Environmental Earth Sciences (2019) 78:586

https://doi.org/10.1007/s12665-019-8604-6

ORIGINAL ARTICLE

Coupling HEC‑RAS and HEC‑HMS in rainfall–runoff modeling

and evaluating floodplain inundation maps in arid environments: case
study of Ain Sefra city, Ksour Mountain. SW of Algeria
Derdour Abdessamed1 · Bouanani Abderrazak2

Received: 20 March 2019 / Accepted: 23 September 2019 / Published online: 28 September 2019
© Springer-Verlag GmbH Germany, part of Springer Nature 2019

Abstract
Ain Sefra watershed is located in the south-west of Algeria that has an area of 1957 km2. In its downstream part located
Ain Sefra city where wadi Ain Sefra (junction of wadi Breidj and Tirkount) takes place. It crosses the completely urbanized
areas where it was liable to many inundations causing lots of losses either economic or loss of life, these damages resulted
from growth population and the spreading of the city over the natural space of this wadi. In this paper, the methodological
approach adopted, focused on the hydrologic modeling through the Hydrologic Engineering Center’s Hydrologic Modeling
System and the hydraulic modeling under the Hydrologic Engineering Center’s River Analysis System with combination
of Watershed Modeling System model and Geographic Information System. The aim of this study is to analyze the inun-
dation behavior of Ain Sefra city during extreme flood events by considering concrete retaining walls existence built by
local authorities and without it. In this case, three types of simulation are performed with the return periods of 10, 100, and
1000 years. The hydraulic modeling revealed that the existence of retaining walls resulted in decrease of flood zone area,
and so much less lands are endangered by floods, so that damages of flood at the study area decrease clearly, but they still
been insufficient for all return periods. The simulations also highlighted that the region most affected by the flood is the
downtown area. Finally, a protection concept is proposed to ensure a better passage of floods in Ain Sefra city, to reduce the
risk and to protect the city against the floods.

Keywords Ain Sefra · Hydrological modeling · Hydraulic modeling · Floods · HEC-HMS · HEC-RAS

Introduction risk of inundation is a crucial question nowadays (Zhang

et al. 2014). According to the UN Office for disaster risk
More people are affected by extreme weather events, espe- reduction, the Centre for Research on the Epidemiology
cially floods and drought, than by any other type of natural of Disasters: during the last 20 years, flooding has been
hazard. Flood is one of the most frequently occurring and the most common natural disaster by far, accounting for
higher impact natural disasters, that caused losing of lives 43% of all recorded events natural disasters that occurred
and properties, and it can be the highest killer weather phe- around the world between 1995 and 2015 which affected a
nomena. Floods can occur in rivers, reservoirs, and coastal further 2.3 billion people (Guha-Sapir et al. 2016). Global
areas with serious damages, when flow exceeds the capac- floods and extreme rainfall events have surged by more than
ity of the river channel. How to estimate and mitigate the 50% this decade, and they are now occurring at a rate four
times higher than in 1980 (EASAC 2013), and the number
of floods per year rose to an average of 171 in the period
* Derdour Abdessamed 2005–2014, from an annual average of 127 in the previous
samederdour@yahoo.fr decade (Guha-Sapir et al. 2016).
1 The floods represent the most widespread catastrophes
Laboratory for the Sustainable Management of Natural
Resources in Arid and Semi‑arid Zones, University Center in the Mediterranean (Ballais et al. 2005). According to the
Salehi Ahmed of Naama, 45000 Naama, Algeria Plan Bleu which is acting as a Mediterranean Observatory
2
Laboratory No. 25; BP119, University of Tlemcen, on Environment and Sustainable Development, 249 destroy-
13000 Tlemcen, Algeria ing floods touched the Mediterranean circumference in these

13
Vol.:(0123456789)
586 Page 2 of 17 Environmental Earth Sciences (2019) 78:586

20 last years (1996–2016). They affected 7861802 people, Hydrological and Hydraulic Model for Flood Prediction in
causing 3474 dead and economic losses which would rise Vietnam. Gül et al. (2010) combined the Hydrologic Engi-
with more than 25 billion euros (EM-DAT 2016). The ris- neering Center Hydrologic Modeling System (HEC-HMS)
ings and the floods in the Mediterranean often originate to and River Analysis System (HEC-RAS) to examine the
the torrential rains. The highest flood frequency corresponds potential serviceability of a planned dam in the Bostanli
to northern Mediterranean countries (Spain and Italy). How- basin in Turkey. Bonnifait et al. (2009) used TOPMODEL
ever, they are less frequent in southern Mediterranean coun- (TOPography based hydrological MODEL) with a hydraulic
tries (northern Africa), but are usually catastrophic with a model for reconstructing a catastrophic flood event in the
very high number of casualties (Llasat et al. 2010). Algeria Gard region, France. Thakur et al. (2017) coupled HEC-
is one of the Mediterranean countries vulnerable to flooding RAS and HEC-HMS to evaluate floodplain inundation map
risks linked to climatic hazards. Where, they have marked in Illinois State in USA. Berni et al. (2009) used hydro-
as one of the most frequent natural disasters and the most logic and hydraulic modeling for reliable flood risk mitiga-
destructive. According to Algerian Civil Defense Directory, tion activities in the Upper Medium Tiber River basin in
floods killed more than 1000 people in Algeria in the last Italy. Hashemyan et al. (2015) combined HEC-HMS and
20 years. While the flood of 10 November 2001 at Bab El HEC-RAS models in GIS to simulate flood in Khoshke
Oued district of Algiers is the deadliest with 781 deaths, 115 Rudan river in Fars province, Iran. Knebl et al. (2005) used
missing, and 3721 destroyed houses (Menad et al. 2012). In NEXRAD rainfall, GIS, and HEC-HMS/RAS in San Anto-
addition, there are more than 700 sites at risk of flooding in nio River Basin in USA.
Algeria, of which 100 at high risk systematically involving Many studies have been conducted also in Algeria to
a priority intervention, according to the Ministry of Water modulate rainfall–runoff relations and to calculate the catch-
Resources. The flood damages will be increased over the ments area and the flash floods discharges for different pur-
years due to population growth and socio-economic devel- poses like runoff harvesting or creating flash floods hazards
opment, and the climate change due to the global warming maps. As example: Derdous et al. (2015) used GIS-based
effect. Therefore, it is necessary to define a methodology approach for the prediction of the dam break flood hazard
to predict the flash floods in this region, to protect the city in Zardezas reservoir in Skikda. Derdour et al. (2018a) mod-
against inundations. The widely used approach to determine eled rainfall–runoff relations using HEC-HMS in a semi-
flash floods occurrence and the relationships between rain- arid region in Ain Sefra watershed. Skhakhfa and Ouerdachi
fall and runoff data is the flood inundation modeling (Der- (2016) used HEC-HMS model for the hydrological modeling
dour et al. 2017). of wadi Ressoul watershed. Astite et al. (2015) cartographed
Flood inundation modeling requires distributed model flood hazard by overflowing rivers using hydraulic modeling
predictions to inform major decisions relating to planning and geographic information system in Oued El Harrach.
and insurance (Bates 2004). In the recent years, flood haz- Yamani et al. (2016) combined HEC-HMS and HEC-RAS
ard assessment has considerably improved especially due to mapping of vulnerability of flooded area in arid region in
to the use of geographic information systems (GIS) in inte- Ghardaia city. Atallah et al. (2018) simulated the extreme
gration with hydrological and hydraulic modeling. The flood events in Sidi Bel Abbes city. Laouacheria and Man-
GIS environment has the ability to extract the hydrologi- souri (2015) used HEC-HMS model by employing fre-
cal variables needed from a good-quality digital elevation quency storm to simulate the runoff in a small urban catch-
models (DEM), such as catchments shapes, flow directions, ment in the North-East of Algeria. Bekhira et al. (2018) used
slopes, path lengths, and watershed delineation (Jenson and GR model to predict of floods in the Wadi Bechar watershed.
Domingue 1988; Wilson and Gallant 2000). Assessments This paper focuses on the simulation of the occurrence
of flood affected areas resulting from extreme precipitation of floods in Ain Sefra city which is situated in the western
and changing land use can be helpful for better understand- part of Ksour Mountains, this area subjected to severe flash
ing the flood events (Dawadi and Ahmad 2012; Mosquera- floods since the creation of the modern city in 1882 as pre-
Machado and Ahmad 2007; Ahmad and Simonovic 2006). vious records in 1904, 1990, 2000, and 2007, and the most
Efforts have also been made to integrate some hydrological recent flood was in 2014, by constructing hydrologic and
models with the GIS environment. Most of these models are hydraulic models using HEC-HMS, HEC-RAS, WMS, and
physically based distributed models, e.g., HEC-HMS, HEC- GIS, and it is based on four models, a flood hazard map
RAS, and WMS. This integration allows assessment and will be developed to help the decision-makers to identify
prediction of the impact of watershed management practices the areas within the city with high risk of flooding and to
(Zhang et al. 2008; Wheater et al. 1999; Verma et al. 2010; test the impact of developments carried out, which can help
Arnold et al. 1998). them in adopting appropriate plans to manage the risk and
Some noteworthy examples over the world are as fol- to take the necessary measures to reduce the effects of floods
lows. Dang Thanh and Florimond (2017) combined the to eliminate the life losses and to limit the material damages.

13
Environmental Earth Sciences (2019) 78:586 Page 3 of 17 586

Study area from its origin in the north side of Morghad Mountain,
and receives wadi Sfissifa (wadi: local term of channel
The region of Ain Sefra is situated in the Ksour mounts, at that is dry except in the rainy season); together, they form
the south-west of Algeria. It is located between longitudes wadi of Breidj, and they pursue the same direction to meet
(1°3′54″ and 0°26′05″W) and latitudes (32°27′5.76″ and wadi Tirkount. These two wadis (Breidj and Tirkount)
33°2′11.5″N) (Fig. 1). The Ksour mounts is a rather high cross through the city of Ain Sefra, and meet in the center
mountainous region formed by extended chain mounts of the urban area to form wadi Ain Sefra, that is run to
towards the South-west and North-east. According to the the south at downstream of the study area while receiv-
Directorate of programming and budget monitoring of the ing other wadis as Tiout and Sam, to become at the end
wilaya of Naama, the total population of Ain Sefra city is wadis Namous that is lost in the great western Erg in the
around 73,211 inhabitants, with a density of 72.85 (Hab/ Algerian Sahara at 370 km from its origin. The watershed
km2) (DPSB 2018). of Ain Sefra is moderately elongated and characterized by
Ain Sefra watershed has an area of 1957 ­km2; it is as a a fairly widespread network, steep slopes, and relatively
landlocked basin surrounded by Ksour Mountains. With impermeable terrains with variety lithology, providing a
67 km of length, the longest flow path runs southwestward favorable environment for the development of flashfloods
(Derdour et al. 2017). The occupation of watershed ground

Fig. 1  Location map of the study area

13
586 Page 4 of 17 Environmental Earth Sciences (2019) 78:586

shows that the main dominant lands in the basin are the Ain Sefra watershed is sub divided into two major sub-
grasslands (56.2%), the forest land (29.2%), bare areas basins (Breidj and Tirkount).
(6.6%), and outcrop lands (4.4%). The rest are in minor Table 1 summarizes the principal characteristics of each
proportion: agricultural lands (1.8%), sand lands (1.6%), sub-basin. Ain Sefra city presents the outlet of the water-
and urban areas (0.3%) (Derdour et al. 2017). The domi- shed; the majority of the activities are concentrated on the
nant geologic formation is the sandstone rocks (Jurassic level of the wadis of Breidj and Tirkount which generates
and Cretaceous), but they are usually clayey interspersed significant damage in case of rising.
by marl or compact quartz, their permeabilities are gen- Topographic data describe the geometry of the simu-
erally very low, except for the Albian sandstone which lated watershed system. The hydrological modeling does
constitute the most important groundwater aquifer in the not require a very fine spatial resolution; that is why,
region which is characterized by confined to semi confined we use digital elevation models (DEMs) with 30 × 30 m
conditions (Rahmani et al. 2017). Ain Sefra watershed is resolution sourced from Shuttle Radar Topographic Mis-
characterized by relatively steep slope upstream, the slope sion (SRTM). On the other hand, the hydraulic modeling
decreases downstream from the confluence of wadi Sfissifa requires a finer spatial scale. Thus, the accuracy of the
with wadi Breïdj. The ground elevation varies from 1058 calculation of water-surface elevation closely on the accu-
to 2213 m above the mean sea level. The mean elevation racy of the topographic data, and given the absence of
is 1334 m and the mean basin slope is 5.8 m/km. There topographic maps of a high spatial resolution of Ain Sefra
are three types of slopes: a relatively low slope between (0 city, we carried out a land survey along the wadis (Breidj,
and 6%) at low altitudes (1058–1400 m), occupying 79% Tirkount and Ain Sefra) using a Topcon-type GPS land
of the total area of the watershed; a moderately steep slope surveying (Fig. 2). The Topcon GPS station is able to per-
between (6 and 25%) at moderate altitudes (1400–1600 m), form measurements with extreme accuracy, “horizontal
where area occupied is about 12%; more steep slopes (over resolution of 0.004 m, and altimeter accuracy of 0.01 m”.
25%) at altitudes (1600–2213 m) occupying 8% of the total
area of the basin, these slopes give the wadi a violent and
torrential type.

Table 1  Morphometric Sub-catchment Area ­(km2) Perimeter (km) Compactness Channel slope % Stream
parameters of sub-basins index (KC) length
(km)

Tirkount 579.55 143.34 1.66 9.56 44.51

Breidj 1373.44 269.2 2.03 9.31 67

Fig. 2  Measures at the stream

centerline (a); measures above
the retaining walls (b)

13
Environmental Earth Sciences (2019) 78:586 Page 5 of 17 586

Climate conditions rains. As an example: the 1990 flood which occurred on

5th May was catastrophic where Ain Sefra city received
Meteorological data for the study area are obtained from 113.4 mm of rainfall, while the month average rainfall is
a weather station installed at Ain Sefra city. The station only 20.5 mm. Table 2 summarizes the most significant
WMO code is 605600 and its coordinates are 32°45′N, and floods.
0°35′39″W. The weather station is managed by the ONM The analysis of peak flows measured over a period of
(National office of Meteorology). The series of data avail- 32 years (from 1977 to 2008) at the stream gaging sta-
able from 1980 to 2014 provided by this station shows that tion of Ain Hadjadj (20 km downstream from Ain Sefra)
the climate of Ain Sefra is characterized by high temporal is based on the data collected from the National Hydric
and spatial irregularity. The annual average rainfall is about Resources Agency (ANRH). (Figure 3) noting that no data
185 mm, where the maximum annual rainfall is 439.8 mm, available after 2008 in this station.
registered in 2008 (Fig. 3). The average monthly rainfall is The morphology of the watershed helps to amplify
of the order of 30.64 mm for the wettest month (October), the peak flows observed at the outlets which vary from
and it is 3.92 mm for the driest month (July). The rainfall is a few ­m 3/s to 751 m 3/s where the maximum is regis-
very irregular and the rain storms are rare but intense which tered in October 24th, 2000. From 16 registered events
initiate and control the occurrence of flash flood event. The (Q max > 100 m 3/s), October is reported as a high-risk
analysis of the series indicates that the climate of the area of month with 11 floods that is represented 69% of the floods
Ain Sefra is typically semi-arid to arid. It is divided into two in the series. Also, the region has experienced three peak
main seasons: one moderate season with often rain intensi- flow surges exceeding 500 m3/s.
ties (from October to March) and another hot and dry season Floods in Ain Sefra are highly variable and irregular,
(from April to September) (Derdour et al. 2018a). short, and stormy, with very high peak flows if compared
Between 1980 and 2014, the annual temperatures at to the average modules, and they are characterized by a
the meteorological weather station of Ain Sefra city vary very fast rise, leading severe damages in Ain Sefra city,
between 15.5° C and 20.5° C, where the average annual tem- and a slow decline followed by a prolonged dry period
perature is around 17.5° C. (from April to September). Autumn is reported as risky
season. However, the flow in dry periods demonstrates
Overview of the flooding of the region of Ain Sefra a groundwater potential, which is produced by the Atlas
relief that shows a fairly strong hydrological relationship
Ain Sefra is one of the Algerian cities, which is affected between surface water and groundwater. The downtown
by risings which generally generate floods due to the over- of Ain Sefra represents the most exposed zone with the
flows of the wadi crossing the city. Throughout history, effects of the floods whose consequences are catastrophic.
Ain Sefra city witnessed flooding during several years: The known zone is located at the junction of wadi Breidj
1904, 1949, 1977, 1990, 2000, 2007, and 2014. Floods in and Tirkount.
the Ain Sefra watershed are exclusively result from heavy

Fig. 3  Distribution of peak flow

from 1977 to 2008

13
586 Page 6 of 17 Environmental Earth Sciences (2019) 78:586

Table 2  The most significant floods in Ain Sefra city

Date Description of the disaster

October 10, 1904 Significant casualties, about 26 dead including the famous writer Isabelle Ebehardt. The French general Girault denounced
the stupidity of the architects, engineers and officers who threw the first foundations of the modern city. For him, the
disaster was not only due to fatality (Girault 1905)
March 17, 1941 Exceptional floods. Extensive material damage was registered at the city of Ain Sefra (Dubief 1955)
May 20, 1949 Major flood, it causes important damage to several places in the city of Ain Sefra (Dubief 1955)
October 26, 1977 With a peak flow estimated at 373 m3/s, it caused significant material damage in the popular neighborhoods of the city of
Ain Sefra. The flood lasted nearly 35 h; the peak flow was reached after 8 h, while the recession took 27 h to return to the
initial flow (three times the rise time)
May 05, 1990 Huge flooding. Very important material damage; water levels reached almost 2 m in the neighborhoods of the downtown of
Ain Sefra. The gauging station registered a peak flow of 392.72 m3/s and a volume of 27.71 million cubic meters during
the flood period
October 22, 2000 A family of five people died; several inhabitants were forced to leave the places giving place to the furies of the waters. The
streets and alleys were completely paralyzed, important damage. The rise of floods lasted 6 h and the recession took 29 h
to return to the initial flow (five times the rise time). Flow rates reached a record with a peak of 750.8 m3/s
November 14, 2004 The torrential rains that hit the area caused extensive material damage in the populous areas of the city of Ain Sefra. The
streets and alleys were completely paralyzed, significant material damage. With a peak of 206 m3/s
October 19, 2007 Major damage in downtown Ain Sefra. Many bridges at the watershed level have been damaged, with a peak of 221.7 m3/s
May 31, 2011 Considerable damage in downtown Ain Sefra
October 1, 2014 An 11-year-old child swept away by the floods. Extensive damage

Methodology • The second step, defining basin delineation, stream net-

work, topography, and geometric characteristics by DEM
The physically based conceptual approach with available processing, using ARC GIS and WMS model (Myronidis
data that provides more insight into the physical nature of et al. 2016).
the watershed (surface and topography) and meteorologi- • The third step, modeling rainfall/runoff relation, using
cal (rainfall and runoff) has been adopted. In our study, HEC-HMS model. Which includes four steps (1) sub-
flood inundation modeling was conducted using HEC- basin delineation; (2) hydrologic parameter estimation;
HMS after passing with WMS (Watershed Modeling (3) model set-up by building basin model, meteorologi-
System) model, to construct an HEC-RAS flow model. cal model, and control specifications; (4) calibration and
The advantage of using WMS is its interchangeability validation (Scharffenberg and Fleming 2010).
with GIS, HEC-HMS, and HEC-RAS. WMS can perform • The fourth step, creating the triangulated irregular net-
operations such as automated basin delineation, geometric work (TIN) of the city of Ain Sefra with high resolution
parameter calculations, GIS overlay computations, cross- (altimeter accuracy of 0.01 m) and rich in topographic
section extraction from terrain data, floodplain delineation, details established from topographic land surveys. That
mapping, and storm drain analysis (Myronidis et al. 2009). represents the topography of the study area.
The first step in the development of a floodplain analy- • The fifth step, extracting cross-sections from the TIN in
sis is to collect highly detailed information grouped into WMS,
three categories: hydrologic, hydraulic, and topographic. • The sixth step, importing the hydrologic modeling results
Hydrologic and hydraulic data consist of rainfall, dis- to be used as input data in hydraulic model HEC-RAS.
charge hydrographs, and stage historical records. They • The seventh step, creating a conceptual model such as
are obtained from the National Hydric Resources Agency the river reaches, the stream centerline, the main chan-
(ANRH) and the National Office of Meteorology (ONM) nel banks, the cross-section lines, and the material zones
at Ain Sefra city. which are called channel geometry.
The proposed method consists of ten steps after data • The eighth step, defining the network schematic on HEC-
collection: RAS using Manning’s roughness coefficient (n) for every
land-use type.
• The first step is to build the intensity curve–dura- • The ninth step, applying a steady gradually varied flow
tion–frequency (IDF), and then estimates the peak dis- analysis which consists of peak discharge information,
charges (Qmax) using a statistical method for different boundary conditions, and flow regime to compute water-
average recurrence interval. surface elevation along the channel geometry of wadi
Breidj, Tirkount, and Ain Sefra.

13
Environmental Earth Sciences (2019) 78:586 Page 7 of 17 586

• The final step, floodplain delineation is conducted to Description of the software used
obtain the flood inundation area involving a flood depth
and a flood extent for the wadis Breidj and Tirkount and Runoff is modeled using the Hydrologic Engineering
their junction wadi Ain Sefra using WMS and HEC-RAS. Center’s Hydrologic Modeling System (HEC-HMS),
In this case, three types of simulation will be carried out: version 4. HEC-HMS is a physically based, semi-distrib-
(1) determining the flood zones with the return periods of uted hydrologic model developed by the US Army Corps
10, 100, and 1000 years without taking in consideration of Engineers (Scharffenberg and Fleming 2010). It is
hydraulic developments which are done in the city of Ain designed to simulate the complete hydrologic processes of
Sefra, by the water resources directorate; (2) determining dendritic watershed systems. The software includes many
the flood zones with the same previous return periods but traditional hydrologic analysis procedures such as event
with taking into account hydraulic developments; (3) the infiltration, unit hydrographs, and hydrologic routing.
third simulation taking into account the solutions that we HEC-HMS also includes procedures necessary for con-
propose to reduce flood damage in the area. tinuous simulation including evapo-transpiration, snow-
melt, and soil moisture accounting. Advanced capabilities
The methodology of the study is illustrated in Fig. 4. are also provided for gridded runoff simulation using the
linear quasi-distributed runoff transform (ModClark).
Supplemental analysis tools are provided for model opti-
mization, forecasting streamflow, depth-area reduction,

Fig. 4  Methodology adopted for

the study

13
586 Page 8 of 17 Environmental Earth Sciences (2019) 78:586

assessing model uncertainty, erosion and sediment trans- Results and discussion
port, and water quality. HEC-HMS is designed to be
applicable in a wide range of geographic areas for solving During the land survey of Ain Sefra city along the minor and
the widest possible range of problems. This includes large major riverbeds, 1925 points were registered on a linear of
river basin water supply and flood hydrology, and small 3.3 km to obtain the triangulated irregular network (TIN) of
urban or natural watershed runoff. Hydrographs produced the city with high resolution (Fig. 5). This TIN resolution is
by the program are used directly or in conjunction with very favorable to be applied for a floodplain area to assess
other software for studies of water availability, urban flood hazards.
drainage, flow forecasting, future urbanization impact, A total of 55 cross-sections were taken over the reaches
reservoir spillway design, flood damage reduction, flood- modeled as shown in Fig. 6. The great number of cross-sec-
plain regulation, and system operation (Scharffenberg and tions was chosen for more details of flood maps. Roughness
Fleming 2016). The program has an extensive array of coefficients used in the study area were 0.04 for cropland,
capabilities for conducting hydrologic simulation. Many 0.08 for urbanized area, 0.035 for rivers, and 0.01 for con-
of the most common methods in hydrologic engineering crete in the retaining walls zones according to Henderson
are included in such a way that they are easy to use. The (1966).
program does the difficult work and leaves the user free
to concentrate on how best to represent the watershed
environment (Scharffenberg and Fleming 2016). Analysis of rainfall and runoff data
Flood propagation and inundation levels in the flood-
plain are simulated with the Hydrologic Engineering The records from Ain Sefra rainfall station were obtained
Center’s (HEC) River Analysis System (HEC-RAS) and analyzed to establish the intensity–duration–frequency
developed by the Hydrologic Engineering Center of the (IDF) curves based on extreme value, to evaluate the water-
US Army Corps of Engineers (USACE). HEC-RAS cal- shed reaction to a given rainfall event. In this study, we
culates one-dimensional steady and 1D and 2D unsteady assumed that the entire watershed would receive the same
flow. HEC-RAS is an integrated system of software, amount of rainfall. The rainfall IDF results adopted for the
designed for interactive use in a multi-tasking, multi- area for various storm durations and average recurrence
user network environment. The system is comprised interval (ARI) are shown in Fig. 7.
of a graphical user interface (GUI), separate hydraulic After having tested several statistical methods employed
analysis components, data storage and management capa- in flood frequency analysis (FFA) (exponential, GUV,
bilities, graphics, and reporting facilities. The HEC-RAS gamma, Gumbel, normal, and log-normal), peak discharges
system contains four hydraulic analysis components for: of wadi Ain Sefra during 32 years fit well with the Gum-
(1) steady-flow water-surface profile computations; (2) bel approach (Derdour et al. 2018b). This statistical method
one and two-dimensional unsteady-flow simulations; (3) employed allows us to estimate the peak discharges (Qmax)
movable boundary sediment transport computations; and for different average recurrence interval (Table 3).
(4) water temperature and constituent transport modeling
(Brunne 2016).
The hydraulic model requires as input and the output Hydrologic modeling results
hydrographs from HMS; its parameters are representative
cross-sections for each sub-basin, including left and right The hydrologic modeling is based in a previous study done
bank locations, roughness coefficients (Manning’s n), and by Derdour et al. (2018a); which simulate runoff in Ain
contraction and expansion coefficients. Roughness coeffi- Sefra watershed through the employing of the Hydrologic
cients, which represent a surface’s resistance to flow and Engineering Center—Hydrologic Modeling System (HEC-
are integral parameters for calculating water depth, were HMS) where the frequency storm is used for the meteoro-
estimated by combining land-use data with tables of Man- logical model, the Soil Conservation Service curve number
ning’s n values. The version 5.0 of HEC-RAS supports (SCS-CN) is selected to calculate the loss rate and Soil Con-
steady- and unsteady-flow water-surface profile calcula- servation Service unit hydrograph method has been applied
tions, sediment transport computations, and water-quality to simulate the runoff rate.
analyses. The software also contains tools for performing The calibrated HEC-HMS model was used to calculate
inundation mapping directly inside the software. A full direct peak discharges for wadi Ain Sefra and for the two
description of the model and its computational schemes ungauged sub-catchments (Breidj and Tirkount) for various
is given by Brunne (2016). average recurrence intervals (10, 100 and 1000 years). The
Nash–Sutcliffe efficiency coefficient was 0.95, which indi-
cates that the hydrological modeling results are satisfactory

13
Environmental Earth Sciences (2019) 78:586 Page 9 of 17 586

Fig. 5  TIN of the study area

and accepted for simulation of rainfall–runoff (Derdour et al. 1 profile every 60 m. To have an even finer space step,
2018a). Table 4 shows the peak discharge flowing from each we created cross profiles by interpolation of the extracted
catchment. profiles, and we added a profile as soon as the distance
between two profiles exceeded 20 m, so the flood path is
Hydraulic modeling results defined through 165 cross-sections.
The cross-sections width is 100 m for wadi Breidj and
For hydrodynamic modeling, the model HEC-RAS was Tirkount, and 150 m for wadi Ain Sefra which represent
considered. This model solves the problems of the flows the downtown zone. Roughness coefficients (Manning’s)
on free surface permanent or non-permanent. It solves used in the study area were 0.04 for cropland, 0.08 for
the equation of unidimensional energy, the losses being urbanized area, 0.035 for rivers, and 0.01 for concrete
evaluated by the formula of friction at the bottom of Man- in the retaining walls zones (Henderson 1966). The flow
ning–Strickler and by formulas of contraction/expansion regime in HEC-RAS is set to be mixed. The upstream
of the flow. Hydraulic modeling is used to produce flood boundary condition is set as a critical depth, while the
hazard maps in Ain Sefra city. The peak discharge data downstream end is set as a normal depth. After fulfilling
for each catchment calculated with HEC-HMS are used as all the input data and boundary conditions in HEC-RAS,
an input to the HEC-RAS to model the water depths and the simulation is done as a steady-state analysis.
along the flood path. The result was imported and readied The flood hazard maps were prepared by considering
by WMS for all return periods. In HEC-RAS, the flood of concrete retaining wall existence and without it. In this
path is defined through 55 cross-sections extracted from case, three types of simulation are performed:
the TIN of the study area. The initial model has on average

13
586 Page 10 of 17 Environmental Earth Sciences (2019) 78:586

Fig. 6  Extracting cross-sections of the study area

Fig. 7  Curves IDF “intensity–

duration–frequency”

First simulation 1000 years without taking into consideration retaining walls
built by local authorities in their flood-fight effort. The peak
The purpose of the first simulation is to determining the discharge calculated by HEC-HMS was inputted into and
flood hazard map with the return periods of 10, 100, and generated by HEC-RAS; the result was imported and readied

13
Environmental Earth Sciences (2019) 78:586 Page 11 of 17 586

Table 3  Peak discharges with Flow ­(m3/s) when average width of the flooded areas reached during the 100-year flood
Gumbel approach recurrence interval (years) and 1000-year flood was 501 m and the 600 m, respectively.
10 100 1000
Second simulation
463 845 1120

After the flood of 19/10/2007, and to stabilize the banks

against erosion and to prevent overflowing, local authori-
by WMS for all return periods. The WMS interpolate the ties of wilaya of Naama, built concrete retaining walls in
results of water-surface elevation data and delineate the several vulnerable areas in Ain Sefra city. The realized linear
flood inundation, as shown in Figs. 8, 9, and 10 for 10-year, is about 1.9 km and total cost of the operation is 3.2 million
100-year, and 1000-year flood. This simulation revealed that Euros (Figs. 11, 12). The aim of the second simulation is to
the region most affected by the flood is the downtown area. testing the performance of concrete retaining walls already
This region represents the junction of wadis Breidj and Tirk- built by local authorities in their flood-fight effort.
ount; it is marked by low altitudes. The water depth in this Depictions of the 10 years, 100 years, and 1000 years
area exceeds 4.44 m downstream for a 10-year flood, and the of flood hazard maps were simulated to provide a

Table 4  Peak discharge (­ m3/s) Average recurrence Hypothetical storm Simulated peak discharges (HEC-HMS) ­(m3/s)
for the sub-catchment interval years (mm)
Wadi Breidj Wadi Tirkount Wadi Ain Sefra

10 56.9 301.1 227.0 425.8

100 106 641.9 473.0 904.3
1000 166 943.5 687.3 1328.3

Fig. 8  Flood inundation map in Ain Sefra city (return period 10 years)

13
586 Page 12 of 17 Environmental Earth Sciences (2019) 78:586

Fig. 9  Flood inundation map in Ain Sefra city (return period 100 years)

representation of the floodplain boundaries that are com- for the return periods of 10,100 years. For our study, we
monly used for flood insurance purposes. The model HEC- proposed the combination of four types of development:
RAS shows its capacity as representation of extended of the
flood and to specialize well the elements exposed at this • Raising the retaining walls in the overflow areas, follow-
risk in three-dimensional view (3D) and by consequence to ing the 10-year and 100-year flow simulations (from 0.3
lead to a tool of decision-making integrated well. Figure 13 to 0.8 cm depending on the overflow areas).
demonstrates the flood inundation map of the study area for • Building retaining walls in areas that have not benefited
the 10-year and 100-year flood. In these sections, we remark from development (right bank of the wadi Tirkount, and
that the retaining walls are generally adequate for 10-year left bank and part of the right bank of the wadi Breidj).
floods, where wadi Ain Sefra has no overflow. However, • Ecological engineering, by the planting plants in the
there are some overflows for the 100-year floods. While stream bed, introduced in HEC-RAS by changing of the
seen in Fig. 14, the protections done are still insufficient for Manning values (n = 0.03).
1000-year flood for the three wadis (Breidj, Tirkount and • Widen the stream bed from their natural form to a rela-
Ain Sefra). These simulations revealed that the existence of tively rectangular shape and clean up the wadis, because
retaining walls resulted in decrease of the flood zone area, they are used as deposition area of construction rubble
and so much less lands are endangered by floods, so that and garbage dump. These deposits have the direct conse-
environmental effects and damages of flood at the study area quence of reduction of sections of the wadi. The deposit
decrease clearly. of garbage dump in the wadi must be forbidden.

Third simulation The superposition of the flood hazard maps for the two
simulated scenarios (before and after proposing develop-
The aim of the third simulation is to test the solutions that ments) (Fig. 15) shows clearly the effect of the proposed
we propose to reduce flood damage in the city of Ain Sefra developments on reducing the risk of flooding over the city.

13
Environmental Earth Sciences (2019) 78:586 Page 13 of 17 586

Fig. 10  Flood inundation map in Ain Sefra city (return period 1000 years)

Fig. 11  Retaining walls in wadi Breidj Fig. 12  Retaining walls in wadi Ain Sefra

The flood inundation map in 3D shows the reassured pas- Conclusion

sage of the 100-year flood at the level of the city of Ain
Sefra (Fig. 16). It should be noted that the fact of protecting Ain Sefra is one of the Algerian cities that had been expe-
the city from the risk of all frequencies (1000-year flood for rienced several devastating floods since the creation of
example) is an infinite cost, and the most efficient arrange- the modern city in 1882. GIS techniques and watershed
ment is that which ensures the optimum protection against modeling are highly effective research tools in mapping,
the risk.

13
586 Page 14 of 17 Environmental Earth Sciences (2019) 78:586

Fig. 13  Flood inundation map in Ain Sefra city in 3D (return period 10 and 100 years)

Fig. 14  Flood inundation map in Ain Sefra city in 3D (return period 1000 years)

investigation, and modeling surface runoff, especially in and GIS. The synthesis of the risks related to the floods is
arid regions where lack of proper hydrometeorological based on three principal concepts: hydrological, hydrody-
data is available. This paper has presented an effective namic, and cartographical. In the hydrological modeling,
approach to estimate the peak discharge of flows developed the HEC-HMS model was applied to Ain Sefra watershed
from high rainfall events uniformly cover Ain Sefra water- to predict the surface runoff after passing with WMS. The
shed, and to predict and assess downstream flood hazard peak discharges obtained for the 10-, 100-, and 1000-year
by the integration of hydrologic and hydraulic modeling storms are, respectively, 425.8, 904.3, and 1328.3 m 3/s.

13
Environmental Earth Sciences (2019) 78:586 Page 15 of 17 586

Fig. 15  Flood inundation map in Ain Sefra city (return period 100 years) after and before the proposed developments

Fig. 16  Flood inundation map in Ain Sefra city in 3D view (return period 100 years) after the proposed developments

13
586 Page 16 of 17 Environmental Earth Sciences (2019) 78:586

The results obtained are very satisfactory. Therefore, run- authorities that accompany the various development projects
offs generated by frequency storm method are invaluable in the city of Ain Sefra.
for the hydraulic modeling. To generalize and examine the reliability of the approach
The hydraulic modeling carried out by the HEC-RAS followed, the methodology of coupling of hydrological and
model based on the data of the hydrological modeling and hydraulic models can be adapted for the use of other catch-
on the TIN of the city of Ain Sefra created following the ment area of arid zones. It can form an important tool for
land survey allowed us to highlight several scenarios over- the management of flood control and for real-time simula-
flow at the city of Ain Sefra. Flood inundation maps were tion of inundation in Algeria to prevent or reduce damage in
prepared by considering of concrete retaining walls exist- terms of lives, property, and infrastructure. Further research
ence and without it. In this case, three types of simulation must be done on sediment transport using numerical mod-
are performed: els, since these models use water depths and velocities as
an input.
• For the first simulation, that represents simulation with-
out taking into consideration retaining walls built by
local authorities, we remark that the city of Ain Sefra Compliance with ethical standards
is vulnerable to flooding even for a 10-year flood and
the flood zone is expanding from the 10-year flood to Conflict of interest The authors declare that they have no conflict of
interest.
the 1000-year flood. This simulation revealed that the
region most affected by the flood is the downtown area
which represents the junction of wadis Breidj and Tirk-
ount. References
• For the second simulation, and to test the performance of
Ahmad S, Simonovic SP (2006) An intelligent decision support system
concrete retaining walls already built by local authorities
for management of floods. Water Resour Manag 20(3):391–410.
in their flood-fight effort. We remark that the retaining https​://doi.org/10.1007/s1126​9-006-0326-3
walls are generally adequate for 10-year floods, where Arnold JG, Srinivasan R, Muttiah RS, Williams JR (1998) Large
wadi Ain Sefra has no overflow. However, there are some area hydrologic modeling and assessment—Part 1: model
development. J Am Water Resour Assoc 34:73–89. https​://doi.
overflows for the 100-year floods, while the protections
org/10.1111/j.1752-1688.1998.tb059​61.x
done remain insufficient for 1000-year for the three wadis Astite SW, Medjerab A, Belabid N, El Mahmouhi N, El Wartiti M,
(Breidj, Tirkount, and Ain Sefra). These simulations Kemmou S (2015) Cartography of flood hazard by overflowing
revealed that the existence of retaining walls resulted in rivers using hydraulic modeling and geographic information sys-
tem: Oued El Harrach case (North of Algeria). Revista de Telede-
decrease of flood zone area, and so much less lands are
teccion 44:67–79. https​://doi.org/10.4995/raet.2015.3985
endangered by floods, so that environmental effects and Atallah M, Hazzab A, Seddini A, Ghenaim A, Korichi K (2018) Inun-
damages of flood at the study area decrease clearly. dation maps for extreme flood events: case study of Sidi Bel
• The third simulation is based on the diagnosis of hydrau- Abbes city, Algeria. J Water Land Dev. https​://doi.org/10.2478/
jwld-2018-0021
lic modeling, which made it possible to target the over-
Ballais JL, Garry G, Masson M (2005) Contribution of hydrogeomor-
flow sites of wadis. The aim is to take into account devel- phological method to flood hazard assessment: the case of French
opment proposals that minimize the risk of flooding by Mediterranean region. CR Geosci 337(13):1120–1130. https:​ //doi.
taking protection measures along the wadi banks, such org/10.1016/j.crte.2005.06.010
Bates PD (2004) Remote sensing and flood inundation modeling.
as: elevating the retaining walls, creating new protection
Hydrol Process 18:2593–2597. https​://doi.org/10.1002/hyp.5649
walls, cleaning the wadis, and recalibrating the minor Bekhira A, Habi M, Morsli B (2018) Hydrological modeling of floods
riverbeds. This simulation is devoted clearly the effect in the wadi Bechar watershed and evaluation of the climate impact
of the proposed developments on reducing the risk of in arid zones (southwest of Algeria). Appl Water Sci. https​://doi.
org/10.1007/s1320​1-018-0834-3
flooding over the city. It should be noted that the fact of
Berni N, Brocca L, Giustarini L, Pandolfo C, Stelluti M, Melone F,
protecting the city from the risk of all frequencies (1000- Moramarco T (2009) Coupling hydrological and hydraulic mod-
year flood for example) is an infinite cost, and the most eling for a reliable flood risk mitigation activities in the Upper-
efficient arrangement is that which ensures the optimum Medium Tiber River basin. Geophys Res Abstr 11:9498. https​://
doi.org/10.1061/41114​(371)241 (EGU2009)
protection against the risk.
Bonnifait L, Delrieu G, Le Lay M, Boudevillain B, Masson A, Bel-
leudy P, Gaume E, Saulnier GM (2009) Distributed hydrologic
The obtained results have led us to conclude that the and hydraulic modelling with radar rainfall input: reconstruction
simulations made by the chosen models are encouraging. of the 8–9 September 2002 catastrophic flood event in the Gard
region, France. Adv Water Resour 32:1077–1089. https​://doi.
This work is a first contribution to the study of floods in
org/10.1016/j.advwa​tres.2009.03.007
the Ain Sefra watershed. It is a work whose results are a Brunne (2016) HEC-RAS River Analysis System, v5: users manual.
tool for guidance, and decision support to local and regional USACE, Hydrologic Engineering Center, Davis

13
Environmental Earth Sciences (2019) 78:586 Page 17 of 17 586

Dang Thanh M, Florimond D (2017) A combined hydrological and Menad W, Douvinet J, Beltrando G, Arnaud-Fassetta G (2012) Evalu-
hydraulic model for flood prediction in Vietnam applied to the ating the influence of urbanised areas in view of a remarkable
Huong River basin as a test case study. Water 9(11):879. https​:// meteorological hazard: the flood event of 9–10 November 2001
doi.org/10.3390/w9110​879 at Bab-el-Oued (Algiers, Algeria). Géomorphologie 18:337–350.
Dawadi S, Ahmad S (2012) Changing climatic conditions in the https​://doi.org/10.4000/geomo​rphol​ogie.9954
Colorado River basin: implications for water resources man- Mosquera-Machado S, Ahmad S (2007) Flood hazard assessment of
agement. J Hydrol 430:127–141. https​://doi.org/10.1016/j.jhydr​ Atrato River in Colombia. Water Resour Manag 21(3):591–609.
ol.2012.02.010 https​://doi.org/10.1007/s1126​9-006-9032-4
Derdour A, Bouanani A, Baba-Hamed K (2017) Hydrological mod- Myronidis D, Emmanouloudis D, Stathis D, Stefanidis P (2009) Inte-
eling in semi-arid region using HEC-HMS model. Case study in grated flood risk mapping in the framework of E.Y. directive of
Ain Sefra watershed, Ksour Mountains (SW, Algeria). J Fundam the assessment and management of flood risks. Frasenlus Environ
Appl Sci 92:1027–1049. https​://doi.org/10.4314/jfas.v9i2.27 Bull 18(1):102–111
Derdour A, Bouanani A, Baba-Hamed K (2018a) Modelling rainfall Myronidis D, Stathis D, Sapountzis M (2016) Post-evaluation of flood
runoff relations using HEC-HMS in a semi-arid region: case study hazards induced by former artificial interventions along a Coastal
in Ain Sefra watershed, Ksour Mountains (SW Algeria). J Water Mediterranean settlement. J Hydrol Eng. https​://doi.org/10.1061/
Land Dev 36:45–55. https​://doi.org/10.2478/jwld-2018-0005 (ASCE)HE.1943-5584.00014​13
Derdour A, Bouanani A, Baba-Hamed K (2018b) Floods typology Rahmani A, Bouanani A, Kacemi A, Baba-Hamed K (2017) Contribu-
in semiarid environment: case of Ain Sefra watershed (Ksour tion of GIS for the survey and the management of water resources
Mountains, Saharaian Atlas, SW of Algeria). J Fundam Appl Sci in the basin “Belhandjir-Tirkount” (Ain Sefra) mounts of Ksour.
10(3):430–447. https​://doi.org/10.4314/jfas.v10i3​.29 Saharian Atlas-Algeria. J Fundam Appl Sci 9(2):829–846. https​
Derdous O, Djemili L, Bouchehed H, Tachi SE (2015) A GIS based ://doi.org/10.4314/jfas.v9i2.14
approach for the prediction of the dam break flood hazard. A case Scharffenberg W, Fleming M (2010) Hydrologic modeling system
study of Zardezas reservoir “Skikda, Algeria”. J Water Land Dev HEC-HMS v3.5: users manual. USACE, Hydrologic Engineer-
27:15–20. https​://doi.org/10.1515/jwld-2015-0020 ing Center, Davis
DPSB (2018) Monographie de la wilaya de Naama. Directorate of Scharffenberg W, Fleming M (2016) Hydrologic modeling system
Programming and Budget Monitoring of the wilaya of Naama. HEC-HMS v4.2: users manual. USACE, Hydrologic Engineer-
Statistical report ing Center, Davis
Dubief J (1955) Essai sur l’hydrologie superficielle au Sahara. PhD Skhakhfa ID, Ouerdachi L (2016) Hydrological modelling of wadi Res-
Thesis, University of Algiers, Algiers, Algeria soul watershed, Algeria, by HEC-HMS model. J Water Land Dev
EASAC (2013) Trends in extreme weather events in Europe: impli- 31:139–147. https​://doi.org/10.1515/jwld-2016-0045
cations for national and European Union adaptation strategies. Thakur B, Parajuli R, Kalra A, Ahmad S, Gupta R (2017). Coupling
Policy report 22 HEC-RAS and HEC-HMS in precipitation runoff modelling and
EM-DAT (2016) International disaster database Centre for Research evaluating flood plain inundation map. 240–251. Sacramento,
on the Epidemiology of Disasters (CRED). Office of US Foreign California: World Environmental and Water Resources Congress.
Disaster Assistance https​://doi.org/10.1061/97807​84480​625.022
Guha-Sapir D, Hoyois Ph, Below R (2016) Annual disaster statistical Verma AK, Jha MK, Mahana RK (2010) Evaluation of HEC-HMS
review 2016: the numbers and trends. CRED, Brussels and WEPP for simulating watershed runoff using remote sens-
Girault E (1905) Une Colonie d’enfer chroniques d’un voyage en Algé- ing and geographical information system. Paddy Water Environ
rie en 1904. Librairie internationaliste, Alfortville 8:131–144. https​://doi.org/10.1007/s1033​3-009-0192-8
Gül GO, Harmancioglu N, Gül A (2010) A combined hydrologic and Wheater HS, Jolley TJ, Onof C, Mackay N, Chandler RE (1999)
hydraulic modelling approach for testing efficiency of structural Analysis of aggregation and disaggregation effects for grid-based
flood control measures. Nat Hazards 54:245–260. https​://doi. hydrological models and the development of improved precipita-
org/10.1007/s1106​9-009-9464-2 tion disaggregation procedures for GCMs. Hydrol Earth Syst Sci
Hashemyan F, Khaleghi MR, Kamyar M (2015) Combination of HEC- 3:95–108. https​://doi.org/10.5194/hess-3-95-1999
HMS and HEC-RAS models in GIS in order to simulate flood Wilson JP, Gallant JC (2000) Digital terrain analysis. In: Wilson JP,
(case study: Khoshke Rudan river in Fars province, Iran). Res J Gallant JC (eds) Terrain analysis: principles and applications.
Recent Sci 4(8):122–127 Wiley, New York, pp 1–27
Henderson PM (1966) Open-channel flow: New York. MacMillan, Yamani K, Hazzab A, Sekkoum M, Toumi S (2016) Mapping of
Basingstoke, p 522 vulnerability of flooded area in arid region. Case study: area of
Jenson S, Domingue J (1988) Extracting topographic structure from Ghardaïa-Algeria. Model Earth Syst Environ 2:1–17. https​://doi.
digital elevation data for geographic information system analysis. org/10.1007/s4080​8-016-0183-x
Photogramm Eng Remote Sens 54:1593–1600 Zhang XS, Srinivasan R, Debele B, Hao FH (2008) Runoff simulation
Knebl MR, Yanga ZL, Hutchisonb K, Maidmentc DR (2005) Regional of the headwaters of the Yellow River using the SWAT model
scale flood modeling using NEXRAD rainfall, GIS, and HEC- with three snowmelt algorithms. J Am Water Res Assoc 44:48–61.
HMS/RAS: a case study for the San Antonio River basin sum- https​://doi.org/10.1111/j.1752-1688.2007.00137​.x
mer 2002 storm event. J Environ Manag 75:325–336. https​://doi. Zhang W, Yanhong X, Yanru W, Hong P (2014) Modeling sediment
org/10.1016/j.jenvm​an.2004.11.024 transport and river bed evolution in river system. J Clean Energy
Laouacheria F, Mansouri R (2015) Comparison of WBNM and HEC- Technol 2:175–179. https​://doi.org/10.7763/jocet​.2014.v2.117
HMS for runoff hydrograph prediction in a small urban catchment.
Water Resour Manag 29:2485–2501. https​://doi.org/10.1007/ Publisher’s Note Springer Nature remains neutral with regard to
s1126​9-015-0953-7 jurisdictional claims in published maps and institutional affiliations.
Llasat MC, Llasat-Botija M, Prat MA, Porc F, Price C, Mugnai A,
Lagouvardos K, Kotroni V, Katsanos D, Michaelides S, Yair Y,
Savvidou K, Nicolaides K (2010) High-impact floods and flash
floods in Mediterranean countries: the FLASH preliminary
database. Adv Geosci 23:47–55. https​://doi.org/10.5194/adgeo​
-23-47-2010

13