Modeling Earth Systems and Environment

https://doi.org/10.1007/s40808-023-01808-0

ORIGINAL ARTICLE

Modeling of watershed intervention techniques to rehabilitate

sediment yield hotspot areas in Hare watershed, rift valley basin,
Ethiopia
Mesfin Amaru Ayele1 · Abebe Temesgen Ayalew1

Received: 1 March 2023 / Accepted: 27 May 2023

© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2023

Abstract
Soil erosion and sedimentation are broadly recognized as one of the major earth and environmental problems. Increase of
agricultural activities, poor farming system, deforestation and overgrazing causing serious soil erosion and sediment yield
in the watershed. Modeling watershed management practices play significant role on soil erosion reduction, increase soil
moisture content and decrease sediment production. Thus, the main objective of this study was modeling of watershed
intervention practices to rehabilitate the soil erosion and sediment hotspot areas. Twenty four suspended sediment samples
were collected from Hare River and its concentration was determined in laboratory. The sediment concentration was then
converted into sediment load in ton per day and sediment-rating curve was developed. Eighteen years sediment data was
generated using developed sediment rating curve. Monthly flow and sediment data calibrated and validated in Soil and
Water Assessment Tool-Calibration and Uncertainty Program (SWAT-CUP). Model performance was checked and found
very well. Spatial variability map of the sediment yield in the watershed was developed. Twelve sediment hotspots sub-
watersheds which produce sediment from high to severe (20.38 to 61.99ton/ha/year) were identified. For those twelve
sediment hotspot sub-watersheds, four watershed intervention techniques were applied. After application of strip cropping,
residue management, contouring and terracing, sediment production was reduced by 47.42%, 5.54%, 69.22%, and 84.88%,
respectively. Among the intervention techniques terracing found to be the most watershed treatment approach in this study.
The research finding is vital to environmental protection authority, decision makers and scientific community to undertake
intervention techniques for soil erosion hotspot sub-watersheds.

Keywords Hare watershed · SWAT · Sediment yield · Hotspot sub-watersheds · Intervention techniques

Introduction (Opeyemi et al. 2019). The International Soil Reference and

Information Centers survey report indicated that soil erosion
Soil erosion and sedimentation are the major environmental accounts about nine million hectares of land is degraded
problems in the world (Borrelli 2017; Degife et al. 2021; all over the world (ITPS 2015; Kuti and Ewemoje 2021).
Deshmukh et al. 2019; Mohammed et al. 2021; Tamiru and Deforestation, poor farming system, and expansion of agri-
Wagari 2021). Globally, about 84% of land degradation cultural activities have been causing thoughtful soil ero-
occurs due to soil erosion and poor land use management sion and sediment yield (Balabathina et al. 2019; Gashaw
et al. 2020; Tsegaye and Bharti 2021; Woldemariam et al.
2018; Zerihun et al. 2018). Around 1.9 billion tons of fertile
agricultural lands were annually washed by soil erosion in
Mesfin Amaru Ayele Ethiopian high lands (Semu 2018).
[email protected]
According to Dinesh et al. (2021), soil erosion causes
Abebe Temesgen Ayalew the removal of rich inorganic matter and nutrients result-
[email protected]
ing in the decrease of agricultural productivity, reduction of
1
Faculty of Hydraulic and Water Resources Engineering, Arba farmers’ incomes, land desertification, and potential risk to
Minch University, Water Technology Institute, Arba Minch, human lives. Erosion also damages the natural environment,
Ethiopia

13
 Modeling Earth Systems and Environment

for instance, distortion of landscape (landslides, slope insta- to rehabilitate the soil erosion and sediment hotspot areas
bility, land degradation), river course aggradation, which using Soil and Water Assessment Tool (SWAT). Specifi-
causes the river bed to increase, reducing the transport cally, (1) to quantify total volume of sediment yield at outlet
capacity of the flow and increasing flood risk (Mithlesh et of Hare watershed; (2) to develop spatial variability of sedi-
al. 2022). ment map in watershed scale and (3) to identify sediment
To ensure sustainable land and water resources in a par- hotspot areas and application of intervention strategies.
ticular watershed, efficient watershed intervention tech- Thus, the research outputs is crucial to design and imple-
nique need to be carefully considered (Tripathi et al. 2004). ment appropriate management intervention; rehabilitate
In Ethiopia, significant soil erosion and sediment interven- the soil erosion and sediment hotspot areas; increase soil
tion activities were implemented during the 1970 and 1980s fertility and agricultural productivity; ensure food security;
by mobilizing farmers through their peasant associations, decrease the severity of the sedimentation problem in Lake
mainly in food for work programs (Woldeamlak 2007). In Abaya and for the environment restoration.
the recent time, the government has also been conducted
soil erosion and sediment intervention activities with inte-
grated and participatory watershed development approaches Materials and methods
to improve rural livelihoods and environmental protection
(Molla and Sisheber 2017). Indigenous soil and water con- Watershed under study
servation terraces are a well-developed practice in differ-
ent parts of Ethiopia, namely in Konso, South Shewa, and Hare watershed is the sub watershed of Lake Abaya basin
Harangue plateau (Beshah 2003). In Konso zone, the ero- and situated in the southern part of Ethiopia and covers
sion management practice like terracing with stone wall drainage area of 187.14 km2. The geographical extent of the
constructed and it is effective for a long period of time Hare watershed is roughly between 6°4’00’’ to 6°17’30’’ N
(Amanuel 2018; Wiebke 2003). latitudes and 37°27’00’’ to 37°45’00’’ E longitudes (Fig. 1).
The watershed management practices with tree conserva- It is categorized at the upstream part as steep valleys and
tion and acidic soil amendments were carried out in south- increasingly fluvial plain until it enters the receiving Lake
ern Ethiopia (Yoseph et al. 2019). However, there is limited Abaya (Yisehak 2021).
scientific research on watershed management interven-
tion in Ethiopia and the conducted studies are not uniform Suspended sediment data collection and analysis
(Assefa and Hans 2014). Southern Ethiopian highlands are
considered as high population, high rainfall, sloppy land, Primary suspended sediment data collection is needed due
and high soil erosion which reduce agricultural productiv- to there is no available sediment recorded data for model
ity; endangering food security and increasing river sedi- calibration and validation. The guideline to select the appro-
mentation (Zerssa et al. 2021). According to Biniyam and priate site selection, record length, and frequency of record
Kemal (2017), Hare watershed was utilized for agriculture, for sediment sample collection and analysis was based on
over-grazing with the density of livestock and fast- popula- the recommendation by (Ellison et al. 2014; Tran and Tuong
tion growth. There is also forest deforestation and steeper 2014; Yang et al. 2020). Materials used for suspended sedi-
slopes/hill lands are being used for agriculture deprived of ment data measurement in the laboratory were 0.45 μm
protective actions and much of the fertile soils have been diameter filter paper, oven dry (used to remove moisture
washed away from the watershed. Thus, poor land manage- content from the filter paper), desiccator (used to prevent fil-
ment in this watershed leads to serious soil loss and land ter paper from atmospheric pressure contact), suction filter
degradation, which causes substantial sediment carriage in (used to filter suspended sediment load from the suspended
the Hare River. The suspended sediment load from Hare sediment samples) and electronic balance (used to measure
River eventually enters into Lake Abaya that increases sedi- suspended sediment load).
mentation, rise of Lake level and causes flooding around the Twenty four suspended sediment sample data were col-
lake shore. Earlier studies in Hare watershed were available lected 3 times in a day using a 500ml bottles from the Hare
regarding the change of land use/cover impact on river dis- River for both rainy and dry seasons and coded for labo-
charge (Kassa and Gerd 2015); impacts of climate change ratory tests (Table 1). In the laboratory, the 0.45 μm filter
on rainfall, flood frequency, and water resources potential paper was prepared and placed in the oven dried for 24 h.
(Biniyam and Abdella 2017). However, they did not include After one day, the cooled filter paper was removed from the
sediment evaluation and management intervention tech- oven dry and again placed in the desiccator for 30 min to
niques to address the problem. The main objective of this prevent the filter paper from atmospheric contact and then
study was modeling of watershed intervention practices measuring the weightage of each filter paper in grams was

13
Modeling Earth Systems and Environment

Fig. 1 (a) Major river basins in Ethiopia; (b) Rift valley basin and (c) Hare watershed

occurred. To measure the sediment weight for each sample, Where, C is the suspended sediment concentration in (mg/l),
filter paper placed on the filter cape and the suspended sedi- m is the weight of sample and filter paper in gram (gm), wf is
ment sample added in the filter cape. Then, waited until the weight of filter paper in gram (gm), and vs. is the volume
all clear water passes into the filter glass and the sediment of the sample in liter (l).
load was remaining on the filter paper. After separating the Table 1 shows the low sediment yield was recorded in
sediment weight, the sample was put into a safe place until the dry season (e.g. for sample numbers S23, S21, and S24,
all samples were carried out by giving a code number for the suspended sediments were 13.8 mg/l, 35.6 mg/l and
each sample. Finally, measuring weightage for each sample 42.2 mg/l respectively). However, in rainy season, high sus-
with a combination of filter weightage was occurred. Over- pended sediment concentration was observed for the sample
all, suspended sediment weight measurement procedure in number S3, S7, and S9 were 1580.4 mg/l, 1416.4 mg/l, and
the laboratory with photo representation was illustrated in 1318.4 mg/l respectively.
Fig. 2.
The net weightage of each suspended sediment concen- Suspended sediment load
tration (mg/l) was calculated by using Eq. 1 (Zakwan 2019).
  Sediment concentration needs to be changed into sediment
m − wf load in tons per day using a mathematical curve fitting
C= 1000 (1)
vs method (Eq. 2) (Morris and Fan 1998).

S = 0.0864 ∗ Q ∗ C (2)

13
 Modeling Earth Systems and Environment

Table 1 Suspended sediment concentration (mg/l) from the field and laboratory measurement
Sample Code Date Depth (m) Width (m) Velocity (m/s) Flow (m3/s) Filter weight Sample Suspended
(gm) weight (gm) sediment
conc. (mg/l)
S1 12/07/2019 0.72 2.53 0.77 1.4026 0.7621 1.3117 1099.2
S2 13/07/2019 0.75 3.45 0.75 1.9406 0.7532 1.2641 1021.8
S3 14/07/2019 0.65 2.87 0.7 1.3059 0.773 1.5632 1580.4
S4 17/07/2019 0.61 1.78 0.78 0.8469 0.7362 1.3021 1131.8
S5 19/07/2019 0.62 3.82 0.74 1.7526 0.7247 1.2945 1139.6
S6 20/07/2019 0.55 3.67 0.65 1.3120 0.714 1.1871 946.2
S7 22/07/2019 0.54 2.69 0.8 1.1621 0.703 1.4112 1416.4
S8 11/08/2019 0.52 4.23 0.71 1.5617 0.7623 1.2362 947.8
S9 14/08/2019 0.57 1.56 0.51 0.4535 0.7866 1.4458 1318.4
S10 16/08/2019 0.48 1.8 0.55 0.4752 0.7761 1.2541 956
S11 18/08/2019 0.5 2.87 0.62 0.8897 0.789 1.2022 826.4
S12 19/08/2019 0.46 5.43 0.7 1.7485 0.8123 1.1892 753.8
S13 21/08/2019 0.45 2.54 0.55 0.6287 0.75 0.9252 350.4
S14 24/08/2019 0.36 1.92 0.45 0.3110 0.741 0.8923 302.6
S15 26/08/2019 0.35 3.46 0.5 0.6055 0.8143 0.9585 288.4
S16 19/10/2019 0.45 4.25 0.47 0.8989 0.8021 0.9373 270.4
S17 21/10/2019 0.4 6.51 0.44 1.1458 0.7246 0.9095 369.8
S18 24/10/2019 0.34 5.82 0.45 0.8905 0.753 0.8751 244.2
S19 14/01/2020 0.3 4.23 0.48 0.6091 0.7046 0.7547 100.2
S20 17/01/2020 0.25 3.82 0.49 0.4680 0.6934 0.7843 181.8
S21 19/01/2020 0.28 2.45 0.46 0.3156 0.6864 0.7042 35.6
S22 16/02/2020 0.35 3.48 0.52 0.6334 0.7546 0.7897 70.2
S23 18/02/2020 0.38 1.78 0.44 0.2976 0.7571 0.764 13.8
S24 22/02/2020 0.26 2.67 0.47 0.3263 0.719 0.7401 42.2

Fig. 2 Suspended sediment weight measurement in the laboratory

13
Modeling Earth Systems and Environment

Table 2 Measured suspended sediment load (tons/day) Developing sediment rating curve
Sample number Date Q (m3/s) C (mg/l) S (ton/day)
S1 12/07/2019 1.4026 1099.2 133.2092
Once the suspended sediment concentration is converted
S2 13/07/2019 1.9406 1021.8 171.3252
into sediment load, the relation between the flow and the
S3 14/07/2019 1.3059 1580.4 178.3093
sediment load was used to develop the sediment rating curve.
S4 17/07/2019 0.8469 1131.8 82.8186
S5 19/07/2019 1.7526 1139.6 172.5651 The collected suspended sediment load was a short period
S6 20/07/2019 1.3120 946.2 107.2603 record. Therefore, it is crucial to produce the long-period
S7 22/07/2019 1.1621 1416.4 142.2118 sediment load by correlating the flow with the short period
S8 11/08/2019 1.5617 947.8 127.8888 sediment record using a sediment rating curve. According to
S9 14/08/2019 0.4535 1318.4 51.6572 Khaleghi and Varvani (2018), the sediment rating curve is
S10 16/08/2019 0.4752 956 39.2508 a correlation between the flow and sediment load. After the
S11 18/08/2019 0.8897 826.4 63.5254 sediment rating curve (Fig. 3) is developed, continues long
S12 19/08/2019 1.7485 753.8 113.8742 period records of flow are changed into the records of sedi-
S13 21/08/2019 0.6287 350.4 19.0321 ment load using the sediment rating curve equation (Eq. 3)
S14 24/08/2019 0.3110 302.6 8.1320
suggested by (Morris and Fan 1998) as:
S15 26/08/2019 0.6055 288.4 15.0877
S16 19/10/2019 0.8989 270.4 21.0000
S = yQX (3)
S17 21/10/2019 1.1458 369.8 36.6079
S18 24/10/2019 0.8905 244.2 18.7877
S19 14/01/2020 0.6091 100.2 5.2733
S20 17/01/2020 0.4680 181.8 7.3503 Where, y and x are regression constants, S is sediment
S21 19/01/2020 0.3156 35.6 0.9706 load and Q is the discharge.
S22 16/02/2020 0.6334 70.2 3.8415 For Hare River, continuous eighteen years daily sediment
S23 18/02/2020 0.2976 13.8 0.3549 load for model calibration and validation was produced by
S24 22/02/2020 0.3263 42.2 1.1896 using the sediment rating curve equation (Eq. 4).

Where, 0.0864 is conversion factor, Q is the flow (m3/s), SSedimentload = 47.29Q2.56 (4)
S is the sediment load in (ton/day) and C is the suspended
sediment concentration (mg/l). The converted suspended
sediment load is shown in Table 2.

Fig. 3 Developed sediment rating curve

13
 Modeling Earth Systems and Environment

Hydro-meteorological data collection and analysis Soil and Water Assessment Tool (SWAT)

Eighteen years stream flow record was obtained from Ethio- SWAT model is a physically based, semi-distributed, long-
pia Ministry of Water and Energy. Climatological data such term simulation, deterministic, and originated from agri-
as rainfall, minimum and maximum temperature, sunshine cultural models with spatially distributed parameters and
hours, wind speed, and relative humidity were collected operating on a daily time (Arnold et al. 2012; Santhi et al.
from Ethiopian Meteorology Institute in the time range 2001). SWAT model has been used worldwide and consid-
from 1987 to 2016. There were three meteorological sta- ered as adaptable environmental model that can be used
tions in and nearby Hare watershed and their data availabil- to evaluate the biophysical impacts of intensification of
ity illustrated in Table 3. interventions at the watershed scale, which supports more
effective watershed management and the development of
Geospatial data collection and analysis better-informed policy (Hailu et al. 2016; Sofonyase and
Abate 2020). The model has been widely applied for the
Digital elevation model (DEM) defines the elevation of the simulation of runoff, sediment yield, nitrogen and phos-
watershed at a detailed spatial resolution (Roostaee and phorus losses from watersheds in different geographical
Deng 2020). DEM of 12.5 m x 12.5 m resolution for Hare locations, with varying soils, land use and management con-
watershed was downloaded from https://asf.alaska.edu/ ditions over long periods of time (Abebe 2019; Dulo and
on December 15, 2021. The downloaded DEM grids were Abate 2020; Neitsch et al. 2011; Samaneh et al. 2016). Sev-
mosaicked using ArcGIS 10.1 software and used in Soil eral researchers for instances Betrie et al. (2011); Dessie et
and Water Assessment Tool (SWAT) model for watershed al. (2014); Getahun et al. (2021); Wolde and Bogale (2017);
delineation and for further analysis (Arnold et al. 2012). The Yesuf et al. (2015) were proved the applicability of SWAT
study area elevation ranges approximately from the lowest model in the Ethiopian watersheds.
point 1166 m to the highest point 3481 m above the mean
sea level (Fig. 4a). The Hare watershed delineated with area Model sensitivity, calibration and validation
of 187.14km2 having 19 sub-watersheds (Fig. 4b).
Most of the upper part of the Hare watershed is covered The model applicability would be evaluated by the method
by moderately cultivated land, and the lower part is covered of sensitivity analysis, calibration and validation (Abba-
by intensively cultivated and shrub land (Kassa and Gerd spour et al. 2015; Kumar et al. 2017; Scott-Shaw et al.
2015). The lookup table of land use/cover with percentage 2020; White and Chaubey 2005). To understand the model
of area coverage has been prepared for SWAT 2012 data- parameter behavior with respect to the model outcome, a
base (Table 4). sensitivity analysis was performed. The main indication of
Watershed response to rainwater events depends on the sensitivity analysis is to identify the most watershed con-
landscape and type of soils (Dulo and Abate 2020; Shres- trolling model parameters (White and Chaubey 2005). The
tha et al. 2008). According to Raffa and Chiampo (2021), eight more influential (watershed controlling) flow and
soil texture, soil hydraulic conductivity, available moisture sediment parameters from high to medium sensitive were
content, and bulk density are the main physical properties, identified and used for further iterations in the calibration
whereas organic carbon content and electrical conductivity periods (Table 6). SWAT model was calibrated for flow and
are under the chemical properties of the soil. The shape file sediment from a period of 1990–2001 with two years warm-
of soil for Hare watershed was clipped from Ethiopia soil up period (1988–1989). The test of the model on an inde-
map. Soil types, area coverage and SWAT model database pendent set of data is model validation (Chandra et al. 2014;
code illustrated in Table 5. Sao et al. 2020). Six years (2002–2007) observed flow and
sediment data were used for SWAT model validation.

Table 3 Meteorological stations in and around the Hare watershed

No Gauging Latitude (m) Longitude (m) Elev. (m) Elements
Stations RF T max T min RH SH WS
1 Arba Minch 340637.01 670048.55 1207 √ √ √ √ √ √
2 Chencha 342900.72 688840.67 2632 √ √ √ N/A N/A N/A
3 Dorze 341782.21 684420.58 2505 √ √ √ N/A N/A N/A
Note: Elev. = Elevation, RF = Rainfall, Tmax = Maximum Temperature, Tmin = Minimum Temperature, RH = Relative Humidity, SH = Sunshine
Hour, WS = Wind Speed, √ = data available and N/A = data not available

13
Modeling Earth Systems and Environment

Fig. 4 (a) DEM of Hare watershed and (b) Delineated Hare watershed

Table 4 Land use/cover lookup table SWAT model performance indicators

Land use/cover Redefined land Area %Area SWAT
use/cover (km2) code A model must be evaluated on the extent of its accuracy,
Forest Forest-Mixed 2.88 1.51 FRST consistency and adaptability (Cerqueira et al. 2020; Cheng
Intensively Agricultural 15.92 8.36 AGRL and Morgan 2013). A forecast efficiency criterion is there-
Cultivated Land-Generic
fore necessary to judge the performance of the model (Nijat
Moderately Agricultural 151.83 79.75 AGRR
Cultivated Land-Row et al. 2016). Evaluating performance of a hydrologic model
Crops requires objective estimates of the closeness of the simulated
Riparian Vegetation Cabbage 0.11 0.06 CABG behavior of the model to observations (Cheng et al. 2017).
Shrub land Range-Brush 19.64 10.32 RNGB In this study, the performance of the SWAT model was car-
ried out by considering the following objective functions:
Table 5 The dominant types of soil, SWAT code, and area coverage of Nash-Sutcliffe efficiency: The Nash-Sutcliffe efficiency
Hare watershed (ENS) used to evaluate the overall agreement of the shape of
Original soil type Area(km2) %Area SWAT code the simulated and observed hydrograph (Abebe and Gebre-
Dystric fluvisols 21.85 11.31 Dystricf
mariam 2019). ENS measures the efficiency of the model by
Dystric nitisols 20.86 10.79 Dystricn
relating the goodness of fit of the simulated data to the vari-
Eutric fluvisols 25.14 13.01 Eutricf
ance of the measured data (Lin et al. 2017).
Eutric nitisols 0.02 0.01 Eutricn
Orthic acrisols 125.38 64.87 Orthica n
(qoi − qsi)2
ENS = 1 − i=1
n 2 (5)
i=1 (qoi − q̄o )

13
 Modeling Earth Systems and Environment

n  
Table 6 Sensitive flow and sediment parameters qsi − ni=1 qoi
P BIAS = i=1 
n ∗ 100 (7)
Parameters Description Range Cali- Fitted Rank i=1 qOi
value bra- values
tion
range Root mean square error to observation standard devia-
ALPHA_BF Alpha base 0–1 0–1 0.145 1 tion ratio (RSR): It is error index statistics and ranges from
flow reces- zero to one; closer to zero indicating higher accuracy of
sion constant
the model performance and one indicate a poor model per-
CANMX Maximum 0–10 0–10 2.688 2
canopy
formance (Singh et al. 2004; Vazquez-Amábile and Engel
storage 2005). RSR is calculated as shown in Eq. 8.
CN2 SCS runoff 35–98 ±25% 0.126 3 
curve n 2
number RMSE i=1 (q oi − qsi)
RSR = =  (8)
USLE_P USLE sup- 0–1 0–1 0.045 4 ST DEV ob n − 2
i=1 (q oi − s )
port practice q
factor
SLSUBBSN Average 10–150 10–150 10.319 5
slope length Where, qsi is the simulated discharge (m3/sec), qoi is the
SOL_Z Soil depth 0-3000 0–3000 0.348 6 measured discharge (m3/sec), q̄s is the average simulated
(for each
discharge (m3/sec) and q̄o is the average measured dis-
layer)
SURLAG Surface run- 0–10 0–10 9.586 7
charge (m3/sec).
off lag time
GWQMN Threshold 0-5000 0–5000 2374.738 8
depth of Results and discussion
water in
the shallow
aquifer for Stream flow calibration and validation with SWAT
return flow model
to occur
The SWAT model performance was evaluated with the
Nash-Sutcliffe efficiency (ENS) can vary between one and model performance indicators (Table 7) and visual inspec-
-∞ and performs best when a value of one is generated. Val- tion of the hydrograph between the observed and simulated
ues between 0.80 and 0.90 indicate that the model performs values (Fig. 5). In the stream flow calibration and validation
very well and values between 0.90 and 1 indicate that the process, the SWAT model performance indicators (Table 7)
model performs extremely well (Nash and Sutcliffe 1970). as R2 = 0.78, ENS = 0.76, RSR = 0.57 and PBIAS = -12.5%
Coefficient of determination (R2): R2 expresses the mea- for calibration and R2 = 0.89, ENS = 0.75, RSR = 0.44 and
sure of how well the model trends to reproduce the mea- PBIAS = 10.2% for validation. This model performance val-
sured data over a specified period of time and for a specified ues shows the SWAT model performed very well (Moriasi et
time step (Chicco et al. 2021; Legates and McCabe 1999). al. 2007; Santhi et al. 2001). The values of the uncertainty
The range of values for R2 is 1.0 (best) to 0.0 (poor correla- measure indicated that the P-factor of 0.73 and R-factor of
tion) and calculated as using Eq. 6. 0.42 for calibration and P-factor of 0.74 and R-factor of 0.52
n for validation were obtained (Table 7). This indicates that
2
2 [ i=1 (qsi
− q̄s )(qoi − q̄o )] about 73% of the calibration data and 74% of the valida-
R = n 2 n 2 (6) tion data are bracketed by the 95PPU (95% prediction of
i=1 (qsi − q̄s ) i=1 (qoi − q̄o )
uncertainty band) with a strength of 0.42 and 0.52 R-fac-
Percent bias (PBIAS): According to Gupta et al. (1999), the tor respectively and this is a tolerable level of uncertainty
tendency of predicted value greater/lesser than measured (Achamyeleh et al. 2019; Kumar et al. 2017).
one is checked by percent bias (PBIAS). For well-perform- Figure 5 shows the peak observed and simulated stream
ing model, the absolute value of PBIAS is minimum (Hugo flow obtained in August 1998 for calibration period and also
et al. 2019) and given by. the peak observed and simulated stream flow obtained in

Table 7 Results of SWAT model performance indicators for observed and simulated monthly stream flow calibration and validation
Gauging Stations Simulation period Uncertainty measures Model performance indicators
P- factor R- factor R2 ENS RSR PBIAS
Outlet of watershed Calibration (1990–2001) 0.73 0.42 0.78 0.76 0.57 -12.5
Validation (2002–2007) 0.74 0.52 0.89 0.75 0.44 10.2

13
Modeling Earth Systems and Environment

Fig. 5 Monthly observed and simulated hydrographs for calibration and validation period

Table 8 SWAT model performance indicators for observed and simulated sediment
Gauging Stations Simulation period Uncertainty measures Model performance indicators
P-factor R-factor R2 ENS RSR PBIAS
Outlet of watershed Calibration (1990–2001) 0.78 0.46 0.79 0.73 0.45 11.4
Validation (2002–2007) 0.75 0.35 0.87 0.72 0.52 16.5

September 2004 for validation period. In general, the hydro- is due to the stream flow and sediment yield are directly
graph (Fig. 5) showed that the SWAT model underestimated proportional to each other (Duguma et al. 2020; Kero et al.
the stream flow in some years and overestimated in most of 2023). Additionally, the calibrated and validated sediment
the year for both calibration and validation processes. yield hydrograph (Fig. 6) has shown that the model under-
estimated in some years and overestimated sediment yield
Calibration and validation of sediment yield with in most of the years. This is due to using uncertainty analy-
SWAT model sis, about 22% and 25% of calibration and validation data,
respectively, were uncertain. The uncertainty may be from
Simulation of sediment yield using SWAT model carried the model input data, the model itself, and/or model param-
out through calibration and validation process (Boufala et eters in SUFI2 algorithm (Achamyeleh et al. 2019).
al. 2022; Kumar et al. 2017). The calibration and validation The quantified total volume of annual sediment yield
result of sediment (Table 8) indicated that very good model after calibration and validation at the outlet of Hare water-
performance level with R2 = 0.79, ENS =0.73, RSR = 0.45 shed were 24,368.10 tons/year (observed) and 24,824.08
and PBIAS = 11.4% for calibration. Also, R2 = 0.87, ENS = tons/year (simulated). This result shows a large volume of
0.72, RSR = 0.52, and PBIAS = 16.5% for validation (Mori- sediment load produced from the study watershed, espe-
asi et al. 2007; Santhi et al. 2001). The uncertainty measures cially from upland watershed of study watershed and needs
showed that about 78% of the calibrated sediment data and watershed intervention techniques (Temesgen et al. 2022).
75% of validated sediment data bracketed by the 95% pre- Before developing watershed intervention techniques,
diction of uncertainty band, and it is also a tolerable level of evaluating the spatial delivery of sediment yield for each
uncertainty (Achamyeleh et al. 2019; Kumar et al. 2017). sub-watershed level and classifying the sediment hotspot
The performance of the model in the sediment calibration sub-watershed was undertaken (Gebrekidan et al. 2020).
and validation process was tested with the statistical indica-
tors and the result was illustrated in Table 8.
Similar to stream flow, Fig. 6 shows the peak observed
and simulated sediment yield occurred in August 1998
for calibration and in September 2004 for validation. This

13
 Modeling Earth Systems and Environment

Fig. 6 Observed and simulated monthly sediment hydrographs for calibration and validation

Table 9 Average annual sediment production variability of each sub- 3, 12 and 14 were very high; sub-watersheds 9, 17, 18, 4 and
watershed 15 were high. Moreover, sub-watersheds 2, 16, 8, 1, and 11
Sub-watershed Sediment pro- Sub-watershed Sediment were falling into moderate and sub-watersheds 13 and 19
duction (ton/ production
ha/year) (ton/ha/year) were classified as low sediment producing sub-watersheds.
1 14.41 11 13.61
2 17.43 12 36.65 Watershed intervention techniques
3 42.14 13 10.87
4 21.68 14 33.03 Application of watershed intervention techniques for the
5 43.38 15 20.38 whole watershed is uneconomical, time consuming and
6 60.62 16 16.71 impractical (Subhasis et al. 2012; Temesgen et al. 2022;
7 54.84 17 25.07 Tesfu and Olkeba 2020). Thus, a prior point before the
8 15.10 18 21.72 development of intervention techniques is the identifica-
9 27.27 19 3.54
tion of sediment hotspot sub-watersheds. Out of the total
10 61.99
19 sub-watersheds, 12 sub-watersheds which produce sedi-
ment yield from high to severe with an average sediment
Modeling spatial distribution of sediment yield in production of 37.39 tons/ha/year were identified and had
sub-watershed level 71.23% of watershed area coverage. Hence, these identi-
fied sediment hotspot sub-watersheds need the application
Sediment yield spatial distribution in each sub-watershed of suitable watershed management practices (Gashaw et al.
was not uniform that’s why the governing issues, for exam- 2020; Gebrekidan et al. 2020). In this study, four watershed
ple, soil erodibility, rainfall distribution, land use/cover, intervention techniques (scenarios) such as scenario I (Strip
topography, and type of soil (Belayneh et al. 2021). For cropping), scenario II (Residue management), scenario III
each sub-watershed, the spatial variability of sediment yield (Terracing), and scenario IV (Contouring) were developed.
delivery was illustrated in Table 9. Baseline scenario: It was assumed to reflect the current
Using the average annual sediment yield result (Table 9), land management practices without conservation measures
the sediment source map (Fig. 7) was generated based on the and used as a reference for comparison of the effective-
sediment yield potential. According to Gashaw et al. (2017), ness of the developed watershed intervention techniques
the sediment load classification is severe (≥ 50 tons/ha/year), (Gebrekidan et al. 2020). In the baseline scenario, twelve
very high [30–50 tons/ha/year), high [18–30 tons/ha/year), hotspot sub-watersheds were identified for simulation of
moderate [11–18 tons/ha/year) and low [0–11 ton/ha/year). four selected watershed intervention techniques (scenarios).
Based on the above classification, the sediment yield spa- Each scenario was then run annually for the same simula-
tial distribution in Hare watershed was classified as follows. tion period to provide a consistent basis for comparison of
Sub-watersheds 10, 6 and 7 were severe; sub-watersheds 5, the scenario results.

13
Modeling Earth Systems and Environment

Fig. 7 Sediment yield spatial dis-

tribution map of Hare watershed

Scenario-I (Strip cropping): According to USDA-NRCS increases water infiltration, and reduces water runoff. Both
(2011), erosion-resisting crops were planted (strip cropping) the soil residue flagged (SO_RES_FLAG) for conservation
across the slope of the field to arrest and infiltrate upcom- practice and soil residue (SO_RES) were adjusted and sim-
ing soil erosion. In this study, strip cropping was simulated ulated in SWAT annually to understand the effect of residue
annually with SWAT model by adjusting the Manning’s management (Gamachu et al. 2017; Onduru et al. 2002).
value (STRIP_N) for overland flow in the strip cropped Scenario-III (terracing): The construction of an earthen
fields to represent increased surface roughness in the direc- embankments (terracing) across the slope of the agricul-
tion of runoff and SCS curve number II (STRIP_CN) to tural land decreases the slope length and steepness that in
account for increased infiltration. In addition, USLE crop- turn reduces soil erosion and sediment production in the
ping factor (STRIP_C) for strip-cropped fields to reflect the watershed (Arnold et al. 2012; Chuxiong et al. 2021). For
average value for multiple crops within the field and the this study, terracing parameters such as USLE practice
USLE practice factor (STRIP_P) adjusted and run annually (TERR_P), factor curve number (TERR_CN), and slope
by SWAT for strip-cropped fields to represent the effects of length (TERR_SL) were adjusted and then simulated annu-
strip-cropping conditions on soil erosion and sediment yield ally by SWAT (Kero et al. 2021; Mwangi et al. 2015).
(Duguma et al. 2020; Hamza et al. 2019). Scenario-IV (contouring): Contours are oriented at the
Scenario-II (Residue management): Residue manage- right angle to the field slope at any point (Pyry et al. 2017).
ment controls the amount and distribution of crop residue The application of contouring in farmland increases the
on the soil surface (USDA 1996). Covering crop residue on surface roughness; it reduces the surface runoff and sedi-
cultivation land can provide protection of soil from rain ero- ment production (Duguma et al. 2020; Hamza et al. 2019).
sion, adds organic matter to the soil, conserves soil moisture, In this research work, two contour parameters such as curve

13
 Modeling Earth Systems and Environment

Table 10 Developed watershed intervention techniques with relation Baseline condition shows the existing watershed charac-
to baseline conditions
teristics without application of intervention techniques and
Sediment Watershed intervention techniques
hotspot Baseline Strip Residue Con- Ter-
used as a reference point to understand the impact of inter-
sub-watersheds condition cropping manage- tour- rac- vention techniques (Boufala et al. 2022; Gebrekidan et al.
ment ing ing 2020). After application of strip cropping, residue manage-
10 62.00 20.49 57.94 12.39 3.76 ment, contouring, and terracing, the sediment yield produc-
6 60.62 21.00 55.39 11.94 4.45 tion in the Hare watershed is 52.58%, 94.46%, 30.78%, and
7 54.85 20.27 53.46 12.26 3.43 15.12% respectively (Table 10). This shows that there can
5 43.38 18.68 41.10 10.93 2.57 be sediment yield reduction by 47.42%, 5.54%, 69.22%,
3 42.14 21.63 37.06 12.97 4.04
and 84.88% after application of strip cropping, residue man-
12 36.65 25.13 34.83 13.65 5.79
agement, contouring and terracing, respectively.
14 33.03 29.86 33.03 11.44 18.65
As shown in Fig. 8, the sediment yield reduction with
9 27.27 5.95 26.68 14.57 8.91
17 25.07 20.47 22.41 12.10 2.62
strip cropping at sub-watershed 9 was better than the other
18 21.72 19.40 21.45 11.23 2.42 intervention techniques. This indicated that the sub-water-
4 21.68 16.98 20.16 9.68 1.57 shed 9 have available strip plants, better surface roughness,
15 20.38 16.09 20.38 4.98 9.55 and increased infiltration rate (Hamza et al. 2019). The sedi-
Average sedi- 37.39 19.66 35.33 11.51 5.65 ment yield reduced from 37.39ton/ha/year to 19.66ton/ha/
ment yield (ton/ year (47.42%) after application of strip cropping (Table 10)
ha/year)
and the result is almost agreed with Duguma et al. (2020).
Sediment pro- 100 52.58 94.46 30.78 15.12
duction (%)
Sediment yield reduction at sub-watershed 14 and 15 was
Sediment reduc- 0 47.42 5.54 69.22 84.88 better with the treatment of contouring than the others
tion (%) (Fig. 8). After application of contour, the sediment yield
reduced by 69.22% (37.39ton/ha/year to 11.51ton/ha/year)
number (CONT_CN) and USLE practice factor (CONT_P) (Table 10) and this value is relatively comparable to those
were adjusted and simulated annually with SWAT model to of Duguma et al. (2020). Treatment of soil erosion hotspot
realize the effects of contouring in the sediment production. region with contour mainly depends on the slope, eleva-
After application of each watershed intervention tech- tion and farming system across the contour line (Pyry et al.
niques, the simulated Sed_Out exported to the Microsoft 2017). To this regard, the sub-watersheds 14 and 15 gave a
Excel sheet and the average annual sediment yield calcu- better response to the application of contour (Fig. 8). The
lated for each sub-watershed. The developed watershed slight amount of sediment reduction was observed after
intervention techniques and their sediment reduction in application of residue management in all sediment hotspot
relation to baseline/existing conditions were illustrated in sub-watersheds (Fig. 8). This is due to residue management
Table 10. mainly protects wind erosion rather than water erosion to
cultivated land (Gamachu et al. 2017; Onduru et al. 2002).

Fig. 8 Sediment mitigation measures for sediment hotspot sub-watersheds

13
Modeling Earth Systems and Environment

Thus, residue management is inappropriate intervention cover, soil, sediment production rate were visibly shown in
technique for the soil erosion hotspot areas in the study the following Fig. 9; Table 11.
watershed. In this study, effective soil erosion and sediment
yield reduction was revealed after application of terracing in
entire hotspot sub-watersheds except sub-watersheds 14, 9, Conclusion
and 15 (Fig. 8). About 84.88% (37.40ton/ha/year to 5.65ton/
ha/year) of sediment yield reduction was shown after appli- Watershed intervention techniques can play a significant
cation of terracing in the sediment hotspot sub-watersheds role in reduction of soil erosion, land degradation, and sedi-
(Table 10), and the result is similar to those of Mwangi et ment yield in the watershed. In this study, Soil and Water
al. (2015). Therefore, terracing is suggested to be employed Assessment Tool (SWAT) model was used to identify soil
in all hotspot sub-watersheds excluding sub-watersheds 14, erosion and sediment yield hotspot areas and to select effec-
9 and 15 to solve the soil loss and sediment problems in the tive watershed intervention techniques for environmental
study watershed. The research findings are very important to protection. Hare watershed delineated at the outlet point
governmental and non-governmental organizations, police with the total area of 187.14 km2. Thirty years climate data
and decision makers, researchers, Gamo zone and Chencha (1987–2016) was used for the simulation of SWAT model.
woreda environmental protection authority, and local farm- Twenty four field sediment sample data were collected and
ers to reduce adverse future soil loss and sediment impacts analyzed in the laboratory and then a sediment rating curve
in the study watershed. was developed. Eighteen years daily sediment data was
For implementation purpose, a clear illustration of water- generated using the developed sediment rating curve. The
shed intervention techniques with their respective sediment watershed controlling flow and sediment parameters which
hotspot sub-watersheds map, areas, locations, land use/ played a great role in the calibration process were deter-
mined through model sensitivity analysis. The calibration
Fig. 9 Overlaid map of sedi-
ment hotspot sub-watersheds
with watershed intervention
techniques

13
 Modeling Earth Systems and Environment

Table 11 Identified twelve sediment hotspot sub-watersheds locations, land use/cover, soil, sediment production and watershed intervention tech-
niques
Sub-watershed Area Location (Kebele) Dominant land Dominant soil type Sediment pro- Watershed
(km2) use/cover duction (ton/ inter-
ha/year) vention
techniques
10 13.99 Malida, Olootsaritae, Amaranbodo AGRR dystricf, dystricn, orthica 61.99 Terracing
6 4.02 Genido and Zalo FRST and AGRR dystricn, orthica and dystricf 60.62 Terracing
7 25.48 Gena Kera, Tiba, Dogema and AGRR dystricf, dystricn and orthica 54.85 Terracing
Webera
5 18.94 Shaye, Losha, Kale, Shale, Meshe AGRR dystricf, dystricn and orthica 43.38 Terracing
and Dalonazara
3 1.32 Upper Dalonazara FRST and AGRR dystricn, orthica and dystricf 42.14 Terracing
12 15.69 Shama, Gulitaguye, Upper Bele and AGRR dystricf, dystricn and orthica 36.65 Terracing
Upper Kola Shara
14 6.39 Tegecha, and Dega Ocholo AGRR, RNGB orthica 33.03 Contouring
9 5.07 Elu and Godiye FRST and AGRR orthica and dystricn 27.27 Strip
cropping
17 5.33 Lower Kola Shara AGRL, AGRR, eutricf, orthica, dystricn 25.07 Terracing
RNGB
18 27.97 Kola Shara, Lower Dega Ocholo, AGRL, AGRR eutricf, orthica and dystricn 21.72 Terracing
Chano Doriga and Chano Cheliba and RNGB
4 4.51 Denibo AGRR Dystricf, orthica 21.68 Terracing
15 4.57 Bele AGRR orthica 20.38 Contouring
Note: AGRR = moderately cultivated, AGRL = intensively cultivated, FRST = forest, RNGB = shrub land, and CABG = riparian vegetation
and validation were carried out from 1990 to 2001 (12 years) to rehabilitate the soil erosion hotspot areas and to sustain
and 2002–2007 (six years) respectively, on a monthly basis the future land and water resources in the study watershed.
of stream flow and sediment data in SWAT- CUP. The model
efficiency was tested with percent bias (PBIAS), Nash-Sut- Acknowledgements Our thanks go to Arba Minch University Water
Resources Research Center for providing support during the field and
cliffe model efficiency (ENS), root mean square error (RSR), laboratory data analysis.
and coefficient of determination (R2). In the discharge and
sediment yield calibration and validation process, the val- Funding The research was supported by Arba Minch University, Wa-
ues of ENS, R2, RSR and PBIAS were vary in the ranges of ter Resources Research Center.
0.72–0.76, 0.78–0.89, 0.44–0.57 and − 12.5 - +16.5 respec-
tively. The above results showed that for both flow and Data Availability The collected, analyzed data and the finding of this
research are available from the corresponding author upon rational
sediment simulation, the SWAT model performance is very request.
well. Quantified total amount of sediment delivery from the
of Hare watershed into Lake Abaya were 24,368.10 tons/ Code Availability (software application or custom code) Not appli-
year (observed) and 24,824.08 tons/year (simulated). Spa- cable.
tial sediment production rate in the watershed ranges from
3.54 to 61.99 tons/ha/year. From the total 19 sub-water- Declarations
sheds, twelve sediment hotspot areas (sub-watersheds) were
Conflict of interest There is no conflict of interest between the authors.
identified and covered 71.23% of the total watershed area.
Watershed intervention techniques (mitigation measures)
such as strip cropping, residue management, contouring,
and terracing were considered to understand the soil loss References
and sediment reduction capability. After application of strip
Abbaspour KC, Rouholahnejad E, Vaghefi SR, Srinivasan R, Yang H,
cropping, residue management, contouring, and terracing, Klove B (2015) A continental-scale hydrology and water quality
the soil loss and sediment production reduced by 47.42%, model for Europe: calibration and uncertainty of a high-resolu-
5.54%, 69.22% and 84.88% respectively. From the soil ero- tion large-scale SWAT model. J Hydrol 524:733–752. https://doi.
org/10.1016/j.jhydrol.2015.03.027
sion and sediment intervention techniques, terracing was
Abebe TA (2019) Calibration, validation and performance evaluation
found to be the best watershed intervention approach in this of SWAT model for sediment yield modeling in megech reservoir
study. The result of this work is valuable for environmen- watershed, Ethiopia. 12(3–4):21–23. https://doi.org/10.2478/jen-
tal protection authority, local government and community geo-2019-0009. Journal of Environmental Geography

13
Modeling Earth Systems and Environment

Abebe T, Gebremariam B (2019) Modeling runoff and sediment yield Chuxiong D, Guangye Z, Yaojun L, Xiaodong N, Zhongwu L et al
of Kesem dam watershed, awash basin, Ethiopia. SN Appl Sci (2021) Advantages and disadvantages of terracing: a compre-
1:446. https://doi.org/10.1007/s42452-019-0347-1 hensive review. Int Soil Water Conserv Res 9(3):344–359 ISSN
Achamyeleh GM, Leon DR, Yali EW (2019) Techniques for calibra- 2095–6339. https://doi.org/10.1016/j.iswcr.2021.03.002
tion and validation of SWAT model in data scarce arid and semi- Degife A, Worku H, Gizaw S (2021) Environmental implications of
arid catchments in South Africa. J Hydrology: Reg Stud. https:// soil erosion and sediment yield in Lake Hawassa watershed,
doi.org/10.1016/j.ejrh.2019.100621 south-central Ethiopia. Environ Syst Res 10:28. https://doi.
Amanuel K (2018) Transforming Konso towards Green Economy org/10.1186/s40068-021-00232-6
through Integrated Land Management. GRIN Verlag, Munich. Deshmukh SS, Wayal AS (2019) Sediment Yield Estimation Using
https://www.grin.com/document/437812 RS and GIS for Upper Karha Watershed Maharashtra of India.
Arnold JG, Kiniry JR, Srinivasan R, Williams JR, Haney EB, Neitsch https://doi.org/10.1007/s40030-018-00355-7
SL (2012) Soil and Water Assessment Tool: Input/ output docu- Dessie M, Verhoest2 NE, Pauwels VR, Admasu T, Poesen J, Adgo E et
mentation. Version TR-439. Texas Water Resources Institute, al (2014) Analyzing runoff processes through conceptual hydro-
College Station, pp 1–650 logical modeling in the Upper Blue Nile Basin, Ethiopia. Jour-
Assefa E, Hans B (2014) Long-term indigenous soil Conservation nal of Hydrol. Earth Syst. Sci., 18, 5149–5167, 2014. https://doi.
Technology in the Chencha Area, Southern Ethiopia: origin, char- org/10.5194/hess-18-5149-2014
acteristics. Sustain Ambio 43(7):932–942 PMC4190142. https:// Dinesh B, Rajeev J, Raju RR, Nripesh A (2021) Impact on Agricul-
doi.org/10.1007/s13280-014-0527-6 tural Productivity by Using the RMMF Model and Local Percep-
Balabathina V, Raju RP, Mulualem W (2019) Integrated remote sens- tion: for Rangun Watershed of Mid-Hills, Nepal, Applied and
ing and GIS-based universal soil loss equation for soil erosion Environmental Soil Science, ID 5747138, 10 pages. https://doi.
estimation in the Megech River Catchment, Tana Lake Sub- org/10.1155/2021/5747138
basin, Northwestern Ethiopia. Am J Geogr Inf Syst 8(4):141–157. Duguma TA, Wakigari SA, Dilgasa EA (2020) Analysis of ambo water
https://doi.org/10.5923/j.ajgis.20190804.01 supply diversion weir sedimentation and assessing impact of
Belayneh Y, Donghong X, Baojun Z, Yong Y, Muhammad AB et al land management practice through hydrological studies. J Water
(2021) Spatial distribution of soil erosion and sediment yield in Resource. https://doi.org/10.1007/s40899-020-00455
the Koshi River Basin, Nepal: a case study of Triyuga watershed. Dulo H, Abate B (2020) Estimation of runoff and sediment yield using
J Soils Sediments. https://doi.org/10.1007/s11368-021-03023-9 SWAT model: the case of Katar Watershed, Rift Valley Lake
Beshah T (2003) Understanding farmers: explaining Soil and Water Basin of Ethiopia. Int J Mech Eng Appl 8(6):125–134. https://doi.
Conservation in Konso, Wolaita and Wello, Ethiopia, vol ISBN org/10.11648/j.ijmea.20200806.11
9789058087959–245. Tropical Resource Management Papers Ellison CA, Savage BE, Johnson GD (2014) Suspended-sediment
Documents concentrations, loads, total suspended solids, turbidity, and par-
Betrie GD, Mohamed YA, Griensven AV, Srinivasan R (2011) Sedi- ticle-size fractions for selected rivers in Minnesota, 2007 through
ment management modelling in the Blue Nile Basin using SWAT 2011: U.S. Geological Survey Scientific Report format 2013–
model. Journal of Hydrol. Earth Syst. Sci., 15, 807–818, 2011. 5205, 43pp. https://doi.org/10.3133/sir20135205
https://doi.org/10.5194/hess-15-807-2011 Gamachu A, Shimelis B, Olkeba T (2017) Assessing the effect of soil
Biniyam Y, Kemal A (2017) The impacts of Climate Change on Rain- and water conservation practices on runoff and sediment yield
fall and Flood frequency: the case of Hare Watershed, Southern on Hunde Lafto watershed of Upper Wabi Shebelle Basin. Civ
Rift Valley of Ethiopia. J Earth Sci Clim Change 08(01). https:// Environ Res. 9(9). ISSN 2224–5790
doi.org/10.4172/2157-7617.1000383 Gashaw T, Tulu T, Argaw M (2017) Erosion risk assessment for priori-
Borrelli P, Robinson DA, Fleischer LR (2017) An assessment of the tization of conservation measures in Geleda watershed, Blue Nile
global impact of 21st century land use change on soil erosion. basin, Ethiopia. Environ Syst Res 6:1. https://doi.org/10.1186/
NatCommun 8. https://doi.org/10.1038/s41467-017-02142-7 s40068-016-0078-x
Boufala MH, El Hmaidi A, Essahlaoui A, Chadli K, El Ouali A, Lah- Gashaw T, Worqlul AW, Dile YT, Addisu S, Bantider A, Zeleke G
jouj A (2022) Assessment of the best management practices under (2020) Evaluating potential impacts of land management prac-
a semi-arid basin using SWAT model (case of M’dez watershed, tices on soil erosion in the Gilgel Abay watershed, upper Blue
Morocco). Model Earth Syst Environ 8:713–731 Nile basin. Heliyon 24(8):e04777. https://doi.org/10.1016/j.heli-
Cerqueira V, Torgo L, Mozetič I (2020) Evaluating time series fore- yon.2020.e04777
casting models: an empirical study on performance estimation Gebrekidan W, Ermias T, Amare B, Yihun TD (2020) Prioritization
methods. Mach Learn 109:1997–2028. https://doi.org/10.1007/ of watershed management scenarios under climate change in
s10994-020-05910-7 the Jemma sub-basin of the Upper Blue Nile Basin, Ethiopia.
Chandra P, Patel PL, Porey PD, Gupta ID (2014) Estimation of J Hydrology: Reg Stud 31(100714):2214–5818. https://doi.
sediment yield using SWAT model for Upper Tapi basin. ISH J org/10.1016/j.ejrh.2020.100714
Hydraulic Eng 20:3 291–300. https://doi.org/10.1080/09715010 Getahun GW, Zewdu EA, Mekuria AD (2021) Impacts and uncer-
.2014.902170 tainties of climate change on stream flow of the Bilate River
Cheng Y, Morgan DL (2013) Classification accuracy and consistency (Ethiopia), using a CMIP5 general circulation models ensem-
of computerized adaptive testing. Behav Res 45, 132–142 (2013). ble. Int J Water Resour Environ Eng. https://doi.org/10.5897/
https://doi.org/10.3758/s13428-012-0237-6 IJWREE2020.0973
Cheng KS, Lien YT, Wu YC et al (2017) On the criteria of model Gupta HV, Sorooshian S, Yapo PO (1999) Status of automatic cali-
performance evaluation for real-time flood forecasting. Stoch bration for hydrologic models: comparison with multilevel
Environ Res Risk Assess 31:1123–1146. https://doi.org/10.1007/ expert calibration. J Hydrologic Eng 4(2):135–143. https://doi.
s00477-016-1322-7 org/10.1061/(ASCE)1084-0699(1999)4:2(135)
Chicco D, Warrens MJ, Jurman G (2021) The coefficient of determina- Hailu KA, Stefan S, Feras Z, Nigus DM, Andreas K (2016) Modeling
tion R-squared is more informative than SMAPE, MAE, MAPE, streamflow and sediment using SWAT in Ethiopian Highlands.
MSE and RMSE in regression analysis evaluation. PeerJ Comput Int J Agric & Biol Eng 9(5):51–66 Open Access at. http://www.
Sci 5:7:e623. https://doi.org/10.7717/peerj-cs.623 ijabe.org

13
 Modeling Earth Systems and Environment

Hamza B, Rachid M, Rachid M, Khadija A (2019) Use of a cali- Nijat M, David E, Peter F, Peter L (2016) Evaluating Forecasting
brated SWAT model to evaluate the effects of agricultural BMPs Methods by Considering Different Accuracy Measures. Proce-
on sediments of the Kalaya river basin (North of Morocco). Int dia Computer Science, Volume 95, Pages 264–271, ISSN 1877
Soil Water Conserv Res 7(2):176–118. https://doi.org/10.1016/j. – 0509, https://doi.org/10.1016/j.procs.2016.09.332
iswcr.2019.02.002 Onduru DD, Gachimbi L, Maina F, Muchena FN, De Jager A (2002)
Hugo HC, Adriana MC, João HM, Azeneth S et al (2019) Hydrologic Sustaining agricultural production in the semi-arid areas of east-
modeling for sustainable Water Resources Management in Urban- ern Kenya: a case study of Mbeere District. INMASP Report No.
ized Karst Areas. Int J Environ Res Public Health 16(14):2542. Ke-03. ETC-EA, KARI (NARL) and LEI-DLO
https://doi.org/10.3390/ijerph16142542 Opeyemi OA, Abidemi FH, Victor OK (2019) Assessing the impact of
ITPS IT (2015) Status of the food and agricultural organization of the soil erosion on residential areas of Efon-Alaaye Ekiti, Ekiti-State,
United States Nigeria. Int J Environ Plan Manag 5(9). http://www.aiscience.
Kassa T, Gerd F (2015) Impact of Land Use / Cover Change on Stream org/journal/ijepm
flow: The Case of Hare River Watershed, Rift valley basin, Ethio- Pyry K, Christian K, Juha O (2017) A design of contour generation
pia. https://www.researchgate.net/publication/266335014 for topographic maps with adaptive DEM smoothing. Int J Car-
Kero AA, Abdulkerim BS, Dawit YM (2023) Sediment yield tography 3(1):19–30. https://doi.org/10.1080/23729333.2017.13
responses to land use land cover change and developing best 00998
management practices in the upper Gidabo dam watershed. Sus- Raffa CM, Chiampo F (2021) Bioremediation of Agricultural Soils
tainable Water Resources Management. https://doi.org/10.1007/ Polluted with Pesticides: a review. Special issue in Bioengineer-
s40899-023-00850-1 ing and Remediation of Polluted Environments 8(7):92. https://
Khaleghi MR, Varvani J (2018) Sediment rating curve parameters doi.org/10.3390/bioengineering8070092
relationship with Watershed characteristics in the Semiarid Roostaee M, Deng Z (2020) Effects of Digital Elevation Model
River Watersheds. Arab J Sci Eng 43:3725–3737. https://doi. Resolution on Watershed-Based Hydrologic Simulation. Water
org/10.1007/s13369-018-3092-7 Resource Manage 34:2433–2447. https://doi.org/10.1007/
Kumar N, Singh SK, Srivastava PK, Narsimlu B (2017) SWAT model s11269-020-02561-0
calibration and uncertainty analysis for stream flow prediction of Samaneh M, Farshad K, Mojgansadat A, Farhad K (2016) Using
the tons River Basin, India, using sequential uncertainty fitting SWAT model to determine runoff, sediment yield and nitrate
(SUFI-2) algorithm. Model Earth Syst Environ 3(1):1–13. https:// loss in Gorganrood Watershed. Iran Ecopersia 4(2):1359–1377.
doi.org/10.1007/s40808-017-0306-z https://doi.org/10.18869/modares.Ecopersia.4.2.1359
Kuti IA, Ewemoje TA (2021) Modeling of sediment yields using the Santhi C, Arnold JG, Williams JR, Dugas WA, Srinivasan R, Hauck LM
soil and water assessment tool (SWAT) model: a case study of (2001) Validation of the SWAT model on a large river basin with
the Chanchaga watersheds. Nigeria” 13. https://doi.org/10.1016/j. point and nonpoint sources. J Am Water Resour Assoc 37:1169–
sciaf.2021.e00936 1188. https://doi.org/10.1111/j.1752-1688.2001.tb03630.x
Legates DR, McCabe GJ (1999) Evaluating the use of goodness-of- Sao D, Kato T, Tu LH, Thouk P, Fitriyah A, Oeurng C (2020) Evalua-
fit measures in hydrology and hydro climatic model validation. tion of different objective functions used in the SUFI-2 calibration
Water Resour Res 35(1):233–241 process of SWATCUP on water balance analysis: a case study of
Lin F, Xingwei C, Huaxia Y (2017) Evaluating the use of Nash- the Pursat River basin. Cambodi Water 12(10):2901. https://doi.
Sutcliffe efficiency coefficient in goodness-of-fit measures for org/10.3390/w12102901
daily runoff simulation with SWAT. J Hydrol Eng. https://doi. Scott-Shaw BC, Hill TR, Gillham JS (2020) Calibration of model-
org/10.1061/(ASCE)HE.1943-5584.0001580 ing approach for sediment yield in a wattle plantation, Kwa-
Mithlesh K, Ambika PS, Narayan S, Sonam SD, Sanjay KR, Balram P Zulu-Natal, South Africa. Water SA 46:171–181. https://doi.
(2022) Global-scale application of the RUSLE model: a compre- org/10.17159/wsa/2020.v46.i2.8232
hensive review. Hydrol Sci J 67(5):806–830. https://doi.org/10.1 Semu AA (2018) The impact of soil and water conservation for
080/02626667.2017.1404068 improved agricultural production in Ethiopia. J Agric 1:9–12.
Molla T, Sisheber B (2017) Estimating soil erosion risk and evaluating https://tinyurl.com/y5cc9yw2
erosion control measures for soil conservation planning at Koga Shrestha G, Peter DS (2008) Carbon accumulation and storage in
watershed in the highlands of Ethiopia. Solid Earth 8:13–25. semi-arid sagebrush steppe: Effects of long-term grazing exclu-
https://doi.org/10.5194/se-8-13-2017 sion. Agriculture, Ecosystems and Environment 125 (2008) 173–
Moriasi DN, Arnold JG, Liew M, Bingner RL, Harmel RD, Veith TL 181. https://doi.org/10.1016/j.agee.2007.12.007
(2007) Model evaluation guidelines for systematic quantification Singh J, Knapp HV, Demissie M (2004) Hydrologic modeling of the
of Accuracy in Watershed Simulations. Trans ASABE 50:885– Iroquois River watershed using HSPF and SWAT. ISWS CR
900. https://doi.org/10.13031/2013.23153 2004-08. Champaign, Ill.: Illinois State Water Survey. www.sws.
Morris G, Fan J (1998) Reservoir Sedimentation Handbook. McGraw- uiuc.edu/pubdoc/CR/
Hill Book Co, New York, p 805 Sofonyase B, Abate B (2020) Estimation of sediment yield using Swat
Mwangi JK, Shisanya CA, Gathenya JM, Namirembe S, Moriasi DN Model: a case of Soke River Watershed, Ethiopia. Int J Eng Res
(2015) A modeling approach to evaluate the impact of conserva- Technol. Vol. 9 Issue 12. IJERTV9IS120252
tion practices on water and sediment yield in Sasumua Water- Subhasis G, Pouyan AN, Sean WN, Zhen Z (2012) Analysis of best
shed, Kenya. J Soil Water Conserv 70(2):75–90. https://doi. management practice effectiveness and spatiotemporal variabil-
org/10.2489/jswc.70.2.75 ity based on different targeting strategies. Journal of hydrologi-
Nash JE, Sutcliffe JV (1970) River flow forecasting through con- cal process. Volume 28, 431–445 (2014). https://doi.org/10.1002/
ceptual models: part 1.A discussion of principles. J Hydrology hyp.9577
10(3):282–290 Tamiru H, Wagari M (2021) Spatio-temporal analysis of sediment
Neitsch SL, Arnold JR, Kiniry JR, Williams JR (2011) Soil and yield estimation using Integrated RUSLE and GIS technique:
Water Evaluation Tool (SWAT) Theoretical Documentation Ver- case of Nashe Dam Reservoir, Abay Basin, Ethiopia. Preprints
sion2009. Texas Water Resources Institute Technical Report 2021020480. https://doi.org/10.20944/preprints202102.0480.v1
No.406, Texas A&M University System, College Station, Texas Temesgen K, Abdisa T, Chelkeba BT (2022) Prioritization of sus-
ceptible watershed to sediment yield and evaluation of best

13
Modeling Earth Systems and Environment

management practice: a case study of Awata River, Southern Woldemariam G, Iguala A, Tekalign S, Reddy R (2018) Spatial mod-
Ethiopia, Applied and Environmental Soil Science. Article ID eling of soil erosion risk and its implication for conservation
1460945:16pages. https://doi.org/10.1155/2022/1460945 planning: the case of the Gobele watershed, East Hararghe Zone,
Tesfu TA, Olkeba TL (2020) Sediment yield estimation and effect of Ethiopia. Land 7(1):25. https://doi.org/10.3390/land7010025
Management Options on Sediment Yield of Kesem Dam Water- Yang H, Li B, Zhang C, Qiao H, Liu Y, Bi J, Zhang Z, Zhou F (2020)
shed, Awash River Basin, Ethiopia. https://doi.org/10.1016/j. Recent spatio-temporal variations of suspended sediment con-
sciaf.2020.e00425. Scientific African centrations in the Yangtze Estuary. Water 12(3):818. https://doi.
Tran VT, Tuong (2014) An Analysis of The Suspended Sediment Rat- org/10.3390/w12030818
ing Curve Parameters In The Upper Mississippi River Basin At Yesuf HM, Assen M, Alamirew T, Melesse AM (2015) Modeling of
The Monthly And Annual Levels. https://Trace.Tennessee.Edu/ sediment yield in Maybar gauged watershed using SWAT north-
Utk_Gradthes/2854 east Ethiopia. CATENA 127:191–205. https://doi.org/10.1016/j.
Tripathi MP, Panda RK, Raghuwanshi NS (2004) Development of catena.2014.12.032
effective management plan for critical sub watersheds using Yisehak B (2021) Prediction of flood frequency under a changing
SWAT model. Department of Soil and Water Engineering Faculty climate, the case of Hare watershed, Rift Valley Basin of Ethio-
of Agricultural Engineering Indira Gandhi Agricultural Univer- pia. Sustain Water Resour Manag 7:9. https://doi.org/10.1007/
sity Raipur 492:006. https://doi.org/10.1002/hyp.5618 s40899-021-00492-1
Tsegaye L, Bharti R (2021) Soil erosion and sediment yield evalua- Yoseph M, Aster G, Getahun HF (2019) Farmers’ indigenous knowl-
tion using RUSLE and GIS-based approach in Anjeb watershed. edge of tree conservation and acidic soil amendments: the role of
Northwest Ethiopia SN Appl Sci 3:582. https://doi.org/10.1007/ “baabbo” and “Mona” systems: Lessons from Gedeo community,
s42452-021-04564-x Southern Ethiopia. Cogent Food & Agriculture 5:1. https://doi.
USDA-NRCS (United States Department of Agriculture- Natural org/10.1080/23311932.2019.1645259
Resources Conservation Service) (2011) Strip Cropping. Avail- Zakwan M (2019) Comparative analysis of the novel infiltration model
able at: http://www.sera17.ext.vt.edu/Documents/BMP_strip_ with other infiltration models. Water and environmental Journal
cropping.pdf page 1–13. https://doi.org/10.1111/wej.12435
USDA (United States Department of Agriculture) (1996) AREI/ Pro- Zerihun M, Mohammedyasin MS, Sewnet D, Adem AA, Lakew M
duction management. Available at: http://www.ers.usda.gov/pub- (2018) Evaluation of soil erosion using RUSLE, GIS and remote
lications/arei/ah712/AH7124-2.pdf sensing in Nort west parts of Ethiopia. Geoderma Reg 12:83–90.
Vazquez-Amábile GG, Engel BA (2005) Use of SWAT to compute https://doi.org/10.1016/j.geodrs.2018.01.002
groundwater table depth and stream flow in the Muscatatuck Zerssa G, Feyssa D, Kim DG, Eichler-Löbermann B (2021) Chal-
River watershed. Trans ASAE 48(3):991–1003 lenges of Smallholder Farming in Ethiopia and Opportunities by
White KL, Chaubey I (2005) Sensitivity analysis, calibration, and vali- adopting Climate-Smart Agriculture for sustainble agriculture.
dations for a multisite and multivariable SWAT model 1. JAWRA Agriculture 11:192. https://doi.org/10.3390/agriculture1103019
J Am Water Resour Association 41(5):1077–1089
Wiebke F (2003) Case Study: The Agricultural System of the Konso Publisher’s Note Springer Nature remains neutral with regard to juris-
in South Western Ethiopia. http://fwu.fb10.uni-siegen.de/bkd/ dictional claims in published maps and institutional affiliations.
FWU_WRP.htm
Wolde K, Bogale G (2017) Effect of land use land cover dynamics Springer Nature or its licensor (e.g. a society or other partner) holds
on hydrological response of watershed: case study of Tekeze exclusive rights to this article under a publishing agreement with the
Dam watershed, northern Ethiopia. International Soil and author(s) or other rightsholder(s); author self-archiving of the accepted
Water Conservation Research, p 116. https://doi.org/10.1016/j. manuscript version of this article is solely governed by the terms of
iswcr.2017.03.002 such publishing agreement and applicable law.
Woldeamlak B (2007) Soil and water conservation intervention with
conventional technologies in northwestern highlands of Ethiopia:
Acceptance and adoption by farmers. Land Use Policy 24:2, 404–
416. https://doi.org/10.1016/j.landusepol.2006.05.004

13