NUMERICAL SIMULATION OF BLUE NILE BASIN USING DISTRIBUTED HYDROLOGICAL

MODEL
Hadir Abd-El Moneim1, Mohamed R. Soliman2, Hossam M. Moghazy3
1,
Postgraduate, Faculty of Engineering, Alexandria University, Egypt.
2
Assistant Professor, Faculty of Engineering, Alexandria University, Egypt on leave to Arab Academy for Science and
Technology &Maritime Transportation, Alexandria, Egypt.
3
Professor of Irrigation Engineering and Drainage, Faculty of Engineering, Alexandria University, Egypt.

ABSTRACT
The Nile River is an international river and its branches pass through eleven
countries with two major tributaries; the White Nile and the Blue Nile. It is considered
the main water resource and life artery for its downstream countries Egypt and Sudan. In
this research, the Hydrological River Basin Environmental Assessment Model (Hydro-
BEAM) is used to simulate the surface discharge in the Blue Nile Basin during the period
(1995-1999) using a long‐term global atmospheric reanalysis product, namely Japanese
25‐year reanalysis (JRA-25). These data include precipitation, temperature, pressure,
wind speed, specific humidity, downward short‐wave radiation, upward short‐wave
radiation, downward long‐wave radiation and upward long‐wave radiation. The
simulated discharge was examined and calibrated at Khartoum, Sennar dam and El Deim
gauging stations. Good agreement obtained between the observed and the simulated
discharge. The model used in this study provides a viable tool to assess and predict the
availability of water resources under future challenges imposed by climate change or
proposed and existing Ethiopian plans for new dams construction.
Keywords: Blue Nile River - Hydrologic model - Hydro-BEAM - JRA25

1. INTRODUCTION important tributary of the Nile River which

provides about 85 % of the Nile flow during high
The Nile River is considered the longest flow season [3]. However, 60% of this large
river in the World, where it extends from 4o S to streamflow is shared between Egypt and Sudan.
31o N with total length about 6800 km [1].
According FAO 2007, the catchment area of Nile The Blue Nile Basin suffers from many
River is about 3.1 million m3, which represents problems including the selection of suitable
about 10 % of Africa’s area [2]. It originates strategies for water management. In addition, the
from Victoria’s Lake and passing through eleven need for powerful hydrological models for flash
countries before reaching the Mediterranean Sea. floods and water resources management is
The Blue Nile River is one of the main tributaries hindered by the lack of observational data and
of the Nile River originating from the Ethiopian high-quality data. On the other hand, Egypt has
plateau and flowing into Sudan, where it meets recently been subjected to significant scarcity in
the White Nile at Khartoum to form the Main water resources due to increasing in water
Nile. The main Nile travels north to Egypt and demands with the continuous growing of
finally ends in a large delta and flows into the population, agricultural expansion, and industrial
Mediterranean Sea. It is considered the most development.


Email: hadir_eng@yahoo.com
1
 Bashar and Zaki (2005), used the catchments for better assessment and prediction
Hydrologic Modeling System (HMS) model for of simulation of the hydrological responses [7].
continuous hydrologic simulation of the Blue
Nile to evaluate the performance of the HMS The need to evaluate the predicted flow
with the Soil Moisture Accounting (SMA) due the contribution of the Blue Nile River
algorithm on the Blue Nile. They calibrated the reaching Egypt is crucial under critical future
simulation results for the period 1990-1994 and challenges including climate change and the
then validated the model using the period 1995- effects of dam construction. Thus, the main
1996. The results were reasonable and objective of this study is to investigate the
acceptable taking into consideration the lumped potential use of hydrological distributed model
time - invariant parameters used in the (Hydro-BEAM) for simulating the surface
calibration and the validation of the model. [4] discharge of the Blue Nile Basin under the main
forces of precipitation, pressure, wind speed,
Abdel-Aziz et al. (2011) used a specific humidity, downward short‐wave
distributed hydrological model based on remote radiation, upward short‐wave radiation,
sensing data to assess water resources in the Blue downward long‐wave radiation and upward
Nile River Basin. The Distributed Hydrological long‐wave radiation.
model (DHM), which was employed in this
study, is a grid-based geomorphology-based
hydrological model that uses governing 2. STUDY AREA
equations to describe hydrological processes.
Global datasets were used to generate the input Blue Nile basin, which has been selected
for simulating daily and monthly stream flows. as the study area Figure 1 is located between
Additionally, it used ground observed latitude 16 2’ N and 7 40’ N, and longitude 32
precipitation from few rain gauge stations. The 49’ E and 39 30’ E in the Ethiopian Highlands
simulated discharge examined at Khartoum [8]. The watershed of the Basin has a range of
gauging stations. [5]. sizes, slopes, topography, climatic patterns,
geological formations, drainage patterns,
Abdel-Aziz et al. (2014) used a DHM vegetation cover, soils and anthropogenic
model in the Blue Nile basin to forecasts activities.
hydrological process scenarios and predicted
scenarios of precipitation from two general
circulation models, CCSM3 model and
Miroc3.2-hires. They simulated river discharge
by the DHM using the observed rainfall from
1976 to 1979, while future precipitation is
simulated from 2011 to 2040, and then discharge
scenarios were predicted [6].
Gerbre (2015), applied HEC-HMS3.5
model for simulation of the upper Blue Nile
Basin. HEC-HMS3.5 is a semi-distributed
hydrological model. In this study, they calibrated
rainfall-runoff relation of the basin using HEC-
HMS 3.5 model from 1988-2000. After
calibration the model they validated the results
from 2001-2005. The results from this model are
satisfactory and acceptable. They concluded
HEC-HMS hydrological model can be used to
Figure 1 Map present the location of Blue Nile
model the upper Blue Nile River basin Basin

2
 Blue Nile River is considered the most Geographical Information System (GIS)
important tributary of the Nile River, which technique, (2) Surface runoff and stream routing
provides about 85% of the flow to the main Nile modeling based on the kinematic wave
River during high flow season. It originates from approximation, (3) Canopy interception losses,
Lake Tana’s outlet, traveling southward in the and (4) Groundwater modeling based on the
Ethiopian highlands then northwest for about linear storage model. The watershed is divided
900 kilometers from Lake Tana until up to the into meshed cells with multi layers and each
Sudanese border and ends when it meets the mesh contains information such as surface
White Nile at Khartoum, Sudan [9]. The runoff, land use, slope direction, and the absence/
catchment area of the Blue Nile basin is about presence of a channel [12]. The spatial resolution
325000 Km2 [6]. used to model the Blue Nile Basin is (5 km × 5
km).
Blue Nile’s topography is divided into
two features; the flat topography in the lowlands Hydro‐BEAM was originally developed
in Sudan, and the rugged topography, in the by Kojiri et al (1998) [13] as a tool to assist in
Ethiopian highlands where the high steep simulating long-term fluctuations in water
mountainous region combined cover about 65% quality and quantity in rivers through an
of the Basin area [10]. The Ethiopian highlands understanding of the hydrological processes that
are concentrated at elevations of 2000-3000 m, occur within a watershed. It has been used in a
with several peaks up to 4000 m or more [6]. pioneering work on comparative hydrology,
where a methodology for assessing the similarity
Rainfall over the Blue Nile basin varies
between watersheds was proposed Park et al.
significantly with altitude and is considerably
(2000) [13] to investigate sediment transport
greater in the Ethiopian highlands than on the
processes in the large watershed of the Yellow
Plains of Sudan. Rainfall ranges from nearly
River, China Tamura and Kojiri (2002) [13] .The
2,000 mm/yr in the Ethiopian Highlands to less
original Hydro-BEAM model with humid
than 200 mm/yr at the junction with the White
conditions configurations was adopted for flash
Nile [11].
floods simulation in the arid areas in wadi system
The highest mean annual temperatures by Saber et al. (2010) [13].
occur in the northeastern clay plains of Sudan. In
Sudan, daily minimum and maximum 3.2 Basic structures of Hydro-BEAM.
temperatures in January are 14°C and 33°C, and
those in May are 24°C and 44°C, respectively. Hydro‐BEAM is a distributed model
The area located in the highlands of Ethiopia is consisting of meshed cells with multi layers. The
characterized by lower minimum mean monthly cells are given by discrediting a river basin into
temperatures that range between 3oC and 21oC, square elements. Each of mesh cells has a
and occur between December and February [11]. hydrological structure of Layer A to D in the
vertical direction Figure 2. Horizontal
Evapotranspiration spatial distributions discharges from Layer A, B, C and D flow into a
are similar to rainfall and temperature with river. In this study, the model consists of three
considerable variations across the basin and layers A, B and C Deeper seepage and long‐term
higher correlation with altitude [11]. groundwater storage were not considered in this
analysis. In Layer A, running water is saturated
3. METHODOLOGY and relatively fast interflow with a higher
permeability in soil. In Layer B, moving water is
3.1 Hydro-BEAM Model
unsaturated and slow interflow with a non‐higher
permeability in soil. In Layer C, streaming water
Hydro-BEAM is a physical-based
is groundwater directly flowing into a river.
distributed hydrological model that is adapted
simulate surface runoff in Blue Nile Basin. It
consists of: (1) the watershed modeling using

3
 Figure 3 Digital Elevation Model of Blue Nile Basin

Figure 2 Conceptual representation of Hydro-BEAM [12]

3.3.2 Watershed Delineation
The watershed delineation of the target
3.3 Watershed Modeling basin determines the watershed, sub basins
boundaries and the main streams and the streams
GIS technique is used to serve as a network by using ArcMap 10.1 GIS platform and
watershed modeling tool in the modeling SRTM (30 m resolution) DEM [14]. Figure 4 (a)
process. Some geomorphologic information such [13] shows the steps of watershed delineation.
as area of watershed, perimeter, and main The generated streams are then compared with
channel length are calculated. The following the available satellite images, the geological map
processes have to be considered in the watershed and Google Earth to verify the results. The
modeling such as: results of watershed delineation and streams
network determine of Blue Nile Basin are shown
• Determination of the watershed boundary in Figure 4 (b).
location,
• Division of the watershed into a regular grid of
mesh cells (1 km or 2 km, etc.),
• Determination of a flow routing network based
on mesh cell elevation as provided by a Digital
Elevation Model (DEM) and checked against
topographical maps.
3.3.1 Digital Elevation Model
Topographic data was obtained by the (a)
Shuttle Radar Topography Mission (SRTM).
The used data set is SRTM void filled 1 arc-
second elevation data. The study area is
characterized by middle range with elevation
range from 4218 m above sea level (a.s.l) at
upstream to 393 m to downstream a.s.l and with
general slope towards the North West where
Blue Nile River meets the White Nile River at
Khartoum to form the Main Nile River as shown
in Figure 3. As well, DEM data are used to (b)
determine and delineate the watershed, sub
Figure 4 Flow chart of watershed modeling steps & Blue Nile
basin, the stream networks of the target basin, Basin and Streams network
flow direction, and mesh slope identification.

4
3.3.3 Flow routing map 3.4 PREPARATION OF INPUT CLIMATIC
There are two well known types of flow DATA
routing system to determine the water flow
direction in a watershed: four directions, and In order to apply DHM to a large river
eight directions routing systems. Hydro-BEAM basin, it is necessary to obtain the distributed
was mainly developed to use a four-direction input datasets. Therefore, we use a long‐term
flow routing map. The function of a flow routing global atmospheric reanalysis data, namely
map is to define a downstream destination for the (JRA25) (Japanese 25‐year reanalysis), which
discharge resulting from every cell in the was completed using the japan meteorological
watershed till reaching the furthest downstream agency (JMA) numerical assimilation and
mesh cell located at the watershed mouth. The forecast system [15]. The JRA25 reanalysis data
DEM elevations of the corners of each mesh cell is used to overcome the scarcity of the observed
is used to estimate flow direction from any given data in the study area. The Soil‐Vegetation‐
mesh cell as described in Figure 5. Atmosphere Transfer (SVAT) model applied to
obtain input variables for the hydrological
simulation [16]. Moreover, it can define the
potential evaporation, Ep, as the evaporation
from a continuously saturated imaginary surface.
Meteorological parameters of JRA25 include
precipitation, temperature, pressure, wind speed,
specific humidity, radiation, upward short‐wave
radiation, downward long‐wave radiation and
upward long‐wave radiation expect, evapo-
transpiration. The SVAT model was applied to
obtain input variables for the hydrological
simulation. There are five stages for preparation
input data to export climatic input and land use
[17].

3.5 LAND USE CLASSIFICATION

Hydro-BEAM is set to use five categories

of land use types Table 1, where they are
grouped and represented as a percentage land
cover of the total area of the mesh cell.

Table 1 land use types of modified Hydro-BEAM [12].

Figure 5 Schematic diagram represent the determination of

the flow direction [12]

The flow path of each mesh is decided

based on the elevation values of each corner. On
the other side, the perpendicular direction of
slope of the two halves of the mesh is estimated Land use types in the target basin were
based on dividing of each mesh into distinguished into the flowing three categories as
subdivisions. The flow direction depends on the follow; forests, grass and desert. The results of
direction of its slope in each mesh. land use distribution in the target basins were

5
depicted as Forests, grass, and desert as shown in 3.7 LINEAR STORAGE MODEL
Figure 6.
To predict the base flow process more
realistically, by applying a multi‐layer linear
storage function model to the B and C layers of
Hydro‐BEAM. The continuity equation and
dynamic equation of the linear storage function
model are as follows:
𝑑𝑆
=𝐼−𝑂 (5)
𝑑𝑡
𝑂 = 𝑘𝑆 (6)
Where, S is water storage (m), I is inflow (m/s),
O is outflow (m/s), k is the runoff coefficient
(1/s).
In the case of the B layer, the coefficient k
corresponds to k1 and k2. In the case of C layer,
Figure 6 the land use results after the reclassification to three
types
the value corresponds to k3. The inflows to the B
and C layer (hereafter IB and IC) are determined
3.6 THE KINEMATIC WAVE MODEL as follows:
𝐼𝐵=(1−𝑓)𝑟 (7)
The kinematic wave runoff model is
applied for surface runoff and stream routing 𝐼𝐶=𝑘2×𝑆𝐵 (8)
modeling under the assumption of the river
Where, SB (m) is water storage in the B layer.
channel triangle cross section. In Hydro-BEAM,
the integrated model of kinematic wave is used
4. RESULTS AND ANALYSIS
for overland flow and A layer flow as expressed
in Eqs (1), (2), (3) and (4). In this study, surface runoff is estimated
in Blue Nile using the topographical and climatic
𝜕ℎ 𝜕𝑞 data of period (1995 -1999). Figure 7 represents
+ = 𝑓𝑟 (1)
𝜕𝑡 𝜕𝑥 the location of points which the results of
discharge were calculated.
5
𝑞 = 𝛼(ℎ − 𝑑)3 + 𝑎ℎ (ℎ > 𝑑) (2)

𝑞 = 𝑎ℎ (3)

√𝑠𝑖𝑛𝜃 𝑠𝑖𝑛𝜃
𝛼= , 𝑎 = 𝐾𝑠 (4)
𝑛 𝛾

Where, r is input effective rainfall intensity

(m/s), which corresponds to the sum of rainfall
infiltration into the soil after extracting evapo-
transpiration losses. f is the direct runoff ratio,
which is equal to the soil saturation ratio in the
upper soil, A layer (0–1). Figure 7 Locations of Stations along the Blue Nile Basin

6
 The simulation has been performed using 31/12/1999). In 1995, 1996 and 1997 at
Hydro-BEAM based on JRA-25 climatic data in Khartoum station, the base flow prediction is
the target wadi basin. The results of simulation generally reasonable. However, results of peaks
were calibrated with the observed data available are underestimated. Peaks are either
in the three gauging stations shown in Figure 10. underestimated, or overestimated in different
The calibration performance of the Hydro- stations at some years. This may be due to lack
BEAM is carried out by comparing of the of accuracy in the climatic data. Meanwhile,
monthly simulated runoff with the observed data. better agreement is noticed for other months
To assess the performance of the model during the year. Whereas, in 1997, the maximum
predictability of representing the hydrological peak of simulation is 3978.16 m3/sec, however,
simulation of the reality of the basin. Discharge the maximum peak of observed data is 4336.67
of each mesh is computed all over the basin. The m3/sec. In addition, good agreement noticed
simulated hydrographs of surface runoff show between the maximum peak of simulation and
the flow features of surface runoff at the Blue the observed data in 1996 and 1997 at D.S.
Nile Basin. In addition, the simulated discharges Sennar dam station. At El Deim station, results
coincide in their characteristics with the are the same as previously in 1997-1998
observed ones as shown in Figure 8 and Figure denoting satisfied fitting between the results and
9. the observed. The model performance was
evaluated using coefficient of determination R2,
Figure 8 and Figure 9 represent the the results obtained are satisfactory and
monthly hydrograph for the calibration of the acceptable to simulate the basin runoff for future
simulated discharges and the observed flow projection Table 2.
during the simulation period (1/1/1995-

Khartoum Station
(a) (b)

(c) (d)

Figure 8 Simulated hydrographs for the period (1995-1999) at the target basin

7
 D.S. Sennar Dam Station
(e) (f)

(g)

(h)

EL Deim Station
(b)

(a)

(d)

(c)

Figure 9 Simulated hydrographs for the period (1995-1999) at the target basin

8
 Table 2 the simulation results of the period from (1995-1999) at the Blue Nile Bain

Khartoum Station

Obs. data Hydro-BEAM

Time to Total Time to Total

Peak Peak
Peak Volume Peak Volume R2
(m3/sec) (m3/sec)
(month) (Mm3/year) (month) (Mm3/year)
1995 5451.43 August 29.86 3486.34 August 20.09 0.79

1996 7135.00 August 45.28 5222.20 August 46.09 0.76

1997 4336.67 August 34.15 3978.16 August 30.27 0.84

1998 6718.57 August 57.58 11202.3 August 78.89 0.94

1999 7032.00 August 55.17 7323.59 August 56.32 0.74

D.S. Sennar Dam Station

Obs. data Hydro-BEAM

Time to Total Time to Total
Peak Peak
Peak Volume Peak Volume R2
(m3/sec) (m3/sec)
(month) (Mm3/year) (month) (Mm3/year)
1995 5117.5 August 25.4 3091.483 August 16.84 0.97

1996 4544.93 August 45.02 5657.5 August 38.6 0.99

1997 3539.47 August 28.6 3660 August 27.69 0.92

1998 6608.89 August 50.81 9804.042 August 65.76 0.95

1999 5502.31 August 36.59 6424.187 August 42.62 0.55

Distributed Maps of the simulated distributed maps are very helpful in flash flood,
discharge in Blue Nile Basin present the water resources, urban and land use
variation of discharge during the months of the management. Moreover, it can be also useful in
year. In January and February (dry Seasons), the detecting flash flood prone areas and
discharge is very low. After this dry season, the consequently mitigate and manage flash flood
discharge gradually starts to increase until it disasters in arid and semi-arid areas.
reaches its peak in August and then it decreases Furthermore, the application of such DHM
again in December as depicted in Figure 10. models can be valuable tools for wadi
Therefore, it is obvious that the discharge development and land use management to
increases in summer especially in August with a identify the best location for residential, touristic,
noticeable reduction in winter. The presented industrial, and agricultural activities [13].

9
Figure 10 Distribution maps of the simulated discharge at Blue Nile Basin during the months of the year for the simulation period
(1995-1999)

10
 5. CONCLUSIONS Blue Nile," Journal of Water Resource and
Hydraulic Engineering (JWRHE), 2005.
In this study, a physical-based distributed
hydrological model (Hydro-BEAM) has been
[5] Abd El-Aziz O. R., Valeriano O. C. S., El-Zeir
proposed for simulation the surface runoff in Blue
Nile Basin. It used to simulate the surface runoff of M. A. and Ookawara S., "Simulation Of Blue
the Blue Nile basin for the period (1995-1999). The Nile River Using A Distributed Hydrological
results of the model simulation were calibrated with Model And Global Data Sets," Conference
observed data to evaluate the model performance. Paper, Published on ResearchGate, 2011.
The advantages and findings of this work are
summarized as follow: (1) The overall performance [6] Abdel-Aziz O. R., "Flood forecasting in Blue
of the Hydro-BEAM model using JRA-25 Nile basin using a process-based hydrological
reanalysis data showed acceptable agreement with model," INTERNATIONAL JOURNAL OF
observed river discharge, (2) Simulation of the ENVIRONMENT, pp. ISSN 2091-2854,
surface runoff has been successfully achieved at Volume-3, Issue-1, Dec-Feb 2013/14.
Blue Nile Basin using Hydro-BEAM, (3) The
contributed water of the Blue Nile Basin towards the [7] Gebre S. L., "Application of the HEC-HMS
main Nile River has been evaluated during high Model for Runoff Simulation of Upper Blue
flows, (4) Understanding how to build and work
with a new generation of more complex distributed Nile," Hydrol Current Res 6, 2015.
hydrologic models, and (5) The potential use of
hydrological model to assess and predict the water [8] Mostafa H., Roushdi M. and Kheireldin K.,
availability in Nile river basins under future "Statistical Analysis of Rainfall Change over
projected climate change. the Blue Nile Basin," Conference paper
(Published on ResearchGate), 2016.

6. REFERENCES [9] Samy A., Valeriano O. C. S. and Negm A.,

"Variability of Hydrological Modeling of the
Blue Nile," World Academy of Science,
[1] Mostafa H., Saleh H., Sheikh M. El and Engineering and Technology International
Kheireldin K. , "Assessing the Impacts of Journal of Environmental, Chemical,
Climate Changes on the Eastern Nile Flow at Ecological, Geological and Geophysical
Aswan," Journal of American Science , 2016. Engineering, vol. 9, p. No: 3, 2015.

[2] Tesemma Z. K., "Long Term Hydrologic [10] Gebrehiwot S. G., Ilstedt U., Gardenas A. I.
Trends In The Nile Basin," Master's Thesis the and Bishop K., "Hydrological characterization
Faculty of the Graduate School of Cornell of watersheds in the Blue Nile Basin,
University, 2009. Ethiopia," Hydrology and Earth System
Sciences, 2011.
[3] MWRI, The Nile Basin : Bahr el Jebel after
the sudden rise in the Equatorial Lakes, vol. [11] Awulachew S. B., McCartney M., Steenhuis
XII, Cario: not published (available in MWRI T. S. and Ahmed A. A., "A Review of
library), 2014. Hydrology, Sediment and Water Resource
Use in the Blue Nile Basin," p. International
[4] Bashar K. E. and Zaki A. F., "SMA Based Water Management Institute. 87p. (IWMI
Continuous Hydrologic Simulation Of The Working Paper 131), 2008.

11
[12] Saber M., "Hydrological Approaches of Wadi
System Considering Flash Flood in Arid
Regions," Graduate School of Engineering
Kyoto University, 2010.

[13] Abd-El Fattah M., Kantoush S. and Sumi T.,

"Integrated Management Of Flah Flood in
Wadi System Of Egypt :Disaster Prevention
and Water Harvesting," Annuals of Disas.
Prev. Res. Inst. , Kyoto Univ. No. 58 B, 2015.

[14] "http://explorer.usgs.gov/," Dem. [Online].

[15] "https://rda.ucar.edu/datasets?," JRA25

Reanalysis Climatic Data. [Online].

[16] Sato Y., Ma X., Xu J., Matsuoka M., Zheng

H., Liu C. and Fukushima Y., "Analysis of
long-term water balance in the source area of
the Yellow River basin," HYDROLOGICAL
PROCESSES, 24 July 2007 in Wiley
InterScience.

[17] Sato Y., Abd-El Fattah M., Saber M.,

Michihiro Y., Hamaguchi T., Tanaka K.,
Suzuki Y., Kantoush S. A. and Sumi T.,
"Hydrological Modeling for the Nile River
using Hydro-BEAM: Program Manual.,"
Water Resources Research Center, Disaster
Prevention Research Institute, Kyoto ,611-
0011, JAPAN, 2012.

12