Soil Science Society of America Journal 63:1319-1326 (1999)
© 1999 Soil Science Society of America
DIVISION S-6-SOIL & WATER MANAGEMENT & CONSERVATION
Bromide Leaching on a Piedmont Toposequence
G.L. Olsona and
D.K. Casselb
a Lockheed Martin Idaho Technologies Co., P.O. Box 1625, Idaho Falls, ID 83415-2107 USA
b Dep. of Soil Science, North Carolina State Univ., Raleigh, NC 27695-7619 USA
keith_cassel{at}ncsu.edu
 |
ABSTRACT
|
|---|
Solute leaching on the field-scale is a poorly understood, complex process affected by local soil variation and landscape position. We hypothesized that Br ion leaching is a function of landscape position in a well-structured, clayey soil on a Piedmont toposequence that had been in pasture for the previous 30 yr. Dry KBr, mixed with sand at a ratio of 12 g sand to 1 g KBr, was surface-applied at a rate of 314 kg Br ha-1 on 24 May 1994 along two transects and allowed to move into the soil under natural rainfall conditions. Soil cores 0.90- and 2.00-m-long were taken 13 June and 20 Dec. 1994, respectively (corresponding to 20 and 210 d after Br application and 15 and 63 cm rain, respectively). The cores were subdivided into increments 10 or 15 cm, oven dried, and analyzed for Br. Centers of Br mass at both sampling times were significantly deeper in the footslope position (31 and 82 cm for June and December, respectively) compared with the shoulder and linear slope positions combined (25 and 70 cm for June and December, respectively), which was possibly due to lower clay contents (44 vs. 50% clay) and lower water retention (37 vs. 43 cm in the top meter in December) for the footslope vs. the linear and shoulder slopes combined, respectively. Predicted leaching depths were calculated from measured soil water content profiles and were positively correlated with observed depths to the center of Br mass for the Dec. 20 sampling (r2 = 0.35, P < 0.007). Anion leaching may be partially controlled by landscape position, and soils susceptible to initial rapid leaching may not necessarily be susceptible to sustained rapid leaching throughout the year.
Abbreviations: AOV, analysis of variance • Ksat, saturated hydraulic conductivity • Q25, depth to the plane where 25% of the Br leached • Q50, depth to the plane where 50% of the Br leached • Q75, depth to the plane where 75% of the Br leached • IQR, interquartile range, IQR = Q75–Q25
|
INTRODUCTION
|
|---|
THE RELEVANCE OF SOIL-BASED GUIDELINES to the minimization of the leaching of agricultural chemicals hinges on our ability to predict leaching at the field scale. Quisenberry et al. (1993) proposed a soil classification system for determining solute leaching susceptibility. This system is based on nearly two decades of research on water and solute transport conducted in the southeastern USA. Soils are divided into groups based on texture of the A horizon because clay content influences soil aggregation, which favors preferential flow. Soils are then subdivided into groups based on properties of the subsoils (e.g., structure, clay type). In sandy soils, solute and water tend to move uniformly though the A horizon, and displacement of antecedent soil solution by invading water and dissolved solutes is nearly complete. Movement of solutes though clayey soils is less well defined, and the authors indicate that the clayey group will have to be further subdivided based on factors such as soil structure, aggregation, and clay type.
Although this system for classifying soils is highly desirable, natural variability of soil properties and the confounding influence of soil management likely will interfere with the interpretive value of the system. In general, high field variability and complex soil reactions, such as anion exclusion, anion exchange, and clogging of pores, pose problems for generalizing solute movement patterns across soil types (Jaynes and Rice, 1993).
Preferential movement of a solute can be greater than expected, and the depth of penetration of an anion after a short time can be surprising. While examining Br transport on a clay soil under natural rainfall conditions, Bronswijk et al. (1995) found that the maximum mean Br concentration was located 55 cm below the soil surface in response to only 34 mm of net precipitation following the Br application. In contrast, predicted preferential flow though veins in saprolite in the North Carolina Piedmont was not observed in a study by Williams and Vepraskas (1994), thus challenging the policy of restricting the installation of septic tanks over this material.
Predicting solute transport in field soils is complex; however, soil properties are somewhat predictable in a landscape. This recognized predictability is used in creating soil maps at the soil phase level. Daniels et al. (1985) demonstrated that certain landscape positions in the North Carolina Piedmont are typically associated with certain erosion classes and depth to the B horizon. Stone et al. (1985) showed that with increasing erosion on Piedmont soils, there was an associated loss of organic matter, a decrease in thickness of the Ap horizon, and an increase in total soil water content in the profile.
The somewhat predictable variability of soil properties with landscape position is expected to cause differences in anion transport. The objective of this study was to test our hypothesis that bromide transport is a function of landscape position in a well-structured, clayey soil on a mulched, nonvegetated Piedmont toposequence.
|
Materials and methods
|
|---|
Site Description
The experiment was conducted at the North Carolina Piedmont Research Station (35.4°N, 80.4°W, 220 m elev.) about 10 km west of Salisbury in Rowan County, North Carolina. At this location, mean annual precipitation is 115 cm and mean annual potential evapotranspiration exceeds rainfall from June though September. The 9-ha study site was situated on an east facing slope bounded on the upper end by a mixed deciduous forest, and at the lower end by a drainage ditch. The site had not been tilled during the past 30 yr and was continuously covered with fescue (Festuca spp.), which was harvested for hay. The soil is an eroded Hiwassee (fine, kaolinitic, thermic Typic Rhodudults) clay loam with some inclusions of Mecklenburg (fine, mixed, active, thermic Ultic Hapludalfs).
Twenty 3-m by 3-m plots were established on two transects (Fig. 1) . Each plot was oriented with its upper boundary along the contour. One transect, containing Plots 1 though 9, extended from the highest point in the field to the lowest; a second transect was longer and included Plots 9 though 20. Distance between adjacent plots ranged from 10 to 50 m.
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 1 Wire frame diagram of the study site constructed from elevation data. Plot locations are designated S, L, I, and F for shoulder, linear, interfluve, and foot slopes, respectively
|
|
All plots were classified by landscape position (Table 1)
. Four plots were located on the footslope position. Footslopes occur at the lowest position on the landscape, are less likely to be eroded, and generally support more vegetation than other landscape positions. Eleven plots were located on linear slopes that lie along hillsides or slopes and are subject to some erosion. Four plots were located on the shoulder slopes, which are situated on convex-shaped breaks of hills and are subject to erosion. Soil erosion exposes portions of the clay-rich Bt horizon, which impacts infiltration rates and other soil properties. Because only a small fraction of the total area of the field was occupied by interfluves, only one plot was located at this position. Interfluves are characterized as local high flat or nearly flat positions.
Plot Preparation and Bromide Application
The fescue on each plot was killed prior to Br application by spraying it with 4.7 L ha-1 glyphosate [isopropylamine salt of N-(phosphono-methyl)glycine] (Monsanto, St. Louis), and then the plot was mowed. Mowing the dead vegetation prior to Br application promoted uniform spreading of the Br salt over the soil surface and also provided mulch on the soil surface to minimize evaporation. On 24 May 1994, dry KBr, mixed with sand at a ratio of 12 g sand to 1 g KBr, was applied to the plots using a tractor-pulled Model 1006T Gandy fertilizer spreader (Gandy Co., Owatonna, MN). Past experience has shown that this ratio of sand to bromide provides a uniform Br application. Tractor speed and the spreader delivery system were calibrated to deliver the Br to each plot at a rate of 314 kg Br ha-1 (i.e., approximately 31 g Br m-2). A battery-operated recording rain gauge was installed in the field between the two transects.
Soil samples for Br analysis were taken 20 and 210 d after Br application (13 June and 20 Dec. 1994, respectively), which corresponds to 15 and 63 cm of cumulative rainfall after Br application, respectively (Fig. 2)
. Three soil cores were collected from each plot using a truck-mounted hydraulic probe. Core diameter was 6 cm for the June sampling and 5 cm for the December sampling. Holes were tightly back-filled with a mixture of sand and soil taken from the field.
The three soil cores from each plot for the June sampling were taken 1 m apart in the middle of the plot, along the contour. The cores were 90 cm long and were separated into 15-cm increments (n = 360). The three plot replicates for each depth increment were composited to give one sample per depth increment per plot (i.e., 20 plots x 6 depths, n =120). The three soil cores per plot for the December sampling were 200 cm long. These cores were taken 1 m apart along the contour on the down slope side of the plot. The top 60 cm of each core was sectioned into 15-cm increments and the remainder into 10-cm increments (i.e., 20 plots x 18 increments x 3 replicates, n = 1080). Samples where not composited for the December sampling, thus allowing variance component analysis (discussed below).
All soil samples were oven-dried and ground to pass through a 2-mm sieve before Br extraction. The June samples were first analyzed using an Orion Br-selective electrode (Orion Research, Boston, MA). A 10.0-g sample of dry soil was mixed with 1 mL of solution of 5 mol L-1 NaNO3 and 50 mL of distilled water and shaken for 20 min. Due to interference of an undetermined nature, the electrode results were unacceptable. The electrode readings did not stabilize during analysis, and duplicate analyses of the same samples on the same day were routinely 15% different, despite constant calibration of the instrument with control samples. A colorimetric procedure with a Lachat Quik Chem IV flow injection autoanalyzer (Lachat Instruments, Mequon, WI) was then used to analyze the samples.
For the colorimetric procedure, a 10.0-g sample of ground, oven-dried soil was weighed into a plastic vial. Fifty mL of distilled water and 1 mL of solution of 5 mol L-1 NaNO3 were added to the sample, shaken for 20 min, and filtered through #42 filter paper. The filtrate was poured into a glass tube for use with the colorimeter, which was equipped with a filter to remove organic molecules. This colorimeter was calibrated prior to sample analysis using eight standard solutions that ranged from 0.1 to 10 mg L-1 Br. Soil solution samples with a Br concentration exceeding 10 mg L-1 were diluted and reanalyzed. A 4 mg L-1 check standard of Br was analyzed after each group of eleven soil samples. If the check standard failed (>0.3 mg L-1 deviation), a new calibration curve was constructed before analysis would be resumed.
While colorimetry proved adequate for analysis of most soil samples, there was some interference on some soil surface samples, for which the filtrate was tinted yellow. These samples, although few in number, were not ignored, but were analyzed with the ion chromatograph. For ion chromatography, 10.0-g samples of oven-dried soil were shaken mechanically for 20 min with a 50-mL solution of 800 g L-1 Mg SO4. The solution was filtered using #42 filter paper and analyzed along with standards. Comparison of Br concentrations for the three analytical procedures vs. soil depth for Plot 3 (Fig. 3)
shows that results for the ion chromatograph and the colorimeter were similar and that the electrode results were consistently higher, particularly in the iron- and aluminum-rich subsurface samples. Gravimetric water content was measured for all soil samples. Gravimetric water content was converted to volumetric water content using measured bulk density values.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 3 Comparison of ion chromatograph (IC), colorimeter (Lachat Instruments, Mequon, WI), and ion selective electrode results from Plot 3 (linear slope)
|
|
Soil Physical Properties
Particle-size distribution was determined by the hydrometer method (Gee and Bauder, 1986) on 15-cm soil increments from the June samples. Saturated hydraulic conductivity (Amoozegar, 1988), bulk density, and porosity and volumetric soil water content at soil water pressure heads of 0, -60, -100, and -335 cm (Danielson and Sutherland, 1986) were determined by horizon on undisturbed soil cores that were collected from each plot on 14 June 1995. Soil cores for physical property determination were 10 cm in diameter and 8 to 10 cm long.
Calculating and Predicting Bromide Leaching Depth
The depth to the center of Br mass (Q50) is defined as the depth to which 50% of the measured Br leaches. The Q50 corresponds to the piston front of water when there is no preferential flow or dispersion (Smiles et al., 1978, 1981; Bond and Phillips, 1990), and it is a useful parameter for comparing Br leaching depths among plots, sample units, or treatments. The Q50 was calculated for each core by determining the cumulative amount of Br recovered in all sections of a soil core, dividing that value by 2, and then estimating by interpolation of the depth above which this cumulative amount of Br was stored.
Estimates of solute leaching depth are often needed. Most of the time data on actual evaportranspiration losses from the soil profile is not available. To predict the depth of leaching for each plot, effective precipitation estimates (i.e., precipitation minus estimated evaporation) are necessary. The effective precipitation estimates were derived by calculating the mean monthly potential evaporation (based on daily temperature data) using the Thornthwaite method (Thornthwaite, 1948), then reducing the potential evaporation by a specified amount to reflect the reduced evaporation that is due to mulch cover. We assumed for the first 2 d after a rainfall event (dates and amounts of rainfall are shown in Fig. 2) that evaporation from the soil surface was 50% of the calculated daily potential evaporation rate. This adjustment is based on the assumptions that both the mulch cover and the lack of live vegetation on plots reduce evaporation from the wet soil surface. From 2 to 10 d after a rainfall event, evaporation was estimated to be 25% of the daily evapotranspiration rate. If more than 10 d passed without additional precipitation, evaporation from the soil surface was assumed to be zero. Effective precipitation was 11.5 cm for the June sampling and 41.4 cm for the December sampling. In retrospect, semiweekly monitoring of soil water by neutron attenuation in each plot would have eliminated the necessity to estimate the effective precipitation.
The depth of leaching for each plot was predicted using the above calculated effective precipitation together with the measured soil water content data at the time of sampling. The volumetric soil water content (gravimetric water content multiplied by bulk density) was used to determine water storage in each soil decrement. For piston displacement, we assumed this amount of water would have to be displaced by the effective precipitation. For example, a 1-cm by 1-cm by 7.5-cm-deep volume of soil with a water content of 0.20 cm3 cm-3 would store 7.5 cm3 x 0.20 = 1.5 cm3 of water. In general, each depth increment of soil holds a different equivalent depth of water because bulk density and soil water content change with depth. The depth to which 11.5 cm of precipitation would penetrate the soil, if piston flow is assumed, is equal to the depth associated with the cumulative water storage of 11.5 cm3. Cumulative water storage downward from the soil surface was calculated for each plot and interpolated to determine the depth above which 11.5 cm3 of water was stored.
Statistical Procedures
Analysis of variance (AOV) was used to test the landscape-position effect for leaching depth (Q50), percent Br recovery, and some soil physical properties. For our AOV comparisons, analyses were conducted under the assumptions that landscape position is the primary criterion for classifying the plots, and that the experimental design is completely randomized. The transect (or block) effect was evaluated and found to have negligible contribution to the statistical model.
After selecting the two transects, we chose plots well-distributed along the transects, then identified landscape positions. This approach led to unequal replication for each treatment (landscape position) and transect (Table 1), but a better representation of the positions as they occur in the landscape. Every time the AOV was significant, a linear contrast (Steel and Torrie, 1980) was used to examine whether the footslope plots were significantly different from the linear and shoulder slope plots combined (i.e., footslopes vs. linear and shoulder slopes).
Simple and multiple regression analyses (SAS, 1989) were used to develop an equation to describe the observed Q50s. The simple regression used only the predicted depth of leaching to estimate Q50. If the predicted leaching depth was significantly correlated with the Q50 (P < 0.10), a multiple regression equation was developed using the following additional variables: Ksat of the surface horizon, Ksat of the second horizon, and maximum clay content within the top 1 m of soil. The term with the highest P value was dropped and the model was rerun until all terms had a P value < 0.15 (i.e., backward stepwise elimination method).
Variance components were calculated for Q50 for the December sampling using SAS Proc Mixed (SAS, 1989). This procedure allows both random and fixed terms in the model and accounts for correlated errors. Because the December samples consisted of three cores from each plot, it was possible to calculate within-plot variability (from the cores), plot-within-position variability, and position variance using the SAS mixed model. Landscape position was specified as a fixed variable, and plot-within-position was specified as a random variable; the residual was the within-plot variability.
|
Results and discussion
|
|---|
Mean percent Br content recoveries were 78 and 75% for the June and December samplings, respectively. Percent of Br content recovery was not significantly different by landscape position for any sampling time; however, mean Br concentrations were consistently higher in the greatest depths in the footslope position for both sampling times, suggesting we did not sample footslopes deeply enough.
The Br concentration distribution through time for Plot 16 at the shoulder slope position (Fig. 4)
is typical of most plots. The Br bulge became wider as the peak concentration moved deeper into the soil. Plot 3 at the linear slope (Fig. 5)
is an exception in that the width of the Br peak remained about the same in December as in June. The trend of deeper leaching and a wider peak through time is shown by landscape position in Fig. 6a and 6b .
View larger version (40K):
[in this window]
[in a new window]
|
Fig. 6 (a) Bromide concentration vs. depth for each landscape position in June. Mean Br recoveries were 93% for footslopes, (SD = 24), 76% for linear slopes (SD = 17), and 70% for shoulder slopes (SD = 15). (b) Bromide concentration vs. depth for each landscape position in December. Mean Br- recoveries were 69% (SD = 12)
|
|
The Q50 values are significantly deeper in the footslope than at the linear and shoulder slopes combined for both sampling times (P < 0.05). For June, average Q50s were 31 cm for footslopes, 25 cm for linear slopes, and 28 cm for shoulder slopes. For December, average Q50s were 82 cm for footslopes, 71 cm for linear slopes, and 66 cm for shoulder slopes. Increased leaching at the footslopes may result from a combination of more water reaching the footslopes via runoff plus adequate drainage from soil properties and landscape factors. While the contributions of runoff and drainage were not directly measured, the Ksat and clay content data in Table 2
suggest that the footslopes had greater potential for deeper leaching than the linear slopes did. Mean Ksat of the surface soils (Table 3)
tended to be higher in the footslopes than in the linear and shoulder slopes, although the contrast was not statistically significant. Maximum clay contents in the subsoil were lower (P < 0.001) and water contents at field capacity in December were lower (P < 0.01) in the footslopes than in the linear and shoulder slope positions combined (Table 3). Because less water is needed to displace resident soil water in the footslope than in the linear or shoulder positions, footslopes have the potential for greater leaching under the same rainfall conditions.
The variance component analysis of Q50 for December samples indicated that the within-plot variability was almost three times greater than variability among landscape plots or among positions. The high within-plot variability implies that Q50 varied so much from core to core that it is difficult to detect landscape position effects unless the Q50 differences are large, or numerous cores per plot are analyzed for Br. High within-plot variability likely stems from the inherent variability of soil properties that influence Br transport, but variability arising from nonuniform application of KBr to the soil surface cannot be discounted.
To further evaluate the Br distributions, we calculated the amount of water held within the depth that is associated with the Q50 of each plot and then compared this amount of water among landscape positions. Similar values would result when the same amount of water was used to leach Br to variable depths. For each plot, the cumulative water contents were calculated to the depth of Q50. For example, for Plot 1 in June, the Q50 was 27 cm, which is associated with 8.9 cm of water; i.e., the top 27 cm of soil held 
8.9 cm of water (4.8 cm in the 0–15 cm increment, which had a water content of 0.34 cm3 cm-3, plus 4.1 cm water in the 15–27 cm increment, which had a water content of 0.32 cm3 cm-3). For Plot 4 in June, the Q50 was also 27 cm, which corresponds to 10.0 cm of water for that plot. From this example, it is apparent that equal Q50s do not correspond to equal amounts of stored water; 8.9 cm of water is required for Plot 1, whereas 10.0 cm of water is required for Plot 4 to move Br to the same depth.
For both times, the footslopes appeared to hold slightly more water within the depth to Q50 than did the linear and shoulder slopes (Table 4)
, but the statistical contrasts were not significant. In June the water retention values associated with the Q50s were 10.8 cm for the footslope, 8.4 cm for linear, and 9.8 cm for the shoulder slope positions (effective precipitation in June was 11.8 cm). Water retention values in December were 29.9, 26.9, and 27.7 cm for the footslope, linear slope, and shoulder slopes, respectively (effective precipitation was 41.1 cm). The similarities in water retention values among landscape positions despite the multitude of potentially compounding factors suggests that effective precipitation and soil water content information are useful for estimating leaching depth.
View this table:
[in this window]
[in a new window]
|
Table 4 Mean values of water retention (cm) in the depths associated with Q50s (cm) in June and December by landscape position
|
|
Observed Q50s and predicted Q50s based on the water balance were highly significantly correlated for December samples (r2 = 0.35, P < 0.007, Fig. 7)
. A high r2 value for a linear regression of predicted to observed Q50, which passes through the origin, would be expected if the Br were not adsorbed to the soil solids, but moved through the soil as a front, pushed by the invading rainwater as it displaces resident water. The significant correlation at December suggests that the easily measured soil water content and effective precipitation parameters are useful for determining the relative leaching potential of a soil within a landscape.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 7 Predicted Q50 vs. observed Q50 for Br displacement at 20 locations on a Piedmont landscape. The regression line on the left omits the two very high, observed Q values. The one on the right includes the values. The slope >1 indicates Br retardation
|
|
The calculated Q50 water retention values were lower than the effective precipitation. Factors that may have contributed to the relatively high water retention values in the depth to Q50 are: (i) effective precipitation estimates were too high because of underestimation of evaporation from the soil surface; (ii) soil water content estimates were too high because of overestimation of bulk density; and (iii) Br leaching is retarded by chemical processes in the soil such as anion adsorption. This latter possibility is credible in light of the work by Bellini et al (1996) on anion repulsion in a Cecil soil (Typic Kanhapludult) in Georgia. They reported that anion (NO-3) retention increases as the pH decreases. Jury et al. (1991) indicate that the depth of water required to move an adsorbed solute to a given depth under piston flow is R times the depth of water required for a nonadsorbed solute. The slope of the predicted Q50 vs. observed Q50 curve in Fig. 7 is 1.49 and is equivalent to the retardation factor R. Although we did not measure pH in the subsoil, it typically is between 5 and 6. Bellini et al. (1996) found a NO-3 retardation factor of 1.5 to 2 in this pH range for the Cecil soil.
|
Conclusions
|
|---|
Several statistical analyses were used to evaluate the effect of landscape position on Br leaching in a well-structured soil on a Piedmont toposequence. Soil occurring on the linear and shoulder slope positions have significantly higher percentages of clay in the subsoil than soils on footslopes and are more effective in retarding the rate of Br leaching. Theoretically, increases in clay content are generally associated with the following soil properties that might impede Br leaching: (i) Increased specific surface area in the solid fraction of an iron- and magnesium-rich soil that is associated with an increased potential for transport retardation by anion exchange reactions on the positively charged sites; (ii) increased tortuosity within the soil matrix, which would lengthen the flow path for solutes under unsaturated conditions; and (iii) higher soil water contents, and hence, more water to displace during the leaching process. The observed shallower leaching on the clay-rich shoulder and linear slopes compared with the footslopes confirms this expectation.
Because soil properties related to leaching vary systematically on landscapes, the soil-based leaching classification system proposed by Quisenberry et al. (1993) might be modified to incorporate landscape position effects. Alternatively, water content in the top 1 m of soil at field capacity may be a useful indicator of a soil's leaching potential. Further definition of the macropores and their connectivity within the soil profile are needed to better estimate that fraction of Br carried deeper in the profile by preferential flow. Fine-tuning the leaching classification system might improve our ability to minimize the potential for contamination of groundwater. While runoff considerations may lead a field manager to apply more sludge to the footslopes than the shoulder or linear slopes, our results suggest that the footslopes are more susceptible to leaching losses than linear and shoulder slopes. Furthermore, depth to groundwater tends to be shallower on the footslopes than on linear slopes, amplifying the leaching risk on the footslopes.
|
ACKNOWLEDGMENTS
|
|---|
We wish to acknowledge Ellis Edwards for his extensive field and laboratory assistance and Guillermo Ramirez for assistance in analyzing the soil samples for Br and trouble-shooting the Br analysis procedures.
|
NOTES
|
|---|
Contribution from the Dep. of Soil Science, North Carolina State Univ. The use of trade names in this publication does not imply endorsement by the North Carolina Agricultural Research Services of the products named, nor criticism of similar ones not mentioned.
Received for publication February 3, 1998.
|
REFERENCES
|
|---|
- Amoozegar A. Preparing soil cores collected by a sampling probe for laboratory analysis of soil hydraulic properties. Soil Sci. Soc. Am. J. 1988;52:1814-1816.[Abstract/Free Full Text]
- Bellini G., Sumner M.E., Radcliffe D.E., Qafoku N.P. Anion transport through columns of highly weathered acid soil: Adsorption and retardation. Soil Sci. Soc. Am. J. 1996;60:132-137.[Abstract/Free Full Text]
- Bond W.J., Phillips I.R. Ion transport during unsteady water flow in an unsaturated clay soil. Soil Sci. Soc. Am. J. 1990;54:636-645.[Abstract/Free Full Text]
- Bronswijk J.J.B., Hamminga W., Oostindie K. Field-scale solute transport in a heavy clay soil. Water Resour. Res. 1995;31:517-526.
- Daniels R.B., Gilliam J.W., Cassel D.K., Nelson L.A. Soil erosion class and landscape position in the North Carolina Piedmont. Soil Sci. Soc. Am. J. 1985;49:991-995.[Abstract/Free Full Text]
- Danielson R.E., Sutherland P.L. Porosity. In: Klute A., ed. Methods of soil analysis. Part 1. 2nd ed. Agron. Monogr. 9. Madison, WI: ASA and SSSA, 1986:443-461.
- Gee G.W., Bauder J.W. Particle-size analysis. In: Klute A., ed. Methods of soil analysis. 2nd ed. Agron. Monogr. 9. Madison, WI: ASA, 1986:383-411.
- Jaynes D.B., Rice R.C. Transport of solutes as affected by irrigation method. Soil Sci. Soc. Am. J. 1993;57:1348-1353.[Abstract/Free Full Text]
- Jury W.A., Gardner W.R., Gardner W.H. Soil physics, 5th ed New York: John Wiley & Sons, 1991.
- Quisenberry V.L., Smith B.R., Phillips R.E., Scott H.D., Nortcliff S. A soil classification system for describing water and chemical transport. Soil Sci. 1993;156(5):306-315.
- SAS. SAS User's guide: Statistics. Cary, NC: SAS Inst, 1989.
- Smiles D.E., Perroux K.M., Zegelin S.J., Raats P.A.C. Hydrodynamic dispersion during constant rate absorption of water by soil. Soil Sci. Soc. Am. J. 1981;45:453-458.[Abstract/Free Full Text]
- Smiles D.E., Philip J.R., Knight J.H., Elrick D.E. Hydrodynamic dispersion during absorption of water by soil. Soil Sci. Soc. Am. J. 1978;42:229-234.[Abstract/Free Full Text]
- Steel R.G.D., Torrie J.H. Principles and procedures of statistics: A biometrical approach. New York: McGraw-Hill, 1980.
- Stone J.R., Gilliam J.W., Cassel D.K., Daniels R.B., Nelson L.A., Kleiss H.J. Effect of erosion and landscape position on the productivity of Piedmont soils. Soil Sci. Soc. Am. J. 1985;49:987-991.[Abstract/Free Full Text]
- Thornthwaite C.W. An approach toward a rational classification of climate. Geogr. Rev. 1948;38:55-94.
- Williams J.P., Vepraskas M.J. Solute movement through quartz-diorite saprolite containing quartz veins and biological macropores. J. Environ. Qual. 1994;23:810-815.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
D. E. Clay, Z. Zheng, Z. Liu, S. A. Clay, and T. P. Trooien
Bromide and Nitrate Movement through Undisturbed Soil Columns
J. Environ. Qual.,
January 1, 2004;
33(1):
338 - 342.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. S. Strock, D. K. Cassel, and M. L. Gumpertz
Spatial Variability of Water and Bromide Transport Through Variably Saturated Soil Blocks
Soil Sci. Soc. Am. J.,
November 1, 2001;
65(6):
1607 - 1617.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
