SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Nemati, M. R.
Right arrow Articles by Delaporte, L.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Nemati, M. R.
Right arrow Articles by Delaporte, L.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Nemati, M. R.
Right arrow Articles by Delaporte, L.
Related Collections
Right arrow Solute Transport Models
Right arrow Colloid-Facilitated Transport
Right arrow Soil Physics
Soil Science Society of America Journal 67:694-702 (2003)
© 2003 Soil Science Society of America

DIVISION S-1—SOIL PHYSICS

Contamination by Slaked Fragments with Sorbed Compounds in a Structured Soil

M. R. Nemati*,a, O. Bantonb, J. Caronc and L. Delaportea

a Université du Québec, INRS-Eau, 2800 rue Einstein, C.P. 7500, Sainte-Foy (QC), Canada, G1V 4C7
b Université du Québec, INRS-Eau, 2800 rue Einstein, C.P. 7500, Sainte-Foy (QC), Canada, G1V 4C7, or Laboratoire d'Hydrogéologie, Université d'Avignon, 33 rue Pasteur, 84000, Avignon, France
c Département des Sols et de Génie Agroalimentaire, FSAA, Université Laval. Québec (QC), Canada, G1K 7P4

* Corresponding author (reza.nemati{at}fafard.qc.ca)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
An adequate understanding of the mechanisms involved in solute transport through porous media is needed to help design management strategies aimed at controlling ground water contamination. The objectives of this study were, first, to evaluate the effect of soil structure changes on the processes of water and solute transport, and second, to assess the contribution of the fine particles detached during rapid wetting (slaked fragments) to sorbed solute transport. A laboratory study was conducted on a silty loam soil wetted at three rates (slow, medium, and rapid) after adding a soluble compound (Br) and a highly adsorbed compound (radioactive 137Cs) to sieved aggregates that were then deposited on the untreated soil surface. Soil particle migration (obtained from 137Cs measurements), Br transport, mean weight diameter (MWD), and wetting rate were measured following the wetting events. The results showed that, from the soil surface down to a depth of 100 mm, sorbed soil particle transport occurred in significantly greater amounts under the rapid wetting treatment than under the slow and medium wetting treatments. For one single rainfall event, the sorbed fraction that left the surface represented about 0.3% of the surface applied amount. This occurred despite the fact that soil disintegration following the rapid wetting process increased the residual mass of Br in the soil from 77% (slow wetting) to 88% (fast wetting) and decreased the speed of water and Br transport through the soil profile. These results imply that contaminants sorbed onto external aggregate surfaces could be transported through the soil profile more quickly and in greater quantities than predicted by conventional contaminant transport models.

Abbreviations: HEW, height of equivalent water • MWD, mean weight diameter • pdf, probability distribution function • PVC, polyvinyl chloride • TDR, time domain reflectometry


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
THE PRESENCE OF CONTAMINANTS in the ground water in many agricultural areas has raised concerns about the hazardous effects of such compounds on human health and the quality of the environment (Sprague et al., 2000). An adequate understanding of the mechanisms involved in contaminant transport through the soil is needed to design management strategies aimed at reducing environmental contamination. A number of parameters can influence the transport processes throughout the soil profile; these include the properties of the soil components, the contaminant chemistry, and the characteristics of the liquid phase (Laird et al., 1994). The stability of the soil pore network during water infiltration is one of the soil properties that may play an important role in solute and contaminant transport.

The stability of the soil pore network may be affected by the disruptive action of rapid wetting (Collis-George and Greene, 1979; Or, 1996; Nemati et al., 2000a; Nemati et al., 2000b), which in turn may affect the transport processes. Aggregate disintegration during rapid wetting is a form of soil structural degradation that is frequently encountered. Air compression and differential swelling have been presented as the two principal mechanisms involved in aggregate disintegration during rapid wetting (Kemper and Rosenau, 1984). When water completely surrounds an air-dried aggregate in which the pore space is occupied by air, water moves toward the aggregate center and compresses the trapped air. When the internal pressure from the trapped air exceeds the cohesion between the microaggregates, the aggregates release slaked fragments, mainly around their periphery. This process takes place rapidly, generally within 10 s (Zaher, 2001). At the same time, differential swelling because of wetting creates tangential stresses that intensify the disintegration process. Aggregate disintegration reduces soil porosity by decreasing the number of pores, alters the pore-size distribution, and decreases the infiltration rate (Collis-George and Greene, 1979; Kemper et al., 1988; Or, 1996). The microaggregates and the detached particles resulting from aggregate disintegration may clog the pore network and decrease soil permeability. If contaminants are sorbed onto external aggregate surfaces, they may be transported into the subsurface environment, as these peripheral fragments are more likely to slake and be transported in the liquid phase.

Evidence that colloidal particles are mobile in the soil horizons abounds in pedological literature. The presence of argillic horizons and the higher ratio of fine clay to total clay in illuvial horizons compared with eluvial horizons are examples of particle transport resulting from natural pedogenic processes occurring over a long period of time (Nettleton et al., 1975; Birkeland, 1984; Cabrera-Martinez et al., 1989). Argillic horizons result from several processes acting simultaneously or sequentially, including clay dispersion, clay dissolution, selective erosion, in situ clay formation, and ferrolysis (Soil Survey Staff, 1999).

However, the effects of particles detached during a single wetting event (colloidal particles) on solute transport processes are not well known and have not been quantified. The objectives of the present study were, first, to evaluate the effect of soil structure changes on the processes of solute transport, and second, to assess the importance of fine particles detached during aggregate disintegration in the transport of sorbed compounds.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Experimental Setup
A Platon silty loam soil (Haplorthod) from St-Augustin, Quebec, Canada was studied. The sampling site is in a cool, humid area with a yearly mean precipitation of about 1200 mm. Tillage operations had been performed regularly on this sampling site until 1991, when they were then discontinued. This soil type was chosen because preliminary tests showed that it was well structured (columnar at the bottom and granular-coarse at the top) and had a moderate infiltration rate affected by soil structure, an important parameter in this study.

Rapid Wetting Experiment
Polyvinyl chloride (PVC) cylinders, 5 mm thick and measuring 350 mm high with a 150-mm i.d., were used to extract soil cores from the field with minimum disturbance. The samples were collected from the surface layer of the soil in October 1999. A wood board and a sledgehammer were used to drive the cylinders into the soil to a depth of 300 mm. No compaction was observed during sampling. All 12 samples were taken from an area of 1 m2. Before sampling, the wall of each cylinder was perforated with two vertical rows of ten tapped holes; one row for time domain reflectometry (TDR) probes to monitor water content, and the other row for the insertion of tensiometers (to monitor the matric potential) or lysimeters (to sample the soil solution). Each TDR probe consisted of three 145-mm long stainless steel rods, measuring 2 mm in diameter and spaced 20 mm apart. The tensiometers were 80 mm long and 8 mm in diameter, and could be transformed into lysimeters when required. Holes were made at depths of 25, 50, 75, 100, 125, 150, 175, 200, 225, and 275 mm (Fig. 1). To remove the sampling cylinder casings without deforming the soil cores, the wall of each cylinder was cut vertically and then secured with rubber bands once the core extracted. The soil cores were stored at 4°C to reduce microbial activity.



View larger version (37K):
[in this window]
[in a new window]
  		Fig. 1. Structural setup of the experimental stand. A soil under study is kept in a polyvinyl chloride (PVC) sampling cylinder. An array of ten mini-tensiometers and ten time domain reflectometry (TDR) moisture mini-probes is installed, marking off ten monitored layers of soil.

 
Before starting the wetting experiment, three intact cores were selected randomly and soil characterizations were performed on each of these soil cores. First, the rubber bands around each cylinder were removed. Then a short PVC cylinder (25 or 50 cm long) of the same diameter as the soil column was placed above the sampling cylinder and the soil core was pushed from underneath into the short PVC cylinder. Once the desirable length of soil core had been inserted into the short PVC cylinder, the remaining soil was cut away using a sharp knife. Samples were taken at depths of: 0 to 25, 25 to 50, 50 to 75, 75 to 100, 100 to 125, 125 to 150, 150 to 175, 175 to 200, 200 to 250, and 250 to 300 mm. Each sample was mixed manually to obtain a uniform mixture and then weighed. A 20-g subsample of the field-moist soil from each depth was taken to measure the initial water content. The sample taken at each depth was divided into three parts and each part passed through one of the following sieves: 4, 2, and 0.5 mm. The soil properties characterized were particle-size distribution, bulk density, particle density, pH, cation-exchange capacity (CEC), organic C, organic matter content, MWD of aggregates, background Br concentration, and background radioactive 137Cs concentration. The characteristics of the soil used in this study are summarized in Table 1.


View this table:
[in this window]
[in a new window]
  		Table 1. Selected physical and chemical properties of Platon silty loam soil (Haplorthod), before tracer (Br, 137Cs) applications.

 
The soil particle-size distribution was determined using the hydrometer method (Day, 1965). The soil bulk density was measured using the Blake and Hartge method (1986a) where both the volume of the short PVC cylinder and the dry mass of the soil are known. The pycnometer method (Blake and Hartge, 1986b) was used to determine the soil particle density. Soil pH was measured using a pH meter after demineralized water had been added in a 1:2 (soil/water, weight/volume) ratio. The CEC was determined using the ammonium acetate method (McKeague, 1978) and the organic C with the Walkley-Black method (Nelson and Sommers, 1982) using a conversion factor of 1.333 (corresponding to the assumption that 75% of the total organic C is effectively oxidized). The organic matter content was then determined using a second conversion factor of 1.724 (corresponding to the assumption that total organic C represents 58% of the organic matter).

To determine the initial MWD of the aggregates at each depth, a 20-g sample of field-moist aggregates (smaller than 4 mm) was placed on a nest of sieves (sized 2.0, 1.0, 0.5, and 0.1 mm) and vibrated for 2 min at a frequency of 8 cycles per second. The weight of the soil retained on each sieve was recorded on an oven-dry basis. No correction was made for the presence of sand and gravel in the different size fractions. In this paper, aggregate-size distribution therefore refers to both aggregates and primary particles. The MWD was computed as follows:

[1]
where Wj is the proportion of the total sample weight occurring in fraction j, and Xj is the mean sieve size for fraction j. The initial water content of the sieved soil was determined gravimetrically. To reduce the effect of water content on the MWD during dry sieving, an attempt was made to measure the initial and the final MWD at similar water contents.

Bromide concentration was determined using a modified specific electrode method (Onken et al., 1975). A 25-g sample of the oven-dried 2-mm sieved soil was mixed with 55 mL of water in a 250-mL Erlenmeyer flask. The flask was agitated for 30 min on a wrist-action shaker and the suspension was then filtered through two superimposed Whatman #42 filter papers (Whatman Ltd, Maidstone, UK) at a suction of -33 kPa. One milliliter of total ionic strength adjuster (5 M NaNO3) was added to the 55 mL of filtrate. The Br concentration was measured using a Br electrode (Orion, Boston, MA). A calibration curve was established for each series of measurements since it was observed that the calibration changed from day to day.

An automatic {gamma} counter (Wallac 1480 WIZARD, PerkinElmer Life Sciences Inc., Boston, MA) was used to determine the 137Cs content. The 137Cs measurement was performed on 10-g sample of the air-dried 2-mm sieved soil. Counting times ranged from 3600 to 30000 s, depending on the 137Cs activity of the soil. This was sufficient to obtain a counting error of <10% at the 95% confidence level.

For the wetting experiment, nine intact cores (three treatments x three repetitions) were labeled randomly for the wetting treatments to be applied. The treatments and measurements were performed one block at a time. The samples were saturated by capillary rise, first using a tension table and then by saturation from underneath, raising the water level at a rate of 25 mm d-1. Time domain reflectometry probes and tensiometers were installed and each cylinder was drained by gravity. The soil cores were then further dried using an air-blowing system at the top of the cylinders to speed up the drying process and obtain a potential of -10 kPa or less at a depth of 20 cm. The TDR probes were read with a computer-controlled TDR system (Tektronix Metallic TDR Cable Tester, 1502b, Tektronix, Beaverton, OR), and tensions measured with a pressure transducer. These parameters were recorded by a Campbell CR10 datalogger using multiplexers (Campbell Scientific, Logan, UT). The water content and matric potential of the soil profiles were therefore monitored every 30 min for about 10 d. The drying process was stopped when a volumetric water content of about 0.15 cm3 cm-3 was obtained at a depth of 20 cm. This procedure allowed the top surface of the sample to dry to a water content corresponding to field conditions occurring during the summertime. In this case, the gradient of volumetric water content for a depth of 0 to 20 cm ranged between 0.01 and 0.15 cm3 cm-3.

During the drying process, soil shrinkage occurred at the surface, resulting in a space of about 5 mm between the soil core and the wall of the sampling cylinder. Consequently, TDR probes, tensiometers, and the sampling cylinder were removed with minimum disturbance from each soil core after drying. A thin, 80-mm wide paraffin roll was then immediately firmly wrapped around the core when still warm to ensure a tight fit. The sampling cylinder was then reinstalled around the core and secured with the rubber bands. The application of a paraffin roll around the soil core was intended to prevent water, solute (Br), and soil particle transport (137Cs) through the space between the soil core and the cylinder wall during the wetting experiment. Because of the elasticity of the paraffin roll, soil swelling was not restricted during wetting.

The soil surface was then prepared by removing the first 15 mm of topsoil from each cylinder and passing it through a 4-mm sieve. A 200-g sample (oven-dried basis) of 4-mm sieved soil was first moistened slowly to a gravimetric water content of about 25 ± 1% by spraying distilled water on the dried soil to prevent slaking. Then, a solution composed of 0.353 g (200 kg ha-1) of KBr and 10 000 Bq of 137CsCl dissolved in 50 mL of distilled water was uniformly sprayed on the soil and the soil was then mixed. The mixed soil was air dried to mimic soil surface drying and then transferred back to the designated cylinders. Cesium-137 was used because it is strongly adsorbed on soil particles (inorganic and organic) and becomes nonexchangeable (Lomenick and Tamura, 1965; Ritchie and McHenry, 1990). These properties and the fact that it is biologically and chemically stable make it unique as a tracer for studying physical particle transport in the soil environment (Ritchie and McHenry, 1990). The Br ion is a nonvolatile, dissolved solute that neither reacts nor adsorbs to soil solids. Hence it is often used in solute transport studies to trace the mobile phase through the soil.

After preparing the soil cores, the water volume required to saturate the dried soil core (i.e., the height of equivalent water, HEW) was calculated as follows:

[2]
where {theta}s (cm3 cm-3) is the water content of saturated soil, {theta}d (cm3 cm-3) is the water content of dried soil, and z (cm) represents the soil thickness.

After calculating the HEW, the TDR probes and the lysimeters were re-installed on each core. Using a tension infiltrometer and a water volume equal to the HEW, the cores were submitted to three different wetting treatments: a slow wetting (with a water potential of -1.0 kPa), a medium wetting (with a water potential of -0.2 kPa), and a rapid wetting (with a water potential of 0 kPa). The treatments were replicated three times in a completely randomized block design for a total of nine cylinders (three treatments, three replications). Three different rates of wetting were used to produce three different soil structural levels. Before wetting, a thin capillary mat (a geotextile, 2 mm thick and 433 g m-2 of density) was placed between the soil surface and the porous bottom surface of the tension infiltrometer (131.5-mm diameter) to obtain good contact and also to absorb the kinetic energy of the water. During wetting, readings of the water content were recorded at each depth every 3 min. Water content plotted as a function of time produced a curve with two distinct zones: an initial zone showing a sharp increase in water content with time, and a second zone showing a slower increase in water content with time. The wetting rate ({Delta}{theta}/{Delta}t) was determined by calculating the slope of the linear regression line for the zone with the sharpest increase in water content.

After the wetting treatments were applied, the cores were replaced on the tension tables adjusted at -0.7 kPa, and then one of three levels of rainfall (10, 30, and 50 mm) was simulated for each replicate. The rainfall simulations were performed using a tension infiltrometer adjusted at -0.7 kPa so as to minimize the effect on soil structure. The inflow (entrance-flux density) and the outflow were recorded each day. After every 10 mm of water was applied, the soil solution was sampled using the mini-lysimeters to measure the concentration of transported Br at each depth.

After completion of the rainfall simulation, each soil core was sampled by pushing the soil core from underneath into a short PVC cylinder. Samples (slices) were taken at intervals of 12.5 mm between 0 and 300 mm for a total of 24 samples. These samples were used to measure the resident Br concentration, transported 137Cs concentration, bulk density, and final MWD.

To determine the real concentration of transported 137Cs (Bq) for each depth, the background 137Cs concentration at each depth was subtracted from the final 137Cs concentration following wetting. The percentage of transported 137Cs for each depth was estimated from the proportion of the real concentration of transported 137Cs at each depth to the total concentration of 137Cs added to the soil surface (10 000 Bq). Then these results were converted to weight of transported particles per unitary surface (g m-2).

The depth distribution of the Br was expressed as a probability distribution function (pdf). The theoretical basis behind the approach chosen to describe the observed travel depth distribution has been fully discussed by Jury and Roth (1990) for stochastic-convective flow. A bimodal travel depth pdf was used, combining the two log-normal pdfs fri (z,t) described by Eq. [3] into a mixed distribution model (Hawkins, 1972) in Eq. [4]:

[3]

[4]
where µi and {sigma}i are the mean and the standard deviation characterizing the slow- (i = 1) and the fast-flow regions (i = 2), and z is the travel depth of the solute. The weighting factor ß therefore represents the importance of the slow-flow region relative to the fast-flow region. The percentage of recovery, %R, between the soil surface and any given depth L was calculated from Eq. [5]:

[5]
where m is the total mass of Br that was initially applied. The SAS system for Windows (SAS Institute, 1996) was used for the statistical analysis of the results for each individual depth. The wetting rate data were log-transformed to make them fit a normal distribution before running the statistical analyses. The results obtained for initial and final MWD were analyzed statistically using covariance analyses to eliminate the effect of water content on the MWD measurements. For the Br concentration and the simulation data, the analyses were performed on the summary statistics describing the shape of the depth distribution curves (µi, {sigma}i, ß). A protected least-significant test was used for multiple comparisons (at P = 0.05) in this study.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Wetting Rate
The maximum rate of wetting was observed near the soil surface at a depth of 25 mm and the wetting rate gradually decreased down to a depth of 200 mm (Fig. 2). The maximum rates of wetting at a depth of 25 mm were 10-2, 10-3, and 10-5 cm3 cm-3 s-1 for wetting potentials of 0, -0.2, and -1.0 kPa, respectively. The lowest wetting rates at a depth of 200 mm were 10-4, 10-4, and 10-6 cm3 cm-3 s-1 for wetting potentials of 0, -0.2, and -1.0 kPa, respectively. The statistical analyses of the wetting rate values showed that wetting rates decreased significantly (P <= 0.0001) with depth. The effects of the wetting potential treatments (-1.0, -0.2, and 0 kPa) on the wetting rate values were predominantly significant (P <= 0.0001) for all depths in the soil profile. It is therefore evident that the selected wetting potentials produced different water stresses throughout the soil profile between 0 and 200 mm.



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 2. Variation in wetting rates in the soil profile. Each value in this figure is the mean of three replicates. Bars represent the least significant differences (LSD) between treatments at P <= 0.05. Treatments at each depth marked by different letters are significantly different.

 
Aggregate-Size Distribution
The results of MWD measurements before and after the wetting treatments show that the final MWD values were significantly lower than the initial MWD values between a depth of 0 and 150 mm, except for the depths of 6.3 and 43.8 mm (Fig. 3). This decrease in final MWD values may be related to the impact of the wetting treatments and a prolonged rewetting at a high potential on the initially dry soils which could lead to aggregate weakening and disintegration and consequently, a decrease in MWD.



View larger version (35K):
[in this window]
[in a new window]
  		Fig. 3. Variation in mean weight diameter (MWD) in the soil profile, before and after wettings. Initial is the MWD value in the soil profile before wetting. Each value in this figure is the mean of three replicates. The depths are reported as the midpoint. Bars represent the least significant differences (LSD) between treatments at P <= 0.05. Treatments at each depth marked by different letters are significantly different.

 
The statistical analyses revealed no significant treatment effects for the final MWD values at all depths throughout the soil profile. This result suggests that the wetting treatments (slow, medium, and rapid) had the same impact on the final MWD values and therefore the differences between wetting treatments produced no significant differences in aggregate fragmentation into several subunits of large size.

Soil Particle Transport
The 137Cs measurements (adsorbed phase) indicated that the wetting treatments affected soil particle transport throughout the soil profile (Fig. 4). The rapid wetting treatment resulted in significantly more particle transport than did the medium and slow wetting treatments. The statistical analyses showed that down to a depth of approximately 100 mm, soil particle transport was significantly higher under rapid wetting than under medium or slow wetting. Since 137Cs was added to the first 15 mm of sieved soil at the top of the cylinder and the 137Cs concentration was not significantly different in the first two depths (0–25 mm), these data are not presented in Fig. 4. No significant differences in particle transport were observed for depths between 100 and 300 mm. The total quantity of particles transported from the first 15 mm of soil down to a depth of 100 mm following a rapid wetting of a dried soil was 37.6 g m-2 (376 kg ha-1), which represented 0.33% of the mass applied to the 15-mm soil surface layer. The results show that in relation to the total quantity of untransported particles, the quantity of detached particles was very small. It is therefore unlikely that the detached particles could cause significant changes in the final MWD. These results are in agreement with those obtained from the MWD measurements. Together, these results support the idea that very small soil particles are detached from the aggregate perimeter. Other researchers (Le Bissonnais, 1988; Zaher, 2001) have obtained similar results. They reported that even in free water, aggregates disintegrated only along their perimeter and rarely fragmented completely into large subunits. Although the quantity of transported particles seems to be small for a single wetting event, frequent rapid wetting of a dried soil can result in a considerable quantity of transported particles over time. These mobile particles are formed from clay colloids, possibly slaked microaggregates, Fe and Al oxides, carbonates, suspended organic matter, and microbes, and may act as transport agents in the soil (McCarthy and Zachara, 1989; Sprague et al., 2000). Because of their high specific surface area, colloids have a high sorption capacity and can be excellent sorbents for contaminants of low solubility (Sprague et al., 2000). Contaminants may therefore be adsorbed onto these particles and transported through the soil profile.



View larger version (26K):
[in this window]
[in a new window]
  		Fig. 4. Variation in soil particle transport in the soil profile following three wetting treatments. Each value in this figure is the mean of three replicates. The depths are reported at the midpoint. Bars represent the least significant differences (LSD) between treatments at P <= 0.05. Treatments at each depth marked by different letters are significantly different.

 
The same process may occur at the field scale, as evidences about colloidal and fine particles movement abounds in the literature. The results for soil particle transport are also presented for each level of precipitation, corresponding to each individual replicate (Fig. 5 and 6). The results indicate that the precipitation levels (10, 30, and 50 mm) affected soil particle transport in the soil profile. In most cases, the lower level of precipitation (10 mm) produced a significantly greater level of particle accumulation at the 0- to 190-mm depth than the higher levels of precipitation (30 and 50 mm). No significant differences were observed between precipitation levels for depths below 190 mm. The lower rate of particle accumulation with the higher rate of precipitation (50 mm) in the 0- to 190-mm depth, suggests that the particles must have migrated to lower depths or may have been removed from the soil cores with the drained water. The quantity of precipitation after wetting (saturated or near-saturated conditions) thus plays an important role in particle and contaminant transport. The extent of the differences in soil particle accumulation was dependent on the level of precipitation: the more precipitation, the less accumulation of particles in the soil profile.



View larger version (27K):
[in this window]
[in a new window]
  		Fig. 5. Variation in soil particle transport in the soil profile following three levels of precipitation (10, 30, and 50 mm). Each value in this figure is the mean of three replicates. The depths are reported as the midpoint. Bars represent the least significant differences (LSD) between treatments at P <= 0.05. Treatments at each depth marked by different letters are significantly different.

 


View larger version (37K):
[in this window]
[in a new window]
  		Fig. 6. Observed and fitted travel depth probability distribution functions for different wetting treatments and infiltration rates. The curves were obtained by the least-square method: the parameters obtained by the fitting of Eq. [3] and [4] are presented in Table 2.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Least-square estimates of the mixed travel depth probability distribution function parameters derived from the bromide distribution data.

 
Bromide Distribution
Figure 6 shows the distribution of the Br concentration at different depths between 0 and 300 mm for each replicate of the three wetting treatments. The general solute distribution followed a bell-curved leaching pattern typical of a pulse application. A bimodal (and perhaps multimodal in some cases) solute distribution was observed for all treatments, showing a slow-flow region followed by a fast-flow region. To quantify the relative contribution of both the slow- and the fast-flow regions to solute transport, µi, {sigma}i, and ß values were estimated using Eq. [3] and [4]. The estimated parameters are reported in Table 2.

The position of the slow-flow region, described by the mean of µ1, was: exp 1.95, exp 1.73, and exp 1.37 (i.e., a mean travel depth of 70.6, 56.5 and 39.2 mm, respectively) for the slow, medium, and rapid wetting treatments, respectively. The mean of the µ1 values for each precipitation level was: exp 1.23, exp 1.85, and exp 1.97 (i.e., a mean travel depth of 34.3, 63.9, and 71.4 mm, respectively) for 10, 30, and 50 mm, respectively, of simulated rainfall. The statistical analysis showed no significant effect of wetting treatments or precipitation levels on the µ1 values, but a general increase in µ1 value was observed with decreasing wetting rates and increasing precipitation levels.

In the statistical description of solute spreading, it is the fast-flow fraction that spreads out as a function of wetting potential, as shown by the increase in the µ2 values with a decreasing wetting rate. The position of the fast-flow region (µ2) shows more displacement with depth under the slow wetting than under medium or rapid wetting. The mean values for the position of the fast-flow region were: exp 3.13, exp 3.05, and exp 2.76 (i.e., a mean travel depth of 227.8, 211.6, and 157.7 mm, respectively) for the slow, medium, and rapid wetting treatments, respectively. The differences between the positions of the fast-flow regions for wetting treatments were not significant (P = 0.17), but the general trend observed suggests a more conductive pore system under slow wetting than under medium and rapid wettings. This decrease in Br mobility in the medium and rapidly wetted soils was probably because of the clogging of the pore network by the transported soil particles and to the increased tortuosity of the pores following the wetting event. This result suggests that in the slowly wetted soil Br arrives at any given depth after less drainage than in the medium and rapidly wetted soils. The position of the fast-flow region, µ2, also indicates faster movement with the increasing quantity of infiltrated water (precipitation level), with a value varying between exp 2.80 and exp 3.07 (i.e., a mean travel depth between 164.4 and 215.4 mm). The mean values of standard deviation for the fast-flow region, {sigma}2, were 0.346, 0.420, and 0.567 for the slow, medium, and rapid wetting treatments, respectively. This result suggests that the higher rate of wetting creates a greater dispersion or pulse spreading than does the lower rate of wetting.

The fitted ß values, describing the partitioning of the slow- and fast-flow regions, were significantly higher (P = 0.04) for the rapidly wetted soils than for the medium and slowly wetted soils. The mean ß values were 0.07, 0.12, and 0.18 for the slow, medium, and rapid wetting treatments, respectively. For the rapidly wetted soils, the higher ß value suggests less exchange between the slow and the fast regions during Br transport. This decrease in Br mobility may be attributed to the clogging of the pore network by the fine particles detached during rapid wetting (slaked fragments) and the reduced exchange between mobile and immobile water. Precipitation levels had no significant effect on ß values.

The percentage of recovery for the whole profile, %R, suggests that nearly all of the solute applied in the treated soils was still present in the profile after the wetting and precipitation events. No significant effect of wetting treatments was detected on the percentage of recovery, but a general increase in %R was observed with increasing wetting rates. The mean recovery values were 77.0, 81.0, and 87.6% for the slow, medium, and rapid wetting treatments, respectively. The means of the %R values for each precipitation level were not significantly different, varying between 78.4 and 86.8%. The mean Br resident concentration in soil cores was 38.9, 38.6, 47.0 µg per 1 g of dried soil for the slow, medium, and rapid wetting treatments, respectively. In general, the mean Br resident concentration was higher in the rapidly wetted soil than in the medium and slowly wetted soils, but these results were not significantly different. Since equal quantities of Br were initially applied on the surface of all treated soils (0.353 g), the Br resident concentration results indicate that more Br was leached from the slowly and medium wetted soils than from the rapidly wetted soil. The results obtained from the Br resident concentration and recovery measurements were consistent with the leached Br measurements. The results also suggest that the pore network was more conductive in the slowly wetted soil than in the rapidly wetted soil. No significant effects of wetting treatments or precipitation levels were observed for other parameters.

The results of this study therefore suggest that there are two different ways in which slaking affects solute movement in a structured silty loam soil. First, the clogging of soil pores because of slaking will decrease the importance of the fast-flow region relative to the slow-flow region. This mechanism may be the one that controls the transport of nitrate, K, Br and weakly sorbed pesticides. Slaked fragments containing strongly sorbed particles like Cs, P, and some pesticides may also move with the wetting front. This second mechanism may be of considerable importance as these particles can move in macropores and contaminate ground water at a much faster rate than would be predicted using classical solute transport models. Indeed, such theoretical models, whether they are of a Stochastic-Convective or of a Convective-Dispersive type (Jury and Roth, 1990) would predict that a strongly sorbed solute would be retarded relative to Br. Therefore, in the fast wetting treatment, Cs should move slower than Br and faster in the slow wetting treatment relative to the fast wetting treatment. The opposite is observed, Cs moving faster in the fast wetting treatment relative to the slow wetting treatment, a feature clearly different from that of Br and then possibly because of particle displacement, a factor ignored in solute transport models. This behavior of Cs could not be predicted from the observed Br displacement, once slowed proportionally to the Cs-retardation factor. An attempt should therefore be made to incorporate this additional feature because of slaking as a mechanistic component in contaminant transport models.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
For a mobile tracer (Br), the increased rate of wetting decreased the relative importance of the fast- to slow-flow regions, possibly because of pore clogging and the increased tortuosity of the pores. Increased pore tortuosity can also lead to an increase in the residual mass of solute in the soil and decrease the speed of water and solute transport through the soil profile. For a strongly sorbed tracer (137Cs), the increased rate of wetting resulted in the tracer being transported through the soil profile faster and in greater quantities, in contradiction with the conventional contaminant transport models. Indeed, the results showed increased contamination, possibly because of small-slaked fragments.


   ACKNOWLEDGMENTS
 
The authors thank Dr. David E. Elrick, Dr. Marie Larocque, and Dr. Lionel Mabit for their useful comments during the revision of this manuscript. Thanks are extended to Stéfane Prémont for his laboratory assistance. Thanks are also extended to the National Institute of Scientific Research for the financial support accorded to M. R. Nemati.

Received for publication December 31, 2001.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Nemati, M. R.
Right arrow Articles by Delaporte, L.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Nemati, M. R.
Right arrow Articles by Delaporte, L.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Nemati, M. R.
Right arrow Articles by Delaporte, L.
Related Collections
Right arrow Solute Transport Models
Right arrow Colloid-Facilitated Transport
Right arrow Soil Physics

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