| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Agrosphere Inst., ICG-4, Forschungszentrum Jülich GmbH, 52425 Jülich, Germany
b Inst. of Soil Science, Berlin Univ. of Technology, Salzufer 12, 10587 Berlin, Germany
c Inst. of Soil Sciences, University Hannover, Herrenhäuser Straße 2, 30419 Hannover, Germany
d current address, HortResearch, Tennent Drive, Palmerston North, New Zealand
e Thüringer Landesanstalt für Landwirtschaft, Naumburger Strasse 98, 07743 Jena, Germany
f The Helmholtz Centre for Environmental Research– UFZ, Department of Soil Physics, Lysimeter Branch, Dorfstrasse 55, 39615 Falkenberg, Germany
g Fachgebiet Waldernährung und Wasserhaushalt, Technische Universität München, Am Hochanger 13, 85354 Freising, Germany
* Corresponding author (l.weihermueller{at}fz-juelich.de).
Received for publication May 2, 2007.
ABSTRACT
The knowledge of the composition and fluxes of vadose zone water is essential for a wide range of scientific and practical fields, including water-use management, pesticide registration, fate of xenobiotics, monitoring of disposal from mining and industries, nutrient management of agricultural and forest ecosystems, ecology, and environmental protection. Nowadays, water and solute flow can be monitored using either in situ methods or minimally invasive geophysical measurements. In situ information, however, is necessary to interpret most geophysical data sets and to determine the chemical composition of seepage water. Therefore, we present a comprehensive review of in situ soil water extraction methods to monitor solute concentration, solute transport, and to calculate mass balances in natural soils. We distinguished six different sampling devices: porous cups, porous plates, capillary wicks, pan lysimeters, resin boxes, and lysimeters. For each of the six sampling devices we discuss the basic principles, the advantages and disadvantages, and limits of data acquisition. We also give decision guidance for the selection of the appropriate sampling system. The choice of material is addressed in terms of potential contamination, filtering, and sorption of the target substances. The information provided in this review will support scientists and professionals in optimizing their experimental set-up for meeting their specific goals.
THE unsaturated zone is of great importance for matter fluxes in terrestrial ecosystems. The in situ monitoring of the soil solution therefore is of basic interest for studies concerning ecology, water management, agriculture, forestry, and environmental protection. The major aim of soil water extraction is the specification of the quality and quantity of soil pore water for the different scientific and practical questions. The rapid development of analytical methods during the past decades offers the possibility of a comprehensive qualitative and quantitative documentation of the solutes dissolved in the soil water. This requires an optimal sampling system to avoid artifacts and bias from contamination and changes of the extracted soil water. Furthermore, new developments in data analysis, modeling, and monitoring of water and solute flow by non-invasive methods such as time domain reflectrometry (TDR), electrical resistivity tomography (ERT), or ground penetrating radar (GPR) increased the requirements of sampling and changed needs of the experimental design.
Porous cups, porous plates, capillary wicks, resin boxes, and lysimeters are widely used for the in situ monitoring of the soil solution. The advantage of the aforementioned methods is the high temporal and/or spatial resolution of monitoring solute movement and quality, easy installation, low costs, and large experience gained regarding potential sources of error. During the last decades, a wide range of new materials and modifications of these sampling devices were developed for different scientific approaches.
The main objective of this article is to give an overview of the existing in situ extraction devices in the vadose zone, to support the selection of adequate sampling systems for solute concentration and solute transport measurements, as well as mass balance calculation. Therefore, the representativeness, installation effort, maintenance, temporal resolution, and cost expense were listed and discussed.
Conversely, it is not our intention to analyze in full depth the advances made for each of the different sampling devices, but rather give an overview on what is presently available to sample solutes in the vadose zone and to derive needs for future research.
To meet this objective, we start with a general description, implicit advantages and disadvantages of the sampling systems, the installation procedures, a priori conditioning, and technical, as well as interpretational biases of the respective techniques.
In a second part, we address more general issues related to solute extraction and data acquisition. This includes the criteria of experimental design for solute concentration measurements, solute transport monitoring, calculation of mass balances, as well as the time scale of data acquisition and installation costs for the different devices, with regard to the study aims. Finally, we cover specific requirements for sampling the most important vadose zone solutes, including the choice of materials for sampler construction.
Sampling Devices, Their Advantages, Disadvantages, and Limitations
Suction Cups
Various terms have been used in literature to describe this sampling device, such as porous tube (Krone et al., 1951), deep pressure vacuum lysimeter (Parizek and Lane, 1970), vacuum extractor, porous candle, porous cup, or suction cup (Duke and Haise, 1973). Since the suction or porous cup is only part of the whole system (the small porous body at the lower end) the term suction probe is also proposed (Grossmann and Udluft, 1991). Subsequently, we will use the term suction cup for the whole sampling device.
The principle of the porous cup was first described by Briggs and McCall (1904). Since then, suction cups were widely used in different studies to collect soil water for analytical purposes. Different kinds of suction cups made out of diverse materials are described in literature by Barbee and Brown (1986), Dorrance et al. (1991), and Hart and Lowery (1997). In general they all consist of a cylindrical porous cup sealed to a tube. Inside this tube a small tube is inserted to collect the extracted water. Different ceramic materials, sintered materials, and membranes are nowadays in use as suction cup materials (Dorrance et al., 1991).
The installation of the suction cup into the soil profile is rather simple compared to other soil water sampling systems. In general, four installation modes are possible: (i) horizontal, (ii) vertical non-shaft, (iii) vertical, and (iv) vertical in 45° (Mitchell et al., 2001) (Fig. 1 ).
|
For the operation of a suction cup, a negative pressure has to be imposed by applying suction to the cup using a vacuum system. The optimal height of applied suction to the cup, and the optimal operation mode are still under debate (McGuire and Lowery, 1994; Brandi-Dohrn et al., 1996; Weihermüller et al., 2005). In general, the suction to be applied to the porous cup depends on the soil type, the specific amount of water required for analysis, the actual soil water content, and the time of applied suction (Warrick and Amoozegar-Fard, 1977; Weihermüller, 2005).
For the extraction of soil water with porous cups two operation modes are possible:
The tendencies of suction cups to sorb organic and inorganic compounds have been widely discussed in literature by Wagner (1962), Hansen and Harris (1975), McGuire et al. (1992), and Wessel-Bothe et al. (2000) among others. As stated before, new materials allow minimizing the effect of sorption and contamination by using the optimal material (see section "Substance-Specific Requirements of Solute Sampling" and Table 1 ).
|
Weihermüller et al. (2005) also showed in numerical simulations, assuming continuous infiltration that the SCAD, SCED, and SCSA depend on the soil hydraulic properties, the ambient water content, and the extraction time. The results of Weihermüller et al. (2006) indicate that the extraction domain and the sampling area are not constant in time and space for natural soils under atmospheric boundary conditions. This implies that the calculation of mass balances using porous cup data is not valid in its strict sense or is at least problematic.
Jury and Flühler (1992) suggested that samplers (particularly shallow ones) may be unreliable in soils with preferential flow, whereby recording of bypass flow might be a fundamental weakness of the cup sampler or a result of having too few samplers for adequate sampling of the flow regime. Therefore, the question of how many samplers are required remains, and whether it is practical to install and monitor the required number of suction cups (Flemming and Butters, 1995).
In the work of Weihermüller et al. (2006), it has already been shown that soil heterogeneity seems to play an important role for solute sampling using suction cups. Especially since effective transport parameters calculated from suction cup breakthrough curves were prone to considerable uncertainty.
Given the uncertainty regarding the SCED and the SCSA, the interpretation of sampled concentrations with suction cups is still a matter of debate. Particularly in connection with discontinuous operation mode combined with large vacuum applied, solute concentrations of samples should be similar to resident concentrations [M L–3], which were defined as the mass of solute per unit volume of fluid contained in an elementary volume of the system at a given instant (Kreft and Zuber, 1978). Often suction cups, however, are applied to determine the flux or flux concentration [M L–2 T–1] (Kreft and Zuber, 1978) of solutes through a plane of defined surface. Representative concentrations of flowing water are rather sampled in the continuous mode of operation than in the discontinuous mode. To derive solute fluxes these concentrations are multiplied with water fluxes that are frequently derived from numerical models of the soil water balance, which adds a considerable uncertainty to the calculated flux concentrations.
Suction cups are by far the most frequently used technique for extracting soil water. Easy installation and a large treasure of experience are the most important advantages of this technique. Especially, if solute transport and mass balances are in the focus of the experiment, the poor definition of the SCED and the SCSA together with the small cross-sectional area of the cup pose serious limitations on the interpretation of results. Furthermore, the necessity to use independent estimates of soil water fluxes increase the uncertainty of calculated fluxes. Alternative techniques, which aim at avoiding these limitations, have been proposed since the early beginnings of in situ soil water extraction.
Suction Plates
Similar to the terminology used for suction cups, the term suction plate or tension plate lysimeter is used for separate porous extraction plates as well as for the whole sampling device. In general, the porous plate is inserted into a frame connected to a tube for water extraction (Fig. 2
). Porous ceramics, nylon membranes, sintered stainless steel, and glass are available materials for the porous body.
|
Various control options for the tension applied to the porous plate have been proposed. The simplest operation method is zero tension, where no suction will be applied to the device. The main disadvantage of this operation mode is the formation of a saturated zone above the plate resulting in artifacts such as divergent water flow away from the system, and therefore underestimation of the natural water flux (Chiu and Shackelford, 2000). Flury et al. (1999) also showed in numerical simulations that these seepage face conditions not only influence the water flow but also the solute concentration in the sampled leachate.
A second operation mode is characterized by a fixed predefined suction that is applied to the plate. The matric potential of the soil generally varies as a function of space and time. As a result, the soil water regime in direct vicinity of the plate is likely to be different from the fixed suction exerted by the plates. These results in changes of the natural flow field, and therefore, in differences in the solute concentration compared to the freely percolating water in the soil profile (Rhoades and Oster, 1986; Kosugi and Katsuyama, 2004).
To overcome this limitation, the most promising approach to sample representative soil water with porous plates is the application of a suction equivalent to the ambient matric potential at the same depth (van Grinsven et al., 1988; Byre et al., 1999, 2001; Foley et al., 2003; Lentz and Kincaid, 2003; Pelger et al., 2003; Siemens et al., 2003; Barzegar et al., 2004; Kosugi and Katsuyama, 2004; Siemens and Kaupenjohann, 2004; Mertens et al., 2005). The ambient matric potential is generally measured by reference tensiometers and automatically applied to the plates. This control strategy is expected to sample representative water flow and solute concentrations with little disturbance of the natural flow field. Nevertheless, resulting solute concentrations and solute transport parameters calculated from solute breakthrough curves sampled with suction plates should be verified using water balance models, water tracers, or numerical simulations (Siemens et al., 2003; Kosugi and Katsuyama, 2004; Siemens and Kaupenjohann, 2004, Mertens et al., 2005, Kasteel et al., 2006).
In comparison to suction cups, porous plates exhibit a larger sampling area. Therefore, the detection of preferential flow events may be possible using suction plate arrays (Ciglasch et al., 2005). Due to the 2D surface of the plates, the origin of the sampled water and solutes is also better defined as for suction cups, which supports mass balance estimations (Kasteel et al., 2006). These advantages, however, are associated with larger efforts and disturbance of experimental plots for installation as compared to the installation of suction cups.
Pan Lysimeter
Pan or zero-tension lysimeters are passive water samplers, typically in the shape of a pan without large side walls extending above the system (Fig. 3
), which will collect freely percolating soil water (Ebermayer, 1873; Jemison and Fox, 1992). The pan lysimeter system itself can be made of different materials such as steel, stainless steel, glass, ceramic, or plastic material depending on the scientific question and the target substance. The sampling surfaces of the system can exhibit several square meters (e.g., in waste deposal sealings) with standard dimensions being about 0.5 m2. In general, pan lysimeters are placed below the ground surface to capture drainage water. The installation is comparable to the installation of suction plates. They can also be used in shallow depth, e.g., in organic or forest litter layers and then are called humus lysimeters (Marques et al., 1996; Ranger et al., 2001).
|
Due to the design of the sampler and the absence of capillary connection to the soil, pan lysimeters operate reasonably well in soils with large macro-pores near saturation but are much less successful if the soil dries out (Zhu et al., 2002).
Initially, pan lysimeters were used primarily to analyze water quality and only occasionally to quantify drainage rates. More recently, zero-tension lysimeters have been used to estimate drainage rates over a wide range of soil conditions (Chiu and Shackelford, 2000; Zhu et al., 2002; van der Velde et al., 2003, 2004). Because of water divergence, collection efficiencies less than 10% have been noted for pan lysimeters (Jemison and Fox, 1992; Zhu et al., 2002). Therefore, diversion around zero-tension lysimeters can be a significant problem. Flury et al. (1999) also showed in numerical simulations that these seepage face conditions not only influence the water flow, but also the solute concentration in the sampled leachate. The prominent advantages of pan lysimeters are their low cost and easy maintainance. Disadvantageous are their complex installation, that causes considerable disturbance on experimental plots and the divergence of water flow around the system, which prevents quantitative estimates of flux concentrations and, therefore, complicates the interpretation of solute breakthrough and which may even lead to complete failure of the system. Humus lysimeters show less problems with saturation and bypass flow, due to the fact that humus has a more coarse and open-pored structure. Humus lysimeters only have a nylon mesh at the top and are not filled with gravel or other mineral materials, because the water flowing out of the humus layer should not have contact to mineral surfaces which would cause flocculation and/or changes in solute chemistry (Guggenberger and Zech, 1993).
Wick Samplers
Capillary wick samplers or wick lysimeters are sampling devices which sample soil water by the gravitational potential using an inert wick material such as fiberglass (Holder et al., 1991) or rock wool (Ben-Gal and Shani, 2002) (Fig. 4
). By applying a hanging water column using wicks, the drainage water is pulled out of the sampling device while the lower soil boundary is maintained at a pressure less than atmospheric resulting in an unsaturated soil. The degree of unsaturation depends on the material, length, and diameter of the wick, the dimension of the sampling bottle, the flux rate, and the soil type (Holder et al., 1991; Boll et al., 1992, Knutson and Selker, 1994; Rimmer et al., 1995; Zhu et al., 2002; Mertens et al., 2007). The maximally applied suction is about 50 to 60 cm (Holder et al., 1991; Boll et al., 1992; Brandi-Dohrn et al., 1996). For wick-type lysimeters, drainage bypass can be further minimized by placing an extension tube above the wick (Gee et al., 2002, 2003). The extension tube is filled with soil from the excavation of the hole into which the wick sampler is placed. If the properties of the wick-sampler are adjusted to the soil properties a semiqualitative, direct analysis of the water and solute flow is possible (Boll et al., 1992; Knutson and Selker, 1994; Rimmer et al., 1995; Louie et al., 2000; Zhu et al., 2002; Siemens and Kaupenjohann, 2004). In extensive field testing over several years, leachate collection efficiencies (LCEs), defined as measured drainage divided by estimated drainage (obtained from a mass balance of precipitation and evapotranspiration), have been shown to equal or exceed 100% for wick samplers (Louie et al., 2000; Zhu et al., 2002). Gee et al. (2002, 2003) integrated a tipping bucket into their wick sampler to increase the temporal resolution of flux measurements.
|
Overall, wick lysimeters seem to offer a compromise between rather costly tensiometer-controlled suction plate systems and pan lysimeters, which strongly affect the flow field of soil water. It has to be noted, however, that the wick materials used for construction of the lysimeters are produced in large-scale industrial processes for heat insulation purposes. This means that variations of their hydraulic and chemical properties are likely to be large with repect to their use as scientific equipment. This in turn means that substantial testing of hydraulic, chemical, and sorptive properties of the samplers in pre-experiments is inevitable. The still rather small amount of operating experience gained with this technique also requires a sound testing of samplers before their large scale use in experiments.
Resin Boxes
Resin boxes adsorb solutes of percolating soil water reversibly on synthetic exchange resins. Subsequent extraction of these compounds from exchange resins in the laboratory allows estimation of the solute flux. The goal of the application of resin boxes is the estimation of solute fluxes through a commonly horizontal defined soil cross-sectional area. Solute concentrations cannot be monitored with this technique. Resin boxes commonly consist of a pipe of approximately 10-cm length, which is provided with a mesh at its lower end and filled with a mixture of quartz sand or silt and a synthetic exchange resin (Fig. 5
) (Bischoff et al., 1999; Deutsches Patent no. 19726813, 2003). The type of exchange resin depends on the target compound. Strong cation and anion exchange resins but also less polar sorbents for organic compounds can be used. Layers free of exchange resins at the top and the lower end of the boxes minimize passive diffusion of solutes into the box (Arnold, 1996). The stabilization of adsorbed compounds against biological transformations is a prerequisite for extended monitoring periods (Binkley, 1984). Because the hydraulic properties of the resin boxes deviate from those of the surrounding soil, a deviation of the flow field of soil-water from natural conditions around the box can be expected (Arnold, 1996). The magnitude of the potential bias of estimated solute fluxes that is induced by this pertubation of the flow field can be characterized with the help of tracer experiments (Lang and Kaupenjohann, 2004).
|
Resin boxes are cost effective and maintenance free. This allows the use of many boxes to derive representative values and to characterize spatial heterogeneity at the field or watershed scale (Bischoff et al., 2001). Disadvantages include laborious installation. Tracer experiments for quantification of diversions of water flow by the boxes seem necessary due to the given differences between hydraulic properties of the boxes and the surrounding soil.
Lysimeter
In general, lysimeters or soil columns are containers or vessels containing disturbed or undisturbed soils. The optimal surface area and the lysimeter length depend mainly on the scientific question, the filling procedure, the lower boundary, and the location of installation. The base area is strongly connected to the scale of observation, whereby small-scale heterogeneity will be averaged using large base areas. Lysimeters with crop stands should represent the natural crop inventory and the maximal root penetration depth should be taken into account. Lysimeters can be filled with either monolithic or disturbed soil or materials. Disturbed lysimeters can be filled with the disturbed horizons of the natural soil or artificial material can be used instead. If disturbed soils or materials are used, the natural texture and the spatial heterogeneity will be changed, which will result in changes of the water and solute flow (Johnson et al., 1995; Troxler et al., 1998).
The installation of the lysimeters can be in the field (Meissner et al., 2000), at special lysimeter facilities, under controlled conditions in greenhouses, or in laboratories. In the case of installation under natural boundary conditions, the upper surface of the lysimeters should be equal to the top ground surface to minimize microclimatic changes. The space between the lysimeter vessel and the sourrounding soil should be also minimized to reduce artifical temperature gradients within the soil block. The lower boundary can be segmented to obtain information on the spatial heterogeneity of the water and solute fluxes (Schoen et al., 1999). For acquisition of the surface run-off, the lysimeters can also be equipped with run-off/overflow tubes (van Weesenbeck et al., 1998).
By the drainage behavior of water from the system, two types of lysimeters can be distinguished: (i) Free drainage lysimeters, where water is allowed to drain freely through the soil under gravity, or (ii) suction controlled drainage system, where a defined suction is imposed at the lower boundary using suction cups, wick samplers, or porous plates.
In general, a free drainage lysimeter is easier to install and cheaper than the controlled suction system. A major concern of the free drainage lysimeter is that the lower boundary is exposed to the atmospheric pressure, resulting in an evolution of a water-saturated zone at the bottom of the lysimeter before drainage (Abdou and Flury, 2004). This lower boundary imposes temporary anaerobic conditions which may influence degradation, solute transport, and the capillary rise during evapotranspiration (Bergström, 1990; Giessler et al., 1996). In comparison, suction lysimeters are more expensive and difficult to install, especially if they have large surface areas (Bergström, 1990). Another problem with suction controlled lysimeters is that water and solutes can interact with the material used for the suction devices. Also, the natural matric potential, water flow streamlines, and the composition of the leachate can be altered. The drainage patterns of both systems have been compared in several laboratory and numerical experiments, with the general finding that suction lysimeters drain more water continuously and in larger quantities (Colman, 1946; Dowdell and Webster, 1980; Vereecken and Dust, 1998; Abdou and Flury, 2004).
A major concern of the lysimeter concept is that it does not account for lateral water and solute fluxes, and that the vertical boundaries may cause fringe effects and preferential flow paths (Schoen et al., 1999). Titus and Mahendrappa (1996) listed several techniques to minimize the fringe effect in lysimeters.
The general aim of lysimeter studies is the measurement of solute concentrations at the lower boundary (flux concentrations), transport studies, and calculation of mass balances for scientific questions and pesticide registration (European Community, 1995). If the lysimeter is equipped for volatilization measurements a closed mass balance can be calculated even for volatile substances (Führ et al., 1998; Meissner et al., 2000; Wolters et al., 2003). Additionally, information of the actual evapotranspiration can be drawn if the lysimeter is settled on a scale and if the percolate is logged in short time intervals (e.g., using tipped buckets). Moreover, weighable lysimeters with groundwater control are used to measure the soil water balance parameter (e.g., evapotranspiration, capillary rise and groundwater recharge) of sites influenced by groundwater (Bethge-Steffens et al., 2004).
Lysimeters installed into lysimeter facilities allow additional measurements using hydrogeophysical methods (Vereecken et al., 2006) such as ERT (Binley et al., 1996a, 1996b; Slater et al., 2002) or GPR (Schmalholz et al., 2004) for the characterization of the water and solute flow. These techniques may provide additional high spatial and temporal information necessary to describe nonuniform transport within the soil profile.
Criteria for Specifying the Experimental Design
As outlined in the presentation of the single sampling devices, sampling systems differ with regard to the type of information they collect, their resolution in space and time, their cost, and their maintainance requirements. To choose the right sampler type and specify the appropriate experimental design, the exact definition of the study target is a key issue. In general, targets differ in their content of information and can be grouped into three categories, which are: (i) solute concentration, (ii) solute transport, and (iii) mass balance (Fig. 6 ). The solute concentration is defined as the amount of substance within a specific volume of water (e.g., mg L–1). In general, solute transport is described as the relocation of chemicals within the pedon due to water fluxes (e.g., g cm–2 h–1). The solute transport, solute flux, or solute mass density (Js) is mainly driven by the bulk transport or convection of solute moving with the flowing soil water and/or gas phase and diffusion and dispersion processes (Jury and Horton, 2004). In special cases, the solute flow can be also associated to colloidal or particular transport within the water flux. For the mass balance, the difference between the sums of solute input and solute output within a defined environmental compartment has to be calculated (e.g., kg ha–1 yr–1). The differences between input and output can be ascribed to a loss of substance by chemical reactions such as degradation and sorption or to storage within the compartment.
|
|
Substance-Specific Requirements of Solute Sampling
Different physicochemical properties of the test substances in the soil solution cause a large spectrum of potential interactions between the used materials and the test substances regardless of the sampler type. No sampling system enables the optimal sampling of all test substances. Therefore, the sampling system should be optimized to match the respective study goal (Table 1). In general, the sampler equilibrates with the soil solution with time, regardless of the chosen material. Nevertheless, we advise not to rely on equilibration in the field or in laboratory treatments to achieve representative solute sampling, since (i) the required time for equilibration is not well defined, and (ii) the required time for equilibration is not always given. In the following, recommendations for the most frequently sampled target compounds are given. An overview is also presented in Table 1.
Protons (pH)
The partial pressure of CO2 in the soil air has a decisive influence on the pH value of the soil solution via the equilibrium of H2CO3, HCO3–, CO32–, and carbon dioxide. Since the CO2 partial pressure of the soil air is commonly several times greater than in the atmosphere, CO2 will be released from soil solution samples in contact with the atmosphere (Zabowski and Sletten, 1991). This increases the pH value of the sampled soil solution by about 0.3 to 0.5 pH units (Suarez, 1986, 1987; Kaupenjohann and David, 1996). The outgassing of CO2 can be minimized in dual chamber sampling systems as proposed by Suarez (1986).
Furthermore, the pH value of the sampled soil solution can be reduced by the oxidation of Fe2+, Mn2+, or NH4+ in the sampling system.
Nutrients (N, P, K, S, Ca, Mg)
The anion nitrate (NO3–) is characterized by low interaction with most materials. Studies indicate good agreement of nitrate fluxes determined with lysimeters and suction cups for well-drained loamy sand and sandy loam (Goulding and Webster, 1992; Webster et al., 1993). Ammonium (NH4+), in contrast, is subject to cation exchange processes. Ammonium and nitrate concentrations are easily altered by biological processes (e.g., nitrification, N-assimilation). The shortest possible sampling interval (<2 wk) and cool, dark storage of the samples in the field are therefore decisive. The sample solution thus obtained can be stabilized by adding acids (e.g., HCl) into the sampling bottles if this does not disturb the detection of other target data (e.g., pH value, Cl–) and their subsequent analysis. The samples should be kept cool and dark during transport.
The phosphate ion has a strong affinity to metal hydroxides to which it is specifically bound by ligand exchange. Ceramic materials that contain aluminium (hydr)oxide therefore sorb and desorb considerable quantities of PO43– (Bottcher et al., 1984). The sorption and desorption of PO43– in suction cups and plates can be prevented by using porous PTFE (polytetrafluoroethylene, eg., Teflon) or glass, whereby porous PTFE exhibit low air entrance values due to its hydrophobic surface (Bottcher et al., 1984). In general, a significant fraction of the phosphate is translocated by preferential flow events (Anderson and Xia, 2001; Heckrath et al., 1995), and therefore large surface sampling systems, such as lysimeters or suction plates, seem to be more suitable for recording representative PO43– fluxes than suction cups.
The concentrations of K+, Ca2+, and Mg2+ are influenced by cation exchange. Due to the low cation exchange capacity of most ceramics, stainless steel, and glass sinter materials, however, equilibrium between the concentrations of the soil solution and the components of the sampling systems is commonly soon established. Carbon dioxide outgassing may result in the precipitation of calcium carbonate (CaCO3) or magnesium carbonate (MgCO3) in the sampling systems (Schwarz and Miehlich, 1993). New ceramic and glass components may contain considerable quantities of Ca2+, Mg2+, K+, and Na+, and should therefore be cleaned by rinsing with hydrochloric acid (0.1 mol L–1) and deionized water (Grover and Lamborn, 1970; Wessel-Bothe et al., 2000; Weihermüller, 2005).
Trace Metals (Al, Mn, Fe, Pb, Cr, Cu, Ni, Zn)
Aluminium concentration and aluminium speciation are controlled by the ambient soil pH. Changes in the soil pH value induced by soil water extraction should therefore be taken into account and minimized. Aluminium often forms complexes with organic compounds. Therefore, the sampling systems should be characterized by low sorption of dissolved organic substances. Most ceramic materials contain Al (hydr)oxides and are therefore only suitable to a limited extent. Nylon, PE, and PTFE are more suitable for aluminium sampling.
Dissolved iron (Fe2+) and manganese (Mn2+) occur in environments with low redox potentials and were oxidized to Fe- and Mn-oxides after contact with oxygen in the sampling system (Schwarz and Miehlich, 1993). Increase in pH also leads to the precipitation of Fe- and Mn-(hydr)oxides. To minimize the oxidation and precipitation of Fe2+ and Mn2+, it is recommended that gas-tight tubing (e.g., of nylon) should be used and the soil solution in suction cups should be sampled in the shaft. An acidification of the collected soil solution by adding acids (e.g., HCl, HNO3) to pH 2 also stabilizes the Fe2+ concentrations.
Interaction of heavy metals with suction cups materials were summarized by Wenzel and Wieshammer (1995). Heavy metals are specifically sorbed to metal hydroxides. Porous oxide ceramics are therefore only suitable for determining concentrations of heavy metals after long equilibration. This state may not be reached if low concentrations are to be detected. Especially Cu and Pb were sorbed by P80 ceramic cups in tests of Wenzel et al. (1997). Consequently, often membranes or porous bodies of organic polymers are recommended for sampling of trace metals. Not all suction cups, however, made from organic polymeric materials are equally suited for collecting trace metals. Rais et al. (2006) reported that cups made from a mixture of PTFE and a silicate effectively adsorbed Cu and Pb. Copper was also adsorbed by a pure organic polymer used for the production of micro suction cups (Rais et al., 2006). Stainless steel contains a large number of heavy metals (e.g., chromium, vanadium) and should not be used for this reason. More suitable materials are PE, PTFE, and nylon (Wessel-Bothe et al., 2000). The speciation of heavy metals depends on the pH value. Changes in the pH value due to outgassing of CO2 therefore can influence the measured heavy metal concentrations. Furthermore, in the course of a precipitation of carbonates or oxides a coprecipitation of heavy metals in the measuring system may occur (Schwarz and Miehlich, 1993). Heavy metals are complexed by dissolved organic compounds, which increase their overall concentration in the soil solution. Suitable sampling systems should therefore not sorb or release dissolved organic matter in large quantities. In tests performed by Rais et al. (2006) the presence of dissolved organic matter commonly reduced the adsorption of trace metals by various suction cup materials.
Organic Compounds (Pesticides, Organic Pollutants, Xenobiotics, Dissolved Organic Carbon, Organically Bound Nutrients)
Pesticides, xenobiotics, and organic pollutants such as polycyclic aromatic hydrocarbons (PAHs), and polychlorinated biphenyls (PCBs) display a wide range of physicochemical properties (e.g., water solubility and vapor pressure). It is therefore difficult to make universal and comprehensive recommendations for this group of substances. Depending on their water solubility, the majority of substances are in partitioning equilibrium between organic and aqueous phase. Plastics, adhesives, and elastomers absorb organic compounds and are not well suited for the construction of sampling systems for these substances. Glass (Wessel-Bothe et al., 2000) and stainless steel are more suitable.
Depending on the vapor pressure of the compound, volatilization from the sample of soil solution may occur. In this case, dual-chamber systems should be used (Wood et al., 1981; Suarez, 1986, 1987).
Dissolved organic carbon (DOC), or more precisely dissolved organic matter (DOM), is specifically adsorbed to metal hydroxides via its carboxyl and hydroxyl groups. Many ceramic materials therefore sorb considerable quantities of DOM and are only suitable for determining DOC concentrations after a long period of equilibration (Guggenberger and Zech, 1992; Wessel-Bothe et al., 2000). Porous glass (Wessel-Bothe et al., 2000) or stainless steel is more suitable. Adhesives, glues, and elastomers used for the construction of suction cups and plates contain solvents and plasticizers, which are released into the sampled soil solution and may increase the measured DOC concentration (Siemens and Kaupenjohann, 2003). The sampled solution should therefore be protected from contact with adhesives or elastomers. Stainless-steel capillaries and PTFE tubes are thus more appropriate as sampling device material. Since organically bound nutrients are naturally subjected to the same processes as DOM, it is also recommended that glass, stainless steel, and PTFE should be used.
Preferential flow represents the most significant transport mechanism for strongly sorbing pesticides, organic pollutants, DOM, and organically bound nutrients (Jene, 1998; Kaiser et al., 2000; Qualls, 2000, Vereecken, 2005). For this reason, large sampling systems such as lysimeters, wick samplers, or suction plates are appropriate for measuring representative concentrations and fluxes.
Colloids and Microorganisms
A major problem of sampling colloids and microorganisms is the size exclusion by filtration (Bell, 1974; Dazzo and Rothwell, 1974). In most cases, colloids, bacteria, and viruses are therefore sampled with zero-tension systems (Kaplan et al., 1993, 1996; Thompson and Scharf, 1994) or systems with coarse pores (pore diameter >16 µm) (Schäfer et al., 1998). Cigány et al. (2005) tested the suitability of fiberglass wicks for sampling various kinds of specific colloids (feldspathoids, ferrihydrite, montmorrillonite, and kaolinite) and of a mixture of colloids extracted from coarse sand at various flow rates and pH under unsaturated conditions. The permeabilty of the wicks for colloids depended strongly on pH, flow rate, and type of colloid. The authors conclude that colloids can be retained in the wicks under many conditions, so that their use for colloid sampling should be considered with caution. Similarly, Shira et al. (2006) reported a considerable retention of negatively charged silica microspheres and ferrihydrite in fiberglass wicks. They concluded that wicks are probably suited to characterize mobile colloids in the vadose zone qualitatively but could not be expected to sample colloid concentrations quantitatively. Ilg et al. (2007) compared the efficiencies of a zero-tension system, membranes of various mesh size, glass fiber wicks, and porous glass plates for sampling radioactively labeled goethite in a column experiment. They concluded that nylon membranes of 16 µm mesh and zero tension are best suited for sampling goethite colloids. Overall, there is still limited experience with respect to sampling of colloids under unsaturated conditions. Retention of colloids in sampling systems strongly depends on the type of colloids, and geochemical and hydrodynamic conditons that prevail. Hence, individual testing of colloid retention in the sampling system to be used seems inevitable.
Conclusion
Our review of the literature published over the last decades concerning the optimal setup of in situ soil water extraction reveals a broad spectrum of potential artifacts but also numerous ways to minimize them. In general, it seems difficult and perhaps impossible to obtain pore water samples which are not altered or biased by the sampling process. Nevertheless, the review should give advice to minimize alteration and misinterpretation of the achieved data sets. It is imperative to exactly define the goals of planned experiments to make choices in which bias or alterations can be accepted. Cost considerations, however, will dictate a point at which increased sample representativeness is not practical. At any point, all possible alterations and known measurement errors should be well documented during the course of the experiment to trace back biased data. As Dorrance et al. (1991) already stated, the early literature was mainly focused on the alteration of the sampled soil solution by chemical process, and that these alterations might be less significant than the variabilities in extracted soil water concentration due to soil physical heterogeneities and preferential flow. Recent research on the effect of spatial heterogeneity on soil water composition showed that even unaltered soil water samples should only be viewed as temporally representing the sample location and not spatially representing any other point in space or time. Conversely, the application of all the discussed techniques is more or less restricted to small observation areas such as point, plot, or small field scale and it is too time-consuming for soil solution and transport monitoring of larger areas which are in the focus of recent research interest. Clearly, there is a gap between the two scales of interest for regional and global resolutions. At intermediate spatial scales, such as agricultural fields or small catchments, reliance on sparse point measurements might not provide the accurate solute concentration and transport information required at these scales (e.g., for ground water protection). Therefore, there is a need for soil solution measurement and modeling techniques that can provide dense and accurate information at this intermediate scale with a high temporal resolution.
These limitations should not prevent researchers from using in situ extraction systems but illustrate that a combination of in situ sampling with non-invasive geophysical methods or advanced modeling techniques might be required to greatly increase the representativeness of information regarding soil water quality and solute transport.
NOTES
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  E. Bloem, F. A. N. Hogervorst, G. H. de Rooij, and F. Stagnitti Variable-Suction Multicompartment Samplers to Measure Spatiotemporal Unsaturated Water and Solute Fluxes Vadose Zone J., February 5, 2010; 9(1): 148 - 159. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Dahan, R. Talby, Y. Yechieli, E. Adar, N. Lazarovitch, and Y. Enzel In Situ Monitoring of Water Percolation and Solute Transport Using a Vadose Zone Monitoring System Vadose Zone J., November 17, 2009; 8(4): 916 - 925. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Zhu, T. Wang, F. Kuang, Z. Luo, J. Tang, and T. Xu Measurements of Nitrate Leaching from a Hillslope Cropland in the Central Sichuan Basin, China Soil Sci. Soc. Am. J., June 29, 2009; 73(4): 1419 - 1426. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Tuli, J.-B. Wei, B. D. Shaw, and J. W. Hopmans In Situ Monitoring of Soil Solution Nitrate: Proof of Concept Soil Sci. Soc. Am. J., March 1, 2009; 73(2): 501 - 509. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. W. MacDonald, R. R. Rediske, B. T. Scull, and D. Wierzbicki Landfill Cover Soil, Soil Solution, and Vegetation Responses to Municipal Landfill Leachate Applications J. Environ. Qual., August 8, 2008; 37(5): 1974 - 1985. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
