HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by Caron, J. Articles by Trépanier, L. Search for Related Content PubMed Articles by Caron, J. Articles by Trépanier, L. GeoRef GeoRef Citation Agricola Articles by Caron, J. Articles by Trépanier, L.

DIVISION S-1-SOIL PHYSICS

Field Bromide Transport under Transient-State

Monitoring with Time Domain Reflectometry and Porous Cup

^a Département des Sols et de Génie Agroalimentaire, FSAA, Université Laval, Québec, QC, Canada, G1K 7P4
^b Khrystal Engineering, 37, Avenue Aboulbaba El Ansari, 1004 El Menzah VI, Tunis, Tunisie

jean.caron{at}sga.ulaval.ca

	ABSTRACT

TOP ABSTRACT INTRODUCTION Materials and methods Results and discussion Conclusions REFERENCES

 
The contamination of groundwater by excess fertilizer and pesticides is a problem associated with modern agricultural practices. Estimating fluxes of these contaminants to groundwater requires frequent soil water sampling. Time domain reflectometry (TDR) techniques appear well-suited to this purpose because of their potential for automation and the limited calibration work required, despite potential constraints associated with probe geometry, temperature, and the nonspecificity of the probe. The objective of this study was to compare the performance of the TDR technique with porous cup samplers for estimating solute masses and concentrations in a soil at the lysimeter scale. Potassium bromide was applied at the soil surface in a 0.91 by 0.52 m area and solute was sampled with TDR and porous cups. Recovered Br^- masses were calculated based on measured water contents and Br^- concentrations. When the solute was concentrated in the top 0 to 45 cm of soil, the biases for solute mass reached 2.6 and 5.3 times the applied mass for TDR and the porous cup samplers, respectively. As the solute spread out below this depth, the bias with the TDR technique decreased to between 1.17 and 1.27 times the applied mass, whereas for the porous cup samplers, it varied between 0.97 and 1.83 times the applied Br^- mass. Differences in soil structure appeared the most likely explanation for the bias. The study also indicates that field calibration of TDR from porous cup samplers may be difficult to achieve under transient state conditions.

Abbreviations: TDR, time domain reflectometry

	INTRODUCTION

TOP ABSTRACT INTRODUCTION Materials and methods Results and discussion Conclusions REFERENCES

 
ANNUAL FLUXES of agricultural pollutants to the groundwater must be estimated in order to evaluate the impact of agricultural practices on water quality. Solute fluxes to the groundwater can be predicted by fitting models to observed volume averaged concentration data, referred to as resident concentration [C_v(z,t)], or to flux averaged concentration data [C_f(z,t)], and mass transport calculated from these fitted parameters. Relevance of such modelling approaches is based on many assumptions which sometimes do not hold true under a transient state where conditions for multiphase flow exist. Alternatively, fluxes can be evaluated through frequent monitoring and summed across time intervals to estimate mass transfer without an inferred transport model. Different sampling devices have been used to evaluate these fluxes. They can be classified as direct or indirect evaluation methods. Direct evaluation involves tile drainage and field-installed pan lysimeters in which C_f(z,t) and the solute fluxes to a given depth are determined. Indirect methods involve frequent monitoring of either C_f(z,t) or C_v(z,t) and the corresponding soil volumetric water contents in order to evaluate mass variation within the unsaturated zone, which is then coupled to a field water budget to evaluate solute fluxes to the groundwater (Ben Jemia et al., 1997). The devices most frequently used for mass balance evaluations with these indirect methods are porous cup samplers (McGuire and Lowery, 1994) and soil cores.

In unsaturated conditions, the liquid phase in the soil is often classified as mobile (fraction involved in transport) and immobile (fraction involved in storage). While it is obvious that C_v(z,t) is sampled by soil coring, the exact nature of the concentration type obtained with porous cups is not exactly known, such concentrations being sometimes interpreted as flux concentrations and sometimes as resident concentrations (Vanderbrogth et al., 1996). However, both concentrations are functionally linked, for a convective dispersive process, through the following relationship (Jury and Roth, 1990):

(1)

where D is the dispersion coefficient and V is the velocity parameter. Whether it is a flux, a concentration, or a combination of both that is obtained therefore depends on the relative magnitude of the dispersion vs. the convection processes and on the importance of the concentration gradients. Since the concentration gradient tends toward zero as solute spreads out, then the porous cup sampled concentration is likely to become closer to that obtained by soil coring with depth. Because of these differences, the chemical composition of the samples obtained from porous cup samplers depends on the liquid volumes drawn from both liquid phases (Grossmann and Udluft, 1991; Cochran et al., 1970; Zabowski and Ugolini, 1990) and, not surprisingly, Severson and Grigal (1976) showed that the solution concentration obtained from porous cup samplers often differs from the leaching solution. This is therefore a shortcoming associated with porous cup samplers. They have other shortcomings as well. Their use is limited to soil water potential higher than -80 to -100 kPa, and frequent water extraction can alter the natural pattern of flow in the soil. They are also unsuitable for measuring the preferential flow paths (Roth et al., 1990).

Because of these limitations, soil coring may be a more appropriate method for the estimation of preferential flow and mass balance (Roth et al., 1990). However, soil coring is both time- and resource-consuming because the cores must be obtained and the solution extracted through centrifugation or filtration. Moreover, the sampling procedures are destructive and the newly created boreholes can significantly modify the flow pattern. The estimation of flux averaged concentrations is further complicated by the sampling of both liquid phases. As with porous cup samplers, soil coring requires travelling to the site and frequent sampling to estimate fluxes (Ben Jemia et al., 1997). Automation of the estimation of soil resident concentration would be highly advantageous and the TDR technique may offer this possibility. The TDR technique (Topp et al., 1980) has been used to determine the in situ solute resident concentration (Dalton and van Genuchten, 1986; Ward et al., 1994) for solutes for which changes in concentration cause significant changes in the soil electrical conductivity. This technique is suited to the determination of mass transport in soil because it allows the simultaneous determination of water content (needed for the estimation of drainage water fluxes) and solute concentrations in large soil volumes. Moreover, the TDR technique has no cost associated with the analytical determination (Ward et al., 1994), and it is nondestructive, fast, and easily automated. Time domain reflectometry also provides an important advantage for spatial characterization. Indeed, solute transport studies are caught with spatial variability issues, having to deal with a proper characterization of the field scale variability of solute transport parameters (velocity, dispersion, and percentage of mass recovered and transferred), themselves estimated from sampling devices having different sampling volumes (Van Wesenbeeck and Kachanoski, 1991). For a constant distance between sampling zones, small volume samples may show more variability than large volume samples, hence overestimating the true field scale variance characterizing a given parameter at that same sampling scale, a concept referred to as the support effect (Matheron, 1965). Thus, relative to porous cups and small core samplers, horizontally placing TDR probes appears to be a clearly superior technique as it can integrate horizontal scale variations in transport properties (Ward et al., 1995) and sample a larger volume.

However, the technique also has important limitations. It is obviously inappropriate for pesticides and interactive compounds. Even for conservative tracers, it requires a correction for temperature (Heimovaara et al., 1995) and is nonspecific to ionic species, which makes it sometimes sensitive to the anion exchange phenomenon (Vogeler et al., 1996). Separate calibration curves are also needed for specific probe geometry and soil horizons. This problem is easily overcome under steady state, as Ward et al. (1994) have presented simple methods to rapidly obtain specific calibration curves. However, for transient-state conditions, independent calibrations at different water contents and tracer concentrations have to be made either in the field or in the laboratory, possibly at different temperatures (Perrson, 1997). Field site calibration can be performed but imposes important time and material constraints for performing the in situ calibration procedure for each specific probe (Risler et al., 1996; Perrson, 1997). Moreover, the accuracy obtained by such additional work may be of a very limited value when the soil surface is tilled or subject to temporal changes, for example as a result of spring thawing. Alternatively, laboratory calibration requires core sampling or soil repacking, which also results in soil disturbance and increased uncertainty (Mallants et al., 1996) but certainly appears the least time consuming procedure or sometimes the only valuable alternative for field investigations on sites subjected to important bulk density or soil structure changes.

Moreover, under field transient state conditions, the TDR technique must realistically be used with a limited number of electrodes and calibration procedures, which may then very well limit its efficiency. This study investigates the comparative performance of TDR with limited calibration and porous cup samplers, for properly characterizing Br^- mass transport at a lysimeter scale.

	Materials and methods

TOP ABSTRACT INTRODUCTION Materials and methods Results and discussion Conclusions REFERENCES

 
Field Experimental Design
In June of 1994, three draining pans (Fig. 1) consisting of a solid stainless steel pan (91.5 by 52 cm) were inserted at the bottom (100-cm depth) of an undisturbed soil block of sandy loam in St-Pierre, Île d'Orléans, QC, Canada. The soil was a L'Atrée series, consisting of 20 to 30 cm of loamy sand in the A_p horizon and sand from 25 cm downward (Marcoux, 1981). Bulk densities were found to be 1.24 Mg m^-3 for the 0- to 25-cm zone, 1.57 Mg m^-3 for the 20- to 40-cm zone, and 1.55 kg m^-3 for the 40- to 60-cm zone. With a hydraulic jack, the pans were inserted horizontally from an adjacent trench. One at a time, the four lateral surfaces of the blocks were exposed to air by digging a trench, and the exposed surface plastered with a bentonite mud and covered with a polyethylene plastic film to prevent preferential flow along the side or any water exchange between the soil block and the surrounding soil. At the bottom, the polyethylene film was glued to the pans with a silicon sealant. Therefore the soil block was completely insulated from the surrounding soil without being disturbed, and it constituted an independent lysimeter. Each soil block was assumed to be composed of four different layers measuring 22.5, 22.5, 30, and 25 cm in depth. The upper layer corresponded approximately to the plough layer while the three others represented layers of homogenous material. One TDR probe and two 2.4 cm diameter by 5.6 cm long porous cup samplers (no. 655X0-B1M1, Soil Moisture Equipment Corp., Santa Barbara, CA) were inserted horizontally into each soil layer for a total of 12 TDR probes and 24 porous cups. The zone of influence of the cups was determined from Eq. [2] of Hart and Lowery (1997), using the needed parameters for calculation from known experience and from characterization of the site (Majdoub, 1999). Estimate of the zone of influence radius ranged roughly between 0.5 and 3 cm, a value close to the average value of 2.5 cm reported by Hart and Lowery (1997). Consequently, that average value was used for comparison purposes. The zone of influence of TDR probes was determined by immersing the probes horizontally in a distilled water bath and then gradually dropping the water level by 1-cm decrements to determine the distance between the surface and the center of the probe at which the first detectable change in the transit time and the reflection coefficient occurred. The probes were rotated at a 90° angle and the same determination performed. Both numbers came out at 4 to 5 cm. Therefore, the sampled volume was larger than that of the porous cup. For the porous cup, the drawn volume roughly corresponded with (for the total 8 cups) 0.1% of the total lysimeter volume, considering the radius of influence and the length of the cup, while for the four TDR probes, it reached 4.5% of the same volume. The TDR probes were inserted horizontally in the center of the lysimeter and of the layer, while the cups were inserted perpendicular to the probe, at 15 cm from the probe center along the width of the lysimeter, that is outside of the probe zone of influence. Along the length, the two cups were located at a 30-cm distance from the external surface. The TDR probes were made of two stainless steel rods 0.64 cm in diameter, 84 cm long, and 5 cm apart connected to a parallel cable antenna with the connection sealed by an epoxy resin block. The probe cables were connected to an impedance pulse transformer (TP-103, VA-COM, Lowell, MA), hooked up to a Tecktronix 1502B via a multiplexer (SDMX50, Campbell Scientific, Logan, UT) and connected to a CR-10 datalogger (Campbell Scientific) for continuous determination of the volumetric water content at the different depths in the three lysimeters. T-type copper-constantan thermocouples (Campbell Scientific) were placed in the nearby soil profile to measure temperature at depths of 15, 30, 60, and 90 cm during the experiment. The system installation (and the calibration work highlighted below) was completed by the fall of 1994, after making sure that the data monitoring system and lysimeters were working properly.

View larger version (25K): [in this window] [in a new window]   Fig. 1 Schematic representation of the lysimeters with installed instrumentation

Direct Laboratory Calibration
The direct calibration procedure outlined by Ward et al. (1994) was followed for the TDR probe calibration on repacked soil columns. This calibration procedure, after the modifications outlined below, allowed conversion of the TDR signal to solute concentration in transient as well as in steady-state conditions, at any water content. Following the theoretical development of Ward et al. (1994), the change in the resistance at long travel times is related to a specific solute type concentration through:

(2)

where a() and b() are empirical constants relating resistance (R) to solute concentration (C). Constants a and b account for the geometry of the probe and the influence of the solute and soil type on the signal, while R₀ represents the initial soil electrical conductivity. For transient-state flow conditions, precise measurement of (1/R - 1/R₀) requires the adjustment of R₀ according to the soil water content because R₀ varies with soil water content. Equation [2] therefore becomes:

(3)

where R₀() is the soil water content–dependent resistance at each specific measurement depth, determined prior to any tracer application. The determination of 1/R₀() for a TDR probe in the lysimeter can be done using two methods, the first consists of taking different measurements of R at different in a lysimeter before applying the tracer , and obtaining R₀ through a regression of R₀ as a function of . The second method requires the determination of a general equation of R₀ as a function of in a laboratory and the use of this equation in the field based on limited field values of R₀(). The second method is more appealing since obtaining a wide range of R₀() values in the field may be time-consuming and difficult for the lower horizons. The second method also assumes that the slope of the R₀() relationship is linear and independent of the probe used, assumptions that were checked with the data set from this study.

The calibration of the relationship between R and the Br^- concentration was carried out in three steps. In the first step, the calibration of R₀() for different probes was performed. To determine R₀(), a polyvinyl chloride cylinder, 15 cm in diameter and 90 cm long, was filled with soil from a composite sample taken from six subsamples collected on the outside of each lysimeter during installation, at the 25- to 60- and 60- to 90-cm depths (two subsamples per lysimeter). The soil from the top surface was not taken, as it varied in bulk density during the year and represented only a small proportion of the whole sampling volume (20%). We acknowledged that it may have introduced an error, but it was later found to be negligible when leaching started (see below). The soil in the column was then compacted near the measured field bulk density (1.55 Mg m^-3). The column was placed horizontally and stabilized at a volumetric water content of 0.15 m³ m^-3. One after the other, the six probes were inserted into the cylinder and R₀ was determined. Water content was raised by 0.05 m³ m^-3 increments up to 0.35 m³ m^-3 by adding a known amount of water with a syringe whose needle was inserted into 40 equally spaced holes located along the cylinder. Measurements of R₀ with the different probes were taken after each increase once the readings showed constant values. Carefully removing and reintroducing the six different probes did not result in a significant change in the measured volumetric water content. The effect of the removal and reintroduction process was therefore considered to be negligible. The comparison between the regression equations was performed using the procedures outlined by Zar (1974).

In the second step, the determination of a() and b() in Eq. [3] was carried out on five soil cylinders similar to those described above. The soil of the five columns, also from a composite sample taken as above, was oven dried at 70°C and brought to a water content of 0.15 m³ m^-3 by injecting known volumes of KBr solutions of 300, 600, 900, 1200, and 1500 mg L^-1 of Br^- into each column. Syringes were used to inject KBr solutions into 40 equally spaced holes along the columns. The needle was completely driven into each hole and the solution slowly pushed out while gradually moving the needle from the soil cylinder. The TDR probes were inserted into the cylinders and the values of R and measured. Volumetric water content was then increased by 0.05 m³ m^-3 increments and additional R and determinations were carried out up to 0.35 m³ m^-3. Readings were recorded after making sure they showed constant values. To ensure that the sampling zone of the TDR probes remained within the limits of the cylinder, the TDR signals obtained from a sealed cylinder immersed in water were compared with those obtained from the same sealed cylinder in air. Since the volumetric water content as well as the signal attenuation measured on the TDR trace did not change with this change in medium, it was concluded that the probe sampling zone remained within the cylinder diameter, a conclusion confirming the observation made earlier in the TDR probe sampling volume.

In the third step, bulk soil electrical conductivity readings were corrected to a standard reference temperature of 22°C. A temperature compensation factor, referred to as , was measured at different temperatures in a soil column (0.20 m in diameter and 0.915 m long) filled with soil compacted near the field bulk density (1.55 Mg m^-3). The soil had previously been adjusted to a volumetric water content of 0.29 m³ m^-3 and a Br^- solute concentration of 200 mg Br L^-1, representing the average concentrations during the field experiment. A TDR probe similar to those installed in the lysimeters was placed in the soil column and the setup was installed horizontally in a temperature-controlled room where R was measured at 22, 15, 10, 5, and 1°C. The temperature compensation factor was the ratio between 1/R at 22°C and 1/R at the measured temperature.

The relationship of with temperature was fitted with a linear relationship, rated for the temperature range of 1 to 22°C, and was introduced in Eq. [3] to yield:

(4)

Bromide Masses
Bromide concentrations obtained by porous cup samples or TDR probes were calculated assuming a uniform concentration for each layer. Bromide stocks were then calculated from:

(5)

where SBRO is the stock of Br^- (mg) above the pan lysimeter, A is the surface area (cm²) of the lysimeters, _L is the volumetric water content of layer L, T_L is the thickness (cm) of the soil layer, CBRO_L is the Br^- concentration (mg cm^-3), and 1000 is a unit conversion factor. CBRO_L was either estimated from the R values measured with TDR or measured in the soil water samples obtained from the porous cups.

	Results and discussion

TOP ABSTRACT INTRODUCTION Materials and methods Results and discussion Conclusions REFERENCES

 
Direct Laboratory Calibration
As seen in Fig. 2 , the calibration curves were similar to those of Ward et al. (1994) and were linear at low concentrations (<900 mg L^-1) and curvilinear in the 900 to 1500 mg Br L^-1 range. Linear regressions between Br^- concentrations and (1/R - 1/R₀) were fitted for the range of 0 to 900 mg Br^- L^-1, which covered the entire area of interest for the field data from this study. Intercepts [b()] did not statistically differ from 0. The change in a() was modeled with a quadratic equation:

(6)

View larger version (21K): [in this window] [in a new window]   Fig. 2 Relationship between Br- concentration and changes in cable ohmic resistance (R) at different volumetric water contents. R0 represents the initial soil electrical conductivity, prior to solute application at the same water content

(7)

where ß_T is the intercept term.

View larger version (25K): [in this window] [in a new window]   Fig. 3 Variation of 1/R0 at different volumetric water contents for the six different rods used in the calibration procedure

(8)

By introducing Eq. [6] and Eq. [7] into Eq. [3], the soil solution Br^- concentration can be determined from and R measurements (Eq. [9]).

(9)

Equation [9] assumes that R_i measurements have been taken for each probe prior to any tracer applications, so that ß_T can be calculated from Eq. [7]. This form of equation disregards the variation of the intercept term value of Eq. [3], since this term was found to be negligible in this study. Finally, the temperature correction factor for our study is given in Fig. 4 . It presents the values obtained in the column calibration process and the results of the regression analysis. These data showed a linear decrease of the correction factor with increasing temperature. The slope was constant at all temperatures and 0.02°C^-1 for the 1 to 22°C to interval, consistent with estimates obtained by others when temperature approaches 25°C in that temperature range (Heimovaara et al., 1995; Robinson and Stoke, 1959, reported by Heimovaara et al., 1995). Since the soil temperature in the lysimeters ranged from 6 to 16°C during the experiment, the correction factor remained between 1.1 and 1.3.

View larger version (17K): [in this window] [in a new window]   Fig. 4 Temperature compensation factor obtained for five temperatures with the fitted regression equation

View larger version (27K): [in this window] [in a new window]   Fig. 5 Evolution of Br- concentrations according to TDR measurements and porous cups in all three lysimeters

View larger version (26K): [in this window] [in a new window]   Fig. 6 Estimation of Br- masses in all three lysimeters

(10)

where µ_t and _t are the mean and the variance of the distribution for a given time t (Jury and Roth, 1990). A constant k, representing the mass applied per unit area divided by the mean volumetric water content in the lysimeter, was introduced at the numerator of Eq. [10] to take into account that a known mass m was applied instead of a unitary mass pulse input and to convert Br^- liquid concentration into resident concentration data (Vanderbrogth et al., 1996). The fitting of Eq. [10] assumes quasi–steady state flow, uniform water content, and point estimation of Br^- concentrations from TDR readings, assumptions made for the purpose of the simulation only. The stochastic–convective model was used instead of the convective–dispersive equation because it better fitted the data. The curve obtained on Day 140 is reported again in Fig. 7b. The area under the curve on that day was calculated, giving an estimate of the Br^- sampled per unit area. The same calculation was performed using Eq. [5] using the concentration predicted by the model at the four specific sampling depths, and assuming the concentration uniform through all the sampled layer. The ratio of the modeled (area under the curve predicted with Eq. [10]) vs. the approximate mass per unit area (Eq. [5]) is referred to as the bias and reported on Fig. 7b. It is seen that on Day 140, calculation for Eq. [5] generates a bias of 1.07 relative to the exact mass predicted by the model.

View larger version (32K): [in this window] [in a new window]   Fig. 7 (a) Modelled depth distribution of Br- resident concentration determined using TDR on Day 140, 150, and 160 in Lysimeter B, using Eq. [10]. The least square estimated µ parameter is given between parentheses. (b) Predicted depth distribution of Br- concentrations by varying µ only. The ratio of the mass per unit area calculated from the surface under the curve to that calculated from Eq. [5], referred to as bias, is given for each simulation

Secondly, the presence of a bias may be attributed to changes in soil structure (Mallants et al., 1995). Assuming the relationship between bulk electrical conductivity (EC_a) and the electrical conductivity of bulk water (EC_w), itself linked to Br^- concentration, to follow (Rhoades et al., 1976),

(11)

then bias in mass estimation may come from an increase in EC^'_s because of sorbed cations (Vogeler et al., 1996). Since the predicted concentrations from TDR came back to nearly zero once the solute had passed the first horizon, this hypothesis found little support. This nearly zero concentration also showed that the absence of calibration for the top layer had no impact on the bias observed later. On the other hand, the bias may be due to the fact that tortuosity is changed (Mallants et al., 1996). Indeed, a change in the tortuosity, resulting from the differences in the soil structure of the repacked and the undisturbed soil, will cause a change in the slope of the calibration equation represented by T in Eq. [11] but not in the intercept. Therefore, the absence of difference in intercept after the solute had passed support a change in the slope (and therefore of tortuosity) for the cause of overestimation instead of a change in the intercept (as a result of a change in EC^'_s).

Bromide Mass Transfer Variance
For all lysimeters, Br^- had started to leached out on Day 150. Therefore, an estimate of the true variance of the Br^- mass at the lysimeter scale could be obtained and was calculated from the Br^- fraction left in the profile. For that period, the variance was found equal to 440 mg² (Table 1). In comparison, the estimated variance for the TDR was of the same order of magnitude, being 283 mg². In contrast, this estimated variance for the mass estimated from porous cup sampling (3967 mg²) was significantly higher to that of the lysimeter at (with only two degrees of freedom for both the numerator and the denominator) and to that of the TDR at . This suggests that horizontally placed electrodes may adequately estimate the lysimeter scale variance, while the porous cup sampler cannot. Therefore, porous cup sampling increased the mass estimated variance, which appears linked to a smaller sampled volume, an effect known in geostatistics as the support effect (Matheron, 1965).

Practical Implications
One implication for research on solute transport studies is that since the average bias in Br^- mass decreases as the solute spreads out, the zone of influence of the probes must be known and the probes placed in such a manner as to completely sample the different soil zones. This is particularly important to ensure the precision of the probes located close to the surface, but becomes negligible as solute spreads out. Therefore, this may be important for preferential flow study, where a large fraction of the solute remains close to the surface while a small fraction is rapidly leached out of the sampling zone. However, this will have a negligible influence when leaching occurs with an important spreading of the solute through all sampled layers. While for close-to-surface investigations vertical positioning of the probes would solve the probe resolution problem, horizontal probes provide a better assessment of the lysimeter scale variance in mass transfer, since they take into account the interactions between transport zones with different velocities (Ward et al., 1995). Data suggest that TDR probes can provide estimates of the true mass transfer variance at the lysimeter scale, free from the support effect associated with differing sampling volumes.

The results on the bias in Br^- mass also show that a direct laboratory calibration procedure can be successfully used to estimate on-site solute concentrations, within 21% of accuracy limit, as long as the horizontal location of the probes is adequate. The results on bias are in accordance with those of Hart and Lowery (1998), who observed biases in mass (calculated from recovery) between 0.65 and 1.21 times the applied mass. This finding is also in agreement with those of Mallants et al. (1996) and suggests that this bias might be attributed to soil structure changes. Therefore, on sites where calibration are difficult to obtain because of changes in soil structure or because the time required to reach equilibrium is too long for an in situ calibration, biased estimate of solute masses can be obtained with a laboratory calibration. As this bias seems to be related either to the freezing–thawing or the recompaction process, further work should aim at evaluating how this bias can be taken into account from a limited number of measurements, so that laboratory calibration could still be used on site with a greater accuracy. The results on R₀() measurements show that for field calibration (Risler et al., 1996; Perrson, 1997), simple measurements of the resistance prior to the solute application allow a single calibration for a whole set of probes.

The comparison of TDR with porous cup samplers is of importance since many studies have shown a good correlation between both techniques under steady-state conditions. However, in this transient-state study, the TDR technique presents less bias in Br^- mass than the porous cup sampler technique. The lack of relationship between the two techniques shows that it is difficult to base in situ calibration of the TDR probes on the porous cup sampler solutions under transient conditions, and that a prolonged period of equilibrium is required for such a calibration (Perrson, 1997).

Finally, the relatively high accuracy of the TDR technique under transient-state conditions is of interest in solute transport studies, as the estimation of solute fluxes from a combination of the TDR technique and solute concentrations requires the frequent measurement of both the volumetric water content and the solute concentrations (Ben Jemia et al., 1997). In such a case, a multiplex TDR probe system could allow the simultaneous collection of both R and at a high sampling frequency, an impossible procedure with porous cup samplers.

	Conclusions

TOP ABSTRACT INTRODUCTION Materials and methods Results and discussion Conclusions REFERENCES

 
This study showed that Br^- mass estimated from horizontally placed electrodes presented less bias than those estimated from porous cup samples. This bias for TDR was most likely a result of soil disturbance resulting from the laboratory calibration performed on repacked columns. It also showed that the variance estimated from these probe readings within a lysimeter was of the same order of magnitude as that obtained for the whole lysimeters, while that obtained from porous cup sampling was 10 times higher.

	ACKNOWLEDGMENTS

 
The authors are grateful to Daishowa Inc. and to NSERC for providing financial support. Thanks are extended to A. Ward and three anonymous reviewers for comments on the early and final parts of this manuscript, G.C. Topp for helpful discussions, and N. De Rouin for reviewing the manuscript. Special thanks are also expressed to G. Thériault, R. Chabot, and S. Cliche for assistance in both field and laboratory experiments.

Received for publication June 8, 1998.

	REFERENCES

TOP ABSTRACT INTRODUCTION Materials and methods Results and discussion Conclusions REFERENCES

 

• Ben Jemia S., Gallichand J., Caron J. Field water and solute fluxes: optimizing sampling frequency and calculation procedures. Trans. ASAE 1997;40:1623-1630.
• Cochran P.H., Marison G.M., Leaf A.L. Variation in tension lysimeter leachate volumes. Soil Sci. Soc. Am. J. 1970;34:309-311.[Abstract/Free Full Text]
• Dalton F.N., van Genuchten M.Th. The time-domain reflectometry method for measuring soil water content and salinity. Geoderma 1986;38:237-250.
• Grossmann J., Udluft P. The extraction of soil water by the suction cup method: A review. J. Soil Sci. 1991;42:83-93.
• Hart G.L., Lowery B. Axial-radial influence of porous cup soil solution samplers in a sandy soil. Soil Sci. Soc. Am. J. 1997;61:1765-1773.[Abstract/Free Full Text]
• Hart G.L., Lowery B. Measuring instantaneous solute flux and loading with time domain reflectometry. Soil Sci. Soc. Am. J. 1998;62:23-35.[Abstract/Free Full Text]
• Heimovaara T.J., Focke A.G., Bouten W., Verstaten J.M. Assessing temporal variations in soil water composition with time domain reflectrometry. Soil Sci. Soc. Am. J. 1995;59:689-698.[Abstract/Free Full Text]
• Jury W.A., Roth K. Transfer functions and solute movement trough soil: Theory and application. Berlin: Birkhaüser Verlag Basel, 1990.
• Mallants D., Vanclooster M., Toride N., Vanderborght J., van Genuchten M.Th., Feyen J. Comparison of three methods to calibrate TDR for monitoring solute movement in undisturbed soil. Soil Sci. Soc. Am. J. 1996;60:747-757.[Abstract/Free Full Text]
• Majdoub R. Modeling solute transport in a sandy soil using finite difference and element model. Sainte-Foy, QC, Canada: Département des Sols et de génie agroalimentaire. Université Laval, 1999 Ph.D. thesis..
• Marcoux R. Étude pédologique de l'Île d'Orléans, aux Coudres et aux Grues. Fisheries and Food, Canada: Québec Ministry of Agriculture, 1981.
• Matheron G. Les variables régionalisées et leur estimation. Paris: Masson, 1965.
• McGuire P.E., Lowery B. Monitoring drainage solution concentrations and solute flux in unsaturated soil with a porous cup sampler and soil moisture sensors. Ground Water 1994;32:356-362.[ISI]
• Perrson M. Soil solution electrical conductivity measurements under transient conditions using time domain reflectometry. Soil Sci. Soc. Am. J. 1997;61:997-1003.[Abstract/Free Full Text]
• Rhoades J.D., Raats P.A.C., Prather R.J. Effects of liquid phase electrical conductivity, water content, and surface conductivity on bulk electrical conductivity. Soil Sci. Soc. Am. J. 1976;40:651-655.[Abstract/Free Full Text]
• Risler P.D., Wraith J.M., Gaber H.M. Solute transport under transient flow conditions estimated using time domain reflectometry. Soil Sci. Soc. Am. J. 1996;60:1297-1305.[Abstract/Free Full Text]
• Roth K., Flühler H., Jury W.A., Parker J.C. Field-scale water and solute flux in soils. Berlin: Birkhaüser Verlay, 1990.
• Severson R.C., Grigal D.F. Soil solution concentrations: Effect of extraction time using porous ceramic cups under constant tension. Water Resour. Bull. 1976;12:1161-1169.
• Topp G.C., Davis J.L., Annan A.P. Electromagnetic determination of soil water content: measurements in coaxial transmission lines. Water Resour. 1980;16:574-582.
• Van Wesenbeeck I.J., Kachanoski R.G. Spatial scale dependence of in situ solute transport. Soil Sci. Soc. Am. J. 1991;55:3-7.[Abstract/Free Full Text]
• Vanderbrogth J., Vanclooster M., Mallants D., Diels J., Feyen J. Determining convective lognormal solute transport parameters from resident concentration data. Soil Sci. Soc. Am. J. 1996;60:1306-1317.[Abstract/Free Full Text]
• Vogeler I., Clothier B.E., Green S.R., Scotter D.R., Tillman R.W. Characterizing water and solute movement by time domain reflectometry and disk permeametry. Soil Sci. Soc. Am. J. 1996;60:5-12.[Abstract/Free Full Text]
• Ward A.L., Kachanoski R.G., Elrick D.E. Laboratory measurements of solute transport using time domain reflectometry. Soil Sci. Soc. Am. J. 1994;58:1031-1039.[Abstract/Free Full Text]
• Ward A.L., Kachanoski R.G., von Bertoldi A.P., Elrick D.E. Field and undisturbed-column measurements for predicting transport in unsaturated layered soil. Soil Sci. Soc. Am. J. 1995;59:52-59.[Abstract/Free Full Text]
• Wraith J.M., Confort S.D., Woodbury B.L., Inskeep W.P. A simplified waveform analysis approach for monitoring solute transport using time domain reflectometry. Soil Sci. Soc. Am. J. 1993;57:637-642.[Abstract/Free Full Text]
• Zabowski D., Ugolini F.C. Lysimeter and centrifuge soil solution, seasonal differences between methods. Soil Sci. Soc. Am. J. 1990;54:1130-1135.[Abstract/Free Full Text]
• Zar J.H. Biostatistical analysis. Englewood Cliffs, NJ: Prentice Hall, 1974.

This article has been cited by other articles:

				Y. Coquet, C. Coutadeur, C. Labat, P. Vachier, M. Th. van Genuchten, J. Roger-Estrade, and J. Simunek Water and Solute Transport in a Cultivated Silt Loam Soil: 1. Field Observations Vadose Zone J., July 18, 2005; 4(3): 573 - 586. [Abstract] [Full Text] [PDF]

				M. O. Gasser, J. Caron, R. Lagace, and M. R. Laverdiere Predicting Nitrate Leaching under Potato Crops Using Transfer Functions J. Environ. Qual., July 1, 2003; 32(4): 1464 - 1473. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by Caron, J. Articles by Trépanier, L. Search for Related Content PubMed Articles by Caron, J. Articles by Trépanier, L. GeoRef GeoRef Citation Agricola Articles by Caron, J. Articles by Trépanier, L.