Impact
Statement

PDF version
for printing

© 2007 Plant Management Network.
Accepted for publication 17 May 2007. Published 6 August 2007.

Predictability of Crop Production in a Clay Soil Based on a Comprehensive, Post-land-leveling Soil Property Evaluation

Kristofor R. Brye, Department of Crop, Soil, and Environmental Sciences, University of Arkansas, Fayetteville 72701

Corresponding author: Kristofor R. Brye. kbrye@uark.edu

Brye, K. R. 2007. Predictability of crop production in a clay soil based on a comprehensive, post-land-leveling soil property evaluation. Online. Crop Management doi:10.1094/CM-2007-0806-01-RS.

Abstract

Land leveling is a common yet severe soil disturbance in the rice-producing regions of the mid-southern United States. Heavy earth-moving equipment is used to grade fields to improve surface drainage and increase uniformity of flood irrigation applications to crops such as rice (Oryza sativa L.) and soybean [Glycine max (L.) Merr.]. The objective of this study was to determine whether crop yield and dry matter production were predictable based on a comprehensive soil property characterization following land leveling of a clay soil (with maximum cuts of -0.95 ft and fills of +0.19 ft) in the upper Mississippi River Delta region. Land leveling resulted in significantly (P < 0.05) lower soil organic matter, total N and C, C:N ratio, extractable P, fungal biomass, and fungal:bacterial biomass ratio and significantly (P < 0.05) greater bulk density in the top 4 inches. Land leveling also resulted in significantly increased soil property variability. Significant correlations between soil properties and crop responses were generally weak (r < 0.4) and inconsistent across crops and growing seasons. Results indicate that soybean and rice growth and yield response in the first three growing seasons following land leveling were not reliably predicted based on a suite of 25 immediate post-leveling soil physical, chemical, and biological properties. Based on the results of this and a previous study, it appears that the negative effects of land leveling on soil properties may be less in clay than in silt-loam soils. Though land leveling may facilitate the uniform distribution of irrigation waters, the resulting increased spatial variability in crop yield cannot be predicted by crop managers with current soil physical and chemical measurements.

Introduction

Under natural conditions, plant productivity is intimately related to the condition of the soil in which that plant is growing. For example, plant productivity would be expected to be high where sufficient N exists and low where insufficient N exists in the soil. Similarly, plant productivity would be expected to be high where sufficient moisture was present and low where insufficient moisture was present in the soil. This plant-soil relationship can certainly be extended to include the response of agricultural field crops to the soil in which they are grown. However, certain agricultural management practices, such as land leveling, can severely disrupt the near-surface natural condition of the soil (1).

Land leveling is relatively commonplace as a water conservation practice in the mid-southern United States, particularly in regions of rice (Oryza sativa L.) and soybean [Glycine max (L.) Merr.] production (2). Land leveling creates a slight surface gradient to facilitate the uniform distribution of irrigation water. However, to achieve the slight, uniform soil surface gradient, large, heavy machinery are necessary to remove soil from local high spots (i.e., a cut) and add soil to local low spots in a field (i.e., a fill). This removal, addition, and relocation of soil within a field during land leveling activities substantially alters the magnitude and spatial variability and distribution of soil properties throughout the field (1,2,3,4,5).

With increased spatial variability and distributions of soil properties following land leveling, one might reasonably expect that the plant-soil relationship be even more pronounced than under natural, undisturbed conditions. However, Brye et al. (6) demonstrated that a relatively comprehensive, immediate-post-leveling soil property evaluation, which included more than 20 physical, chemical, and biological properties, was unsuccessful at predicting crop response with any degree of confidence in the first (soybean) or second (rice) growing season after shallow-cut land leveling of a silt-loam Alfisol in south-central Arkansas.

The nature of the predominately silty alluvial parent material likely contributed greatly to outcome of the Brye et al. (6) study. Compared to the deep, highly clayey, alluvial Vertisols located nearer to the Mississippi River channel and its floodplain, the soil profile of the silt-loam Alfisol would tend to be more vertically differentiated meaning that there is more vertical soil property change, particularly with particle-size distribution and texture, from horizon-to-horizon in the silt-loam Alfisol than in a highly clayey Vertisol. Thus, one could contend that land leveling would have a greater negative impact on resulting soil properties and crop response in a silt-loam Alfisol than in a highly clayey Vertisol (7). This contention was supported by the results of previous studies (4,5).

Under the assumption that a clayey Vertisol is less prone to severe disturbance by land leveling activities than a silt-loam Alfisol, the objective of this study was to determine whether crop growth and production were predictable with some degree of confidence based on a comprehensive soil property characterization following land leveling of a clay soil in northeast Arkansas. Contrary to the results of Brye et al. (6), it was hypothesized that crop growth and production are correlated to near-surface soil properties immediately following relatively deep-cut land leveling of a clay soil and that crop response following land leveling would be somewhat predictable based on a reasonably comprehensive, post-leveling soil evaluation.

Site Description and Experimental Design

A 12-acre field, previously cropped to soybean, on a Sharkey clay soil (very-fine, smectitic, thermic Chromic Epiaquert) at the Northeast Research and Extension Center (NEREC), Keiser, AR, was land leveled in April 2004 (4). There was a uniform < 2% slope throughout the field prior to leveling such that irrigation water flowed from east to west, and contained north-south-oriented raised beds (< 15-cm-tall) every 12 m throughout the field that were left from previous research studies in this particular field (4,5).

Prior to land leveling, two 50-m wide by 100-m long study areas were established parallel to one another and separated by 25 m within the field. Each study area was divided into ten 10-m wide by 50-m long plots. One study area was used to evaluate the use of poultry litter, while the other study area was used to evaluate deep tillage as potential management practices that could be used to improve soil quality following land leveling in clay soils. Poultry-litter and deep-tillage treatments were randomized within each study area, such that a completely random experimental design resulted with five treatment replications and five control replications in each study area (8).

In addition to the poultry litter and deep tillage treatments, a 50-point grid system was superimposed onto each study area, such that grid points were evenly spaced at 10 m apart from one another. The grid system was established to facilitate soil and plant sampling from the same point in each study area from year to year to allow for the effects of land leveling over time to be evaluated (4,5). Details of the imposition of the poultry litter and deep tillage treatments will not be described here, but can be found in Brye (6). Similarly, additional details regarding the study site and experimental design can be found in Brye (5,6) and Brye et al. (4).

Field Management

Land leveling activities were described in detail by Brye et al. (4), thus only an abbreviated description follows. Land leveling activities began on 18 April and were completed on 20 April 2004. Following initial land leveling activities, the entire field was disked three times and land planned (i.e., floated) twice on 27 May to reduce soil-clod size to an approximate diameter of < 2 cm for a proper seed bed.

A Round-Up-ready (i.e., glyphosate resistant) soybean cultivar was drill seeded at a 7.5-inch row spacing throughout both study areas on 17 June 2004 (6). After emergence, approximately 1 week after planting, a 100 lb/acre rate of triple-super phosphate was manually applied with a hand spreader to both study areas. No K or N was applied to the soybean crop. Soybeans were furrow-irrigated on an as-needed basis throughout the growing season and harvested on 22 October 2004. The entire study area was left fallow over winter.

In 2005, two passes across both study areas were made with a soil conditioner (i.e., Do-All, Forrest City Machine Works, Forrest City, AR) and then land planned twice to prepare a proper seed bed (6). ‘Wells’ rice was drill seeded on 27 April at a 7.5-inch row spacing and at a seeding density of 100 lb/acre. On 9 June, at about the 5-leaf rice stage, a blanket application of N at 167 lb/acre as urea was spread manually across both study areas. Previous soil-test results indicated no additional P or K were needed for optimal rice production. The flood was established on 10 June and released on 26 August in preparation for harvest on 16 September 2005.

In 2006, the study area was prepared in a similar manner to that in 2005. ‘Wells’ rice was drill seeded on 28 April at a 7.5-inch row spacing and at a seeding density of 100 lb/acre. On 13 June, again at about the 5-leaf rice stage, a blanket application of N at 150 lb/acre as urea was spread manually across both study areas. Soil-test results again indicated no additional P or K were necessary. The flood was established on 14 June and released on 8 September in preparation for harvest on 27 September 2006.

Soil Sampling and Analyses

Immediately prior to (17 April) and shortly after land leveling activities were completed (29 and 30 April), elevation was measured using a laser level and stadia rod at each of the 50 grid points in each study area to characterize the relative elevational changes that occurred throughout the entire study area as a result of land leveling (4). Within two weeks following land leveling, soil samples were collected from the top 4 inches (10 cm) from each grid point throughout the entire study area to characterize post-leveling soil physical, chemical, and biological properties (4,5). In addition, pre-leveling soil samples were also collected from the same grid points from the poultry-litter study area only to assess the immediate effects of land leveling on soil properties.

For all soil sampling occurrences, a single 1.9-inch diameter soil core was collected from the 0- to 4-inch depth within an 8-inch radius surrounding each grid point, oven dried at 70°C for 48 h, and weighed for soil bulk density determination (4). The soil-core sampling chamber was beveled to the outside to minimize compaction upon sampling. Oven-dry soil was subsequently crushed and sieved to pass a 2-mm mesh screen for particle-size analysis using the 2-h hydrometer method (9). Oven-dry soil was also used for soil chemical property characterization [i.e., pH, electrical conductivity (EC), extractable nutrients, soil organic matter (OM), and total soil N and C] (5). Soil pH and EC were determined with an electrode on a 1:2 soil-to-water solution. Soil sub-samples were extracted with Mehlich-3 extractant solution (10) in a 1:10 soil-to-extractant-solution ratio and analyzed for extractable nutrients [i.e., P, K, Ca, Mg, Na, S, Fe, Mn, Zn, and Cu] using by inductively coupled argon-plasma spectrophotometry (CIROS CCD model, Spectro Analytical Instruments, MA). Organic matter was determined by weight-loss-on-ignition after 2 h at 360°C. Total soil C and N were determined by high-temperature combustion using a LECO CN-2000 analyzer (LECO Corp., St. Joseph, MI) and used to calculate soil C:N ratios.

A second set of samples consisting of ten 0.8-inch-diameter soil cores were collected and composited from the 0- to 4-inch depth from within the 8-inch radius surrounding each grid point for total fungal and bacterial biomass determinations (4).

Extractable soil nutrient and microbial biomass concentrations, expressed on a mass-per-mass basis, and the soil bulk density measured at each grid point were used to calculate extractable nutrient and microbial biomass contents, expressed on a mass-per-area basis for the top 10 cm (4 inches) of soil.

Plant Response Measurements

At maturity each year, a 3-ft section of the row straddling each grid point was cut at the soil surface and collected for total aboveground dry matter, yield, and partial harvest index determinations. Actual samples were collected on 22 October 2005 (soybean), 16 September 2005 (rice), and 26 September 2006 (rice). Plant samples were dried at approximately 30°C for two weeks in a forced-draft oven and weighed. In 2004, soybean samples were mechanically threshed to separate the seed from the remaining plant material. The seed was collected and weighed for soybean seed yield determination. Similar to Brye et al. (6), for rice grown in 2005 and 2006, all panicles in a sample were cut and removed at the first node and weighed for rice panicle yield determination. Soybean seed and rice panicle yields were divided by total aboveground dry matter to calculate a partial harvest index (PHI) at each grid point (6).

Data Analyses

Pre- and post-leveling soil property means and their variances were compared. Whole-field summary statistics were calculated for post-leveling soil properties and annual crop response variables. Linear correlations between post-leveling soil properties and annual crop response variables (i.e., total aboveground dry matter, yield, and PHI) were initially performed. Multiple linear regression analyses were then performed to demonstrate the predictability of annual crop response based on all 25 post-leveling soil properties measured. All statistical analyses were conducting with Minitab (Minitab 13.31, Minitab Inc., State College, PA).

Land Leveling Effects on Soil Properties

Land leveling smoothed out the raised beds, altered surface drainage to a South-North orientation, and resulted in a similar uniform < 2% slope as existed prior to leveling. Specifically, land leveling resulted in an average surface elevational change of -0.34 ft (i.e., an overall cut), ranging from + 0.19 ft (i.e., a fill) to -0.95 ft, across the entire study area (4,8). This degree of soil surface manipulation represented a significant amount of soil deposition, removal, and relocation throughout the study area. A 1-ft maximum cut in this study represents a relatively common and moderate soil manipulation for the region (4). Cuts of > 3 ft occur on clay soils in the region, but are less common.

Land leveling also resulted in significant changes to near-surface soil properties (Table 1). Numerous soil property magnitudes increased or decreased significantly as a result of land leveling (Table 1). Few near-surface soil properties remained unaffected by land leveling (Table 1). Similar to soil property magnitudes, the variability associated with many soil properties increased significantly due to land leveling resulting in a soil surface across the entire field that was less-uniform after land leveling than before land leveling occurred (Table 1). As hypothesized, it was expected that the degree of subsequent crop uniformity across the entire field would be correlated to the degree of post-leveling soil property uniformity. Whole-field, post-leveling soil property statistics are summarized in Table 2.

Table 1. The effect of land leveling on soil properties in the top 10 cm of a clayey Aquert in the Mississippi River Delta region of northeast Arkansas. Mean values (± standard error) are reported (n = 50). Also indicated is whether the variance changed [i.e., increased, decreased, or not significant (NS)] significantly (P < 0.05) as a result of land leveling.

 ^x Asterisks next to post-leveling means represent a significant difference (P ≤ 0.05) between pre- and post-leveling soil properties.

 ^y Soil physical and biological property means and variance change summarized from Brye et al. (4).

 ^z Soil chemical property means and variance change summarized from Brye (5).

Table 2. Whole-field summary (n = 100) of post-leveling soil physical, chemical, and biological properties in the top 10 cm of a clayey Aquert in the Mississippi River Delta region of northeast Arkansas.

 ^x Standard error (SE).

 ^y Coefficient of variation (CV) based on absolute value of the mean.

Post-leveling Crop Response

Except for a minimal, though statistically significant (P < 0.05), 6.5% larger soybean yield without deep tillage than with deep tillage in the first growing season (i.e., 2004) following land leveling (8), neither poultry litter nor deep-tillage affected crop yields in the three subsequent growing seasons following land leveling (8) (K. R. Brye, unpublished data). However, the annual effects of litter application and deep tillage on a grid-point-by-grid-point basis were still present and remained in the following analysis of post-leveling soil property correlatins with crop response.

In 2004, the first growing season following land leveling, soybean above-ground dry matter averaged 6654 lb/acre, yield averaged 2633 lb/acre (43.9 bu/acre based on 60 lb/bu), and PHI averaged 0.36 across the entire study area (Table 3). Soybean yield immediately following land leveling was substantially greater than the estimated whole-field average of 33 bu/acre prior to land leveling (Sam Atchley, personal communication, 2005) (8). Since irrigation has supplied ample water to recent crops grown in this field prior to land leveling, thus minimizing the potential impact of dis-similar rainfall between years, it is likely that land leveling caused a positive crop response. However, the exact explanation for the crop response is still unclear since near-surface soil bacterial and fungal biomass, organic matter, total C and N, and extractable P all decreased significantly, while bulk density, pH, and extractable K, Ca, and Mg all increased significantly as a result of land leveling (Table 1).

Table 3. Whole-field summary (n = 100) of total aboveground dry matter (AbvDM_tot), dry seed yield, and partial harvest index (PHI) for three consecutive years following land leveling of a clayey Aquert in the Mississippi River Delta region of northeast Arkansas.

 ^w Standard error (SE).

 ^x Coefficient of variation (CV) based on absolute value of the mean.

 ^x Rice yield is based on dry mass of panicles cut above first node.

 ^z Partial harvest index (PHI) for rice calculated based on total aboveground dry matter and dry mass of panicles cut above first node.

 * Asterisk denotes significant difference based on paired t-tests between 2005 and 2006 rice crop.

In 2005, the second growing season following land leveling, rice aboveground dry matter averaged 15,973 lb/acre, panicle yield averaged 8740 lb/acre (194 bu/acre based on 45 lb/bu), and PHI averaged 0.49 across the entire study area (Table 3). Rice had not been grown in this particular field recently, thus there is no field-specific, historic rice yield for comparison.

In 2006, the third growing season following land leveling, rice aboveground dry matter averaged 13,170 lb/acre, panicle yield averaged 6830 lb/acre (152 bu/acre based on 45 lb/bu), and PHI averaged 0.47 across the entire study area (Table 3). Rice aboveground dry matter, panicle yield, and PHI all decreased significantly (P < 0.05) from 2005 to 2006, but each crop response parameter had similar variabilities both years (Table 3).

Post-leveling Soil Property and Crop Response Correlations

Immediate, post-leveling soil property correlations to subsequent crop response was inconsistent, at best, from parameter to parameter and year to year (Tables 4 to 6). In 2004, the first growing season following land leveling, aboveground soybean dry matter, yield, and PHI were generally weakly (0.20 < | r | < 0.65), though significantly (P < 0.05), correlated both positively and negatively with 7, 7, and 15, respectively, of the 25 post-leveling soil properties evaluated (Table 4). Each crop response variable in 2004 was significantly correlated with at least one post-leveling physical, chemical, and biological soil property evaluated.

Table 4. Summary of significant linear correlations (r; n = 100) between immediate post-leveling, top 10 cm soil properties and total aboveground dry matter (AbvDM_tot), yield, and partial harvest index (PHI) for the first year (2004, soybean) following land leveling of a clayey Aquert in the Mississippi River Delta region of northeast Arkansas. Asterisks denote significant correlations at P ≤ 0.05 (*), P ≤ 0.01 (**), or P ≤ 0.001 (***). Non-significant correlations are denoted with NS.

Table 5. Summary of significant linear correlations (r; n = 100) between immediate post-leveling, top 10 cm soil properties and total aboveground dry matter (AbvDM_tot), yield, and partial harvest index (PHI) for the second year (2005, rice) following land leveling of a clayey Aquert in the Mississippi River Delta region of northeast Arkansas. Asterisks denote significant correlations at P ≤ 0.05 (*), P ≤ 0.01 (**), or P ≤ 0.001 (***). Non-significant correlations are denoted with NS.

Table 6. Summary of significant linear correlations (r; n = 100) between immediate post-leveling, top 10 cm soil properties and total aboveground dry matter (AbvDM_tot), yield, and partial harvest index (PHI) for the third year (2006, rice) following land leveling of a clayey Aquert in the Mississippi River Delta region of northeast Arkansas. Asterisks denote significant correlations at P ≤ 0.05 (*), P ≤ 0.01 (**), or P ≤ 0.001 (***). Non-significant correlations are denoted with NS.

Table 7. Linear regression summary of whole-model P and R² values when 25 immediately post-leveling soil properties were used as a comprehensive set of soil evaluation data to predict subsequent crop growth and production parameters [i.e., total aboveground dry matter (AbvDM_tot), yield, and partial harvest index (PHI)].

In 2005, the first rice growing season after land leveling, aboveground dry matter, yield, and PHI were only weakly (0.21 < r < 0.35), though significantly (P < 0.05), positively correlated with 6, 6, and 0, respectively, of the 25 post-leveling soil properties evaluated (Table 5). Crop response was unrelated to any post-leveling soil biological property evaluated in 2005.

In 2006, the second consecutive rice growing season after land leveling, aboveground dry matter, yield, and PHI were again only weakly (0.22 < | r | < 0.37), though significantly (P < 0.05), correlated both positively and negatively with 2, 6, and 10, respectively, of the 25 post-leveling soil properties evaluated (Table 6). At least one of the three crop response variables evaluated in 2006 was significantly correlated with at least one post-leveling physical, chemical, and biological soil property.

The lack of correlation consistency among soil properties and crop responses from year to year was somewhat surprising. For example, as one might expect, the more compacted the soil is (i.e., increasing bulk density), the poorer the crop response. This relationship was shown to exist for soybean in 2004, where aboveground dry matter was weakly (r = -0.20), though significantly (P <  0.05), negatively correlated with soil bulk density, indicating as bulk density increased, aboveground dry matter tended to decrease. However, the relationship was still significant, though opposite for soybean yield and PHI, where both were positively correlated (0.22 < r < 0.53) with soil bulk density. In 2005, rice response was unrelated to soil bulk density. However, in direct contrast to 2004, rice yield and PHI in 2006 were weakly, though significantly (P <  0.05), negatively correlated (r = -0.26 and -0.21, respectively) with soil bulk density.

One would also tend to expect crop response to be consistently correlated with other soil properties such as pH or organic matter. However, crop response was unrelated to soil pH in any of the first three growing seasons following land leveling (Tables 4, 5, and 6). Similarly, crop response was unrelated to soil organic matter in the first two growing seasons, but soil organic matter was weakly, though significantly (P < 0.05), positively correlated (r » 0.23) with rice aboveground dry matter and yield in 2006, the third growing season following land leveling (Table 6).

Walker et al. (11) demonstrated a significant correlation between rice yield and the amount of soil manipulated in an area (i.e., whether the area was a cut or a fill) in a recently leveled clay soil in Mississippi. However, in this study, the estimated amount of soil moved on a grid-point-by-grid-point basis was only weakly, though significantly (P < 0.001) positively correlated with rice yield (r = 0.34) and aboveground dry matter (r = 0.32) in 2005, the second growing season after leveling (Table 5). There was no correlation between estimated soil moved and yield for soybean in 2004 or rice in 2006 (Tables 4, 5, and 6). In contrast, rice PHI in 2006, the third growing season after leveling, was weakly, though significantly (P < 0.01), negatively correlated with estimated soil moved (Table 6), indicating that rice PHI tended to be greater (i.e., greater grain mass per unit of total aboveground dry matter) in areas where soil was removed and tended to be smaller (i.e., less grain mass per unit of total aboveground dry matter) in areas where soil was added.

Based on a multiple linear regression approach using 25 post-leveling soil properties, including physical, chemical, and biological properties, the greatest degree of crop response predictability, as one might expect, was observed in the first growing season after land leveling. The 25-variable regression model was significant for soybean aboveground dry matter (P =  0.007), yield (P < 0.001), and PHI (P < 0.001), but the models only explained between 42 and 58% of the variability in crop response (Table 7). Except for rice aboveground dry matter in 2005 (P = 0.013; R² = 0.401), the second growing season after leveling, the 25-variable regression models were non-significant for all other crop responses (P > 0.09) with only between 22 and 34% of the crop response variability being explained with the comprehensive soil property evaluation (Table 7).

Practical Implications

In theory, land leveling is conducted to facilitate the delivery of irrigation waters to fields that under natural conditions have surfaces that undulate too much to uniformly apply water across the whole field. However, though improved uniformity of applied irrigation waters may be achieved, land leveling severely disrupts the biogeochemical equilibrium of the subsequent plow layer and root zone. Recent evidence exists that demonstrates how relatively shallow-cut land leveling in a silt-loam soil and relatively deep-cut land leveling in a clay soil in the rice-growing region of eastern Arkansas result in more spatially variable soil properties and crop response than existed prior to land leveling. The lack of apparent crop response predictability based on a comprehensive post-leveling soil property evaluation indicates that the post-leveling management of recently leveling fields, regardless of soil texture (i.e., silt loam or clay) may be quite challenging to sustain high productivity beyond the initial few growing seasons. The observations made in this study regarding the lack of consistent soil property correlations to crop response in a land leveled clay soil suggest that possible solutions, such as variable-rate herbicide and fertilizer applications, deep-tillage to alleviate compaction during land leveling, and the addition of organic soil amendments like poultry litter, may not be as effective as once thought at improving the uniformity of crop growth and production in the long term.

Acknowledgments

This research was funded by the Arkansas Rice Research and Promotion Board. Field assistance provided by Mandy Pirani, Matt Cordell, Tarra Verkler, Lee Riley, Mike Duren, and other NEREC station personnel is gratefully acknowledged.

Literature Cited

1. Brye, K. R., Slaton, N. A., and Norman, R. J. 2005. Penetration resistance as affected by shallow-cut land leveling and cropping. Soil Till. Res. 81:1-13.

2. Brye, K. R., Slaton, N. A., Savin, M. C., Norman, R. J., and Miller, D. M. 2003. Short-term effects of land leveling on soil physical properties and microbial biomass. Soil Sci. Soc. Am. J. 67:1405-1417.

3. Brye, K. R., Slaton, N. A., Mozaffari, M., Savin, M. C., Norman, R. J., and Miller, D. M. 2004. Short-term effects of land leveling on soil chemical properties and their relationships with microbial biomass. Soil Sci. Soc. Am. J. 68:924-934.

4. Brye, K. R., Slaton, N. A., and Norman, R. J. 2006. Soil physical and biological properties as affected by land leveling in a clayey Aquert. Soil Sci. Soc. Am. J. 70:631-642.

5. Brye, K. R. 2006. Soil biochemical properties as affected by land leveling in a clayey Aquert. Soil Sci. Soc. Am. J. 70:1129-1139.

6. Brye, K. R., Slaton, N. A., and Chen, P. 2004. Can crop productivity be predicted by soil characterization after land leveling? Online. Crop Management doi:10.1094/CM-2004-0908-01-RS.

7. Norman, R. J., Wilson, C. E., and Slaton, N. A. 2003. Soil fertilization and mineral nutrition in U.S. mechanized rice culture. Pages 331-411 in: Rice: Origin, History, Technology, and Production. C. W. Smith, ed. John Wiley & Sons Inc., New York, NY.

8. Brye, K. R. 2006. Crop response to poultry litter and deep tillage following land leveling in a clay soil. Online. Crop Management doi:10.1094/CM-2006-1215-01-RS.

9. Arshad, M. A., Lowery, B., and Grossman, B. 1996. Physical tests for monitoring soil quality. Page 123-141 in: Methods for Assessing Soil Quality. J. W. Doran and A. J. Jones, ed. SSSA Spec. Publ. 49. SSSA, Madison, WI.

10. Tucker, M. R. 1992. Determination of phosphorous by Mehlich 3 extraction. Pages 6-8 in: Soil and Media Diagnostic Procedures for the Southern Region of the United States. S. J. Donohue, ed. Vir. Agric. Exp. Stn. Ser. Bull. 374. Blacksburg, VA.

11. Walker, T. W., Kingery, W. L., Street, J. E., Cox, M. S., Oldham, J. L., Gerard, P. D., and Han, F. X. 2003. Rice yield and soil chemical properties as affected by precision land leveling in alluvial soils. Agron. J. 95:1483-1488.