Published online 26 May 2006
Published in Vadose Zone J 5:649-656 (2006)
DOI: 10.2136/vzj2005.0068
© 2006 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
ORIGINAL RESEARCH
Colloid and Bromide Transport in Undisturbed Soil Columns
Application of Two-Region Model
Tjalfe G. Poulsena,*,
Per Moldrupa,
Lis W. de Jongeb and
Toshiko Komatsuc
a Institute of Life Sciences, Environmental Engineering Section, Aalborg University, Sohngaardsholmsvej 57, DK-9000 Aalborg, Denmark
b Dep. of Agroecology, Soil Physics and Chemistry Section, Danish Institute of Agricultural Sciences, P.O. Box 50, DK-8830 Tjele, Denmark
c Dep. of Biological and Environmental Sciences, Graduate School of Science and Engineering, Saitama University, 255 Shimo-okubo, Sakura-ku, Saitama, 338-8570, Japan
* Corresponding author (tgp{at}bio.aau.dk)
Received 23 May 2005.
 |
ABSTRACT
|
|---|
Bromide tracer breakthrough and natural soil colloid leaching curves for undisturbed soil columns were used to characterize dissolved and suspended matter transport at the field scale. Data from 33 columns of 20-cm diameter and 20-cm height were used. Columns were collected in a grid of 25 by 30 m at an agricultural field. A two-region (mobile–immobile water phase, MIM) solute transport model was fitted to data. The model was used to estimate bromide and colloid transport parameters including mobile and immobile water contents (
m, im), bromide and colloid advective velocities (vBr, vColl), and mobile–immobile mass transfer coefficients (
br, coll). Both soil physical properties and transport parameters were highly variable across the sampling field. Comparison of bromide transport parameters with basic soil physical properties revealed that vBr was proportional to soil clay content and bulk density (
b), but Br was inversely proportional to these parameters. Colloid transport parameters, vColl and Coll, however, showed only a weak correlation with clay content and b. Also, vColl was typically three to four times higher than vBr. The colloid velocity was generally higher than the bromide velocity, implying size exclusion of colloids. The spatial distributions of soil physical properties, bromide and colloid transport parameters, and leached quantities of particles were compared. The results suggested that bromide and colloid mass transfer (diffusion) were not controlled by the same soil physical conditions, and that soil structure and macropore flow are more important than the quantity of dispersible colloids in controlling colloid leaching.
Abbreviations: MIM, mobile–immobile phases • SSD, sum of squared deviations • WDC, water dispersible colloids • 2MIM, 2 mobile and one immobile phases
|
INTRODUCTION
|
|---|
MOVEMENT OF STRONGLY adsorbing compounds such as P, heavy metals, and many organic chemicals in natural undisturbed soil is partly controlled by mobilization and transport of colloids, to which the chemicals are adsorbed. Several studies have documented the role of colloid movement in the transport of P, poly aromatic hydrocarbons, and pesticides (Kretzschmar et al., 1999; McCarthy and McKay 2004; de Jonge et al., 2004a; Laegdsmand et al., 2004). Understanding colloid mobilization and transport is therefore necessary, for instance, when evaluating the risk of groundwater contamination due to colloid-facilitated P and pesticide leaching in agricultural soils.
Colloid mobilization potential is governed by several factors, including soil clay content and pore-water velocity (Kjaergaard et al., 2004a, 2004b). High clay contents can increase the content of easily water dispersible colloids (WDC), and increased water shear will cause increased mobilization of these colloids. Colloid movement and transport of compounds adsorbed to colloids will depend on the transport parameters governing colloid transport. When mobilized, the colloids are transported in a way similar to transport of dissolved compounds in soil. Transport is thus governed by dispersion, advection, and exchange of colloids between mobile and immobile water phases in the soil (Schelde et al., 2002; Kjaergaard et al., 2004a). Transport of colloids and compounds adsorbed to them may therefore be modeled similarly to transport of dissolved compounds in soils.
Transport of dissolved compounds in soils has been modeled using different strategies, and the selection of model approaches available for transport simulation is extensive. In general transport has been modeled using models that include an immobile water phase in combination with one or more mobile water phases. Examples of studies using models with one immobile and one mobile water phase (MIM models) are van Genuchten and Wierenga (1977), Gaber et al. (1995), and Langner et al. (1999). Two mobile phases and one immobile phase (2MIM model) were considered by Morizawa et al. (1986). Both MIM and 2MIM models have been applied successfully to describe colloid transport in soils (Schelde et al., 2002; Kjaergaard et al., 2004a).
To model transport of colloids and of compounds sorbed to colloids it is necessary to know the values of the governing transport parameters, including mass transfer and dispersion coefficients and how soil chemical and physical properties affect these transport parameters. While values for parameters governing solute transport and their relation to soil properties are fairly well established for a wide selection of solutes encountered in soils, this is not the case for colloids. Kjaergaard et al. (2004a) found that the content of mobile water in which colloid transport takes place is present in pores with diameters >30 µm. This mobile water content was found inversely proportional to clay content for a range of Danish soils with clay contents between 12 and 43%. Apart from these findings, however, the present knowledge about colloid transport parameters in natural, undisturbed soil systems, and their dependence on soil properties is very limited.
Our objective is to evaluate possible relationships between colloid transport parameters and soil physical properties. The evaluation will be based on measurements of colloid and tracer transport and breakthrough in undisturbed soil in combination with numerical transport modeling. Transport parameters to be investigated are mobile water content, advective transport velocity, dispersion, and mass transfer between mobile and immobile water phases. Soil physical properties considered are clay content, content of WDC, soil bulk density, air permeability, and equivalent macro-pore diameter as inferred from air permeability and air-filled porosity.
|
DATA USED
|
|---|
The colloid leaching data used in this study were taken from de Jonge et al. (2004b). They measured natural colloid leaching curves in 42 intact soil columns of 20-cm diameter and 20-cm height. Leached particles with a diameter of <2 µm were considered colloids. The amounts of easily WDC were measured for each collection point. In the present study, bromide tracer breakthrough curves measured on the same columns but not published previously were also used. The columns were collected at an agricultural field at Røgen, Denmark and in a grid arrangement of 25 by 30 m, as shown in Fig. 1
. To avoid effects of tilling on the results, sampling points were displaced randomly 1 m away or kept at the grid crossing. During the transport experiments, columns were irrigated with artificial rain water at an intensity of 10 mm h–1, and a pulse of bromide tracer was added during the initial 10 min of the experiment. Effluent bromide and colloid concentrations (turbidity) were measured as functions of time. Before the experiment columns had been drained to a soil water potential of –20 cm H2O and water content and air permeability (ka20) were measured at that potential. During the experiments, however, the water content increased for two reasons. First, more water was added to the columns than was drained during the experiments. Second, the water flowing through the columns takes up additional space. The total water content used in the modeling was therefore calculated as the sum of the initial water content 20; (at –20 cm potential), the additional water stored in the column through the course of the experiment, and the average volume of the moving water in the column.
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 1. Sampling grid structure and location of samples. Filled circles indicate the 33 samples (columns) used in this study.
|
|
The increase in water content due to storage (storage) in each column was determined as
 | [1] |
where Vin is total cumulative inflow, Vout is total cumulative outflow, and Vcolumn is the total volume of the column. The average increase in water content due to water flowing through column (flow) was determined as
 | [2] |
where T is the duration of the experiment, A is the cross-sectional area of the column, and vmacro is the macropore flow velocity assumed equal to bromide breakthrough time divided by column height as suggested by de Jonge et al. (2004b). In this study breakthrough bromide and colloid data from 33 of the 42 samples were used. For 8 of the 42 columns breakthrough data could not be measured because of clogging and ponding during the transport experiments. Due to problems with accurate determination of bromide effluent mass for one of the remaining columns, breakthrough data for this column were also excluded from this study. The locations of the remaining 33 columns used in this study are shown in Fig. 1; their physical and transport properties are shown in Table 1.
|
MODEL THEORY
|
|---|
Bromide breakthrough and colloid leaching curves showed only single peaks in concentration, unlike the studies of Schelde et al. (2002) and Kjaergaard et al. (2004a) where in some cases twin peaks in concentration were observed. A choice was therefore made to model transport of both bromide and colloids using one mobile and one immobile water phase. Transport of bromide tracer and colloids was assumed to occur by advection and dispersion in the mobile water phase. Mass transfer of colloids and bromide between the mobile and immobile water phases was modeled as first order processes controlled by the difference in concentration between the two phases.
Transport of both colloids and the bromide tracer was modeled using the traditional advection-dispersion equation modified for mass transfer between the mobile and immobile water phases. The differential equation for bromide and colloid transport in the mobile water phase is,
 | [3] |
where Cm is the concentration of bromide or colloids in the mobile water, Cim is the concentration in the immobile water, D is the dispersion coefficient, R is the retardation factor (assumed equal to 1 for bromide transport), v is the pore water velocity, and is the mass transfer coefficient. The dispersion coefficient was determined as
 | [4] |
where 
is the dispersivity coefficient in the mobile water phase. Bromide concentrations in the immobile water were modeled by the following expression:
 | [5] |
where m and im are the mobile and immobile water contents, respectively. Colloid breakthrough curves showed constant concentrations at late times, indicating a constant release of colloids to the mobile water phase (Jacobsen et al., 1997; Motoshita et al., 2001; Kjaergaard et al., 2004b). It was therefore chosen to use a constant colloid concentration in the immobile water. This is equivalent to assuming instant equilibrium between the immobile water and the soil solid phase with respect to colloid release. The content of the immobile water and the pore-water velocity, v, were calculated from the mobile water content as
 | [6] |
 | [7] |
where total is the average, total volumetric water content, and q is the Darcy velocity. The governing equations were solved in Microsoft Excel using a forward time backward space finite difference scheme corrected for second-order numerical errors (Moldrup et al., 1994). A time step of 0.0015 h and a spatial step of 0.0136 m were used.
With total known and q measured during the transport experiments (de Jonge et al., 2004b), the parameters m, im, vBr, and Br were fitted from the bromide tracer breakthrough curves by minimizing the sum of squared deviations (SSD) between measured and predicted bromide effluent concentrations. The value of R for bromide transport was assumed to be 1; thus, the bromide transport velocity vBr and the pore water velocity v were equal. The initial bromide concentration in both water phases was assumed to be 0. The fitted values of m, im, vBr, and were then used together with the corresponding colloid breakthrough data to fit values of the colloid transport parameters Rcoll and coll. The initial amount of colloids in both the mobile and immobile water phases was assumed equal to the total amount of easily WDC measured during the experiments (de Jonge et al., 2004b). Because the columns had been equilibrating for some time before initiating water transport, it was assumed that the initial concentrations of colloids in the mobile and immobile water phases were identical. In five cases (Columns 20, 21, 25, 35, and 42), the measured value of WDC was very low compared with the measured colloid concentration in the effluent. For the remaining 28 columns the ratio between the effluent colloid concentration at 0.5 h after the start of the experiment and WDC was approximately 0.45. This ratio was therefore used to set the initial colloid concentration for the five columns mentioned instead of using WDC as initial condition in the modeling.
|
RESULTS AND DISCUSSION
|
|---|
Bromide and Colloid Breakthrough Curves
Breakthrough and leaching curves for bromide and colloids are shown in Fig. 2
and 3.
Selected types of breakthrough curves illustrating the transport properties of colloids and bromide tracer are shown in Fig. 2. Almost simultaneous breakthrough of bromide and colloids (Fig. 2a and 2b) corresponding to colloid retardation factor, Rcoll > 0.7 for colloid transport was found for approximately 10% of the columns. Enhanced breakthrough of colloids compared with bromide (Fig. 2c and 2d) corresponding to 0.7 > Rcoll > 0.3 was found for approximately 60% of the columns, and very early breakthrough of colloids (Fig. 2e and 2f) corresponding to Rcoll < 0.3 was found for approximately 30% of the columns. Breakthrough of bromide did not occur earlier than colloids for any of the columns; that is, colloid transport velocity, vcoll, was always larger than that of bromide. This implies selective (size) exclusion of colloids (White 1985)—colloids are mainly transported in the larger pores whereas bromide is also transported in smaller pores, resulting in a lower average transport velocity for bromide. Rates of mass transfer also varied greatly among columns, as illustrated in Fig. 2g and 2h. Column 11 shows extensive tailing of both bromide and colloids caused by high mass transfer and Column 27 exhibits only little tailing of both components corresponding to low mass transfer. Breakthrough data for the remaining columns are shown in Fig. 3. In all cases it was possible to get a close model fit to the measured breakthrough data for both bromide and colloids.
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 2. Eight general types of bromide breakthrough and natural colloid leaching curves observed for the 20-cm-high and 20-cm-diameter intact soil columns.
|
|
View larger version (41K):
[in this window]
[in a new window]
|
Fig. 3. Bromide and colloid breakthrough curves for the remaining 25 intact soil columns used in this study.
|
|
Mobile Water Contents
The relationship between mobile water content, m, and the amount of water-filled pores with diameter larger than 150 µm is shown in Fig. 4a
. The amount of water in pores with diameters >150 µm (macro) was determined as the air-filled porosity at a soil water potential of –20 cm H2O (
20) minus the air-filled porosity in the columns during the experiment (exp). For most columns the estimated mobile water content (m) and macro were very similar, indicating that the water flow in these columns mainly occurred in pores with a diameter >150 µm. In a few cases m is higher than macro, indicating that the flow also occurred in the smaller pores. The latter columns are also characterized by the lowest bromide transport velocities (Fig. 4b and 4c).
Advective Bromide and Colloid Transport Velocity
Bromide transport velocity vBr was observed to be proportional to soil bulk density (Fig. 4b) and to clay content since bulk density and clay content were closely correlated. It is very likely that this is caused by the presence of larger cracks in the more dense clayey columns where water and bromide are then transported rapidly. Figure 4c shows only a weak correlation between colloid transport velocity (vcoll) and bromide transport velocity (vBr), and, as discussed earlier, vcoll was always larger than vBr, probably due to size exclusion of colloids. Thus, the main transport pathways for dissolved (bromide) and suspended (colloidal) matter likely are different, in agreement with the discussion by White (1985).
Spatial Distribution of Soil Properties
Distributions of the cumulative amount of leached particles, content of WDC, soil bulk density, and equivalent pore diameter (dpore) of the large pores (taken as the pores that are air-filled at a soil water potential of –20 cm H2O) are shown in Fig. 5a
through 5d, respectively. The equivalent pore diameter is calculated as (Moldrup et al. (2001), combining their Eq. [5] and [11]):
 | [8] |
where 
is total soil porosity and ka20 is soil air permeability at a soil water potential of –20 cm H2O. Both cumulative amounts of leached particles and soil bulk density exhibit larger values in the southern end of the field. The content of WDC and the equivalent diameter of the large pores both vary somewhat more randomly across the field. This indicates that the content of WDC is not strongly correlated with soil physical properties.
Spatial Distribution of Bromide and Colloid Transport Parameters
Distributions of bromide and colloid transport velocities, mobile–immobile water phase mass transfer coefficients, and colloid retardation factors are shown in Fig. 5e through 5i. Bromide transport velocity, vBr (Fig. 5e), is proportional to soil bulk density (and, thus, to clay content) while bromide immobile–mobile mass transfer coefficient, Br (Fig. 5f), is inversely proportional to these two parameters. Because the clayey region does not generally have larger macropores (Fig. 5d), the explanation is likely that the mass transfer of bromide between the mobile and immobile phases is controlled by diffusion within the immobile water. Because the immobile water tortuosity will likely be larger in the more clayey soil, diffusion, and thus mass transfer, is slower. Colloid transport velocity vcoll (Fig. 5g) and colloid mass transfer coefficient coll (Fig. 5h) both have a tendency for being higher in the southern end of the field. This means that, even if colloids are transported rapidly, mass exchange with the soil matrix still occurs, likely because of increased water shear. Colloid retardation factor (Fig. 5g) as well as colloid and bromide dispersivity, (not shown), did not exhibit any clear relationships with basic soil physical properties. It is noted that because of the three missing columns in the lower left corner of the sampling field (Fig. 1) parameter distributions in this area are likely not correct.
Factors Controlling Colloid Mobilization and Transport
By comparing the two local areas indicated in Fig. 5 (labeled Areas I and II, each covering 25% of the sampling field) it is seen that while the cumulative amount of leached particles is low in Area I and high in Area II, the opposite appears to be the case for WDC. Furthermore, the amount of leached particles does not correlate well with either bulk density or clay content, as perhaps would be expected because soils with high clay contents contain more small particles. Thus, soil containing high concentrations of water dispersible colloidal particles does not necessarily exhibit high colloid leaching. The amount of leached particles in the two areas, however, correlates well with vColl and to some degree also with Coll, implying that colloid leaching is high when a high water flow velocity in larger pores is combined with good mass transfer conditions. The amount of leached particles does to some degree correlate with the equivalent pore diameter, again implying that colloid transport mainly takes place in the very large pores. These results show that the soil structure and the related transport parameters were the most important factor for controlling colloid leaching, whereas soil texture and the amount of water-dispersible colloids present in the soil were less important. Colloid retardation factors, Rcoll, are low in both Areas I and II, meaning that size exclusion is pronounced in both areas even though most of the other parameters investigated are different for the two areas. Hence, retardation of colloidal particles cannot be related in any simple way to soil structure and transport characteristics.
|
CONCLUSIONS
|
|---|
A two-region mobile–immobile water transport model was well fitted bromide breakthrough and colloid leaching curves and was used to estimate bromide and colloid transport parameters in undisturbed soil columns.
Colloid transport was faster than bromide for all columns. In general, colloid transport velocities were three to four times larger than those for bromide, with a total range of 1.3 to 21, implying that colloids were transported mainly in large pores and channels (size exclusion). For most columns MIM model mobile water content was close to the amount of water-filled pores with diameter >150 µm, suggesting that water flow in these columns mainly occurred in macropores. Bromide transport velocity vBr was proportional to soil bulk density, likely due to the presence of larger macropores/cracks in the more dense clayey soil columns.
The distributions of soil physical properties and estimated bromide and colloid transport parameters across the sampling field were highly variable. Comparison of the distributions of soil physical properties, transport parameters, and leached quantity of colloids suggests that the most important mechanism controlling colloid leaching is soil structure and related transport parameters, whereas soil texture and the amount of water dispersible colloids are less important.
|
ACKNOWLEDGMENTS
|
|---|
We acknowledge the Japan Society for Promotion of Science (JSPS) Grant-in-Aid for Scientific Research no. 14550545, and a research grant from the Mazda Foundation.
|
REFERENCES
|
|---|
- de Jonge, L.W., C. Kjaergaard, and P. Moldrup. 2004a. Colloids and colloid-facilitated transport of contaminants in soils: An introduction. Available at www.vadosezonejournal.org. Vadose Zone J. 3:321–325.[Abstract/Free Full Text]
- de Jonge, L.W., P. Moldrup, G.H. Rubæk, K. Schelde, and J. Djurhuus. 2004b. Particle leaching and particle facilitated transport of phosphorous at the field scale. Available at www.vadosezonejournal.org. Vadose Zone J. 3:462–470.[Abstract/Free Full Text]
- Gaber, H.M., W.P. Inskeep, S.D. Comfort, and J.M. Wraith. 1995. Nonequilibrium transport of atrazine through large intact soil cores. Soil Sci. Soc. Am. J. 59:60–67.[Abstract/Free Full Text]
- Jacobsen, O.H., P. Moldrup, C. Larsen, L. Konnerup, and L.W. Petersen. 1997. Particle transport in macropores of undisturbed soil columns. J. Hydrol. 196:185–203.[CrossRef]
- Kjaergaard, C., T.G. Poulsen, P. Moldrup, and L.W. de Jonge. 2004a. Colloid mobilization and transport in undisturbed soil columns. I. Pore structure characterization and tritium transport. Available at www.vadosezonejournal.org. Vadose Zone J. 3:413–423.[Abstract/Free Full Text]
- Kjaergaard, C., L.W. de Jonge, P. Moldrup, and P. Schjønning. 2004b. Water-dispersible colloids: Effects of measurement method, clay content, initial soil matric potential and wetting rate. Available at www.vadosezonejournal.org. Vadose Zone J. 3:403–412.[Abstract/Free Full Text]
- Kretzschmar, R., M. Borkovec, D. Grolimund, and M. Elimelech. 1999. Mobile subsurface colloids and their role in contaminant transport. Adv. Agron. 66:121–193.
- Laegdsmand, M., L.W. de Jonge, P. Moldrup, and K. Keiding. 2004. Pyrene sorption to water-dispersible colloids: Effect of solution chemistry and organic matter. Available at www.vadosezonejournal.org. Vadose Zone J. 3:451–461.[Abstract/Free Full Text]
- Langner, H.W., H.M. Gaber, J.M. Wraith, B. Huwe, and W.P. Inskeep. 1999. Preferential flow through intact soil cores: Effects of matric head. Soil Sci. Soc. Am. J. 63:1591–1598.[Abstract/Free Full Text]
- McCarthy, J.F., and L.D. McKay. 2004. Colloid transport in the subsurface: Past, present and future challenges. Available at www.vadosezonejournal.org. Vadose Zone J. 3:326–337.[Abstract/Free Full Text]
- Moldrup, P., T. Olesen, T. Komatsu, P. Schjønning, and D.E. Rolston. 2001. Tortuosity, diffusivity, and permeability in the soil liquid and gaseous phases. Soil Sci. Soc. Am. J. 65:613–623.[Abstract/Free Full Text]
- Moldrup, P., T. Yamaguchi, D.E. Rolston, K. Vestergaard, and J. Aa. Hansen. 1994. Removing numerically induced dispersion from finite difference models for solute and water transport in undisturbed soils. Soil Sci. 157:153–161.
- Morizawa, S., M. Horiuchi, T. Yamaoka, and T. Inoue. 1986. Evaluation of solute transport soil column by multi-component flow model. (In Japanese.) Proc. Environ. Sanit. Rng. Res. 22:9–22.
- Motoshita, M., T. Komatsu, P. Moldrup, N. Ozaki, and T. Fukushima. 2001. Soil colloid mobilization and transport in intact and repacked soil columns during periods of irrigation. JSCE J. Environ. Syst. Eng. 678(7–19):111–122.
- Schelde, K., P. Moldrup, O.H. Jacobsen, H. de Jonge, L.W. de Jonge, and T. Komatsu. 2002. Diffusion limited mobilization and transport of natural colloids in macroporous soil. Available at www.vadosezonejournal.org. Vadose Zone J. 1:125–136.[Abstract/Free Full Text]
- van Genuchten, M.Th., and P.J. Wierenga. 1977. Mass transfer studies in sorbing porous media: II. Experimental evaluation with tritium (3H2O). Soil Sci. Soc. Am. J. 41:272–278.[Abstract/Free Full Text]
- White, D.E. 1985. The influence of macropores on the transport of dissolved and suspended matter through soil. Adv. Soil Sci. 3:95–120.
This article has been cited by other articles:
|
|
|
|
 
T. Harter, E. R. Atwill, L. Hou, B. M. Karle, and K. W. Tate
Developing Risk Models of Cryptosporidium Transport in Soils from Vegetated, Tilted Soilbox Experiments
J. Environ. Qual.,
January 4, 2008;
37(1):
245 - 258.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
