Soil & Tillage Research 51 (1999) 341±356
Ridge tillage and contour natural grass barrier strips reduce
tillage erosion$
B.B. Thapaa, D.K. Cassela,*, D.P. Garrityb
b
a
Department of Soil Science, NCSU, Raleigh, NC 27695-7619, USA
Systems Agronomist and Coordinator, Southeast Asian Regional Research Programme, ICRAF, P.O. Box 161, Bogor, 16001, Indonesia
Abstract
Large amounts of soil are eroded annually from tilled, hilly upland soils in the humid tropics. Awareness has been increasing that
much of this erosion may be due to tillage operations rather than water-induced soil movement. This ®eld study estimated soil
translocation and tillage erosion for four tillage systems on Oxisols with slope gradients of 16±22% at Claveria, Misamis Oriental,
Philippines. Soil movement was estimated using `soil movement tracers' (SMT) which consisted of painted 12-mm hexagonal steel
nuts. The SMT were buried in three replicate plots of the following tillage treatments: (1) contour moldboard plowing in the open
®eld (MP-open); (2) contour ridge tillage in the open ®eld (RT-open); (3) contour moldboard plowing plus contour natural grass
barrier strips (MP-strip); and (4) contour natural grass barrier strips plus ridge tillage (RT-strip). Two hundred SMT were placed at
the 5-cm depth at 5-cm spacings on 10 rows and 20 columns in two microplots within each plot. The microplots were oriented with
the boundaries running downslope and along the contour of each 8-m-wide 38-m-long (downslope) tillage plot. After tilling the
land for four successive corn (Zea mays L.) crops (20 tillage operations), the SMT were manually excavated and their positions
recorded. Recovery of SMT ranged from 82% to 85%. Displacement of SMTwas directly related to slope length, percent slope, and
tillage method. Mean displacement distance of SMT during the four corn growing seasons was 3.3 m for MP-open, 1.8 m for RTopen, 1.5 m for the RT-strip, and 2.2 m for MP-strip. Based on tillage operations associated with two corn crops per year, mean
annual soil ¯ux was estimated to be 241, 131, 158 and 112 kg mÿ1 for MP-open, RT-open MP-strip, and RT-strip, respectively.
Compared to the mean annual soil loss for MP-open of 63 Mg haÿ1, soil loss was reduced by 30%, 45%, and 53% for the MP-strip,
RT-open, and RT-strip systems, respectively. Both ridge tillage and natural grass barrier strips reduced soil displacement, soil
translocation ¯ux, and tillage erosion rates. # 1999 Elsevier Science B.V. All rights reserved.
Keywords: Tillage erosion; Ridge tillage; Grass barrier strips; Soil erosion; Land degradation; Animal powered-tillage; Humid tropics
1. Introduction
Over the last few decades, many researchers have
studied rainfall and runoff water as agents for soil
*Corresponding author.
$
Paper presented at International Symposium on Tillage
Translocation and Tillage Erosion held in conjunction with the
52nd Annual Conference of the Soil and Water Conservation
Society, Toronto, Canada, July 24±25, 1997.
detachment and transport to lower positions on the
landscape. This research led to the development and
re®nement of empirical models such as USLE (Wischmeier and Smith, 1958, 1978), RUSLE (Renard et al.,
1991), and process-based models such as GUEST
(Rose and Freebairn, 1985; Rose, 1988), WEPP (Foster et al., 1989; Nearing et al., 1989), EUROSEM
(Morgan et al., 1990), and MEDALUS (Kirkby et al.,
1993). Even though these models use climatic data,
the condition of the soil surface and its variability
0167-1987/99/$ ± see front matter # 1999 Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 7 - 1 9 8 7 ( 9 9 ) 0 0 0 4 7 - 1
342
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
create uncertainties in the soil loss estimates. The
various soil erosion models have not considered the
effects of tillage on soil erosion.
Tillage is a dynamic process that alters the nature of
the soil surface, detaches and displaces soil aggregates
and clods, and moves or translocates soil to lower
elevations (Powell and Herndon, 1987). This translocation of soil to lower elevations (tillage erosion) has
not been considered to be a major factor contributing
to the overall soil erosion process even though Mech
and Free (1942) reported tillage-induced downslope
soil movement over ®ve decades ago. The new challenge is to model soil erosion by considering water,
tillage, soil creep, wind, and fauna-induced soil transport. In most situations, however, water and tillageinduced soil erosion dominate all other forces of soil
erosion.
Animal-pulled tillage implements and human-powered hoes (Turkelboom et al., 1997; Lewis and Nyamulinda, 1996) are common in present-day farming
systems in the humid tropics. For many soils on
convex slopes, tillage erosion may be responsible
for a signi®cant reduction of topsoil depth and exposure of subsoil (Veseth, 1986). Current knowledge of
tillage erosion is limited, and most of the information
was developed for tillage operations using large tractor-pulled tillage implements and farm machinery on
gently sloping land in North America and Europe
(Lindstrom et al., 1990, 1992; Lobb et al., 1995;
Govers et al., 1994).
On sloping land in the humid tropics, reduction of
net downhill soil movement due to animal or humanpowered tillage implements is essential to sustain crop
production. This is true even for contour alley cropping systems where a reduction in tillage erosion
would reduce the rate of soil scouring in the upper
part of an alley. Alley cropping is a system for which
annual crops are grown in the area or `alleys' between
rows of perennial legume shrubs or trees. This practice
emerged in the early 1970s as a conservation farming
technique on acid, steepland Oxisols and Utisols in the
humid tropics (Benge, 1987; Kang et al., 1990; Nair et
al., 1995). Contour strips of legume shrubs or trees on
sloping land proved to be an effective barrier to waterinduced soil erosion (O'Sullivan, 1985; Lal, 1987;
Young, 1989; Agus, 1993). In the Philippines, Thapa
(1991) found that a 1-m-wide contour strip of the
legume shrub, Desmanthus virgatus, in 12-m-long
plots on a 14±19% slope reduced water-induced soil
loss from 200 to <10 Mg haÿ1 yearÿ1 for land tilled up
and downslope.
On moderately sloping and steep lands, many tillage operations up and downslope still occur, but
plowing across the slope is gaining in popularity
worldwide. Ridge tillage is a conservation tillage
practice in which ridges are reformed atop the planted
row by cultivation, and the ensuing row crop is planted
into ridges formed the previous growing season. The
soil ridges are left from year to year. Crop residue is
left on the soil surface between adjacent ridges and
weeds are controlled with herbicides, interrow cultivation, or a combination of the two. As the crop grows,
soil from between the ridges is thrown up on the ridges
to maintain them. Thapa (1997) evaluated ridge tillage
for the ®rst time on sloping uplands in the humid
tropics and reported signi®cantly higher corn grain
yield for ridge tillage systems at reduced operating
cost compared to moldboard plow tillage systems.
The use of contour grass barrier strips is an alternative for the perennial shrubs or trees in some alley
cropping systems (Thapa, 1997). Contour ridge tillage
in combination with contour natural grass barrier
strips is a simple, practical and effective alternative
conservation method for intensively tilled sloping
uplands in the humid tropics. Information on the
effects of various tillage systems on tillage-induced
soil erosion will provide useful information to further
evaluate their ef®cacy. The objectives of this study
were to evaluate the effects of four tillage systems on
tillage-induced soil translocation downslope (tillage
erosion) and compare the effectiveness of contour
ridge tillage and natural grass barrier strips in reducing
tillage erosion for corn production on highly erodible
steepland soils in the humid tropics.
2. Materials and methods
2.1. Site description
The study began in July 1994 and was superimposed on ongoing experimental plots established in
1992 at Claveria municipality, Misamis Oriental, the
Philippines (88380 N, 1248550 E). The experiment was
located on hillsides ranging from 16% to 22% slope
gradients at ca. 500 m above mean sea level. The soil
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
343
Table 1
Selected soil properties at the Ane-i site, Claveria, Philippines, 1992
Soil
depth
(cm)
Clay
(%)
Bulk
density
(Mg mÿ3)
Base
saturation
(%)
CEC
(cmol‡
kgÿ1)
Exchangeable
aluminum
(cmol‡ kgÿ1)
Soil pH
in H2O
(1 : 1)
Total N
(g kgÿ1)
Bray P
(mg kgÿ1)
Organic
C (g kgÿ1)
0±15
15±37
37±77
78
85
88
0.73
0.85
0.94
33
21
21
9.1
8.2
8.0
1.0
1.6
2.1
4.1
3.7
4.1
2.0
1.0
1.0
5.7
2.0
2.3
16.0
9.0
5.0
is a very ®ne, kaolinitic, isohyperthemic, Lithic Hapludox (Soil Survey Staff, 1992) of volcanic origin
(Table 1). Bulk density was <1.0 Mg mÿ3 throughout
the soil horizons. Pan evaporation in this region
typically exceeds rainfall from January through
May and rainfall exceeds pan evaporation from June
through December. Monthly rainfall and pan evaporation for 1994 and 1995 are presented in Fig. 1. Total
rainfall in 1994 and 1995 was 2020 and 1901 mm,
respectively.
2.2. Tillage and cultural practices
The four tillage systems, each consisting of a
combination of one tillage method and one ®eld
condition, were: (1) contour moldboard plowing in
the open ®eld, the control (MP-open); (2) contour soil
barriers formed by ridge tillage in the open ®eld (RTopen); (3) contour barriers formed by natural grass
strips plus moldboard plowing (MP-strip); and (4)
contour barriers formed by a combination of ridge
tillage and natural grass strips (RT-strip). The 8-mwide 38-m-long plots were arranged in a randomized complete block design with the four tillage
systems replicated three times. The continuity of the
slope length (Fig. 2(A)) was disrupted by contour
natural grass barrier strips in the MP-strip and RTstrip systems (Fig. 2(B)). Five contour natural grass
strips formed four alleyways in each MP-strip and RTstrip experimental plot. Five 50-cm-wide unplowed
contour natural grass strips were established at intervals based on a 1.5 m vertical drop in elevation.
Because these narrow strips were not tilled, natural
grasses were soon established. The grass strip was
dominated by bahia (Paspalum notatum), guinea
(Panicum maximum), and Rottboellia (Rottboellia
cochinchinensis) grasses. Alley width in plots with
grass barrier strips was 9±10 m depending on slope.
By September 1995, distinct terraces had formed
between adjacent grass strips.
Tillage operations to a depth of 20 cm were performed alternately along the contour for all four
Fig. 1. Monthly rainfall and pan evaporation at Claveria, Misamis Oriental, Philippines (IRRI, 1994 to IRRI, 1995).
344
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
Fig. 2. Schematic diagram for (A) open field (MP-open and RT-open) and (B) contour natural grass barrier strip (MP-strip and RT-strip) plots.
systems. For the MP-open system, a single animal
(ox)-pulled, single bottom moldboard plow and harrow (in separate operations) were used to prepare the
land prior to corn seeding. In accordance with local
custom the moldboard plow direction was such to
displace soil downslope. The moldboard plow was
also used to create 15-cm-wide furrows at seeding and
for two `hilling up' operations of the MP-open plots.
Hilling operations, performed primarily to control
weeds and incorporate side-dressed fertilizers, moved
soil both upslope and downslope.
The RT-open system was managed analogous to the
MP-open system for planting and the ®rst hilling
operations. During the second hilling operation, 20cm-high ridges spaced 60 cm apart (distance between
adjacent corn rows) were constructed on the contour
using an ox-pulled, locally-fabricated ridger. The
ridger was a double-blade moldboard plow, which
displaced soil both uphill and downhill from the
furrow to both corn rows. Once formed initially in
1992, the ridges were maintained permanently in the
same position for all four crops. The MP-strip system
was identical to the MP-open system except that
natural grass strips in the former were left in place
permanently after they became established. The RTstrip system was managed using the combination of
practices described for the RT-open and MP-strip
systems.
Corn (Pioneer 3274) was planted within one or two
weeks following harrowing (Table 2). The corn was
hand seeded 2-cm deep and 25-cm apart in each
furrow and covered by foot. Plant population was
67 000 plants haÿ1. Nitrogen, P and K fertilizers
were applied at rates of 80, 30, and 30 kg haÿ1 on
elemental bases, respectively. Hand weeding by
machete in all four systems was performed 40 days
after corn emergence. For RT-open and RT-strip, an
additional hand weeding was necessary two weeks
before crop harvest. The dates and types of implements used for the cultural operations for the four
tillage systems from August 1994 to October 1996 are
given in Table 2.
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
345
Table 2
Dates and types of implements used for four tillage systems for four successive corn crops at Claveria, Misamis Oriental, Philippines
Tillage systema
Crop 1, 1994
MP-open, MP-strip
RT-open, RT-strip
Crop 2, 1995
MP-open, MP-strip
RT-open, RT-strip
Crop 3, 1995
MP-open, MP-strip
RT-open, RT-strip
Crop 4, 1995
MP-open, MP-strip
RT-open, RT-strip
a
Plowing
Harrowing
Planting
Ist Hilling
2nd Hilling
20 August
Moldboard
None
25 August
Harrow
None
9 September
Moldboard
Moldboard
30 September
Moldboard
Moldboard
14 October
Moldboard
Ridger
28 December
Moldboard
None
3 January
Harrow
None
5 January
Moldboard
Moldboard
27 January
Moldboard
Moldboard
12 February
Moldboard
Ridger
2 May
Moldboard
None
7 May
Harrow
None
12 May
Moldboard
Moldboard
5 June
Moldboard
Moldboard
20 June
Moldboard
Ridger
5 September
Moldboard
None
10 September
Harrow
None
22 September
Moldboard
Moldboard
14 October
Moldboard
Moldboard
29 October
Moldboard
Ridger
MP ˆ contour moldboard plowing; RT ˆ contour ridge tillage.
2.3. Soil movement tracers
Prior to inserting soil movement tracers (SMT), the
soil surface elevation was measured at 240 points in
each plot and the average percent slope of each plot
was computed. The SMT were installed on July 1994
at the shoulder slope position in the 8-m-wide 38m-long erosion plots (Fig. 3). The cross slope was
minimal. The SMT were metal 12-mm hexagonal steel
nuts painted with a red or white anti-rust coating. The
SMT were buried in two 90- 95-cm demarcated
microplots. Each microplot was divided into cells
of 10 rows down the hill by 20 columns along the
contour. One SMT was placed at the 5 cm depth at the
nodes of the grid by pressing an index ®nger into the
soil. Red SMT were placed in the left-hand microplot
and white SMT were placed in the right-hand microplot of each plot. The location of all buried tracers was
recorded.
2.4. Excavation of tracers and distance measurement
The SMT were excavated in May 1996 after all
cultural operations and harvests for four successive
corn crops were completed. The SMT were recovered
by manually excavating the entire 8-m-wide 11-mlong plot section. Labor to excavate the tracers for 12
erosion plots required the equivalent of 538 eight-hour
days. Each plot was excavated by systematically
removing 10 cm 10 cm 20-cm deep soil blocks.
The excavation began 0.5 m uphill from row 1 of the
buried SMT. A 20-cm-high metal framework was
pushed into the soil and a metal soil dividing strip
was pushed into the soil to cut the soil 20-cm deep.
Then the soil was cut using a locally fabricated 10cm-wide 20-cm-long knife and removed in 10 cm
10 cm 20 cm deep blocks. Each soil block was
crushed manually and forced through a screen with
10-mm openings. Upslope±downslope and horizontal
coordinates for the recovered SMT were recorded
relative to the references shown in Fig. 3.
Using the center of the original 90- 95-cm tracer
placement area as the origin, the horizontal and
upslope±downslope displacement distance for each
SMT was computed by counting rows and columns.
The actual distance, which is the total distance of each
SMT translocation was calculated using the Pythagorean theorem.
2.5. Statistical analysis and nomogram development
The horizontal, downslope, and actual displacement
distances of all recovered tracers for each erosion plot
were analyzed using the SAS univariate procedure
346
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
Fig. 3. Schematic of area where soil movement tracers were inserted in open field (MP-open, RT-open) and contour natural grass strip (MPstrip, RT-strip) plots. Two hundred red and 200 white 12-mm hexagonal painted stainless-steel nuts were inserted in each plot. Plot dimensions
not to scale.
(SAS Institute Inc., 1988). Analysis of variance was
performed using the generalized linear model (GLM)
procedure (SAS Institute Inc., 1988). Treatment
means were compared by Duncan's multiple range
test. The mean and Q3 displacement distances were
regressed on percent slope for contour moldboard
plowing and ridge tillage.
Nomogram development for each tillage system
involved measurement of the displaced distance of
SMT (discussed above), determination of soil ¯ux,
and computation of tillage erosion rates. The soil ¯ux
(k) was computed as:
k ˆ MD TD BD
(1)
where MD is the mean actual displaced distance
(m yearÿ1), TD the tillage depth of 0.2 m, and BD
is soil bulk density (kg mÿ3). Eq. (1) estimates the soil
¯ux caused by tillage operations based on either the
actual or total displacement distance of SMT assuming
that runoff water, wind, faunal activity, and soil creep
were minimal during the two-year period. Later, in
Fig. 8(A) and (C), we will see that MD can be
expressed in terms of slope gradient using a regression
equation. This soil ¯ux computation is similar to those
of Turkelboom et al. (1997) and Govers et al. (1994),
except that our unit of soil ¯ux, kg mÿ1, represents the
total soil ¯ux for all tillage operations on an annual
rather than a per tillage event basis. Soil ¯ux
(kg mÿ1 yearÿ1) in our case is de®ned as the mass
of soil that moves from a speci®ed location in the ®eld
in response to the summation of an unit length of one
pass of all tillage operations performed annually to
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
grow crops. Mean annual soil loss due to a given
tillage practice refers to the mean annual soil ¯ux that
occurs due to tilling an unit land area with a given
slope length or contour interval.
Mean annual tillage erosion rate (TER) (Mg haÿ1
yearÿ1) was calculated by dividing the soil ¯ux value
by the plot or slope length (L) (downslope for MPopen and RT-open, and alley width or contour interval
distance for the MP-strip and RT-strip systems) and
converting the value to a per ha basis as follows:
TER ˆ
MD…cm† TD…cm† BD…kg=m3 †
L…m†
…1 m†2
…100 cm†
2
Tillage
systema
Recovery
of SMT (%)
Displacedc
(%)
MP-open
RT-open
MP-strip
RT-strip
Standard error
85.2ab
83.7a
81.5a
82.7a
1.2
90.3a
80.0a
80.6a
53.9b
6.1
MP ˆ contour moldboard plowing; RT ˆ contour ridge tillage.
Means in a given column followed by common letters are not
significantly different at the p ˆ 0.05 level.
c
Percent displaced downslope based on row counts only.
b
where TD is 20 cm and 1/2 represents the conversion
of data collected over a two-year period to a one-year
period. Simplifying, the nomogram becomes
(3)
for contour moldboard plowing and
TER ˆ ‰…8:8 S ÿ 5:7†TD BDŠ=20 L
Table 3
Percent recovery of soil movement tracers (SMT) and percentage
that moved downslope from original position after growing four
corn crops at Claveria, Philippines
a
1 Mg 1 104 m2
(2)
ha
103 kg 2
TER ˆ ‰…19:2 S ÿ 72:8†TD BDŠ=20 L
347
(4)
for contour ridge tillage. The function in parentheses
in Eqs. (3) and (4) replaces MD and was derived by
regressing mean displacement distance on percent
slope (S) (presented later in Fig. 8(A) and (C)). Likewise MD in Eq. (2) can be replaced by regression
equations (presented later in Fig. 8(B) and (D)) to
estimate soil erosion rates based on the Q3 distance of
SMT movement.
3. Results and discussions
3.1. Tracer recovery and movement
Percent recovery of the 400 SMT buried per plot
was 82±85% and was not affected by tillage system
(Table 3). However, tillage system did affect the
percent of SMT displaced downslope (p < 0.0001).
Only 54% of the SMT recovered in the RT-strip
system were displaced downslope, but for the other
systems, >80% of recovered SMT were displaced
downslope.
In reality, each SMT likely was displaced from its
original location. Since each SMT was not numbered,
and therefore could not be assigned an exact original
location, it was necessary to measure displacement
distance using the center of the SMT microplot as a
reference (Fig. 3). Consequently, a few SMT for the
ridge tillage system were recovered in a location
having an upslope±downslope component identical
to the original row in which it was placed, but moved
laterally from the original column in which it was
placed. In this situation, calculation of SMT displacement distance relying only on row counts (downslope), such as reported by Turkelboom et al.
(1997), would ignore lateral soil movement and, thus,
underestimate actual SMT displacement. Measurement of downslope displacement distance by Turkelboom et al. (1997) was appropriate because hand
digging on a 60% slope is systematically done by
hoeing down hill. Our situation is different in that
animal-powered, contour plowing can create both
lateral and upslope±downslope soil movement. Plotting of x and y coordinates of each recovered SMT in
our study, in reference to the original position, allows
comparison of both upslope±downslope and horizontal displacement of SMT for the different tillage
systems.
The displaced position of each recovered SMT after
growing four corn crops, in reference to the center of
the 90- 95-cm microplots is shown in Fig. 4 for the
four tillage systems. The greater spread of tracers
downslope for the MP-open compared to the RT-open
system was expected because the soil in the ridges in
the latter tend to reduce the rate of downward movement. Likewise, downhill tracer movement was
greater for the MP-strip than for RT-strip system.
348
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
Fig. 4. Displacement of SMT for four tillage systems on 16% slope after cultivating the land for four corn crops. Total number of recovered
SMTs out of 200 for each color are given in parentheses. The reference point marked by the large dot represents the center of the 90- 95 cm
area of SMT installed microplot. One row ˆ 10 cm, one column ˆ 10 cm.
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
349
Fig. 5. Displacement of red SMT for four tillage systems on 19% and 22% slopes after cultivating the land for four corn crops. Total number
of recovered SMTs out of 200 for each treatment are given in parentheses. The reference point marked by the large dot represents the center of
the 90 95 cm area of SMT installed microplot. One row ˆ 10 cm, one column ˆ 10 cm.
350
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
For each tillage method, SMT movement was less
under grass strips (MP- and RT-strip) compared to the
open ®eld. Tracers for MP-open and MP-strip moved
further downslope than those for RT-open and RTstrip. An increase in percent slope exacerbates downhill movement of SMT for the MP systems (Fig. 5).
Slope steepness had less effect on tracer movement for
the RT systems.
The reference point for calculation of downslope
movement of tracers is between rows 5 and 6 (Fig. 3).
This coincides with row zero in Figs. 4 and 5. The
reference line in the microplot used to calculate
horizontal tracer movement is between columns 10
and 11 (Fig. 3). When we ®rst began to excavate the
SMT, we excavated the border area between tillage
plots in addition to the entire plot area because we did
not know how far the tracers would move horizontally.
After excavation of three plots (RT-open, MP-strip,
and RT-strip in Fig. 4), we discovered that few tracers
moved into the border area, and we discontinued
excavation of the border area.
The relative frequencies of actual SMT displacement distance for the four tillage systems and three
slopes are presented in Fig. 6. Incorporation of ridge
tillage into a tillage system reduces actual SMT movement. Most SMT under the grass strip systems moved
<6 m from the original position, whereas some of the
SMT under open ®eld systems moved 8 m or more.
Observation of relative frequency data in Fig. 6(A, C,
and E) shows that the greatest actual displaced distance for the open system increased with percent
slope. The open system has no barriers to restrict
downslope SMT transport as exists for the grass strip
systems.
Fig. 6. Relative frequency distribution of actual displaced tracers for four tillage systems at three slopes.
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
351
Mean displacement distance increased with percent
slope for each tillage system, but the rate of increase in
displacement distance with respect to slope was higher
for the MP systems compared to the RT systems. The
highest standard deviation occurred for MP-open,
which had the greatest tracer dispersion. The standard
deviation of displacement distance is useful as an
indicator of random dispersion of tracers due to
different tillage systems. Standard deviation for each
treatment is expected to be smallest in the ®rst year of
tillage and to increase as the number of tillage operations increases.
The analysis of variance for mean, Q3, and maximum actual and downslope displacement distances
are shown in Table 5. Highly signi®cant differences in
these statistical parameters as affected by tillage system were observed. In Fig. 7 we see that the greatest
mean and maximum actual displacement distances
occurred for MP-open.
The observed difference in mean SMT displacement distance between MP-open and MP-strip was
due to the contour natural grass barrier strips. The
grass strips reduced the amount of soil moving down-
Lindstrom et al. (1990) reported a strong correlation
between tillage-induced soil movement and percent
slope. Other factors reported to affect tillage-induced
soil movement include slope length (Turkelboom et
al., 1997), speed of tillage operation, size of tillage
implement, and direction of tillage (Lobb et al., 1995).
Factors such as tilling fallow land during high weed
incidence, or ®rst plowing for seed bed preparation,
might contribute to the dispersion of SMT in the MP
systems. Roots and large clods accumulating in front
of the moldboard plow often cause soil to be dragged
along with the plow. In addition, when turning a
moldboard plow at the end of a row, a plowman
applies extra force to free the plow of any accumulated
soil so as to lighten the plow before lifting it from the
furrow. Lifting the plow takes less physical effort
when the plow is turned downhill.
3.2. Displacement distance of tracers
Various statistical parameters based on actual displacement distance for the four tillage systems at 16%,
19%, and 22% slope gradients are shown in Table 4.
Table 4
Summary of actual displacement distances of soil movement tracers for four tillage systems on three slopes
Tillage systema
nb
Mean
Maximum
Q3c
SDd (cm2)
p < We
cm
16% Slope
MP-open
RT-open
MP-strip
RT-strip
341
329
327
334
288
135
195
125
780
462
512
371
361
177
239
170
78
40
48
42
0.0001
0.0880
0.0010
0.0001
19% Slope
MP-open
RT-open
MP-strip
RT-strip
322
327
325
320
319
192
213
165
789
517
508
463
418
230
255
201
90
45
46
44
0.0001
0.0251
0.0001
0.2994
22% Slope
MP-open
RT-open
MP-strip
RT-strip
359
350
326
338
383
211
241
170
1013
542
622
510
470
255
309
211
92
47
53
45
0.0025
0.2678
0.0068
0.4513
a
MP ˆ contour moldboard plowing; RT ˆ contour ridge tillage.
Number of tracers recovered out of 400 SMT buried per plot.
c
The distance beyond which 25% of the tracers were displaced.
d
Standard deviation.
e
Probability test in univariate procedure which is similar to `T' in ANOVA.
b
352
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
Table 5
Mean squares from randomized complete block analysis of variance for various statistical parameters associated with SMT downslope and
actual displacement distances (cm) as a function of tillage system
Source
Downslope displaced distance of tracers
Block
Tillage system
Pooled error
CV (%)
R2
Actual displaced distance of tracers
Block
Tillage system
Pooled error
CV (%)
R2
a
b
df
Mean
Q3a
Maximum
2
3
6
17658.8
17257.0b
334.0
16.5
0.97
5293.8
101913.9b
2257.6
16.0
0.96
9300.0
181355.6b
2257.6
16.0
0.96
2
3
6
4318.2
18224.8b
244.2
7.1
0.98
5581.4
29569.0b
264.8
5.9
0.98
5581.4
29569.0b
264.8
5.9
0.98
The distance beyond which 25% of tracers displaced.
Indicates significance at the p ˆ 0.001 level.
slope and gradually reduced the percent slope in the
alley between adjacent grass strips. Failure to observe
differences in mean SMT displacement distance
between RT-open and RT-strip was due to the large
reduction in SMT movement under ridge tillage alone.
Both ridge tillage systems reduced SMT displacement
distance compared to the MP systems.
3.3. Models for displacement distance of tracers
Linear regression models relating the mean and Q3
parameters for actual SMT displacement distance as a
function of percent slope for two tillage systems are
presented in Fig. 8. The high r2 value for moldboard
plowing in Fig. 8(A) indicates that soil displacement
is highly related to percent slope. The slope of the
curves for contour moldboard plowing (Fig. 8(A) and
(B)) was two- to three-fold higher compared to the
slopes of contour ridge tillage (Fig. 8(C) and (D)). The
low r2 value and gentle slope for the ridge tillage
curves indicate that soil movement for this practice is
not as closely related to percent slope as for moldboard
plowing.
3.4. Soil flux and tillage erosion rates
Fig. 7. Mean (A) and maximum (B) distances of displaced tracers
based on actual and row (downhill) distances for four tillage
systems (MP ˆ moldboard plowing; RT ˆ ridge tillage; O ˆ open
field; S ˆ grass strip). Each value is the mean of three observations.
Means based on actual distance followed by common letters above
the bars are not significantly different at the p ˆ 0.05 level. Vertical
lines at the top of each bar indicate the standard error.
Estimation of soil ¯ux and tillage erosion rates was
based on actual displacement distance of tracers
assuming that the tracers and soil material move at
the same rate. The mean and Q3 for soil ¯ux and
tillage erosion rate were highest for MP-open and
lowest for RT-strip (Table 6). Annual soil ¯ux and
tillage erosion rate were greater for the MP-open
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
353
Fig. 8. Relationship between percent slope and mean and Q3 actual distance of displaced tracers for moldboard plowing and ridge tillage
systems at Claveria, Philippines. The symbols * and ** indicate significance at the p ˆ 0.05 and 0.01 levels, respectively.
compared to the MP-strip, but no differences in soil
¯ux and erosion rates were observed between RT-open
and RT-strip. The relation of soil ¯ux to tillage opera-
tions performed on an annual basis are compatible
with water erosion rates reported in a companion study
(Thapa, 1997). The soundness of using Q3 to estimate
Table 6
Annual soil flux and tillage erosion rates for four tillage systems at Claveria, Philippines
Tillage systema
Annual soil flux rate
Mean
Annual tillage erosion rate
b
Q3
Kg mÿ1 yearÿ1
241ac
131 c
158 b
112 c
6.6
MP-open
RT-open
MP-strip
RT-strip
Standard error
304 a
161 c
195 b
141 c
6.9
df
Mean squares from ANOVA
Blocks
Treatmentsd
Residuals
CV %
2
3
6
2301.6
9714.5
130.1
7.1
2975.5
15760.2
141.0
5.6
63.4
34.4
41.5
29.5
1.7
a
c
b
c
159.5
672.9
9.0
7.1
MP ˆ contour moldboard plowing; RT ˆ contour ridge tillage.
Q3 soil flux and erosion rate are computed based on the distance beyond which 25% of the tracers were displaced.
c
Means in a given column followed by common letters are not significantly different at the p ˆ 0.05 level.
d
Significant at the p ˆ 0.0001 for all soil flux and tillage erosion variables.
b
Q3b
Mg haÿ1 yearÿ1
Source
a
Mean
80.0
42.4
51.4
37.2
1.6
a
c
b
c
205.4
1091.6
9.7
5.9
354
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
Fig. 9. Nomograms to determine tillage erosion rate (TER) for contour moldboard plowing and contour ridge tillage systems as a function of
percent slope and countour interval distance or downhill plot length at Claveria, Philippines. S ˆ percent slope, TD ˆ tillage depth (m),
BD ˆ bulk density (kg mÿ3), and L ˆ slope length or contour interval distance (m).
soil ¯ux and tillage erosion rates as proposed by Thapa
et al. (1999) needs further investigation but the concept appears to be valid.
The 63.4 Mg haÿ1 yearÿ1 of tillage erosion in the
MP-open system is very high and much soil eventually
will be lost from the ®eld if soil continues to move
downslope unimpeded. Installation of contour grass
strip barriers reduced the annual tillage erosion rate
by one-third. The higher tillage erosion rate in MPstrip (42 Mg haÿ1 yearÿ1) compared to RT-strip
(30 Mg haÿ1 yearÿ1) indicates a more rapid rate of
terrace formation for MP-strip. Soil moved from the
upper alley area in the MP-strip system accumulates
on the upslope side of contour grass strips and forms a
terrace more quickly. This process accelerates topsoil
loss in the upper alleyway, which may cause serious
soil fertility degradation in the zone just below the
contour grass strip. Thus, from the ®eld or mass
balance point of view, redistributed soil is not lost
from the ®eld because the soil material remains on the
terrace. However, from a nutrient management point
of view, breaking the continuum of slope length with
contour grass strips creates zones of land with scoured
and nutrient-poor surface soil, and a zone in the lower
alleyway with accumulation of a nutrient rich and
deeper topsoil (Govers et al., 1994; Poesen, 1995;
Garrity, 1996; Turkelboom et al., 1997). We estimate
that as much as 40% of the cropped alleys might
eventually be degraded physically, chemically, or
biologically if the alleys are moldboard plowed frequently. In this study, 42 Mg haÿ1 yearÿ1 of soil in the
MP-strip system moved downhill from the upper alley,
leaving an exposed acidic subsoil with high Al saturation. Although the reduced yields in the upper alleyways may be compensated partially or wholly by
increased yields in the lower alleyways, and the soil
fertility of the upper alleyways may recover gradually
over a number of years, careful management of these
spatially variable areas within an individual alley must
be provided to avoid short- and long-term negative
effects on crop production.
3.5. Nomogram for tillage erosion rates
Nomograms to determine soil losses for contour
moldboard plowing and ridge tillage as a function of
percent slope and plot length (distance between adjacent contour grass strips or length of open plot) are
shown in Fig. 9. Regression equations relating displacement distance to percent slope (Fig. 8(A) and
(B)) were used for the MD term in Eq. (1). Soil
redistribution to lower alleyway zones under moldboard plowing can be as high as 500 Mg haÿ1 yearÿ1
if a farmer cultivates land with a slope of 20% and a
grass strip contour interval or plot length of 5 m.
About 53% of this soil movement can be reduced
by adopting ridge tillage.
4. Conclusions
Intensive moldboard plowing, where practiced, is a
primary factor of land degradation in humid tropical
uplands because the more fertile topsoil is physically
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
displaced downhill exposing less fertile, acidic subsoil. Soil degradation by moldboard plowing and its
long-term consequences on soil physical, chemical,
biological, and hydrological properties has been
omitted from alley cropping research and soil erosion
modeling. This omission may help explain the failure
of alley cropping to meet farmers' expectations to
have higher crop production. The present challenge is
to model soil erosion by considering all soil transport
agents-tillage, water, wind, soil creep, macro-fauna
and human activities. In this study, we assumed negligible rates of soil erosion by wind, soil creep, and
faunal activity. This limitation should be taken into
consideration when extrapolating our results to other
®eld conditions.
Reduction of tillage erosion is indispensable to
maintain soil fertility and sustain crop production
on sloping lands in the humid tropics, especially under
contour hedgerow alley cropping systems. Contour
ridge tillage in conjunction with natural grass barrier
strips gradually reduces the degree of slope, breaks the
continuum of downhill slope length, and reduces the
amount of soil removed from the ®eld.
5. List of symbols
SMT soil movement tracersÐdevices, natural or
manmade, placed in the soil to assess the mass
transport of soil
MP moldboard plow
RT
ridge tillage
soil flux, i.e. the mass of soil that moves
ks
through a vertical cross-section of the plough
layer of unit width due to the combined effects
of all tillage passes performed annually to
grow crops
MD displaced distanceÐthe distance a SMT is
displaced in one year
TD
tillage depth
BD bulk density
TER mean soil loss due to contour moldboard
plowing on ridge tillage contour (Mg haÿ1
yearÿ1)
Q3
the distance beyond which 25% of all
recovered SMT moved
Q1
the distance beyond which 75% of all
recovered SMT moved
S
L
355
percent slope
downslope length or contour interval distance
(m)
References
Agus, F., 1993. Soil processes and crop production under contour
hedgerow systems on sloping Oxisols. Ph.D. Dissertation (Soil
Science), North Carolina State University, Raleigh, North
Carolina, USA.
Benge, M.D., 1987. Agroforestry systems (mimeo). Bur. Sci.
Technol., Agency Int. Dev., Washington, D.C. In: Kang, B.T.,
Reynolds, L., Atta-Krah, A.N., 1990. Alley Farming. Adv.
Agron. 43, pp. 315±359.
Foster, G.R., Land, L.J., Nearing, M.A., Finker, S.C. and Flangan,
D.C., 1989. Erosion component. In: Lane, L.J., Nearing, M.A.
(Eds.), USDA-Water Erosion Prediction Project: Hillslope
Profile Model Documentation 10(1±10), pp. 12.
Garrity, D.P., 1996. Tree-soil-crop interactions on slopes. In: Ong,
C.K., Huxley, P. (Eds.), Tree-Crop Interactions: A Physiological Approach. CAB Int'l and ICRAF, Wallingford, Oxon, UK,
pp. 299±319.
Govers, G., Vandaele, K., Desmet, P., Poesen, J., Bunte, K., 1994.
The role of tillage in soil redistribution on hillslopes. European
J. Soil Sci. 45, 469±478.
IRRI, 1994. Compilation of weather data for 1994. Climate Unit,
Agronomy, Plant Physiology, and Agroecology Division, Los
Banos Laguna, Philippines.
IRRI, 1995. Compilation of weather data for 1995. Climate Unit,
Agronomy, Plant Physiology, and Agroecology Division, Los
Banos Laguna, Philippines.
Kang, B.T., Reynolds, L., Atta-Krah, A.N., 1990. Alley farming.
Adv. Agron. 43, 315±359.
Kirkby, M., Baird, A.J., Diamond, S.M., Lockwood, J.G.,
Mcmahon, M.D., Mitchell, P.L., Shao, J., Sheehy, J.E.,
Thornes, J.B., Woodhard, F.I., 1993. The MEDALUS slope
catena model: a physically based process model for hydrology,
ecology and land degradation interactions. In: Thornes, J.,
Brandt, J. (Eds.), Mediterranean Desertification and Land Use,
Wiley, UK.
Lal, R., 1987. Science. 236, pp. 1069±1076. In: Kang, B.T.,
Reynolds, L., Atta-Krah, A.N., 1990. Alley Farming. Adv.
Agron. 43, pp. 315±359.
Lewis, L.A., Nyamulinda, B., 1996. The critical role of human
activities in land degradation in Rwanda. Land Degradation and
Development 7, 47±55.
Lindstrom, M.J., Nelson, W.W., Schumacher, T.E., 1992. Quantifying tillage erosion rates due to moldboard plowing. Soil Tillage
Res. 24, 243±255.
Lindstrom, M.J., Nelson, W.W., Schumacher, T.E., Lemme, G.D.,
1990. Soil movement by tillage as affected by slope. Soil
Tillage Res. 17, 255±264.
Lobb, D.A., Kachanoski, R.G., Miller, M.H., 1995. Tillage
translocation and tillage erosion on shoulder slope landscape
positions measured using 137Cs as tracer. Canadian J. Soil Sci.
75, 211±218.
356
B.B. Thapa et al. / Soil & Tillage Research 51 (1999) 341±356
Mech, S.J., Free, G.A., 1942. Movement of soil during tillage
operations. Agric. Eng. 23, 379±382.
Morgan, R.P.C., Quinton, J.M., Rickson, R.J., 1990. Structure of
the soil erosion prediction model for the European Community.
In: Proceedings of International Symposium on Water Erosion,
Sedimentation and Resource Conservation, October 1990,
Dehradun, India. Central Soil and Water Conservation Institute,
pp. 49±59.
Nair, P.K.R., Kass, D.L., Kang, B.T., 1995. Nutrient cycling and
soil-erosion control in agroforestry systems. In: Juo, A.S.R.,
Freed, R.D. (Eds.), Agriculture and the Environment: Bridging
food production and environmental food protection in developing countries. ASA Spec. Publ. 60. ASA, Madison, WI, pp.
117±138.
Nearing, M.A., Foster, G.R., Lane, L.J., Finker, S.C., 1989. A
process-based soil erosion model for USDA water-erosion
prediction project technology. Transactions of ASAE 32, 1587±
1593.
O'Sullivan, T.E., 1985. Farming systems and soil management: the
Philippines±Australian development assistance program experience. In: Craswell, E.T., Remenyi, J.V., Nallana, L.G.
(Eds.), Soil Erosion Management. ACIAR Proceedings Series
6. Canberra: ACIAR, pp. 77±81.
Poesen, J., 1995. Soil Erosion in mediterranean environment.
Proceedings of the European School of Climatology and
Natural Hazards Course, held in el Campello, Pueblo
Acantilado, Alicante, Spain 6 to 13 October, 1993, pp. 123±
151.
Powell, G.M., Herndon, L.P., 1987. Maintenance of field
conservation structures. In: Optimum Erosion Control at Least
Cost. Proceedings of a National Symposium on Conservation
Systems. American Society of Agri. Eng., St. Joseph, MI, pp.
374±383.
Renard, K.G., Foster, G.R., Weesies, G.A., Porter, J.P., 1991.
RUSLE, Revised universal soil loss equation. Journal of Soil
and Water Conservation 46, 30±33.
Rose, C.W., Freebairn, D.M., 1985. A new mathematical model of
soil erosion and deposition processes with applications to field
data. In: El-Swaify, S.A., Moldenhauer, W.C., Lo, A. (Eds.),
Soil Erosion and Conservation. Soil Conservation Soc. Am.,
Ankeny, IA, USA, pp. 549±557.
Rose, C.W., 1988. Research progress on soil erosion processes and
a basis for soil conservation practice. In: Lal, R. (Ed.), Soil
Erosion Research Methods. Soil and Water Conservation
Society, Ankeny, IA, USA, pp. 119±140.
SAS Institute Inc., 1988. SAS/STAT1 Users' Guide, Release 6.03
ed. SAS Institute Inc., Cary, NC.
Soil Survey Staff, 1992. Keys to soil taxonomy, fifth edn. SMSS
Technical Monograph. No. 19. Blacksburg, Virginia, Pocahontas Press Inc.
Thapa, B.B., 1991. Evaluation of infiltration, surface runoff and
soil losses at various levels of rainfall in relation to surface
cover, tillage, and soil management practices. M.S. Thesis (Soil
Science), University of the Philippines, Los Bonos Lagura,
Philippines.
Thapa, B.B., 1997. Contour ridge tillage and natural grass barrier
strip effects on soil erosion, soil fertility, and corn production
on sloping oxisols in the humid tropics. Ph.D. Dissertation (Soil
Science), North Carolina State University, Raleigh, North
Carolina, USA.
Thapa, B.B., Cassel, D.K., Garrity, D.P., 1999. Assessment of
tillage erosion rates on steepland Oxisols in the humid tropics
using granite rocks. Soil Tillage Res. 51, 233±243.
Turkelboom, F., Poesen, J., Ohler, I., Van Keer, K., Ongprasert, S.,
Vlassak, K., 1997. Assessment of tillage erosion rates on steep
slopes in northern Thailand. Catena 199, 1±16.
Veseth, R., 1986. Tillage erosion - changing landscapes and
productivity. In: Veseth, R., Wysocki, D. (Eds.), STEEP
Extension Conservation Farming Update, Fall 1986. University
of Idaho, Moscow, ID, pp. 2±7.
Wischmeier, W.H., Smith, D.D., 1958. Rainfall energy and its
relationship to soil loss. Trans. Amer. Geophys. Union 39, 285±
291.
Wischmeier, W.H., Smith, D.D., 1978. Predicting rainfall erosion
lossesÐa guide to conservation planning. USDA, Agriculture
Hand Book No. 537.
Young, A., 1989. Agroforestry for Soil Conservation CAB Int'l.,
Wallingford, UK.