SSSAJ Grow Your Career with SSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 1 January 2007
Published in Soil Sci Soc Am J 71:35-42 (2007)
DOI: 10.2136/sssaj2006.0106
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
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 ISI 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 HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Mertens, J.
Right arrow Articles by Vanderborght, J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Mertens, J.
Right arrow Articles by Vanderborght, J.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Mertens, J.
Right arrow Articles by Vanderborght, J.
Related Collections
Right arrow Vadose Zone Processes and Chemical Transport
Right arrow Water Flow Models
Right arrow Experiment Design

SOIL PHYSICS

Numerical Analysis of Passive Capillary Wick Samplers prior to Field Installation

Jan Mertens*, Jan Diels and Jan Feyen

Soil and Water Management, Katholiek Universiteit Leuven, Celestijnenlaan 200E, B-3001 Heverlee, Belgium

Jan Vanderborght

Agrosphere, ICG-IV, Forschungszentrum Jülich GmbH, D-52425 Jülich, Germany

* Corresponding author (mertensja{at}yahoo.co.nz)


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSIONS
 REFERENCES
 
Accurately measuring water fluxes and associated nutrient or contaminant concentrations through the vadose zone is difficult because an appropriate suction needs to be exerted on the soil to sample water under unsaturated conditions. Passive capillary wick sampling systems are cheap and reliable instruments resulting in acceptable measurements of water fluxes in the vadose zone; however, their success in measuring realistic fluxes depends on their compatibility with the soil and climatic conditions in which they are installed. This study was developed in the preplanning phase of a field experiment with its main objective the monitoring of dissolved organic matter and the associated transfer of Cu2+ and pesticides through the vadose zone. We studied a combination of two-dimensional and axisymmetrical three-dimensional numerical analyses using the HYDRUS-2D software to identify what sampler geometry, wick type, wick length, and number of wicks are most suitable for the soil conditions at the experimental site. An AM3/8HI wick with seasonally varying wick length (40 cm in winter and 100 cm in summer) was found to be most appropriate for the soil and climatic conditions of the experimental field. The numerical analysis indicated that well-designed wick samplers had a negligible effect on the soil moisture content close to the sampler. A double-ring wick sampler is proposed to minimize the effect of the area between the installation pit or trench and the sampler. This approach is easily applicable and transferable to other soil and wick types and climatic conditions. The study emphasizes the suitability of numerical modeling to optimize experimental design before installation.


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSIONS
 REFERENCES
 
Monitoring and measuring techniques that can determine drainage flux and its solute concentrations from undisturbed soil profiles are critical for the understanding of the migration of solutes and selection of possible remediation strategies (Masarik et al., 2004). Leachate sampling in unsaturated conditions remains a challenge since a suction needs to be exerted to sample the water. Traditionally a vacuum source on ceramic cups has been used to sample leachate from within the vadose zone. The derived fluxes are erratic because of the small sampling surface, the uncertainty of the exact soil volume sampled, and the fact that ceramic cups seldom intercept the critical solute plume (Brandi-Dohrn et al., 1996; Weihermuller et al., 2005). Most simple larger samplers are known as pan or zero-tension devices. These devices usually consist of a pan shape filled with some coarse material that intercepts the leachate. For water to flow into the pan, however, a zero-tension boundary condition (seepage face) is required. In soils finer than the material in the pan, this induces divergence away from the sampler and results in an underestimation of the natural drainage flux. In a numerical evaluation by Flury et al. (1999), it was shown that the saturated zone above the sampler not only had a large effect on the sampled flux but also influenced the chemical concentration of the sampled leachate. A more frequent sampling system is the constant-tension lysimeter, with the applied suction usually somewhere between –100 and –300 cm of water. Again, soil water conditions close to the sampler are a function of the applied tension. This may result in differences between the sampled leachate flux and its solute concentration and the undisturbed flux and concentration. Recently, equilibrium tension systems have been developed (Brye et al., 1999, 2001; Siemens et al., 2003; Foley et al., 2003; Pelger et al., 2003; Barzegar et al., 2004; Kosugi and Katsuyama, 2004; Masarik et al., 2004). These systems measure soil tension in the undisturbed soil and match this tension to the vacuum supplied to the sampler. This control strategy results in the rate of water extraction and solute concentration sampled by the lysimeter to be representative of the unsaturated water flux in the undisturbed soil profile at the same depth. For further information on setup and control of these equilibrium tension plate lysimeters (ETPLs), see Masarik et al. (2004). Initial results are very promising and we believe these ETPL systems will be increasingly used. The only, but at the same time significant, disadvantage of these ETPL systems is the very high cost associated with their implementation, operation, and maintenance.

The material here was developed in the preplanning phase of a field experiment proposed in a project investigating the role of dissolved organic matter (DOM) on the leaching of Cu2+ and pesticides in Belgian soils. Given that the project aim is the assessment of the effect of different land use treatments on DOM fluxes and related Cu2+ and pesticide migration rather than the exact estimation of leachate fluxes and solute concentrations, a simple, robust, and cheap sampling device would be adequate. Therefore we believed that wick samplers using a fiberglass wick, maintaining a fixed tension, would be the most appropriate and cost-effective device for collecting the leachate. According to Holder et al. (1991), Boll et al. (1992), Knutson and Selker (1994), and Rimmer et al. (1995), the degree of suction exerted by the wick depends on the wick type, the flux rate, and the soil type.

The main advantages of the wick samplers are that they do not require any kind of suction equipment, they intercept flow from a known volume of soil, and their low cost and ease of installation and maintenance (Knutson and Selker, 1994). Zhu et al. (2002) found the leachate collection efficiencies of wick samplers during a 4-yr period to be 101%, compared with only 40% for pan lysimeters. Gee et al. (2002, 2003) minimized the effect of the constant suction by using confined lysimeter systems that had side walls extending above the sampler. The soil volume confined by the sidewalls reduced the influence of the applied constant tension and therefore on the water flow through the top boundary of the sampler (Lentz and Kincaid, 2003); however, sidewalls introduce the risk of saturation and preferential flow along the walls and complicate the installation of the samplers. Therefore, sidewalls are not further discussed here. Vandervelde et al. (2005) found that water-balance and HYDRUS-1D model estimates of drainage corresponded well with the measurement by nonsuction water flow meters while suction water flux meters overestimated drainage. The latter was probably due to flow convergence induced by the wick and divergence barrier lengths being not properly sized for the flow conditions. This study emphasizes the importance of soil and flow conditions being in harmony with the wick type, length, and diameter. During the design phase of the field experiment, the following four key questions arose:

The questions were addressed using two-dimensional as well as axisymmetrical three-dimensional HYDRUS-2D numerical simulations. Only rarely are studies presented that report on the use of modeling tools to aid in the design and setup of laboratory and field experiments. We believe, however, that a numerical analysis before the setup and installation of experiments is very informative and provides valuable input for the design of experiments. Since time is generally limited before field installation, we wanted to use an easy and relatively fast tool to help identify possible artifacts in the experimental setup. Gee et al. (2002, 2004) successfully applied the STOMP 2D model (White et al., 1995) to identify the importance of different design parameters and boundary conditions on the leachate sampled by a water fluxmeter with divergence control. Abdou and Flury (2004) used the CHAIN_2D code (Simunek and van Genuchten, 1994) to investigate the effect of the bottom boundary condition of lysimeters on water flow and solute transport. Mertens et al. (2005) recently applied the HYDRUS-2D model (Simunek et al., 1999) to evaluate the effects of the design and installation of ETPLs on the leachate volume. To analyze the performance of passive capillary wick samplers before their installation, a similar HYDRUS-2D model setup to the one described in Mertens et al. (2005) was used in this study; however, this study implemented a vertical axisymmetrical three-dimensional model to account for the three-dimensional flow pattern around a circular wick sampler.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSIONS
 REFERENCES
 
Model Setup
The HYDRUS-2D model (Simunek et al., 1999) solves the Richards' equation (Richards, 1931) for two-dimensional vertical flow or three-dimensional axisymmetrical flow using the van Genuchten (1980) and Mualem (1976) soil hydraulic functions. The soil profile of the experimental site where the wick samplers were to be installed is described in the HYDRUS-2D as three different soil horizons: Ap (0–30 cm), Bt (30–55 cm), and B32 (55–300 cm). Soil hydraulic parameters (Ks, the saturated hydraulic conductivity; {theta}s and {theta}r, the saturated and residual water content, respectively [L3 L–3]; {alpha} [L–1] and n [unitless] with m = 1 – 1/n [unitless], shape parameters) of the three horizons were estimated using laboratory measurements (Klute, 1986) in a previous study (Diels, 1994) and are presented in Table 1. The upper atmospheric boundary condition consisted of measured daily rainfall and estimated potential evapotranspiration (Allen et al., 1998) data from a station near the experimental field measured between 10 Jan. 1995 and 31 Dec. 1998. Root water uptake was simulated according to Feddes et al. (1978) using the parameters for grass from the HYDRUS-2D database. A linear root distribution varying between 0 at the bottom of the root zone (30-cm depth) and 1 at the soil surface was incorporated.


View this table:
[in this window]
[in a new window]

  		Table 1. Hydraulic parameters{dagger} (van Genuchten–Mualem model) for the three different soil horizons of the experimental field and the two wick types used in the simulations (saturated hydraulic conductivity [Ks] and saturated volumetric moisture content [{theta}s] of the wicks have been corrected for a sampling area of 707 cm2).

 
Figure 1 presents a sketch of a 30-cm-diameter circular wick sampler as well as the way it is implemented in the HYDRUS-2D software as a vertical axisymmetrical three-dimensional problem. The arrow indicates the vertical axis of symmetry around which the model is rotated. The lower boundary was assumed to be a free-drainage boundary type and the sides of the flow domain were set up as no-flow boundaries. The core of this study was based on the fact that the wick itself can be represented as a porous material within the model. It was incorporated as a fourth "soil type" and laterally delineated by no-flow boundaries. The bottom of the wick was represented as a seepage-face flow boundary, which corresponds to the lower end of the wick where water leaves the wick and is collected into a container. Knutson and Selker (1994) presented the van Genuchten (1980)–Mualem (1976) parameters for different wick types for independent n and m. In this study, only two of those were considered: the Amatex 0.95-cm (3/8-inch) high-density no. 10–864KR-02 (Amatex-Norfab Corp., Norristown, PA) (AM3/8HI) and Pepperell 1.27-cm (1/2-inch) no. 1381 (Pepperell Braiding Co., Pepperell, MA) (PEP1/2). The wick was incorporated as being 30 cm wide across its entire length. This is in contrast with the actual situation, where a wick sampler might be 30 cm wide but the wick(s) leaving the sampler usually has a width of around 1 cm depending on the wick diameter and number of wicks used. To correct for this discrepancy between the model and the actual situation, the Ks and {theta}s of the wicks, as measured by Knutson and Selker (1994), were multiplied by the ratio of wick area to sampler area. Incorporating the real wick dimensions in the model and not scaling its physical parameters would increase the possibility of numerical instabilities due to abrupt shape changes in the finite element model mesh. Additionally, using our approach we could evaluate the effect of using multiple wicks simply by rescaling the wick physical parameters without having to redefine the finite element model domain. The corresponding retention and hydraulic conductivity were refitted using the RETC code (van Genuchten et al., 1991) and the assumption that m = 1 – 1/n for use in HYDRUS-2D. More weight was given to the wetter part of both the retention and conductivity curves in the fitting exercise to achieve reasonable good fits for this part of both curves. Table 1 presents the hydraulic parameters (van Genuchten, 1980; Mualem, 1976) for the three soil horizons and the two wick types used in this study, having corrected the Ks and {theta}s of the wicks with the of wick area/sampler area ratio. A circular wick sampler with a 30-cm diameter, corresponding to an area of 707 cm2, was assumed.


Figure 1
View larger version (70K):
[in this window]
[in a new window]

  		Fig. 1. Reality compared to the three-dimensional axisymmetrical modeled flow domain (vertical cut) and boundary conditions (l_Wick = length of the wick). The arrow indicates the axis of symmetry around which the model is rotated.

 
Initial soil tensions were set to a –50-cm pressure head throughout the entire domain. The numerical water-balance errors were examined and the simulated leachate volumes sampled by the wick were compared for different wick lengths and configurations against the simulated soil water flux at the same depth in a flow domain without a wick. In all results presented here, the first 3 mo of simulations were considered the "initialization" period and not considered in any of the analyses presented. The finite element mesh was generated using the triangulation tool of the MESHGEN-2D embedded in the HYDRUS-2D software. The mesh was refined in the neighborhood of internal boundaries. Depending on the complexity of the model layout (i.e., length of internal boundaries), between 1000 and 7000 nodes were incorporated in the model.


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
SUMMARY AND CONCLUSIONS
 REFERENCES
 
Effect of Distance between Top and Base of the Wick Sampler
Since wick samplers are usually installed through the sidewalls of a soil pit or trench, a "window" needs to be cut through the sidewalls into the undisturbed soil. As shown by Mertens et al. (2005) in a similar numerical study investigating the design of ETPLs, the height of this "window" has to be sufficiently large to minimize the "umbrella" effect, which is the effect on the sampled leachate of the dry zone created by the wick beneath the wick sampler. Mertens et al. (2005) showed that a distance of 50 cm was sufficient to minimize this effect, even for clayey soil types. This analysis was therefore not repeated in this study and 50-cm-high "windows" in which the wick samplers were installed were considered in this numerical exercise. Water-balance errors (i.e., the difference between water input volume as a result of rainfall and water output volume as a result of evapotranspiration and fluxes out of the model domain must be equal to the change in storage volume) were <1% in all simulations reported below.

Effect of Wick Type, Length, and Diameter on Cumulative Leachate Volumes
It was investigated to what extent the wick length influenced the sampled leachate volumes for both wick types. Four different wick lengths were tested: 30, 50, 75, and 100 cm. For each of these wick lengths, a new axisymmetrical three-dimensional model was constructed since the length of the internal no-flow boundaries around the wicks varied with wick length. As explained above, the "window" height was kept at 50 cm and, therefore, the minimal length of the internal no-flow boundaries was kept minimum at 50 cm even if inside a wick length of only 30 cm was modeled. Before the wick simulations, the water captured during the 3-yr period at 40-cm depth by an ETPL was simulated. The modeling of an ETPL requires two simulations. The first one without the ETPL has an observation node, inserted in the model domain at the location where the tension plate will be located. This node records the matric head potential that occurs in the undisturbed soil profile throughout the simulation. This recorded time series of tensions is subsequently applied in the second simulation as a variable-head boundary condition to simulate the leachate volume collected by the ETPL. This volume is referred to as the reference volume.

The cumulative leachate volumes sampled by both wick types compared with the reference volume at 40-cm depth are presented in Fig. 2 . Not surprisingly, longer wicks lead to higher sampling volumes. For wick lengths >30 cm, the AM3/8HI wick type yielded higher volumes than the PEP1/2 wick type at an equal length. The reason for this can be found in examining the conductivity functions of both wick types, presented in Fig. 3 . Note that the plotted wick conductivity functions are area corrected for a 707-cm2 sampler. Although the PEP1/2 wick had a larger saturated conductivity value than the AM3/8HI type, its conductivity decreased steeply with increasing tension. For example, at 40-cm tension, the conductivity of the PEP1/2 wick was even lower than the conductivity of the Ap horizon. This was not the case for the AM3/8HI wick and as a result the AM3/8HI wick type sampled higher leachate volumes than the PEP1/2 type when wick lengths were >30 cm. In the case where the wicks were only 30 cm or shorter, the PEP1/2 wick type sampled slightly more leachate than the AM3/8HI due to its higher conductivity at these low tensions. Differences between sampled volumes for 50-, 75-, and 100-cm length for the PEP1/2 wick type are small compared with the differences in sampled leachate volumes at those lengths using the AM3/8HI wick type. The reason again is the very low unsaturated conductivity of the PEP1/2 wick at tensions >50 cm. This means that once the length of the PEP1/2 wick reaches 50 cm, increasing its length would not increase sampled leachate volumes significantly. Since unsaturated conductivities up to 100-cm tension are still comparable to the soil unsaturated conductivities for the AM3/8HI type, its length up to 100 cm does have a significant impact on the sampled volumes. It is for this reason of more flexibility that the AM3/8HI wick type was selected for the samplers to be installed in the field experiment. Rimmer et al. (1995) stated that for a good match between wick and soil types, (i) their capillary length ({alpha}–1) needed to be similar and (ii) the ratio of their saturated hydraulic conductivities should be similar to the inverse ratio of their cross-section. Following these recommendations and inspecting Fig. 3, it is evident that the AM3/8HI wick type is more suitable for the soil conditions at the experimental site, as used in this numerical analysis.


Figure 2
View larger version (18K):
[in this window]
[in a new window]

  		Fig. 2. Simulated cumulative leachate volumes sampled by (a) the Amatex 3/8-inch high-density wick (AM 3/8 HI) and (b) the Pepperell 1/2-inch wick (PEP1/2) during a 3-yr period compared with the reference.

 

Figure 3
View larger version (23K):
[in this window]
[in a new window]

  		Fig. 3. Hydraulic conductivity functions of the three soil horizons (Ap, Bt, and B32) and the two wick types (Amatex 3/8-inch high-density [AM 3/8 HI] and Pepperell 1/2-inch [PEP 1/2]).

 
An AM3/8HI wick length of 50 cm seemed to get the cumulative volumes right for the first 2 yr but then overestimated the leachate volumes during the last year of the simulations. Since simulated volumes for wick lengths of 50 cm were already higher than the reference volume, the effect of increasing wick diameter (or number of wicks) was not further investigated. This could, however, be easily done by just rescaling the saturated soil moisture content and hydraulic conductivity for the area considered (e.g., just multiply by two if two wicks are to be used). Rimmer et al. (1995) recommended, on the basis of their one-dimensional steady-state model, that the optimal wick length for sandy soils was 30 to 40 cm, while for silt loams, the optimal length was >100 cm. Based on our axisymmetrical three-dimensional transient numerical simulations, we decided that the length of the AM3/8HI wick should be made seasonally variable. As a first estimate, the modeling suggested that a 40-cm-long wick in winter and a 100-cm-long wick in summer might be the most appropriate. We decided, however, that one "sampling" tensiometer would be installed above each wick sampler. The "sampling" tensions measured above the wicks could then be compared to "reference" tensions measured by the "reference" tensiometers installed in the undisturbed soil at the same depth below the ground surface. On the basis of differences between both tensions, corrections can be made to the wick length; however, it is not our intention to correct wick lengths for small differences or on a small temporal resolution but rather change the length between the different seasons.

Figure 4 zooms in on the simulated leachate volumes sampled during a wet period (Fig. 4a) and a dry period (Fig. 4b). From both figures, it is obvious that the wicks had difficulties getting the dynamics right. The wicks tended to work as an "on or off" kind of sampling tool compared with the reference. This is especially true for the shorter wicks, e.g., no leachate sampled during February 1998 using the 30-cm wick length while the reference and 100-cm wick still sampled significant amounts of water. Even longer length wicks, however, had difficulties in getting the recession part of the leachate curves right. This is most obvious during dry periods where, in the undisturbed soil, some water still moved downward through the profile that would never be sampled by a wick sampler. The reason for this is the fact that wick conductivities at tensions higher than –100 cm are much lower than the conductivity of the surrounding soil (see Fig. 3). From this it can be concluded that, when using wick samplers, the small amount of water flowing through the vadose zone during summer periods will not be sampled.


Figure 4
View larger version (15K):
[in this window]
[in a new window]

  		Fig. 4. Simulated daily leachate volumes sampled by a 30- and 100-cm long Amatex 3/8-inch high-density (AM 3/8 HI) wick in (a) a wet period and (b) a dry period compared with the reference.

 
Effect of Wick Samplers on Soil Moisture Content
It is common knowledge that soil moisture content is of primordial importance in microbiological processes—certainly when investigating the transport of DOM, associated pesticides, and Cu2+. For this reason, the effect wick samplers might have on soil moisture in the neighborhood of the wick was investigated. Figure 5 presents the simulated soil tension and soil moisture content at 1 cm above two AM3/8HI wick samplers, one 30 cm long and the other 100 cm long, corresponding to the time periods shown in Fig. 4. Differences between the reference tension and the tensions simulated at 1 cm above the wick are small. Once tensions reached about 100 cm, the soil above the wick remained wetter for both wick lengths since wick conductivities at these tensions are low. Consequently, differences between the reference soil moisture content and the soil moisture content at 1 cm above the wicks was small as well. Maximum deviation between the reference and soil moisture contents above the wicks was <0.05 m3 m–3. Based on these results, we believe that the effect of the wicks on the microbiological process is negligible.


Figure 5
View larger version (31K):
[in this window]
[in a new window]

  		Fig. 5. Simulated (top) soil tension and (bottom) moisture content at 1 cm above 30- and 100-cm long Amatex 3/8-inch high-density (AM 3/8 HI) wicks in (a) a wet period and (b) a dry period compared to the reference.

 
Effect of Boundary Condition between Soil Pit and Wick Sampler
Since wicks are usually installed through a "window" cut in the sidewalls of a soil pit or trench but at some distance away to avoid boundary effects, an inner boundary is created. Figure 6 illustrates what is meant by this "inner boundary condition". As shown for ETPLs (Mertens et al., 2005), the effect of the inner boundary type on the sampled leachate volumes is important. Ideally, this boundary needs to be a free-drainage boundary. A free-drainage condition can only be created by refilling the soil between the access pit and the sampler with the same bulk density as the undisturbed soil. Leaving the void open changes the inner boundary into a seepage-face boundary condition in modeling terms. The effect of a seepage-face condition on the leachate volumes sampled by the wick sampler was investigated.


Figure 6
View larger version (71K):
[in this window]
[in a new window]

  		Fig. 6. Two-dimensional modeled flow domain (vertical cut) and schematization of the inner boundary condition.

 
Figure 6 shows the model setup used for the evaluation of the "inner boundary condition" as implemented in HYDRUS-2D. Introducing this void between the installation trench and the wick sampler implies that the water flow can no longer be accurately described by an axisymmetrical three-dimensional model. Therefore, a two-dimensional model was adopted, realizing that ideally, a three-dimensional model would be required to accurately describe the water flow. In the two-dimensional model, water reaching the inner boundary is either forced through the seepage face or flows into the wick. In reality, however, this water can also flow in the third dimension (perpendicular to the model domain presented in Fig. 6) across the 30-cm-wide inner boundary. This is also the reason that the length of the inner boundary condition was assumed to be only 30 cm since longer seepage-face lengths would increase the need for the presence of the third dimension in the model domain. Still, the results presented based on a two-dimensional analysis should be interpreted as indications of the importance of the inner boundary effects rather than quantitatively being exact. First, a two-dimensional model with an AM3/8HI wick type of 50-cm length was set up and resulted in a total leachate volume of 1331 mm during the entire simulation period. Note that this is remarkably less than the 1910 mm simulated using an axisymmetrical three-dimensional analysis of the same wick sampler layout. The reason is the larger wick effect on the simulated leachate in three dimensions compared with the setup presented in Fig. 6, where the wick was assumed to only affect the water flow in two dimensions. Table 2 shows that the sampled leachate volume increased from 1331 to 1886 mm in the presence of an inner boundary seepage-face condition. This increase in leachate volume is due to the fact that saturated conditions are needed for water to flow through the seepage face, resulting in water converging to the wick.


View this table:
[in this window]
[in a new window]

  		Table 2. Total simulated volume of leachate sampled during a 3-yr period by a single wick sampler with and without the inner boundary condition and by a "double-ring" wick sampler with the inner boundary condition in place.

 
Since it is hardly realizable to refill the soil so it has the same bulk density as the original soil, it was decided that the window should be left open. Not refilling the soil makes the wick samplers remain easily accessible and facilitates changing wick length based on the tensiometer readings. The question was raised whether a kind of a "double-ring" wick sampler could overcome the oversampling problem caused by the seepage face. Therefore, a two-dimensional model was set up with a central AM3/8HI wick sampler (30-cm diameter and 50-cm length) surrounded by an 8-cm-wide outer ring having the same physical properties as the inner wick sampler (Fig. 7 ). Simulated leachate volumes through the left wick (closest to the inner boundary seepage face), middle wick, and right wick are presented in Table 2.


Figure 7
View larger version (34K):
[in this window]
[in a new window]

  		Fig. 7. Water flow velocity vectors around a double-ring wick sampler with an inner seepage-face boundary condition

 
We know that a wick sampler without an inner boundary condition sampled 1331 mm of leachate during the simulation period. Table 2 shows that the left wick sampled 2210 mm during the entire period. The reason for this is the obvious convergence of water from the seepage face to the left wick, as presented in Fig. 7. Fortunately, the leachate volume sampled by the middle wick (1330 mm) was very close to the volume sampled without the seepage face (1331 mm). The simulated volume captured by the right wick was 1430 mm, which is a little different from the volume sampled without the seepage face. This is probably due to (i) small differences between both model setups (e.g., different mesh) and (ii) side effects influencing the flow around the edges of the wick sampler. Since differences were small (<10%), we concluded that an 8-cm-wide wick to the left of the middle wick will capture the convergent flow induced by the seepage face and therefore solves the oversampling caused by this seepage-face inner boundary condition. Simulations additionally suggest that the leachate sampled by both the middle and right wick should not be discarded and can be used for chemical analysis.

Again, it must be stressed that forcing a three-dimensional system to behave as if it were a two-dimensional computational problem might induce artifacts in the solution. Therefore the results presented here should be seen as a qualitative evaluation rather than a true quantification of the inner boundary effect. Additionally, we wanted an easy and relatively fast tool to help identify possible artifacts before the field installation and we believe that our two-dimensional analysis succeeded in that. Since a full three-dimensional analysis before field installation can be difficult and time consuming, it is generally not possible before field installation due to lack of time and money.

Layout of Wick Sampler
Based on the numerical simulations, a wick sampler layout is suggested. First of all, the AM3/8HI fiberglass type was chosen for reasons explained above. Additionally, it was decided that the length of the wick should be made seasonally variable on the basis of differences between reference and sampling tensions. The final shape of the wick sampler layout is presented in Fig. 8 . A rectangular shape was chosen for ease of installation. A narrow rectangular shape enables reducing the width of the "window" (in this case 35 cm) and minimizing the risk of soil collapsing during the installation. Since Wick 3 corresponds to the so-called left wick in the simulations above, it is expected that this wick will capture significantly more leachate than Wicks 1 and 2. Wick 2 will capture the water flow at the edges around Wick 1 and the sampler can therefore be seen as a kind of "double-ring" wick sampler. Simulations suggested that Wicks 1 and 2 should capture very similar amounts of leachate. If this is confirmed in the field experiment, the leachate of both wicks can be used for chemical analysis, and the sampler area will be doubled.


Figure 8
View larger version (32K):
[in this window]
[in a new window]

  		Fig. 8. Top view of the layout of the wick sampler (dashed lines represent the unravelling of the wicks).

 

   SUMMARY AND CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSIONS
REFERENCES
 
The numerical analysis of the performance of passive capillary wick samplers for different configurations was conducted within the framework of the setup of a field experiment for investigating the transfer of DOM and associated Cu2+ and pesticides. The aim of the research was to identify before the installation the configuration of the fiberglass wick sampler that would most likely collect the most realistic fluxes for the given soil conditions. A modeling tool was set up to answer four key questions that would help to define the most appropriate wick sampler design.

The simulations revealed that (i) the AM3/8HI wick type is the most suitable for the soil and climatic conditions where the experimental field will be established; (ii) a wick length of 40 cm in winter and 100 cm in summer yielded realistic leachate volumes; (iii) regular adjustment of the length of the wick on the basis of the difference in tension measured 1 cm above the wick and at a point at the same elevation but placed outside the influence of the wick sampler would considerably improve the performance of the wick sampler; (iv) the effect of the wick samplers on the soil moisture content close to the sampler is negligible, and (v) the effect of the area between the installation pit or trench and the wick sampler can be minimized using a "double-ring" wick sampler. Based on those observations, a final layout of the wick samplers was designed, as depicted in Fig. 8.

The developed modeling tool consists of incorporating the wick sampler into the modeling domain as a porous material with hydraulic properties equal to the hydraulic properties of the fiberglass wick sampler. It is a very effective instrument for answering questions arising during the design phase of field experiments in which wick samplers are planned for the collection of leachate under natural conditions. The approach is easily applicable and transferable to other soil types, wick types, and climatic conditions. Very importantly, the study revealed that through numerical simulations, it is feasible to optimize the field design before equipment installation.


   ACKNOWLEDGMENTS
 
We are grateful to Frans Schoovaerts and Valentijn Tuts for their professional technical assistance in the construction of the wick samplers and their installation. This research was feasible thanks to a GOA grant of the Katholieke Universiteit Leuven, awarded for studying the dynamics of DOM in the vadose zone and the associated transport of Cu2+ and pesticides.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSIONS
 REFERENCES
 
Abbreviations: AM3/8HI, Amatex 3/8-inch high-density wick; DOM, dissolved organic matter; ETPL, equilibrium tension plate lysimeter; PEP1/2, Pepperell 1/2-inch wick.

Received for publication March 8, 2006.


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




This article has been cited by other articles:


Home page
 Vadose Zone JHome page
J. Mertens, V. Tuts, J. Diels, J. Vanderborght, J. Feyen, and R. Merckx
Design and Testing of a Drop Counter for Use in Vadose Zone Water Samplers
Vadose Zone J., April 14, 2008; 7(2): 434 - 438.
[Abstract] [Full Text] [PDF]

Home page
 J. Environ. Qual.Home page
L. Weihermuller, J. Siemens, M. Deurer, S. Knoblauch, H. Rupp, A. Gottlein, and T. Putz
In Situ Soil Water Extraction: A Review
J. Environ. Qual., October 24, 2007; 36(6): 1735 - 1748.
[Abstract] [Full Text] [PDF]

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 ISI 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 HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Mertens, J.
Right arrow Articles by Vanderborght, J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Mertens, J.
Right arrow Articles by Vanderborght, J.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Mertens, J.
Right arrow Articles by Vanderborght, J.
Related Collections
Right arrow Vadose Zone Processes and Chemical Transport
Right arrow Water Flow Models
Right arrow Experiment Design

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