|
|
||||||||
a USDA-ARS, Coastal Plain Research Center, 2611 W Lucas St., Florence, SC 29501-1242 USA
b Clemson Univ., 2200 Pocket Rd., Pee Dee Res. Center, Florence, SC 29506-9706 USA
busscher{at}florence.ars.usda.gov
| ABSTRACT |
|---|
|
|
|---|
Abbreviations: N, no deep tillage S, spring deep tillage (before soybean planting) W, fall deep tillage (before wheat planting) B both spring and fall deep tillage
| INTRODUCTION |
|---|
|
|
|---|
A common deterrent to plant growth in many southeastern Coastal Plain soils is high soil strength. High strength, as measured by cone index, is found throughout the profile but especially in the E horizon, just below the Ap. Strength in the E horizon can restrict root growth even when water content is at field capacity (Campbell et al., 1974) and strength increases as the soil dries. Cone indices up to 20 MPa have been measured in the E horizon (Karlen et al., 1991); this is 10 times the root limiting cone indices reported by Blanchar et al. (1978) and Taylor and Gardner (1963) for soils of similar textures. Though a few roots may find their way through the hard layer (Vepraskas et al., 1995), cone indices of this magnitude generally prevent roots from taking water and nutrients from the E and lower horizons. As a result, yield is reduced (Sojka et al., 1991).
Current practice in the southeastern Coastal Plain is to reduce soil strength by deep tillage. Because the soil reconsolidates between growing seasons (Threadgill, 1982; Busscher et al., 1986), deep tillage is required annually. Deep tillage is recommended either in spring (Threadgill, 1982; Busscher et al., 1986) or fall (Porter and Khalilian, 1995). Though it is not usually recommended twice a year, some producers deep till twice if double cropping. A recommended time and frequency for deep tillage has yet to be established for the southeastern Coastal Plains, and potentially for other similar coastal areas. The timing and effectiveness of deep tillage are important because it is expensive; it requires large tractors (1420 kw per deep tillage shank), 20 to 25 L of fuel per hectare, and 20 to 40 min labor per hectare (Karlen et al., 1991). Timing, frequency, and effectiveness of deep tillage need to be developed.
We hypothesized that for the narrow-row management system, frequency and timing of deep tillage would affect double-crop productivity and soil strength. Our purpose was to determine whether deep tillage in spring, fall, or both gave the greatest reduction in soil strength and the greatest improvement in soybean and wheat yield. Since we planned to measure both soil strength (cone index) and yield, another purpose was to correlate the effects of soil strength reduced by tillage with yield for the narrow-row, deep-tilled, double-cropped management system.
| Materials and methods |
|---|
|
|
|---|
The day before planting either wheat or soybean, we imposed surface tillage and deep tillage treatments onto the plots. The two surface tillage treatments involved not disking (planting into the stubble of the previous season's crop) or disking twice before planting. The four deep tillage treatments included no paratilling (N), paratilling before soybean planting (S), paratilling before wheat planting (W), and paratilling before planting both soybean and wheat (B). Treatments were replicated four times in a randomized complete block design.
Surface tillage, deep tillage, and planting were done in separate operations. All tillage and harvesting equipment followed the same wheel tracks as closely as possible. Surface tillage was done with a 3-m-wide Tufline1 disk (Tufline Mfg. Co., Columbus, GA) pulled by a John Deere 4230 (Deere and Co., Moline, IL) 75-kw tractor with wheels on 1.6-m centers. Deep tillage was done with a two-shank paratill in fall of 1993 and after that with a four-shank paratill (Tye Co., Lockney, TX). Shanks were set 0.66 m apart. The paratill was pulled with a Case 2670 (now Case-IH, Racine, WI) 165-kw, four-wheel-drive tractor with dual wheels on 1.9-m and 3.1-m centers. Shanks deep-tilled soil to approximately 0.4 m (the bottom of the hardpan).
Plots were planted to soft red winter wheat cultivar Northrup King Coker 9134 and `Hagood' soybean, a Maturity Group VII cultivar. Both wheat and soybean were drilled in 19-cm-spaced rows with a 3-m-wide John Deere 750 No-till Planter pulled by a Massey Ferguson 398 (Massey Ferguson, Inc., Des Moines, IA) 60-kw tractor with wheels on 1.9-m centers. Wheat was drilled on 18 Nov. 1993, 23 Nov. 1994, and 21 Nov. 1995 at a rate of 66 seeds m-1 and harvested on 27 May 1994, 30 May 1995, and 24 May 1996. Soybean were drilled on 30 May 1994, 1 June 1995, and 7 June 1996 at a rate of 13 seeds m-1 and harvested on 3 Nov. 1994, 3 Nov. 1995, and 8 Nov. 1996. Whole plant samples for yield of wheat and soybean were harvested from six 1-m sections of row in each plot. Yield data for both were corrected to 130 g kg-1 moisture.
When in wheat, grain for the whole plot was harvested with an Allis Chalmers (now Deutz-Allis, Norcross, GA) F3 Gleaner with a 4-m-wide header and wheels on 2.4-m centers. When in soybean, grain from the whole plot was harvested with an IH (now Case-IH, Racine, WI) 1420 axial flow combine with a 4.0-m wide header and wheels on 2.3-m centers.
Following Clemson soil test recommendations (Clemson University, 1982), P and K were preplant broadcast on all plots at rates of 90 kg ha-1 each before disking for wheat. Ammonium nitrate was broadcast on all plots at 34 kg N ha-1 immediately after planting wheat and at 56 kg N ha-1 as a side-dressing in late February or early March (the stem erect wheat growth stage). Fertilizer was applied with a 3-m-wide Gandy spreader (Gandy Co., Owatonna, MN) pulled by a Massey Ferguson 253 tractor with wheels on 1.9-m centers.
To control weeds, non-disked plots were sprayed with glyphosate [N-(phosphonomethyl)glycine] at a rate of 1.1 kg a.i. ha-1 before wheat planting or Bronco (Monsanto, St. Louis) [glyphosate plus (alachlor) 2-chloro-2',6'-diethyl-N-(methoxymethyl)acetanilide] at a rate of 3.9 kg a.i. ha-1 before soybean planting. Disked plots were sprayed with alachlor preemergence at a rate of 2.6 kg a.i. ha-1 before soybean emergence. To control annual broadleaf weeds and nutsedge (Cyperus esculentus L.), all plots were sprayed with chlorimuron ethyl (2-[[[(4-chloro-6-methoxyprimidin-2-yl)-carbonyl]-amino]sulfonyl]benzoate) at 0.013 kg a.i. ha-1 at 21 d after soybean planting. To control annual grasses, all plots were sprayed with sethoxydims (2-[1-(ethoxyimino)butyl]-5-[2-(ethylthio)propyl]-3-hydroxy-2-cyclohexen-1-one) at 0.2 kg a.i. ha-1 at 30 d after soybean planting.
Cone index data were taken with a 12.5-mm-diameter cone-tipped penetrometer (Carter, 1967) on 21 June 1994, 16 June 1995, and 13 June 1996 in soybean and on 20 Dec. 1994 and 12 Dec. 1995 in wheat. Cone indices were measured by pushing the penetrometer into the soil to a depth of 55 cm at nine positions spaced 9.5 cm apart starting at the middle of the plot and moving outward. Cone index data were digitized into the computer at 5-cm-depth intervals and log transformed before analysis according to the recommendation of Cassel and Nelson (1979). Data for all positions across the plot and depth were combined to produce cross-sectional contours of soil cone indices by the method of Busscher et al. (1986).
In 1993, at the first wheat planting, we did not take cone index data. At that time, we had performed tillage in N and W plots but we could not perform tillage in S and B plots until the following spring. The first set of cone index data were taken in 1994 after the first soybean planting and after all treatments had been imposed.
Gravimetric soil water content samples were taken along with cone indices. They were taken at the first and fifth positions of cone index readings. Water contents were measured at 10-cm-depth intervals to the 60-cm depth. These water contents were taken as representative of the water contents of the plot.
We analyzed cone index and water content data using the ANOVA and the least square mean separation procedures (SAS Institute, 1990). Data were analyzed by a split-split plot randomized complete block design where the first split was position across the row and the second depth. We analyzed yield as a function of cone index using the linear regression procedure, REG (SAS Institute, 1990). Data were tested for significance at the 5% level.
| Results and discussion |
|---|
|
|
|---|
|
Although water contents did not vary with treatment or treatment interactions with depth and date, they varied with depth and date of measurement (Table 1). Water content generally increased with depth. Water content differences among dates of measurement depended on how thoroughly each rain wetted the profile. Because of its lack of variation with treatment, water content was ignored for the analysis of cone indices, except when considering depth and date.
Cone Index
Deep Tillage Treatments
For the deep tillage treatments, the cone indices were generally lowest in the more recently or more frequently deep-tilled plots (Table 2)
. Over the course of the experiment, Treatment B, paratilled at the beginning of both seasons, developed the lowest mean cone indices, when averaged over the whole profile that was measured (Fig. 1 and 2)
. On a season-by-season basis, cone indices for Treatment B were as low as those for the spring-deep-tilled treatment, S, in the spring and as low as (or lower than) those for the fall-deep-tilled treatment, W, in the fall.
|
|
|
In addition to analyzing data for spring tillage vs. fall tillage, we analyzed the data to determine the effects of more recent versus less recent tillage on cone indices. Mean profile cone indices were lower for the more recently tilled treatments. Averaged over all seasons, the treatment with no deep tillage resulted in the highest mean cone index, 1.74 MPa (0.264)2
. The treatment with last season's deep tillage resulted in the next highest mean cone index, 1.30 MPa (0.145). The treatment with the current season's deep tillage resulted in the third highest mean cone index, 0.99 MPa (0.039). The treatment with deep tillage both seasons resulted in about the same mean cone index, 0.94 MPa (0.015,
), as the current season's deep tillage treatment. From this point of view, more recent tillage provides the soil with lower cone indices. Deep tilling in both seasons has no advantage in reducing cone indices over deep tilling in the current (or more recent) season; however, if we continually till in the current season, we will be tilling in both seasons anyway. Also, fall tillage before spring planting a single season crop would not be recommended unless earlier planting, as a result of the earlier tillage, could overcome the potential loss of yield caused by the harder soil.
Surface Tillage Treatments
For the average over the course of the experiment and for two of the five season-by-season measurements, disking compacted the soil more than it loosened it (Table 3)
. Averaged over the course of the experiment, disked treatments developed 60-kPa higher mean profile soil cone indices than non-disked treatments. On a season-by-season basis, disked treatments in spring 1994 and spring 1996 developed higher mean profile cone indices than non-disked treatments. Disking never reduced mean profile cone index (Table 3). Though disking loosened the top 5 to 15 cm of the profile (Fig. 1), it compacted soil below the disked zone to produce mean profile cone indices that were equivalent to or higher than non-disked treatments.
|
When averaged over disked and non-disked treatments, cone indices for deep tilled treatments increased for each 5-cm-depth interval from the soil surface to the 55-cm depth. The cone index averages by 5-cm-depth increment were 0.47 (-0.243), 0.73 (-0.079), 0.93 (0.011), 1.11 (0.084), 1.28 (0.141), 1.40 (0.177), 1.56 (0.221), 1.86 (0.293), 2.20 (0.362), 2.44 (0.406), and 2.56 MPa (0.425, LSD = 0.013 at
). Since water content also increased with depth, cone index increases were due to increased soil strength, not to water content changes (If water contents had been equal, changes would have been greater.).
Position across the Rows
As also seen by Reeves et al. (1990), Wiermann et al. (1999), and others, cone indices varied significantly with position across the rows. Cone indices below the wheel track (position = 76 cm, Fig. 1 and 2) averaged over depths were 0.14 MPa higher (0.854 difference of the logs, LSD = 0.018 at
) than below the non-wheel-track (position = 0 cm). Even for the no-till treatments, cone indices below the wheel tracks [1.88 MPa (0.296)] were higher than below the non-wheel-tracks [1.62 MPa (0.235), LSD = 0.028 at P = 0.05)].
Cone index differences due to position were also a result of the deep tillage vs. non-deep tillage treatments (Fig. 1 and 2). The interaction between deep tillage treatment and position was significant and was a result of the zone in the middle of the measured area (position = 25.3 to 50.6 cm of Fig. 1 and 2) being either disrupted by deep tillage or not disrupted. Disruption resulted in a mean cone index of 0.80 MPa (-0.043) for the deep-tilled treatments vs 1.73 MPa (0.262) for the non-deep-tilled treatments (LSD = 0.036 at P = 0.05). For non-deep-tilled treatments, cone indices in the middle of the measured area are as high in strength as those in the wheel tracks 1.75 MPa (0.267).
Yield
Yield was summarized in Frederick et al. (1998). Briefly, averaged over years, wheat yielded more in the non-disked treatments (3.05 Mg ha-1) than in the disked treatments (2.84 Mg ha-1, LSD = 0.16 at P = 0.05). Soybean also yielded more in the non-disked treatments (4.60 Mg ha-1) than in the disked treatments (3.78 Mg ha-1, LSD = 0.22 at P = 0.05). With respect to timing of tillage, wheat yielded more for treatments that were tilled more recently; the order was B (3.29 Mg ha-1)
W (3.26 Mg ha-1) > S (2.95 Mg ha-1) > N (2.30 Mg ha-1, LSD = 0.22 at P = 0.05). Soybean yielded more for treatments that were tilled more recently or more frequently in the order B (4.98 Mg ha-1) > S (4.32 Mg ha-1)
W (4.07 Mg ha-1) > N (3.39 Mg ha-1, LSD = 0.31 at P = 0.05).
When both wheat and soybean yields were regressed against mean profile cone indices, shown in Tables 2 and 3, yield decreased with increased cone index. This trend was not significant when data from all seasons were analyzed together. However, the trend was significant at P < 0.01 when data for each season were analyzed separately. The regression coefficients (r2s) ranged between 0.52 and 0.84 (Fig. 3) . Within the range of cone indices measured and on the basis of the slopes of these linear regressions, wheat yields were reduced 1.75 Mg ha-1 in 1994 for every megapascal increase in mean profile cone index and 1.50 Mg ha-1 in 1995. Similarly, soybean yields were reduced 1.55 Mg ha-1 in 1994 for every megapascal increase in mean profile cone index, 1.08 Mg ha-1 in 1995, and 1.81 Mg ha-1 in 1996.
|
| Summary and conclusions |
|---|
|
|
|---|
deep tilled both seasons. Spring only deep tillage maintained lower cone indices in the following double-cropped growing season than fall only deep tillage.
For season-by-season correlations, yields were reduced by an increase in mean profile cone indices. Wheat yields were reduced 1.75 Mg ha-1 in 1994 and 1.50 Mg ha-1 in 1995 for every megapascal increase in mean profile cone index, and soybean yields were reduced 1.55 Mg ha-1 in 1994, 1.08 Mg ha-1 in 1995, and 1.81 Mg ha-1 in 1996. On the basis of the mean profile cone indices given in Table 2 and the mean of the regressions of Fig. 3, soybean yield from the treatment with deep tillage in both spring and fall was reduced by 0.03 Mg ha-1 if deep tilled only in spring, by 0.73 Mg ha-1 if deep tilled only in fall, and by 1.38 Mg ha-1 if not deep tilled. Similarly, wheat yield was reduced by 0.18 Mg ha-1 if deep tilled only in spring, by 0.30 Mg ha-1 if deep tilled only in fall, and by 0.99 Mg ha-1 if not deep tilled. If producers deep till only once a year, spring appears to be better for this soil.SAS Institute Inc 1990
| NOTES |
|---|
|
|
|---|
2 Numbers in parentheses are log transforms. Analyses are based on log transforms. For the transform, 0.1 was added to each reading to prevent taking log(0). ![]()
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
R. L. Baumhardt, O. R. Jones, and R. C. Schwartz Long-Term Effects of Profile-Modifying Deep Plowing on Soil Properties and Crop Yield Soil Sci. Soc. Am. J., May 1, 2008; 72(3): 677 - 682. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. H. Schomberg, R. G. McDaniel, E. Mallard, D. M. Endale, D. S. Fisher, and M. L. Cabrera Conservation Tillage and Cover Crop Influences on Cotton Production on a Southeastern U.S. Coastal Plain Soil Agron. J., August 3, 2006; 98(5): 1247 - 1256. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Kubota, J. Bordon, K. Hoshiba, T. Horita, and K. Ogawa Change in Physical Properties of "Terra Rossa" Soils in Paraguay under No-tillage Soil Sci. Soc. Am. J., August 4, 2005; 69(5): 1448 - 1454. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. G. Hunt, P. J. Bauer, T. A. Matheny, and W. J. Busscher Crop Yield and Nitrogen Accumulation Response to Tillage of a Coastal Plain Soil Crop Sci., September 1, 2004; 44(5): 1673 - 1681. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Bruand, C. Hartmann, S. Ratana-Anupap, P. Sindhusen, R. Poss, and M. Hardy Composition, Fabric, and Porosity of an Arenic Haplustalf of Northeast Thailand: Relation to Penetration Resistance Soil Sci. Soc. Am. J., January 1, 2004; 68(1): 185 - 193. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. E. Herrick and T. L. Jones A dynamic cone penetrometer for measuring soil penetration resistance Soil Sci. Soc. Am. J., July 1, 2002; 66(4): 1320 - 1324. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||