Illinois Fertilizer
Conference Proceedings
January 25-27, 1993
| Home | 1993 Index | Search |
|
|
| Home | 1993 Index | Search |
H.M. Brown, R.G. Hoeft, and E.D. Nafziger1
Proper N application rates are required for continued economic and environmental
viability of U.S. agriculture. Application at rates below that required for
economic optimum will place U.S. farmers at a competitive disadvantage in the
world market. On the other hand, excessive rates of N fertilizers may result
in contamination of ground water..
Experiments were conducted on 77 farms throughout Illinois from 1990 through 1992 to evaluate the efficacy of three N recommendation systems for corn. The first of the three recommendation systems consisted of the current University of Illinois system, in which proven yield (based on a 5-year average) for a field is multiplied by 1.2 lbs N/bushel for continuous corn, with adjustments (downward) if corn follows a legume or if manure has been applied. For corn following soybeans this adjustment is 40 lb N per acre, and for corn following a full stand of alfalfa, the rate is reduced by 100 lb N per acre. The amount of available N contained in manure is subtracted from the recommended rate. This system will be designated the proven yield (PY) system. The second system is modeled after the Michigan State University recommendation in which the amount of N found in the top two feet of soil in early spring is subtracted from the normal recommendation. This system will be designated as PPNT. The third system evaluated is the current Iowa State University recommendation (Blackmer et al., 19911) in which adjustments are made based on soil nitrate levels at presidedress time, when corn is 6 to 12 inches tall. This system will be referred to as PSNT.
Experimental sites were selected to provide a measure of the effect of these recommendation systems under varying conditions, including differences in soil type, previous crop, climatic conditions, and manure management. At each location, soil samples were collected in early spring (late March to early April) in 1 foot increments to a depth of 2 feet and again at N sidedress time (late May to early June) as a single sample to a depth of 1 foot. The samples were kept frozen until analysis for NO3 N, which was done according to Keeney and Nelson, 19822. When corn was 6 to 12 inches tall, N was applied using 6 rates of N evenly spaced from 0 to 100% of the normal recommendation in 1990 and from 0 to 125% of the normal recommendation in 1991 and 1992. The normal recommendation was determined by multiplying the soil's corn yield potential in bushels per acre (as given in Fehrenbacher et al., 19843) times 1.2 pounds of N, minus corrections based on previous crop and/or manure application history. Nitrogen treatments were injected as urea-ammonium nitrate solutions (28% N). Plant populations were thinned to a uniform stand at each location. Yield was determined by hand harvesting 30 feet of each of two rows. In November, soil samples were collected in 1 foot increments to a depth of 3 feet on the control, highest rate, and a mid-rate treatment for determination of NO3-N concentration.
A separate experiment was conducted in 1992 to determine the spatial distribution of nitrate and ammonium concentrations in the soil following anhydrous ammonia injection. Anhydrous ammonia was injected at a depth of 8 inches in late April at rates of 100, 150, and 200 lbs N/acre in a Drummer silty clay loam at Urbana. No crop was grown during the season. Soil samples were collected to a 1 foot depth about every 2 weeks during May and June at distances of 0, 3, 6, 9, 12, and 15 inches from the injection band. Additional samples were collected in October.
All experiments utilized a randomized complete block design with 4 replications
and were analyzed using the appropriate SAS procedures. When the response to
N rate was best described as a quadratic in the form of N = a + bN + cN2, the
economically optimum N rate was determined as N = -(b-R)/2c, where R is the
cost of N in dollars per pound divided by the price of corn in dollars per bushel.
We used R = $.20 / $2.50 = 0.08.
Nitrate concentrations ranged from 1 to 45 ppm NO3-N in the top two feet of soil in early spring, with an average value of 7 ppm. The highest concentration was observed on a field that had received the equivalent of 560 pounds available N per acre as manure in the preceding fall. At presidedress, the average NO3-N value in the top foot of soil was 11 ppm with a range from 0 to 55 ppm. Again, the highest value was associated with a field that had received a high rate of manure application the previous fall. The average NO3-N concentration in the surface foot of soil was higher at sidedress time than in early spring, 11 versus 4 ppm respectively, for sidedress and early spring samples. This increase in concentration from early spring to sidedress time likely resulted from more mineralization than N loss during this period.
Twenty six locations had PSNT NO3-N concentrations above 10 ppm, with 8 of those exceeding 20 ppm. Seven of the eight locations that exceeded 20 ppm had manure applied the previous fall, and there was no significant response to fertilizer N application at any of these seven locations. At the eighth location where NO3-N concentration exceeded 20 ppm, no manure had been applied and there was a significant response to applied N. Response to applied N was observed at only 5 of the 18 locations where the NO3-N concentration was between 10 and 20 ppm.
Significant response to applied N was observed at 44 of the 77 locations evaluated (Table 1). For the responsive locations, the optimum yield obtained was over 20 bushels per acre greater than the yield goal used for calculation of the N requirement. This large difference appeared as a result of the high yield levels of 1992. The average optimum N rate determined in these experiments was within 1 pound per acre of that recommended using the PY method. For continuous corn, the optimum N rate was 1.09 lbs N/bushel of corn produced. For corn after soybeans, the optimum was 1.03 lbs N/bushel, which is again very close to the rate recommended under the PY system when one assumes a 40 lb N/acre credit for the preceding soybean crop. The PPNT and PSNT systems underestimated N needs by approximately 30 and 8 pounds per acre respectively. Use of the PPNT system would have substantially reduced net income, while use of either of the other two systems would have resulted in net return to N very near the maximum.
All three systems overestimated the amount of N needed for optimum yield at the 33 nonresponding locations (Table 2). Closer examination provides some reasons for these overestimates. Nine of the locations that did not respond to fertilizer N had received manure application the previous fall. All three systems predicted that five of these locations would not need additional N. However, all three predicted a need for N at two locations that had received surface applied straw manure. In these two cases, we had neither an accurate estimate of the amount of manure applied nor of the N content of the manure. As a result, a small amount of N was recommended. Based on these limited results, it would appear that the PSNT system might have potential for use in those cases where manure is applied on the surface but not incorporated into the soil quickly enough to prevent volatilization losses.
Alfalfa was the previous crop at four of the non-responding locations. One of these was also drought affected, thus the need for supplemental N was markedly reduced. The three nondrought affected locations produced an average of 187 bushels per acre with no supplemental N, indicating that mineralization of the legume residue was greater than normally expected for those locations. All three systems predicted a response to applied N at these four locations.
Yields were markedly reduced at eight of the locations by drought. Yields averaged only about two thirds of normal expectations for those locations with some being as low as one third of normal. With dry weather as the dominant influence on yield, a lack of response to N rates was understood.
There was no obvious reason why the remaining twelve locations did not respond to applied N (Table 3). Yields at these locations were near expectations, and substantially exceeded expectations in some cases. Six of the twelve followed corn, raising the possibility of N carryover from the prior year. However, neither of the soil samples revealed significant NO3-N levels. Several of the locations had very high P and K soil test values, leading us to suspect past manure applications. If so, these soils may have had higher mineralization rates than if they had not received manure. Since the cooperators reported all manure applications within the three years prior to the study, any manure applied to the test site had to have been applied at least four years prior.
Residual nitrate levels in the surface foot of soil following harvest in the fall increased slightly with increasing N rates at the responding sites (Table 4). At the non-responding sites, nitrate levels at all application rates and at all depths were higher than at the responding sites, indicating some movement of excess N through the profile. These figures were inflated somewhat by the inclusion of the sites that had received manure. Surprisingly, the unexplained non-responding sites were not substantially higher in NO3-N concentration than were the responding sites. In all cases, residual NO3-N concentrations were well below those normally thought to be needed for optimum corn production.
Irrespective of rate of application, there was little indication of NO3-N
movement beyond 6 inches from the point of ammonia application during the first
six to eight weeks after ammonia application (Fig.
1). At all application rates, the concentration was highest in the sample
taken directly over the band, and decreased by one-half to two-thirds in samples
collected only 3 inches from the band. This rapid change in concentration as
one moves away from the band will make it very difficult to attain accurate
results unless one systematically probes every few inches perpendicular to the
direction of application. As an alternative, one might utilize a field test
kit to identify the location of the band and then proceed to sample the field
according to the known location of the band.
Based on these results, we think that both the PY and PSNT systems will provide
an accurate estimate of N needed when a response to N occurs. Use of the PPNT
system showed limited promise on fields that had experienced a yield limiting
environment the prior year that may have resulted in carryover of N. The PSNT
system showed promise for assisting those producers who surface apply manure
without immediate incorporation, or who have limited information available on
the rate or concentration of N being applied in the manure. Even under conditions
of low yields due to unfavorable weather, where one might expect considerable
N to remain unused, we found little NO3-N in the upper three feet of soil following
harvest. Such evidence does not preclude the possible accumulation of organic
forms of N that might be released into the profile at a later date and that
might result in groundwater contamination. Limited movement of nitrates from
the point of ammonia injection in the first few weeks after application will
require a carefully planned sampling program in order to utilize the PPNT system
on field that have received a spring preplant application of N.
Table 4: Effect of N application on residual NO3-N concentration with depth at harvest
Blackmer, A. M., T. F. Morris, D. R. Keeney, R. D. Voss, and R. Killorn. 1991. Estimating nitrogen needs for corn by soil testing. Iowa State University Extension Bulletin Pm-1381. Ames, IA.
Keeney, D.R. and D.W. Nelson. 1982. Nitrogen - Inorganic forms. In A.L. Page (ed.) Methods of soil analysis, Part 2. 2nd Edition, Agronomy 9:711-733.
Fehrenbacher, J.B., J.D. Alexander, I.J. Jansen, R.G. Darmody, R.A. Pope, and M.A. Flock. 1984. Soils of Illinois. Bulletin 778, University of Illinois, Champaign-Urbana, IL.
