| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Environmental Engineering Section, Dep. of Life Sciences, Aalborg University, Sohngaardsholmsvej 57, DK-9000 Aalborg, Denmark
b Dep. of Agroecology, Danish Institute of Agricultural Sciences, P.O. Box 50, DK-8830 Tjele, Denmark
c Currently Danish Institute of Agricultural Sciences, Department of Agroecology, P.O. Box 50, DK-8830 Tjele, Denmark
* Corresponding author (C.Kjaergaard{at}agrsci.dk).
Received 3 July 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
= –2.5 (wet), –100 (moderately wet), and –15500 hPa (dry) followed by 5 d of irrigation at 1 mm h–1 and applying a suction of 5 hPa at the lower boundary. The results revealed that (i) colloid leaching from the initially wet and moderately wet soils decreased with increasing clay content, (ii) colloid leaching from the initially dry soils was low and independent of clay content, and (iii) the leaching of total organic C (TOC) consisted mainly of dissolved organic C (DOC), and drying to –15500 hPa increased the leaching of C. In situ colloid mobilization and leaching was related to measurements of low-energy water-dispersible colloids (LE-WDC). Results indicate that in situ colloid mobilization from the initially wet and moderately wet 12% clay soils subjected to matrix-dominated flow behavior was controlled mainly by the time-dependent increase in colloid dispersion, while colloid mobilization from the initially dry soils was limited by the strong and persistent association created during the drying. In the more clayey soils, which were dominated by preferential flow, a lower displacement of high–ionic strength soil water with low–ionic strength rainwater may contribute to the inherently lower dispersibility in controlling colloid mobilization.
Abbreviations: BTC, breakthrough curve • DOC, dissolved organic C • EC, electrical conductivity • IMP, initial matric potential • LE-WDC, low-energy water-dispersible colloids • NTU, nephelometric turbidity units • POC, particulate organic C • SAR, sodium adsorption ratio • TOC, total organic C • WDC, water-dispersible colloids
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Natural structured soils differ from the homogeneous systems in two important aspects. First, the complex association of colloids in aggregates of varying size and stability determine the inherent ability of colloids to disperse in response to infiltration of low–ionic strength rainwater. Second, pore structure has a profound effect on the active flow volume of the infiltrating water, affecting both in situ colloid mobilization and the subsequent transport of mobilized colloids. Dispersion is the ultimate state of aggregate breakdown that results in release of colloids as a consequence of expanding double layers and dominating repulsive forces as described by the DLVO theory (Derjaguin and Landau, 1948; Verwey and Overbeek, 1948). The fraction of colloids that disperses in water (water-dispersible colloids, WDC) has been used as an input parameter for predicting colloid leaching and colloid-facilitated transport through the vadose zone (Jarvis et al., 1999; Villholth et al., 2000). The importance of clay mineralogy (e.g., Frenkel et al., 1978; Seta and Karathanasis, 1996) and solution chemistry (e.g., Rengasamy, 1983) to colloid dispersibility are well documented. In addition, total clay content (e.g., Pojasok and Kay, 1990; Brubaker et al., 1992; Kjaergaard et al., 2004a); soil matric potential (e.g., Pojasok and Kay, 1990; Kjaergaard et al., 2004a); and management in terms of crop sequence, application of organic manures, soil tillage, and field traffic (e.g., Watts et al., 1996a, 1996b; Schjønning et al., 2002) affect the dispersibility of colloids from aggregates.
It is well known that structural macropores may provide preferential pathways for water flow and suspended colloids (e.g., Smith et al., 1985; Camobrecco et al., 1996), but only few studies have investigated the actual in situ colloid mobilization and transport in undisturbed soil columns (Jacobsen et al., 1997; de Jonge et al., 1998, 2000; Lægdsmand et al., 1999; Schelde et al., 2002; de Jonge et al., 2004) or in field studies (Ryan et al., 1998; El-Farhan et al., 2000; Villholth et al., 2000; Petersen et al., 2003). These studies have verified the importance of solution ionic strength and pH in colloid mobilization in structured soils (de Jonge et al., 1998), while the importance of physical perturbations is less clear. Several studies have evaluated the role of colloid mobilization by hydraulic shear following rapid infiltration of water (Jacobsen et al., 1997; Ryan et al., 1998; Schelde et al., 2002; Petersen et al., 2003); these studies all indicated that in situ colloid mobilization in structured soils is not governed by shear stress. The importance of flow transients was demonstrated by El-Farhan et al. (2000), who observed that all peak particle concentrations occurred near either the rising or falling limb of the water flux hydrograph. The authors attributed this to the movement of air–water interfaces during imbibition and drainage, but Schelde et al. (2002) concluded in their study that this was not a dominant process in controlling colloid mobilization. Jacobsen et al. (1997) suggested that the amount of mobilized particles was limited by colloid diffusion, and Schelde et al. (2002) demonstrated that colloid mobilization in structured soils was a time-dependent and possibly a diffusion-limited process. Most of the studies performed on undisturbed soil columns identified the macropores as the source of in situ mobilized colloids, but the lack of knowledge concerning the diffusive displacement of high–ionic strength resident water with low–ionic strength rainwater between immobile and mobile water regions does not elucidate the actual role of preferential flow on in situ colloid mobilization.
The primary objective of this study was to further evaluate the processes controlling in situ colloid mobilization in structured soils. In structured soils inter- and intraaggregate porosity affects the accessibility of the colloids to the infiltrating water and ion diffusion. On the basis of the conceptual perception that most colloids are associated in aggregates, we hypothesize that in situ mobilization of colloids may be controlled by (i) the diffusive displacement of high–ionic strength intraaggregate water with low–ionic strength infiltration water (Fig. 1a) , (ii) the actual dispersibility of the colloids (Fig. 1b), and (iii) the diffusion of colloids from the immobile intraaggregate water to the mobile convective water (Fig. 1c). In this study we investigated colloid mobilization and transport in undisturbed soil columns exposed to a low and frequently occurring irrigation intensity (and with a weak suction applied to) the lower boundary of the soil column, simulating a situation of water infiltration through the upper unsaturated soil horizon where the macropores are not necessarily fully functional. We focus on total clay content and IMP as key parameters affecting in situ colloid mobilization and transport because of the influence of these parameters on colloid dispersibility, shrinkage–swelling phenomena, the active flow volume, as well as their role in promoting transient flow conditions along wetting fronts when water infiltrates dry soils.
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
= –2.5, –100, and –15500 hPa), ranging from near saturation to the crop wilting point. The experiment was performed with three replicates of each combination. The pretreatment procedure involved (i) saturation of the soil columns by slow capillary wetting on tension tables with electrolyte solution having a chemical composition similar to natural rainwater, (ii) drainage to –2.5 and –100 hPa on tension tables and drying to –15500 hPa by passing through dry air, and (iii) incubation for 14 d at 10°C.
Column Leaching Experiment and Analysis of Colloid Suspensions
The soil column setup and details of the experimental procedure are explained and illustrated in the companion paper Kjaergaard et al. (2004b). In short, the leaching experiment involved irrigation at 1 mm h–1 with an electrolyte solution of a chemical composition similar to natural rainwater and sampling of effluent at a suction of 5 hPa applied to the lower boundary of the soil. A cloth was placed on the top of the soil core to avoid erosion by raindrop impact. Effluent fractions of 15 mL were collected continuously, weighed, and analyzed for colloid concentration, electric conductivity (EC), pH, and TOC. Colloid concentration was determined by turbidity using a Hach 2100AN turbidimeter (Hach, Loveland, CO). Turbidity values (expressed in nephelometric turbidity units, NTU) were converted to colloid concentration using regression equations based on suspensions of LE-WDC separated from each clay soil. The procedure for separating LE-WDC is described by Kjaergaard et al. (2004a). Calibration curves were made from stock LE-WDC suspension for each clay soil. A linear correlation existed between colloid concentration and turbidity at turbidity values 
100 NTU, while nonlinearity was observed for all clay soils at colloid concentrations >100 NTU. Electric conductivity was measured directly in all samples, and suspension pH was measured on every 10th sample. Total organic C was measured in all samples by C combustion using a Total Organic Carbon Analyzer (TOC-5000A, Shimadzu Scientific, Kyoto, Japan) equipped with a suspended particle kit and coupled to an auto sampler with magnetic stirring. The fraction of leached particulate organic C (POCleached) was estimated from the ratio of POC/LE-WDC measured by Kjaergaard et al. (2004a), assuming that the fraction of colloids leached (Colleached) resembles the fraction of LE-WDC:
 | [1] |
The fraction of DOC was calculated as the difference between TOC and POC. Leached colloid concentration, TOC, and POC were plotted against number of eluted pore volumes (V/V0) where V is the outflow volume (m3), and V0 is the water-filled porosity (m3 m–3) at –5 hPa.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
30 µm was initially drained, and the soils were allowed to moisten by irrigating with rainwater (Fig. 2b and 2e). This implies that when leaching started from these soils, the resident soil water in the pore fraction 30 µm had already been displaced by more dilute low–ionic strength rainwater compared with the soils at IMP –2.5 hPa. Effluent was collected after 2 to 6 h of irrigation. When a 5-hPa suction was applied to the bottom of the soil columns, a steep increase in effluent flow rate was observed, reflecting the drainage of pores with an equivalent diameter >600 µm. After application of suction, the effluent flow rate rapidly stabilized in the range ±0.5mL h–1 from the mean. The rapid stabilization of the effluent flow rate for the soils incubated at IMP –2.5 and –100 hPa indicates that no further wetting of the soil columns took place after application of suction.
|
0.2 µm was initially drained, and 8 to 12 h of irrigation was applied to the heavy clay soil (43% clay), compared with 20 to 22 h of irrigation to the 12% clay soils, before effluent was collected. When suction was applied to the 12% clay soils, there was a small increase in effluent flow rate, followed by a moderate, rapid buildup of the flow rate to a constant outflow, except for one of the replicates that had a markedly lower effluent flow. No increase in effluent flow rate was observed when suction was applied to the high clay soil (43% clay), and the flow rate increased steadily during the next 21 to 25 h until a constant outflow occurred (Fig. 2f).
A thorough analysis of the active flow volume in these soil columns in terms of tritium (3H2O) breakthrough experiments was presented by Kjaergaard et al. (2004b). Here it was found that the 12% clay soils exhibited matrix-dominated flow behavior, while soils with 18% clay exhibited an increasing degree of preferential flow with increasing clay content. This is illustrated by the number of pore volumes eluted when 12.5% of the applied tritium has been leached (Fig. 3)
. Drying to –15500 hPa and rewetting significantly reduced the degree of preferential flow for all soils (Fig. 3) and resulted in matrix-dominated flow behavior at 18 and 24% clay. No differences in the breakthrough curves (BTCs) were found between soils with IMP –2.5 and –100 hPa.
|
24% clay) colloid leaching generally displayed similar behavior among clay contents and the different IMPs; effluent concentrations decreased rapidly to constant low values below 20 mg L–1 within the first pore volume of leaching (Fig. 4g–4l).
|
24% clay) high initial concentrations (30–70 mg L–1) were generally observed in all soils, declining to fairly constant values (15–40 mg L–1) within one pore volume of leaching (Fig. 5g–5l). The estimated amount of POC clearly indicated that DOC constituted the main fraction of C leached from all soils irrespective of initial moisture conditions, with POC contributing to <10% of the total amount of C leached. These results are in agreement with results of Lægdsmand et al. (1999), who demonstrated that 83 to 99% of all TOC leached from an undisturbed sandy loam soil is present as DOC.
|
24% clay.
|
24% clay) compared with higher clay contents. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
In the present study the high initial flush of colloids followed by a rapid decline to a constant low level observed mainly at clay contents 24% clay (Fig. 4g–4l) coincidences with the application of suction and the simultaneous drainage of pores with an equivalent diameter >600 µm. Thus, it is highly likely that the initial flush of colloids we observed at the beginning of the leaching experiment, in agreement with other studies (Jacobsen et al., 1997; Ryan et al., 1998; Schelde et al., 2002), resulted from the buildup of colloid concentration in the mobile region during nonflow periods allowing time for colloid diffusion from immobile to mobile water as suggested by Schelde et al. (2002). The combination of a high degree of preferential flow (i.e., a large fraction of immobile water) and the rapid decline to constant low concentrations at clay contents 24% clay, may indicate that the supply of colloids was limited at least partly by either a low-diffusive displacement of high–ionic strength soil water with low–ionic strength rainwater impeding colloid dispersion, or by colloid diffusion from immobile to mobile water. In contrast, the matrix-dominated flow behavior at 12% clay and a minor degree of preferential flow at 18% clay combined with the increase in colloid leaching, at least during part of the experiment, suggest that the larger active flow volume results in a larger source area of colloids, which may partly explain the higher colloid leaching in these soils.
High initial flush of colloids has also been attributed to flushing of colloids sorbed to moving air–water interface (e.g., El-Farhan et al., 2000). El-Farhan et al. (2000) attributed peaks in particle concentrations occurring near the rising or falling limb of the water flux hydrograph to the movement of air–water interfaces during imbibition and drainage. Not considered by these authors was that changes in pore water velocity during imbibition and drainage have changed the active flow volume and thereby the source area of colloids. At present no studies on undisturbed structured soils have documented the actual role of moving air–water interfaces. Schelde et al. (2002) concluded that even if air–water interfaces contributed to the initial flush of colloids, this was not a dominant process in colloid mobilization compared with the time-dependent and rate-limiting processes of colloid release. In this study, the observed decline in colloid concentration coinciding with the rewetting of the initially dry soils at 12 and 18% clay does not agree with the suggestion that transient flow regimes increase colloid mobilization.
The observed leaching patterns with either increasing of declining colloid concentrations indicate the role of the active flow volume on in situ colloid mobilization, but it was not possible to explain the differences in colloid leaching behavior observed for the replicate soil columns at 12% clay, nor to explain the differences in colloid leaching between different IMPs at 12% clay from the knowledge of the active flow volume in these soils (Kjaergaard et al., 2004b). The differentiation of replicate soil columns with respect to leaching pattern, and the differences among soils with different IMPs suggest differences in aggregate stability and colloid dispersibility.
Role of Colloid Dispersibility
Investigation of colloid dispersibility (Kjaergaard et al., 2004a) demonstrated that the amount of LE-WDC decreased with increasing clay content as a result of increased aggregate stability and decreased with decreasing IMP as a consequence of increased interparticle bonding or cementation upon drainage and drying (Fig. 6c). When drying was severe (IMP –15500 hPa), colloid dispersibility was low and independent of clay content, and the response of colloid dispersibility to drainage and drying persisted even after 1 wk of slow rewetting. These results may at least partly explain the observed decline in colloid leaching as clay content increases from the initially wet and moderately wet soils, and the clay-independent leaching after drying to –15500 hPa (Fig. 6a). Several possible mechanisms of colloid mobilization may explain the initial high leaching of colloids when rewetting the dry soils (Fig. 4). First, when dry aggregates are rewetted, they may break down by a combination of physicochemical dispersion and slaking, which is disintegration caused by compression of entrapped air during wetting (Le Bissonnais, 1996). The results from Kjaergaard et al. (2004a) documented that the amount of LE-WDC from these soils significantly increased as a result of aggregate breakdown by slaking upon fast rewetting. Second, it is possible that the initial high leaching of DOC (Fig. 5) may facilitate the mobilization of mineral colloids. The rapid depletion of mobile colloids indicates, however, that the drying has introduced very strong associations among colloids, and these associations are not easily broken even after prolonged leaching of low–ionic strength rainwater.
The measurements of LE-WDC documented that colloid dispersibility decreased as the IMP decreased (Fig. 6c), but the actual mobilization and leaching of colloids from the undisturbed soil columns was largest from the initially moderately wet soils (Fig. 4 and 6a). The explanation for this deviation is differences in the pretreatment procedure between the two types of experiments. As explained earlier in this paper, when leaching started from the initially moderately wet undisturbed soil columns the soil water in the pore fraction 30 µm was already displaced with more dilute low–ionic strength rainwater compared with the initially wet soils. This implied that the initially moderately wet soils were already more dispersed, explaining the higher colloid leaching from these soils (Fig. 4b and 4e). When measuring LE-WDC the soil samples incubated at –2.5 hPa were shortly drained to –100 hPa before measurement and subsequently resaturated with low–ionic strength water, to have similar treatments with respect to changes in electrolyte composition for the samples at –2.5 and –100 hPa (Kjaergaard et al., 2004a). This means that the measurements of LE-WDC reflect the sole effect of initial soil water content on colloid dispersion, while differences in colloid leaching between initially wet and moderately wet samples are influenced by differences in the initial pore water ionic strength.
Measurements of LE-WDC (Fig. 6c) also demonstrated an effect of IMP on colloid dispersibility for all clay soils. The accumulated colloid leaching (Fig. 6a), however, showed no effect of IMP at higher clay contents (24%). This is probably a consequence of pores >30 µm constituting a decreasing fraction of the pore volume with increasing clay content (Kjaergaard et al., 2004b), which consequently minimizes the effect of drainage of this pore fraction.
Mechanisms for In Situ Colloid Mobilization
To evaluate the time dependency of colloid mobilization, the cumulative mass of colloids has been plotted against the square root of time for the different IMPs at 12, 18, and 28% clay (Fig. 7)
. In addition, the time-dependent changes in the electric conductivity of the effluent water are illustrated. Generally, the curves for accumulated colloid mass leached seemed nonlinear within the first part but attained linearity after some time, indicating that colloid mobilization after some leaching is rate limited. Several authors demonstrated that colloid leaching in structured soils is time dependent and possibly limited by diffusion (Jacobsen et al., 1997; Lægdsmand et al., 1999; Schelde et al., 2002). However, dispersion of colloids from aggregates may also be time dependent in a dynamic system, where the ionic strength of the intraaggregate soil water gradually decreases due to continued infiltration and displacement of high–ionic strength soil water with low–ionic strength rainwater.
|
We find that straining was not likely to be the cause of the reduced leaching of colloids from the higher clay soils. First, measurements of LE-WDC showed that the inherent dispersibility of these soils was already very low (Fig. 6c). Second, the results from the 3H2O BTCs indicated that only a minor fraction of large pores contributed to the active flow volume of these soils, minimizing the possibility of significant straining (Kjaergaard et al., 2004b). In contrast, it is important to note that we observed the largest colloid leaching from soils having matrix-dominated flow behavior. This suggests that colloid transport is not exclusively related to large structural macropores. The role of preferential flow deduced from this study may also explain the generally observed colloid leaching behavior from structured soils (e.g., Jacobsen et al., 1997; Ryan et al., 1998; Schelde et al., 2002). In these studies preferential flow was promoted by zero-tension lower boundaries and high irrigation intensities promoting a high degree of preferential flow, which can result in the macropores being the primary source of mobile colloids and lead to a rapid depletion in effluent colloid concentration. During declining flow intensities (e.g., El-Farhan et al., 2000) or nonflow periods (e.g., Schelde et al., 2002), the rate of ion displacement among immobile and mobile water regions would increase, resulting in supply or replenishment of mobile colloids.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
24% clay) or stronger colloid associations upon drying (IMP –15500 hPa for all soils). Finally, colloid dispersibility should be considered a key parameter when developing predictive models for in situ colloid mobilization and colloid-facilitated transport in the vadose zone. The agreements between the LE-WDC estimates and the actual mobilization of colloids indicates that measurements of LE-WDC may be a valuable tool for risk assessments with respect to colloid mobilization.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. M. Levin, J. S. Herman, G. M. Hornberger, and J. E. Saiers Colloid Mobilization from a Variably Saturated, Intact Soil Core Vadose Zone J., April 27, 2006; 5(2): 564 - 569. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Schelde, L. W. de Jonge, C. Kjaergaard, M. Laegdsmand, and G. H. Rubaek Effects of Manure Application and Plowing on Transport of Colloids and Phosphorus to Tile Drains Vadose Zone J., March 8, 2006; 5(1): 445 - 458. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Kjaergaard, H. C. B. Hansen, C. B. Koch, and K. G. Villholth Properties of Water-Dispersible Colloids from Macropore Deposits and Bulk Horizons of an Agrudalf Soil Sci. Soc. Am. J., November 1, 2004; 68(6): 1844 - 1852. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. W. de Jonge, C. Kjaergaard, and P. Moldrup Colloids and Colloid-Facilitated Transport of Contaminants in Soils: An Introduction Vadose Zone J., May 1, 2004; 3(2): 321 - 325. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Kjaergaard, T. G. Poulsen, P. Moldrup, and L. W. de Jonge Colloid Mobilization and Transport in Undisturbed Soil Columns. I. Pore Structure Characterization and Tritium Transport Vadose Zone J., May 1, 2004; 3(2): 413 - 423. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Laegdsmand, L. W. de Jonge, P. Moldrup, and K. Keiding Pyrene Sorption to Water-Dispersible Colloids: Effect of Solution Chemistry and Organic Matter Vadose Zone J., May 1, 2004; 3(2): 451 - 461. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. W. de Jonge, P. Moldrup, G. H. Rubaek, K. Schelde, and J. Djurhuus Particle Leaching and Particle-Facilitated Transport of Phosphorus at Field Scale Vadose Zone J., May 1, 2004; 3(2): 462 - 470. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME |
