Using time-domain reflectometry to characterize shallow solute transport in an oak woodland hillslope in northern California, USA

Page 1

HYDROLOGICAL PROCESSESHydrol. Process. 16, 2921–2940 (2002)Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/hyp.1047

Using time-domain reflectometry to characterize shallowsolute transport in an oak woodland hillslope in

northern California, USA

Chris G. Campbell,* Masoud Ghodrati and Fernando Garrido†

Ecosystem Sciences Division, Department of Environmental Science Policy and Management, University of California, Berkeley, 151Hilgard Hall, Berkeley, CA 94720-3110, USA

Abstract:

The natural heterogeneity of water and solute movement in hillslope soils makes it difficult to accurately characterizethe transport of surface-applied pollutants without first gathering spatially distributed hydrological data. This studyexamined the application of time-domain reflectometry (TDR) to measure solute transport in hillslopes. Three differentplot designs were used to examine the transport of a conservative tracer in the first 50 cm of a moderately slopingsoil. In the first plot, which was designed to examine spatial variability in vertical transport in a 1Ð2 m2 plot, a singleprobe per meter was found to adequately characterize vertical solute travel times. In addition, a dye and excavationstudy in this plot revealed lateral preferential flow in small macropores and a transport pattern where solute is focusedvertically into preferential flow pathways. The bypass flow delivers solute deeper in the soil, where lateral flow occurs.The second plot, designed to capture both vertical and lateral flow, provided additional evidence confirming the flowpatterns identified in the excavation of the first plot. The third plot was designed to examine lateral flow and onceagain preferential flow of the tracer was observed. In one instance rapid solute transport in this plot was estimatedto occur in as little as 3% of the available pore space. Finally, it was demonstrated that the soil anisotropy, althoughpartially responsible for lateral subsurface transport, may also homogenize the transport response across the hillslopeby decreasing vertical solute spreading. Copyright 2002 John Wiley & Sons, Ltd.

KEY WORDS soil solute transport; hillslope runoff; hydrological pathways; measurement technology; time-domainreflectometry; tracers; electrical conductivity

INTRODUCTION

Non-point source (NPS) pollution has been recognized as a major contributor to the pollution of surfaceand ground water world-wide (Duda, 1993). As a result, there is a continuing need to develop methods toassess and manage NPS pollution. In order to model the production of NPS pollution, relationships betweenpollutant resident time and degradation of labile contaminants during transport must be understood (Fetter,1977; Verchot et al., 1997; ). Estimates of solute residence time are necessary as dilution and degradation ofpollutants can increase with time, depending on interactions with the soil matrix, resulting in lower pollutantconcentrations reaching streams. The residence time of a pollutant in a watershed is controlled by runoff rateand transport pathways initiated at the surface of the hillslope. Therefore, to better understand NPS pollutiontransport, the factors controlling hillslope runoff and throughflow (rapid downslope flow through hillslopes)must be determined. As the partitioning between surface runoff and subsurface flow occurs in the first fewcentimetres of the soil and at boundaries between soil horizons, detailed characterization of transport processesin hillslope soils is warranted.

* Correspondence to: Chris G. Campbell, Water Guidance and Monitoring Group, Lawrence Livermore National Laboratory, P.O. Box 808,L-627, Livermore, CA 94551, USA. E-mail: [email protected]† Present address: Centro de Ciencias Medioambientales, CSIC, Serrano 155 dup., E-28006-Madrid, Spain.

Copyright 2002 John Wiley & Sons, Ltd.

Page 2

2922 C. G. CAMPBELL, M. GHODRATI AND F.GARRIDO

The initial partitioning between runoff pathways begins at the soil surface with infiltration. Infiltration capac-ity depends on soil characteristics, moisture regime, rainfall intensity and surface cover, including vegetationand litter (Blackburn, 1975). Runoff pathways on hillslopes include: surface runoff, as Hortonian overland flowor as saturation overland flow (Dunne and Black, 1970), and throughflow, as subsurface flow (Turton et al.,1992) or rapid baseflow (Sklash et al., 1986). Subsurface flow may occur in a soil horizon with lower bulkdensity underlain by a more dense layer (Luxmoore et al., 1990; Wilson et al., 1990), or it may occur as flow insoil macropores (McDonnell, 1990; Tsuboyama et al., 1994). All of these processes occur at various temporalscales, which creates the complication of discriminating between old and new water that mix in macropores(McDonnell, 1990) and at perched water tables (Sklash et al., 1986). As a result, current understanding ofhillslope solute transport has been limited to watersheds in which these processes have been measured.

A technique commonly used to study water and solute transport in hillslope soils has been to dig a trench andcollect the throughflow moving out of a soil pedon (Dunne and Black, 1970; McDonnell, 1990; Turton et al.,1992; Tsuboyama et al., 1994; Newman et al., 1998). Some trenches have been left open to the atmosphere(Turton et al., 1992; Tsuboyama et al., 1994), others filled with gravel or sealed with cement walls (Dunne andBlack, 1970; McDonnell, 1990; Newman et al., 1998). This approach has been successful, but the requiredsemi-permanent infrastructure may be expensive and there is also a potential to change the drainage boundaryconditions and artificially influence transport processes in the measurement plot. This creates the need for insitu non-destructive measurement techniques to characterize throughflow in hillslopes.

Time-domain reflectometry (TDR) systems measure the transport of a salt tracer, in situ and at a relativelyhigh spatial and temporal resolution. These probes may be calibrated to measure both water content andbulk soil electrical conductivity (EC), which may be used to monitor the transport of a salt tracer (Daltonand van Genutchen, 1986; Kachanoski et al., 1992). The utility of TDR probes to measure solute transportin soils has been examined predominantly in repacked and intact laboratory column studies (Mallants et al.,1996, 1994; Persson and Berndtsson, 1998; Campbell et al., 1999). Field studies on level soils also have beenperformed and are becoming more common (Kachanoski et al., 1992; Rudolph et al., 1996; Jacques et al.,1998; Radcliffe et al., 1998; Garrido et al., 2001). The application of TDR for solute transport studies inhillslopes would be a non-destructive alternative to drainage pits, with the potential to collect data at multiplesampling locations.

We performed a series of solute transport studies in three plots of different designs on hillslopes in anoak-woodland–rangeland watershed. Our objectives were to:

1. examine the utility of TDR probes to measure solute transport in a moderately sloped clay loam soil;2. gather detailed measurements and observations of transport rate and pathways in a hillslope soil;3. characterize the spatial variability in solute transport in the plots.

MATERIALS AND METHODS

The experiments were performed on hillslopes at the University of California Russell Reservation located8 km north of Lafayette, in the Bay Area of northern California (Figure 1). Elevations in the watershed rangefrom 230 to 360 m above mean sea level, with slopes from 2 to 75%. The bedrock units consist of UpperMiocene to Lower Pliocene marine and non-marine sedimentary rocks, which probably are Briones Sandstone,Hombre Sandstone and Rodeo Shale formations belonging to the San Pablo and Monterey Groups (Wagner,1978). Soil depths on the hillslopes range from less than 1 m on the higher slopes (>40%) to as much as5 m on the lower slopes at 20 to 30% (Salve, 1995). Studies were performed on three different plots ina hillslope with Botella clay loam soil (mixed thermic Pachic Argixeroll) and slopes ranging from 25% to35%. The hillslope is a mosaic of oak woodlands containing coast live oak (Quercus agrifolia) and Californiabay (Umbellularia californica) and California annual grassland which historically has been grazed by cattleand horses. For the duration of the study the site was fenced to protect it from disturbance. The region has

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 3

CHARACTERIZING SHALLOW SOLUTE TRANSPORT 2923

SacramentoReno

CA

LIFOR

NIA

, USA

Happy

Valley

Rd.

Hwy. 13

Hwy. 24 Lafayette

Berkeley

San

Francisco

Bay

580

University of California

Russell Reservation

Happy ValleyRd.

Interstate 80

San Pablo and BrionesReservoirs

Approx. 8 km

Santa Rosa

San Jose

PacificOcean

Los Angeles

Las Vegas

Plot # 2 Plot # 3

Plot # 1

BARN

Russell Reservation Entrance

SanFrancisco

Figure 1. Location of the University of California Russell Reservation and the study hillslope and plots

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 4

2924 C. G. CAMPBELL, M. GHODRATI AND F.GARRIDO

Table I. Selected soil characteristics of Botella clay loam at the field site

Property Value Coefficient of variation (%)(n D 24)

Classificationa Fine loamy, mixed thermicPachic Argixerolls

Sand (%)b 37Ð2 25Silt (%) 43Ð8 18Clay (%) 19Ð0 37pH 6Ð45 NAElectrical conductivity (EC) (mS cm�1)c 0Ð55 NABulk density (g cm�3) 1Ð01 14

a Classification at the series level.b Using the pipette method from Methods of Soil Analysis (Gee and Bauder, 1986).c Soil EC from saturation extract in laboratory.

a mediterranean climate characterized by dry summers and wet mild winters (Salve, 1995). Selected soilcharacteristics from the sites are listed in Table I.

Time-domain reflectometry system theory and operation

Time-domain reflectometry is based on the transmission of an electromagnetic pulse through a coaxial cablewith a probe at the end, in contact with the soil (Nadler et al., 1991). The travel time and attenuation of theelectromagnetic pulse may be related to volumetric water content and bulk soil electrical conductivity (EC),respectively. The travel time of the pulse is used to estimate the apparent permitivity (Ka) from which watercontent can be calculated using Topp’s equation or a soil specific calibration (Campbell et al., 1999). In addi-tion, the attenuation of the pulse, impedance loading (ZL), is estimated from the wave form produced by theelectromagnetic pulse generator. The ZL value is related inversely to bulk EC of the soil. Given relationshipsbetween the bulk soil EC and the EC of the soil solution, salt solutions can be used as tracers by calibratingthe TDR to measure concentration.

Methods proposed for the calibration of ZL measurements by the TDR to solute concentration have beenexamined by Mallants et al. (1996). This study used a similar methodological approach to Campbell et al.(1999) and Garrido et al. (2001). Concentration was estimated by calibrating the ZL measurements fromthe TDR to the electrical conductivity of the solution in the soil pore space. First the TDR system wascalibrated for bulk EC measurements by immersing the TDR probes into solutions of known concentrationand fitting the ZL values to the equation proposed by Nadler et al. (1991) to estimate a geometric constant(Kc) value

ECbulk D KcZ�1L �1�

Using Equation (1) the TDR signal is calibrated to measure bulk soil EC. Bulk soil EC was linearly relatedto the soil solution EC using a two-point calibration performed by saturating the soil with two solutions ofCaCl2 (Kim et al., 1998). Although Mallants et al. (1996) suggest calibrating TDR probes at the point ofmeasurement, that was not a possibility in this study given the large number of probes and measurementlocations. Thus the final calibration was performed in the laboratory using intact columns collected fromthe plots. All 16 probes were installed in four large diameter (approximately 26 cm) intact columns forcalibration.

The relationship between solute concentration and soil solution EC is also linear within the range of tracerconcentrations used in this study (Campbell et al., 1999). Normalized resident concentration (Cr/Co) wascalculated by

Cr

CoD EC

ECoD [ECz,t � ECi]

[ECo � ECi]�2�

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 5

CHARACTERIZING SHALLOW SOLUTE TRANSPORT 2925

where ECz,t is the EC measured at time t at a given depth z, ECo is the EC of the initial concentration of thetracer pulse, and ECi is the baseline EC before and after the tracer application (Kachanoski et al., 1992). Thevalue of ECo was determined in the laboratory for each probe using large (26 cm diameter) intact undisturbedsoil columns collected from the plots after each experiment. In addition, TDR measurements may be sensitiveto ambient temperature, so a linear correction technique, which was similar to temperature corrections inlaboratory EC meters proposed by Heimovaara et al. (1995), was applied to the measurements. Soil and airtemperature were measured using HOBO Temperature probes (Onset Computer Corp., Bourne, MA).

The TDR system used in this study consists of three pronged TDR probes attached to a 16 channelmultiplexer by coaxial cables. The system is controlled and data collected by a computer and supportingsoftware (Dynamax Co., Houston, TX). The pulse generator, computer, and power source were housed nextto the plots in a tent with a portable heater for protection from cool temperatures typical of the end of thewet season (January to March). The plots were instrumented with 16 vertically installed 20-cm-long TDRprobes placed either at the surface measuring at 0 to 20 cm depth or at 30 cm measuring at 30 to 50 cmdepth (Figure 2).

Plot designs and tracer studies

The positions of the three plots on the hillslope are shown in Figure 1. Solute transport studies wereperformed in three plot designs under a steady-state water content (Figure 2). The first plot was designedto test the application of TDR to measure transport in the hillslope soil (Figure 2a). The TDR probes wereinstalled in a grid pattern to examine variation in the vertical movement of water and solutes in the soil. Inthis plot the tracer pulse was applied over the entire plot area and then leached. The second plot was designedto examine both vertical and lateral solute movement (Figure 2b). A line pulse was applied to the upslope endof the plot and leached while the downslope probes monitored the progression of the tracer. The third plotwas designed to examine lateral flow and its variability across the slope (Figure 2c). All plots were irrigatedfor more than 24 h during the rainy season after the soil was near storage capacity to ensure a steady-statewater content had been reached. Volumetric water content was checked by examining the apparent permitivity(Ka) measurements of the TDR probes.

The approach to the plot studies was to perform tracer miscible displacements or the application of a definedmass of tracer in solution that is pushed forward (or deeper) by another (background) solution (Jury et al.,1991). The temporal measurement of the tracer concentration is called a breakthrough curve (BTC). Underconstant water content and irrigation, the volume of water is proportional to time, which can be inferred usingthe pore volume, the volume of solution necessary to fill the entire pore space of the measurement volume(Jury et al., 1991). The BTC is a representation of the transport processes acting on the conservative soluteduring transport. As a result, the BTC is a baseline transport characteristic of the soil under the prescribedconditions.

Irrigation

Owing to differences in elevation and limitations of the irrigation system the plots were irrigated at ratesranging from 2Ð0 to 3Ð25 cm h�1. The irrigation system consisted of 8 to 16 RainbirdTM sprinklers attached toa pressurized hose connected to a water source. The pressure head from the hose was constant so the amountof water flow was dependent on elevation. This dependency was controlled as much as possible with a valveon the hosing, however, differences in irrigation rates did exist. As a result, the impact of three differentirrigation rates (3Ð1, 1Ð8 and 1Ð2 cm h�1) on transport was examined in the laboratory soil columns used forTDR calibration.

The irrigation was maintained at a constant rate, which was determined by measuring precipitation atfive points across the plot surface during the studies. The irrigation rate was 3Ð25 cm h�1, 2Ð9 cm h�1 and2Ð0 cm h�1 for plots 1, 2 and 3, respectively. The tracer was added to the irrigation water and appliedin the same manner. The steady-state studies at the end of the wet season simulate the time and ambient

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 6

2926 C. G. CAMPBELL, M. GHODRATI AND F.GARRIDO

c.

Muliplexer

1.2 m

1.0 m

0.5 m

0.2 m

Sprinklers

12a13a

15a 16a14a

7a5a 6a

8a1a 2a

4a

10a11a

9a

3a

a.

b.

0.5 m

0.5 m

2.0 m

3.0 m

3b2b1b

9b

7b5b 6b

10b

15b 16b

13b 14b

12b11b

8b

1.0 m

9.0 m

1c 2c 3c 4c 5c 6c 7c 8c

9c 10c 11c 12c 13c 14c 15c 16c

20-25%

20 cm Long

TDR Probes

ExcavationSteps

16 Channel

Tektronix 1502C

Pulse Generator

PortableComputer

Slope Class

25-30%

Tracer Application Area

Slope Class

25-30%

Slope Class

TDR at 0-20 cm depth

TDR at 30-50 cm depth Tracer Application Area

Figure 2. Study plot designs where (a) is plot 1, (b) is plot 2 and (c) plot 3, as referred to in the text

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 7

CHARACTERIZING SHALLOW SOLUTE TRANSPORT 2927

soil moisture conditions during which there would be the greatest risk for a surface-applied pollutant (e.g.fertilizers, pesticides, or contaminant spills) to runoff (as throughflow) and reach local water bodies.

Plot 1—vertical flow

In order to establish the variability in vertical solute transport in the plot, two consecutive miscibledisplacement studies were performed, each consisting of 5-cm pulses (tracer pulse volume D surface area ð5 cm, or c. 75 L) of 15 g L�1 CaCl2, applied to the plot and a 20 cm boundary area. Each pulse was appliedat the same irrigation rate and through the same sprinklers using a portable electric pump connected to thetracer solution container. After each pulse was applied, the irrigation system was reconnected to the originalhose containing background fresh water. During the first miscible displacement the computer connected tothe TDR stopped recording data (probably as a result of a power surge), however, as most of the BTC wasmeasured it was used to examine variability between consecutive BTCs measured in the same plot.

After the second pulse application of CaCl2 was completed, a third and final 6 cm pulse of red foodcolouring was applied to the plot in the same manner as the CaCl2. Once all the red dye was applied andallowed to redistribute (c. 15 min), the plot was excavated at 20 cm intervals upslope and to a depth of 50 cm.Observations of the presence or absence of the dye in the excavated soil profiles were drawn in a 10 by 10 cmcell grid on graph paper and photographs were taken following the procedures outlined in Ghodrati and Jury(1990). The observations, or traces, represent the dye present in each 100-cm2 cell after the soil was carefullycut away with a shovel. Structural features such as the large macropores and cracks, as well as visible lateralflow in the soil also were recorded.

Plot 2—downslope lateral flow

Next, to examine both vertical and lateral transport, the downslope 6-m2 plot was selected on the hillslopeadjacent to the first plot. The TDR probes were installed at two soil depths, measuring 0–20 cm and 30–50 cmin the soil. The probes were arranged in three equally spaced lines 1 m apart (Figure 2b). The probes at30–50 cm were installed by excavating the soil with a post hole digger to a depth of 30 cm and theninserting the prongs of the TDR into undisturbed soil. A layer of silica flour slurry was poured over theprobes to prevent preferential downward flow and the excavated soil was placed on top and packed firmly.The repacked soil was allowed to settle to its original level for more than a week in each case.

Once the plot was at steady-state water content under constant irrigation, the tracer solution was appliedalong a line at the top (uphill side) of the plot (Figure 2b). A total of 300 L of 15 g L�1 CaCl2 was appliedin a strip about 15 cm wide.

Plot 3—across-slope lateral flow

Finally, to specifically examine lateral flow, the site for the third plot (9 m2) was selected on the samehillslope as the second (Figure 1). The 16 TDR probes were installed in a line transect at two depths 0–20 cmand 30–50 cm. The deeper probes were installed in the same manner as in the previous plot. The probes werespaced 50 cm apart laterally with probes at alternating depths (Figure 2c). The tracer (300 L of 15 g L�1

CaCl2) was applied in a strip 50 cm upslope from the probes using the same irrigation system and electricpump described previously.

Data Analysis

The TDR system collected data at 10-min (plot 1) and 6-min (plots 2 and 3) intervals. The TDR data wereconverted to relative tracer concentration and plotted over time to construct solute BTCs. These BTCs, ascharacteristics of transport in this soil, are treated as statistical distributions that may be described using thefirst and second moments, or the times corresponding to the statistical mean and variation of the distribution(Jury and Roth, 1990). The BTCs in this case are used as transfer functions under stochastic-convection theory,

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 8

2928 C. G. CAMPBELL, M. GHODRATI AND F.GARRIDO

or the functional relationship describing the change in the input irrigation to produce an output response (Juryand Roth, 1990). This approach to solute transport in soil science is similar to the unit hydrograph conceptin surface water hydrology. Calculation of these moments and details on the transfer function approach maybe found in numerous references (Jury and Sposito, 1985; White et al., 1986; Jury and Roth, 1990; Valocchi,1990; Mallants et al., 1994; Persson and Berndtsson, 1998).

The BTCs from all the plots were then analysed by calculating the first (�) and second (�2) moments.Moment analysis is a method to quantitatively represent a statistical probability distribution using characteristicvalues (Valocchi, 1990). Although there are an infinite number of moments that may be calculated, the first twomoments are usually adequate to describe statistical distributions. Using the BTCs as probability distributionsof solute travel time in the soil, the first temporal moment (�) is

� D∫ 1

0t1f�t� dt �3�

where t is time on the x axis of the BTC. This temporal moment characterizes the mean displacement timeof the solute. The second temporal moment (�2) represents the spreading of the distribution along the x axis.This value is defined as

�2 D∫ 1

0�t � ��2f�t� dt �4�

where � is the first temporal moment and t is again time on the x axis (Mallants et al., 1994). Thereforewithin stochastic-convection theory, the BTCs are the relative amount of stream tubes transporting solute at agiven travel time during the miscible displacement. The moments therefore may be thought of as representingthe probability distribution of these travel times over the time axis.

In addition to the moment analysis, another characteristic of BTCs is the total mass of the tracer (or areaunder the probability distribution). The tracer mass commonly is presented as the mass measured versus themass applied, or mass recovery (M/Mo). The equation is

M/Mo D∑

[�Cr�z,t�/Co�Vw�t�] �5�

where Cr�z,t�/Co is calculated for Equation (2) and Vw�t� is the volume of water moving past the probe during

each sampling period (Campbell et al., 1999).Using the estimated moments, the variability in the transport parameters between measurement points was

examined in all the plots. As plot 1 was designed to provide preliminary information on transport variabilityand consistency, analysis was performed to determine the number of probes necessary to adequately representthe ensemble plot average within a given accuracy using �, �2 and M/Mo (Jury, 1985). In addition, a Spear-man’s Rank test was used to examine the temporal stability in the ranked responses measured between the twomiscible displacements in the first plot (Vachaud et al., 1985; Campbell et al., 2001). This test simply com-pares the ranked order of a data series for two different times or reproductions of the series. In this case, the testexamines if the order of � and �2 for all the probes in the first BTC is the same order measured in the secondBTC. The more similar the ranked values of the two BTCs, the closer to 1 the Spearman’s coefficient will be.

Specifically, Spearman’s test uses a rank Rij of the measured variable (in this case �ij and �2ij) and Rij0

the rank of the same variable at the same location (i) at a different time (j0). The Spearman’s rank correlationcoefficient is

rs D 1 �6

n∑iD1

�Rij � Rij0�2

n�n2 � 1��6�

where n is the sample size. If rs is close to 1, the variable is temporally stable. Critical rs values may befound in standard texts on statistics, below which the difference observed with time is considered significant(Daniel, 1995).

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 9

CHARACTERIZING SHALLOW SOLUTE TRANSPORT 2929

Various additional characteristics of the solute BTCs measured by the TDR probes may be used to quantifyand identify the occurrence of preferential flow in the soil. First the shape of the BTC and time to peak soluteconcentration may be compared with estimates using the convection (advection) dispersion equation. If theaverage time to peak concentration of a conservative tracer is faster than one pore volume, then the flow wasfaster than a well-mixed miscible displacement. In addition, although this was not observed in this study theBTC shape also may indicate preferential flow by having multiple peaks. This type of indicator was observedin BTCs from hillslope studies by Hornberger et al. (1991).

A third and more quantitative indicator of preferential flow may be used to estimate the mobile phase watercontent or the proportion of the water filled pores actively participating in transport (�st) (White et al., 1986).The resulting equation for a conservative tracer for BTCs measured under steady-state conditions is

�st D �tŁmedianJw�/L �7�

where tmedian is the median travel time, Jw is the constant input flux, and L is the travel length (L).

RESULTS AND DISCUSSION

Time-domain reflectometry system performance

The TDR system performed well both in the laboratory calibration columns and in the heterogeneoushillslope soils in all three study plots. The installation and maintenance of both the probes and the irrigationsystem was straightforward and the temperature correction had little influence on the overall shape of theBTCs. The tracer fronts used for calibration of the 16 TDR probes in the intact laboratory columns areshown for each irrigation rate in Figure 3. These cumulative BTCs demonstrate that although the systemis heterogeneous, measuring a consistent calibration point (ECo) is possible. Similar to Tsuboyama et al.(1994), no relationship was found between the three irrigation rates (3Ð1 cm h�1 in 3a, 1Ð8 cm h�1 in 3b, and1Ð2 cm h�1 in 3c) and BTC shape (Figure 3). It is also clear from the shape of the BTCs that some probesreach the maximum concentration faster than others, demonstrating that soil heterogeneity existed even inthese laboratory columns.

Plot design 1

In the first plot, the first miscible displacement appears similar to the complete second tracer pulse (Figure 4),although the data were incomplete owing to equipment failure. The average mass recovery was 86% for thecomplete tracer pulse test, with a range from 46% to 113% (Table II). The coefficients of variation (CVs) inthe mass recoveries (M/Mo) were 16% for the first BTC and 22% for the second. The BTCs, as used in thisstudy, represent the distributions of travel times required for the tracer to leave the measurement volume ofthe TDR probes. The low CVs demonstrate that the measurement volume of the probes is large enough toinclude plot-scale heterogeneity.

The EC of the initial tracer concentration was high enough to increase the bulk soil EC by a factor ofseven, when the soil was saturated with the tracer. This difference between the EC of the tracer and theinitial soil EC was large enough to dampen the background variability in soil EC. The BTCs resemble alog-normal probability distribution for most of the probes (Figure 4). The pattern of the tracer flow and TDRmeasurements appears consistent among probes on the time-scale of 20 to 40 h. Probes 1a, 5a and 6a hadM/Mo greater than the mean in both miscible displacements and probe 4 consistently measured low M/Mo.There are some discrepancies, such as probes 2a, 8a, 14a and 15a, where the concentration maximums of thefirst BTC were not as high as the second.

The estimated statistic, rs, from the Spearman’s rank test was 0Ð76 for the comparison of M/Mo betweenmiscible displacements. This value is closer to 1 than the critical rs value for this number of samples anddegrees of freedom (0Ð43). The test demonstrates that the rank of the mass recovered at all the probes

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 10

2930 C. G. CAMPBELL, M. GHODRATI AND F.GARRIDO

a.

b.

c.

1.0

0.8

0.6

0.4

0.2

0.00.00 1.67 3.33 5.00 6.67 8.33

Time (hrs)

C/C

oC

/Co

C/C

o

1.0

0.8

0.6

0.4

0.2

0.00.00 4.17 8.33 12.50

Time (hrs)

Time (hrs)

1.0

0.8

0.6

0.4

0.2

0.00.00 4.17 8.33 12.50 16.67

Figure 3. Cumulative breakthrough curves of CaCl2 measured in intact soil columns for TDR system calibration. Plot (a) contains probes 1to 5 performed at 3Ð1 cm h�1, (b) contains, probes 6 to 10 at 1Ð8 cm h�1, and (c) contains probes 11–16 at 1Ð2 cm h�1

in the first miscible displacement is not statistically different from the second. Consequently, the resultsmay be considered temporally stable as defined by Vachaud et al. (1985). Buchter et al. (1995) also foundtemporal consistency in one-dimensional solute transport studies on a stony soil monolith taken from ahillslope.

As the main goal in plot 1 was to examine spatial variability, the estimated number of probes required toobtain a mean within a percentage of the total mean 95% of the time was calculated for each of the transportparameters (Table III). The 16 measurement points used adequately described the M/Mo calculated from thedata at a greater than 90% confidence in the first miscible displacement and greater than 80% for the second.More than six probes per square metre would be necessary to characterize the mass recovery of the tracerwithin 20% of the true mean.

The estimated time to peak solute concentration, assuming a well-mixed miscible displacement (e.g.convection dispersion), under these boundary conditions, is 80 min. The average peak concentration for

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 11

CHARACTERIZING SHALLOW SOLUTE TRANSPORT 2931

020

040

060

080

010

0012

0014

000

200

400

600

800

1000

1200

1400

020

040

060

080

010

0012

0014

000

200

400

600

800

1000

1200

1400

020

040

060

080

010

0012

0014

00

020

040

060

080

010

0012

0014

00

020

040

060

080

010

0012

0014

00

020

040

060

080

010

0012

0014

00

020

040

060

080

010

0012

0014

00

020

040

060

080

010

0012

0014

00

020

040

060

080

010

0012

0014

00

020

040

060

080

010

0012

0014

00

020

040

060

080

010

0012

0014

00

020

040

060

080

010

0012

0014

00

020

040

060

080

010

0012

0014

000

200

400

600

800

1000

1200

1400

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

C/Co C/Co C/Co C/Co

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Pro

be

1aP

rob

e 5a

Pro

be

9aP

rob

e 13

a

Pro

be

14a

Pro

be

15a

Pro

be

16a

Pro

be

10a

Pro

be

11a

Pro

be

12a

Pro

be

6a

Pro

be

7a

Pro

be

8a

Pro

be

2a

Pro

be

3a

Pro

be

4a

Tim

e (m

in)

Tim

e (m

in)

Tim

e (m

in)

Tim

e (m

in)

Figu

re4.

Bre

akth

roug

hcu

rves

(BT

C)

ofC

aCl 2

from

plot

1,w

here

the

tria

ngle

sar

eth

efir

stin

com

plet

eB

TC

san

dop

enci

rcle

sar

eth

eco

mpl

ete

seco

ndB

TC

s

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 12

2932 C. G. CAMPBELL, M. GHODRATI AND F.GARRIDO

Table II. Transport parameters for the breakthrough curves (BTCs) from each pulse appli-cation on the first hillslope plot

Probe � �2 M/Mo

(normalized time) (dimensionless variable) (mass recovery)

BTC 1 BTC 2 BTC 1 BTC 2 BTC 1 BTC 2

1a 5Ð44 5Ð52 0Ð673 0Ð789 60 1092a 5Ð70 5Ð44 0Ð936 0Ð929 49 1033a 4Ð99 4Ð8 0Ð709 0Ð849 64 864a 4Ð91 4Ð76 0Ð661 0Ð713 42 465a 5Ð21 5Ð25 0Ð842 1Ð039 61 1026a 5Ð28 5Ð26 0Ð808 0Ð992 60 1007a 5Ð52 5Ð48 0Ð98 1Ð099 67 838a 5Ð06 4Ð77 0Ð666 0Ð791 56 1139a 5Ð02 5Ð19 0Ð744 0Ð898 57 9310a 5Ð54 5Ð53 0Ð884 0Ð767 51 8111a 5Ð42 5Ð58 0Ð743 0Ð733 53 6612a 4Ð95 4Ð75 0Ð743 0Ð678 54 5413a 5Ð06 4Ð96 0Ð831 0Ð733 66 7614a 5Ð74 5Ð16 0Ð745 0Ð900 37 7815a 5Ð59 5Ð22 0Ð824 0Ð720 42 8616a 5Ð30 5Ð32 0Ð704 0Ð825 56 96Mean 5Ð29 5Ð19 0Ð78 0Ð84 55 86SD 0Ð28 0Ð30 0Ð10 0Ð13 18Ð92 8Ð78CV(%) 5 6 12 15 16 22

Table III. The estimated number of samples required to producea mean with a probability of 95% that it lies within 5%, 10%

and 20% of the true mean from the first plot

Breakthrough Confidence � �2 M/Mo

curves limit (%)

1 5 5 26 4710 2 7 1220 1 2 3

2 5 7 31 8810 2 8 2220 1 2 6

the 16 probes occurred at 55 min, indicating that the solute was not well mixed throughout the total wettedpore space. Similar results have been found in other hillslope runoff studies (Hornberger et al., 1991; Turtonet al., 1992). This study design has the added advantage of quantitatively estimating the amount of pore spaceparticipating in solute transport (�st). The �st values calculated from the TDR BTCs ranged from 0Ð22 to 0Ð38,which are less than the 0Ð43 total wetted pore space. This demonstrates that in first 20 cm from the surfacethere is not complete mixing in the transport volume. However, flow is distributed over most of the availablepore space near the soil surface.

Dye study and excavation

The dye application and excavation revealed a more complete picture of the flow processes occurringduring the miscible displacement of the tracer in plot 1. Traces of the five consecutive soil cut faces recorded

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 13

CHARACTERIZING SHALLOW SOLUTE TRANSPORT 2933

as presence or absence of red dye in a 10 ð 10 cm grid are in Figure 5. The tracer entered the soil fairlyuniformly from the sprinkler application. From there it began to be focused into preferential flow pathways.Lateral flow was evident along preferential flow pathways during the soil excavation (dye gushing out of thecutfaces). This convergence of solute into a few pathways has been observed both in undisturbed soil columnsfrom hillslopes (Hornberger et al., 1991) and in other field studies on level (zero slope) soils (Flury et al.,1994, 1995; Ghodrati and Jury, 1990; Garrido et al., 2001).

In their study of a hillslope in the Bear Brook catchment of Maine, USA, Hornberger et al. (1991) pointedout that preferential flow occurred in flow pathways that could not be distinguished morphologically from thesurrounding soil matrix. Similar results were found in this investigation, as seen in Figure 5, cut 5. The crackmarked with ‘CR’ in the figure was only visible at the soil surface not deeper in the soil profile, however, itclearly produced preferential flow as deep as 40 cm into the soil. As mentioned by Hornberger et al. (1991),if preferential flow is not isolated to identifiable soil morphological characteristics, mechanistic modelling ofsolute transport may be difficult. However, BTCs measured with in situ probes such as TDR may be used forempirical transfer function modelling of solute transport in hillslopes.

Fluehler et al. (1996) described three zones in the soil profile, which included: (i) somewhat uniform flowbeing focused into preferential flow pathways in zone 1, (ii) the preferential flow zone 2, and (iii) distributedflow in zone 3, where dispersion and greater lateral spreading of solute occurs. Applying Fluehler’s zonesto the dye and excavation, it appears that transport in the beginning of zone 3 was measured by the TDRprobes in this study at 30 to 50 cm depth. It is in this zone where many small-diameter pores transportedthe tracer laterally in the hillslope. A schematic description of these zones is shown in Figure 6. The patternobserved in the field excavation follows Fluehler’s conceptual model of the zones. However, the pattern ismodified to include a larger lateral flow component, where vertical flow in the upper soil horizons flow intopreferential flow paths that carry the solute deeper in the soil profile to a dispersed lateral flow zone. Thepattern shown in Figure 6 has been described in several hillslope studies (McDonnell, 1990; Buchter et al.,1995; Newman et al., 1998; Noguchi et al., 1999). As suggested by Fluehler et al. (1996), identification ofconsistent flow patterns, as in Figure 6, is necessary to develop appropriate models. In this study and theseexamples, modelling water flow as predominantly one-dimensional either vertically or laterally may be anoversimplification.

Plot design 2

Using the information provided by the plot 1 design, a second design was used with the goal of measuringboth the lateral and vertical partitioning of a surface-applied solute. In the plot 2 design, tracer concentrationswere detected in the first row of TDR probes (probes 1b, 2b and 3b), but not in other probes placed in thefirst 20 cm of the soil (probes 5b to 10b) (Figure 7). Alternatively, tracer concentrations were detected by allthe probes at 30 to 50 cm depth (probes 11b to 16b). These results are in agreement with the flow patternidentified in the excavation study of plot 1 (Figure 6). The tracer moves vertically to a depth of about 50 cmand then flows laterally. In fact, excavation of two 100-cm-wide by 60-cm-deep segments in this plot revealedflow patterns similar to those shown in Figure 5 (data not shown).

Notice the large relative concentrations and longer residence times of tracer measured by the probes at 30to 50 cm depth directly under the tracer application (probes 11b and 12b, Figure 7). This is evidence that thesolute is rapidly moving vertically into this area, where it collects and is temporarily retained before movingout of the area as lateral flow. This type of lateral flow over a soil layer of lower permeability has beendescribed in other hillslope studies (Luxmoore et al., 1990; Wilson et al., 1990; Tsuboyama et al., 1994).However, in contrast to these studies lateral flow in this hillslope does not move predominately as matrix flowin the soil, but rather as preferential flow. As evidence of rapid lateral flow in macropores, the mass recoveredat the downslope probes at 30 to 50 cm depth was low, suggesting that the transport was only occurring in aportion of the volume-averaged measurement area of the TDR probe. As the TDR takes a volume-averagedmeasurement, the low mass recovery may result if a major portion of the flow moved beneath the TDR probes

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 14

2934 C. G. CAMPBELL, M. GHODRATI AND F.GARRIDO

CUT 1

0 10 20 30 40 50 60 70 80 90

10 D D D D D D D D DW

20 M5 D D D D

30

40 M1 M0.2 M0.2 M0.2 M0.2

50

CUT 2 -- 10 cm back up slope0 10 20 30 40 50 60 70 80 90

10 D D D D D D D D D ONLY 1st 5 cm

20 D M5 D D D many worm holes 0-50cm

30 Msm

40 Msm Msm Msm Msm Msm Msm

50

CUT 3 -- 20 cm back from 20 10 20 30 40 50 60 70 80 90

10 D D D D D D D D

20 D M7 D D D D

30 Msm Msm

Msm40 M1 M2

50

CUT 4 -- 20 cm back from 3

0 10 20 30 40 50 60 70 80 90

10 D D D D5 D5 D5 D D D

20 D D D D D D

30 D Msm Msm Msm

40 Msm Msm Msm Msm Msm

50

CUT 5 -- 20 cm back from 40 10 20 30 40 50 60 70 80 90

10 D D D D D D CR D

20 D D D D D CR CR

30 M10 D CR CR

40 Msm Msm Msm MsmCR

Msm Msm50 CR evidence of lateral old water flowin a crack and worm holes

Mx -- macropore with x diameter (sm-small<0.1) CR -- crackD -- distributed throughout area D5 -- only in the upper 5cm of area

Figure 5. Traces of the distribution of the red dye on the soil cut faces progressively upslope each 20 cm

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 15

CHARACTERIZING SHALLOW SOLUTE TRANSPORT 2935

Macropores of various sizes from worms to rodents

0 - 10 Similar to “Attractor Zone” cm Fluehler et al. (1996)

10 - 40 Similar to “Transmission Zone” cm Fluehler et al. (1996)

> 40 Similar to “Dispersion Zone” cm Fluehler et al. (1996)

Crack from soil swelling

More lateral macropore flow anddistribution of tracer above a more densesoil horizon

Figure 6. A conceptual model of transport pathways based on the dye study and related to flow zones described by Fluehler et al. (1996)

or the lateral flow is isolated to macropores comprising a small volume of the total pore space, resulting in asmaller EC signal.

The transport parameters for the probes are in Table IV. Examination of these parameters suggests thatin order for the tracer to reach the downslope distances of 100 and 200 cm at the rate measured, the activetransport volume (�st) must have been between 0Ð02 and 0Ð08. Thus, what the TDR probes are measuring atthese points is tracer that moved preferentially at average flow rates ranging from 6Ð8 to 27Ð9 cm h�1. Thisis a high rate compared with the irrigation flux of 2Ð9 cm h�1. The total porosity (�) was estimated from theaverage bulk density (�b), from Table I (� D ��b/�m C 1), where �m is the mineral density estimated to be2Ð65 g cm�3 (Jury et al., 1991). Compared with the estimated porosity (62%), the �st values are extremelysmall (c. 3%), which suggests that the solute must have reached the downslope probes as macropore flow. Themacropore flow in worm holes or decaying root channels, as seen in the excavation of the first plot (Figure 5),promotes rapid subsurface flow, which can increase the risk of pollutant transport to adjacent water bodies.

Plot design 3

The third plot design was used to examine lateral flow and its variability. The BTCs from the TDR probesin this plot (Figure 8) provide more evidence of lateral subsurface flow and support for Fluehler’s proposedtransport pattern. The shape of the BTCs are similar to each other at 30 to 50 cm depth, which is interestingconsidering that the probes are measuring transport at points 1 m apart, and which must have occurred throughdifferent cracks and macropores. The BTCs measured by eight probes in the upper 20 cm of the soil (Figure 8)reveal that the transport response is more variable in this upper soil layer than at 30–50 cm depth. Manyprobes detected no tracer (probes 4c, 5c and 8c) in spite of the close proximity of the application strip. Theprobes with the highest tracer concentrations and estimated mass recovery (probes 12c and 14c) tend to be theprobes underneath those that did not detect the tracer (probes 5c and 8c). These results support the focusing

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 16

2936 C. G. CAMPBELL, M. GHODRATI AND F.GARRIDO

0 500 10001500 2000 2500

0 500 1000 1500 2000 2500

0 500 1000 1500 2000 2500

0 500 1000 1500 2000 2500

0 500 1000 1500 2000 2500

0 500 1000 1500 2000 2500 0 500 1000 1500 2000 2500

time (min.)

C/Co

C/Co

C/Co

C/C

o0.6

0.5

0.4

0.3

0.2

0.1

0

0.6

0.5

0.4

0.3

0.2

0.1

0

0.6

0.5

0.4

0.3

0.2

0.1

0

Probe 1b

Probe 11b

Probe 13b

Probe 15b Probe 16b

Probe 14b

Probe 12b

Probe 2b Probe 3b

0.7

0.60.5

0.4

0.3

0.2

0.1

0

0.7

0.60.5

0.4

0.3

0.2

0.1

0

0.2

0.1

0

0.15

0.05

0 500 1000 1500 2000 2500

0.2

0.1

0

0.15

0.05

0.2

0.1

0

0.15

0.05

0 500 1000 1500 2000 2500

time (min.)

0.2

0.1

0

0.15

0.05

Figure 7. Breakthrough curves of CaCl2 from plot 2, where probes 1 to 3 were placed at 0–20 cm and 11 to 16 were placed at 30–50 cmdepth

process in Fluehler’s zone 1 where, in this case, the tracer was applied in a strip and was then funnelled intovertical and lateral preferential flow pathways that bypass the upper soil horizon.

The mass recoveries (M/Mo) and other transport parameters (�, �2 and �st) in Table V show a similarpattern. The M/Mo values at each depth are different. The first moments for probes in the first 20 cm are morevariable than those at greater depth. This might seem counterintuitive, that tracer transport would becomehomogenized with depth, however this is the pattern seen here and can be explained using Figure 6, where thesolute meets a less permeable soil layer at depth. The solute, which has been flowing in preferential pathwayswith irregular geometry, reaches a less permeable layer and collects, causing dispersion to decrease. This isthe same response observed at probes 11b and 12b in plot 2. The process of solute moving into a zone whereit is momentarily retained and dispersion decreases is called solute compression and it has been documented

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 17

CHARACTERIZING SHALLOW SOLUTE TRANSPORT 2937

Table IV. Transport parameters from the downslope lateral flow plot (plot 2)

Probe Depth Downslope � �2 �st

(cm) distance (cm) (cm h�1) (cm2 h�1)

1b 0–20 0 11Ð18 71Ð94 0Ð262b 0–20 0 7Ð04 44Ð17 0Ð153b 0–20 0 7Ð54 59Ð28 0Ð164b 0–20 50 — — —5b 0–20 100 — — —6b 0–20 100 — — —7b 0–20 100 — — —8b 0–20 200 — — —9b 0–20 200 — — —10b 0–20 200 — — —11b 30–50 0 12Ð57 72Ð78 0Ð0312b 30–50 0 14Ð76 64Ð17 0Ð0813b 30–50 100 7Ð66 40Ð00 0Ð0514b 30–50 100 8Ð36 39Ð44 0Ð0515b 30–50 200 10Ð33 32Ð50 0Ð0216b 30–50 200 7Ð18 17Ð42 0Ð02

Average 9Ð62 49Ð08 0Ð09SD 2Ð75 19Ð02 0Ð08CV(%) 29 39 90

in field studies at soil horizon boundaries (Ellsworth and Jury, 1991). Solute compression is demonstrated inthis plot by the smaller amount of solute spreading deeper in the soil compared with near the soil surface.For example, the second moment of probes 1c and 2c are much larger than those probes at 30 to 50 cm depth(Table V).

In combination with the previous results, the transport volume estimates (�st) suggest that the upper soilzone (0–20 cm) includes Fluehler’s zones 1 and 2, whereas zone 3 is occurring somewhere around 30–50 cmdepth. In a manner similar to Garrido et al. (2001), the �st values are larger in the upper soil, averaging 0Ð112,compared with 0Ð099 at 30–50 cm depth. Although the difference in �st of the upper and lower probes isnot statistically significant, the similar timing of solute concentration peaks is evidence of preferential flowoccurring in the upper zone.

It would seem very unlikely that the transport zones described in Figure 6 occur at a consistent soildepth throughout the entire hillslope. The description of these zones is applied to the data mainly to helpclarify processes occurring in the plots. The depths and consistency of the flow pattern would depend onseveral factors, including the downslope soil characteristics, water table height and hillslope shape andposition.

CONCLUSIONS

This study examined the application of an in situ technique, TDR, to measure solute transport in varioushillslope soil plots in a rangeland watershed. The TDR system performed well in all the plots in thisheterogeneous clay loam soil. In addition, TDR was effective in tracking solute transport in this mixedoak-woodland and grassland hillslope. In the first plot, designed to examine variability in vertical transport,the transport parameters were quantified within 90% of the mean and the first temporal moment (�) waswithin 95% of the mean, so travel times were characterized by as few as 1 probe m�2 to within 20% of themean. The transport parameters (�, �2 and �st) and measured mass recovery were all temporally stable onthe scale of hours.

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 18

2938 C. G. CAMPBELL, M. GHODRATI AND F.GARRIDO

Probe 1c Probe 5c Probe 9c Probe 13c

Probe 10c Probe 14c

Probe 11c Probe 15c

Probe 12c Probe 16c

Probe 2c Probe 6c

Probe 3c Probe 7c

Probe 4c Probe 8c

0.4

0.3

0.2

0.1

0

C/C

oC

/Co

C/C

oC

/Co

Probes 0-20 cm Deep

0 240 480 720 960 12000.4

0.3

0.2

0.1

00 240 480 720 960 1200

0.4

0.3

0.2

0.1

00 240 480 720 960 1200

0.4

0.3

0.2

0.1

0

0.4

0.3

0.2

0.10

0 240 480 720 960 1200 0 240 480 720 960 1200

0.4

0.3

0.2

0.10

0 240 480 720 960 12000.4

0.3

0.2

0.10

0 240 480 720 960 1200

0.4

0.3

0.2

0.10

0 240 480 720 960 12000.4

0.3

0.2

0.10

0 240 480 720 960 1200

0.4

0.3

0.2

0.10

0 240 480 720 960 12000.4

0.3

0.2

0.10

0 240 480 720 960 1200

0.4

0.3

0.2

0.10

0 240 480 720 960 1200

0.4

0.3

0.2

0.1

00 240 480 720 960 1200

0.4

0.3

0.2

0.1

00 240 480 720 960 1200

0.4

0.3

0.2

0.1

00 240 480 720 960 1200

0.4

0.3

0.2

0.1

00 240 480 720 960 1200

time (min) time (min) time (min) time (min)

Probes 30-50 cm Deep

Figure 8. Breakthrough curves of CaCl2 from plot 3, where probes 1 to 8 were placed at 0–20 cm and 9 to 16 were placed at 30–50 cmdepth

Table V. Transport parameters from breakthrough curves collected inthe lateral flow plot (plot 3)

Probe � �2 �st M/Mo

(cm h�1) (cm2 h�1)

1c 6Ð21 58Ð3 0Ð10 302c 8Ð16 97Ð7 0Ð13 343c 5Ð77 31Ð3 0Ð10 144c NA NA NA 05c NA NA NA 06c 6Ð74 29Ð3 0Ð11 807c 7Ð30 27 0Ð12 798c NA NA NA 09c 8Ð69 34Ð3 0Ð11 5310c 7Ð23 30Ð5 0Ð10 7511c 8Ð60 31Ð5 0Ð11 5712c 7Ð94 22Ð9 0Ð05 17013c 7Ð34 27 0Ð10 12514c 7Ð50 27Ð8 0Ð10 15415c 9Ð00 22Ð8 0Ð12 12916c 7Ð84 28Ð2 0Ð10 147

Average 7Ð84 36Ð05 0Ð11 71Ð69SD 1Ð40 20Ð54 0Ð03 58Ð22CV(%) 18 57 27 81

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 19

CHARACTERIZING SHALLOW SOLUTE TRANSPORT 2939

In plot 2, measuring vertical and lateral transport, the only shallow probes that measured the tracer werethose directly under the application point. All probes at 30–50 cm measured lateral transport of the tracer. Thepotential for rapid lateral flow was evidenced by the average velocity calculated for the downslope probes,which ranged from 6Ð8 to 27Ð9 cm h�1. Compared with the porosity (62%), the �st values calculated for thedownslope probes (2% to 8%) are an extremely small portion of the pore space available for solute transport.These results provide convincing evidence that rapid downslope solute transport occurs.

Results from the third plot, examining lateral flow across the hillslope, provided more evidence that transportoccurred in a pattern like that described by Fluehler et al. (1996). In this pattern, solute uniformly enters thesoil and is focused into preferential flow pathways, producing bypass flow deeper in the soil profile. Thesolute then collects over a less permeable layer and flows laterally. A majority of the mass applied to plot 3was recorded as lateral flow by the probes installed at 30 to 50 cm depth. There was clear evidence of thesignificance of both preferential and lateral flow pathways, which have the potential to cause rapid runoff of asurface-applied solute. In addition, it was demonstrated that the soil anisotropy may not only promote lateralflow, but also may decrease solute spreading, which creates a more homogenized lateral flow response, asmeasured by the TDR probes at 30 to 50 cm depth.

ACKNOWLEDGEMENTS

The authors wish to acknowledge the Horton Graduate Research Award administered by the AmericanGeophysical Union as well as the University of California Sustainable Agriculture Research and EducationProgram for their support of this research. We also would like to thank Brad Peconom for his help and supportat the University of California Russell Reservation.

REFERENCES

Blackburn WH. 1975. Factors influencing infiltration and sediment production of semiarid rangelands in Nevada. Water Resources. Research11(6): 929–937.

Buchter B, Hinz C, Flury M, Fluehler H. 1995. Heterogeneous flow solute transport in an unsaturated stony soil monolith. Soil ScienceSociety of America Journal 59: 14–21.

Campbell CG, Ghodrati M, Garrido F. 1999. Comparison of time domain reflectometry, fiber optic mini-probes, and solution samplers forreal time measurement of solute transport in soil. Soil Science 164(3): 156–170.

Campbell CG, Ghodrati M, Garrido F. 2001. Temporal consistency of solute transport in a heterogeneous field plot. Soil Science 166:491–506.

Dalton FN, Van Genutchen MTh. 1986. The time domain reflectromtery method for measuring soil water content and salinity. Geoderma38: 237–250.

Daniel WW. 1995. Biostatistics: a Foundation for Analysis in Health Sciences , 6th edn. Wiley: New York; 613–622, 756.Duda AM. 1993. Addressing nonpoint sources of water pollution must become an international priority. Water Science and Technology 28:

1–11.Dunne T, Black RD. 1970. An experimental observation of runoff production in permeable soils. Water Resources Research 6: 478–490.Ellsworth TR, Jury WA. 1991. A three-dimensional field study of solute transport through unsaturated, layered, porous media 2. Character-

ization of vertical dispersion. Water Resources Research 27(5): 967–981.Fetter CW. 1977. Attenuation of waste water elutriated through glacial outwash. Ground Water 15: 365–371.Fluehler H, Durner W, Flury M. 1996. Lateral solute mixing processes—a key for understanding field-scale transport of water and solutes.

Geoderma 70: 165–183.Flury M, Fluehler H, Jury WA, Leuenberger J. 1994. Susceptibility of soils to preferential flow of water: a field study. Water Resources

Research 30(7): 1945–1954.Flury M, Leuenberger J, Studer B, Fluehler H. 1995. Transport of anions and herbicides in a loam and sandy field soil. Water Resources

Research 31(4): 823–835.Garrido F, Ghodrati M, Campbell CG, Chendorain M. 2001. Detailed characterization of solute transport processes in a heterogeneous field

soil. Journal of Environmental Quality 30(2): 573–583.Gee GW, Bauder JW. 1986. Particle-size analysis. In Methods of Soil Analysis. Part 1, 2nd edition, Agronomy Monographs 9 , Klute A (ed.).

ASA and SSA: Madison, WI; 399–403.Ghodrati M, Jury WA. 1990. A field study using dyes to characterize preferential flow of water. Soil Science Society of America Journal

54: 1558–1563.Heimovaara TJ, Focke AG, Bouten W, Verstraten JM. 1995. Assessing temporal variations in soil water composition with time domain

reflectometry. Soil Science Society of America Journal 59(3): 689–698.

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)

Page 20

2940 C. G. CAMPBELL, M. GHODRATI AND F.GARRIDO

Hornberger GM, Germann PF, Beven K. 1991. Throughflow and solute transport in an isolated sloping soil block in a forested catchment.Journal of Hydrology 124: 81–99.

Jacques D, Kim DJ, Diels J, Vanderborght J, Vereecken H, Feyen J. 1998. Analysis of steady state chloride transport through twoheterogeneous field soils. Water Resources Research 34(10): 2539–2550.

Jury WA. 1985. Spatial Variability of Soil Physical Parameters in Solute Migration: a Critical Literature Review. Electrical Power ResearchInstitute EA-4228, Report to the Electrical Power Research Institute; 3–1.

Jury WA, Roth K. 1990. Transfer Functions and Solute Movement through Soil . Birkhauser Verlag: Basel.Jury WA, Sposito G. 1985. Field calibration and validation of solute transport models for the unsaturated zone. Soil Science Society of

America Journal 49: 1331–1341.Jury WA, Gardner WR, Gardner WH. 1991. Soil Physics , 5th edn. Wiley: New York; 328.Kachanoski RG, Pringle E, Ward A. 1992. Field Measurement of solute transport travel times using time domain reflectometry. Soil Science

Society of America Journal 56: 47–52.Kim DJ, Vanclooster M, Feyen J, Vereecken H. 1998. Simple linear model for calibration of time domain reflectometry measurements on

solute transport. Soil Science Society of America Journal 62: 83–89.Luxmoore RJ, Jardine PM, Wilson GV, Jones JR, Zelanzny LW. 1990. Physical and chemical controls of preferred path flow through a

forested hillslope. Geoderma 46: 139–154.Mallants D, Vanclooster M, Meddahi M, Feyen J. 1994. Estimating solute transport in undisturbed soil columns using time-domain

reflectometry. Journal of Contaminant Hydrology 17: 91–109.Mallants D, Vanclooster M, Toride N, Vanderborght J, Feyen J. 1996. Comparison of three methods to calibrate TDR for monitoring solute

movement in undisturbed soil. Soil Science Society of America Journal 60: 747–754.McDonnell JJ. 1990. A rationale for old water discharge through macropores in a steep, humid catchment. Water Resources Research 26(11):

2821–2832.Nadler A, Dasberg S, Lapid I. 1991. Time domain reflectometry measurements of water and electrical conductivity of layered soil columns.

Soil Science Society of America Journal 55: 938–943.Newman BD, Campbell AR, Wilcox BP. 1998. Lateral subsurface flow pathways in a semiarid ponderosa pine hillslope. Water Resources.

Research 34(12): 3485–3496.Noguchi S, Tsuboyama Y, Sidle RC, Hosoda I. 1999. Morphological characteristics of macropores and the distribution of preferential flow

pathways in a forested hillslope. Soil Science Society of America Journal 63: 1413–1423.Persson M, Berndtsson R. 1998. Estimating transport parameters in an undisturbed soil column using time domain reflectometry and transfer

function theory. Journal of Hydrology 205: 232–247.Radcliffe DE, Gupte SM, Box JE Jr. 1998. Solute transport at the pedon and polypedon scales. Nutrient Cycling in Agroecosystems 50:

77–84.Rudolph DL, Kachanoski RG, Celia MA, LeBlanc DR, Stevens JH. 1996. Infiltration and solute transport experiments in unsaturated sand

and gravel, Cap Cod, Massachusetts: Experimental design and overview of results. Water Resources Research 32(3): 519–532.Salve R. 1995. Spatial and temporal variations in flow in a rangeland catchment in California. Doctoral Thesis, University of California at

Berkeley; 17–23.Sklash MG, Stewart MK, Pearce AJ. 1986. Storm runoff generation in humid headwater catchments 2. A case study of hillslope and

low-order stream response. Water Resources Research 22(8): 1273–1282.Tsuboyama Y, Sidle RC, Noguchi S, Hosoda I. 1994. Flow and solute transport through the soil matrix and macropores of a hillslope

segment. Water Resources. Research 30(4): 879–890.Turton DJ, Haan CT, Miller EL. 1992. Subsurface flow responses of a small forested catchment in the Ouachita Mountains. Hydrological

Processes 6: 111–125.Vachaud G, Passerat De Silans A, Balabanis P, Vauclin M. 1985. Temporal stability of spatially measured soil water probability density

function. Soil Science Society of America Journal 49: 822–828.Valocchi AJ. 1990. Use of temporal moment analysis to study reactive solute transport in aggregated porous media. Geoderma 46: 233–247.Verchot LV, Franklin EC, Gilliam JW. 1997. Nitrogen cycling in piedmont vegetated filter zones: II. Subsurface nitrate removal. Journal of

Environmental Quality 26: 337–347.Wagner JR. 1978. Late Cenozoic history of the Coast Ranges east of San Francisco Bay. Doctoral thesis, University of California at Berkeley;

17–130.White RE, Dyson JS, Haigh RA, Jury WA, Sposito G. 1986. A transfer function model of solute transport through soil 2. Illustrative

applications. Water Resources Research 22(2): 248–254.Wilson GV, Jardine PM, Luxmoore RJ, Jones JR. 1990. Hydrology of a forested hillslope during storm events. Geoderma 46: 119–138.

Copyright 2002 John Wiley & Sons, Ltd. Hydrol. Process. 16, 2921–2940 (2002)