SINET:Ethiop. J. Sci.

, 46(1): 1-16, 2023 ISSN: 0379–2897 (PRINT)

© College of Natural and Computational Sciences, Addis Ababa University, 2023 eISSN: 2520–7997

Date received: August 30, 2022; Date revised: March 23, 2023; Date accepted: April 01, 2023
DOI: https://dx.doi.org/10.4314/sinet.v46i1.1

Soil Erosion Estimation Using Revised Universal Soil Loss Equation Integrated with
Geographic Information System by Different Resolution Digital Elevation Model Data in
Weyto Sub-Basin, Southern Ethiopia

Shimeles Damene *1 and Poshendra Satyal 2

1 Center of Environment and Sustainable Development, College of Development Studies, Addis

Ababa University, Addis Ababa, Ethiopia. E-mail: shimelesdamene@yahoo.com,
shimeles.damene@aau.edu.et
2 Science, Policy and Information Division, BirdLife International, Cambridge, UK. E-mail:

poshendra.satyal@birdlife.org

ABSTRACT: Soil erosion is a global environmental challenge for developing countries including
Ethiopia that require regular monitoring to take corrective measures. In this context, this study was
focused on estimating soil erosion using the Revised Universal Soil Loss Equation ( RUSLE)
integrated with Geographical Information System (GIS) technique for which it applied 30 m and 200
m resolution Digital Elevation Model (DEM) data to generate slope gradient and length. Rainfall
erosivity, soil erodibility, land cover/use and management factors data were obtained from
existing studies and field-based assessments where the data were used to estimate the soil erosion
using RUSLE model in ArcMap under two different DEM resolution scenario. The model estimated
an average of 1.38 and 1.86 million tons of annual soil loss by water using 200 and 30 meters
resolution DEM data, respectively, while keeping other factors constant. The erosion estimated
using higher (30 m) resolution DEM data was more realistic than low (200 m) resolution data , as the
higher resolution DEM data allowed less generalization. In high resolution DEM data, the slopes
generated were also more in line with ground reality. Based on the case study of Weyto sub-basin
in Southern Ethiopia, we thus conclude that the GIS technique and remote sensing data can be used
in RUSLE based erosion risk prediction for large areas even at basin, sub-basin and macro watershed
level. We suggest that the accuracy of the prediction can be improved by using high resolution
(large scale) input data disaggregated by micro- and sub-watersheds.

Keywords/Phrases: Ethiopia, Geographical Information System, Land degradation, Modelling,

Revised Universal Soil Loss Equation, Soil Erosion

topsoil from crop lands of the Ethiopian

INTRODUCTION highlands. Various studies in Ethiopia indicated
that most farmlands are severely affected by
Nearly one-third of the global land used for water erosion, particularly in the highlands (e.g.,
agriculture has been affected by soil degradation Hurni, 1993; Bojö and Cassells, 1995; Lulseged
in the historic past, where most of this damage Tamene et al., 2006; Shimeles Damene et al.,
was caused by water and wind erosion (Nana- 2012; Habtamu Sewnet and Amare Sewnet, 2016;
Sinkam, 1995; Scherr, 1999; Hurni et al., 2008). Balabathina et al., 2019; Nyssen et al., 2019;
The problem is persisting in developing Atoma et al., 2020; Yared Mesfin et al., 2020). The
countries mainly in Sub-Sahara African (SSA) soil erosion induced land degradation imposed
nations. The drivers of soil erosion by water in chronic food insecurity across Ethiopia
the subcontinent are attributed to natural (e.g. highlands (Bekele Shiferaw and Holden, 1999;
topographic features) and manmade (e.g. land Kebrom Tekle, 1999), resulting in economic
use pattern) causes (Nana-Sinkam, 1995; Scherr, losses of about US$ 106 million annually at
1999; FAO, 2004; Vlek et al., 2008). Like most SSA national level (Bojö and Cassels, 1995) and also
nations, soil erosion induced land degradation in caused various environmental hazards including
Ethiopia is exceedingly high, where the problem recurrent drought (Bekele Shiferaw and Holden,
is acute in tits highlands (Hurni, 1993; Bekele 1999; Kebrom Tekle, 1999). Consequently, for
Shiferaw and Holden, 1999; Shimeles Damene et combating the land degradation in the
al., 2013; 2020). For instance, Hurni (1993) highlands, various efforts have been put in place
estimated 42 ton per ha annual losses of fertile by Ethiopia’s Ministry of Agriculture over the
_____________________
*Author to whom correspondence should be addressed.
2 Shimeles Damene and Poshendra Satyal

past four decades, mainly on cultivated lands. Renard et al., 1997). In this regard, Hurni (1985)
These interventions on farmlands include developed reference criteria for the different
construction of physical measures like contour erosion factors to estimate soil erosion under
bund, hillside terraces, check dams and Ethiopian condition. This days researchers have
biological measures such as tree planting at been integrating RUSLE model with the
homestead and on farms, afforestation and Geographic Information System (GIS) and
closure of degraded lands for self-restoration Remote Sensing (RS) techniques to estimate the
(Shimeles Damene et al., 2013; Hurni et al., 2016). soil erosion by water at affordable cost and time
Understanding and mitigating erosion coupled (Fu et al., 2005; Kouli et al., 2009; Demirci and
with the associated land and environmental Karaburun, 2012; Habtamu Sewnet and Amare
degradation is critical because of its possible Sewnet , 2016; Ganasri and Ramesh, 2016; Ayele
adverse effects, such as loss of nutrients, river Desalegn et al., 2018; Bouhadeb et al., 2018;
and reservoir siltation, water quality Asnake Yimam and Amare Bantider 2019;
degradation, and decreases in land productivity López-García et al., 2020; Yared Mesfin et al.,
(Bagherzadeh, 2014). Moreover, in connection to 2020).. The model is also widely used as flexible
the construction of small to large-scale irrigation tool (Kumar and Kushwaha, 2013;
dams and huge hydropower generating dams Panditharathne et al, 2019) that has been adapted
(such as the Great Renaissance, Gilgel Gibe I –IV, to landscape and watershed scales, combined
Koysha, Melka Wakena and Tana Beles), the with GIS and RS techniques (Panditharathne et al,
country has been striving its best to restore the 2019). Nonetheless, most studies do not compare
environment and protecting dams from siltation the RS data at different resolutions on the
through implementing integrated watershed precision of soil erosion prediction as the model
management (IWSM) e.g soil and water has been accused of having error in estimation of
conservation (SWC) interventions and planting sediment redistribution in small drainage units
billions of tree seedlings. Although the country before the runoff heading to large drainage
has been investing considerable finance and system (Nearing, 1997; Cohen et al., 2005;
mobilizing local people in IWSM, SWC and tree Kinnell, 2005; Nearing, et al., 2005; Panagos et al.,
planting, the result of the intervention is not 2015b). In this regard, high resolution RS data
supported by regular monitoring and often lacks such as digital elevation model (DEM) from
a robust tool and methodology to track changes which the topographic factor (slope length and
and understand possible impacts. Despite the gradient) is calculated, could help to divide the
various efforts of the government in protecting watersheds into small compartments and thus
land from soil erosion induced land degradation, might reduce the effect of aggregated estimation
very limited studies have been conducted yet to errors. It is well known that the topographic
test and identify the best input data use in factors (including slope length and gradient) is
estimating soil loss in Ethiopia. the most sensitive in soil loss prediction
The universal soil loss equation (USLE) modelling (Panditharathne et al, 2019).
developed by Wischmeier and Smith (1965; 1978) Therefore, this study is aimed at evaluating the
and modified by Renard et al., (1997) that come RUSLE model using 30 m and 200 m resolutions
as the revised universal soil loss equation (RUSLE) DEM data to demonstrate the difference in the
which enhanced the prediction power of the erosion estimation prediction power. For the
model for erosion estimation. Thus RUSLE purpose, we have used the case study of Weyto
become a comprehensive mathematical model sub-basin in Southern Ethiopia.
that uses different soil erosion factors to estimate
soil loss although it still has same limitations.
Despite its persisting limitations, the RUSLE MATERIALS AND METHODS
model is still in use to estimate soil erosion by
water, particularly at a watershed level (Panagos Description of Weyto sub-basin
et al., 2015a; Phinzi and Ngetar, 2019; Almaw Weyto sub-basin is found in the Southern
Fenta et al., 2020). RUSLE parameters can be Nations, Nationalities, and Peoples' Region
developed based on small-scale studies of (SNNPR) of Ethiopia (Figure 1). Geographically,
agricultural plots (Benavidez et al., 2018), using the sub-basin is located between 5°23´00´´ and
this model in large-scale conditions can prove 6°15´00´´ North latitude and 36°35´00´´ and
extremely different from the small agricultural 37°25´00´´ East longitude and covers a total area
plot conditions, and hence the model may lead to of 438,384 ha.
error extrapolation (Wischmeier and Smith, 1978;
SINET: Ethiop. J. Sci.,46(1), 2023 3

Figure 1. Location map of Weyto sub-basin.

The sub-basin is characterized by a wide range (7.6%) and very few (<1%) Arenosols (Halcrow
of biophysical features in terms of climate, agro- and GIRD, 2007).
ecology, soil, geology, land use/cover, drainage The sub basin has a bimodal rainfall pattern
pattern and density. The study area has diverse that annually varies from 678 mm at the Weyto
landform and geology. The topography has meteorological station (at 570 metre a.s.l.
various characteristics such as plain, valley elevation.) to 2,107 mm at the Gerese station (at
plains, plateaus, ridges, hills, medium and high an elevation of 2,329 m.a.s.l) (Figure 2). Nearly
mountains. The sub-basin consists of 17 major- 70% of the rainfall occurs in the first (March to
and sub-watersheds (Figure 1). The watersheds May) and second (September to November)
in the highland areas (e.g., Tsfitso-Sosa, Dencha, rainy seasons which contributes about 24% to
Uba Shafa and Upper Bezo) have narrow and 50% and 19% to 35% of annual rainfall
steep landforms. In contrast, watersheds located respectively. The mean annual temperature
in the lowland areas (e.g. upper and lower varies from 16°C in the highlands to 28°C in the
Weyto, Lower Bezo) are relatively wider in size lowlands. As shown in Figure 2, the mean
and characterized by gentle slope. The geology is monthly minimum temperature ranges from
predominantly volcanic and sedimentary 10.2°C (at 2,280 m a.s.l.) to 23.6°C (at 1,158 m
formation such as unwelded pumiceus a.s.l.) (). The mean monthly maximum
pyroclastic, ignimbrite, tuff, water lain temperature reaches over 34°C in the lowlands
pyroclastic and undivided alluvial fluvitile and between December and March and the
lacustrine sediments; and meta sedimentary temperature sometime rises as high as 40°C.
genesis originated from biotite, quartz, feldspar, According to the local classification system
gneiss, granite rhyolite and trachyte (Halcrow (Hurni, 1998), the sub-basin lies within four
and GIRD, 2007). The diverse geology and agro-ecological zones namely: kolla (warm),
geomorphology have resulted in a variety of soil weyna dega (mild), dega (cool) and wurch (cold)
types, where the major soil types are Cambisols that cover 63%, 26%, 10.5% and 0.5% area,
(66.4%), Nitisols (15%), Luvisols (10%), Vertisols respectively.
4 Shimeles Damene and Poshendra Satyal

Figure 2. Annual rainfall and temperature records by meteorological stations in and around Weyto sub-basin.

Subsistence aagriculture is the main economy steep landscapes, patches of grasslands and Enset
activityof the local people. The majority (70%) of altogether have been reducing runoff velocity
people depend on mixed crop and livestock and soil erosion rate.
production (sedentary farming agriculturalists) The watersheds in the south-eastern parts
and the remaining (30%) were agro-pastoralists particularly in Konso areas (e.g., Keseba and
and pastoralists. The sedentary farming Keselte watersheds) are degraded as they have
agriculturalist communities inhabited the low vegetation cover and are heavily cultivated.
highland and midland areas. Crop production is However, farmers of the areas are actively
diversified and includes a variety of crops: engaged in terrace construction to protect the
cereals, fruits, vegetables, root crops, Enset farmlands from soil erosion. Watersheds in the
(Ensete ventricosum), cash crops, particularly lowland areas and valley plains (upper and
spices (e.g., black cardamom, ginger) and coffee. lower Weyto) are characterized by flat slope and
The lowland areas are predominantly occupied covered by dense bush, woodlands and forests.
by pastoralist and agro-pastoralist communities. Thorny, deciduous lowland trees (mainly Acacia
The watersheds have different land use/cover species) are dominant vegetation with some
characteristics. Most watersheds in the highland broad-leaved trees (like Ficus and Palm trees) as
and mountainous areas are characterized by riparian vegetation along the Weyto River and
agro-forestry based crop cultivation. In these its major tributaries. The lowlands have very
areas, farmers mostly plant perennial crops, sparse population who depend on pastoralist
particularly Enset, integrated with cereals (like and agro-pastoralist livelihood systems, mainly
barley, wheat), legumes, root crops and producing grazing and browsing animals
vegetable-based farming, even on the steep (Figure. 3c & d). The indigenous pastoralist
landscapes. These watersheds also have patchy communities to some extent are also involved in
grasslands, scattered and small grove of trees are hunting of wild pigs. In these areas, some settlers
found on farmlands at the homesteads, along the who come from neighbouring districts are found
river courses and on very steep landscapes to be practicing crop production through shifting
(Figure. 3a & b). The trees along the river sides, cultivation by clearing the natural vegetation.
SINET: Ethiop. J. Sci.,46(1), 2023 5

(a)
(b)

(c) (d)

Figure 3. Major land use: Enset (a) and cereal (b) based farming systems in the highlands and land covers: forests (c),
papyrus grass stripe (d) in Weyto sub-basin.

Methods arguably soil erosion modelling in general, stem

Application of GIS for soil erosion estimation from several factors, including: the inability of
models to capture the complex interactions
The RUSLE has now been one of the widely
involved in the soil loss, the low availability of
used soil erosion model to predict soil erosion by
long-term reliable data for modelling, and the
water. However, the model has been questioned
lack of soil erosion observational data for model
for its accuracy, as it has various limitations in
validation, especially in data-scarce
output quality which is determined by input
environments (Benavidez et al., 2018). Therefore,
data and where some of the input data involve
in order to enhance the model erosion prediction
expert judgment and hence are subject to
power our analysis combined RUSLE with
personal bias and error. The model has less
Geographic Information System (GIS) techniques
sensitivity to rainfall than land cover (Nearing,
and Remote Sensing (RS) data. Moreover, we also
1997). Also, it does not consider the effects of
applied Hurni (1985) indexes for the different
gully erosion and redistribution of eroded soil
soil erosion factors to fit the RUSLE model with
within small land units. Hence, the model is
Ethiopian condition. . Therefore, in our analysis,
mainly used to estimate annual soil loss over
we have used RUSLE model together with Hurni
long periods for smaller areas, as it lacks runoff
(1985)’s erosion factor indices by integrating with
factors and mostly gives exaggerated estimation
the GIS techniques and RS data, mainly the
(Kinnell, 2005). Moreover, it disregards ground
Digital Elevation Model (DEM with 30 and 200 m
conditions like plant litter, which are important
resolution) as part of the input data. GIS
in monitoring water infiltration and thereby soil
technique has been applied to generate and
erosion (Nearing, et al., 2005), and also does not
transform erosion factors, and perform the
estimate sediment pathway along hill slopes
modelling.
(Cohen et al., 2005). On the whole, the
In the process of analysis, the different input
uncertainties associated with the RUSLE, and
parameters of RUSLE: Rainfall erosivity (R), soil
6 Shimeles Damene and Poshendra Satyal

erodibility (K), slope gradient (S) and length (L)

treated as topographic (LS) factor, land cover (C) Determining rainfall erosivity (R )
and management (P) factors were generated. In Rainfall erosivity, R (MJ mm ha-1 h-1 yr-1) is the
order to simplify the analysis, land cover and product of the kinetic energy of rainfall (E) and
management factors were merged into one the maximum intensity of rain in 30 minutes in
before use in the model. As some of land cover cm hour-1 (I30). Rainfall intensity records are not
practices are part of land management factors, available in most rain gauge stations of Ethiopia.
we combined both while generating raster data Hence regression equation (Eq.2) developed by
using the tow maps (land cover and Hurni (1985) for Ethiopian condition was used to
management). Application of the GIS techniques determine R values for mean annual rainfalls
involved assigning the erosion factor by instead of I30. Therefore, 34 years (1980 to 2014)
inserting index value for spatial feature data. annual rainfall data of 10 stations (found in and
Then, the feature data (shape files) containing the around the sub-basin) combined with agro-
index values were converted into raster data ecological map (Hurni 1998) developed from
while also performing the vis-à-vis process as DEM were used to generate rainfall and erosivity
detailed in the subsequent sub-sections below. map of the sub-basin. In this process the data
Finally, all erosion factors data transformed into analysis was performed using Excel and the final
raster formats were used to model the erosion result transferred into raster map. After
risk using the RUSLE model (Eq.1) in the raster preparation of the erosivity map, the shape file
calculator tool of ArcMap 10.3.1. (feature data) was converted into raster data
format so as to make it ready as input data for
Eq. 1 RUSLE modelling.
where:
A = Total soil loss (t/ha/yr) Eq. 2
R = Rainfall erosivity factor
where, R is rainfall erosivity factor and P is mean
K = Soil erodibility factor
annual precipitation (mm)
LS = Topographic (L= Slope length and S= Slope gradient) factor
C = Land cover factor Soil erodibility factor (K)
P = management factor
Soil erodibility (K) factor is a measure of the
Data source and process of determining erosion susceptibility of soil particles to detachment and
factors transport by rain drop and runoff (USDA, 2011).
According to the models used in Wischmeier and
Data sources
Smith, (1965, 1978) and Renard et al. (1997), K
The modelling was done using 1:250,000 scale factor depends on various physicochemical
maps of soil erodibility, rainfall ersosivity, land properties of soils such as the organic matter
cover and management factors. On the other content, texture, permeability and structure that
hand, slope length (L) and gradient (S) factors can range from 0.7 for the most fragile soils to
aggregated as topographic (LS) factor were 0.01 for the most stable soils. Zhang et al. (2008a)
generated from 200 m and 30 m resolutions of estimated K values on 13 runoff plots of soils in
the Digital Elevation Model (DEM) data. The eastern China that ranged from 0.007 to 0.02 MJ
erosion factors data such as soil survey, land mm ha-1 h-1 yr-1. Nevertheless, there is as such no
use/cover were obtained from Halcrow and GIRD standard K index to be used under different soils
(2007), soil and water conservation data were and settings so far. Based on Wischmeier and
compiled from the records of district agriculture Smith (1965; 1978) method, Hurni (1985)
offices while the interpretation of satellite data developed a simplified K factor index based on
was assisted by high resolution [60 cm] Google soils colour, assuming that soil colour has
earth data and field observation. DEM data of intrinsic relation with the above soil properties
Shuttle Radar Topography Mission (SRTM) was for Ethiopian case. Hence the soil map of the
accessed from: https://earthexplorer. usgs.gov. sub-basin adopted from Halcrow and GIRD (2007)
Rainfall data of ten meteorological stations was used to produce soil colour map and
obtained from the Ethiopian Meteorological assigned with K values as per Hurni (1985) i.e. K
Agency (EMA) were used to generate rainfall value of 0.15, 0.20, 0.25 and 0.30 for black, brown,
erosivity factor map. These data were then used red and yellow colour soils respectively. The K
to generate different erosion factors. The details values were assigned on the feature data (shape
of the estimation methods are discussed below.
SINET: Ethiop. J. Sci.,46(1), 2023 7

file) and then transformed into raster data to raster data and made ready for RUSLE based
make it ready for RUSLE based modelling. calculation.

Land cover (C) and management practice (P) Topographic (LS) factor
factors Slope length (L) and gradient (S) are among
The land cover factor is based on clean tilled the factors that affect soil erosion, which together
and continuous fallow conditions as C represents are referred to as topographic (LS) factors
the effects of plants, soil cover, soil biomass, and (Krusekopf, 1943; Hurni 1985). In soil loss
soil disturbing activities on erosion. The P factor estimation, slope length determination is often a
represents the ratio of soil loss from lands treated complicated process and hence combined form
with soil conservation practices (such as specifically single value estimated by Hurni
contouring and/or strip-cropping) to that with (1985) for slope length and gradient is used
straight row farming up-and-down slope (1.00). (Bagegnehu Bekele and Yenealem Gemi, 2021).
In this analysis, the C and P factors were To overcome the challenge, unlike USLE, the
modified to Ethiopian condition as per the RUSLE takes a number of considerations. For
recommendation of Hurni (1985). Therefore, land example, the RUSLE considers runoff differences
use/cover (LULC) map developed by Halcrow over catchment such as runoff channelled into
and GIRD (2007) was used to generate C and P rills and gullies (as rill erosion is a major
factor data which were mapped in one. component in the RUSLE), soil saturation
Accordingly, 5 LULCs were identified and resulting from long duration rainfall that carries
provided with C values as adopted from Hurni more runoff and creates greater erosion, soil
(1985) i.e. forestland (0.001), woodlands (0.005), deposition at concave slope landform and also
bushland (0.01), grassland with fragmented takes into account of converging and diverging
farmlands/woodlands (0.05) and farmlands terrain (Renard et al., 1997). Therefore, Hurni
(0.0975 to 0.135 depending on the management (1985) recommended the combined LS
type). The land management practices and (topographic) factor, based on which this study
corresponding P values for farmlands include: also used the combined LS factor for various
terracing and agro-forestry supported slope gradients (%). In this exercise, two
smallholder farmlands (0.0975), agro-forestry resolutions (200 m and 30 m) DEM data were
based smallholder’s farmlands (0.105), enset used to determine the LS factor. The DEM data
based smallholder’s farmlands (0.12), were transformed into feature data (shape file) so
smallholder’s farmlands mixed with patchy as to assign erosion index values for ranges of
wood/bush lands (0.12) and counter cultivation slope gradients as given in Table 1 and
based smallholder farmlands (0.135). Finally, reclassified into new class as per Hurni (1985)
after C and P index values were assigned to each recommendation. Then, LS feature data (shape
mapping unit in the feature dataset (shape file) file) carrying erosion index values were
in ArcMap 10.3.1, the map was converted into transformed back into raster format to make it
ready for the RUSLE model-based calculation in
ArcMap 10.3.1.

Table 1. Slope gradient (%) and LS factor index.

Slope gradient (%) <2 4 6 8 13 25 40 55 100 >100

LS factor index 0.19 0.38 0.66 1.14 1.90 3.80 6.08 7.98 10.45 19.00

Sources: Hurni (1985) developed erosion factor indexes under Ethiopian condition; adapted from Wischmeier and Smith (1965; 1978)

Field data collection and map verification socio-economic situation, erosion extents and
Following interpretation and map production degree of severityy. The assessment was done
using the different remote sensing data, in situ throughout the delineated watershed areas.
field visit and data collection were carried out in Transect walks were made to check polygons
two rounds of wet and dry seasons. The field with unique characters/tones, erosion hotspots,
data collection involved observation and cross- unique land uses/land covers, in different
checking of interpreted information (polygon) agronomic, conservation (soil and water) and
from maps, images and DEM data to ascertain the land management practices. Geographical
actual ground condition on biophysical and Positioning System (GPS) was used to locate the
8 Shimeles Damene and Poshendra Satyal

polygons and features of the base map on the which are located in the central parts of the sub-
ground and vice versa. basin receive relatively low rainfall and thus
Moreover, focus group discussions (FGDs) were have low erosivity index and are characterized
held with selected community members (at 15 by flat and gently undulating landscapes.
localities) along the transect walk and key Therefore, the calculation considered this fact
informant interviews (KIIs) were also carried out supported by ground (rainfall) data while
with selected districts agriculture offices. In inputting index values for R.
order to collect relevant secondary data,
structured formats were developed and Soil erodibility factor (K):
distributed to districts and zones agriculture In general, as shown in Figure 4b, the major
offices in 2015. The primary and secondary field soil types of the sub-basin are: Cambisols
data as well as very high (0.6 m) resolution (63.2%), Nitisols (17.4%), Luvisols (11.1%), and
Google Earth map were used to verify the Vertisols (8.3%). Thus, nearly three fourth
produced maps and to assign erosion factor (71.5%) of the sub-basin soils possess slight
values of the various polygons of the input maps (8.3%) to moderate (8.3%) erodibility and the
used in RUSLE-based erosion modelling. remaining parts are characterized by relatively
high (0.18 to 0.2) to very high erodibility index
(Panagos et al., 2015a). Soils with high to very
RESULTS AND DISCUSSION high erodibility are Luvisols and Nitisols, which
are located in the central north and northeast
Erosion factors analysis results parts of the sub-basin. In contrast, areas with
Rainfall erosivity factor (R): slight to moderate erodibility are covered by
Rainfall erosivity values were estimated from Cambisols and Vertisols. In this regard,
mean annual rainfall data of 10 meteorological literatures suggest that loam and fine sand
stations found in and around the sub-basin and textured soils are the most erodible that have fine
agro-ecological map. As discussed earlier, the clay and loam particles, which can easily be
modelling used mean annual total precipitation. transported even under low runoff velocity and
Our analysis revealed that erosivity ranged from soils coarser than fine sand settle at short
373 to 1176 MJmm/ha/h/yr (Figure 4a). The distance from the venue of detachment
study sub-basin rainfall is expected to cause very (Wischmeier and Smith, 1965; 1978; Hurni 1985;
high to moderate erosivity in the mountainous Duiker et al., 2001; Zhang et al., 2008a).
areas. In convers, the lowlands and midlands

Figure 4. Maps of rainfall erosivity (a), soil erodibility (b) and land cover and management practice factors (c) for the Weyto
sub-basin.
SINET: Ethiop. J. Sci.,46(1), 2023 9

Land cover (C) and management practice (P) farming have a considerable role to reduce soil
factors: erosion, thus these land units have low erosion
As shown in Figure. 4c, about 43.9% of the sub- risk compared to farmlands without such
basin is covered by forestlands (2.7%), management practice. In this regard, Young
woodlands (15.8%), bushlands (21.2%) and (1989) underlined that agro-forestry system
grasslands with fragmented farmlands/ reduces soil erosion through maintaining soil
woodlands (4.1%). Hence, these land units have organic matter, improving soil chemical,
minimal soil erosion risk due to good land cover. biological and physical properties thereby
On the other hand, farmlands that cover about enhancing efficient nutrient recycling within
56.1% of the sub-basin have high soil erosion pedological system even from the substrata. On
risk, which are located in the northeast and the other hand, runoff generated from
northwest parts. Studies in highland of Ethiopia fragmented farmlands located within forest,
showed that most farmlands are severely wood- and bush-lands might not continue with
affected by water erosion, (e.g., Hurni, 1993; Bojö erosive velocity and the runoff will not have
and Cassells, 1995; Shimeles Damene et al., 2012; cumulated effect to cause significant erosion risk
Balabathina et al., 2019; Atoma et al., 2020; Yared on the farmlands located at the down slope
Mesfin et al., 2020). However, the erosion risk position. The field visits and satellite images also
under different land management practices of depict that the farmlands in the south eastern
the farmland has considerable variation. parts mainly in Konso areas have well developed
Accordingly, as our FGDs, KIIs and field terraces to minimize soil erosion (Figure 5). Here,
assessment revealed, farmers in the highlands it is worthy to mention that the Konso people are
have been practicing different land management known traditionally for well-developed terraces.
activities to enhance sustainable land use and The United Nations Educational, Scientific and
crop production. From the farmlands of the sub- Cultural Organization (UNESCO) have registered
basin, nearly half (46%) are characterized by the traditional terracing practice of Konso as a
Enset (Ensete ventricosum, Musaceae) and agro- world heritage, which is estimated to be older
forestry based farming mixed with homestead than 400 years (Watson and Currey 2009).
tree and fruit planting. The agro-forestry based

Figure 5. Terracing practice in Konso area.

Topographic (LS) factor: risk or significantly higher LS values. As shown

The LS factor analysis output under the two in Figure 6, parts of the sub-basin are
scenarios, using 30 m and 200 m resolution DEM characterized by depositional or very low soil
data is given in Figure 6 and Table 2. The erosion hazard areas, which are located along the
analysis shows that the sub-basin has a very valley plain following the Weyto River and its
complex landscape and topography. About 7% major tributaries. This is in accordance with the
of the sub-basin have <2% slope gradients, findings of Habtamu Sewnet and Amare Sewnet,
whereas considerable parts (about 20%) of the (2016). (2016) who reported that the LS factor has
sub-basin are characterized by steep (>25%) a significant influence on soil loss at the upper-
slope gradients that can facilitate high erosion slope as opposed to lands at the lower landscape
10 Shimeles Damene and Poshendra Satyal

in the watershed. The soils in the low lying areas surfaces drainage; thus they are depositional
are dominated by Vertisols and Cambisols, than being exposed to erosion (Nyssen et al.,
which are characterized by poor internal and 2019).

Figure 6. Weyto sub-basin topographic (LS) factor map calculated using 200 m (a) and 30 m (b) resolutions DEM data.

As indicated in Table 2 and Figure 6, the low vanishing of steep slopes and micro-relief
resolution (small scale, i.e., 200 m) DEM data features that tends to lengthen the flow path, and
exaggerate the slope values in level to sloping hence increasing the catchment areas (Wilson
(<25% slope) landforms compared to high and Allant, 2000). According to the FAO (2006)
resolution (large scale, i.e., 30 m resolution) DEM classification, steep lands have over 30% slope
data. Analysis of 200 m resolution DEM data that include high gradient escarpment, hills,
revealed that the coverage area of land with mountains and valley landforms. This highlights
<25% slope gradient is 5.7% greater than area that the estimation of LS factor values is
generated from 30 m resolution DEM data. In influenced by the resolution of DEM data under
contrast, low resolution (200 m) DEM data use. Clinometers based slope measurement
underestimate slope values at steep terrains applied to verify slope data (generated using 200
(>25% slope) compared to high (30 m) resolution and 30 m resolutions DEM data) showed
DEM data. Thus, the area covered by steep slope considerable generalization and aggregation at
(>25%) landform generated from 200 m lower slope percentage in low than high
resolution DEM data is 5.7% lower than that of 30 resolution DEM data. Therefore, LS estimation
m resolution DEM data. This is in accordance using high resolution data might be more
with the findings of Zhang et al. (2008b) who realistic to the natural slope condition. Hence, we
stated that a coarse resolution DEM generates suggest that a better estimation of erosion risk
more generalized terrain by maintaining only the can be done using the high resolution DEM data.
major relief features. These results in the
SINET: Ethiop. J. Sci.,46(1), 2023 11

Table 2. Weyto watershed topographic (LS) factor and area covered by units of factors.

No Slope range Combined Area covered under

(%) LS factor* 200 m resolution DEM 30 m resolution DEM data (B) Difference
data (A) (A-B)
Hectare % Hectare % Hectare %
1 < 2% 0.19 41,822 9.5 21,871.5 5.0 19,951 4.6
2 2 – 4% 0.38 58,665 13.4 48,696.2 11.1 9,968 2.3
3 4 – 6% 0.66 49,467 11.3 48,867.8 11.1 599 0.1
4 6 – 8% 1.14 39,574 9.0 41,995.4 9.6 -2,422 -0.6
5 8 – 13% 1.90 69,661 15.9 73,719.8 16.8 -4,059 -0.9
6 13 – 25% 3.80 106,536 24.3 105,583.5 24.1 953 0.2
7 25 – 40% 6.08 59,232 13.5 60,007.2 13.7 -775 -0.2
8 40 – 55% 7.98 11,157 2.5 22,868.6 5.2 -11,712 -2.7
9 55 – 100% 10.45 2,270 0.5 13,766.2 3.1 -11,496 -2.6
10 >100% 19.00 - - 1,007.8 0.2 -1,008 -0.2
Total 438,384 100 438,384 100.0
*Adopted from Hurni (1985)

Erosion hazard analysis the low resolution DEM data underestimated

The sub-basin soil loss rate through water erosion risk of steeper slope areas due to
erosion was calculated using the RUSLE by relatively low level of generalization by the
integrating it with the GIS technique using two higher resolution DEM data. Total area estimated
different resolutions DEM data. As discussed to have moderate to very high soil erosion risk
earlier, the different factors were mapped and using 200 m and 30m resolution DEM data
converted into raster data. Raster data output of account for about 1.9% (8,330 ha) and 7.6%
the soil erosion factors (R, K, LS, C and P) were (33,407 ha), respectively. In general, the erosion
imported into ArcMap 10.3.1 and the erosion rate risk analysis using the two different resolutions
was estimated using the raster calculator. In this DEM data revealed that over 90% of the sub-basin
analysis, the 200 and 30 m resolution DEM data does not have concerning soil erosion rate, which
yielded slightly different two soil erosion risks. was estimated within tolerable soil loss limit, i.e.,
In general, the soil erosion estimate using the below 10 t/ha/yr. This study is in harmony with
higher resolution (30 m) was higher than low the findings of various studies focused on GIS
resolution (200 m) DEM data. The maximum and RS-based RUSLE model prediction (e.g., Fu et
erosion estimate using 30 and 200 m resolution al., 2005; Kouli et al., 2009; Demirci and
DEM data were 36t/ha/yr and 21t/ha/yr, Karaburun 2012; Habtamu Sewnet and Amare
respectively (Figure 7). As shown in Table 3, area Sewnet, 2016; Gaubi et al., 2017; Bouhadeb et al.,
mapped using 200 m and 30 m resolution DEM 2018; Ayele Desalegn et al., 2018; Balabathina et
data as natural to very slight erosion risk account al., 2019; Phinzi and Ngetar, 2019; Atoma et al.,
for about 89.3% and 78%, respectively. Nearly 2020) that reported low soil erosion losses at
11.3% area was generalized in low slope range in steep slope land under good vegetation cover
case of low resolution (200 m) DEM data and land management practices.
compared to 30 m resolution data. In converse,

Table 3. Weyto sub-basin soil loss (A) in t/ha/y using 200 m and 30 m resolution DEM data.

N Soil loss Erosion class Area covered under

o ranges 200 m resolution 30 m resolution DEM Difference
(t/ha/yr) DEM data (A) data (B) (A-B)
Hectare % Hectare % hectare %
1 0-5 Natural - very slight 391,459 89.3 342,124 78.0 49,335 11.3
2 5 – 10 Slight 38,595 8.8 62,853 14.3 -24,258 -5.5
3 10 – 15 Moderate 7,736 1.8 25,182 5.7 -17,446 -3.9
4 15 – 20 High 583 0.1 7,286 1.7 -6,703 -1.6
5 20 – 36 Very high 11 0.0 939 0.2 -928 -0.2
Total 438,384 100 438,384 100

Note: The soil erosion class is adopted from Asnake Yimam andAmare Bantider, (2019) and contextualized to local condition.

In total, the average annual soil loss from 200 m resolution DEM data was estimated 1.38
Weyto sub-basin through water erosion using million tons, while the estimate using 30 m
12 Shimeles Damene and Poshendra Satyal

resolution DEM data was 1.86 million tons under use/cover and management practices. The
the current land use/land cover (LULC) and land model predicted moderate to severe soil erosion
management practice. Generally, the analysis risk in the northwest, northeast and southeast
revealed more generalization and watersheds, while the rest parts are likely prone
underestimation of erosion risk using low to low risk (Figure 7). The low soil loss rate is
resolution DEM data. This implies that the mainly associated to good vegetation cover over
accuracy of soil loss estimation decreases owing large area. Agro-forestry and SWC practices in
to the coarse (low) resolution of DEM (Mondal et some areas also contributed to lesser estimate of
al., 2017). the model on farmlands. Out of 17 identified
The landmasses having low erosion risk either major (sub) watersheds, Dencha and Bera
have good vegetation cover (bush-, wood-lands, watersheds have very high soil erosion hazards,
and forests) or receive relatively low rainfall or followed by Upper Bezo, Konte and also part of
have level topography. Areas mapped under Lomate, Keselte, Keseba and Zororo watersheds.
moderate to very high soil erosion rate are Any change in LULC and land management
mainly farmlands. On the other hand, farmlands practices would potentially alter the soil loss
mapped under slight erosion risk are protected rate. LULC shift to crop production, particularly
from erosion hazards with different management on lands over 3% slope should be accompanied
practices like terracing, agro-forestry and Enset by appropriate SWC and other land management
based farming practices, where the factors are practices, otherwise soil degradation through
captured in the model. The current RUSLE, GIS erosion can be significantly increased (Shimeles
technique and RS data based modelling of Damene et al., 2012). In converse, good land
erosion risk results show similarity with earlier management practices like terracing, agro-
studies (e.g. Habtamu Sewnet and Amare forestry practices and shift from annual to
Sewnet, 2016; Gezahegn Weldu et al., 2018; perennial crops in sloppy area could reduce
Yared Mesfin et al., 2020). accelerated soil erosion via water or runoff
(Young, 1989; USDA, 2011). In light of these
Erosion hazard hotspots and implication for findings, we recommend that land use
watershed development governance needs to be improved through strict
As our GIS based modelling revealed, there enforcement of Ethiopia’s existing policies such
were significant differences in spatial soil loss as the rural land use and management rules
across the sub-basin. Although the sub-basin has (Negarit Gazeta No. 456/2005) and Community-
considerably large (over 80%) undulating to Based Participatory Watershed Development
steep landform, soil erosion rate is not as such a Guideline (MoARD, 2005) so as to minimize
severe problem over large area owing to the land potential danger on land resources.

Figure 7. Weyto sub-basin soil erosion (t/ha/y) calculated using 200 m (a) and (b) 30 m resolution DEM data.
SINET: Ethiop. J. Sci.,46(1), 2023 13

CONCLUSION REFERENCES

The results from our analysis demonstrates the 1. Asnake Yimam Yesuph, Amare Bantider Dagnew,
usefulness of combining the GIS technique with (2019). Soil erosion mapping and severity
RS data such as the DEM for RUSLE based soil loss analysis based on RUSLE model and local
perception in the Beshillo Catchment of the
estimation and in identifying water erosion
Blue Nile Basin, Ethiopia. Environmental
hotspot areas. The findings from the study can Systems Research 8(17). https://doi.org/
provide some useful insights to a range of 10.1186/s40068-019-0145-1
relevant actors, particularly the land use 2. Atoma, H., Suryabhagavan, K.V., Balakrishnan, M.
planners, development actors; policy and (2020). Soil erosion assessment using RUSLE
decision makers so as to conduct regular erosion model and GIS in Huluka watershed, Central
risk monitoring and devise appropriate actions. Ethiopia. Sustainable Water Resources
However, as our analysis shows, the quality of Management 6(12). https://doi.org/
erosion risk estimation depends on scale and/or 10.1007/s40899-020-00365-z
resolution of input data, i.e., erosion factor data 3. Bagherzadeh, A. (2014). Estimation of soil losses by
such as the DEM. The analysis reveals that erosion USLE model using GIS at Mashhad plain,
Northeast of Iran, Arab. Journal of
estimation using higher (30 m) resolution DEM
Geosciences 7: 211–220. https://doi.org
data is more realistic as the slope generated is /10.1007/s12517-012-0730-3
more in line with ground reality and allows
4. Balabathina, V., Raju, R.P., Mulualem, W. (2019).
lesser generalization. Based on our findings from Integrated Remote sensing and GIS-based
the case of Weyto sub-basin, we thus conclude Universal Soil Loss Equation for Soil Erosion
that the GIS technique and RS oriented RUSLE Estimation in the Megech River Catchment,
model can be used for soil erosion risk prediction Tana Lake Sub-basin, Northwestern Ethiopia.
in large areas even at basin, sub-basin and macro American Journal of Geographic Information
watershed level. However, the accuracy of the System 8(4):141-157. https://doi.org
prediction can be enhanced using very detailed /10.5923/j.ajgis.20190804.01
input data, i.e., high resolution RS (DEM and 5. Bagegnehu Bekele, Yenealem Gemi, Y. (2021). Soil
satellite data) or large-scale maps of other erosion risk and sediment yield assessment
with universal soil loss equation and GIS: in
erosion factors data.
Dijo watershed, Rift valley Basin of Ethiopia.
Our field survey reveals that areas identified as Modelling Earth Systems and Environment
high erosion risk hotspot areas have very high 7:273–291. https://doi.org/10.1007/s40808-
water erosion, including severe gully formation 020-01017-z
and land slide. Moreover, number of pocket 6. Benavidez, R., Jackson, B., Maxwell, D., Norton, K.
areas with very high erosion evidence has been (2018). A review of the (Revised) Universal
mapped as slight and moderate erosion risk Soil Loss Equation ((R)USLE): with a view to
areas. The generalization and underestimation of increasing its global applicability and
erosion prediction is mainly emanated from improving soil loss estimates. Hydrology and
Earth System Science 22(11): 6059–608.
generalized input data that have very low scale
https://doi.org/10.5194/hess-22-6059-2018
(resolution), making it difficult to disaggregate
7. Bojö, J., Cassells, D. (1995). Land degradation and
the data for necessary details. In light of these
rehabilitation in Ethiopia: A reassessment.
findings, we suggest that the application of more AFTES, Working Paper 17, World Bank.
detailed soil, LULC, land management (such as https://documents1.worldbank.org/curated
SWC practices and topographic or LS factors) /en/381421468744300592/pdf/multi-
might help to yield better erosion risk estimation. page.pdf Accessed 30 April 2021.
As topography (particularly the slope of the 8. Bouhadeb, C.E., Menani, M.R., Bouguerra, H.,
land) plays an important role in soil erosion, use Derdous, O. (2018). Assessing soil loss using
of high resolution DEM data might also improve GIS based RUSLE methodology. Case of the
the model estimation capacity. Therefore, erosion Bou Namoussa watershed – North-East of
risk analysis result of high resolution (scale) Algeria. Water and land development 36(1-3):
27–35. https://doi.org/10.2478/jwld-2018-
input data might help planers and researchers to
0003
give practical and accurate advices for
9. Cohen, M.J., Shepherd, K.D., Walsh, M.G. (2005).
practitioners and policy makers.
Empirical reformulation of the universal soil
loss equation for erosion risk assessment in a
tropical watershed. Geoderma 124 (3-4):235–
252. https://doi.org/10.1016/j.geoderma.
2004.05.003
14 Shimeles Damene and Poshendra Satyal

10. Shimeles Damene, Asmamaw Bahir A, Villamor scale using RUSLE and GIS: A case study in
GB (2020) The Role of Chomo grass the loess plateau of china. Land Degradation
(Brachiaria humidicola) and exclosures in & Development 16:73–85. https://doi.org
restoring soil organic matter, total nitrogen /10.1002/ldr.646
and associated functions in degraded lands in 20. Ganasri, B.P., Ramesh, H. (2016). Assessment of
Ethiopia. Regional Environmental Change soil erosion by RUSLE model using remote
20:92. https://doi.org/10.1007/s10113-020- sensing and GIS - A case study of
01680-z Nethravathi Basin. Geoscience Frontiers
11. Shimeles Damene, Lulseged Tamene, Vlek, P.L.G. 7(6):953-961. https://doi.org/10.1016/j.gsf.
(2012). Performance of farmland terraces in 2015.10.007
maintaining soil fertility: A case of Lake 21. Gaubi, I., Chaabani, A., Ben Mammou, A.,
Maybar watershed in Wello, northern Hamza, M.H. (2017). A GIS-based soil erosion
highlands of Ethiopian. Journal of Life prediction using the Revised Universal Soil
Science 6(55):2151-2161. https://doi.org/ Loss Equation (RUSLE) (Lebna watershed,
10.17265/1934-7391/2012.11.010 Cap Bon, Tunisia). Natural Hazards 86:219–
12. Shimeles Damene, Lulseged Tamene, Vlek, P.L.G. 239. https://doi.org/10.1007/s11069-016-
(2013). Performance of exclosure in restoring 2684-3
soil fertility: A case of Gubalafto district in 22. Habtamu Sewnet Gelagaya, Amare Sewnet Minale,
North Wello Zone, northern highlands of (2016). Soil loss estimation using GIS and
Ethiopia. CATENA 101:136-142. https:// Remote sensing techniques: A case of Koga
doi.org/10.1016/j.catena.2012.10.010 watershed, Northwestern Ethiopia.
13. Demirci, A., Karaburun, A. (2012). Estimation of International Soil and Water Conservation
soil erosion using RUSLE in a GIS Research 4(2):126-136. https://doi.org/
framework: a case study in the 10.1016/j.iswcr.2016.01.002
Buyukcekmece Lake watershed, northwest 23. Halcrow Group Limited and Generation Integrated
Turkey. Environmental Earth Rural Development (GIRD) Consultants,
Sciences 66:903–913. https://doi.org/ 2007. Rift Valley Lakes Basin Integrated
10.1007/s12665-011-1300-9 Resources Development Master Plan Study
14. Ayele Desalegn, Abrham Tezera Gessesse, Fikrey Project, Part II Sector Assessments, Volume
Tesfay, F. (2018). Developing GIS-Based Soil 2:– Natural Resources. Annex A: Geology
Erosion Map Using RUSLE of Andit Tid and Seismology; Annex C: Soils and Land
Watershed, Central Highlands of Ethiopia. Evaluation, Annex D: Land Use and Land
Journal of Scientific Research & Reports Cover and Annex F: Livestock and
19(1):1-13. https://doi.org/10.9734/JSRR Rangeland. Addis Ababa, Ethiopia
/2018/40841 24. Hurni, H. (1998). Agro-ecological belts of Ethiopia:
15. Duiker, S.W., Flanagan, D.C., Lal, R. (2001). Explanatory notes on three maps at a scale of
Erodibility and infiltration characteristics of 1:1,000,000. Soil Conservation Research
five major soils of southwest Spain. Catena Program Research report. https://edepot.
45:103–121. https://doi.org/10.1016/S0341- wur.nl/484855. Accessed 20 May 2021
8162(01)00145-X 25. Hurni, H. (1993). Land degradation, famines and
16. FAO (2004) Methodological framework for land resource scenarios in Ethiopia. In: Pimentel
degradation assessment in dry-lands. (ed.) World soil erosion and conservation.
ftp://ftp.fao.org/agl/agll/lada/LADA- Cambridge University Press, Cambridge, 27-
Methframwk-simple.pdf Accessed 25 July 62.http://dx.doi.org/10.1017/CBO978051173
2021. 5394.004
17. FAO (2006) Guidelines for soil description, Fourth 26. Hurni, H. (1985). Erosion-Productivity-
edition. Italy, Rome. https://www.fao.org Conservation Systems in Ethiopia.
/3/a0541e/a0541e.pdf Accessed 20 May Proceedings of 4th International Conference
2021. on Soil Conservation, Maracay, Venezuela, 3
18. Almaw Fenta, Tsuneka, A., Nigussie Haregeweyn, (9):654-674
N., Poesen, J., Tsubo, M., Borrelli, P., 27. Hurni, H., Berhe, W.A., Chadhokar, P., Daniel, D.,
Panagos, P., Vanmaercke, M., Broeckx, J., Gete, Z., Grunder, M., Kassaye, G. (2016). Soil
Yasuda, H., Kawai, T., Kurosaki, K. (2020). and Water Conservation in Ethiopia:
Land susceptibility to water and wind Guidelines for Development Agents, Second
erosion risks in the East Africa region. revised edition. Bern, Switzerland: Centre for
Science of The Total Environment 703:135016 Development and Environment (CDE),
https://doi.org/10.1016/j.scitotenv.2019.135 University of Bern, with Bern Open
016 Publishing (BOP).
19. Fu, B.J., Zhao, W.W., Chen, L.D., Zhang, Q.J., Lu, 28. Hurni, H., Herweg, K., Portner, B., Liniger, H.
Y.H., Gulinck, H., Poesen, J. (2005). (2008). Soil Erosion and Conservation in
Assessment of soil erosion at large watershed Global Agriculture. In: Braimoh, A.K. and.
SINET: Ethiop. J. Sci.,46(1), 2023 15

Vlek, P.L.G (Eds.), Land Use and Soil 39. Nyssen, J., Tielens, S., Tesfamichael
Resources 41-71. https://doi.org/10.1007/ Gebreyohannes, Tigist Araya, Kassa Teka,
978-1-4020-6778-5_4 Van de Wauw, J., Degeyndt, K.,
29. Kinnell, P.I.A. (2005). Why the universal soil loss Descheemaeker, K., Kassa Amare, Mitiku
equation and the revised version of it do not Haile, Amanuel Zenebe, Munro, N.,
predict event erosion well. Hydrological Walraevens, K., Kindeya Gebrehiwot, K.,
Process 19(3):851–854. http://dx.doi.org/ Poesen, J., Frankl, A., Alemtsehay Tsegay,
10.1002/hyp.5816 Deckers, J. (2019). Understanding spatial
patterns of soils for sustainable agriculture in
30. Kouli, M., Soupios, P., Vallianatos, F. (2009). Soil
northern Ethiopia’s tropical mountains. PLoS
erosion prediction using the Revised
ONE 14(10): e0224041. https://doi.org/
Universal Soil Loss Equation (RUSLE) in a
10.1371/journal.pone.0224041
GIS framework, Chania, Northwestern Crete,
Greece. Environmental Geology 57:483–497. 40. Panagos, P., Borrelli, P., Poesen, J., Ballabio, C.,
https://doi.org/10.1007/s00254-008-1318-9 Lugato, E., Meusburger, K., Montanarella, L.,
Alewell, C. (2015a). The new assessment of
31. Krusekopf, H.H. (1943). The Effect of Slope on Soil
soil loss by water erosion in Europe. Environ.
Erosion. University of Missouri College 'of
Sci. Policy, 54:438–447. https://doi.org
agriculture agricultural experiment station,
/10.1016/j.envsci.2015.08.012
Research bulletin 363. https://core.ac.uk/
download/pdf/62789983.pdf Accessed 10 41. Panagos, P., Ballabio, C., Borrelli, P., Meusburger,
January 2021 K., Klik, A., Rousseva. S., Tadić, M.P.,
Michaelides, S., Hrabalíková, M., Olsen, P.,
32. Kumar, S., Kushwaha, S.P.S. (2013). Modelling soil
Aalto, J., Lakatos, M., Rymszewicz, A.,
erosion risk based on RUSLE-3D using GIS in
Dumitrescu, A., Beguería, S., Alewell, C.
a Shivalik Sub-watershed. Journal of Earth
(2015b). Rainfall erosivity in Europe. Science
System Science 122(2): 389-398.
of The Total Environment 511(1): 801-814.
https://doi.org/10.1007/s12040-013-0276-0
https://doi.org/10.1016/j.scitotenv.2015.01.0
33. López-García, E.M., Torres-Trejo, E., López-Reyes, 08
L., Flores-Domínguez, Á.D., Peña-Moreno,
42. Panditharathne, D.L.D., Abeysingha, N.S.,
R.D., López-Olguín, J.F. (2020). Estimation of
Nirmanee, K.G.S., Mallawatantri, A. (2019).
soil erosion using USLE and GIS in the
Application of Revised Universal Soil Loss
locality of Tzicatlacoyan, Puebla, México. Soil
Equation (RUSLE) Model to Assess Soil
and Water Research, 15 (1):9–17.
Erosion in “Kalu Ganga” River Basin in Sri
https://doi.org/10.17221/165/2018-SWR
Lanka. Applied and Environmental Soil
34. Mondal, A., Khare, D., Kundu, S., Mukherjee, S., Science 2019:4037379. https://doi.org/
Mukhopadhyay, A., Mondal, S. (2017). 10.1155/2019/4037379.
Uncertainty of soil erosion modelling using
43. Phinzi, K., Ngetar, N.S. (2019). The assessment of
open source high resolution and aggregated
water-borne erosion at catchment level using
DEMs. Geoscience Frontiers 8(3):425-436.
GIS-based RUSLE and remote sensing: A
https://doi.org/10.1016/j.gsf.2016.03.004
review. International Soil and Water
35. Nana-Sinkam, S.C. (1995). Land and environmental Conservation Research, 7(1):27–46.
degradation and desertification in Africa, https://doi.org/10.1016/j.iswcr.2018.12.002
ECA/FAO. http://www.fao.org/docrep
44. Renard, K.G., Foster, G.R., Weesies, D.K., McCool,
/X5318E/x5318e02.htm.
Yoder, D.C. (1997). Predicting Soil Erosion by
36. Nearing, M. (1997). A Single, Continuous Function Water: A Guide to Conservation Planning
for Slope Steepness Influence on Soil Loss. with the Revised Universal Soil Loss
Soil Science Society of America Journal , Equation (RUSLE). Agricultural Handbook,
61(3). https://doi.org/10.2136/sssaj1997. Number 703, Washington DC.
03615995006100030029x https://www.ars.usda.gov/arsuserfiles/640
37. Nearing, M.A., Jetten, V., Baffaut, C., Cerdan, O., 80530/rusle/ah_703.pdf Accessed 18 May
Couturier, A., Hernandez, M,, Le Bissonnais, 2021
Y., Nichols, M.H., Nunes, J.P., Renschler, 45. Scherr, S.J. (1999). Soil degradation, a threat to
C.S., Souche`re, V,, van Oost, K. (2005). developing-country food security by 2020?
Modeling response of soil erosion and runoff Food, agriculture and the environment
to changes in precipitation and cover. Catena discussion paper 27. International Food
61:131–154. https://doi.org/10.1016/j.catena. Policy Research Institute 2033 K Street, N.W.
2005.03.007 Washington DC 20006-1002 USA.
38. Negarit Gazeta, 2005. Federal Democratic Republic https://acervo.socioambiental.org/sites/def
of Ethiopia rural land administration and use ault/files/documents/T1D00022.pdf.
proclamation. Negarit Gazeta, No 456/2005, Accessed 5 January 2021
Addis Ababa. 46. Bekele Shiferaw, Holden, S.T. (1998). Resource
degradation and adoption of land
16 Shimeles Damene and Poshendra Satyal

conservation technologies in the Ethiopian 53. Wilson, J.P., Gallant, J.C. (2000). Terrain Analysis:
highlands: A case study in Andit Tid, North Principles and Applications, New York.
Shewa. Agricultural Economics 18:233-247. 54. Wischmeier, W.H., Smith, D.D. (1978). Predicting
https://doi.org/10.1111/j.1574-862.1998. Rainfall Erosion Losses. A Guide to
tb0050 Conservation Planning. The USDA
47. Lulseged Tamene., Vlek, P.L.G., Park, S.J., Dikau, Agricultural Handbook No. 537, Maryland.
R. (2006). Analysis of factors determining 55. Wischmeier, W.H., Smith, D.D. (1965). Prediction
sediment yield variability in the highlands of Rainfall Erosion Losses from Cropland East
northern Ethiopia. Geomorphology, 76(1):76- of the Rocky Mountains: A Guide for
91. https://doi.org/10.1016/j.geomorph. Selection of Practices for Soil and Water
2005.10.007 Conservation. Agricultural Handbook, No.
48. Kebrom Tekle, (1999). Land degradation problems 282.
and their implications for food shortage in 56. Gezahegn Weldu Woldemariam, Anteneh
south Wello, Ethiopia. Environmental Derribew Iguala, Solomon Tekalign, Reddy
Management, 23(4):419-427. https://doi.org R.U. (2018). Spatial Modeling of Soil Erosion
/10.1007/s002679900197 Risk and Its Implication for Conservation
49. Yared Mesfin Tessema, Jasi´nska, J., Lemma Tiki Planning: the Case of the Gobele Watershed,
Yadeta, Switoniak, M., Puchalka, R., Eyob East Hararghe Zone, Ethiopia. Land 7(25).
Gebrehiwot Gebregeorgis, (2020). Soil Loss https://doi.org/10.3390/land7010025
Estimation for Conservation Planning in the 57. Young, A. (1989). Agroforestry for soil
Welmel Watershed of the Genale Dawa conservation; CAB International,
Basin, Ethiopia. Agronomy 10(6):777. International Council for Research in
https://doi.org/10.3390/agronomy10060777 Agroforestry.http://apps.worldagroforestry.
50. USDA (2011). National Soil Survey Handbook, 430- org/downloads/Publications/PDFS/03_Agr
VI, Department of Agriculture, Natural oforestry_for_soil_conservation.pdf.
Resources Conservation Service. Accessed 05 February 2021
http://soils.usda.gov/technical/handbook/ 58. Zhang, K.L., Shu, A.P., Xu, X.L., Yang, Q.K., Yue,
Accessed 17 October 2020. B. (2008a). Soil erodibility and its estimation
51. Vlek, P.L.G., Le, B.Q., Lulseged Tamene, (2008). for agricultural soils in China. Journal of Arid
Land decline in land-rich Africa. A creeping Environments 72(6):1002-1011. https://
disaster in the making. Center for doi.org/10.1016/j.jaridenv.2007.11.018
development research (ZEF) University of 59. Zhang, J.X., Chang, K., Wu, J.Q. (2008b). Effects of
Bonn, Science Council and CGIAR. DEM resolution and source on soil erosion
https://www.fao.org/3/i0056e/i0056e00.ht modelling: a case study using the WEPP
m. Accessed 10 February 2021 model. International Journal of Geographical
52. Watson, E.E., Currey, J. (2009). Living terraces in Information Science 22(8):925-942.
Ethiopia: Konso landscape, culture and https://doi.org/10.1080/13658810701776817
development, University of Cambridge.