero-N check. Additionally, nitrapyrin as Stay-N 2000 was included with the applied N for the 0.8, 1.0, 1.2, 1.4, and 1.6 lb N per bushel application rate treatments. A summary of those treatments appears below.
| Nitrogen Treatment (lb N per bu) |
Nitrapyrin | N Application Rate (lb N per acre) | ||
|---|---|---|---|---|
| Normalized Yield Productivity | ||||
| Low | Average | High | ||
| Check | - | 0 | 0 | 0 |
| 0.8 | - | 60 | 90 | 120 |
| 1.0 | - | 90 | 120 | 150 |
| 1.2 | - | 120 | 150 | 180 |
| 1.4 | - | 150 | 180 | 210 |
| 1.6 | - | 180 | 210 | 240 |
| 1.8 | - | 210 | 240 | 270 |
| 0.8 | + | 60 | 90 | 120 |
| 1.0 | + | 90 | 120 | 150 |
| 1.2 | + | 120 | 150 | 180 |
| 1.4 | + | 150 | 180 | 210 |
| 1.6 | + | 180 | 210 | 240 |
All nitrogen treatments were replicated three times within a randomized complete block arrangement in each of the productivity zones. Individual plot sizes were 15 feet (six rows) wide by 35 feet long. The nitrogen source was 28 percent UAN solution knifed in with an alternate row- spaced shank applicator. Application of N was made at the five-leaf stage of development. Measurements taken included stand counts, ear leaf N tissue at silking, and grain yield and moisture at maturity. Table 1 provides additional experimental details concerning dates, cropping information, and precipitation.
The growing season was nearly ideal in 2001 for corn production at this location. As seen in Table 1, rainfall was below normal in April, allowing early season field work to proceed rapidly and corn planting to occur by late April. Near normal rainfall was received throughout the growing season, with the important month of July being nearly two inches above normal. No particular heat stress, insect/disease pests, or severe storms impacted the crop.
Corn yield results closely paralleled the previously observed pattern of soil productivity in the field (Table 2). Yield increased from an average of 127 bu/ac for 60 percent of normalized yield (NY) to 184 bu/ac for the 110 percent NY regions. No yield difference was noted between the 110 and 120 percent of NY points in the field. Ear leaf N composition followed in a very close order with the yield trend in low and high productivity soils of the field, ranging from 2.36 to 2.89 percent N. When uniform rate and the two variable rate N methods were evaluated, very few significant differences were noted among any of the crop measurements, including yield (Table 3). Only at the 110 percent of NY locations was there a difference between rate of N methods. There was a 6 bu/ac advantage for uniform N over VRN-Old. This was somewhat surprising because VRN-Old received slightly more fertilizer N than the uniform N rate treatment. No differences were observed in ear leaf N or stalk NO3-N at harvest as a function of the method of N rate application.
Strip yields over the entire field for the N application methods were observed as shown below: A two bu/ac average increase in yield was noted for VRN-New over VRN-Old. It is likely that most of the increase was a result of more N being applied in low-productivity regions with the VRN-New treatment. When all factors are evaluated, this slight increase in yield is likely non-significant when evaluated agronomically or economically. It should be noted that the N rate per acre was identical for VRN-Old and VRN-New, but the crop fertilized by both of these methods received 5 lb N/ac less than the uniform N rate method.
| Strips | Acres | N Applied (lb N/ac) |
Yield (bu/ac) |
|---|---|---|---|
| Uniform N | 11.14 | 158 | 169.6 |
| VRN-Old | 11.18 | 153 | 168.3 |
| VRN-New | 11.06 | 153 | 170.2 |
The response of corn grain yield to increasing N rates in low, average, and high normalized yield productivity zones is given in Figure 3, Figure 4, and Figure 5, where without and with nitrapyrin (Stay-N) inclusion was evaluated. At all three productivity levels, inclusion of the nitrification inhibitor resulted in a higher yield at the calculated optimum N rate of application. It should also be noted that the calculated optimum N rate for all levels of productivity was well below the 1.2 lb N/bu rate of N that would be recommended for corn grown in each of these zones. This suggests that either N losses were very low in 2001 or there was a large pool of mineralizable N that was released into the soil in 2001. Figure 6 shows an overall summary of the corn yields as affected by N rates and productivity levels.
Table 2. Soil test and crop response measurements at the 21 sampling points, Franklin Co., IL, 2001
Figure 2. Predicted N to apply at Franklin Co., IL, 2001
Figure 3. Effect of nitrogen rate on corn grain yield, Franklin Co., IL, 2001 (Yield Level = High)
Figure 5. Effect of nitrogen rate on corn grain yield, Franklin Co., IL, 2001 (Yield Level = Low)
Figure 6. Effect of nitrogen rate on corn grain yield, Franklin Co., IL, 2001 (Combined)
1 E.C. Varsa is Professor and T.D. Wyciskalla is Researcher, Dept. of Plant, Soil and General Agriculture, Southern Illinois University, Carbondale, IL; S.A. Ebelhar is Agronomist and C.D. Hart is Visiting Research Specialist, Dept. of Crop Sciences, University of Illinois, Dixon Springs Agricultural Center, Simpson, IL.
Carr, P.M., G.R. Carlson, J.S. Jacobsen, G.A. Nielsen, and E.O. Scogley. 1991.
Farming soils, not fields: A strategy for increasing fertilizer profitability.
J. Prod. Agric. 4:57- 61.
Redulla, C.A., J.L. Havlin, G.L. Kluitenberg, N. Zhang, and M.D. Shrock. 1996.
Variable nitrogen management for improving groundwater quality. pp. 1101-1110.
In P.C. Robert et al. (ed.), Proceedings of the Third International Conference
on Precision Agriculture. Minneapolis, MN. June 23-26, 1996. ASA, CSSA, and
SSSA, Madison, WI.
Sawyer, J.E. 1994. Concepts of variable rate technology with considerations for fertilizer application. J. Prod. Agric. 7: 195-206.
