Soil & Tillage Research 112 (2011) 47–57

Impact of tillage on solute transport in a loamy soil from leaching experiments

A. Besson a,b,*, M. Javaux a, C.L. Bielders a, M. Vanclooster a

a Earth and Life Institute, Environmental Sciences, Universite Catholique de Louvain, Croix du Sud 2, BP 2, B-1348 Louvain-la-Neuve, Belgiumb Soil Science Unit, INRA 2163 Avenue de la Pomme de Pin, CS 40001, Ardon, 45075 Orleans Cedex 2, France

A R T I C L E I N F O

Article history:

Received 9 June 2010

Received in revised form 31 October 2010

Accepted 8 November 2010

Available online 8 December 2010

Keywords:

Tillage/reduced tillage

TDR breakthrough curves

Undisturbed lysimeters

Solute dispersivity

Mixing regime

Dirac response

Transfer function

A B S T R A C T

Soil tillage practices can affect water flow and solute transport processes dynamically in space and in

time. However, the relationship between tillage practices and flow and transport in soils is not yet well

understood. Within this paper, we analyze the short term impact of the conversion from conventional

mouldboard ploughing (CT) to reduced disc harrowing (RT) on the solute transport process within a

loamy soil. Solute breakthrough experiments at two flow rates were performed on 2 undisturbed

lysimeter collected in a CT and RT field plot. Solute transport parameters were estimated using transfer

function theory. Important differences in solute transport were observed between the RT and the CT

treatments. The CT treatment exhibited a rapid, more homogeneous and less dispersive solute transport

as compared to the RT treatment. These results are explained by the changes in soil structure due to

tillage and compaction. The dominant transport was identified as being a stochastic-convective process

in both lysimeters. The similarity of the mixing regime for the two soil columns can be explained by

preferential solute trajectories activated within structural macropores as a result of the high flow rates

applied. We show that the relationship between tillage practices and transport is complex, not only scale

and time dependent but also influenced by the boundary conditions and tillage practices.

� 2010 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Soil & Tillage Research

journal homepage: www.elsev ier .com/ locate /s t i l l

1. Introduction

Soil is a key component of terrestrial ecosystems where waterrunoff, infiltration, drainage and storage interact with chemicalmovement. Flow and transport in soil is complex, space and timescale dependent and subject to human influence. Soil tillage, forinstance, has a major control on flow and transport. Indeed, themechanical perturbation, aiming at developing desirable soilconditions for a seedbed and establishing specific surface configu-ration for planting, irrigation, drainage or harvesting operations, canhave a considerable impact on the soil structure, and hence, on soilhydraulic functioning (Kepner et al., 1978). Tillage may consist of awide range of practices, ranging from minimum and reduced tillage,say harrowing, and no-till practices for conservation systems inwhich a substantial part (at least 30%) of the soil remains covered by

Abbreviations: CT, conventional tillage; RT, reduced tillage; TDR, Time Domain

Reflectometry; T1, T2, T3, TDR transects; zm, mean soil depth; Jw1, steady state flow

rate equal to 5 cm per day (cm d�1); Jw2, steady state flow rate equal to 20 cm per

day (cm d�1); DL, longitudinal dispersion coefficient; v, mean solute velocity; lL,

longitudinal hydrodynamic dispersivity; Crt*, time-normalized resident concentra-

tion; superscript ‘‘l’’, local parameters; superscript ‘‘m’’, depth averaged parameters;

R, retardation factor; BTCs, breakthrough curves.

* Corresponding author at: Soil Science Unit, INRA 2163 Avenue de la Pomme de

Pin, CS 40001, Ardon, 45075 Orleans Cedex 2, France. Tel.: +33 02 38 41 78 00;

fax: +33 02 38 41 78 79.

E-mail address: Arlene.Besson@orleans.inra.fr (A. Besson).

0167-1987/$ – see front matter � 2010 Elsevier B.V. All rights reserved.

doi:10.1016/j.still.2010.11.001

previous crop residues (Holland, 2004), to mouldboard ploughing asin traditional (or conventional) systems.

Different tillage practices can explain the differences observed insoil structure and, consequently, in water flow and chemicaltransport in the soil. Several authors underline the benefits ofreduced tillage on hydraulic functioning. Plant water uptake, soilwater storage, soil water infiltration and transmission are improvedin many conservation systems, mainly as a consequence of themodification in soil physical and hydraulic properties (van Doren andAllmaras, 1978; Hatfield et al., 2001; Turner, 2004; De Vita et al.,2007; Moret and Arruea, 2007; Casa and Lo Cascio, 2008; Strudley etal., 2008). Tracer studies also show that conservation tillage generallypromotes more rapid leaching of non-reactive solutes (Gish et al.,1991; Shipitalo and Edwards, 1993; Shipitalo et al., 2000) andpesticides (Alletto et al., 2010 for an extensive review). Indeed,macroporeflow dominates generally solutetransport in conservativetilled soil as compared to conventionally tilled soil (Petersen et al.,2001; Vervoort et al., 2001; Kulli et al., 2003; Vogeler et al., 2006).Conventional tillage generally reduces solute transport by cuttingfunctional macropores (Jarvis, 2007). As suggested by the studies ofVanclooster et al. (2005) and Javaux et al. (2006), cutting functionalmacropores by intense cultivation will also affect flow pathways, andgeneral solute mixing regime at the scale of the soil profile.

Yet, the relationship between tillage and flow and transport isnot unambiguous. Whereas the abovementioned studies andadditional numerical studies suggest greater solute leaching under

Table 1Soil texture of the A and Bt horizons of Eutric Luvisol.

Depth (m) Clay (g kg�1) Silt (g kg�1) Sand (g kg�1)

A horizon 0–0.3 156 657 187

Bt horizon 0.3–1 250 489 261

A. Besson et al. / Soil & Tillage Research 112 (2011) 47–5748

conservation tillage (Masse et al., 1996; Isensee and Sadeghi,1997), other studies show significantly opposite findings with noeffect of tillage and even increased flow and transport inconventionally tilled soils (Granovsky et al., 1993; Levanonet al., 1993; Clay et al., 1998).

Differences in climate, in soil type, in physico-chemical soilproperties, in soil historical management, in crop type residues, inexperimental design and in space–time variation of initial soilconditions, can overwhelm tillage effects (Logsdon et al., 1993;Logsdon and Jaynes, 1996; Alletto et al., 2010). In addition theimpact of tillage depends on intrinsic soil properties, on tillagecharacteristics such as tillage type, depth and speed and the levelof exerted mechanical stress. In most experimental studies, theimpact of two extreme tillage systems, i.e. conventional versusconservative, on flow and transport has been analyzed. Fewstudies analyze the impact of the mechanical stress applied tosoils, i.e. for instance primary tillage (mouldboard ploughing)versus secondary tillage (harrowing), on flow and transport,notwithstanding tillage implements, and then the tillage intensi-ty, have been reported to have an important impact on flow andtransport (Jarvis, 2007).

Given the above mentioned debate, more detailed studies onthe impact of tillage on flow and transport, in particular themechanisms responsible for the control of tillage on flow andtransport, are needed. Such studies will allow elucidate thebenefits of conversion tillage on soil functioning as compared toconventional tillage.

The objective of this study was to compare the impact ofconventional mouldboard tillage (CT), and reduced tillage (RT),consisting of harrowing without prior ploughing, on the solutetransport process within a loamy soil. Our study focused on theshort-term effects of tillage on the solute transport process, i.e. thecomparison of the effect of changes in level of mechanical stress(CT and RT) on solute transport after one single reduced tillageoperation (RT) following a long term CT. From laboratorycontrolled transport experiments on undisturbed soil lysimeterscollected in field plots subject to a long term CT treatment,effective solute transport parameters were estimated. Thevariability of transport properties in terms of tillage practicewas analyzed at local and depth averaged scale for two differentflow rates.

[(Fig._1)TD$FIG]

Fig. 1. Soil structure as schematized

2. Materials and methods

2.1. Column sampling and soil characteristics

The experimental study was performed in the laboratory ontwo, 0.5 m3 undisturbed lysimeters (0.8 m diameter � 1 m height)sampled in a Eutric Luvisol (FAO, 2006) in Louvain-la-Neuve,Belgium. The soil sampling technique was described by Vancloos-ter et al. (1995). The soil profile encompasses (1) a loamy Ahorizon, susceptible to structure deterioration when tilled, giventhe high silt content (about 66%) and (2) an argic Bt horizon with ahigher clay content (about 25%). Under field conditions, the soil is awell drained loamy soil. Soil texture is summarized in Table 1.

The two soil columns were sampled in April 2007 on bare soilfrom the same field, which had been previously cultivated with asilage maize crop under conventional tillage. One day prior tosampling, part of the field was subjected to conventional tillage(CT), i.e. conventional mouldboard ploughing (0–30 cm) followedby disc harrowing (0–10 cm). Maize stalks were incorporated intothe topsoil during tillage. The other part of the field was subjectedto superficial and reduced tillage (RT) consisting only in discharrowing (0–10 cm). One column (CT) was sampled from theconventionally tilled part and another column (RT) from thesuperficially tilled part.

A visual description of soil structure was realized on two soilpits dug in the CT and RT parts, as shown schematically in Fig. 1.The top 10 cm of the soil had a fine, porous and rather uniformstructure for CT soil but encompassed a mix of porous zones anddense clods for the RT topsoil. In the lower part of the loamy Ahorizon (15–30 cm depth), undecomposed organic residues,many earth-worm holes and roots, dense clods (0.1–30 dm3) oraggregates tightly packed were encountered for both tillagesystems with a 5-cm thick plough pan identified at about�35 cmdepth. However, the no-tilled horizon of RT soil presented lessorganic residue and was more strongly consolidated. To complete

from in situ visual descriptions.

A. Besson et al. / Soil & Tillage Research 112 (2011) 47–57 49

the physical description of soils, bulk density measurements wereperformed by means of undisturbed soil cores (100 cm3) sampledin situ at several soil depths and close to the soil lysimeters, i.e. inthe ploughed (CT) plot (n = 39) and in the unploughed (RT) plot(n = 30). Bulk densities variability is used to comfort solutetransport results as described in Section 4.

2.2. Equipment

In the laboratory, each lysimeter was equipped to monitorelectrical impedance, soil permittivity, soil temperature and outletand inlet flow (Fig. 2).

Electrical impedance and permittivity were monitored every15 min at 6 soil depths using twelve Time Domain Reflectometryprobes (TDR). TDR probes consisted of three stainless steel rods(42.5 cm long and 0.5 cm diameter) with 2.5 cm rod spacing. Theywere all horizontally inserted into the soil along three verticaltransects spaced 1208 apart (Fig. 2). The first transect (T1)consisted of six probes placed at �15, �30, �45, �60, �75 and�90 cm depth from the soil surface, the second (T2) consisted ofthree probes placed at �13, �43 and �73 cm depth and the third(T3) consisted of three probes placed at �17, �47 and �77 cmdepth. This specific spatial configuration initially developed byJavaux (2004b) aimed at increasing the horizontal spatialresolution of measurements at the soil profile scale. In thisway, each of the three soil depths (zm) of about �15 cm, �45 cmand �75 cm were monitored by three probes.

TDR measurements were performed using a TDR 100 TimeDomain Reflectometer connected to a SDMX50 multiplexer system(Campbell Scientific Inc., UK) which was controlled by a data loggerand PC computer. From the TDR signals, soil permittivity andimpedances were recorded every 15 min. Soil permittivity valueswere converted into soil volumetric water content using Topp’scalibration model (Topp et al., 1980).

Knowing that the variability of soil temperature, for instanceinduced by water application, can impact electrical impedance, in

[(Fig._2)TD$FIG]

Fig. 2. Experimental set up for transport ex

addition to TDR probes, four temperature NTC probes (NegativeTemperature Coefficient thermistors) were horizontally insertedinto each column at �15, �45, �75 and �90 cm depth (Fig. 2).Temperature measurements were recorded every 15 min. Meantemperatures at �30 and �60 cm depth were calculated frommeasurements performed respectively at �15, �45 cm depth andat �45, �75 cm depth. The electrical impedance dataset was thencorrected for the temperature effect by the well known Keller andFrischknecht equation (Keller and Frischknecht, 1966) at 25 8C,with a temperature coefficient equal to 0.025.

2.3. Upper and bottom boundary conditions

The incoming flow rate was controlled during the infiltrationexperiments by a water applicator on each column. The waterapplicator was made of a square reservoir (80 cm � 80 cm � 1 cm)at the bottom of which 280 needles (0.5 mm diameter) were placedon a regular grid (5 cm � 5 cm) (Fig. 2). The applicators werecontinuously fed with water (780 mS/cm) by means of a pressurepump and were placed at 30 cm height from the soil surface. Onecentimeter of fine gravel was spread over the soil surface in orderto avoid surface sealing by the falling raindrops. The lysimeterbottom consisted of a plate with an outlet gate embedded at itscentre from which a drainage tube was connected to a tippingbucket device (GME, Type PR12). The number of tips was recordedevery 30 min. The outflow time series were then obtained bytransforming the tipping frequency into an instantaneous flow rateand used to verify whether steady state had been achieved. Thebottom boundary water condition was a seepage face.

2.4. Transport experiments

After sampling, the soil columns were stored for one year in thelaboratory (15–20 8C) without any disturbance and irrigation.During storing, soil columns were closed to restrict soilevaporation. Initial soil water content, i.e. before any infiltration

periments in undisturbed lysimeters.

A. Besson et al. / Soil & Tillage Research 112 (2011) 47–5750

experiments, was relatively wet, close to the water field capacity.Unsaturated flow experiments were then performed on thelysimeters. A first constant flow rate Jw1, equal to 5 cm per day(cm d�1), was imposed at the soil surface until a steady state flowrate was obtained. Slightly mineralized water (780 mS cm�1) wasused. Secondly, while maintaining the steady state flow rate Jw1, aDirac like pulse of CaCl2 solution (0.5 mol l�1) was applieduniformly at the soil surface. The application duration was equalto 10 h. Subsequently, the slightly mineralized water(780 mS cm�1) was applied with the same flow rate Jw1.

Once the electrical conductivity of the outlet flow reached itsinitial value (roughly 780 mS cm�1), a second constant water flowrate Jw2, equal to 20 cm per day (cm d�1), was imposed at the soilsurface. At steady-state flow regime, a second Dirac 0.5 mol l�1

CaCl2 pulse was performed during 3 h. Thereafter, irrigation wascontinued with the slightly mineralized water.

In all cases, the total volume of the solute pulse is approxi-mately equal to 10 dm3 and the duration of solute application wasadjusted as a function of Jw to be small compared with the totalduration, so that we could mathematically interpret it as a Diracdelta function. The soil water contents estimated from TDRmeasurements were relatively constant and close to saturation allalong the soil profiles (Table 2). During the experiments, the timevariations of water content did not exceed 0.008 cm3 cm�3.

2.5. Estimating solute transport properties at local and depth

averaged scale

We used the one dimensional convection-dispersion equation(CDE) to characterize the solute transport mixing regime in aneffective way:

dC

dt¼ DL

d2C

dz2� v

dC

dz(1)

with C (ML�3) the solute concentration, DL (L2T�1) the longitudinaldispersion coefficient, v (LT�1) the mean solute velocity, and t (T), z

(L) the time and space coordinates. Neglecting chemical diffusion,the longitudinal hydrodynamic dispersivity lL (L) was calculatedas a function of the travel depth z:

lLðzÞ ¼DLðzÞvðzÞ

(2)

A constant lL along the solute travel path would indicate that themixing regime can be described as a Fickian, concentrationgradient-driven process (Dagan and Nguyen, 1989). In such case,the mixing regime is called convective-dispersive (CD). On thecontrary, when lL increases with travel depth, the mixing regime isstochastic-convective (SC) (Simmons, 1982).

We used the analytical solution of Eq. (1) implemented inMatlab 7.6 (Mathworks Inc., Natick, USA, 2008) in a toolbox calledCASlib (library for Convection-dispersion Analytical Solutions)developed by Javaux (2004a). We assumed that the bottomboundary did not affect solute concentration inside the columnsand hence considered that the solution for a semi-infinite soil

Table 2Mean values of water content (cm3 cm�3) calculated from three TDR measurements

at the depths of CT and RT soils for water flow rates of 5 and 20 cm per day (cm d�1).

Standard deviations are given in italic.

Soil depth (cm) 5 cm d�1 20 cm d�1

CT RT CT RT

�15 0.43� 0.02 0.41� 0.04 0.45�0.01 0.40�0.03

�45 0.39� 0.02 0.40� 0.01 0.41�0.01 0.46�0.03

�75 0.40� 0.01 0.45� 0.01 0.41�0.01 0.45�0.004

column was applicable. For steady-state flow and for a soil profileinitially free of solute and with uniform water content (BoundaryValue Problem), the concentration distribution at any given time(t) and depth (z) was generalized from the following convolution:

Crðz; x; tÞ ¼Z t

0Cr

inð0; tÞGrðt � t; z; xÞdt (3)

where r, the superscript for resident type; x refers to the radialposition of the TDR probe around the monolith; Cr

in, the inputfunction; t, the time lag; G, is the integral kernel for the third typeupper boundary condition. For a non-reactive tracer, the traveltime probability density function (pdf) Gr was given by Jury andRoth (1990) for BVP with a Dirac input function such that:

G r ¼ vffiffiffiffiffiffiffiffiffiffiffipDLtp exp �ðz�Þ

2

4DLt

" #� v2

2Dexp

DL

� �er fc

zþffiffiffiffiffiffiffiffiffiffi4DLtp� �

(4)

Following Vanderborght et al. (1996), we use the integral kernel Gr

to correspond to the travel time probability density function (pdf),equal to the time-integral-normalized resident (r) concentrationCrt* written as (Vanderborght et al., 1996):

Crt�ðz; x; tÞ ¼ Crðz; x; tÞR10 Crðz; x; tÞ dt

(5)

As described by Javaux and Vanclooster (2003a), the time-normalized resident concentration Crt* method avoids the impactof the uncertainty on the probe calibration (Mallants et al., 1996).Indeed, the TDR readings without any calibration can then bedirectly related to Crt* such as:

Crt�l ðz; x; tÞ ¼

1=Z�ðz; tÞR10 1=Z�ðz; tÞdt

(6)

where Z*(z,t) is the difference between actual and initial TDRimpedance. In the absence of calibration, the time-normalizedresident concentrations were not interpreted in terms of total massof solute crossing the sampling TDR window. The mean time-integral normalized resident concentration Crt�

m at each depth zm

was obtained by averaging the time series of three TDR readings ateach of the �15, �45 and �75 cm depths, such that:

Crt�m ðzm; tÞ ¼

P3x¼1 Cl

rt�ðz; x; tÞR10

P3x¼1 Cl

rt�ðz; x; tÞdt(7)

Solute transport parameters v and DL were then estimated byfitting the analytical solution Eqs. (3) and (4) to the time-integral-normalized concentrations obtained from the TDR time series atlocal (Eq. (6)) or depth averaged scale (Eq. (7)). The correspondingdispersivity lengths were calculated from Eq. (2). The retardationfactors R, expressed in pore volume, expresses the ratio betweenpiston-flow velocity and the mean solute transport velocity:

R ¼ Jw

vuðzÞ dz(8)

with Jw, the water flow, v, the solute mean velocity (in Eq. (4)) andu, the soil water content. Local scale parameters are derived frombreakthrough curves (BTC) corresponding to local TDR samplingvolumes positioned at locations x and at depths z (transects T1, T2and T3). In addition to the vertical variability, the horizontalvariability in local parameters is analyzed.

Depth averaged solute transport parameters were derived fromCrt�

m and thus from the mean of the three BTCs at depth zm. Weassumed that Crt�

m was representative for the cross-section of thelysimeter and integrated, hence, the horizontal variability at thescale of the cross section of the lysimeter. In the following, we willdenote local parameters with superscript ‘‘l’’ and depth averagedparameters with superscript ‘‘m’’.

[(Fig._3)TD$FIG]

0

0.0001

0.0002

0.0003

0.0004

0.0005

0.0006

0 5000 10000 15000 20000 25000 30000Time min

Crt*

Measured data Modelled data

Fig. 3. Local BTCs measured and modeled for TDR transect T1 of the CT soil at a flow rate of 5 cm per day (cm d�1).

Table 3Normalized estimation errors (NRMSE) in percentage between measured BTCs and

estimated BTCs obtained for CT and RT soils at water flow rates of 5 and 20 cm per

day (cm d�1). Datasets with NRMSE values marked by an asterisk were not

analyzed.

Soil depths (cm) 5 cm d�1 20 cm d�1

CT RT CT RT

Local-scale

Trans. T1 �15 2 3 6 2

�30 2 3 6 4

�45 6 9 6 6

�60 3 5 5 8

�75 2 No data 11 8

�90 6 14 No data 12

Trans. T2 �13 4 2 4 5

�43 1 8 2 15*

�73 3 6 5 12

Trans. T3 �17 2 4 3 8

�47 7 5 13 7

�77 9 3 20* 12

Depth averaged-scale

Depth average �15 1 2 4 4

�45 5 5 6 5

�75 2 9 8 22*

A. Besson et al. / Soil & Tillage Research 112 (2011) 47–57 51

2.5.1. Error calculations on estimates and parameter uncertainties

The adequacy between measurements and BTC estimates isanalyzed from normalized estimation errors (NRMSE) calculatedfor each dataset and expressed in percentage, such as:

NRMSE ¼ 1

Crt�max � Crt�

min

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiPn1 ðC

rt� � Crt�^Þ

2

n

vuut(9)

with Crt*, the time-integral-normalized concentrations calculatedfor the n measured data; Crt�

max � Crt�min, the range of Crt* data and Crt�

^,

the time-integral-normalized concentrations obtained from theestimated transport parameters following the inversion process.

The parameter uncertainty was determined on the basis of theparameter variance–covariance matrix (Kool and Parker, 1987).The parameter confidence intervals were approximated by linearregression analysis for the nonlinear problem (Lambot et al., 2002;Javaux and Vanclooster, 2006). Approximate confidence intervalsCIapp of the estimated parameters are expressed as:

CIapp ¼ tn�c1�a=2

ffiffiffiffiffiffiC p

q(10)

where Cp is the diagonal of the variance–covariance matrix of theparameter p, c is the matrix size, t is the value of the Studentdistribution with (n � c) degrees of freedom and confidence level(1 � a). In the analysis a was equal to 0.05, say 95% confidenceinterval.

3. Results

Fig. 3 shows an example of local BTCs measured and modeledfor one TDR transect. For clarity, only fitted BTCs will be shownhereafter. In general the CDE solution fitted well the experimentalBTCs. As shown in Table 3, normalized estimation errors (NRMSE)indicate small residual variance with values generally lower than10%. However, some datasets presented considerable noise or wereincomplete for technical reasons (‘‘No Data’’ in Table 3). Thedatasets with a NMRSE equal or higher than 15% were therefore notconsidered in the further analysis.

3.1. Analysis of local scale transport parameters

At the local scale and irrespective of flow rates and soil depths,differences in shapes of BTCs are observed between the CT and theRT soils as shown for transect T1 in Fig. 4. The long tailing in BTCsobserved in the RT soil suggests the implication of lowpermeability zones in transport, in particular for the topsoil, asopposed to the more symmetrical BTCs of CT soil.

The local solute velocity v profiles are shown in Fig. 5. It isobserved that CT velocities are significantly higher than RT velocityvalues. Close to the soil surface (around 15 cm depth), the RT soilexhibits large horizontal variations in local transport velocities.

The same is visible in Fig. 6, where local retardation factorslarger than 1 reflect retarded transport as compared to the waterflow, whereas factors smaller than one suggest rapid solutetransport. For the low flow rate, it is observed that the CTretardation factors are generally lower than 1, while the transportin the RT soil is generally delayed (R > 1). For high flow rate, RT andCT both show retardation along the whole profile, with again RTretardation factors generally higher than those for the CT soil. Ingeneral, for a given depth (around 15, 45 and 75 cm), there is lesshorizontal variability in retardation for the CT soil than for the RTsoil. The CT transport is relatively more homogeneous than the RTtransport, particularly when flow rate was equal to 20 cm per day(cm d�1) (Figs. 5 and 6).

Additionally, CT dispersivity profiles show more homogeneoustransport with lower and less scattered dispersivity values for agiven depth (Fig. 7). RT dispersivity lengths are generally higherand more spread. In the top 45 cm of the soil, the dispersivitylengths range from 3 to 18 cm for CT against 10 to 60 cm for RT. Thevalues of the latter are more dispersed with outliers exceeding thetravel distance. The general trend of the dispersivity profiles is

[(Fig._4)TD$FIG]

0.00000

0.00005

0.00010

0.00015

0.00020

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

0.00000

0.00005

0.00010

0.00015

0.00020

0.00025

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10

RT soil, 5 cm d-1RT soil, 20 cm d-1

0.0000

0.0001

0.0002

0.0003

0.0004

0.0005

0.0006

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

0.0000

0.0002

0.0004

0.0006

0.0008

0.0010

0.0012

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10

-15 cm-30 cm-45 cm-60 cm-75 cm-90 cm

CT soil, 5 cm d-1 CT soil, 20 cm d-1

*rtlC

*rtlC

Pore volume Pore volume

Soil depths in columns

Fig. 4. Local BTCs modeled for T1 transects of CT and RT soils at flow rates of 5 and 20 cm per day (cm d�1).

[(Fig._5)TD$FIG]

a- 5 cm d-1

b- 20 cm d-1

0

-10

-20

-30

-40

-50

-60

-70

-80

-90

-100

0 0.005 0.01 0.015 0.02 0.025 0.03

v l cm min -1

Soil

dept

hz

cm

Soil CT Soil RT

0

-10

-20

-30

-40

-50

-60

-70

-80

-90

-100

0 0.005 0.01 0.015 0.02 0.025 0.03

v l cm min -1v l cm min -1

Soil

dept

hz

cm

Soil CT Soil RT

0

-10

-20

-30

-40

-50

-60

-70

-80

-90

-100

0 0.005 0.01 0.015 0.02 0.025 0.03v l cm min -1

Soil

dept

hz

cm

Soil CT Soil RT

0

-10

-20

-30

-40

-50

-60

-70

-80

-90

-100

0 0.005 0.01 0.015 0.02 0.025 0.03v l cm min -1v l cm min -1

Soil

dept

hcm

Soil CT Soil RT

Fig. 5. Mean solute velocity vl estimated from local BTCs for CT and RT soils at flow rates of 5 cm per day (a) and 20 cm per day (b).

[(Fig._6)TD$FIG]

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5

-13

-15

-17

-30

-43

-45

-47

-60

-73

-75

-77

-90

Soil

dept

h cm

.

Pore volume

CT soil

RT soil

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5

-13

-15

-17

-30

-43

-45

-47

-60

-73

-75

-77

-90

Soil

dep

th c

m

.

Pore volume

CT soil

RT soil

a- 5 cm d-1 b- 20 cm d-1

Fig. 6. Local retardation factors for CT and RT solute transports at flow rates of 5 cm per day (a) and 20 cm per day (b).

A. Besson et al. / Soil & Tillage Research 112 (2011) 47–5752

probably a slight increase with depth for CT, indicating imperfectlateral mixing (transport in parallel stream tubes). For the RT soil,no trend is clearly visible.

The local BTCs and related solute transport parameters suggesta relevant impact of tillage practices on solute transport. Indeed

the RT and CT soils exhibit differences in terms of transportparameters (velocity and dispersivity) and of BTC shapes and meantravel time of solute. Parameter values are highly variablehorizontally, in particular for the RT soil. This hampers theinterpretation in term of mixing regime.

[(Fig._7)TD$FIG]

-100-90-80-70-60-50-40-30-20-10

01009080706050403020100

RT soil 5 cm d-1

CT soil, 5 cm d-1

RT soil, 20 cm d-1

CT soil, 20 cm d-1

λLl cm So

il de

pth

z cm

.

Fig. 7. Local dispersivity calculated from transport parameters DLl and vl for CT and

RT soils at flow rates of 5 and 20 cm per day (cm d�1).

[(Fig._9)TD$FIG]

5.554.543.532.521.510.50

-15

-45

-75

Soil

dept

h cm

.

Pore volume

CT soil, 5 cm d-1

RT soil, 5 cm d-1

CT soil, 20 cm d-1

RT soil, 20 cm d-1

Fig. 9. Depth averaged retardation factors for CT and RT solute transports at flow

rates of 5 and 20 cm per day (cm d�1).

A. Besson et al. / Soil & Tillage Research 112 (2011) 47–57 53

To better understand the mixing regime in the lysimeter, thedepth average analysis is fundamental since it integrates the soilhorizontal variability.

3.2. Analysis of depth averaged transport parameters

Depth averaged BTC curves were obtained from horizontallyaveraged local scale BTCs at three soil depths (�15, �45 and�75 cm) (Fig. 8). As expected from the analysis of local scale curvesand irrespective of flow rates and soil depths, differences in shapesof BTCs are obtained between the CT and the RT soils. The longtailing of BTCs as observed in RT soil suggests the implication ofzones with relative low permeability in transport. As shown inFig. 9, the depth averaged retardation factors for the RT soil exceed1 for both flow rates and are higher than the CT factors. Thissuggests a retarded solute transport in the RT soil as compared tothe water flow and the more rapid CT transport. This is consistentwith the results observed at the local scale. The correspondingdepth averaged parameters vm and lLm, are shown in Fig. 10. Aconstant velocity profile is observed for both treatments due to therelatively homogeneous water content profile for both flow ratesand soils (Table 2). Depth averaged velocity is higher for the CTthan for the RT soil and depends also on the flow rate.

The depth averaged dispersivity lengths are again lower for CTthan for RT. In the CT soil, solute transport is then more rapid andless dispersive than in the RT soil. The results of the depth average

[(Fig._8)TD$FIG]

0.0000

0.0001

0.0002

0.0003

0.0004

0.0005

0.0006

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

0.00000

0.00005

0.00010

0.00015

0.00020

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Pore volume

*rtmC

*rtmC

-15

-45

-75

RT soil, 5 cm d-1

CT soil, 5 cm d-1

Soil depths in columns

Fig. 8. Depth averaged BTCs modeled for three depths zm of CT

analysis regarding transport parameters comfort those obtained atthe local scale. Additionally, both profiles exhibit an increasingtrend of dispersivity with flow rate and with depth, whichcharacterizes a stochastic convective mixing regime.

4. Discussion

4.1. Mixing regime and flow rate influence

Independent of flow rates and tillage operations, the dis-persivity length increases with depth is a characteristic of thestochastic-convective (SC) process. This shows that solutes wereincompletely mixed laterally between pore water regions. Hence,the mixing time must have been longer than the average traveltime. For this transport process type (or mixing regime) theclassical convection dispersion equation (CDE) fails to describethe observed dispersion evolution with depth. Stream tubemodels, which assume a constant velocity and no lateral solutemixing, would be more adapted to predict solute transport (Jury,1982; Simmons, 1982). However, as pointed out by Vanderborghtet al. (2001), lL can be used as an apparent dispersivity coefficientfor the mixing regime, when obtained by fitting Eq. (4) to BTCs atdifferent depths.

0.0000

0.0002

0.0004

0.0006

0.0008

0.0010

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10

0.0000

0.0001

0.0002

0.0003

0.0004

0.0005

0.0006

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10

Pore volume

CT soil, 20 cm d-1

RT soil, 20 cm d-1

and RT soils at flow rates of 5 and 20 cm per day (cm d�1).

[(Fig._10)TD$FIG]

-80-70-60-50-40-30-20-10

00 0.005 0.01 0.015 0.02 0.025 0.03

Soil

dept

h zm

cm

vm cm min-1

RT Soil, 5 cm d-1

CT soil, 5 cm d-1

RT soil, 20 cm d-1

CT soil, 20 cm d-1

-80-70-60-50-40-30-20-10

00 10 20 30 40 50 60 70 80

Soil

dept

h zm

cm

λLm cm

a. b.

Fig. 10. Depth averaged transport parameters estimated from BTCs (Fig. 8) for CT and RT soils at flow rates of 5 and 20 cm per day (cm d�1). Error bars correspond to

approximate confidence intervals. (a) Mean solute velocity, vm; (b) Dispersivity, lLm.

[(Fig._11)TD$FIG]

Bulk density g cm-3

Topsoil

Ploughed or not layer

Plough pan

A horizon

Bt horizon

RT soil

CT soil

0-5

-10-15-20-25-30-35-40-45-50-55-60-65-70-75-80-85-90

1.3 1.35 1.4 1.45 1.5 1.55 1.6 1.65 1.7 1.75

Soil

dept

hcm

0-5

-10-15-20-25-30-35-40-45-50-55-60-65-70-75-80-85-90

1.3 1.35 1.4 1.45 1.5 1.55 1.6 1.65 1.7 1.75

Fig. 11. Bulk densities measured on soil cores sampled in CT (n = 39) and RT (n = 30)

soils at several soil depths.

A. Besson et al. / Soil & Tillage Research 112 (2011) 47–5754

The SC transport process could be explained by the high watercontent obtained along the soil profiles at the column scale (Table2) as a consequence of high flow rates imposed at the soil surface.Close to saturation, preferential flow paths such as wormholes,interraggregate cracks, large voids, macropores are likely to beactivated, thereby bypassing locally the solution in the soil matrix.This will generate more heterogeneous transport characterized bylong tailing in the BTCs and exchange between more and lessmobiles regions. Our results suggest the impact of flow rates onvelocities and also on the values of depth averaged dispersivity at aspecific soil depth. Whatever tillage operations initially applied onsoils, depth averaged dispersivity is higher at a flow rate of20 cm d�1 than at 5 cm d�1. This is consistent with the work ofVanderborght (1997) who obtained increased dispersivities withan increasing flow rate. Nevertheless this result contrasts withseveral other studies which reported that the plume dispersivitydecreases with increasing saturation degree and, hence, with flowrate (Jardine et al., 1993; Maraqa et al., 1997; Padilla et al., 1999).However, our result can only be interpreted as a trend. Indeed,water content was not substantially different between the twowater flow rates in the present experiment.

4.2. Impact of soil tillage on solute transport in relation to the soil

structure

Whereas the structure of the deeper soil layers, e.g. the Bthorizon, is relatively similar for both soils with many rootschannels and earth-worm holes crossing an overall dense andsilty–clay matrix, the soil surface, i.e. the A horizon, encompassessignificant structural differences depending on the tillage opera-tions (Fig. 1). The analysis of bulk densities in depth, as shown inFig. 11, enables to comfort visual morphological descriptionsrealized in-field. As far as the bulk density of the Bt horizon isconcerned, mean values are 1.59 � 0.04 g cm�3 for CT (n = 24) and1.57 � 0.05 g cm�3 for RT (n = 12) (Fig. 11). No significant differences(Student test, 5% probability) were obtained between CT and RT at thelevel of the entire profile and for the Bt horizon. On the contrary, thebulk density of the A horizon in the RT soil (1.59 � 0.11 g cm�3,n = 14) is higher and more variable than in the CT soil(1.46 � 0.07 g cm�3, n = 12). It is consistent with numerous short-term studies that show that the bulk density increase in conserva-tional tilled soils as compared to the conventional tilled soils (Starr,1990; Alegre et al., 1991; Osunbitan et al., 2005). We precise still thatbulk densities were local measurements obtained on soil coressampled in-field. This means that, for instance, they do not inform onthe presence of large voids, though acting considerably on solutetransport. Consequently bulk density are interpreted and used as aproxy of the soil structure in the next.

Depending on the tillage operations and natural pedogenetichorizonation, three main morphological structures can be identi-fied at the soil profile scale.

4.3. The topsoil (0–15 cm depth)

Although the first 10 cm of the soils were harrowed for bothsoils, the harrowing operation was not performed under thesame soil condition. Indeed, prior to harrowing, the structure ofthe soil surface in the CT soil is yet strongly disturbed anddisaggregated due to the initial ploughing as opposed to theunploughed RT soil. As a consequence, the RT topsoil is moreheterogeneous, encompasses dense clods separated by porouszones whereas a finer, more uniform and porous structure isobtained in the CT topsoil (Fig. 1). Consequently, as shown inFig. 11, RT bulk densities in the topsoil are higher (mean value:1.57 � 0.11 g cm�3, n = 8) than in the CT soil (1.44 � 0.08 g cm�3,n = 6) and more variable as suggested by standard errors. Thisdifference in soil structure has a drastic impact on solute transport.At �15 cm depth, the solute movement in the CT soil is more rapid,homogeneous and less dispersive than in the RT soil (Figs. 6 and9). The latter soil is further characterized by BTCs that reflect a lowpermeability zone probably induced by the presence of denseclods. Coquet et al. (2005) also observed such results for acultivated silt loam soil and attributed it to the umbrella (orshadow) effects of compacted soil clods.

A. Besson et al. / Soil & Tillage Research 112 (2011) 47–57 55

4.4. The recently ploughed or unploughed layer (15–30 cm depth)

The remainder of the A horizon also exhibits a difference interms of soil structure and solute transport.

For CT, the intense soil perturbation due to the ploughinginvolved the burying of undecomposed organic residues and thefragmentation of the initial soil structure. As a consequence,organic residues such as maize stalks, dense clods with variablesizes, cracked or not, and voids are juxtaposed in an overall porousmatrix that encompasses many roots and earth-worms holes(Fig. 1). The CT bulk densities obtained from soil cores sampled inthe 15–30 cm layer are in general relatively low, smaller than1.5 g cm�3 with a mean value of 1.48 � 0.06 g cm�3 (n = 6) (Fig. 11).Solute movements remain rapid, slightly dispersive and homoge-neous, following numerous preferential flow paths such as worm-holes, interraggregate cracks, large voids and macropores. Asreviewed by Jarvis (2007), macropore flow can still be generated inconventionally ploughed soils under intense rain along ped surfacesand loose volumes between aggregates. In addition, enhancedpreferential solute transport directed towards the incorporatedstraw can also occur as shown by Kasteel et al. (2007).

Conversely, the no-tilled layer of the RT soil encompassesaggregates tightly packed with also many roots and earth-wormsholes. The RT bulk densities are high, variable and larger than1.55 g cm�3 with a mean value equal to 1.61 � 0.11 g cm�3 (n = 6)(Fig. 11), probably as a result of the soil natural consolidation orcompaction due to traffic. In this context, the macropore domain(structural pore domain) is modified mainly by deforming theaggregates during compaction (Gupta et al., 1989), by reducing thepore space between aggregates (Kooistra, 1994) and by inducing theformation of relict structural pores within the textural network(Bruand and Cousin, 1995; Bruand et al., 1997). Richard et al. (2001)suggest that a smaller proportion of the water contributes to thewater movement in the compacted layer due to the relict structuralpores. The decrease of the macropore space combined with thedistortion of structural porosity due to compaction could affectsolute transport, i.e. it could increase the dispersivity and theretardation. We show that, for the RT soil, solute transport reachesslowly the base of the A horizon and is more dispersive andheterogeneous than for the CT soil. Transport is still stochastic-convective. Solutes bypass the solution in the dense matrix ofaggregates and follow rather preferential paths such as wormholesand activated macropores. It is consistent with the work of Coquet etal. (2005) who show that very little water and bromide penetrate thelarge compacted zones under the wheel tracks. The compacted clodsin the A horizon should act as low-permeability barriers that divertwater and solute flow around them and finally accentuate thetransport heterogeneity. Kulli et al. (2003) showed that compactionunder wheel tracks can promote preferential flow into earthwormburrows. Solute fluxes can be redirect laterally towards morepermeable and porous zones. Consequently, like at �15 cm depth,the umbrella effects of compacted soil clods are yet visible on BTCs at�30 cm and �45 cm depth.

4.5. The soil layering and specific morphological features such as the

plough pan (35–100 cm depth)

For both soils, the solute behaviour induced by the differencesin topsoil structure (0–15 cm depth), seems to persist in theremainder of the A horizon and also in the deep no tilled naturalhorizons. Below 45 cm depth, solute transport is rapid, lessdispersive and homogeneous for the CT soil and heterogeneous forthe RT soil. This demonstrates the limited influence of the soillayering and the plough pan on solute transport. Both soils arecomposed of an upper loamy A horizon (66% in loam) overlying anargic Bt horizon with a higher clay content (25%). They both

present a plough pan at �35 cm depth. Impermeable andcemented layers such as plough pans can act as flow barriers,and induce lateral solute mixing and solute redistribution(Vanderborght et al., 2001; Janssen and Lennartz, 2008).Moreover, the interfaces between different soil layers canenhance lateral solute mixing (Roth et al., 1991). Yet, suchresults were not observed for both soils in the present experiment.For both soils, solute transport process is SC. This could beexplained by the high flow rates applied (5 cm d�1 and 20 cm d�1)which probably exceeded the saturated hydraulic conductivity ofthe soil matrix that is expected in medium to fine textured soilmaterials. Under field conditions, irrigation larger than about1 mm h�1 has been shown to generate non-equilibrium flow andtransport through active macropores for loamy-clayed soils (Gishet al., 2004). The macropore flow, resulting from root channelsand wormholes that are not sealed by ploughing and that cross theBt horizon and the plough pan, is important and reduces lateralsolute mixing. In such case, the soil layering and the plough pan donot show strong impermeable behaviour (Sander and Gerke,2007; Roulier et al., 2002).

Hence, the results observed at �15 cm depth are persistent indepth: solute transport remains relatively homogeneous, rapid andless dispersive in the deepest layers of the CT soil and, conversely,heterogeneous, slow and more dispersive in the RT soil.

4.6. Retardation factor of solute transport and preferential flow

pathways

The analysis of retardation factors allows analyzing theimportance of preferential flow during flow and transportexperiments (Allaire et al., 2002). R values obtained at thelocal-scale and at depth averaged scale suggest differencesbetween treatments in transport variability and mixing level.Whereas transport is relatively homogeneous in CT soil (R close to1), the overall solute migration in RT soil is largely retarded andmore heterogeneous.

Vanderborght and Vereecken (2007) reviewed different trans-port experiments in undisturbed soils at different scales. Forexperimental conditions relatively similar to those of the presentstudy, retardation values were generally close to 0.8–1, but extremevalues close to 2.5 were also obtained. As shown by Vanderborghtet al. (2001) for Belgian soil types, lateral mixing is promoted whencemented soil layers are present in the soil profile. Tightlycompacted zones and clods may force solute transport to occur insmaller low mobile transport domains (Heitman et al., 2007), whichmay create the apparent retardation effect.

Moreover, fast or preferential flow (R close or smaller than 1)has been assumed to be mainly related to macroporous flow, i.e.

flow along ped faces, loose volumes or interraggregate cracks. Thisis consistent with findings from leaching experiments realizedunder conventional system and persistent intense rain (Petersenet al., 1997; Gjettermann et al., 1997; Schwartz et al., 1999).Several studies showed that intense ploughing can disrupt and, tosome extent, destroy the continuity of biopores (Shipitalo et al.,2000; Vervoort et al., 2001). We, however, cannot exclude thecontribution of regenerated earthworm holes on preferentialsolute transport for both CT and RT. The two soil columns werestored during one year before the solute transport experimentswere performed. During such a time lapse wormholes can bereestablished and structural change can be induced. Likewisestructural changes could also occur during the experiment. Indeedretardation factors obtained at 20 cm d�1 tend to be higher than at5 cm d�1. The first infiltration experiment realized at 5 cm d�1

could modify the soil structure and therefore could impact thesecond experiment realized at 20 cm d�1. Then we cannot excludean eventual retroaction between soil structure changes and

A. Besson et al. / Soil & Tillage Research 112 (2011) 47–5756

transport according to the experimental procedure. For instance,Kodesova et al. (2009) showed that water flux is affectedsignificantly during the experiment due to aggregate breakdownin the soil. Soil particles leaching as observed for Luvisol (FAO, 2006)could also modify macropore volume and then transport properties.Given the high flow rate applied in our experiments, these structuralchanges could be significant for both soils. In addition tomacroporous flow, additional preferential flow phenomena suchas fingered flow can be invoked to explain the observations.Structural discontinuities related to undisturbed and dense seedbedaggregates located in a general porous matrix can generate localsaturated zones due to capillary barriers (Javaux and Vanclooster,2004). Flow fingers could then emerge as a consequence of wettingfront instabilities induced by structural interfaces.

Yet other results could be encountered under differentexperimental conditions. Indeed our study was performed withrelatively high flow rates, and hence, near-saturated soils wherethe macropore flow dominate the transport process. As mentionedabove, the dispersivity depends on the actual water potential andapplied flow rates (Javaux and Vanclooster, 2003b). It seemstherefore that the spatial variability in macroscopic hydraulicproperties and in different macroscopic preferential flow paths,influence our results on the transport process. Results might bedifferent for lower flow rates and for drier soils. In such cases,solute trajectories could be mainly situated within the microporedomain (i.e. the lacunar pore domain), lateral solute mixing couldbe enhanced and the impact of soil layering on the relevanttransport could be significant.

Further experiments, such as dye tracer experiments, shouldallow evaluating the mechanisms explaining the fast or preferen-tial flow and demonstrate for instance to what extent regenerationof macropores occurred in the lysimeter. However, such dye tracerexperiments were beyond the scope of the present study.

5. Conclusions

In this study, we analyze the short term impact of theconversion from conventional mouldboard ploughing (CT) toreduced disc harrowing (RT) on the solute transport process withina loamy soil. Solute breakthrough experiments at two flow rateswere performed on 2 undisturbed lysimeters collected in a CT andRT field plot. Solute transport parameters were estimated usingtransfer function theory.

It have been shown that, compared to the CT soil, transport in theRT soil was more heterogeneous with higher dispersivity lengths,smaller local velocities and larger retardation factors. This could beexplained by the differences in soil structure due to tillage practicesimplemented in-field. Soil disturbance and the soil structurepulverization were more intense in the CT soil than in the RT soil.This last one exhibited then several dense seedbed aggregates whichrestricted and diverged solute leaching.

For the RT soil, the presence of a plough pan on solute transportwas visible, inducing an apparent retardation on solute transport.It also appeared that that tillage treatment effect in the surfacehorizons on solute transport is persistent in depth.

In all cases, the increase of the dispersivity with depth suggeststhat the dominant mixing regime is stochastic-convective, i.e.

little mixing occurs between fast flow paths and soil matrix.However, we pointed out that the solute transport is stronglydependent on the boundary conditions such as the flow rate. Ourexperiments were performed with rather high flow rates forwhich the macropore domain is largely activated. The impact oftillage on the transport process might be different for lower flowrates. In addition the complex relation between soil structure andtransport should further be explored by taking into account thecumulative effects of tillage combined with different incidental

soil processes, such as soil natural reconsolidation, soil crackingand sealing, compaction due to traffic, wetting/drying or plantgrowth, in space and in time. For instance, our study focused onshort-term tillage effects and could be improved by event-basedand seasonal/annual analysis. The persistence of short-termresponses due to the soil management on solute transport couldalso be clarified from a long-term study based on suchmechanistic approach.

Acknowledgements

The authors wish to thank G. Rentmeesters from ‘‘Genie Rural,UCL, Louvain-la-Neuve’’ and C. Legout from ‘‘LTHE, Grenoble,France’’. They are grateful to the FSR and FNRS for financial support.

References

Alegre, J.C., Cassel, D.K., Amezquita, E., 1991. Tillage systems and soil properties inLatin America. Soil and Tillage Research 20, 147–163.

Allaire, S.E., Gupta, S.C., Nieber, J., Moncrief, J.F., 2002. Role of macropore continuityand tortuosity on solute transport in soils: 1. Effects of initial and boundaryconditions. Journal of Contaminant Hydrology 58, 299–321.

Alletto, L., Coquet, Y., Benoit, P., Heddadj, D., Barriuso, E., 2010. Tillage managementeffects on pesticide fate in soils: a review. Agronomy for Sustainable Develop-ment 30, 367–400.

Bruand, A., Cousin, I., 1995. Variation of textural porosity of a loamy-clay soil duringcompaction. European Journal of Soil Science 46, 377–385.

Bruand, A., Cousin, I., Le Lay, C., 1997. Formation of relict macropores in a clay-loamy soil by wheel compaction. In: Shoba, S., Gerasimova, M., Miedema, R.(Eds.), Soil Micromorphology: Studies on Soil (Diversity, Diagnostics and Dy-namics). Printing Service Centre Van Gils B.V., Wageningen, pp. 247–254.

Casa, R., Lo Cascio, B., 2008. Soil conservation tillage effects on yield and water useefficiency on irrigated crops in central Italy. Journal of Agronomy and CropScience 194, 310–319.

Clay, S.A., Clay, D.E., Koskinen, W.C., Berg, R.K., 1998. Application method: impactson atrazine and alachlor movement, weed control, and corn yield in three tillagesystems. Soil and Tillage Research 48, 215–224.

Coquet, Y., Coutadeur, C., Labata, C., Vachier, P., van Genuchten, M.Th, Roger-Estrade, J., Simunek, J., 2005. Water and solute transport in a cultivated siltloam soil 1. Field observations. Vadose Zone Journal 4, 573–586.

Dagan, G., Nguyen, V., 1989. A comparison of travel time and concentrationapproaches to modeling transport by groundwater. Journal of ContaminantHydrology 4, 79–91.

De Vita, P., Di Paolo, E., Fecondo, G., Di Fonzo, N., Pisante, M., 2007. No-tillage andconventional tillage effects on durum wheat yield, grain quality and soilmoisture content in southern Italy. Soil and Tillage Research 92, 69–78.

FAO, 2006. World Reference Base for Soil Resources. FAO, Rome, Italy.Gjettermann, B., Nielsen, K.L., Petersen, C.T., Jensen, H.E., Hansen, S., 1997. Prefer-

ential flow in sandy loam soils as affected by irrigation intensity. Soil Technol-ogy 11, 139–152.

Gish, T.J., Helling, C.S., Mojasevic, M., 1991. Preferential movement of atrazine andcyanazine under field conditions. Transactions of the ASAE 34, 1699–1705.

Gish, T.J., Kung, K.-J.S., Perry, D.C., Posner, J., Bubenzer, G., Helling, C.S., Kladivko, E.J.,Steenhuis, T.S., 2004. Impact of preferential flow at varying irrigation rates byquantifying mass fluxes. Journal Environmental Quality 33, 1033–1040.

Granovsky, A.V., McCoy, E.L., Dick, W.A., Shipitalo, M.J., Edwards, W.M., 1993. Waterand chemical transport through long term no-till and plowed soils. Soil ScienceSociety America Journal 57, 1560–1567.

Gupta, S., Sharma, P.P., Defranchi, S.A., 1989. Compaction effects on soil structure.Advances in Agronomy 42, 311–338.

Hatfield, J.L., Sauer, T.J., Prueger, J.H., 2001. Managing soils to achieve greater wateruse efficiency: a review. Agronomy Journal 93, 271–280.

Heitman, J.L., Gaurb, A., Hortonc, R., Jaynesd, D.B., Kaspard, T.C., 2007. Field mea-surement of soil surface chemical transport properties for comparison ofmanagement zones. Soil Science Society of America Journal 529–536.

Holland, J.M., 2004. The environmental consequences of adopting conservationtillage in Europe: reviewing the evidence. Agriculture Ecosystems and Envi-ronment 103, 1–25.

Isensee, A.R., Sadeghi, A.M., 1997. Interactions of tillage and rainfall on atrazineleaching under field and laboratory conditions. Chemosphere 34, 2715–2723.

Janssen, M., Lennartz, B., 2008. Characterization of preferential flow pathwaysthrough paddy bunds with dye tracer tests. Soil Science Society of AmericaJournal 72, 1756–1766.

Jardine, P.M., Jacobs, G.K., Wilson, G.V., 1993. Unsaturated transport processes inundisturbed heterogeneous porous media: I. Inorganic contaminants. SoilScience Society of America Journal 57, 945–953.

Jarvis, N.J., 2007. A review of non-equilibrium water flow and solute transport in soilmacropores: principles, controlling factors and consequences for water quality.European Journal of Soil Science 58, 523–546.

A. Besson et al. / Soil & Tillage Research 112 (2011) 47–57 57

Javaux, M., 2004a. CASlib: A Matlab Library for Convection-dispersion AnalyticalSolutions. Reference and User’s Manual. Universite catholique de Louvain,Louvain la Neuve, Belgium.

Javaux, M., 2004b. Solute transport in a heterogeneous unsaturated subsoil: experi-ments and modeling. PhD Thesis. Dep. AGRO/MILA, Universite catholique deLouvain, Louvain la Neuve, Belgium, 194 pp.

Javaux, M., Vanclooster, M., 2003a. Robust estimation of the generalized solutetransfer function parameters. Soil Science Society of America Journal 67, 81–91.

Javaux, M., Vanclooster, M., 2003b. Scale- and rate-dependent solute transportwithin an unsaturated sandy monolith. Soil Science Society of America Journal67, 1334–1343.

Javaux, M., Vanclooster, M., 2004. In situ long-term chloride transport through alayered nonsaturated subsoil. 2. Effect of layering on solute transport processes.Vadose Zone Journal 3, 1331–1339.

Javaux, M., Vanclooster, M., 2006. Scale-dependency of the hydraulic properties of avariably saturated heterogeneous sandy subsoil. Journal of Hydrology 327,376–388.

Javaux, M., Kasteel, R., Vanderborght, J., Vanclooster, M., 2006. Interpretation of dyetransport in a macroscopically heterogeneous unsaturated subsoil with a 1-Dmodel. Vadose Zone Journal 5, 529–538.

Jury, W.A., 1982. Simulation of solute transport using a transfer function model.Water Resources Research 18, 363–368.

Jury, W.A., Roth, K., 1990. Transfer Functions and Solute Movement Through Soil:Theory and Applications. Birkhauser Verlag, Basel, 226 pp.

Kasteel, R., Garnier, P., Vachier, P., Coquet, Y., 2007. Dye tracer infiltration in theplough layer after straw incorporation. Geoderma 137 (3–4), 360–369.

Keller, G.V., Frischknecht, F.C., 1966. Electrical Methods in Geophysical Prospecting.Pergamon Press, Oxford.

Kepner, R.A., Bainer, R., Barger, E.L., 1978. Principles of Farm Machinery. Wesport,CT, USA.

Kooistra, M.J., 1994. Effects of compaction on soil microstructure. In: Soane, B.D., vanOuwerkerk, C. (Eds.), Soil Compaction in Crop Production. Elsevier, Amsterdam,pp. 91–111.

Kool, J.B., Parker, J.C., van Genuchten, M.Th., 1987. Parameter estimation forunsaturated flow and transport models – a review. Journal of Hydrology 91,255–293.

Kodesova, R., Nadia Vignozzi, N., Rohoskova, M., Hajkova, T., Kocarek, M., Pagliai, M.,Kozak, J., Simunek, J., 2009. Impact of varying soil structure on transportprocesses in different diagnostic horizons of three soil types. Journal of Con-taminant Hydrology 104, 107–125.

Kulli, B., Gysi, M., Fluhler, H., 2003. Visualizing soil compaction based on flowpattern analysis. Soil and Tillage Research 70, 29–40.

Lambot, S., Javaux, M., Hupet, F., Vanclooster, M., 2002. A global multilevel coordi-nate search procedure for estimating the unsaturated soil hydraulic properties.Water Resources Research 38, 1222–1239.

Levanon, D., Codling, E.E., Meisinger, J.J., Starr, J.L., 1993. Mobility of agrochemicalsthrough soil from two tillage systems. Journal of Environmental Quality 22,155–161.

Logsdon, S.D., Jaynes, D.B., 1996. Spatial variability of hydraulic conductivity in acultivated field at different times. Soil Science Society of America Journal 60,703–709.

Logsdon, S.D., Jordahl, J.L., Karlen, D.L., 1993. Tillage and crop effects on ponded andtension infiltration rates. Soil and Tillage Research 28, 179–189.

Mallants, D., Vanclooster, M., Toride, N., Vanderborght, J., van Genuchten, M.Th.,Feyen, J., 1996. Comparison of three methods to calibrate TDR for monitoringsolute movement in undisturbed soil. Soil Science Society of America Journal 60,747–754.

Maraqa, M.A., Wallace, R.B., Voice, T.C., 1997. Effects of degree of water saturationon dispersivity and immobile water in sandy soil columns. Journal of Contami-nant Hydrology 25, 199–218.

Masse, L., Patni, N.K., Jui, P.Y., Clegg, B.S., 1996. Effluent quality and chemical lossesunder conventional and no tillage – Part 2: Atrazine and metolachlor. Transac-tions of the ASAE 39, 1673–1679.

Mathworks Inc., 2008. Matlab. The Language of Technical Computing. ReleaseR2008a. Mathworks Inc., Natick, MA, USA.

Moret, D., Arruea, J.L., 2007. Dynamics of soil hydraulic properties during fallow asaffected by tillage. Soil and Tillage Research 96, 103–113.

Osunbitan, J.A., Oyedele, D.J., Adekalu, K.O., 2005. Tillage effects on bulk density,hydraulic conductivity and strength of a loamy sand soil in southwesternNigeria. Soil and Tillage Research 82, 57–64.

Padilla, I.Y., Yeh, T.-C.J., Conklin, M.H., 1999. The effect of water content on solutetransport in unsaturated porous media. Water Resources Research 35, 3303–3313.

Petersen, C.T., Hansen, S., Jensen, H.E., 1997. Depth distribution of preferential flowpatterns in a sandy loam soil as affected by tillage. Hydrology and Earth SystemSciences 1, 769–776.

Petersen, C.T., Jensen, H.E., Hansen, S., Bender Koch, C., 2001. Susceptibility of asandy loam soil to preferential flow as affected by tillage. Soil and TillageResearch 58, 81–89.

Richard, G., Cousin, I., Sillon, J.F., Bruand, A., Guerif, J., 2001. Effect of compaction onsoil porosity: consequences on hydraulic properties. European Journal of SoilScience 52, 49–58.

Roth, K., Jury, W.A., Flulher, H., Attinger, W., 1991. Transport of chloride through anunsaturated field soil. Water Resources Research 27, 2533–2541.

Roulier, S., Angulo-Jaramillo, R., Bresson, L.-M., Auzet, A.-V., Gaudet, J.-P., Bariac, T.,2002. Water transfer and mobile water content measurement in a cultivatedcrusted soil. Soil Science 167, 201–210.

Sander, T., Gerke, H.G., 2007. Preferential flow patterns in paddy fields using a dyetracer. Vadose Zone Journal 6, 105–115.

Schwartz, R.C., McInnes, K.J., Juo, A.S.R., Cervantes, C.E., 1999. The vertical distribu-tion of a dye tracer in a layered soil. Soil Science 164, 561–573.

Simmons, C.S., 1982. A stochastic-convective transport representation of dispersionin one-dimensional porous media systems. Water Resources Research 18,1193–1214.

Shipitalo, M.J., Edwards, W.M., 1993. Seasonal patterns of water and chemicalmovement in tilled and no-till column lysimeters. Soil Science Society ofAmerica Journal 57, 218–223.

Shipitalo, M.J., Dick, W.A., Edwards, W.M., 2000. Conservation tillage and macroporefactors that affect water movement and the fate of chemicals. Soil and TillageResearch 53, 167–183.

Starr, J.L., 1990. Spatial and temporal variation of ponded infiltration. Soil ScienceSociety of America Journal 54, 629–636.

Strudley, M.W., Green, T.R., Ascough II, J.C., 2008. Tillage effects on soil hydraulicproperties in space and time: state of the science. Soil and Tillage Research 99,4–48.

Topp, G.C., Davis, J.L., Annan, A.P., 1980. Electromagnetic determination of soil watercontent: measurements in coaxial transmission lines. Water Resources Re-search 16, 574–582.

Turner, N.C., 2004. Agronomic options for improving rainfall-use efficiency of cropsin dryland farming systems. Journal of Experimental Botany 55, 2413–2425.

van Doren, D.M., Allmaras, R.R., 1978. The effect of residue management practice onthe soil physical environment, microclimate and plant growth. In: Crop ResidueManagement Systems. American Society of Agronomy, Madison, WI.

Vanclooster, M., Javaux, M., Vanderborght, J., 2005. Solute transport in soil at thecore and field scale. In: Anderson, M.G., McDonnell, J.J. (Eds.), Encyclopedia ofHydrological Sciences. Wiley.

Vanclooster, M., Mallants, D., Vanderborght, J., Diels, J., Feyen, J., 1995. Monitoringsolute transport in a multi-layered sandy lysimeter using time domain reflec-tometry. Soil Science Society of America Journal 59, 337–344.

Vanderborght, J., 1997. Experimental and Numerical Study on Non-reactive SoluteTransport in Soils. Katholieke Universteit Leuven, Leuven, Belgium, 310 pp.

Vanderborght, J., Vanclooster, M., Mallants, D., Diels, J., Feyen, J., 1996. Determiningconvective lognormal solute transport parameters from resident concentrationdata. Soil Science Society of America Journal 60, 1306–1317.

Vanderborght, J., Vanclooster, M., Timmerman, A., Seuntjens, P., Mallants, D., Kim,D.J., Jacques, D., Hubrechts, L., Gonzalez, C., Feyen, J., Diels, J., Deckers, J., 2001.Overview of inert tracer experiment in key Belgian soil types: relation betweentransport, and soil morphological and hydraulic properties. Water ResourcesResearch 37, 2873–2888.

Vanderborght, J., Vereecken, H., 2007. Review of dispersivities for transport model-ing in soils. Vadose Zone Journal 6, 29–52.

Vervoort, R.W., Dabney, S.M., Romkens, M.J.M., 2001. Soil Science Society of AmericaJournal 65, 1227–1234.

Vogeler, I., Horn, R., Wetzel, H., Krummelbein, J., 2006. Tillage effects on soil strengthand solute transport. Soil and Tillage Research 88, 193–204.