Вы находитесь на странице: 1из 11

The effect of tillage depth, tillage

speed and soil condition on chisel


tillage erosivity.
ABSTRACT: in this study, a series of tillage experiments were set up to
investigate the effect of variations in tillage speed on net soil displacement
and associated tillage erosion rates for up and downslope chisel tillage of a
loamy soil under two different soil conditions: 1). a consolidated soil under
stubble vegatation (primary pass), 2). a freshly tilled, loosened soil
(secondary pass). the experimental results show that the average
displacement distance is not only a function of slope gradient, but also of soil
condition, tillage depth, and speed. a model incorporating these additional
effect was proposed and validated using data alvaliable in literature:
variations in displacement distance can be succesfully predicted, but their
absolute magnitude in probably also controlled by tillage implement
geometry, which is at present not incorporated into the model. the fact that
displacement distance are not affected by slope gradient only complicates
the calculation of tillage erosivity. however, if only the controlled variations
in tillage depth and speed are accounted for, the use of a single tillage
transport coeficient k is still posibble without an unacceptable loss of
accuracy. using the model, a series of nomograms have been developed
that allow one to evaluate the effect of tillage depth and / or speed on soil
erosivity of a given tillage operation. these nomograms are a valueable tool
for evaluating possible strategies to remediate tillage erosion due to chisel
tillage. finally, experiment results show that chisel tillage in the Belgian
Loam Belt is very erosive, leading to annual erosios rates exceeding 2
mm/yr (0.08 in yr) locally.

Keywords: Chisel tillage, modeling, soil erosion, soil translocation, tillage


erosion
Past experimental data show that tillage operations are responsible for
the movement of substantial quantities of soil material, resulting in tillage
erosion. As early as 1934, Nichols and Reed executed experiments to study
the movement of soil material as a result of tillage. They showed that soil is
moved in the direction of tillage during moldboard tillage. Mech and Free
(1942) were the first to show, experimentally, that soil movement is related
to slope gradient; in other words, harrowing and chiseling operations along
the contour lines result in a downslope movement of soil material that is far
from insignificant. Weinblum and Stekelmacher (1963) showed that
moldboard tillage along the contours resulted in a net downhill movement of
soil material.

After these early attempts to experimentally study tillage translocation


and tillage erosion, no systematic research was carried out for several
decades. A renewed interest in soil translocation and erosion resulting from
tillage operations occurred in the late 1980s and early 1990s. Lindstrom et
al. (1990, 1992) were the first to carry out a systematic series of tillage
erosion experiments. They reported on the significant relationship between
soil translocation and slope gradient for moldboard tillage. More recently,
consistent experimental data were published on the relationship between soil
translocaion and slope gradient for various agricultural conditions in Europe
(Kiburys 1989; Govers et al. 1994; Guiresse and Revel 1995; Revel et al.
1993; Poesen et al. 1997), North America (Lobb 1991; Lobb et al. 1995;
Montgomery 1993; Montgomery et al. 1999) and the tropics (Turkelboom et
al. 1997, 1999; Thapa et al. 1999a, 1999b).
Based on this relationship, Lindstrom et al. (1992) developed a simple
soil translocation model and showed that tillage operations can be held
responsible significant soil erosion (surface denudation) on convex slope
positions, while accumulation of soil material occurs in concavities. Govers et
al. (1993, 1994) proposed to predict the net unit soil transport rate Qs,net
(i.e., the net downslope movement of soil material due to a specific tillage
operation) by a diffusion type model. To relate the net unit soil transport rate
to the slope gradient S, they introduced the diffusion coefficient k, a term
Grovers el al. (1999) later changed to tillage transport coefficient) expressed
as
Q = kS (1)
This tillage transport coefficient is a measure of the tillage erosivity for
a given tillage operation and can be used to calculate tillage erosion rates-high values correspond to high tillage erosion rates. In subsequent years,
experimentally derived values of the tillage transport coefficient k have
become available for a variety of tillage tools under different tillage
conditions (Table 1). The k value has become a standard measure to describe
the tillage erosivity of a given tillage operation.
The k value listed in Table 1 are only valid for a specific combination of
tillage speed, tillage depth, implement type, and soil condition. Only limited
attention has been given to the effects of the various controlling factors,
other than slope gradient and implement type, on the process of tillage
erosion. In most experimental studies, these variables were kept constant or
were at least assumed to be constant (Lobb et al. 1999). Van Muysen et al.
(1999) have shown that soil condition is important-moldboard tillage of
loose, pretilled soil results in tillage erosion rates that are 3 to 4 times higher
compared to tillage of a consolidated grass fallow soil.

The effect of tillage speed on tillage translocation in the direction of


the tillage has been investigated by Sharifat and Kushwaha (1997) for tillage
with a sweep, while Sohne (1960) reported on the effect of tillage speed on
lateral soil translocation during moldboard tillage, and noted that higher
tillage speeds resulted in larger lateral displacement distances. These
studies, however, did not include the effect of slope gradient, so their results
cannot be used to assess the effect of tillage speed on tillage erosion rates.
Lobb et al. (1999) investigated the effect of tillage depth and tractor
speed on tillage translocation. Although their results suggested that these
effects were significant, the identified relationships were neither consistent
nor strong. They attributed this to the confounding role of other factors, such
as the tractor-implement match and the responsiveness of the tillage
operator. However, the limited number of experiments that were carried out
with a single tillage implement, as well as the fact that no systematic
variation of tillage depth and speed was built into the experimental design
may also have played a role.
Information on the effects of tillage speed, tillage depth, and soil
condition on soil translocation by tillage is a prerequisite for the development
of more sophisticated tillage erosion models, allowing erosion rates to be
predicted and possible conservation strategies to reduce tillage erosion to be
developed and evaluated. Therefore, a series of tillage experiments were set
up using a chisel plough. The main objective of these experiments was to
investigate the effect of tillage speed, tillage depth, and initial soil condition
at the moment of tillage, on the translocation and erosion of soil material by
tillage.
Materials and Method
Study area. Experiments were conducted in July and August of 1997
at a study site in the Municipality of Huldenberg (ca. 10 km east of Brussels)
and situated in the Belgian loam belt. This are is characterized by a rolling
topoghraphy with slope gradients upto 0.30 m m-1 (984 ft). Annual
precipitation ranges between 700 mm and 800 mm (27.58 in and 31.52 in)
and is distributed throughout the year, with the maxima in July-August and
November-January. Maenannual temperature in the study area is 9.5 oC (49
F). The main soil types in the area are Luvisols, Cambisols, and Regosols
(FAO 1990) which have developed in loess deposits.
The land use in the area is characterized by dryland cropping where
cereals make up a large part of the crop rotation, together with maize,
potatoes and sugarbeet. After harvesting the cereal crop, the fields re tilled
with a chisel plough during late summer or early fall. Many farmers till their
land two times at this stage in order to improve seed bed conditions for
consecutive seeding of grasses or leguminosae (soil protectors during winter

time) or for winter wheat as the next crop. As a result of this specific tillage
sequence, soil conditions at the moment of tillage vary cansiderably. The soil
is strongly consolidated at the time of the first tillage operation after the
harvest of cereals--the upper layer is very loose at the time of the second
tillage pass because this tillage is carried out immediately after the first one.
Experimental setup. A hillslope was selected and a detailed
topographycal map of the area was constructed using an automatic
theodolite (Figure 1). Two sets of experiments were conducted on the
hillslope, each with a different tillage treatment. The first set is referred to as
the 'stubble treatment'. Here, the field of winter wheat (Triticum aestivum L)
was harvested one week prior to the experiments and no other farming
operations were carried out. Therefore, the soil was under stubble vegetation
and strongly consolidated. On this treatment, five slope profiles (S1-S5) were
selected. The second treatment is reffered to as the 'pretilled treatment'.
Here, the soil had already been tilled a first time with the chisel plough. For
this treatment, two profiles (T1 and T2) were delineated.
On each profile, 10 to 12 strip plots ca. 1.5 m (4.92 ft) wide
perpendicular to the tillage direction and on varying slope gradients, were
established. Positioning of the strip plots on the different profiles ensured
that half of the strip plots were tilled in the upslope direction and half of the
strip plots in the downslope direction (Figure 1).
For the measurement of soil translocations, numbered aluminium
cubes with a side length of 1.5 cm (0.59 in) were used as tracers. The tracers
were inserted into the soil following the hole drilling technique described by
Govers et al. (1994). On each strip plot, a series of 15 holes with a diameter
of ca. 2 cm (0.79 in) were drilled at intervals of ca. 10 cm (3.94 in). The holes
were drilled to a depth of ca. (0.98 ft) for profiles S1 to S5 and to a depth of
ca. 35 cm (1.15 ft) for profiles T1 and T2. A tracer marked with a unique
number was inserted in each hole and its location precisely recorded using
an automatic theodolite. Next, the hole was filled with fine white sand over a
depth of ca. 5 cm (1.97 in) and another tracer was inserted and it position
recorded. This procedure was repeated until the hole was completely filled. A
total of 56 tracer strips were prepared for the stubble treatment (with an
average of 43 tracers per strip) and 24 strips for the pretilled treatment, with
ca. 40 tracers per strip (Table 2).
Both treatments were tilled using a 3 m (9.84 ft) wide chisel with 13
fixed tines, arranged in three rows of 5, 5, and 4 tines, respectively, with 0.2
m (0.66 ft) between the tines. The chisel was pulled by a 85 kW tractor. After
the tillage operation, the areas immediately up and downslope of the original
location of the tracer strips were carefully excavated and the position of each
tracer recorded. A recovery rate in excess of 84% for the strips of the stubble
treatment 91% for the pretilled treatment was obtained (Table 2).

Tillage speed and depth were kept within the range farmers apply
under present agricultural practice in the area. The original aim of the
experiments was to control tillage equipment of the tractor and varying the
tillage speed between the different profiles. Visual observation during the
experiments showed that tillage depth and tillage speed varied cosiderably
within and between treatments. Therefore, tillage speed and depth were
measured on each strip.
Tillage speed was estimated during the experiment by measuring the
time needed for the tractor to travel 5m across the strip. tillage depth was
estimated as the average difference in elevation between the bottom of the
plough layer, as recorded after excavation of the tracers, and the elevation
of the soil surface for the same location, as derived from the detailed digital
elevation model. for each the strips, bulk density and gravimetric soil
moisture content (three sample per strip) where determined for the top 20
cm, while texture and organic matter content were determined for the top 10
cm (one sample per strip).
From this data, individual horizontal displacement distances of the
tracers were calculated. only those tracers that were situated in the plough
layer were used in these calculation as these were the only ones which were
subject to movement during a tillage pass. no corrections were made for lost
tracers. for each strip, the mean displacement distance in the direction of
tillage was calculated and these data were used in further analyses.
RESULTS AND DISCUSSION
Site description. an overview of the summary statistics for the soil
characteristics is given in table 2. for the stubble treatment, the soil was
rather compacted, resulting in a mean bulk density of the top 20 cm of 1560
kg with a standard deviation of 95 kg. due to dry weather in the antecedent
period, the soil was dry with a mean gravimetric moisture content of 0,14 g.
as the soil was tilled once prior to the experiments of the pretilled treatment.
a reduction of the mean bulk density to 1250 kg was measured. due to some
rainfall in the period between the two sets of experiments, mean gravimetric
moisture content of the soil increased to 0,21 g. the texture of the soil is silt
loam (clay, silt, sand), while the mean organic matter content of the soil is
0,95%.
Tillage erosion intensity. the most common way to analyze tillage
experimental data is by plotting the relationship between the mean
displacement distance of the tracer in the tillage direction and the slop
gradient in the direction of tillage, whereby the slope gradient is taken
negative for downslope tillage and positive for upslope tillage. if the
relationship between the displacement distance and the slope gradient is

linear and it is assumed that tillage depth is constant, k can be estimated as


BLA BLA RUMUS
YANG BELUM DITRANSLET
where k= the tillage transport coefficient
d= tillage depth
b= the slop of the linear regression equation of the relationship between
average soil displacement and slope gradient.
rho= bulk density of the soil
The relationship between displacement distance and slope gradient is
not always linear: it may be different for up and downslope tillage or for
steep slopes; in such cases, other methods may be used to estimate k.
The relationship between displacement distance and slope gradient for
our tillage experiments are shown. as the two sets of experiments were
conducted under two different soil conditions, data were grouped according
to the soil condition : one group of data from the experiments conducted a
stubble soil and one group from the experiments conducted on a stubble soil
and one group from the experiments on a pretilled soil. the experimental
results show that, both for the stubble and pretilled treatment, the mean
displacement distance of soil after one tillage pass is significantly related to
slope gradient. these relationships can be described by linear regression
equations:
for the stubble treatment
RUMUS
Values for the tillage transport coefficient k can then be calculated
from equations 3 and 4 in combination with the data for the bulk density and
average tillage depth. resulting k values are 225 kg m-1 per tillage operation
for the stubble treatment and 550 kg m-1 for the pretilled treatment. these
values are in the same order of magnitude as those reported in the
literature.
At first glance, the observed differences in k values can be attributed
to the initial soil conditions. Due to the pretillage operations, the soil was
loosened, resulting in significantly lower [1250 kg m-3 (78.18 lb ft-3)] average
bulk density of the pretillage treatment compared to the stubble treatment
[1560 kg m-3 (97.44 lb ft-3)]. Furthermore , the loose condition of the soil
allowed the farmer to plough the soil at a significantly greater tillage depth
on the pretillage soil (Table 2). However, when calculating the tillage
transport coefficient k using equation 2, these two compensate for each
other (i.e.,ca. 30% higher tillage depth on the pretillage treatment is
compensated for by the ca. 20% lower bulk density on the pretillage
treatment compared to the stubble treatment).

The observed differences in k value are mainly due to differences in


values for the regression coefficient B in equation 2. Thus B value comprises
the effect of all possible factors controlling the relationship between soil
displacement and slope gradient, including tillage depth and tillage speed.
In this perspective, it is interesting to note that there exists a large
scatter in displacement distances for a given slope gradient (Figure 2). This
scatter can mainly be attributed to observed differences in tillage depth and
tillage speed between individual data points. The present day modeling
approach however, does not allow these effects to be taken into account
directly when calculating the relationship between soil displacement and
slope gradient, using equation 1. Only the average tillage depth is used in
equation 2 to calculate the tillage transport coefficient k, the differences in
tillage depth between individual data points are not included. Furthermore,
the effect of tillage speed is not taken into account at all. Therefore, we
included variations in tillage depth and tillage speed between individual data
points directly into te calculation of the tillage translocation coefficient k.
New methodology to calculate the tillage transport coefficient
k. A simple correlation analysis reveals that there are significant effect of
tillage speed and depth on the average displacement distance. To include
these effects in a predictive model, one could use a multiply linear regression
approach (e.g., Lobb et al. 1999; however, such a model structure does not
take into account the fact that tillage erosion is basically a gravity-driven
process, primarily controlled by slope gradient. Therefore, this different
model structure is proposed
................(5)
Where

d = average soil translocation distance (m)


S = slope gradient (m m-1)
D = tillage depth (m)
V = tillage speed (m s-1)
C = describes the soil condition
A, B, , , are regression coefficients
Similar to equation 1, equation 5 considers soil displacement to be
primarily controlled by slope gradient as a linear function. However, the
exact shape of this relationship depends on the tillage depth, tillage speed,
and soil condition. By using the ratio of the tillage dapth and speed (0.1 m
and 1 m s-1, respectively, for the reference tillage depth and tillage speed),
these factors are introduced into equation 5 as dimensionless scaling
factors. The soil condition factor C in equation 5 incorporates all the effects
of the condition of the soil (e.g., consolidation, soil moisture, rooting,
stoniness) at the moment of tillage, on soil displacement for a given tillage
speed and tillage depth. The exponents of these scaling factors (exponents
, , for tillage depth, tillage speed, and soil condition, respectively)
describe the magnitude of the effect of these factors on the relation between
soil displacement and slope gradient.

Using this model structure a linear regression analysis was carried out
(proc. NLIN, SAS Inst. 1989). In this regression analysis, the soil condition
factor C was by convention taken to be 2 for the stubble treatment and C = 1
for the pretilled treatment. As a result, the relationship between soil
translocation and slope gradient, tillage depth, tillage speed and soil
condition could be described as
RUMUS
Equation 6 clearly shows a negative relationship exists between soil
displacement distance an tillage depth. As the exponent a is negative,
increasing tillage depths result in lower soil displacement distances. This can
be explained as follows: the tracers (and soil) at or near the surface move
over relatively large distances, no matter what the tillage depth is. This is
because their movement is not hampered by the normal stresses exerted by
the overlaying soil mass. At greater depth, the displacement distance is
relatively smaller, as the soil movement is limited by more important normal
stresses. This results in a negative relationship between displacement
distance and initial tracer depth (Figure 3).
Similar observations have been made by Sharifat et al. (1994) and
Kouwenhoven and Terpstra (1970). Consequently, average displacement
distance are high when only the upper centimeters are tilled, while they
decrease with increasing tillage depth. Another factor that may play a role is
the shape of the chisel tines. The implement used in this study was eqquiped
with a small duckfoots on each tine. The effect of the duckfoots on horizontal
soil displacement may be largest when the implement is operated at shallow
tillage depths, again due to lower normal stresses.
The effect of tillage speed is different. The positive value of means
that large values for d occur when tillage speed is increased. S the exponent
of the C factor is negative, one can conclude that, for a given combination
of tillage depth and tillage speed, soil will move ca. 40% furthermore during
one pass on a pretilled soil (where C = 1), compared to tillage of a
consolidated stubble soil .
In order to validate equation 6, we used the data collected by Quine et
al. and Poesen in the Guadalentin area (southeast Spain), as they provided
detailed data on tillage depth and tillage speed variations for individual data
points. In fig.4, the relationship between the simulated displacement
distance, as calculated using equation 6, and the measured displacement
distance d is presented for the datasets of Poesen an Quene. The use of
equation 6 allows one to predict variations in displacement distances for
slope gradients up to ca. 0,25 m m-1 (the range of slope for which equation 6
is calibrated).
Poesen and Quine found that for steeper slopes, the relationship
between displacement distance and slope gradient might be different than
for slopes < 0,25 m. Suggest that on very steep slopes, additional
translocation may occur during downslope tillage due to sliding and rolling of

the soil. However, the present study suggest that variations in tillage depth
play a key role. Tillage depth were very shallow on steep slopes, which, in
agreement with equation 6, resulted in very high displacement distance. A
limitation of equation 6 is that negative displacement distances are predicted
for upslope tillage on very steep slopes. However, this limitation is of the
oretical rather than practical nature as upslope tillage on such steep slope is
very unusual : predicted displacement distances could be set to zero for
these cases.
Although equation 6 allows one to predict the variations in
displacement distances pengarang, there exists a systematic deviation from
the line of perfect agreement. In order words, predicted displacement
distance are sistematically lower than observed displacement distances for
both datasets. These deviations frm the 1:1 line in both cases may, to a large
extent, be attributed to differences in tillage implements and soil conditions.
Both pengarang used a 2,4 m wide duckfoot chisel with 11 tines. The
maximum width of the duckfoot was 0,26 m. Thus, the implement coverage
(i.e., the relative area covered by the implement if it would be projected on a
vertical plane behind the implement) could be estimated as ca. 0,60. For the
chisel used in this study the value would be ca. 0,20. It may, therefore, be
expected that a single tillage pass would result in much larger displacement
distances in the case of the Guadalentin experiments. In addition, the
presence of the wide duckfoot on each tine and the high consentration of
large rock fragments in the Guadalentin experiment result in large rock
fragment trapped between the tines. This may result in soil being
transported further compared to our experiments where no rock fragments
were present and duckfoots were much smaller.
Aplication
The major advantage of incorporating the effect of tillage speed and
tillage depth in the analysis of soil displacement distances and the
calculation of the tillage transport coefficient is that it becomes possible to
evaluate the effect of these factors on tillage translocation and tillage
erosion. Using equations 5, 9, and 10, approximate k values can be
calculated for any combination of tillage depth, speed, bulk density, and soil
condition, once the exponents , B, and and the coefficients A and B are
known. These simulations have been done for two soil conditions similar to
the village experiments (Figure 8a and 8b).
It is clear from these figures that the tillage transport coefficient k
increases with increasing tillage speed, and this effect is more pronounced
for greater tillage depths. Similarly, k values increase with increasing tillage
depth, and this effect is more pronounced for higher tillage speed.
Furthermore, k values for a secondary tillage operation are in all cases higher
than for a first tillage pass on a consolidated soil.
The major advantage of these figure is that they allow one to quickly
asses the tillage transport coefficient k due to chisel tillage for various

conditions, and this not only for a single pass but also for a sequence of
tillage operations, in a case where two consecutive tillage operations are
carried out, the resulting total k value can be obtained by adding up the k
values for each single tillage operation.
From the point of view of soil conservation, these figure allow one to
quickly evaluate possible strategies to minimize tillage erosion. For example,
assume the k value associated with a tillage pass is on loosened soil at a
depth of 0,27 m (0,89 ft) is 400 kg m -1 (268,9 lb ft -1) and the farmer wants
to reduce the associated tillage erosion by 20% without a reduction in tillage
speed [2.0 m s -1 (6,56 ft s -1)] because this would increase the cost of the
tillage operation. Based on equation 6,9 ,and 10 one can calculate that
reducing the tillage depth to ca. 0,19 m (0,62 ft) will result in a reduction of
the k value to ca. 320 kg m -1 (215,1 lb ft -1).
Such strategies may be used in conservation studies to reduce or
minimize tillage erosion, especially in areas where topsoil depth is a limiting
factor for crop production, as is the case in several mediterranean areas.
Conclusion
Most experimental studies on tillage erosion have hitherto focussed on
the erosivity of a given tillage operation under certain soil condition.
However, the experiments presented in this study were designed to
systematically study the effect of soil condition, tillage depth, and speed no
soil translocation by chisel tillage.
Our data clearly show that all these factors have a significant effect on
soil translocation. The data were used to construct a simple model to predict
the average displacement distance from slope gradient, tillage depth, tillage
speed, and soil condition. both the magnitude of soil translocation as well as
its variation with slope are positively related to tillage depth .Also, tillage of a
loose, pretilled soil results in higher displacement distances and a larger
variation with slope than tillage of a consolidated soil under stubble
vegetation.
The proposed model was validated using data presented by poesen et
al.(1997) and quine et al.(1999). This validation show that variation in
displacement distances could be predicted; however absolute magnitudes of
displacement distance were underestimated, which may be due to
differences in implement characteristic.
If variation in tillage depth and speed and important, then they should
be accounted for when assessing tillage erosivity, otherwise, erroneous
results may be obtained. Two components may be distinguished in these
variation. First, the tractor operator may select a given tillage depth and
speed may occur; the latter appear to be systematically related to slope
gradient.
The use of a single k value for practical purposes allows one to take
into account the controlled variations in average speed and depth but
neglects the systematic effect of slope gradient on uncontrolled variations.

However, this simplification leads to errors which are well below 20% for
normal tillage operations. Therefore, nomographs predicting the tillage
transport coefficient k as a function of preset tillage speed and depth may be
a useful tool in developing strategies to minimize tillage erosion.
Use of this simplified model also shows that chisel tillage as applied in
the Belgian Loam Belt is very erosive and may lead to local denudation rates
exceeding i mm (0.04 in) per tillage operation. As chisel tillage is usually
combined with moldboard ploughing, total annual tillage erosion rates can
exceed 3 mm/yr (0.12 in/yr).

Вам также может понравиться