A Participatory Assessment of Soil Erosion and Farm Management Practices in Northwest Ethiopia

Mehretie Belay1 and Woldeamlak Bewket2

1
PhD student, Department of Geography and Environmental Studies, Addis Ababa University, P.O. Box 1176, Addis Ababa,
Ethiopia, Email: belaymehrete@yahoo.com, Fax: 0587711764
2
Associate professor, Department of Geography and Environmental Studies, Addis Ababa University, P.O. Box 150372, Addis
Ababa, Ethiopia, Email: wbewket@geog.aau.edu.et

ABSTRACT
Soil erosion is a widespread problem on cultivated fields in northwest Ethiopia. Plot level survey studies of soil erosion and
conservation are few and far fewer have involved farmers in their assessments of the erosion process and farmers’ conservation
efforts. This paper presents the outcome of a farmer-participatory research conducted at two rural communities, Dubi and Gayta,
in Dangila Woreda (district), in the northwestern highlands of Ethiopia. The study estimated the extent of soil erosion from tree
root exposure measurements and identified farmers’ soil and water conservation (SWC) practices by categorizing the farmers
into three income groups: poor, medium and rich households. Data were collected from 31 plots between May and October
2010. Descriptive statistics and analysis of variance (ANOVA) were used to analyze the data. The results indicate average rates
of soil erosion to be about 1.26 mm year-1, but rates varied from 1.94 mm year-1 on seriously affected sites to 0.21 mm year-1 on
the relatively less affected ones. No statistically significant difference was observed in soil loss between the three household
income groups. The farmers used contour farming, traditional ditches, grass and tree planting for SWC purposes. The study
concluded that as the extent of soil erosion is highly variable spatially, plot and location specific SWC measures that are
designed by considering farmers’ indigenous knowledge will be required to control soil loss in the study area. This study
demonstrates that participatory plot level tree root exposure assessment provides useful information for soil and water
conservation planning.
Key words: Soil erosion; farmer participation; tree root exposure; conservation; Ethiopia.

INTRODUCTION

In Ethiopia, agriculture is the main source of livelihood, problem to sustainable agriculture. Even though soil
employment, and foreign exchange earnings. It supports erosion is a serious environmental and economic
the livelihood of about 90% of the poor and generates problem in Ethiopia, available studies on soil erosion
90% of the national export trade and greater than 40% of measurement or estimation are quite few, and many are
the Gross Domestic Product (Diao, 2010). But, the watershed or plot scales to “represent the diverse
agricultural sector in Ethiopia is confronted with diverse environments of the country” (Bewket and Sterk, 2003).
environmental problems. Land degradation in the form Herweg (1992) has noted that the model of test plots is
of soil erosion causes a severe damage on crop lands, questionable, because they are confined to limited areas
particularly in the highlands. The steep terrain, erosive that might not represent wide spatial segments.
rains, improper use of land and water, rapid population Similarly, Stocking and Murnaghan (2001) state that
growth and dependence on fragmented subsistence experimental plots and quantitative models are
farming are major causes of the ongoing land ‘researcher-centered’ and vague to farmers in less
degradation (Berhane and Mekonnen, 2009). According developed countries to easily assess the level of erosion
to Constable (1984), some 1.9 billion tons of soil are on farm plots. Bewket and Sterk (2003) also note that
removed each year from the Ethiopian highlands. Most plot and watershed level measurements do not show the
of this loss was estimated to occur from cultivated fields extent of eroded soil from individual farm fields due to
where the soil rate was estimated at 100 t ha-1 year-1. gap in scale of measurement and applied methodologies.
Other studies have indicated much lower rates of soil Stocking and Murnaghan (2001) argue that participatory
loss. For instance, Bewket and Sterk (2003) indicated survey methodologies instead integrate local farmer
soil loss rates ranging between 18 to 79 t ha-1 year-1 from experiences with scientific methods and provide
two micro-watersheds in northwestern highlands, and opportunity for smallholders to easily estimate soil
Nyssen et al. (2007) reported 57.3 t ha-1 year-1 average erosion from their farm plots and thus are farmer-
sheet and rill erosion in Tigray, northern Ethiopia. All focused designs. Consideration of farmers’ experiences,
studies of soil erosion in Ethiopia however agree that the knowledge and views can provide essential inputs to the
rate of soil erosion is high and constitutes a serious success of conservation interventions. The approach can
even provide researchers the chance to learn from may also develop clear and practical awareness on the
farmers how they realize and control soil erosion part of participant farmers. Tree root exposure
(Stocking and Murnaghan, 2001). For instance, Abera measurements and computations do not require complex
and Belachew (2011) have observed farmers classifying techniques and models. They are simple, cheap and
soils based on fertility and designing their farming flexible to explore with both semi-quantitative and
practices using their indigenous know-how in qualitative methods. However, the methodology has
southeastern Ethiopia. Similarly, Okoba and Sterk some limitations. During tillage operations exposed
(2006) identified eleven erosion indicators through the roots are covered with soil and make identification
use of farmers’ indigenous knowledge and participation difficult. Weed heaps collected at the base of some trees
in Kenya. also make difficult to notice root exposures.
One method of the participatory approaches is field Nevertheless, owing to the less cost and time required,
survey of tree root exposures (Stocking and Murnaghan, the ease of measurements and computations, there is a
2001). Tree root exposure shows the removal of top soil possibility to consider it as one appropriate method to
covering the root part of a tree due to sheet or rill assess soil erosion. The aim of this paper was to assess
erosion. As trees are frequent and familiar biological the extent of soil erosion using tree root exposure
features, the condition can be easily perceived and measurements and examine farmers’ SWC practices in
measured by farmers using simple tools like a ruler. This the north western part of Ethiopia.

MATERIALS AND METHODS livelihood of the people in both sites. Crops and
community grazing fields occupy large areas while
Site description forests comprise small proportions. Tef (Eragrostis tef)
The study was carried out in two Rural Kebele in Gayta and maize (Zea mays) in Dubi are leading crops
Adminstrations (RKAs, the lowest tiers in Ethiopia’s in area coverage and quantity of output. Vegetables and
administrative structure) named Gayta and Dubi in fruits are important crops cultivated using traditional
Dangila woreda, northwestern Ethiopia (Figure1). The irrigation around homesteads.
two RKAs cover 2332 and 2358 ha, respectively, and
experience slight differences in altitude and local Data and methods
climatic conditions. They form part of the northwestern Erosion assessment by measuring tree root exposures
highlands with elevations varying between 2100 m to has been conducted by many researchers in different
2300 m asl in Gayta and 1850 to 2255 m asl in Dubi. countries. To mention some, Dunne et al. (1979) used
The local relief of both areas is broken by small streams the method and estimated average soil erosion rates at 8
and gullies that often fill with rainwater during the rainy mm year-1 on basement rocks and 14.7 mm year-1 on the
season. The climate is moist Weina-Dega (sub-tropical) lava plateaus in Kenya. Bodoque et al. (2005) applied
with a mean annual temperature of 170C and annual the method in their study in Central Spain and estimated
rainfall of 1578 mm as measured at Dangila town. mean erosion rates at 1.7-2.6 mm year-1 in one of their
Based on color, the local people identify three soil study sites (Senda Schmidt) and 1.1-1.8 mm year-1 in
types: forefor (red color), mezega (black) and bunama another site (Monterrubio). They used dendro-
(grey-brown) as dominating the areas. The red soils chronological analysis to determine the age of the trees
(which belong to the nitosols group) commonly occur on they used as references. But, Stocking and Murnaghan
hilly and sloping parts in about half of the study sites. (2001) note that tree-ring dating is not usually accurate
They exhibit a clay-loam texture and are most because tree-rings are not annually created all the time,
intensively cultivated but also most seriously eroded. particularly in the tropics and subtropics. Trails and
The black soils (vertisols group) are more prevalent in trekking grounds were main focus areas of Bodoque et
Dubi and often cover low lying landscapes. The grey- al. (2005) study, while grazing fields were the focus
brown soils (luvisols group) frequently occupy the areas of the study by Dunne et al. (1979). This study
pediments. Croton (Croton macrostachyus), Acacia differs from these two studies in that it focuses on farm
(Acacia lahai), Eucalyptus (Eucalyptus camaldulensis), plots and it involved farmers in the survey process. Also,
Cordia (Cordia africana), Albizia (Albizia gummifera), the ages of measured trees were determined by
Terminalia (Terminalia brownie), Sesbania (Sesbania consulting the owners who have planted them. Thus, it
sesban) and Justicia (Justicia schimperiana) form the was designed based on the ‘farmer-perspective’ model
dominant vegetation types in the areas. Crop-livestock suggested in Stocking and Murnaghan (2001). The
mixed subsistence farming is the basic source of
2 

 
approach involved farmer participation in identification, measurement, mapping and ranking of affected areas.

Figure. 1. Location map of the study area.

The study was based on participatory field ownership and annual food production1. Based on this,
observation, formal and informal discussions with the 22 plots were selected from the poor and medium
local farmers and plot level field measurements. For this, farmers (11 from each) and the rest (9 plots) were from
first two Farmers’ Groups (FGs), each containing five the rich farmers (Table 2).
members, were set up in each study RKA. These FGs Assessment of tree root exposures was conducted in the
discussed, with facilitation by the lead author, on the presence of plot owners and FGs. A detailed observation
problems of soil erosion and SWC practices and of the surrounding biophysical environment of the farms
identified four villages in their respective RKAs as was made before measurement of exposed roots. The
severely affected, moderately affected, slightly affected size of sampled fields and age of the target trees were
and non-affected areas, which were then ranked into four first recorded by consulting plot holders. Local slope
soil loss severity classes and mapped by the FGs and the gradients were determined using clinometers. Soil types
researcher (Figure 2). For each site, presence of tree root and their depths, and availability of any erosion control
exposures, gullies, rock-outcrops, soil color and soil practices were recorded through observation and field
depth were assessed by a joint transect walk of the measurements.
researcher and FG members. Soil depth was measured
from gully cuts and by digging using local tools. Soil
texture was determined by finger-feel method in situ.
Following mapping and ranking, 31 plots (17 in Gayta                                                             
and 14 in Dubi) were selected on purpose to represent 1
Poor: households owning < 1 ha of land, do not own
three broadly defined household income (wealth) groups any or have only one ox and do not produce sufficient
(HIGs), namely poor, medium and rich households as food for themselves annually. Medium: households who
stratified by the FGs by using landholding sizes, oxen own 1-2 ha of land, two or three oxen and produce
sufficient food for one year. Rich: households with >2
ha of land, own more than three oxen and produce
surplus annually.
3 

 
 Figure.2. Sketch map prepared by FGs and the researcher, May 2010.

Next to recording of the surrounding geographical plot. The size of each measured plot was about 0.25 ha.
features, a close observation was made on tree root Of the total 108 measured trees, 37, 42 and 29 belonged
morphology to detect any changes in color, texture and to poor, medium and rich households, respectively.
structure of the tree stem based on methods cited in Average rates of annual soil loss were computed for
Dunne et al. (1979). Then heights (length from the upper each tree by dividing the measured exposed root height
part of the root to the present ground level) of exposed to its respective age as suggested by Stocking and
roots were measured using ruler following methods Murnaghan (2001). This has provided the estimate of
suggested in Stocking and Murnaghan (2001). Based on lost soil in terms of millimeters per annum for each
the measurements, data were gathered from 108 trees of measured tree. Average erosion rate for all trees was
different species, namely, croton (Croton then calculated by adding the values of all measured
macrostachyus), acacia (Acacia lahai), eucalyptus trees and dividing the sum to their total number. A one-
(Eucalyptus camaldulensis), albizia (Albizia gummifera), way ANOVA was employed to examine the mean
terminalia (Terminalia brownie) and sesbania (Sesbania difference in erosion, root exposure and tree age among
sesban) from May to October 2010. The minimum the 31 farm plots. A two-way ANOVA was used to see
number of trees observed per plot was two and the the mean difference in soil loss among the study sites,
maximum was five. On average, about three trees far slope categories, soil types and farm fields of the
apart at least 8 to 10 meters were assessed from each different HIGs.

RESULTS AND DISCUSSION and 16.6 years, respectively. Exposed root height ranged
from 2 mm to 61 mm, while tree age varied between 3
Erosion assessment and 38 years. From these measurements, on average 1.26
Table 1 presents summary statistics about tree age, root mm of soil depth or 16 tons of soil per hectare has been
exposure and soil loss for the measured farms. Average lost annually (Table 2). This rate is less than the
exposed root height and mean tree age were 20.7 mm maximum tolerable erosion (which is 18 t ha-1 year-1) but
over the minimum rate (i.e. 2 t ha-1 year-1) cited in
Berhane and Mekonnen (2009). The figure is much Table 2. Rates of soil erosion by study site and village
higher if compared to the 6 t ha-1 year-1 tolerance erosion RKAs Mean Std. Range
rate predicted by Hurni (1983) for areas located 2000- &Villages mm (t ha-1 mm year-1
2500 m a.s.l in the Simen Mountains of Ethiopia. -1 -1 * mm year-1
year year )
Conversely, it is much lower when compared with the
Gayta
average erosion rates estimated in northwest Ethiopia,
Giorgis 1.1 14 0.41015 0.57-1.93
18-79 t ha-1 year-1 in Bewket and Sterk (2003) and 93 t
Ashina 1.52 20 0.23245 1.15-1.83
ha-1 year-1 in Bewket and Teferi (2009) in the Chemoga
Gishen 1.51 20 0.36353 0.83-1.94
watershed as well as 27 t ha-1 year-1 in Tibebu et al
Selassie 0.55 7.2 0.36472 0.13-1.25
(2010) near Lake Tana. But, it can be taken as modest in
Total 1.15 15 0.51448 0.13-1.94
the context of a moderate erosion rate of 16-50 t ha-1
year-1 mentioned in Desta et al (2000) for the western Dubi
zones of Amhara region, Ethiopia. Soil loss rates ranged one 0.21 3 0.08348 0.12-0.43
from 0.12 mm to 2.8 mm year-1 and significantly Two 1.94 25 0.66832 0.77-2.80
differed across the measured fields as shown by the Three 1.72 22 0.69851 0.50-2.71
results of the one-way ANOVA (Table 1). Four 1.51 20 0.45639 0.67-2.14
Total 1.37 18 0.86240 0.12-2.80
Table 1. Summary statistics of tree age, root exposure and Grand total 1.26 16 0.71498 0.12-2.80
* -1 -1 -1
soil loss and results of one-way ANOVA mm year values converted to t ha year based on1 mm
Statistics Variables =13 t ha-1 suggested in Stocking and Murnaghan (2001)
Tree age Root exposure Soil loss considering a bulk density of 1.3 for recently cultivated
(years) (mm) (mm year-1) tropical soils mentioned in Fitzpatrick (1992).
N 108 108 108
Mean 16.56 20.72 1.26 Soil erosion variations across the study sites
Std. 6.71 14.59 0.71 Villages Two, Three and Four in Dubi RKA, and
Minimum 3.00 2.00 0.12 Ashina and Gishen in Gayta RKA face the largest soil
Maximum 38.00 61.00 2.80 removals (Table 2). These villages all lie on degraded
CV* 0.41 0.70 0.57 parts of the two communities. They are dominated by
F 2.37 3.56 6.71 shallow red soils and steep slopes. The highest erosion
P 0.001 0.000 0.000 rates occurred in areas characterized by steep
topography, and coarse and shallow soils. The lowest
*Coefficient of variation
erosion rates were observed as expected in what were
Bodoque et al. (2005) estimated mean annual soil loss classified by the FGs as non-affected villages (village
rate as ranging between 1.1 to 2.6 mm in their study in One in Dubi and Sellasie in Gayta). The results in
Central Spain. These rates are almost nearer to the rates general indicate that soil erosion differs at micro-levels.
recorded in this research (0.12-2.80 mm year-1). The Results of the two-way ANOVA (Table 5) show
slight differences may be caused by variations in local significant mean erosion difference across the villages at
climate, geographic location and the methodology used. P < 0.0001. The variations are expected as the villages
Ethiopia is a tropical country while Spain is in the differ in slope gradient, soil type and overall extent of
temperate zone. Trail was the base of measurement in land degradation and the results matched with the ranks
Bodoque et al. (2005) and they used ring dating methods given by the FGs.
while this study relied on recording age of trees by Soil loss variations by slope and soil type
consulting tree owners on farm plots. Natural erosion Table 3 summarizes the proportion of measured farms
rates (1.24-1.90 mm year-1) cited in Krusic (1990) for and soil removals at the various slope categories and soil
the Hiking trails, New Hampshire (USA) is not much far types. The results of the two-way ANOVA (Table 5)
from the findings of this research. But, their trail erosion indicate that there are significant soil loss differences
rates ranging 6.6-7.5 mm year-1 is much greater from the between the three slope categories (at P < 0.0001).
natural rate cited in their paper and from the results Sloping areas that comprised 36% of the measured farms
recorded in this paper. Such variations are expected faced the largest soil removal (52% of the total soil loss).
because the natural rates are resulted from purely natural In this area mean soil loss has been about 1.77 mm year-
geological factors and the rates cited in this paper are
initiated by tillage and natural factors.

5 

 
1
. Soils here are coarse textured and shallow in depth. same soil at Hunde-Lafto, southeast Ethiopia. The
Hence, slope and soil properties have contributed to the reason for the mentioned differences may probably be
high rate of soil erosion by water. Average soil loss was the methodology applied and the topographical and
moderate (about 1.2 mm year-1) on gently sloping mid- climatic variations of the research areas. It is obvious
stream areas where 45% of measured farms are located. that this research was based on field survey while those
These areas accounted for about 35% of the total soil cited in Herweg and Ludi (1999) were based on test plot
loss. experiment. With regard to soil type, higher mean
Erosion rates range from 1.7-36.4 t ha-1 year-1 on red erosion rates were recorded on farms with red soils
soil areas, 1.7-35.8 t ha-1 year-1 over brown soils and 1.6- where 55% of the total farms are found (Table 3). These
25.2 t ha-1 year-1 at areas covered with black soils (Table soils are coarse and shallow occupying steep slopes.
4). These results are by far very low compared to 17-176 They are intensively cultivated for long time and are
t ha-1 year-1 for red and 59-167 t ha-1 year-1 for brown significantly affected by erosion. Black soils that make
soils (luvisols) cited in Herweg and Ludi (1999) in up 16% of the measured farms contributed to some 23%
Anjeni, northwest Ethiopia. The results for the black of the total soil loss. These soils face low erosion rates
soils; however, indicate larger rates as compared to 0-16 as they occur in level topography.
t ha-1 year-1 recorded in (Herweg and Ludi, 1999) for the

Table 3. Proportion of measured farms and soil loss by slope, soil type and HIGs
Variables Measured farms (%) Total soil loss (%)
Gayta Dubi Total Gayta Dubi Total
Slope Plain 18 21 19 22 6 13
Gentle 59 29 45 34 38 35
Steep 23 50 36 44 56 52
Soil types Red 65 43 55 44 43 43
Black 12 21 16 36 15 23
Brown 23 36 29 20 42 34
HIGs poor 35.3 35.7 35.5 27 35 32
Medium 35.3 35.7 35.5 38 31 34
Rich 29.4 28.6 29.0 35 34 34

Soil loss variations by household income groups

Soil erosion rates were compared by the income status
of the households cultivating the sample fields. It was Table 4. Soil loss rates by slope, soil type and HIGs
found that in the plain areas, farms cultivated by the Variables Mean Std. Range
poor households suffered more than those cultivated by mm t ha-1 mm mm year-1
the other income groups. On gently sloping lands, both -1 -1 -1
year year year
the poor and medium household farms faced large soil Slope
erosion rates of about 1.2 mm year-1. In steep slopes,
Plain 0.44 6 0.43816 0.12-1.90
farms cultivated by the rich households suffered the
Gently sloping 1.20 16 0.53190 0.13-2.47
most . But, the results of the two-way ANVOA (Table 5)
Steep 1.77 23 0.60329 0.50-2.80
show statistically non-sgnificant differences in soil loss
Soil type/depth
by the household income groups. Household income is
thus not an important factor in determing soil erosion Red/shallow 1.46 19 0.56556 0.13-2.80
rates from individual farm plots. Black/deep 0.80 10 0.58037 0.12-1.94
Brown/moderate 1.17 15 0.85004 0.13-2.75
Soil and water conservation practices
Farmers in the study areas use different soil and water HIGs
conservation measures. These include contour farming, Poor 1.19 15 0.82740 0.13-2.71
terracing, check dams, water ways, cutoff drains, grass Medium 1.29 17 0.63780 0.12-2.75
strips, and planting trees. Rich 1.30 17 0.68418 0.13-2.80
Table 5. Results of the two-way ANOVA northern Ethiopia. They are important measures in steep
Soil loss (mm Exposed root slope areas. But, farmers in the study villages complain
Grouping
year-1) length (mm) that terraces and bunds take part of their plots and host
Variables F Sig. F Sig. pests that attack crops. Such complaints were also
Study site 0.001 0.973 7.37 0.008 reported in Herweg (1992). But, Vancampenhout et al.
Slope 32.88 0.000 14.17 0.000 (2006) reported that yields increased by 7% through
Interactions 4.77 0.010 1.09 0.340 application of stone bunds and no yield reduction
Study site 1.15 0.287 1.72 0.192 occurred due to 8% land occupied by stone bunds.
Soil type 8.25 0.000 7.38 0.001 Water ways and cutoff-drains
Interactions 7.83 0.001 7.51 0.001 Water ways (locally known as feses) are commonly
Village 21.99 0.000 9.042 0.000 constructed diagonally from the top to the bottom edge
HIGs 0.242 0.786 0.340 0.713 of a farm to drain out excess water. Cutoff drain, locally
Interactions 2.099 0.020 1.544 0.113 called tiras boy is however built horizontally at the head
Slope 36.6 0.000 11.24 0.000 of farms to collect and channel upcoming runoff. Water
HIGs 0.254 0.776 0.301 0.740 ways were observed in 53% of the farms in Gayta and in
Interactions 0.223 0.925 0.513 0.726 50% farms in Dubi. Cutoff drains were observed in 47%
of the farms in Gayta and in 36% of the fields in Dubi.
Using water ways is common in level and water-lodged
Contour plowing farms. In total, over 50% of the measured farms were
Contour plowing is a very common traditional practice observed having water ways, while cutoff drains were
elsewhere in Ethiopia. It was expectedly used by all of found in only 42% of the fields. As they are easy to
the farmers sampled for this study (data not shown here). construct using a plough during tillage operations, they
Plowing is the practice of land preparation before are exercised by many farmers.
sowing to make the soil more porous to water seepage,
minimize weed emergence and improve land Table 6. Proportion of farms with SWC structures by study
productivity (Abera and Belachew, 2011). In the study site (%)
villages, farm lands are commonly tilled three to seven SWC practice Gayta Dubi Total
times depending on the requirement of the specific crop. Soil terraces 29 36 32
Millet and maize fields are often ploughed three to five Stone bunds 12 29 19
times while tef fields are tilled four to seven times. As Grass strips 41 36 39
Bewket (2003) notes however repeated tillage and fine Ditches/water ways 53 50 52
seed-bed formation also prepares the soil to more Cutoff drains 47 36 42
erosion. Nyssen et al. (2000) remarked that traditional Trees 100 100 100
contour farming practices encourage gradual down-slope No of farms 17 14 31
soil translocation and material accumulation in the lower
farm margins. Therefore, there is the need to assess the
Grasses and trees
benefits and adverse effects of the traditional practice of Grasses and trees are planted to serve numerous
repeated tillage. purposes such as: control of soil erosion, source of
Terraces and bunds animal feed, fuel and construction material. They are
Terraces and bunds were applied in 32% and 19% of all also sold in towns to generate income. Almost all of the
the farms, respectively (Table 6). In Dubi, 36% of the measured farms contain trees and 39% of them contain
farms had soil terraces and 29% contained stone bunds. grass strips (Table 6). Trees were found at the centers as
The proportion of these SWC structures is less in Gayta well as at the margins of the fields. Grasses commonly
compared to Dubi. Soil terraces were seen on 29% and occupied farm boundaries and planted on strips across
stone bunds on only 12% of the assessed farms in Gayta. the centers of few fields. Both grasses and trees were
These SWC structures were generally applied on steep adopted by many farmers because they require less labor
slope farms. They were first introduced to the area in the and provide diverse benefits. Grass and tree barriers in
1980’s. Soil, stone, wood or a combination of both are general deter the speed of running water and improve
used in construction of terraces and bunds. soil stability and sediment deposition.

Both terraces and bunds are important to cultivate grass

CONCLUSIONS
but they require large labor to build. Vancampenhout et The objective of this study was to assess the extent of
al. (2006) for instance indicated spaces occupied by soil erosion using tree root exposure measurements and
stone bunds were used to grow fodder grasses in Tigray,
7 

 
examine farmers’ SWC practices in the northwestern Since trees are familiar objects on farm plots farmers can
part of Ethiopia. The results show that the annual easily and quickly measure exposed roots and assess the
average rate of soil loss in the study area is about 16 tons amount of soil removed by dividing the measured
per hectare or a soil depth equivalent of 1.26 mm. Soil lengths to the tree ages. This farmer participatory
loss rates significantly differed within farm fields, research approach is thus practical, simple to implement
villages, soil types and slope categories (at P < 0.001). in local contexts and help to reverse the long existing
This implies that SWC and land management practices “scientist-to-farmer” to a “farmer-to-farmer” approach.
should be designed and implemented considering land The participatory survey methodology, therefore, can be
slope, soil type and topographic location of farms. Each taken as a good alternative to complex models to
particular site thus requires specific SWC measure that estimate soil erosion rates at the farm level, where other
fits its unique characteristics. Contour farming, ditch more reliable measurements are unavailable.
construction, grass and tree planting were observed most
widely practiced SWC structures by the local farmers. ACKNOWLEDGEMENTS
Future SWC and land management should focus on We thank the farmers who owned the surveyed farms
integrated interventions that involve biological and and the participant farmers (FGs) for their kind and
structural measures and participation of land users. enthusiastic cooperation for the study. We also thank
The results obtained in this participatory research experts at the district office of agriculture and rural
indicated that the “farmer-perspective” framework is a development for providing some useful information to
good mechanism to estimate soil erosion rates on the study. The study was financially supported by the
cultivated fields by simply measuring exposed roots. School of Graduate Studies of Addis Ababa University.

REFERENCES Constable, M. 1984. The degradation of resources and

Abera, Y. and Belachew, T. 2011. Local perceptions of an evaluation of actions to combat it.
soil fertility management in southeastern Ethiopian Highlands Reclamation Study
Ethiopia. International Research Journal of (EHRS) Rural Development Strategy. Draft of
Agricultural Science, 1:064-069. Part II. Ministry of Agriculture/FAO, Addis
Berhane, G.and Mekonnon, K. 2009. Estimating soil Ababa.
loss using Universal Soil Loss Equation Desta, L., Kassie, M, Benin, S. and Pender, J. 2000.
(USLE) for soil conservation planning at Land degradation and strategies for
Madego watershed, northern Ethiopia. Journal sustainable development in the Ethiopian
of American Science, 5:58-69. highlands: Amhara Region, Socio-economic
Bewket, W. 2003. Towards integrated watershed and Policy Research Working Paper 32. ILRI
management in highland Ethiopia: The (International Livestock Research Institute),
Chemoga watershed case study. PhD Thesis, Nairobi, Kenya. 122 pp.
Wageningen University, The Netherlands. Diao, X. 2010. Economic importance of agriculture for
Bewket, W., Sterk, G. 2003. Assessment of soil erosion sustainable development and poverty
in cultivated fields using a survey reduction: The case study of Ethiopia. Global
methodology for rills in the Chemoga Forum on Agriculture, 29-30 Nov. 2010,
watershed, Ethiopia. Agriculture, Ecosystems OECD headquarters, Paris.
and Environment, 97:81-93. Dunne, T., Dietrich, W. E. and Brunengo, M. J.1979.
Bewket, W. and Teferi, E. 2009. Assessment of soil Rapid evaluation of soil erosion and soil life-
erosion hazard and prioritization for treatment span in the grazing lands of Kenya.
at the watershed level: Case study in the Proceedings of the Canberra Symposium, Dec.
Chemoga watershed, Blue-Nile basin,
1979, IAHS-AISH Publ. No.128.
Ethiopia. Land Degradation and Development,
20:609-622. http://iahs.info/redbooks/a128/iahs_128_0421.
Bodoque, J. M., Diez-Herrero, A., Martin-Duque, J. F., pdf, accessed Sept. 2010.
Rubiales, J. M., Godfrey, A., Pedraza, J., Fitzpatrick, E.A. 1992. An Introduction to soil science.
Carrasco, R.M., and Sanz, M.A. 2005. Sheet 2nd ed. John Wiley and sons, Inc. New York.
erosion rates determined by using Herweg, K., 1992. A Survey method for soil erosion
dendrogeomorphological analysis of exposed assessment and conservation control, In:
tree roots: Two examples from Central Spain. Hurni, H., Tato, K. (Ed.), Erosion,
Catena, 64:81–102. Conservation and Small-Scale Farming.
ISCO/WASWC. pp. 1-13.
8 

 
Herweg, K. and Ludi, E. 1999. The Performance of Nyssen, J., Poesen, J., Gebremichael, D.,
selected soil and water conservation measures. Vancampenhout, K., Daes, M., Yihdego, G.,
Case studies from Ethiopia and Eritrea. Govers, G., Leirs, H., Moeyersons, J., Nudts,
Catena, 36: 99-114. J., Haregewein, N., Haile, M. and Deckers, J.
Hurni, H.1983. Soil erosion and soil formation in 2007. Interdisciplinary on-site evaluation of
agricultural ecosystems: Ethiopia and stone bunds to control soil erosion on crop
northern Thailand. Mountain Research and land in northern Ethiopia. Soil and Tillage
Development, 3:131-142. Research, 94:151-163.
Krusic, P. 1990. Soil erosion and tree root response Stockign, M. and Murnaghan, N. 2001. Hand book for
along hiking trails in the White Mountains of the field assessment of land degradation.
New Hampshire: A study performed during Earthscan publications Ltd. London.
the first North American dendroecological Tibebu, T.Y., Abiy, A. Z., Zegeye, A. D., Dahlke, H. E.,
fieldweek. Pinkham Notch Camp, Gorham, Easton, Z. M., Tilahun, S. A., Collick, A. S.,
New Hampshire, Aug. 12-19, 1990. Kidanu, S., Moges, S., Dadgeri, F. and Steenhuis,
Nyssen, J., Poesen, J., Haile, M., Moeyersons, J. and T. S. 2010. Surface and Sub-surface flow effect on
Deckers, J. 2000. Tillage erosion on slopes permanent gully formation and upland erosion
with soil conservation structures in the near Lake Tana in the northern highlands of
Ethiopian highlands. Soil and Tillage Ethiopia. Hydrology and Earth System Sciences,
Research, 57:115-127. 14:2207-2217.
Okoba, B.O. and Sterk, G. 2006. Farmers’ identification Vancampenhout , K., Nyssen, J., Gebremichael, D., Deckers,
of erosion indicators and related erosion J., Poesen, J., Haile, M. and Moeyersons, J., 2006.
damage in the central highlands of Kenya. Stone bunds for soil conservation in the northern
Catena, 65:292-301. Ethiopian highlands: Impacts on soil fertility and
crop yield. Soil and Tillage Research, 90:1-15.

9