Content uploaded by Maritxu Guiresse
Author content
All content in this area was uploaded by Maritxu Guiresse on Sep 19, 2017
Content may be subject to copyright.
Soil & Tillage Research 35 (1995) 157-166
Erosion due to cultivation of calcareous clay soils on
hillsides in south-west France. II. Effect of
ploughing down the steepest slope
M. Guiresse”, J.C. Revel
Laboraioire Inghierie Agronornique, ENSAT, 145 Av. de Muret 31076 Toulouse, France
Accepted 13 April 1995
Abstract
The erosion of calcareous clay soils from Terrefort in south west France is not exclusively due to
run-off water. An experiment was carried out in situ to quantify the effect of ploughing on the
displacement of soil down the slope. In s field experiment, a trench 0.4 m x 0.4 m x 20 m was dug
parallel to the contours and filled with crushed gravel. Gravel distribution on the surface and within
the soil mass was examined after ploughing downhill and then uphill. Mathematical analysis of the
data showed that a much greater total soil mass was displaced downwards by ploughing downhill
than by ploughing uphill. This resulted in soil removal from the top of the plot (0.32-1.62 kg m-*
for a single ploughing) and much greater accumulation at the base of the slope than estimated for
surface water erosion (i.e. 0.04-0.25 kg m-* year-‘), according to measurements made by other
authors in the same region.
Keywords: Erosion; Soil on slope; Tillage; Field experiment; Ploughing
1. Introduction
The soils on the hillsides of south west France have developed on impermeable Tertiary
sandstone. Downcutting by rivers since the end of the Tertiary has produced a hilly relief
(Revel, 1982) in which the soil’s spatial organisation with regard to chronosequence and
toposequence is highly complex. Revel and Rouaud ( 1985), Rouaud ( 1987) and Revel et
al. ( 1990) showed that run-off water was not the only erosive agent and that cultivation
techniques have contributed strongly to downward soil displacement. The present mesorelief
and eroded areas may be explained from a model based on the ancient contour ploughing
* Corresponding author
0167.1987/95/$09.50 0 1995 Elsevier Science B.V. All rights reserved
SSDIO167-1987(95)000483-l
158 M. Guiresse. J.C. Revel/Soil & Tillage Research 35 (1995) 157-166
method, in which the soil was tipped over and downwards (Revel and Guiresse, 1995).
After mechanisation, ploughing had to be carried out in the direction of the slope. The
low power of the first tractors only permitted downhill ploughing, which resulted in soil
accumulation at the base of the plot. When ploughing is performed under idea1 moisture
conditions, the soil’s structural stability should, in theory, ensure sufficient cohesion of the
aggregates to allow the ploughed portion to be overturned on the spot without any isolated
particles or small aggregates being displaced down or up the slope according to the direction
of ploughing. Because of their velocity, the implements used during ploughing project some
of the aggregates in the direction of movement. When the tractor goes downhill, a weight
effect adds to this phenomenon, whereas the effect is curtailed in the uphill direction.
Ploughing such soils leads in the long term to downward displacement of soil towards the
base of the slope. It is this phenomenon that we intend to quantify here.
2. Material and methods
2. I. Choice of tracer
To examine the effect of ploughing on soil displacement down the slope, a tracer has to
be introduced homogeneously into the soil above a reference level, at a known concentration,
and to a depth at least equal to that of ploughing. Choice of the tracer presents several
difficulties. Whether the tracer is a particulate material or chemical compound, radioactive
or not, it must not be transported in suspension or in solution by rainwater. Subsequent
extraction of the tracer from the soil and its determination should also be easy. Use of a
solid tracer poses the problem of homogenisation. Numerous large aggregates exist in clay
soils and, if these are crushed to ensure good tracer distribution, the structure and mechanical
properties of the final material no longer have anything in common with those of the soil in
situ. The same problem of homogenisation is encountered if the tracer is dissolved in a
stable organic solvent, insoluble in water. The organic solvent may also considerably modify
soil structure (Henin, 1960).
Despite numerous trials, an ideal tracer that fits all of these criteria has not been found.
We finally selected crushed siliceous grave1 4-6 mm in diameter, that we put in a trench.
The soils under test did not contain any such material.
2.2. Experimental conditions
The experimental plot was situated on a hillside with a uniform slope of 11”. The soil
was a calcic cambicsol (according to FAO taxonomy) with a sandy clay loam texture, high
organic matter content for the region ( 17 g kg-‘) and high structural stability (Zs=O.5;
Henin, 1960). A trench was dug parallel to the contour, the dimensions (0.4 m X 0.4 m X 20
m) dictated by the experimental conditions: the length of 20 m was a multiple of the machine
width, the depth of 0.4 m was slightly greater than the ploughing depth, and the width of
0.4 m corresponded to the dimension of the mechanical digger scoop used to prepare the
trench.
M. Guiresse, J.C. Revel/Soil & Tillage Research 35 (1995) 157-166 159
First the soil was taken out of the trench and removed from the plot. The trench was then
filled with gravel after protecting the surrounding soil with tarpaulin to avoid contamination.
Traditionally, the soil is ploughed both uphill and downhill along the steepest slope. Thus
the estimated amount of soil displaced downwards is the difference between the volume of
soil displaced downhill and the volume of soil displaced uphill. We therefore carried out
two treatments on 12 m widths. The first treatment was a single downhill ploughing which
was carried out with a Huard plough with three frames and five shares 0.4 m apart, drawn
by a Fiat tractor 980 DT 100 hp, at a speed of 7 km hh’. The estimated ploughing depth
was 0.3 m. The second treatment, a single uphill ploughing with speed reduced to 6.2 km
h - ‘, was carried out nearby. After ploughing the soil was rolled and levelled with the tractor
to facilitate sampling. Then, we sampled the soil surrounding the trench to quantify the
gravel displacement due to both the treatments, uphill and downhill ploughing.
2.3. Mode of sampling
First, gravel displacement in the trench due to ploughing was initially observed on the
soil surface. A visual, systematic count was made on the soil surface, using 5 cm square
quadrats. Complete analysis requires a consideration of the mass of soil. For this, depth
samples were taken from positions where the surface distribution was about average. To
avoid the effect of position in relation to passage of the tractor, rectangular parallelepipeds
with lengths equal to the distance between two ploughshares (0.4 m) were sampled. The
samples were 0.1 m in height and 0.1 m in width so as to provide samples of reasonable
weight for data analysis. Each sample was weighed then sieved to determine the weight of
the constituent gravel (I’,). The nature of these operations precluded replication in the
initial phase. The total bulk densities of the gravel (D, = 1.33 Mg m- 3, and soil (D, = 1.63
Mg m-“) were measured to calculate the soil weight equivalent (P) of the gravel. The
need for corrections for water content was circumvented by measuring densities D, in the
field moist state. To convert our results to a dry soil, D, was measured in the dry state.
P=P,(D,/D,) in kg (S,,)
(1)
The sum of all the P values gave the weight of the gravel carried out of the trench,
expressed in soil weight equivalent (S,,,,) .
2.4. Limits of the experimental protocol
First, to determine the total amount of soil displaced across a reference level, the tracer
should have been introduced into a soil strip that was at least as wide as the maximum
displacement of the particles. To correct for this, a mathematical model was used to extrap-
olate the results obtained for a width of 0.4 m to a strip in which the width was equal to the
greatest distance covered by the gravel. Second, the texture, structure and mechanical
properties of the gravel differed considerably from those of the soil. The particulate nature
of the gravel certainly amplified the effect of implements on particle migration. This exper-
iment nevertheless provides an upper estimate of the degree of downward soil displacement
due to ploughing. Finally, the trench produced a discontinuity in the mechanical resistance
160 M. Guiresse, J.C. Revel /Soil & Tillage Research 35 (1995) 157-166
to passage of the implements. Because of the difference between the mechanical behaviour
of soil and gravel we reduced the trench width as much as possible.
3. Results
3.1. &&ace gravel distribution
The observations of gravel displacement on the soil surface showed that the upward and
downward displacements of gravel were clearly defined and could attain considerable
distances from the sides of the trench (2.2 m with downhill ploughing and 1.4 m with uphill
ploughing), A greater quantity of gravel was displaced downward than upwards. Consid-
erable erosion would thus occur if ploughing was carried out solely down the slope and this
would not be compensated by uphill passage of the same implement.
3.2. Gravel distribution within the soil muss
3.2.1. Sampling results
First, we examined the results in each horizontal soil layer at O-O. 1,O. l-O.2 and 0.2-0.3
m depths. This gave a random distribution from the trench (Revel et al., 1990,1993). Such
considerable dispersal of the gravel could be explained by the displacement of particles in
all three spatial directions. To suppress vertical variability, the sum of the amounts of gravel
found at three different depths (O-O. 1,O. l-0.2, and 0.2-0.3 m) was considered as a function
of distance from the trench. The results represented in Fig. 1 exhibit a modal distribution
which is a function of distance from the trench. Before continuing the data analysis, the
reliability of the measurements was tested by verifying that the total volume of gravel
sampled after one uphill ploughing ( T,,,) and one downhill ploughing (r,) corresponded
to the volume of gravel initially placed in the trench (7’). This gave
T,,, = 0.045
m3
and
Td = 0.054
m3
while
T= (0.4) (0.4) (0.3) = 0.048
m3
The difference between
T,,,
and
T,
and between
Td
and
T
reflected the degree of accuracy
of the measurements (6% and 12% respectively). This result was considered satisfactory
given the experimental conditions. Fig. 1 shows that the gravel migrated further within the
soil mass after downhill ploughing than after uphill ploughing. Using Eq. ( 1 ), we calculated
the sum of all
P
values, and so we found the mass of soil displaced without any extrapolation.
The results are given in Table 1.
M. Guiresse, J.C. Revel /Soil & Tillage Research 35 (1995) 157-166 161
A : Downhill ploughing
1 looo Weight of gravel expressed in soil weight equivalent in grammes
T
n
UPHILL
10000 --
9000 --
6000 --
5000 --
4000 --
3000 --
2000 --
1000 --
DOWNHILL
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Distance to LR line in decimeter
B : Uphill ploughing 11000 T Weight of gravel expressed in soil weight equivalent in grammes
-
10000
9000
DOWNHILL
;$I
8000
7 0
6000 --
5000 --
4000 --
3000 --
2000 --
1000 --
UPHILL
L
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Distance to LR line in decimeter
Fig. 1. Gravel distribution in the mass of soil either side of the trench after downhill ploughing (A) and uphill
ploughing (B) as a function of distance from the reference line LR, which was the lower or upper limit of the
trench. respectively.
3.2.2. Extrapolation
of
the results
The objective of this work was to determine the total mass of soil (M) displaced under
the effect of implements across a linear metre in relation to a reference line (LR) . In this
experiment, the reference line (LR), in the case of downhill ploughing, was the lower limit
162 M. Guiresse. J.C. Revel/Soil & Tillage Research 35 (199.5) 157-166
Table I
Mass of soil (kg) displaced along the slope by one downhill ploughing and one uphill ploughing with and without
extrapolation
Without extrapolation With extrapolation
Downhill ploughing 181.5 306.8
Uphill ploughing 110.3 144.6
Difference 71.2 162.2
of the trench, and the upper limit in the case of uphill ploughing. As the samples had been
taken over a width (L) of 0.4 m, the factor 1
lL
was applied to convert the results to a strip
1 m in width for later use in comparisons with other implements.
It is apparent from Fig. 1 that the particles exhibited variable displacement. Some moved
a long distance, others remained nearer the trench. Thus, at a given point, the sampled soil
contained undisplaced soil, displaced gravel, and also soil from above the point under
consideration. Finally, all particles which migrated at most 0.4 m were counted in this
experiment as the soil was marked over a distance of 0.4 m beyond the line (LR). In
contrast, not all the material that moved a greater distance was taken into account. The
results need to be extrapolated for complete analysis which requires use of a model. The
functional principle of this will now be described.
It is presumed that the particles (M) of soil displaced across LR under the effect of
implements, can be divided into different classes (Mi) depending on the distance travelled
(i) . To simplify the expression, i is given in decimeters so that the distances are expressed
in whole numbers, this being a prerequisite of the indexed scoring system used in the model.
This gives
where Mj is the total mass of particles carried i dm across LR and
D
is the maximum distance
in decimetres travelled by the gravel.
To calculate M, all values of Mi have to be determined between i= 1 and
D.
For each
value of Mi, the displaced particles are characterised by j, their position in relation to the
line LR prior to displacement. This gives
M, = Cmi,j (3)
where
mjj
is the mass of soil carried over a distance
i
and coming fromj dm above the line
LR (Fig. 2))
i
and j are expressed in decimeters. For one value of
i,
j cannot be greater than
i.
So we have
j-i
M,=m;,l +mi.z+mi.3+ . . . +mi,;= Cm;,- (4)
j=l
It is assumed that soil displacement down a constant slope obeys a steady state (i.e. the
mass of soil displaced over distance i is identical at any point on the slope irrespective of
its position). This means that whatever the value ofj, and for a fixed value of
i, m,,j
is a
M. Guiresse. J.C. Revel/Soil & Tillage Research 35 (1995) 157-166 163
= m 20.1
P
19
= m
19,1+
m 20.2
+ m
20.4
0 :I
Pi
:2
mIJ :3
Fig. 2. Diagrammatic representation of use of the model in downhill ploughing: 1, trench; 2, soil weight equivalent
taken at i dm from LR; 3, soil weight equivalent carried over distance
i
and coming from j dm above the line LR.
constant. All the terms in the summation in Eq. (4) are therefore equal and this gives
Mj=imj
(5)
Use of this model in our data analysis is illustrated by Fig. 2, taking the case of downhill
ploughing in which D = 20 dm. In this case at a position 20 dm downhill from the trench it
is presumed that all gravel taken from this site (Pzo) comes from the first decimeter inside
the trench, represented by Pzo = mzo., . Sample
PI9
is taken 1 dm nearer to the trench and
contains gravel from the first decimeter within the trench which has been displaced over 19
dm
(m,9.,)
plus gravel from the second decimeter displaced over 20 dm (m&, (i.e.
Ply = mly.l + mzoTz). If we have a steady state situation this implies that m20,1 = mzoe2. Finally,
P,,
-
P,,
gives the value m,,,,.
Thus, if one stands at a distance i from the trench, the gravel taken at this position (Pi)
consists of gravel from the first decimeter within the trench which has been displaced over
i dm ( rn;. , ) plus gravel from the second decimeter within the trench which has been displaced
over i + 1 dm ( mj+ ,,*), etc., giving
P, = mi., + (mi+ 1.2) + (mi+2,3) + (++3,4)
164 M. Guiresse, J.C. Revel/Soil & Tillage Research 35 (1995) 157-166
Table 2
Estimation of downward soil displacement due to ploughing and comparison with rain-induced soil erosion
Slope length
(ml Soil
displaced
(kg mm’)
After pluughing
100
250
500
1.62
0.65
0.32
Esrimafes
of
soil erosion by rainfall
Roose and Cavalie (1988) 0.01425
Rouaud (1987) 0.04-0.17
Etchanchu ( 1988)
0.02
“Total amount of soil displaced after one downhill ploughing and one uphill ploughing on a slope of 18% in a I-
m-wide plot.
Calculation of all these equations from
i
= 1 to
i= D-
3 and assuming that
nkl =
nh.2
= mi,3 = w4, produces a system of
D
equations with
D
unknowns. By resolving
this system, all the
mi
terms are determined and thence all the il4i classes and finally the total
M. Resolution of the system is without difficulty in the case of downhill ploughing and gives
the sum Md as 306.8 kg (S,,,,) . In the case of uphill ploughing, in contrast, the curve shown
in Fig. 1 (b) had to be slightly smoothed to solve the equations and determine the sum
which was 144.6 kg (S,) . Finally, the effect of one single uphill ploughing and one single
downhill ploughing is the difference Md - M,,, that is 162.2 kg (S,,,,).
The results in Table 1 show that the downward displacement of soil towards the base of
the slope as a result of downhill ploughing is not compensated for by uphill ploughing. The
mathematical model produces a result that is more than twice that obtained without extrap-
olation.
4. Discussion
We compared our results with measurements of rain-induced soil erosion. For this our
results had to be expressed in the same way as those of other authors (i.e. in tonnes of soil
displaced per hectare, or in kilograms per square metre). The plot that we studied was
situated along the greatest slope, as is most frequently encountered in the region. Under
such conditions, the amount of soil displaced down a hillside, parallel to the line of slope,
with constant slope, is independent of the length of the hillside and only dependent on plot
width. However, the mass of soil displaced by ploughing, related to the surface unit
ploughed, is inversely proportional to plot length. Some examples of calculations showing
values of 0.32-l 62 kg me2 are given in Table 2. These results are compared with different
measurements of water-induced erosion (Table 2) which were obtained as follows.
First, on the same plot as our trial, Roose and Cavalie (1988) measured sheet erosion,
brought about by a sheet or film of water produced with a rain simulator. They found that
M. Guiresse, J.C. Revel/Soil & Tillage Research 35 (1995) 157-166 165
the rainfall which has only occurred once in the last 10 years has an intensity of 40 mm h- ’
and produced scarcely visible effects on the landscape (0.01-0.25 kg m-*) compared with
our results.
Second, gully erosion in plots in which the slope varied from 3 to 25% was studied by
Rouaud ( 1987). Rills had been produced during spring storms by rainfall intensities of 4-
10 mm h- ’ which were average for the region. Their erosive effect was quantified by careful
measurement of the volumes hollowed out down the slopes and of the materials accumulated
at the bottom of each channel. On the steep slopes, the rills were anastomosed and had
excavated the soil to depths of 1.4-5.2 cm. On the gentler slopes, the furrows were rectilinear,
covering a smaller surface, but cutting deeper (i.e. 10-14 cm). It should also be noted that
in this experiment only a very small amount of the eroded material (i.e. between 11 and
29%) was redeposited at the base of the plot, the remainder being removed from the
watershed by surface water. Such run-off produced less total damage (0.04-O. 17 kg m-*)
than ploughing.
Finally, Etchanchu ( 1988) calculated the overall erosion produced by river action on a
watershed. Even when the relationship established by Robinson (1977) was taken into
account the calculated erosion (0.02 kg me2) was very slight.
5. Conclusions
In this experiment ploughing up and down a slope had a very marked erosive effect on a
gravel-filled trench. Extrapolation of the results to soil only provided an approximate
estimate of the effect of ploughing on downward soil displacement. In addition, certain
farmers only plough downhill and not always under good soil moisture conditions. This
results in considerable soil displacement towards the bottom of the hillside.
In the long term, mechanical erosion produced by implements has a preponderant impact
on surface relief in this landscape, especially on levelling and soil distribution. The displaced
materials remain on the slope but extensive surface areas may be deeply scoured. Revel and
Rouaud ( 1985) showed that in the Verrneil watershed (692 ha), 58% of the soil surface
was removed to a depth of 1.08 m and that a thickness of 1.47 m of soil was accumulated
over only 37% of the surface, This indicates an intense modification of the watershed.
In the region under study, other authors measured the impact of water erosion. The results
are variable but the greatest value (0.25 kg m-‘) was found after a rainfall event 10 years
ago which was still less than the amount of soil eroded by ploughing, even in the most
favourable case of a very long plot (0.32 kg mP2). In addition, this latter result was an
annual one and did not take other cultivation practices (disc ploughing, harrowing, etc.)
into account.
Thus, in this area, ploughing has a greater impact on downward soil displacement than
water erosion. In the context of management practices aimed at soil protection, any conser-
vation methods must provide solutions to limit this phenomenon: minimal soil tillage or
even zero tillage, high uphill speed, reduced downhill speed, different shaped implements.
New experiments are underway which should provide a better understanding of the impact
of cultivation techniques on soil erosion, notably by indicating the role of different para-
166 M. Guiressc. J.C. Revel /Soil & Tillage Research 35 (1995) 157-166
meters such as the degree and form of the slope as well as the type of implements used and
their velocity.
References
Etchanchu, D., 1988. Gdochimie des eaux du bassin de IaGaronne. Transferts de ma&es dissoutes et patticulaires
vers I’ocean Atlantique. These Univ. Toulouse 3, 178 pp.
Henin, S., 1960. Cours de physique du sol. Tome 1. ORSTOM Editest, 159 pp.
Revel, J.C., 1982. Formation des sob sur marries. Etude d’une chronosequence et d’une toposequence complexes
dans le Terrefort toulousain. These INP Toulouse, 209 pp.
Revel, J.C. and Guiresse, M.. 1995. Dry erosion of calcareous clay soils on the hillsides of south west France, I.
Effect of former farming practices. Soil Tillage Res., 35: 147-155.
Revel, J.C. and Rouaud, M., 1985. Mecanismes et importance des remaniements dam le Terrefott toulousain
(Bassin Aquitain, France). Pedologie, XXXV, 2: 171-189.
Revel, J.C., Coste, N., Cavalie, J. and Costes, J.L., 1990. Premiers r&mats experimentaux sur I’entrainement
mecanique des terres par le travail du sol dam le Terrefort toulousain (France). Cah. ORSTOM, s&e
pedologie, Vol. XXV, no. I et 2, pp. 111-l 18.
Revel, J.C., Coste. N.. Cavalie. J. and Coste;J.L., 1993. Erosion hydrique et entrainement mecanique des terres
par les outils dans les coteaux du Sud-Ouest de la France. La ntcessite d’btablir un bilan avant tome mesure
anti-erosive. In: S. Wicherek (Editor), Farm Land Erosion: In temperate Plains Environment and Hills. Proc.
of a Symposium, 25-29 May 1992, Paris, France, pp. 551-562.
Robinson, A.R., 1977. Relationship between erosion and sediment delivery. Proc. of Paris Symp., July 1977.
International Association of Hydrological Sciences, Pub. No. 122, UNESCO, Paris, pp. 159-167.
Roose. E. and Cavahe, J., 1988. Nouvelle strategic de gestion conservatoire des eaux et des ~01s. Application en
France et en Afrique Occidentale. In: S. Rimwanich (Editor), Land conservation for future generations, Proc.
of ISCO 5, 18-29 January 1988. Bangkok, pp. 913-924.
Rouaud, M., 1987. Evaluation de I’erosion quatemaire, des remaniements de versant et de l’brosion en rigole dans
le Terrefort toulousain. These 3eme cycle Univ. Toulouse 3,320 pp.