Geoderma 116 (2003) 29 – 46

www.elsevier.com/locate/geoderma

Tillage and traffic effects on soil hydraulic

properties and evaporation
J.F. Sillon a, G. Richard a,*, I. Cousin b
a
INRA, Unité d’Agronomie de Laon-Péronne, 02007 Laon Cedex, France
b
INRA, Unité de Science du Sol-SESCPF, Centre de Recherche d’Orléans, BP 20619,
45166 Olivet Cedex, France

Abstract

Evaporation is a major component of the water loss from the soil whose structure is modified by
traffic and tillage. This study was undertaken to analyse, in field conditions, the effect of tillage and
traffic on soil structure and evaporation, and to determine the role of the change in hydraulic
properties on soil drying using water transfer model. Three structures of the ploughed layer were
formed in a loess soil (Luvisol Orthique) and a calcareous soil (Rendzina): a fragmentary structure
created by deep soil tillage in autumn or in spring (rotary tiller at 30 cm depth), a compacted
ploughed layer created by compaction under wet conditions. The bulk density varied from 1.16 to
1.63 Mg m
3 in the loess soil, from 1.00 to 1.45 Mg m
3 in the calcareous soil. Evaporation was
calculated from the change in soil water content and matric water potential profiles measured during
the spring season. Soil hydraulic properties were estimated using an inverse modelling method
applied to field measurements of water content and water potential or the Wind method. Soil
structure greatly affected the drying of the calcareous soil: the evaporation of the compacted plot was
about two times that of the tilled plot. The compacted plot dried out homogeneously, with a soil
surface which remained wet. Evaporation mainly concerned the first 15 cm of the ploughed layer
created by autumn or spring tillage. This effect of soil structure on evaporation was not observed in
the loess soil. The unsaturated hydraulic conductivity was higher in the compacted plot than in the
tilled plots in the calcareous soil. It was similar in the three plots of the loess soil, because of the
formation of relict structural pores by compaction. Experimental and numerical results showed that
unsaturated hydraulic conductivity is of major concern in soil drying and that the albedo and surface
roughness have minor effects if any. The possible relict structural pores have to be characterised in
various soils, as a function of soil sensitivity to compaction, traffic and tillage conditions.
D 2003 Elsevier Science B.V. All rights reserved.

Keywords: Soil structure; Soil drying; Retention curve; Hydraulic conductivity; Water transfer model

* Corresponding author. Tel.: +33-3-23-23-64-86; fax: +33-3-23-79-36-15.

E-mail address: richard@laon.inra.fr (G. Richard).

0016-7061/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0016-7061(03)00092-2
30 J.F. Sillon et al. / Geoderma 116 (2003) 29–46

1. Introduction

Tillage alters the structure of cultivated soils and the localisation of crop residues and
thus affects, in conjunction with climate, the soil water balance. Tillage can influence the
evaporation process because of its effects on the physical properties of the soil surface
(albedo, roughness, mulch) and on the tilled layer (hydraulic properties). Numerous
experimental studies have compared the effects of conventional tillage to those of no
tillage practices on soil water regimes (Negi et al., 1981; Gauer et al., 1982; Johnson et
al., 1984; Radke et al., 1985; Wu et al., 1992; Benjamin, 1993). Water evaporation is
generally reduced by no-tillage because of the mulch formed by crop residues at the soil
surface. Thus the soil surface water content is generally greater on no-tilled plots than on
conventional tilled plots during dry periods (Radke et al., 1985). On the contrary, soil
tillage has been shown to increase short-term evaporation (Holmes et al., 1960; Allmaras
et al., 1977) and reduce long term evaporation (Gill et al., 1977), as a function of
evaporative demands (Mwendera and Feyen, 1994). The effect of various tilled layer
structures, without considering the effect of crop residues, on soil drying has been
investigated less. Similar soil surface water contents were obtained in no-tilled and
conventional tilled plots by Gauer et al. (1982) when wheat straw was burned. Richard
and Cellier (1998) have found that surface water content decreased more rapidly after a
superficial tillage of bare soil. Warkentin (1971) and Agrawal and Phogat (1998)
observed an increase of volumetric water content with the degree of soil compactness.
Several attempts have been made to apply evaporation and soil water transfer models to
various tillage conditions (Hammel et al., 1981; Linden, 1982; Culley et al., 1987). Two-
dimensional models have also been developed to study the effect of ridge tillage
(Benjamin et al., 1990) or partial soil surface cover with mulch (Chung and Horton,
1987). The use of these water transfer models implies estimating the hydraulic properties
of tilled soil, which are very important in simulating the soil water regime (Linden, 1982;
Witono and Bruckler, 1989). However, measurements of hydraulic conductivity as a
function of soil water content are particularly scarce and mainly concern only saturated
conditions. The objective of the work was therefore (1) to study, in field conditions, the
effect of tillage and traffic on soil structure and soil drying; (2) to characterise the change
in hydraulic properties by soil structure; (3) to analyse the change in soil drying using
water transfer model.

2. Material and methods

2.1. Field experiment

The study was conducted on two of the main soils in northern France, a loess soil and a
calcareous soil. The loess soil and the calcareous soil were, respectively, a Luvisol
Orthique and a Rendzina in the FAO classification, a Typic Hapludalf and a Typic Rendoll
in the USDA classification. The field experiments were carried out in spring, from March
to June 1995 for the loess soil, and from March to June 1996 for the calcareous soil. They
were performed on a 25  30 m2 field at the experimental Research Centre of INRA-Mons
 J.F. Sillon et al. / Geoderma 116 (2003) 29–46 31

en Chaussée (Somme) for the loess soil, and at the experimental Research Centre of
INRA-Châlons (Marne) for the calcareous soil.
The two experiments were carried out following a wheat crop grown in fields which are
moldboard ploughed annually. The wheat straw was burned after wheat harvesting. Each
experimental field was tilled in September with a digging machine (rotary tiller Dufalo,
FALC, Firenze, Italia, tillage depth = 0.3 m) to obtain as a fine soil structure as possible
without large clods in the ploughed layer. Three plots (25 m length, 10 m width),
corresponding to a range of soil structures from a massive ploughed layer to a loose
ploughed layer, were created in spring by varying traffic and tillage conditions. A fully
compacted ploughed layer was produced using a tractor (8.3 t with rear tyres of 65 cm
width inflated to 200 kPa) in wet conditions (water potential of
5 kPa) in March
(compacted soil or C). This was done by driving the tractor across the plot completely
covering the soil surface. A highly porous ploughed layer was produced by a new passage
of the digging machine (tillage depth = 0.3 m) under good weather conditions in April
(spring-tilled soil or ST). At this time, the mean water potential from 5 to 30 cm depth was

40 kPa in the loess soil and
60 kPa in the calcareous soil. No tillage and traffic were
performed in the third plot which therefore corresponded to an autumn-tilled soil (AT).
The effect of winter climate on the autumn-tilled soil was assumed to produce intermediate
porosity in the ploughed layer in spring, because of compaction and re-consolidation by
winter rainfall. The winter rainfall was 380 mm on the loess soil and 340 mm on the
calcareous soil.

2.2. Measurements

Measurements of textural properties in each plot were made to be sure of the textural
homogeneity of the field in each experimental site (Table 1). The soil particle density
(Table 1) was established for each treatment in the laboratory by pycnometer measure-
ments (four repetitions per treatment).
Bulk density was measured in the field using a gamma-ray probe with six
replications per treatment. Measurements were made every 5 cm of depth in the 5 –
60 cm layer. Soil cores (100 cm3) were sampled just under the soil surface to evaluate

Table 1
Physical properties of the loess and calcareous ploughed layers
Soil type Treatmenta Clay Silt Sand CaCO3 Organic Soil particle
< 2 Am 2 – 50 Am 50 – 2000 Am (g kg
1) carbone Densityb
(g kg
1) (g kg
1) (g kg
1) (g kg
1) (Mg m
3)
Loess C 154 790 51 5 10 2.68
AT 144 799 52 5 10 2.67
ST 140 803 52 5 10 2.69
Calcareous C 108 175 15 702 20 2.67
AT 109 179 17 695 20 2.67
ST 111 182 15 692 20 2.67
a
C: Compacted plot, AT: Autumn-tilled plot; ST: Spring-tilled plot.
b
Standard deviation: 0.02 Mg m
3.
32 J.F. Sillon et al. / Geoderma 116 (2003) 29–46

the bulk density of the first 5 cm of soil. The thickness of the ploughed layer was
estimated by measuring the distance from the bottom of the ploughed layer (located
with the change in soil color) to the soil surface every 5 cm over two soil profiles 5 m
in width. The aggregate size distribution was assessed within the ploughed layer by
sieving six soil samples of 0.014 m3 per treatment. Soil was sampled at a water content
around 0.2 g g
1 using a cylinder of 25 cm diameter and was sieved without drying.
Soil surface roughness was measured using a 2-m long pinmeter (the distance between
two pins was 1 cm). The surface roughness was considered as the standard deviation of
the differences between the height readings and a plane of best fit calculated for each
measurement (Currence and Lovely, 1970). All these measurements were performed in
May in the middle of each experiment. Continuous measurements of reflected solar
radiation in each plot (one pyranometer per plot, CM6 pyranometer, Kipp and Zonen,
Delft, Holland) and incident solar radiation (one pyranometer per experiment) allowed
the calculation of soil albedo under various weather conditions and soil surface water
content.
Soil drying was analysed using measurements of soil water content and water potential.
Changes in the water potential profile were measured daily using mercury tensiometers
installed at 2, 5, 10, 17, 25, 40 and 60 cm depths (three replicates per depth, 13 mm
diameter, Objectif K, Tours, France). Changes in water content profiles were characterised
by gravimetric measurement at 2 – 3 day intervals during drying periods. Measurements
were made every centimetre in the 0 –5 cm layer to provide an accurate soil water
description near the surface, every 5 cm in the 5 –35 cm layer, and every 10 cm in the 35–
65 cm layer (six replications per treatment). Evaporation was deduced from the water
content and the water potential profile using the ‘‘zero flux plane’’ method (Vachaud et al.,
1978). The change in soil moisture was characterised by the water ratio or volumetric
water content. Water ratio is the water volume for a given volume of the solid phase: it
does not depend on the change in soil bulk density by tillage and on particle density due to
soil constitution. Water ratio (#) was calculated from gravimetric water content (w) and
soil particle density (qs) measurements using the following expression:

# ¼ ðqs =qw Þw ð1Þ

where qw is the water density.

Relationships between matric water potential (w) and water ratio (#) were established
from field data for water potential >
80 kPa, and from laboratory data for water
potential <
80 kPa. Two layers were considered: the ploughed layer and the subsoil
(35 – 60 cm). Field data were pairs of values (w, #) measured within the same sub-layer at
the same time for each treatment. Laboratory data were obtained on small aggregates (2– 3
mm diameter) with Richard’s press method (Klute, 1986) and the salt solutions method
(O’Brien, 1948). Field and laboratory data were fitted in the same way with the equation
proposed by Van Genuchten (1980) to describe water-retention properties:
 m
1 1
Se ¼ with m ¼ 1
ð2Þ
1 þ ðawÞn n
 J.F. Sillon et al. / Geoderma 116 (2003) 29–46 33

where Se is the effective water content, w is the matric potential (m), a (m
1) and n are
fitting parameters. Se can be expressed using the volumetric water content or water
ratio:

#
#r h
hr
Se ¼ ¼ with h ¼ #q=qs ð3Þ
#s
#r hs
hr

where #r and #s denote residual and saturated water ratio, hr and hs. denote residual and
saturated volumetric water content, q denotes the soil bulk density. #s, #r, a and n were
estimated by minimising the sum of squared differences between the measured and
simulated water ratio for each measured water potential. Two sets of parameters were
established, either for water potential <
80 kPa or for water potential >80 kPa,
thereby imposing the continuity of the w(#) relationship.
Relationships between hydraulic conductivity (K) and water ratio (#) were estab-
lished in the laboratory using the Wind method (Wind, 1968) for water potentials
>
80 kPa. Five undisturbed soil samples (15 cm diameter and 7 cm height) were
taken from the 5 – 22 cm layer of each treatment. The cores were saturated using the
method of Richard et al. (2001b) to reproduce field rewetting conditions as accurately
as possible. The saturated cores were then subjected to evaporation, and the water
potentials were recorded. For a soil core, the Wind method gives a set of (#, K) pairs of
values, where K is the unsaturated hydraulic conductivity. For each treatment, we
calculated the mean hydraulic conductivity K from five soil cores for every increment
of 0.005 m3 m
3 for #.

2.3. Modelling of soil water transfer

The mechanistic model of Witono and Bruckler (1989) (named the TEC model) was
used (1) to estimate the K(#) relationship of each ploughed layer outside the tensiometric
range of water potential with an inversion procedure, and (2) to calculate the change in soil
water content as a function of climate conditions, initial conditions and soil parameters.
This model describes the water and heat transfer within the soil, and the energy balance at
the soil surface. It is based on heat and mass flow theory in partially saturated porous
media (Philip and De Vries, 1957). The non-linear partial differential equations of the soil
model describing changes in water potential and temperature are solved by a Galerkin
finite element method with time steps < 600 s.
The ploughed layer and the subsoil of each treatment were described by their dry bulk
density, thermal, hydrodynamic and gas diffusion properties (see Richard et al., 2001b).
Mean dry bulk density was deduced from field measurements. The water content-water
potential relationships used in the model were obtained from field and laboratory data
described by the van Genuchten formula (Eq. (3)) applied to two ranges of water
potential: <
80 and >
80 kPa. The relationships between water content and hydraulic
conductivity K[#] was described by a log –polynomial equation (third degree) as proposed
by Chanzy and Bruckler (1993):

log10 ½Kð#Þ ¼ a0 þ a1 # þ a2 #2 þ a3 #3 ð4Þ

34 J.F. Sillon et al. / Geoderma 116 (2003) 29–46

Thermal properties were estimated using the model of De Vries (1963). The water
vapour diffusion coefficient was calculated as a function of air-filled porosity, by using
the relationships of Bruckler et al. (1989). The soil surface of each treatment was de-
scribed by albedo, aerodynamic roughness and emissivity. Soil emissivity was equal to
0.95. Aerodynamic roughness (z0) was taken as a tenth of surface roughness (Brutsaert,
1982). Albedo depended on the water ratio of the superficial layer (#0 – 1 cm) as
proposed by Idso et al. (1975). Boundary conditions consisted of hourly meteorological
data (solar radiation, air temperature, air moisture and wind speed) measured at 2 m
height, and of daily measurements of water potential and temperature at 60 cm depth.
Initial conditions were obtained from field measurements of the water content and
temperature profile.
The model was applied to the 0 – 25 cm layer of each treatment to estimate the K(#)
relationship under dry conditions, in order to complete the estimation made by the Wind
method which concerned only water potential >
80 kPa. The four coefficients of log –
polynomial K(#) relationship were obtained by fitting the water content calculated
during a 7-day dry period with that observed daily in the 0 –1, 1 –2, 2 –3, 3 –4, 4 –6, 6–
10, 10 – 15, 15 – 20 and 20 –25 cm layers (least-squares fitting procedure, Marquart,
1963).

3. Results

3.1. Ploughed layer structure

In the calcareous soil, the structure of the autumn-tilled plot (AT) remained fine
and fragmentary in spring, with a low bulk density and few clods with a diameter >40
mm (Table 2). Consequently, the spring tillage, leading to the spring-tilled plot (ST),
did not lead to great change in bulk density and aggregate size distribution. In the

Table 2
Structural properties of the loess and calcareous ploughed layers
Date Thicknessa Mean bulk Aggregates
(cm) density < 2 mm >40 mm
(Mg m
3) (kg kg
1) (kg kg
1)
Calcareous soil ATb 17/10/1995 35.5b 1.06b 0.32b 0.04b
ST 04/04/1996 34.1b 1.00c 0.36b 0.01b
C 22/03/1996 24.9ac 1.45a 0a 1a
Loess soil AT 10/10/1994 33.0b 1.28b 0.18b 0.19b
ST 21/04/1995 34.0b 1.16c 0.27c 0.08c
C 13/03/1995 25.9a 1.63a 0a 1a
Standard error 0.8 0.01 0.01 0.02
a
Distance between the soil surface and the bottom of the ploughed layer.
b
C: Compacted plot, AT: Autumn-tilled plot; ST: Spring-tilled plot.
c
Column values within each soil type followed by the same letter are not significantly different using the
Tukey’s test ( P = 0.05).
 J.F. Sillon et al. / Geoderma 116 (2003) 29–46 35

Table 3
Surface properties of the loess and calcareous ploughed layers
Surface roughnessa Albedowetb Albedodryb
(cm)
Calcareous soil ATa 1.20 0.18 0.33
ST 1.23 0.18 0.26
C 0.30 0.22 0.33
Loess soil AT 1.28 0.17 0.31
ST 1.12 0.14 0.29
C 0.37 0.18 0.35
a
C: Compacted plot, AT: Autumn-tilled plot; ST: Spring-tilled plot.
b
Standard error was not defined for surface roughness and equal to 0.005 for albedo.

loess soil, the spring-tilled plot had a lower bulk density than the autumn-tilled plot
and most of the clods with a diameter >40 mm disappeared after tillage. Indeed, the
structure of the autumn-tilled plot (AT) of the loess soil was coarser than that of the
calcareous soil and the effect of spring tillage was more pronounced. Soil surface
roughness and albedo were quite similar in all the tilled plots (Table 3).
The increase in bulk density of the ploughed layer (between C and AT plots) by
compaction was 0.39 and 0.35 Mg m
3 in the calcareous and in the loess soils,
respectively (Table 2). The thickness of the compacted ploughed layer (C plot) was
10.6 and 7.1 cm less than that of the AT ploughed layer in the calcareous and in the
loess soil, respectively (Table 2). Compaction produced a massive structure in the
ploughed layer of the two soil types and no fine aggregates can be observed in the
compacted plots (Table 2). Compaction caused a strong decrease in soil surface
roughness in the two soil types which was associated with a slight increase in wet
and dry albedo (except for dry albedo of calcareous soil) (Table 3).

3.2. Soil drying

We have selected two drying periods (about 10 days, Table 4) which followed a
rainy period to study, in each experimental site, soil drying as a function of tillage and

Table 4
Soil evaporation and Penman reference evaporation during drying periods
Experiment Period Penman Soil evaporation (mm day
1)
evaporation
Ca AT ST
(mm day
1)
Calcareous soil 1996 22/3 ! 2/4 (11 days) 1.6 1.9 ab F 0.4c 0.8 b F 0.2 n.c.d
4/6 ! 13/6 (9 days) 5.1 2.9 a F 0.5 1.4 b F 0.1 1.3 b F 0.1
Loess soil 1995 4/4 ! 14/4 (10 days) 1.8 1.2 a F 0.2 1.8 a F 0.3 n.c.d
1/5 ! 9/5 (7 days) 4.2 1.2 a F 0.2 1.4 a F 0.2 1.7 a F 0.3
a
C: Compacted plot, AT: Autumn-tilled plot; ST: Spring-tilled plot.
b
Line values followed by the same letter are not significantly different using the Tukey’s test ( P = 0.05).
c
Standard errors (number of replications = 6).
d
Treatment was not created for this period.
36 J.F. Sillon et al. / Geoderma 116 (2003) 29–46

traffic. The first drying period (early spring) corresponded to a low Penmam reference
evaporation, i.e. 1– 2 mm per day. The second drying period (late spring) corresponded
to a high Penmam reference evaporation, i.e. 4 –5 mm per day. Spring tillage was only
studied during the second period.
Table 4 shows the effect of tillage and traffic on mean daily evaporation. In the
calcareous soil, evaporation from the compacted treatment was higher than that of the
autumn-tilled treatment during the two drying periods. The difference was more than 1
mm per day. Evaporation from the autumn-tilled treatment and from the spring-tilled
treatment was similar during the second drying period. In the loess soil, there was no
significant effect of tillage and traffic on evaporation, although evaporation was lowest

Fig. 1. Change though time of the volumetric water content profile in each treatment during the two drying
periods in the calcareous soil (a) and in the loess soil (b). Standard deviation on volumetric water content per layer
was in the range 0.007 – 0.0150 m3 m
3. Values in the legend are dates.
 J.F. Sillon et al. / Geoderma 116 (2003) 29–46 37

Fig. 1 (continued).

in the compacted plot. Evaporation was higher during the second drying period than
during the first one in the compacted and autumn-tilled plots of the calcareous soil. It
was similar during the two periods in the loess soil.
Fig. 1 indicates the change through time in volumetric water content profiles in the
three treatments of each soil type during the two drying periods. For each drying
period, the zero flux plane was below 30 cm after 2 days in the different plots,
indicating that evaporation dominated infiltration in the ploughed layer. In the
calcareous soil and for the first drying period, the water volume was quite
homogeneously distributed along the profile from 0 to 0.4 cm depth. During the
whole drying period studied, the water content remains constant at depths greater than
15 cm and decreases only in the top 15 cm: the top 15 cm was the only depth that
contributed to evaporation. During the second drying period, the whole profile, depths
as great as 40 cm lost water by evaporation, although the drying phenomenon was
38 J.F. Sillon et al. / Geoderma 116 (2003) 29–46

more considerable in the top 5 cm. The spring-tilled plot exhibited the same
behaviour as the autumn-tilled plot. In the compacted horizon, the initial water
content at the beginning of the first drying period was not homogeneously distributed
over the 40 cm, the bottom of the profile being wetter. During this drying period, the
water content decreased in the 0 –40 cm depth range, meaning that the whole profile
contributed to water loss by evaporation. The same phenomenon was observed during
the second drying period, although the initial water profile was more homogeneously
distributed. The contribution of the deep layers (>15 cm) to evaporation was less
pronounced in the autumn-tilled plot than in the compacted plot. The water volume
that was evaporated from the autumn-tilled plot mainly came from the superficial sub-
layers ( < 15 cm). In the loess soil, the contribution of each sub-layer within the
whole ploughed layer was quite similar between the three plots, with greater water
loss in the superficial layer < 5 cm.

Fig. 2. Comparison of the water ratio profile between the autumn-tilled plot and the compacted plot or spring-
tilled plot, at the beginning and at the end of each drying period in the calcareous and loess soils. Standard
deviation on water ratio measurements per depth was in the range 0.01 – 0.02 m3 m
3. C: Compacted plot, AT:
Autumn-tilled plot; ST: Spring-tilled plot. (i1) means the first day the first drying period, (f1) means the last day
the first drying period, (i2) means the first day the second drying period, (f2) means the last day the second drying
period.
 J.F. Sillon et al. / Geoderma 116 (2003) 29–46 39

The consequences of the change in evaporation behaviour on the water ratio profile
are indicated on Fig. 2. In the calcareous soil, the water ratio profile on the first day of
the first drying period was quite similar in the compacted and autumn-tilled plots
because this period began the day when the compacted plot was created. Water ratio
profiles in the two plots were very different 10 days after: the 0– 6 cm layer was still
wet in the compacted plot whereas this layer was dry in the autumn-tilled plot. In
contrast, the water ratio of the soil layer >7 cm depth was higher in the autumn-tilled
plot than in the compacted plot. At the beginning of the second period, the water ratio

Fig. 3. Water retention curve, for water potential >
80 kPa, obtained from pairs of field values (W, #) in the
three treatments of each soil type. Horizontal bars indicate standard errors. For each treatment, we calculated the
mean water ratio # for every increment of 5 kPa of water potential. C: Compacted plot, AT: Autumn-tilled plot;
ST: Spring-tilled plot. n C, D AT, ST.
40 J.F. Sillon et al. / Geoderma 116 (2003) 29–46

profile was similar in the autumn-tilled and spring-tilled plots, but the water ratio
was lower in the compacted plot. Ten days afterwards, there was no difference
between the two tilled plots. Only the water ratio of the first centimetre of the
compacted plot was higher than that of the autumn-tilled plot. The water ratio for
layers >1 cm was lower in the compacted plot than in the autumn-tilled plot and the
difference in water ratio between these two plots remained constant.
In the loess soil, the water ratio on the first day of the first drying period was
already lower in the compacted plot and in the autumn-tilled plot. The first drying
period that we investigated could not immediately follow the creation of the compacted
treatment because of bad weather conditions, in contrast with what we were able to do
in the calcareous soil. After 10 days of drying, the shape of the water ratio profile was
very similar between the two plots, with a superficial layer (0– 10 cm) drier than the
deep layers (>10 cm). At the beginning of the second drying period, the water ratio of
the compacted plot was still lower than in the autumn-tilled plot. The spring-tilled plot
had the lowest water ratio near the soil surface. After 10 days of drying, the shape of
the water ratio profile was very similar between the three plots, with a superficial layer
(0– 10 cm) drier than the deep layers (>10 cm).

3.3. Hydraulic properties

Regarding the water retention curves, field pairs of values (#, W) are presented in Fig. 3
for water potential >
80 kPa and Van Genuchten’s parameters are given in Table 5. In the
calcareous soil, the compacted plot retained less water than the tilled plots for water
potential >
10 kPa. No difference of water retention was observed between the three
plots for water potential <
10 kPa. In the loess soil, the compacted plot retained less
water than the autumn– spring plot for water potential >
30 kPa. For water potential
<
30 kPa, the compacted plot retained more water than the autumn-tilled plot. The
spring-tilled plot retained less water than the autumn-tilled plot for water potential <
10
kPa.

Table 5
Estimated parameters of the water potential – water ratio relationship using the Van Genuchten model (1980),
(Eq. (2))
]
l,
80] kPa ]
80,
5] kPa
2
#s #r a n r #s #r a n r2
3
3
1 3
3
1
m m m m m m
a
Calcareous soil C 1.87 0.02 3.59 1.40 0.99 0.60 0 0.48 1.13 0.51
AT 1.73 0.01 6.67 1.34 0.99 0.67 0 4.17 1.08 0.83
ST 1.97 0.01 6.97 1.36 0.99 0.69 0 3.65 1.11 0.84
Loess soil C 0.53 0 0.03 1.65 0.99 0.59 0 0.26 1.15 0.63
AT 0.48 0 0.02 1.72 0.99 0.66 0 0.21 1.50 0.88
ST 0.50 0 0.02 1.69 0.99 0.68 0 0.85 1.20 0.93
a
C: Compacted plot, AT: Autumn-tilled plot; ST: Spring-tilled plot.
 J.F. Sillon et al. / Geoderma 116 (2003) 29–46 41

Figs. 4 and 5 indicate the change in hydraulic conductivity as a function of water

ratio obtained with the WIND method (Fig. 4) or the inverse method (Fig. 5). Table 6
gives the four parameters of the log – polynonimal relationships K[#]. The two
methods, based on totally independent data, gave similar curves, in shape and
classification, in the tensiometric range, i.e. water ratio >0.5 m3 m
3. In the
calcareous soil, the compacted plot had a higher hydraulic conductivity in the whole
range of water ratio. The two tilled plots had similar hydraulic conductivity. The ratio
of the hydraulic conductivity of the compacted plot to that of the tilled plot was about
3 for a water ratio of 0.55 m3 m
3. In the loess soil, hydraulic conductivity was

Fig. 4. The logarithm of hydraulic conductivity, measured in m s
1, against the water ratio obtained from pairs of
values (#, K) from the Wind method in the three treatments of each soil type. C: Compacted plot, AT: Autumn-
tilled plot; ST: Spring-tilled plot. n C, D AT, ST. Vertical bars indicate F standard deviation.
42 J.F. Sillon et al. / Geoderma 116 (2003) 29–46

Fig. 5. The logarithm of hydraulic conductivity, measured in m s
1, against the water ratio expressed by log-
polynomial relationships obtained with an inverse method (TEC model). C: Compacted plot, AT: Autumn-tilled
plot; ST: Spring-tilled plot. — C, – – – - AT, – – – ST.

similar in the three plots for a water ratio >0.3. For dry conditions, i.e. water ratio
< 0.3, hydraulic conductivity was lowest in the spring-tilled plot and highest in the
compacted plot.

Table 6
Estimated parameters of the hydraulic conductivity – water ratio relationship using a log – polynomial relationship
(third degree, Eq. (4))
a0 a1 a2 a3 r2

1
log10 [m s ]
a
Calcareous soil C
17.39 19.30
8.23 1.83 0.94
AT
19.38 22.72
13.71 7.36 0.95
ST
20.09 23.26
8.02 0.69 0.93
Loess soil C
15.70 17.25
25.53 26.42 0.96
AT
16.64 21.67
26.09 19.43 0.96
ST
17.58 20.37
14.31 7.72 0.96
a
C: Compacted plot, AT: Autumn-tilled plot; ST: Spring-tilled plot.
 J.F. Sillon et al. / Geoderma 116 (2003) 29–46 43

4. Discussion

In the calcareous soil, the change in bulk density led to a considerable change in soil
drying, i.e. evaporation and soil water content profile: evaporation was 1 mm per day
greater in the compacted and dense plot than in the tilled and porous plot. Moreover, the
soil surface remained wet during a longer period in the dense plot. This result can be
linked to the higher hydraulic unsaturated conductivity in the dense plot: a high hydraulic
conductivity transfers water more efficiently from deep layers to the soil surface which can
remain wet. Therefore the first stage of evaporation, when evaporation remains high, is
longer. An increase of unsaturated hydraulic conductivity due to an increase of bulk
density was already observed by Tamari (1994) on artificial soil cores. It can be explained
by a greater contact surface between soil aggregates in a dense soil than in a porous soil
(Gupta et al., 1989), if the structure of the aggregates does not change with tillage and
traffic. The similarity between the retention curves of the dense and porous plots, in a
range of water potential where water is mainly located within the soil aggregates (water
potential <
20 kPa), suggests that the aggregate structure was not affected by tillage and
traffic in the calcareous soil. The spring-tilled and autumn-tilled plots had similar
structures, and consequently similar hydraulic properties and drying behaviours. This is
probably due to the fact that calcareous soil is a stable soil (Richard et al., 1995),
insensitive to soil structure degradation by rainfall, and thus to tillage timing.
Our results in the loess soil are very different than those in the calcareous soil. In spite
of a wide range of bulk density (from 1.16 to 1.63 Mg m
3), drying curves during the
spring season were very similar for the three plots. However, soil water content at the
beginning of the two drying periods of our experiment differed between the compacted
and autumn-tilled plots: the compacted plot was dryer than the autumn-tilled plot. The
unsaturated hydraulic conductivity curve was similar in the compacted and autumn-tilled
plots. It seemed reasonable to expect that the difference in initial water content has
influenced the soil drying curves. We analysed the effect of hydraulic properties from the
compacted plot and autumn-tilled plot with the same initial conditions on soil drying, by
using the mechanistic TEC model. Fig. 6 shows the change in water ratio profile in the
compacted and autumn-tilled plots when the same water ratio profile was considered at the
beginning of the first drying period. Mean day evaporation remained similar in the two
plots (1.3 and 1.4 mm in the compacted and autumn-tilled plots, respectively). The water
ratio profile had the same shape in the two plots but with a higher water ratio in the
compacted plot in the first 15 cm. We also checked that the difference in albedo and
roughness between the compacted plot and the autumn-tilled plot had little effect on soil
drying. Two main points can be mentioned:

First, the experimental and numerical results that we have obtained in the calcareous
soil and in the loess soil shows that soil drying is mainly determined by hydraulic
conductivity. This confirms the theoretical work of Linden (1982) who compared the
effect of albedo, roughness and hydraulic properties on soil drying using a numerical
model of soil water transfer and evaporation.
Second, the water ratio in the compacted plot was lower than in the autumn-tilled plot at
the beginning of the first drying period. The evaporation rate between the two plots was
44 J.F. Sillon et al. / Geoderma 116 (2003) 29–46

Fig. 6. Water ratio profile in the compacted plot and autumn-tilled plot calculated during the first drying period in
the loess soil using the water transfer model (TEC model). C: Compacted plot, AT: Autumn-tilled plot.

similar during the two drying periods. Therefore, it appears that real water infiltration
between the day of the creation of the compacted plot and the first day of the first
drying period was probably different in the compacted and the autumn-tilled plots. This
could be attributed to a difference in soil surface characteristics: slaking at the soil
surface was observed and runoff probably occurred in the compacted plot in spite of a
very low slope. We did not investigate water infiltration, but these results show that the
analysis of soil drying as a function of soil structure must take into account soil surface
characteristics.

The effect of bulk density on the hydraulic properties of the loess soil was analysed by
Richard et al. (2001a) using mercury porosimetry and scanning electron microscopy. They
showed that the change in the soil hydraulic properties could be related to the formation of
relict structural pores (Bruand and Cousin, 1994) by compaction. Relict structural pores
are structural pores which have been distorted by compaction in the field during traffic and
which are accessible only through the necks of the lacunar pores. On the one hand, water
retained by soil would increase in a compacted soil because of the contribution of the
volume of relict structural pores to water retention (in a range of water potential where
classical structural pores do not contribute to water retention). This would explain the
change in the retention curve between the compacted plot, the autumn-tilled plot and the
spring-tilled plot. On the other hand, the water retained in the relict structural pores may
not contribute to the water movement because these pores are accessible only through the
necks of lacunar pores and behave as reservoirs. As a consequence, a smaller proportion of
the water contributes to the water movement in the compacted plot than in the tilled plots.
The similarity of the relations between hydraulic conductivity and the water ratio whatever
the soil bulk density may be due to the two opposing influences of compaction on
hydraulic conductivity, increase the contact surface between the aggregates on the one
hand; decrease the proportion of the water that contributes to water transfer on the other
hand.
 J.F. Sillon et al. / Geoderma 116 (2003) 29–46 45

5. Conclusion

The effect of soil structure, caused by tillage and traffic, on soil drying depends on
hydraulic conductivity: an increase in unsaturated hydraulic conductivity by an increase in
bulk density leads to an increase in evaporation, in the water contribution of deep soil
layers and in the period of time when soil surface remains wet. The albedo and surface
roughness have minor effects if any. Soil drying also depends on water infiltration which
determines the soil water content profile at the beginning of each drying period. The
increase in unsaturated hydraulic conductivity by an increase in bulk density depends on
soil type: it was observed in the calcareous soil, but not in the loess soil. This may be due
to the formation of relict structural pores during compaction in the loess soil. We must
therefore also investigate the possible relict structural pores in various soils, as a function
of soil sensitivity to compaction and traffic conditions.

Acknowledgements

The work was done as part of the requirements for the PhD by J.F. Sillon. It was
partially supported by the ‘‘Chambre d’Agriculture de la Marne’’ and the ‘‘Conseil
Régional de Picardie’’. The authors thank D. Boitez, C. Dominiarczyk and F. Mahu for
their helpful technical assistance and K. Hodson O. Parkes for checking the English text.

References

Agrawal, R.P., Phogat, V.K., 1998. Soil management for improving the quality of physically constraint soils. 16th
World Congress of Soil Science, vol. 1. Proc., 20 – 26 August 1998, Montpellier (France).
Allmaras, R.R., Hallauer, E.A., Nelson, W.W., Evans, S.D., 1977. Surface energy balance and soil thermal
property modifications by tillage-induced soil structure. Tech. Bull., 306.
Benjamin, J.B., 1993. Tillage effects on near-surface soil hydraulic properties. Soil Tillage Res. 26, 277 – 288.
Benjamin, J.G., Blaybock, A.D., Brown, H.J., Cruse, R.M., 1990. Ridge tillage effects on simulated water and
heat transport. Soil Tillage Res. 18, 167 – 180.
Bruand, A., Cousin, I., 1994. Variation of textural porosity of a loamy – clay soil during compaction. Eur. J. Soil
Sci. 46, 377 – 385.
Bruckler, L., Ball, B.C., Renault, P., 1989. Laboratory estimation of gas diffusion coefficient and effective
porosity in soils. Soil Sci. 147 (1), 1 – 10.
Brutsaert, W., 1982. Evaporation Into the Atmosphere. Reidel Ed. 299 pp.
Chanzy, A., Bruckler, L., 1993. Significance of soil moisture with respect to daily soil evaporation. Water Resour.
Res. 29, 1113 – 1125.
Chung, S.O., Horton, R., 1987. Soil heat and water flow with a partial surface mulch. Water Resour. Res. 23,
2175 – 2186.
Culley, J.L.B., Larson, W.E., Allmaras, R.R., Schaffer, M.J., 1987. Soil – water regimes of a Typic Haplaquoll
under conventional and no-tillage. Soil Sci. Soc. Am. J. 51, 1604 – 1610.
Currence, H.D., Lovely, W.G., 1970. The analysis of soil surface roughness. Trans. ASAE 13, 710 – 714.
De Vries, D.A., 1963. Thermal properties of soils. In: Van Wijk, W.R. (Ed.), Physics of Plant Environment,
vol. 1. Academic Press, London, pp. 57 – 109.
Gauer, E., Shaykewich, C.F., Stobbe, E.H., 1982. Soil temperature and soil water under zero tillage in Manitoba.
Can. J. Soil Sci. 62, 311 – 325.
46 J.F. Sillon et al. / Geoderma 116 (2003) 29–46

Gill, K.S., Jalota, S.K., Prihar, S.S., Chauphary, T.N., 1977. Water conservation by soil mulch in relation to soil
type, time of tillage, tilth and evaporativity. J. Indian Soc. Soil Sci. 25 (4), 360 – 366.
Gupta, S., Sharma, P.P., DeFranchi, S.A., 1989. Compaction effects on soil structure. Adv. Agron. 42, 311 – 338.
Hammel, J.E., Papendick, R.I., Campbell, G.S., 1981. Fallow tillage effects on evaporation and seedzone water
content in a dry summer climate. Soil Sci. Soc. Am. J. 45, 1016 – 1022.
Holmes, J.W., Greacen, E.L., Gurr, C.G., 1960. The evaporation of water from bare soils with different tilths. 7th
International Congress of Soil Science, vol. 1 (14). Madison, WI, USA.
Idso, S.B., Jackson, R.D., Reginato, R.J., Kimball, B.A., Nakayama, F.S., 1975. The dependence of bare soil
albedo on soil water content. J. Appl. Meteorol., 109 – 113.
Johnson, M.D., Lowery, B., Daniel, T.C., 1984. Soil moisture regimes of three conservation tillage systems.
Trans. ASAE, 1385 – 1395.
Klute, A., 1986. Water retention: laboratory methods. In: Klute, A. (Ed.), Methods of Soil Analysis: Part 1, 2nd
ed. Agronomy, vol. 9, pp. 635 – 662.
Linden, D.R., 1982. Predicting tillage effects on evaporation from the soil. Predicting Tillage Effects on Soil
Physical Properties and Processes. Vol. ASA Special Publication Number 44, 117 – 132.
Marquart, D.W., 1963. An algorithm for least-squares estimation of non-linear parameters. SIAM J. Appl. Math.
11, 431 – 441.
Mwendera, E.J., Feyen, J., 1994. Effects of tillage and evaporative demand on the drying characteristics of a silt
loam: an experimental study. Soil Tillage Res. 32, 61 – 69.
Negi, S.C., Raghavan, G.S.V., Taylor, F., 1981. Hydraulic characteristics of conventionally and zero-tilled field
plots. Soil Tillage Res. 2, 281 – 292.
O’Brien, F.E.M., 1948. The control of humidity by saturated salt solutions. J. Sci. Instrum. Phys. Ind. 25,
73 – 76.
Philip, J.R., De Vries, D.A., 1957. Moisture movements in porous materials under temperature gradients. Trans.
Ann. Geophys. Union 38, 222 – 232.
Radke, J.K., Dexter, A.R., Devine, O.J., 1985. Tillage effects on soil temperature, soil water, and wheat growth in
South Australia. Soil Sci. Soc. Am. J. 49, 1542 – 1547.
Richard, G., Cellier, P., 1998. Effect of tillage on bare soil energy balance and thermal regime: an experimental
study. Agronomie, 161 – 178.
Richard, G., Boiffin, J., Duval, Y., 1995. Direct drilling of sugar beet into a catch crop: effects on soil physical
conditions and crop establishment. Soil Tillage Res. 34, 169 – 185.
Richard, G., Cousin, I., Sillon, J.F., Bruand, A., Guérif, J., 2001a. Effect of compaction on soil porosity:
consequences on hydraulic properties. Eur. J. Soil Sci. 52, 49 – 58.
Richard, G., Sillon, J.F., Marloie, O., 2001b. Comparison of inverse and direct evaporation methods for estimat-
ing soil hydraulic properties under different tillage practices. Soil Sci. Soc. Am. J. 65, 215 – 224.
Tamari, S., 1994. Relations between pore-space and hydraulic properties in compacted beds of silty-loam
aggregates. Soil Technology 7, 57 – 73.
Vachaud, G., Dancette, C., Sonko, S., Thony, J.L., 1978. Méthode de caractérisation hydrodynamique in situ d’un
sol non saturé. Ann. Agronom. 29, 1 – 36.
Van Genuchten, M.T., 1980. A closed form equation for predicting the hydraulic conductivity of unsaturated
soils. Soil Sci. Soc. Am. J. 44, 892 – 898.
Warkentin, B.P., 1971. Effects of compaction on content and transmission of water in soils. Compaction of
Agricultural Soils, ASAE, 126 – 153.
Wind, G.P., 1968. Capillary conductivity data estimated by a simple method. In: Rijtema, P.E., Wassink, H.
(Eds.), Water in the Unsaturated Zone, vol. 1. Proc. Wageningen Symp. June 1966. Int. Assoc. Scientific
Hydrol., Gentbrugge, Belgium, pp. 181 – 191.
Witono, H., Bruckler, L., 1989. Use of remotely sensed soil moisture content as boundary conditions in soil-
atmosphere water transport modelling. Water Resour. Res. 25, 2423 – 2435.
Wu, L., Swan, J.B., Paulson, W.H., Randall, G.W., 1992. Tillage effects on measured soil hydraulic properties.
Soil Tillage Res. 25, 17 – 33.