Effect of land use land cover dynamics on

International Soil and Water Conservation Research 5 (2017) 1–16

HOSTED BY

Contents lists available at ScienceDirect

International Soil and Water Conservation Research journal homepage: www.elsevier.com/locate/iswcr

Original Research Article

Effect of land use land cover dynamics on hydrological response of watershed: Case study of Tekeze Dam watershed, northern Ethiopia☆

MARK

⁎

Kidane Weldea, , Bogale Gebremariamb a b

Alamata Agricultural Research Center, Natural Resources Management Research Core Process, Alamata, Ethiopia Arbaminch University, Institute of Technology, Department of hydraulic Engineering, Arbaminch, Ethiopia

A R T I C L E I N F O

A BS T RAC T

Keywords: LULC dynamics Stream flow Sediment yield Tekeze dam SWAT

Land use change is a very important issue considering global dynamics and their response to hydrologic characteristics of soil and water management in a catchment. A significant land use change has been observed in the Tekeze dam catchment. The main objective of this study was to estimate the potential impacts of the land use land cover (LULC) dynamics on hydrological response (stream flow and sediment yield). This was done by integrating SWAT model with GIS. The simulation and sensitivity analysis for each land use was done by dividing the catchment in to 47 sub-catchments and assigning HRUs based on multiple HRU definition. After a sensitivity analysis, calibration and validation of SWAT model, the impact of LULC dynamics on hydrological response were evaluated with three scenarios (climate of 2000s & 2008 LULC, climate of 2000s & 1986 LULC and climate of 1980s & 1986 LULC). In the Tekeze dam watershed, land cover change had a beneficial impact on modeled watershed response due to the transition from grass and shrub land to agricultural land. Simulation results for the Tekeze dam watershed indicates that increasing bare land and agricultural areas resulted in increased annual and seasonal stream flow and sediment yield in volumes. The mean annual stream flow was increased by 6.02% (129.20–137.74 m3/s) and the impact on sediment yield amounts to an increase of 17.39% (12.54–15.18 t/ha/yr) due to LULC dynamics. The hydrological response was more sensitive to LULC dynamics for the months of August to October than others in the year. These results demonstrate the usefulness of integrating remote sensing and distributed hydrologic models through the use of GIS for assessing watershed conditions and the relative impacts of land cover transitions on hydrologic response in a continuous manner.

1. Introduction Catchments are sensitive to land use land dynamics induced by human activities (Bosch & Hewlett, 1982).land cover changes are predicted to have important effect on river flows and sediment yields from specified catchment (Xing, Jianchu, Shive, Aggarwal & Jiatong, 2009). Hydrological response dynamics (an integrated indicator of watershed conditions) and water response in different river basins are attracted by changes in land use and climate (Wang, Liu, Kubota, & Chen, 2007). Land use activities, development and management of water resource are interdependent. Sedimentation in water resources is the outcome of the land erosion in its catchment area. Land erosion fundamentally has impact on physical and chemical characteristics of soils and causes on-site nutrient loss and off-site sedimentation and nutrients enrichment of water resources (Upadhya, Pandey, Upadhyay, Bajpai, 2012). To identify, prioritize and compare watersheds those that are

sensitive to change and to help management attempts to minimize undesired effects requires, improved assessment and understanding of the relationship among habitat change, land use change, runoff and water quality at landscape scale (Chhabra et al., 2006). Watershed process are highly dynamic in both space and time (Bloschl & Sivapalan, 1995). General statements about land water interactions need to be continuously questioned to determine whether they represent the best available information and the available information supports decision making processes for developmental activities in a sustainable way (FAO, 2002). Regional-scale hydrological models can play a vital role in river basin management. They simulate impacts on possible future changes of LULC and help to find measures improving adaptive capacity of river basins (Valentina et al., 2014). Expansion of agriculture, urbanization, deforestation and the day to day activities of mankind resulted to temporal and spatial change in land use land cover have affected water flow path ways and water balance (Rawat & Manish, 2015). Developing countries like Ethiopia,

☆ ⁎

Peer review under responsibility of International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Corresponding author. E-mail address: [email protected] (K. Welde).

http://dx.doi.org/10.1016/j.iswcr.2017.03.002 Received 22 November 2016; Accepted 9 March 2017 Available online 11 March 2017 2095-6339/ © 2017 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

Fig. 1. Location map of Tekeze dam watershed with digital elevation model.

There are a lot of evidence for the application SWAT for hydrological response modeling under different land uses and related issues (Asres & Awulachew, 2010; pp. 10; Fiseha, Setegn, Melesse, Volpi, & Fiori, 2012; Mengistu & Sorteberg, 2012; Nathan and Bosch, 2011).

where there agriculture serves as backbone of the economy and ensure wellbeing of the people, the adverse effects of land use land cove change are diverse. Beside to this various water resource development sectors (hydropower, irrigation, urban and rural water supply etc.) have persistently been affected by both temporal and spatial changes of LULC (Nigussie & Yared, 2010). The Tekeze Hydropower dam is the tallest arch dam in Africa, generating 300 MW power from a 180 m dam height but its reservoir is threatened by massive sedimentation. Even though the design period of the dam is for 50 years, it will not to exceed 25 years to be filled by sediment (Aforki, 2006). The sustainability of Tekeze dam reservoir is under the risk of sedimentation. Beside to this the life expectations of many reservoirs in the area built for irrigation or water supply in the dry season are threatened by massive sedimentation (Vanmaercke et al., 2010). But still little is known about the amount and dynamics of sediment transport in the Northern Ethiopian Highlands. Now a day runoff and soil erosion in catchment areas and its subsequent deposition in rivers, lakes and reservoirs are of great worry of humanity (Vipul, Tiwari, & Manjushree, 2010). This study i.e estimating the effect of land use land cover dynamics on the response of Tekeze dam catchment hydrology using models can motivate the policy makers and experts to formulate and implement effective, appropriate and sustainable response strategies to minimize the undesirable effects of land use land cover changes. Even though a number of watershed models (empirical and physically based) are available (Arnold, 1998) SWAT 2009 model was used for this study.

2. Materials and methods 2.1. Case study watershed The Tekeze Dam watershed, part of the Tekeze river basin covers an area of about 29,404 km2 and is situated between 11° 39’ 32.17″ and 13° 27′ 15.96′’ North latitude, and 37° 33′ 27.63′’ and 39° 40′ 7.24′’East longitude in the north-western part of Ethiopia (Fig. 1). Administratively, the major part of Tekeze dam watershed is in Amhara regional state and small part in Tigray regional state. It is much diversified watershed in terms of topography, climate land use and socio-economics. The average altitude is 2064 m with a mean slope of 12.5%. The Tekeze dam watershed has an annual rainfall ranging between 775–1220 mm. The majority of the area characterized by a semiarid climate with moderate rainfall and most of the total annual rainfall is received during one rainy season (June to September). There is high temporal variation rather than spatial variation of rainfall in the study area. It has high diurnal change in temperature i.e. there is high variation between the daily maximum and minimum temperature with an average temperature of 18 °C. According to the Ethiopian Ministry of water resources (MoWR) (2008), the local climate of the Tekeze dam

2

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

Fig. 2. Digital elevation model of Tekeze dam catchment.

station were obtained from National Meteorological Stations of Ethiopia. Since relative humidity, wind speed and solar radiation data records were limited for all the stations except for the Lalibela, Gonder and Mekele stations weather generator capabilities of SWAT model was used to generate those data by using Lalibela station records. Daily stream flow records (1978–2006) at Emba Madre gauging station was obtained from the hydrology department of Ministry of Water Resource, Irrigation and Energy of Ethiopia (MWRIEE). The sediment concentration record is a challenge to obtain since measurements on sediment concentration taken by the MWRIEE is in a noncontinuous time step. Hence the sediment data was prepared through a sediment rating curve using a series data record for 100 days in 2005 and 2006 at Tekeze dam site from MWRIEE.

watershed can be divided into three agro-ecological zones Dega (high altitude), Woina Dega (mid altitude) and Kolla (low altitude). Tekeze River basin have high seasonal variability and about 70% of the total runoff occurs during the main rainy season in the period June to October. The Tekeze River alone contributes 13% and 22% of the total annual flow of the Nile water during the dry and flood season respectively (Degefu, 2003). The major land use and land cover classes of the basin includes intensively cultivated land (7%), sparsely cultivated (58%), open woodland (12%), open grass land (5%), sparsely vegetated (0.2%), complex land (15%), and others (2.8%). 2.2. SWAT model and inputs SWAT is a semi-distributed and physically based watershed model that operates at a continuous time-step (Arnold et al., 2012). The model is designed to simulate the effects of changes in the catchment management practices on surface water and groundwater hydrology, diffuse pollution and sediment erosion within catchments (Sam, He, & Kevin, 2016). Two kinds of data; spatial data and temporal data are required by SWAT model. Spatial data include a digital elevation model (DEM), land-use map and soil map. The temporary data include hydrological data (stream flow & sediment yield) and climatic data (precipitation, solar radiation, relative humidity, wind speed and temperature). Within SWAT, a catchment is divided into multiple sub-catchments which are then further divided into Hydrologic Response Units (HRUs) that consist of homogeneous land use, slope and soil characteristics. The simulation processes of watershed using SWAT are split into two phases: as (i) land- based phase and (ii) Routing phase (channel-based phase). The land-based phase controls the loadings like runoff, sediment, nutrient and pesticides. While, the channel based routs the loadings throughout the stream network (Sam et al., 2016).

2.2.2. Geographical or spatial datasets The digital elevation model (DEM) of Tekeze dam watershed (Fig. 2) was obtained by downloading from ASTER GDEM website http://gdem.ersdac.jspacesystems.or.jp/ with 30 by 30 m DEM resolution. This DEM was used to delineate the catchment and the drainage patterns of the surface area analysis. Sub basin parameters such as slope length of the terrain, slope gradient and the stream network characteristics such as channel length, slope, and width were derived from this DEM. It was also used to determine the hydrological parameters of the catchment such as flow accumulation, direction, and stream network. A digitized soil map of Tekeze dam watershed was obtained from Ministry of Agriculture of Ethiopia in the form of shape file. The map of soil types within the sub-catchment was then derived from this National Soil Map vector dataset. The shape file was converted in to grid format using Arc Gis 9.3 as shown in Fig. 3. The soil data needed into physical and chemical characteristics of the soil which both play a large role in determining the movement of water and air within the HRU. The properties required by SWAT for each layer of each soil type include the depth of soil layer, soil texture, hydraulic conductivity, bulk density and organic carbon content and soil depth for the different layers of soil were obtained mainly from Tekeze River basin integrated development master plan and major soils of the world (FAO, 2002). The digital land use/ land cover data of the study area was obtained

2.2.1. Hydro-meteorological data The long-term records (1978−2013) meteorological data was collected from six stations (Gonder, Lalibela, Maichew, Mekele, Samre and May-Tsebri) which lie inside and on the boarder of the study watershed. The observations of meteorological variables of each

3

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

Fig. 3. Soil map of Tekeze dam catchment.

meters contribute most to the output variance due to input variability (Holvoet, Van Griensven, Seuntjens, & Vanrolleghem, 2005). The sensitivity analysis method implemented in SWAT model is called the Latin hypercube One- At- a –Time (LH-OAT) design as proposed by Mories (1991). Sensitivity analysis was then performed to identify those parameters that model outputs were sensitive to. In general, a parameter should be included in calibration if sensitivity analysis identifies that there is a 95% probability that the sensitivity of a variable to a particular parameter is significant. Stream flow sensitivity analysis followed by sediment yield sensitivity analysis was performed for each time reference land uses (1986 LULC and 2008LULC). For each reference land uses stream flow sensitivity analysis was done with 12 number of interval within latin hypercube for a total of 27 flow parameters (324 iterations). Model sediment parameter sensitization analysis for Tekeze dam watershed was also done through 240

from Ministry of water resource and ministry of agriculture of Ethiopia. For comparative use of the land use land cover evolution, LULC of 1986 was obtained from ministry of water and LULC of 2008 was obtained from ministry of Agriculture with the same scale (Figs. 4 and 5). The LULC of the study area is categorized into nine and eight groups for 1986 LULC and 2008 LULC respectively. Even though there have been marked changes in coverage but in both reference land uses cultivated land, bare land and shrub land were the dominant land uses in the study area. 2.3. SWAT model simulation, sensitivity analysis, calibration and validation 2.3.1. Model parameterization and sensitivity analysis Parameter sensitivity analysis provides insights as to which para-

Fig. 4. Historical land use land cover map (1986 LULC).

4

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

Fig. 5. Current land use land cover map (2008 LULC).

Table 1 Summery LULC of Tekeze dam watershed for different period of time. Area of Referenced Year (km2)

SWAT Code

FRSD BRLA AGRC PAST FRSD FRSE RNGB WATR FRST

LULC Type

1986

Total (%)

2008

Total (%)

%change

Afro-alpine Bare land Cultivation Grassland Natural Forest Plantation Shrub land Water Woodland Total

309.74 1193.74 7934.59 6740.01 22.57 32.17 12615.69 4.45 551.5 29404.47

1.05 4.06 26.98 22.92 0.08 0.11 42.9 0.02 1.88 100

– 1458.38 10436.27 5760.72 365.54 107.22 10961.72 192.81 121.82 29404.48

– 4.96 35.49 19.59 1.24 0.36 37.28 0.66 0.41 100

– 0.90 8.51 −3.33 1.17 0.26 −5.62 0.64 1.46

After running of the model for analysis of results, simulated stream flow and sediment yield were evaluated by visual inspection and quantitative statistics i.e. to evaluate how the model simulates well. For quantitative statistics the model performance was evaluated using three statistical criteria, the coefficient of determination (R2), NashSutcliffe efficiency (NSE) and percent bias (PBIAS) as recommended by Moriasi et al. (2007). NSE is a normalized statistic that describes the relative magnitude of the residual variance as compared to the observed and demonstrates how well the plot of observed versus simulated value fits the 1:1 line. The Nash-Sutcliffe Efficiency (NSE) coefficient proposed by Nash and Sutcliffe (1970) is defined by Eq. (1). R2 ranges from 0 to 1 and explains the proportion of variance in the observed data with higher value indicating less error variance. R2 is defined by Eq. (2). PBIAS measures the average tendency of the simulated data to be larger or smaller than their observed counterparts and is defined by Eq. (3). A positive value PBIAS indicates model under estimation bias and negative value indicates model over estimation bias. In general model simulation can be judged as satisfactory if NSE > 0.4 and R2 > 0.5 and PBIAS ± 25% for stream flow and PBIAS ± 55%

iterations (12 parameter *20 iteration per parameter) for each LULC maps. The sensitivity of parameters were categorized in to classes of small 0 < RS < 0.05, Medium 0.05 < RS < 0.2, High 0.2 < RS < 1, very high RS > 1.0 according to Lenhart, Eckhardt, Fohrer, and Frede (2002). Flow and sediment parameters were selected for calibration those their value ranges between very high to medium classes of sensitivity class above.

2.3.2. Model calibration, evaluation and validation methods Before calibration proceeds the performance of the model was evaluated for the initial simulation with the model default parameter values. But the default SWAT simulation result was with discrepancy between measured and simulated outputs. Hence both automatic and manual calibrations was done respectively. SWAT model calibration for stream flow and sediment yield were performed for 1986 LULC and 2008 LULC separately at the watershed outlet (dam axis). Only sensitive parameters were included in the calibration of the model at a monthly time-step against observations of discharge and sediment yield loads recorded at the outlet of the Tekeze dam watershed.

5

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

Table 2 flow parameter sensitivity analysis result. 1986 LULC

2008 LULC

SWAT code

RS

Rank

Sensitivity class

SWAT code

RS

Rank

Sensitivity class

CN2 Alpha_Bf Gwqmn Sol_Z Esco Sol_Awc Blai Canmx Revapmn Ch_K2 Gw_Revap Epco Sol_K Slope Gw_Delay Ch_N2

0.472 0.311 0.299 0.117 0.116 0.11 0.084 0.069 0.0516 0.0464 0.0399 0.0198 0.0161 0.0146 0.0104 0.0079

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

High High High Medium Medium Medium Medium Medium Medium Medium Small Small Small Small Small Small

CN2 Alpha_Bf Gwqmn Esco Sol_Awc Sol_Z Blai Soil_K Gw_Revap Epco Canmx Revapmn Ch_K2 Epco Ch_N2 Slope Gw_Delay Sol_Alb

0.562 0.551 0.322 0.263 0.117 0.115 0.0798 0.0691 0.0642 0.0467 0.0455 0.0273 0.0149 0.0138 0.0113 0.0042 0.0033 0.0024

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

High High High High Medium Medium Medium Medium Medium Medium Medium Small Small Small Small Small Small Small

Note: RS, is relative sensitivity: small to negligible 0 < RS < 0.05, Medium (0.05 < RS < 0.2. High 0.2 < RS < 1, very high RS > 1.0.

Table 3 Calibrated flow parameters. 1986 LULC

2008 LULC

Rank

Parameter

Allowable range

Calibrated value

Parameter

Allowable range

Calibrated value

1 2 3 4 5 6 7 8

CN2 Alpha_Bf Gwqmn Sol_Z Esco Sol_Awc Blai Canmx

0–100 0–1 0–5000 0–3000 0.01–1 0–1 0–1 0–10

−18% 0.782 2800 8.5% 0.46 2.5% 0.09 0.043

CN2 Alpha_Bf Gwqmn Esco Sol_Awc Sol_Z Blai Soil_K

0 −100 0–1 0–5000 0–1 0–1 0–3000 0–1 0–100

−14.5% 0.778 3100 0..51 −3.5 +5.5% 0.09 18

Fig. 6. Observed and simulated monthly flow hydrograph of calibration period (1978–1982) for 1986 LULC.

⎡ ∑ n (Y − Y ) 2 ⎤ obs sim ⎥ NSE = 1−⎢ in=1 ⎢⎣ ∑i =1 (Yobs − Yobs )2 ⎥⎦

for sediment yield (Ajai, Mohd, Isaacc, & Denisc, 2014). For the visual inspection, the scatter plots were used. n

R2 =

∑i =1 (Yobs −

Yobs )(Ysim − Ysim )

n

2

∑1 (Yobs −

Yobs )

n

∑i (Ysim − Ysim )

2

(2)

n

(1)

PBIAS =

6

∑i =1 (Yobs − Ysim ) n

∑i =1 Yobs

×100% (3)

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

Fig. 7. Observed and simulated monthly flow hydrograph of calibration period (1996−2002) for 2008 LULC.

Fig. 8. Observed and simulate monthly flow hydrograph of validation period (1983–1985) for 1986 LULC.

Fig. 9. Observed and simulate monthly flow hydrograph of validation period (2003–2006) for 2008 LULC.

Where Ysim and Yobs are the simulated and observed values respectively, Yobs is the mean of n observed values; and Ysim is the mean of n simulated values. Model validation was done to ensure that the calibrated set of parameters performs reasonably well under an independent data set. In order to utilize any predictive watershed model for estimating the effectiveness of feature potential management practices the model was validated against an independent dataset without adjusting calibrated parameters. The period of 1983–1985 and 2003–2006 daily stream flow and sediment yield data were used for model validation of selected

Table 4 Stream flow calibration and validation model performance statistics result. Parameters

R2 ENS PBIAS Time Step

1986 LULC

2008 LULC

Calibration (1978–1982)

Validation (1983–1985)

Calibration (1996–2002)

Validation (2003–2006)

88 85 −9.6 Monthly

85 88 −5 Monthly

85 84 −4.1 Monthly

85.34 80.42 −6.85 Monthly

7

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

Table 5 Sediment parameters sensitivity analysis result. 1986 LULC

2008 LULC

SWAT_code

RS

Rank

Class

SWAT_code

RS

Rank

Class

USL_P Spcon Slope SOL-AWC Spexp SOL_K BLAI USLE_C Sol_Alb Ch_erod

0.873 0.802 0.657 0.186 0.077 0.063 0.0319 0.0266 0.0238 0.0207

1 2 3 4 5 6 7 8 9 10

High High High Medium Medium Medium Small Small Small Small

USL_P Spcon Slope Ch_Cov Sol_AWC SOL_K Spexp USLE_C USLE_K BLAI SOL_Alb

0.927 0.811 0.557 0.177 0.066 0.055 0.0328 0.0266 0.0215 0.0205 0.0182

1 2 3 4 5 6 7 8 9 10 11

High High High Medium Medium Medium Small Small Small Small small

Table 6 Calibrated sediment Parameters. 1986 LULC

2008 LULC

Rank

Parameter

Allowable range

Calibrated value

Parameter

Allowable range

Calibrated value

1 2 3 4 5 6 7 8

USL_P Spcon Slope SOL_AWC Spexp SOL_K Biomix –

0–1 0.0001–0.01 0–1 0–1 1–2 0–1 0–100 –

0.69 0.0083 +5% +8% 1.25 35 0.09 –

USL_P Spcon Slope Ch_Cov Sol_AWC SOL_K Spexp USLE_C

0–1 0.0001–0.01 0–1 0–1 0–1 0–100 1–2 0–1

0.65 0.0049 +4.5% 0..51 +7% 45.5 1.25 0.3

Fig. 10. Observed and simulated monthly sediment yield hydrograph of calibration period (1978−1982) for 1986 LULC.

Fig. 11. Observed and simulated monthly sediment yield hydrograph of calibration period (1996−2002) for 1986 LULC.

8

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

Fig. 12. Observed and simulated monthly sediment yield hydrograph of validation period (1983–1985) for 1986 LULC.

Fig. 13. Observed and simulated monthly sediment yield hydrograph of validation period (2003−2006) for 2008 LULC.

3. Result and discussions

Table 7 Sediment yield Calibration and validation model performance statistics result.

3.1. Analysis of land use changes Parameters

R2 ENS PBIAS Time step

1986 LULC

2008 LULC

Calibration (1978–1982)

Validation (1983–1985)

Calibration (1996–2002)

Validation (2003–2006)

89 86 −5.67 Monthly

83 85 −6.85 Monthly

88 89 −7 Monthly

83 82 −7.85 Monthly

The dominant land use types of the Tekeze dam watershed in both reference land uses are agricultural, shrub and grass lands, which in total account over 91% of the total area. But their individual percentage coverage of these dominant land uses in each reference land use is different (Table 1). 3.2. Sensitivity analysis, calibration and validation under land use dynamics

Table 8 Model response to land use change. Temporal Land use

Mean annual stream flow (m3/s)

Annual sediment yield (ton/ ha)

Scenario I Scenario II Scenario III Change (I-II) (%) Change (I-III) (%)

137.74 129.20 127.67 6.20% 7.31%

15.18 12.54 11.87 17.39 21.80

3.2.1. Stream flow sensitivity analysis Sensitivity analysis results of SWAT model stream flow parameters identified as significant for a period five years (1978−1982) and seven years (1996−2002) and shows a range of small to high sensitivity class for 1986 LULC and 2008 LULC respectively (Table 2). The top three sensitive flow parameters have the same rank and sensitivity class for both LULC reference years. The variation in sensitivity level of flow parameters for the two reference land uses occurs for those parameters which their sensitivity index laid in medium and low class. Almost similar results was obtained by Ying, Chen, Wang, and Peng (2011).

flow and sediment parameters for 1986 and 2008 LULC respectively in monthly time scale.

3.2.2. Stream flow calibration and validation The periods 1978 – 1982 and 1996 – 2002 were used for stream calibration of 1986 LULC and 2008 LULC respectively. These periods was selected for model calibration as meteorological and stream flow records during this period were complete and include both high and

9

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

Fig. 14. Monthly simulated stream flow variation under land use dynamics for the period of (1996–2002) for 1986 and 2008 LULC.

Fig. 15. Monthly simulated sediment yield variation under land use dynamics for the period of (1996–2006) for 1986 and 2008 LULC.

low flow conditions comparatively. Eight flow parameters which represent groundwater, soil, runoff, evaporation and channel components of the watershed hydrological process were calibrated and set their value as shown in Table 3. Flow hydrographs were developed to compare observed and simulated flow values for the calibration periods of each LULC in monthly time scale (Figs. 6 and 7). Both the calibration and validation periods, respectively, indicates that the model achieved a relatively good fit between predictions and observations. Figs. 6 and 7 shows the simulated flow value for both 1986 LULC and 2008 LULC are slightly less than that of the measured value at peak flow months (august). But the simulated flow is slightly larger than the measured value at low flow months (January to May). Generally the model slightly over estimate mean monthly stream flow for each specified land uses reference years. The temporal variation of monthly observed and simulated results for validation periods of 1986 LULC and 2008 LULC are depicted also in Figs. 8 and 9. A reasonable match can be noticed from these figures, giving more support to ward utilizing SWAT model for Tekeze dam watershed and can achieve the intended modeling objectives. Depending on the model performance indicators; correlation coefficient (R2), Nash-Sutcliffe simulation efficiency (ENS) and percent of bias (PBIAS) summarized in the Table 4, the model stream flow simulation provides confidence for the further application of the model to assess stream flow hydrologic response analysis due to spatial temporal variability of the watershed characteristics. Because, observed flow data during calibration and validation periods, respectively, indicates that the model achieved a relatively good fit between predictions and observations.

3.2.3. Sediment yield sensitivity analysis Ten out of 12 analyzed SWAT sediment flow parameters that directly govern the sediment yield and transport in the catchment were found to be sensitive for 1986 land uses, but 11 parameters were found to be sensitive for 2008 LULC (Table 5). It should be noted that these parameters can be categorized into two groups as upland and channel factors. Some channel factors (Spcon and Ch_cov) become more sensitive than some upland parameters (soil erodibility and initial residue cover). This might be due to sediment loads are sustained during the dry days, even though it is low in magnitude. Similar result also shows by Ashgar Asghar, Ali, Shamsolah, and Ahmad, (2012). Whereas, the higher importance related to soil erosion USLE support practice and USLE soil erodibility factors was expected. The high class sensitive sediment parameters are the same for both 1986 LULC and 2008 LULC including their rank. But for the medium and lower sensitive parameters their rank and class of sensitivity are interchanged for 2008 LULC. 3.2.4. Sediment yield calibration and validation SWAT model calibration and validation for monthly sediment yield were conducted after the model was calibrated and validated for stream flow (Table 6). Because prior to sediment calibration stream flow of the catchment has to be calibrated first as sediment outflow from each HRU is governed by soil, hydrologic and hydraulic parameters such as soil erodibility, surface runoff, stream discharge and stream flow velocity (Holvoet et al., 2005). Figs. 10 and 11 represents the observed and simulated monthly sediment yield hydrograph of calibration periods for 1986 LULC and 2008 LULC respectively. Like the calibration period, the sediment yield hydrograph of measured and simulated output in monthly time setup during the

10

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

validation period (1983–1985) for 1986 LULC & (2003–2006) for 2008 LULC respectively shows a good agreement (Figs. 12 and 13). Values of Correlation coefficient (R2), Nash-Sutcliffe simulation efficiency (ENS) and percent of bias (PBIAS) summarized in Table 7, confirms reasonable monthly sediment yield results of the model simulation (both calibration and validation period). This provides a confidence that further application of SWAT model to assess sediment yield hydrologic response analysis due to spatial temporal variability of the watershed could have minimum bias.

pattern of the sediment outflow from the entire basin of the study area also indicates a seasonal variation of the sediment out flow for both land use land cover reference years (1986 LULC and 2008 LULC) using 2000s climatic data (Fig. 15). The land use land cover dynamics shows a higher effect on the peak sediment yield season (August to September) at the season when stream flow is also maximized. The maximum monthly sediment yield difference between 1986 and 2008 LULC occurs at months of September and august as 1.18 and 1.17 t/ha respectively.

3.3. Modeling stream flow and sediment yield response to land use dynamics

4. Conclusions Tekeze dam watershed had experienced significant changes in the land use land cover over the 22 year interval. Among the different types of land use, which show a large alteration are cultivated land and bare land (increased by 8.51% and 0.9%) respectively. The main cause of this significant change was due to expansion of intensive agricultural practice in the area which later causes a rapid reduction of shrub land and grassland by 5.62% and 3.33% respectively. Forest, water body and plantation were also shown a large dynamics during the comparison of the two times reference land uses. For both reference land uses (1986 LULC & 2008 LULC), sensitive parameters of stream flow and sediment yield were the same even though same parameters sensitivity rank varies. Hence, these calibrated parameters can be used for further future hydrological and environmental studies in the Tekeze basin without needing to do sensitivity analysis. More ever, the applicability of the SWAT model in simulating sediment discharge and stream flow dynamics of Tekeze dam catchment has validated based on the satisfactory values of the statistical measures of the model efficiency. Hence, the model simulation results provide confidence for the further application of the model to assess the hydrologic response analysis due to spatial and temporal variability of the catchment characteristics will have minimal bias within Tekeze basin. The mean annual stream flow and annual sediment yield of the watershed shows an increase in average annual stream flow from 129.20 m3/s to 137.74 m3/s (6.20% increment). Similarly, Sediment yield change shows an increment of 17.39% (from 12.54 to 15.17 t/ha). These increments of stream flow and sediment yield were highly observed during August to October (heavy rainfall seasons). This is the direct relationship of sediment yield and runoff. I.e. sediment yield is a function runoff and other process happening in the catchment. Comparatively the land use dynamics have a higher effect on sediment yield than stream flow. Hence watershed conditions should be continuously assessed for better management of the specific watershed.

3.3.1. Establishing scenarios for assessing impacts of land-use To analyze the effect of land use land cover dynamics of the catchment on stream flow and sediment yield three scenarios were developed. Hence the one factor at a time approach (Li, 2009) by changing one factor at each time while keeping others constant was applied to analyze impacts of land-use changes. Scenario I: Climate of 2000s and land use of 2008 (Base line) Scenario II: climate of 2000s and land use of 1986 (land use change) Scenario III: climate of 1980s and land use of 1986 (land use and climate & other factors change) The climate data sets were separated in to two periods, 1978 – 1989 (representing the 1980s) and 1996–2009 (representing 2000s). The relative change of annual stream flow and sediment yield due to land use land cover dynamics of Tekeze dam watershed shows higher influence over sediment yield than stream flow. Simulation analysis of the effect of land use and land cover change on the annual stream flow and sediment yield of the Tekeze dam watershed using SWAT2009 over the 22 years (1986 LULC-2008 LULC) difference is summarized as shown in Table 8. Mean annual stream flow and sediment yield of Tekeze dam watershed was increased by 6.20% and 17.39% respectively due to land use land cover change only. The main contributing reasons for this change are the expansion of intensive agricultural lands and expansion of bare land and this in turn made the catchment susceptible to surface runoff. On the other hand, the combined effect of land use change, climate variability and other factors made the average annual stream flow and sediment yield to increase by 7.31% and 21.8% respectively. This change may mainly enforced by increment of dynamic rainfall events during wet season. An attempt was also done to show the effect of land use dynamics on the monthly basis of the stream flow and sediment yield. The land use dynamics have a higher effect during the peak stream flow season (August) and the medium flow months (September–November) (Fig. 14). This in turn shows as the surface runoff response to rainfall mainly depends on the characteristics of the watershed, the land use dynamics create a lower effect on the stream flow of the dry season, which is mainly the base flow as compared to the stream flow of the rainy season. A wide range of changes in stream flow is shown between August and October as compared 2008 LULC with 1986 LULC, the increase in monthly stream flow is likely to reach up to 28.67% in September under the same climatic conditions. The simulated temporal

Acknowledgments We gratefully acknowledge Ministry of water, irrigation and energy (MoWIE), National meteorological service agency (NMA), Water works design and supervision Enterprise (WWDSE), and Ministry of Agriculture and rural development (MoARD) of Ethiopia for their all rounded support and cooperation in availing the necessary data. Our deepest gratitude also extends to Alamata Agricultural Research center for financial support of this study.

Appendix See Appendix Tables A1–A5 here.

11

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

Table A1 Mean monthly flow (m3/s) of Tekeze River at Dam site (1978−1985 and 1996–2006). Year

1978 1979 1980 1981 1982 1983 1984 1985 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

Jan

Feb

Mar

Apr

Month May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

3.55 2.48 4.62 2.41 0.84 3.04 5.24 15.62 3.76 0.25 13.62 18.12 10.54 53.52 5.79 6.67 10.28 13.89 15.18

2.28 1.77 2.57 1.16 0.16 2.70 5.16 9.95 1.98 0.47 14.00 10.63 4.24 33.24 14.10 2.84 8.18 13.52 10.92

18.66 1.63 2.80 1.30 3.19 2.94 2.85 12.24 2.16 32.60 62.93 36.38 5.10 33.55 11.34 5.46 9.22 9.92 11.65

12.60 4.40 3.51 9.09 5.21 6.40 7.72 9.50 2.77 89.25 53.08 20.79 18.29 31.64 5.62 2.83 13.79 5.91 11.38

39.05 11.47 16.22 5.53 9.80 6.14 2.11 19.51 13.15 23.29 68.97 43.37 54.27 30.10 5.43 4.85 16.97 9.62 11.18

25.98 67.28 29.29 36.29 1.47 15.70 34.19 40.82 25.83 29.43 131.96 99.93 58.42 66.59 12.78 47.05 57.59 32.02 34.85

197.31 612.68 185.66 220.03 356.56 256.12 274.70 205.53 149.69 378.86 342.18 321.13 500.49 444.00 223.75 498.70 228.95 280.15 137.95

487.54 379.34 280.24 252.95 830.50 640.57 881.81 535.41 709.87 721.19 698.55 386.04 631.66 877.28 658.81 773.61 427.21 513.36 384.62

149.27 167.91 118.57 55.85 182.93 156.07 572.43 166.35 316.48 263.61 248.57 109.30 670.22 546.50 210.07 275.28 181.70 349.12 342.50

49.43 33.48 17.53 7.42 45.48 28.11 96.22 41.76 18.81 57.34 82.41 131.91 195.08 214.26 67.97 70.35 35.00 118.94 170.72

17.60 12.19 6.60 6.09 10.79 12.14 38.36 27.99 3.73 20.43 46.25 76.54 107.54 104.99 20.73 17.42 22.09 25.15 62.77

5.05 6.52 3.60 2.01 4.72 7.44 26.44 13.36 1.26 17.48 37.09 29.29 54.24 58.76 10.40 11.56 20.32 16.57 21.30

Table A2 Measured sediment concentration of Tekeze River at Dam site. Date

Flow (m3/s)

Sediment Conc.

Date

Flow (m3/s)

Sediment Conc.

8-Aug−05 10-Aug−05 11-Aug−05 15-Aug−05 16-Aug−05 18-Aug−05 31-Aug−05 1-Sep−05 2-Sep−05 3-Sep−05 4-Sep−05 5-Sep−05 6-Sep−05 7-Sep−05 8-Sep−05 9-Sep−05 10-Sep−05 12-Sep−05 13-Sep−05 14-Sep−05 15-Sep−05 16-Sep−05 17-Sep−05 21-Sep−05 22-Sep−05 23-Sep−05 24-Sep−05 25-Sep−05 26-Sep−05 27-Sep−05 28-Sep−05 29-Sep−05 30-Sep−05 1-Oct−05 2-Oct−05 3-Oct−05 4-Oct−05 5-Oct−05 6-Oct−05 10-Oct−05 4-Mar−06 5-Mar−06 6-Mar−06 7-Mar−06 8-Mar−06

201.13 296.97 471.26 301.83 391.44 275.46 327.66 935.3 642.83 245.38 245.38 225.88 242.82 323.62 364.67 364.67 280.64 200.13 168.55 149.01 148.75 136.72 127.58 89.23 89.15 97.47 92.06 84.01 68.49 51.18 59.36 52.66 54.65 43.56 44.44 38.71 34.71 31.1 38.42 33.44 3.52 3.32 2.74 2.9 2.8

11159.11 20809.48 25664.38 12464.03 19943.11 11744.3 6919.25 17572.44 10429.88 4851.64 4034.06 4354.92 3006.56 5188.46 5653.2 6440.13 3911.59 4864.7 3655.89 3941.45 3013.05 3101.31 3371.3 1175.26 757.54 1460.21 1740.21 1489.85 1419.57 1413.6 1176.76 1190.78 1662.36 1500.95 3087.2 1229.66 1994.25 2973.45 1246.44 375.34 882.27 348.98 338.68 409.74 271.89

11-Oct−05 12-Oct−05 13-Oct−05 14-Oct−05 16-Oct−05 17-Oct−05 18-Oct−05 19-Oct−05 20-Oct−05 21-Oct−05 22-Oct−05 23-Oct−05 24-Oct−05 25-Oct−05 26-Oct−05 27-Oct−05 28-Oct−05 29-Oct−05 30-Oct−05 31-Oct−05 1-Nov−05 2-Nov−05 3-Nov−05 4-Nov−05 6-Nov−05 7-Nov−05 8-Nov−05 9-Nov−05 10-Nov−05 11-Nov−05 12-Nov−05 13-Nov−05 14-Nov−05 15-Nov−05 16-Nov−05 17-Nov−05 18-Nov−05 28-Feb−06 1-Mar−06 2-Mar−06

38.05 37.8 31.95 31.2 35.47 27.76 30.04 32.96 27.83 26.39 23.98 32.16 24.3 28.3 24.7 24.49 25.67 24.95 25.38 23 21.71 21.28 22.8 20.95 21.64 20.68 19.63 19.47 18.3 18.3 18.57 17.8 16.9 16.37 16.64 16.5 16.94 3.27 2.97 2.69

438.5 2826.21 883.57 608.3 566.88 639.14 1349.48 300.35 789.83 359.66 995.64 440.04 219.97 228.5 426.26 1065.79 1696.49 708.55 1472.14 724.39 428.97 727.08 648.83 242.32 741.88 901.03 742.04 1542.89 565.53 926.16 470.8 677.58 603.15 725.95 442.68 686.83 429.93 350.29 375.46 3427.17

12

13

1 2 3 1 1 1

Orthic luvisols

Chromic vertisol Eutric regosols Eutric nitisols

Vertic cambisols

1 2 3 4

1 2 3

Dystric nitisols

Cambic arenosols

1 2 3

chromic cambisols

1 2 1

1 1

Leptosols Orthic solonchak

Eutric cambisols

NLAYERS

Soil name

D D B

C

B

C

D

D

B

D B

HYDGRP

1500 1574 1500

1651

736.6

1320

900

1651

1651

500 1500

SOL_ZMX

Table A3 Soil parameters of the study area used in SAWT data base.

SILSL C C

FSL-FSL-SIC

SIL-CN-SIL-CNV

C

SIC-C

SICL-C-C

LS-S-SL

CL CL

TEXTURE

177.8 558.8 1651 1500 177.8 1500

279.4 457.2 711.2 736.6

200 900 1320

203.2 609.6 1651

203.2 500 1651

500 1000

SOL_Z1

1.15 1.3 1.55 1.35 1.35 1.35

1.25 1.35 1.83 2.5

1.5 1.46 1.12

1.35 1.35 1.35

1.7 1.7 1.6

1.44 1.43

SOL_BD1

0.19 0.15 0.18 0.1 0.19 0.1

0.19 0.14 0.09 0.01

0.2 0.18 0.14

0.2 0.15 0.13

0.14 0.06 0.13

0.19 0.2

SOL_AWC1

210 150 0.47 90.5 1.7 65.9

26 15 6.8 500

33.6 40 0.39

7.9 0.23 0.12

700 500 25

287 307

SOL_K1

3.49 1.02 0.15 1.9 5.81 1.9

3.49 0.29 0.29 0.1

1.63 1.1 2.2

5.81 1.94 0.65

1.74 0.58 0.19

1.8 1.8

SOL_CBN1

8.5 8.5 45 50 62.5 25

11.5 11.5 11.5 5

21 13 42

35 75 75

3 3 10.5

35 33

CLAY1

26.89 26.89 47.63 23 26.24 27.18

56.14 56.14 56.14 25

33 46 30

47.76 17.58 17.58

16.7 1.51 17.58

38 35

SILT1

64.61 64.61 7.37 10 11.26 27

32.36 32.36 32.36 70

46 7.42 28

17.24 7.42 7.42

15.85 55 7.42

27 12

SAND1

1.1 1.24 0 5.25 0 10.9

7.51 20.93 51.86 98

1.2 0 1.94

1.29 1.29 1.29

0 0 0

1.09 3.02

ROCK1

0.01 0.03 0.17 0.14 0.01 0.14

0.01 0.13 0.13 0.19

0.01 0.01 0.1

0.01 0.01 0.07

0.01 0.16 0.16

0.13 0.13

SOL_ALB1

0.28 0.32 0.49 0.25 0.49 0.25

0.32 0.28 0.2 0

0.31 0.49 0.38

0.49 0.49 0.49

0.49 0.43 0.49

0.21 0.21

USLE_K1

0 0 0 0 0 0

0 0 0

0 0 0

0 0 0

0 0 0.43

0 0

SOL_EC1

K. Welde, B. Gebremariam

International Soil and Water Conservation Research 5 (2017) 1–16

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

Table A4 Variables and their statistical values of Lalibela station used for weather generator. Variable Name

Value

Variable Name

Value

Variable name

Value

Variable name

Value

TMPMX1 TMPMX2 TMPMX3 TMPMX4 TMPMX5 TMPMX6 TMPMX7 TMPMX8 TMPMX9 TMPMX10 TMPMX11 TMPMX12 PR_W1_1 PR_W1_2 PR_W1_3 PR_W1_4 PR_W1_5 PR_W1_6 PR_W1_7 PR_W1_8 PR_W1_9 PR_W1_10 PR_W1_11 PR_W1_12 PCPMM1 PCPMM2 PCPMM3 PCPMM4 PCPMM5 PCPMM6 PCPMM7 PCPMM8 PCPMM9 PCPMM10 PCPMM11 PCPMM12 SOLARAV1 SOLARAV2 SOLARAV3 SOLARAV4 SOLARAV5 SOLARAV6 SOLARAV7 SOLARAV8 SOLARAV9 SOLARAV10 SOLARAV11 SOLARAV12

25.8 27.42 27.22 26.85 26.69 25.15 20.16 19.93 22.71 24.7 25.11 25.14 0.05 0.05 0.13 0.16 0.12 0.17 0.7 0.63 0.21 0.06 0.06 0.04 9.26 7.93 42.8 39.78 24.86 57.7 284.8 256.8 46.08 15.93 35.79 4.8 20.43 22.21 21.78 21.98 21.01 18.53 15.23 16.19 18.32 19.66 20.07 19.88

TMPMN1 TMPMN2 TMPMN3 TMPMN4 TMPMN5 TMPMN6 TMPMN7 TMPMN8 TMPMN9 TMPMN10 TMPMN11 TMPMN12 PR_W2_1 PR_W2_2 PR_W2_3 PR_W2_4 PR_W2_5 PR_W2_6 PR_W2_7 PR_W2_8 PR_W2_9 PR_W2_10 PR_W2_11 PR_W2_12 PCPSTD1 PCPSTD2 PCPSTD3 PCPSTD4 PCPSTD5 PCPSTD6 PCPSTD7 PCPSTD8 PCPSTD9 PCPSTD10 PCPSTD11 PCPSTD12 DEWPT1 DEWPT2 DEWPT3 DEWPT4 DEWPT5 DEWPT6 DEWPT7 DEWPT8 DEWPT9 DEWPT10 DEWPT11 DEWPT12

13.13 14.02 14.42 14.59 15 14.35 12.06 11.81 12.84 12.63 12.6 12.61 0.38 0.33 0.58 0.52 0.42 0.6 0.88 0.85 0.57 0.48 0.58 0.34 1.75 1.61 4.56 3.47 3.29 5.1 10.47 9.43 3.75 3.12 4.84 0.88 8.17 7.7 9.2 10.12 10.36 11.14 13.37 13.99 12.53 10.07 9.15 8.32

TMPSTDMX1 TMPSTDMX2 TMPSTDMX3 TMPSTDMX4 TMPSTDMX5 TMPSTDMX6 TMPSTDMX7 TMPSTDMX8 TMPSTDMX9 TMPSTDMX10 TMPSTDMX11 TMPSTDMX12 PCPD1 PCPD2 PCPD3 PCPD4 PCPD5 PCPD6 PCPD7 PCPD8 PCPD9 PCPD10 PCPD11 PCPD12 PCPSKW1 PCPSKW2 PCPSKW3 PCPSKW4 PCPSKW5 PCPSKW6 PCPSKW7 PCPSKW8 PCPSKW9 PCPSKW10 PCPSKW11 PCPSKW12 WNDAV1 WNDAV2 WNDAV3 WNDAV4 WNDAV5 WNDAV6 WNDAV7 WNDAV8 WNDAV9 WNDAV10 WNDAV11 WNDAV12

1.19 1.64 1.96 2 1.95 2.62 2.2 1.55 1.83 1.34 1.31 0.95 2.4 2.15 7.65 8.15 5.25 8.9 27.2 26.3 10.9 3.65 3.8 1.6 9.56 8.41 6.32 4.38 7.54 4.58 2.27 2.16 4.07 11.6 4.94 7.59 1.2 1.34 1.45 1.52 1.65 1.32 0.85 0.87 0.92 1.16 1.05 1.14

TMPSTDMN1 TMPSTDMN2 TMPSTDMN3 TMPSTDMN4 TMPSTDMN5 TMPSTDMN6 TMPSTDMN7 TMPSTDMN8 TMPSTDMN9 TMPSTDMN10 TMPSTDMN11 TMPSTDMN12 RAINHHMX1 RAINHHMX2 RAINHHMX3 RAINHHMX4 RAINHHMX5 RAINHHMX6 RAINHHMX7 RAINHHMX8 RAINHHMX9 RAINHHMX10 RAINHHMX11 RAINHHMX12

1.03 1.41 1.49 1.49 1.54 1.73 1.11 1.11 1.19 0.9 1.12 0.97 0 0 0 0 0 0 0 0 0 0 0 0

Note: TMPMX: Average or mean daily maximum air temperature for month (ºC). TMPMN: Average or mean daily minimum air temperature for month (ºC). TMPSTDMX: Standard deviation for daily maximum air temperature in month (ºC). TMPSTDMN: Standard deviation for daily minimum air temperature in month (ºC). PCPMM: Average or mean total monthly precipitation (mm H2O). PCPSTD: Standard deviation for daily precipitation in month (mm H2O/day ). PCPSKW: Skew coefficient for daily precipitation in month. PR_W1: Probability of a wet day following a dry day in the monthPR_W2: Probability of a wet day following a wet day in the month. PCPD: Average number of days of precipitation in month. RAINHHMX: Maximum 0.5 h rainfall in entire period of record for month (mm H2O). SOLARAV: Average daily solar radiation for month (MJ/m2/day). DEWPT: Average daily dew point temperature in month (ºC). WNDAV: Average daily wind speed in month (m/s).

14

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

Table A5 Summery of sediment yield comparison at sub basin level. Sub basin

Scenario 1

Scenario 2

Scenario 3

Sub basin

Scenario 1

Scenario 2

Scenario 3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

13.26 9.56 13.56 8.15 15.70 31.03 17.08 16.70 15.57 30.22 8.58 31.90 28.21 29.46 14.42 7.26 28.79 17.21 32.57 24.56 25.18 7.61 22.54 15.91 8.58 7.36 14.07 12.36 13.64 18.49 20.53 10.86 11.53 12.97 7.86 11.76

12.89 7.78 10.42 6.32 13.09 22.57 14.45 12.67 12.08 25.06 6.30 24.99 25.13 20.35 11.96 5.34 23.93 17.01 21.35 22.91 19.61 4.85 14.91 13.55 7.09 7.99 13.64 8.58 7.54 15.49 19.90 8.74 13.56 12.34 9.32 8.73

7.11 7.34 12.87 6.61 11.29 21.30 13.63 11.95 11.40 23.65 5.95 20.15 23.71 18.50 12.35 4.58 22.58 16.05 23.58 18.78 19.20 5.04 14.07 14.62 6.69 7.54 9.83 12.79 12.16 12.79 21.61 7.81 8.10 11.64 8.79 8.24

37 38 39 40 41 42 43 44 45 46 47

7.03 11.58 8.70 8.03 10.40 8.03 9.75 9.63 7.69 7.70 9.25

5.43 5.87 9.13 7.32 8.28 7.00 7.34 6.08 11.55 10.25 8.67

5.12 5.54 8.61 10.90 8.25 5.96 6.93 5.74 6.91 9.67 10.57

FAO (2002). Major soils of the world land and water digital media series Food and Agricultural Organization of the United Nations Rome: CD-ROM 2002. Fiseha, B. M., Setegn, S. G., Melesse, A. M., Volpi, E., & Fiori, A. (2012). Hydrological analysis of the upper Tiber river basin, central Italy: a catchment modeling approach. Journal of hydrological processes. Holvoet, K., Van Griensven, A., Seuntjens, P., & Vanrolleghem, P. (2005). Sensitivity analysis for hydrology and pesticide supply towards the river in SWAT. Phys.Chem. Earth, 30(8-10), 518–526. Lenhart, T., Eckhardt, K., Fohrer, N., & Frede, G. (2002). Comparison of two different approaches of sensitivity analysis. phys.chem. C. Li, Z. (2009). Impact of land use change and climate variability on hydrology in an agricultural catchment on the Loess Plateau of China. Journal of hydrology. Pp., 272–288. Mengistu, F., & Sorteberg, A. (2012). Sensitivity of SWAT simulated stream flow to climatic changes within the Eastern Nile River basin. Journal Hydrology and earth system sciences.. Moriasi, D.N., Arnold, J.G., Van Liew, M.W., Bingner, R.L., Harmel, R.D., Veith, T.L., 2007. Model evaluation guidelines for systematic quantification of accuracy in watershed simulations. Transaction of ASABE50 (3), 885e900. Morris, M. (1991). Factorial sampling plans for preliminary computational Experiments. Techno metrics, 33(2), 161–174. MoWR, (2008) Tekeze River Basin - Physiography and Climate, Ministry of Water Resources, Addis Ababa, Ethiopia.[Online]: //〈www.mowr.gov.et/index.php? pagenum=3.3 & pagehgt=1000px〉 [Accessed June 1, 2015]. Nash, J. E., & Sutcliffe, J. V. (1970). River flow forecasting through conceptual models: part I. Nathan, S., Bosch, David, A., Davi, M., Haejin, H., & Peter, R. (2011). Application of the Soil and Water Assessment Tool for six catchment s of Lake Erie: Model parameterization and calibration. Great lakes Resarch journal. Nigussie, T., & Yared, A. (2010). Effect of land use land cover management on Koka reservoir sedimentation, Nile basin capacity Building Network Cairo, Egypt: Hydraulic Research institute, Delta barrage. Rawat, J. S., & Manish, K. (2015). Monitoring land use/cover change using remote sensing and GIS techniques: A case study of Hawalbagh block, district Almora, Uttarakhand, India. National Authority for Remote Sensing and Space Sciences. The Egyptian Journal of Remote Sensing and Space Sciences, 18, 77–84.

References Aforki, H. (2006). Sediment studies for Tekeze hydropower development. University Library (BIBSYS), Norwegian Centre for International Cooperation in Higher Education. Ajai, S., Mohd, I., Isaacc, R. K., & Denisc, D. M. (2014). Assessing the performance and uncertainty analysis of the SWAT and RBNN models for simulation of sediment yield in the Nagwa watershed, India. Hydrological science journal, 2(2), 65–74. Arnold, J. (1998). Large area hydrologic modeling and assessment - Part 1: Model development. Journal of the American Water Resources Association, 34(1), 73–89. Arnold, J. G., Moriasi, D. N., Gassman, P. W., Abbaspour, K. C., White, M. J., & Srinivasan, R. (2012). SWAT: Model use, calibration, and validation. Trans. ASABE, 55(4), 1491–1508. Asghar, B., Ali, H., Shamsolah, A., & Ahmad, J. (2012). Identification and prioritization of critical sub basins in highly mountainous watershed using SWAT model. Eurasian journal of soil science.. Asres, M., & Awulachew, S. (2010). A SWAT based runoff and sediment yield modeling: a case study of Gumera catchment in the Blue Nile basin. Ecohydrology for water Ecosystem and society in Ethiopia, pp. Vol.10(No. 2-4), 191–200. Bloschl, G., & Sivapalan, M. (1995). Scale issues in hydrological modelling: a review. , in: Modelling, J. D., Kalma, & Sivapalan, M. (Eds.). (1995). Scale issues in hydrological DOI: 10Ð1016/0022-1694(82)90117-2. New York, New York: Wiley and Sons, 9–49. Bosch, J. M., & Hewlett, J. D. (1982). A review of catchment experiments to determine the effect of vegetation changes on water yield and evapotranspiration. Journal of Hydrology, 55, 3–23 DOI: 10Ð1016/0022-1694(82)90117-2. Chhabra, A., Geist, H., Houghton, R. A., Haberl, H., Braimoh, A. K., Vlek, P. Lambin, E. F. (2006). Lambin, E. F., & Geist, H. J. (Eds.). (2006). Multiple impacts of land use/ cover change. In Land-use and Land-cover Change: Local Processes and Global Impacts 2002. Berlin: Springer, 71–116. Degefu, G.T., 2003. The Nile: Historical, Legal and Developmental Perspective. New York pp.413. [Online]: 〈http://books.google.com/books〉 [accessed, April 15, 2015].

15

International Soil and Water Conservation Research 5 (2017) 1–16

K. Welde, B. Gebremariam

basis of soil erosion hazard using remote sensing and geographic information system. International Journal of Water Resources and Environmental Engineering, Vol. 2(3), 130–136. Wang, G. X., Liu, J. Q., Kubota, J., & Chen, L. (2007). Effect of land-use changes on hydrological processes in the middle basin of the Heihe River, northwest China. Hydrological Processes, 21, 1370–1382 DOI: 10Ð1002/hyp.6308. Xing, M. M., Jianchu, X. U., Shive, P., Aggarwal, & Jiatong, L. (2009). Response of hydrological processes to land-cover and climate changes in Kejie catchment southwest China. Journal of hydrological process.. Ying, Li, Chen, Bao-Ming, Wang, Zhong-Gen, & Peng, Shao-Lin (2011). Effects of temperature change on water discharge, and sediment and nutrient loading in the lower Pearl River basin based on SWAT modelling. Hydrological Sciences Journal, 56(1), 68–83. http://dx.doi.org/10.1080/02626667.2010.538396.

Sam, D., He, Yi, & Kevin, M. (2016). Modelling the impacts of agricultural management practices on river water quality in Eastern England. Journal of Environmental Management, 180, 147–163. Upadhya, R., Pandey, K., Upadhyay, S. K., & Bajpai, A. (2012). Annual Sedimentation Yield and Sediment Characteristics of Upper Lake,Bhopal, India. Journal of Chemical Sciences, Vol. 2(2), 65–74. Valentina, K., Fred, H., Shaochun, H., Cornelia, H., Tobias, V., Hagen, K., & Zbigniew, W. (2014). Modelling climate and land use change impacts with SWIM: lessons learnt from multiple applications. Hydrological Sciences Journal, 10, 125–136. Vanmaercke, M., Zenebe, A., Poesen, J., Nyssen, J., Verstraeten, G., & Deckers, J. (2010). Sediment dynamics and the role of flash floods in sediment export from mediumsized catchments: a case study from the semiarid tropical highlands in northern Ethiopia. Journals of Soils and Sediments, 10, 611–627. Vipul, K., Tiwari, N., & Manjushree, S. (2010). Prioritization of micro catchment s on the

16