| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
a Dep. of Agricultural Economics, The Univ. of Tennessee, 2621 Morgan Cir., Knoxville, TN 37996-4518
b Dep. of Biosystems Engineering and Soil Sci., West Tennessee Research and Education Center, The Univ. of Tennessee, 605 Airways Blvd., Jackson, TN 38301
* Corresponding author (rrobert3{at}utk.edu)
Received for publication June 1, 2006.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Notwithstanding its potential benefits, a crop production system using winter cover crops and no-tillage can affect lint yields and cotton profitability compared with one with no cover and conventional-tillage (Larson et al., 2001b). The buildup of plant materials and surface placement of fertilizer can influence soil properties such as pH. Nitrogen is an important fertilizer input in cotton production. No-tillage, in combination with surface-applied N, can result in the top few inches of the soil becoming more acidic due to nitrification (Ismail et al., 1994; Blevins et al., 1978, 1983). Low pH levels in the soil may affect the productivity of N fertilizers in no-tillage systems. Thus, the relationship between N fertilizer and soil acidity is particularly important with no-tillage. Howard et al. (2001) found that 67 kg ha–1 (60 lb ac–1) of N maximized lint yields on Loring and Lexington silt loams where no-tillage cotton was produced with different winter covers. They also found that 101 kg ha–1 (90 lb ac–1) of N was necessary to optimize lint yields on a Memphis silt loam.
Lime has long been viewed as a crop production input providing certain benefits, but those benefits come with a cost (Bongiovanni and Lowenberg-DeBoer, 1999). If crop yields are increased with lime application and the cost of lime and its application is less than the increase in total revenue from the additional yield, lime can be viewed as profitable. A profit-maximizing farmer will increase the rate of lime application so long as the value of the benefits exceeds the cost of lime and its application (Hall, 1983). This economic consideration suggests that a lime rate higher than the profit-maximizing rate would increase costs without a commensurate increase in revenue. Nonetheless, liming to achieve a specific pH level has long been practiced by farmers in the USA. Woodruff (1967, p. 222) states, "The actual soil pH requirements of crops... are not in close agreement with the soil pH recommendations that are made to farmers by the various advisory services." Most crops require lime when pH drops below 5.0 to reduce the level of harmful Al and Mn, increase uptake of beneficial nutrients such as Ca, Mg, and others, and counteract the acidifying effect of ammoniacal N fertilizers (Bongiovanni and Lowenberg-DeBoer, 1999; Sumner and Yamada, 2002). Since stratification of soil acidity can occur in the top 5 to 10 cm (2 to 4 in) the volume of soil to be affected by lime could be less than that for the typical depth of sampling in Tennessee, which is 15 cm (6 in) (Tyler et al., 2001). An experiment was conducted at the West Tennessee Research and Education Center, Jackson, TN, to evaluate if one-half the University of Tennessee Extension recommended rate of lime would be adequate since a thinner zone of soil required lime compared with the full recommended rate of lime for different cover crops, N rates, and tillage methods. Our null hypothesis was that an evaluation of the experimental data would show no significant difference between yields with the half and full rates of lime.
Concern exists over how to evaluate fertilizer and lime recommendations. The two most prevalent approaches are "feeding the crop" vs. "feeding the soil" (Murdock, 1992). Feeding the crop involves adding just enough fertilizer or lime to meet the needs of the crop. The approach of "feeding the soil" involves applying enough fertilizer or lime to maintain specific soil test levels and/or the correct nutrient balance in the soil (Murdock, 1992). In the past, lime application rates have been determined by the desire to achieve the magic pH level of 6.5 (Shelby, 2000). Applying half the recommended rate of lime addresses the idea of maintaining pH at a level to provide adequate crop response rather than maintaining a specific soil pH. There exists a need to reevaluate the target pH level to determine if the optimal pH level could be lowered without risking negative effects from soil acidity, thereby increasing a farmer's profitability. The objective of this research was to determine how lint yields and cotton profitability are affected by full and half rates of lime and alternative winter cover crops, N rates, tillage methods, and their interactions with lime rates.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
After letting pH deteriorate by delaying the regular application of lime from 1981 through 1994, the cover crop, N fertilizer, and tillage plots were further split into blocks that were randomly assigned a one-time application of two lime rates—100% of the recommended University of Tennessee Extension lime rate and one-half the recommended lime rate. The plot size only allowed one additional split; therefore, instead of including a zero rate treatment, only the half and full recommended rates were chosen for application to each plot where the recorded pH suggested lime was needed. After all the splits were completed the final plot size was 8 m wide by 12 m long. No lime was recommended for plots recording a pH of 6.1 or higher in 1995, so they received no lime. The pH means for treatment combinations ranged from 4.8 to above 6.5. Plots with a water pH 
6.0 were analyzed to determine their buffer values. Water pH and buffer values were then used to assign the full recommended rate of lime as suggested by Savoy and Joines (2001). The full rates of applied lime included 3.4, 4.5, 5.6, 6.7, and 7.8 Mg ha–1 (1.5, 2, 2.5, 3, and 3.5 ton acre–1). Rates applied for one-half the recommended rate were 1.7, 2.2, 2.8, 3.4, and 3.9 Mg ha–1 (0.75, 1, 1.25, 1.5, and 1.75 ton acre–1). The pH range in 1995 and the number of plots for each lime rate are presented in Table 1. Lime was applied on 10 and 11 May 1995 according to the Adams and Evans (1962) buffer test for 0–15 cm (0–6 in) soil depth. Individual plots received lime applications expected to increase pH to 6.5 when the full rate was applied. Lime applications were a one-time event for all plots included in this study. Data from 1995 through 2001 were used to evaluate the effects of full and half lime rates, winter cover alternatives, N fertilization rates, and tillage methods on yield and net revenue.
|
22,275 plants ha–1. A burndown herbicide was used to kill the cover crop before planting cotton in the no-tillage plots. Conventional-tillage plots were disked twice and leveled to destroy the cover crop before planting. Stratification was present in both tillage plots but more prevalent in no-tillage plots. Winter covers were reestablished each season after cotton harvest with seeding rates of 101 kg ha–1 (90 lb acre–1) for winter wheat, 22 kg ha–1 (20 lb acre–1) for hairy vetch, and 17 kg ha–1 (15 lb acre–1) for crimson clover. Nitrogen, in the form of ammonium nitrate (340 g N kg–1), was hand broadcast at planting. Rates of N fertilizer applied to the plots were 0, 34, 67, and 101 kg ha–1 (0, 30, 60, and 90 lb acre–1).
Lint Yield Response Function
A quadratic yield response function was estimated using the 1995–2001 data for each winter cover alternative:
 | [1] |
Quadratic yield response to N was assumed to account for expected diminishing marginal productivity of N as N fertilization increased from 0 to 101 kg ha–1 (90 lb acre–1). The quadratic functional form has been widely used to estimate N fertilizer response (Woodward, 1977). The expected signs of ß1 and ß2 were positive and negative, respectively.
Documentation regarding the relationship between no-tillage and lint yields was mixed. Several studies reported similar or higher lint yields with no-tillage compared with conventional-tillage (Bloodworth and Johnson, 1995; Stevens et al., 1992; Triplet et al., 1996; Hutchinson et al., 1995). Other research documented higher conventional-tillage lint yields compared with no-tillage (Bauer and Busscher, 1996; Burmester et al., 1993). The hypothesized sign of ß3 (the TILL coefficient) was difficult to determine a priori because of these conflicting results. Nevertheless, if ß3 were significantly different from zero, the no-tillage yield response function would have a different intercept than the one for conventional-tillage.
Both the half and full rates of lime were hypothesized to positively influence yields by increasing the soil pH to a level where exchangeable Al and Mn were unavailable. Nonetheless, ß4 (the HRD coefficient) was not expected to be statistically different from zero because it measures the difference between yields resulting from application of the half and full rates of lime, and the two lime rates were hypothesized not to affect lint yield differently. If ß4 were not different from zero, the intercepts for the yield response functions would be the same for the half and full rates of lime. The ß5 coefficient (ZLIME) represented the change in yield when no lime was added to plots with a pH of 6.1 or higher. These zero-lime plots already had sufficiently high pH levels at the beginning of the study and were expected to have higher yields than those with lower pH levels, with other factors constant; thus, ß5 was expected to be positive. Many of the zero-lime plots received either no applied N or 34 kg ha–1 (30 lb ac–1) of N, which contributed to maintaining a higher soil pH.
The time trend (TIME) represents the expected effects over time of lime application on yield. The full yield benefit of lime application in 1995 was not expected in that year but was expected to manifest itself over several years. Accordingly, ß6 was expected to be positive.
If lime application increased the marginal physical product of N fertilizer, the half and full rates of lime would increase yield response to N fertilization. The coefficient for N x HRD (ß7) measures the difference in yield response to N fertilizer when the half and full lime rates are applied. Therefore, ß7 was not expected to be statistically different from zero given the null hypothesis that the half and full rates of lime do not affect yields differently. Similarly, ß8 (the TILL x HRD coefficient) was not expected to be statistically significant.
Without additional lime applications, the zero-lime plots may have become increasingly acidic between 1995 and 2001. On the other hand, these plots typically received zero or low N rates, possibly having little effect on pH. Thus, the marginal physical product of N was expected to decline slightly or remain unchanged, suggesting that the coefficient for N x ZLIME (ß9) would be negative or zero. The sign of the coefficient for TILL x ZLIME (ß10) was difficult to hypothesize a priori since the sign of the coefficient for TILL was indeterminate and plots receiving zero lime had higher pH levels at the beginning of the experiment than those that received half or full rates of lime.
With no-tillage, plant material was left on the surface to decay more slowly and release less N for crop use than with conventional-tillage, where it was incorporated into the soil. Mengel et al. (1992) found available N for crop use was reduced in a no-tillage and cover crop system due to C sequestration and N immobilization. This relationship implies a negative sign for ß11 (the N x TILL coefficient) resulting from a lower marginal physical product of N fertilizer under no-tillage than under conventional-tillage.
The TIME x HRD interaction was not expected to be statistically significant given the hypothesis that the half and full lime rates do not affect cotton lint yields differently; thus, ß12 was expected to be zero. The TIME x ZLIME interaction tests the hypothesis that the time trend is different for the zero-lime plots than the plots that receive half or full lime rates. The sign of the coefficient for TIME x ZLIME (ß13) was hypothesized to be zero or negative because, with no lime, the zero-lime plots were not expected to have an increase in yield similar to the plots that received lime, and yields might actually have decreased slightly if soils in those plots became more acidic over time.
The coefficient for TIME x TILL (ß14) was expected to be positive since an increase in soil productivity due to no-tillage relative to conventional-tillage was expected over time. The relationship between the TIME x N interaction and lint yield was indeterminate.
Profit Maximization
Estimated yield response functions were used to calculate profit-maximizing N fertilizer rates, yields, costs, and net revenues above variable and fixed production costs. Profit-maximizing N fertilization rates were calculated by setting the first derivative of the yield response function with respect to N equal to the price of N divided by the lint price and solving for the quantity of N, which is the first-order condition for profit maximization (Boehlje and Eidman, 1984). Net revenue was calculated as:
 | [2] |
Prices and costs used to calculate profit-maximizing values were expressed in 2004 dollars, so changes in net revenues would reflect changes in profit-maximizing yields rather than inflationary price changes. A lint price of $1.12 kg–1 ($0.51 lb–1), an N fertilizer price of $0.75 kg–1 ($0.34 lb–1) of N, and a lime price of $24.14 Mg–1 ($21.91 ton–1) were used to calculate net revenues. Average prices for 2002 through 2004 were used in these calculations (Tennessee Department of Agriculture, 2002, 2003, 2004). These prices were inflated to 2004 dollars by the Implicit Gross Domestic Product Price Deflator before averaging (Congress of the United States, Council of Economic Advisors, 2005). Average prices were used as proxies for expected future prices. An average of the TIME variable, 3.5 yr, was used to calculate the profit-maximizing lint yields.
Winter cover establishment costs were zero for no cover, 91.39 ha–1 ($37.00 acre–1) for winter wheat, $79.04 ha–1 ($32.00 acre–1) for hairy vetch, and $56.81 ha–1 ($23.00 acre–1) for crimson clover. Cover crop costs consisted of costs for seed, machinery, labor, and interest on variable costs of establishment. Cover seed costs were 101 kg ha–1 (90 lb acre–1) multiplied by a price of $0.14 kg–1 ($0.30 lb–1) for winter wheat, 22 kg ha–1 (20 lb acre–1) multiplied by a price of $0.50 kg–1 ($1.10 lb–1) for hairy vetch, and 17 kg ha–1 (15 lb acre–1) multiplied by a price of $0.41 kg–1 ($0.90 lb–1) for crimson clover. Seed prices for each cover crop were obtained from a 2004 Tennessee Farmers Cooperative (2005) suggested retail price list. Machinery and labor costs for seed establishment assumed a 150-hp tractor and a 6.38-m (21-ft) drill with 178-mm (7-in) row spacing requiring 0.27 h ha–1 (0.11 h acre–1) plus labor at $8.50 h–1 for 0.35 h ha–1 (0.14 h acre–1) (Gerloff, 2004). Consistency between cover establishment costs and other costs in the 2004 Tennessee cotton budgets (Gerloff, 2004) was maintained by charging a nominal interest rate of 8% on cover establishment expenses. Using the nominal interest rate for expenses that occur within the period (2004 in this instance) was recommended by the AAEA Task Force on Commodity Costs and Returns (2000).
Amortized annual lime costs used in the net revenue analysis were $22.55 ha–1 ($9.13 acre–1) and $11.26 ha–1 ($4.56 acre–1) for the full and half rates of lime, respectively. The cost of lime for the half and full rates were amortized over n = 7 yr using the capital recovery method (Boehlje and Eidman, 1984):
 | [3] |
Production practices assumed in the cost and return budgets (Eq. [2]) mirror the no-tillage and tillage practices used in the experiment. The no-tillage budget assumed application of a burndown herbicide to kill the winter cover and weeds before planting and total variable, machinery, interest, and labor expenses and other costs of $902.09 ha–1 ($365.22 acre–1). The conventional-tillage budget assumed two disking operations before planting to kill the winter cover and weeds and to prepare the seedbed, and had total variable, machinery, interest, and labor expenses for total tillage and other costs of $960.48 ha–1 ($388.86 acre–1). The cost difference between the conventional- and no-tillage budgets of $58.39 ha–1 ($23.64 ac–1) was mostly from additional machinery and labor expenses for conventional-tillage relative to no-tillage. Other costs of production that remained constant were taken from the University of Tennessee Extension enterprise budgets for conventional-tillage and no-tillage Roundup Ready cotton (Gerloff, 2004).
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
Diagnostics indicated that multicollinearity may have degraded the standard errors of the coefficients of ZLIME, N x ZLIME, TIME, and TIME x ZLIME in the no cover (condition index = 58) and winter wheat (condition index = 62) covers, both regressions having large variance proportions for the above variables. Thus, hypothesis tests that fail to reject the null hypothesis of no effect may have been compromised by multicollinearity.
An evaluation of F statistics (no cover, F4/432 = 0.12; winter wheat, F4/432 = 0.17; hairy vetch, F4/434 = 0.53; crimson clover, F4/433 = 0.29) for the half lime rate dummy variable and its interactions with N fertilizer, tillage method, and time (the linear combination of HRD, N x HRD, TILL x HRD, and TIME x HRD) was not significantly different from zero at the 0.05 significance level for any of the four cover alternatives. This finding reinforces the hypothesis that within the 7-yr study period, applying the half rate of lime did not affect cotton lint yields differently than applying the full Extension recommended rate of lime. Furthermore, this finding was consistent regardless of cover alternative, tillage method, N fertilization rate, or the number of years after lime was applied (up to 7 yr in this study).
Another important finding is that cotton lint yield response to N fertilizer was not affected by the rate of lime applied, the tillage method, or the number of years following lime application. The interactions between N fertilization and the lime rate, tillage method, and time were not significantly different from zero individually as indicated by t statistics in Table 3, nor were they significantly different from zero collectively for any cover alternative as indicated by F statistics (no cover, F4/432 = 0.15; winter wheat, F4/432 = 1.21; hairy vetch, F3/434 = 0.65; crimson clover, F3/433 = 0.06) for the linear combination of N x HRD, N x ZLIME, N x TILL, and TIME x N.
A significant finding is that tillage method made a difference in affecting cotton lint yields for all four cover alternatives. The TIME x TILL interaction was statistically significant for each cover alternative, suggesting that no-tillage significantly increased lint yields over time compared with conventional-tillage. In addition, the F statistics (no cover, F5/432 = 5.02; winter wheat, F5/432 = 6.57; hairy vetch, F4/434 = 6.53; crimson clover, F5/433 = 7.31) for the linear combination of tillage method and its interactions with the lime rate, the N rate, and the time trend (TILL, TILL x HRD, TILL x ZLIME, N x TILL, and TIME x TILL) was significantly different from zero at the 0.01 significance level for no cover and the 0.001 significance level for the winter wheat, hairy vetch, and crimson clover cover alternatives.
The F statistic for the zero-lime rate dummy variable and its interactions with N fertilizer, tillage method, and time (the linear combination of ZLIME, N x ZLIME, TILL x ZLIME and TIME x ZLIME) was not significantly different from zero at the 0.05 significance level for any cover alternatives.
Profit Maximization
Profit-maximizing N rates, lint yields, costs, and net revenues for the full and half lime rates and tillage methods are presented in Table 4 for each cover alternative. Without exception, no-tillage produced higher net revenues than conventional-tillage for all four cover alternatives. Generally, higher net revenues for no-tillage typically resulted from higher lint yields and lower N fertilizer costs compared with conventional-tillage.
|
Profit-maximizing N rate differences between the half and full rates of lime were minimal for all winter cover crops. Profit-maximizing N rates were highest for no cover and winter wheat and lowest for the legume cover alternatives.
Using the no cover alternative provided the highest net revenue regardless of tillage method or lime rate, while the legume cover alternatives produced the lowest net revenues. Combining conventional-tillage with legume covers resulted in negative net revenues when using either the full or half rates of lime. The legume N from these cover crops did not produce sufficient yield increase to justify the extra cost of cover establishment.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Among the cover alternatives, no cover combined with no-tillage produced the highest net revenues of $135 ha–1 ($55 acre–1) and $168 ha–1 ($68 acre–1) and lint yields of 1000 kg ha–1 (893 lb acre–1) and 1021 kg ha–1 (911 lb acre–1) in the short-to-medium term (7 yr) when using the full and half rates of lime, respectively. Using a winter cover of crimson clover required the least amount of N fertilizer, but resulted in the lowest net revenues and lint yields among the tillage methods and lime rates.
Net revenues achieved with one-half the University of Tennessee Extension recommended rate of lime were either comparable to or greater than the full rate of lime for both tillage methods, suggesting that the increased cost of applying the full rate of lime was not justified by sufficient yield increases. Acidification did not appear to be an issue in the no-tillage plots since no-tillage yields exceeded those received from conventional-tillage plots. On the basis of the findings of this study, which included 7 yr of data, cotton farmers may receive similar or higher net revenue by applying one-half the University of Tennessee Extension recommended rate of lime instead of the full recommended rate. Some cotton farmers, especially those facing uncertain lease agreements, may want to consider applying less than the recommended rate of lime. Note, however, this study did not analyze the cash-flow implications of lime application under a short-term lease. Additional research including more than 7 yr of data would be needed to determine if farmers would benefit from the full rate of lime for a longer period of time than the half rate. While the half-rate of lime may be more profitable for conventional-tillage and no-tillage methods, it is important to consider that this analysis did not test whether a crop response to lime exists, nor did it identify an optimal rate or timing of lime application. Future studies should be designed to include additional multiples of the full recommended lime rate (e.g., zero, one-fourth, one-half, three-fourths, and full) on a range of initial soil pH test levels. This would allow estimation of a soil pH yield response. Also, annual pH tests by plot would allow estimation of a carryover equation. With pH response functions and carryover equations, net present value maximizing lime strategies could be calculated for farmers with various rental arrangements and soil stewardship goals.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
