PAN-AFRICAN UNIVERSITY

INSTITUTE FOR WATER AND ENERGY SCIENCES

(including CLIMATE CHANGE)

Master Dissertation
Submitted in partial fulfillment of the requirements for the Master degree in
[Water Engineering]

Presented by

N’guessan Aimé KONAN

Urban Flood Modeling and Floodplain Mapping Using ArcGIS, HEC-HMS

and HEC-RAS in Abidjan City, Côte D’Ivoire – West Africa: Case Study of
Bonoumin – Rivièra Palmeraie Watershed

Defended on 04/08/2018 Before the Following Committee:

Chair Joseph Adelegan Prof. University T. Limpopo, South Africa

Supervisor Yasser Eldeberky Prof. Helwan University, Egypt
External Examiner Abdessamed Derdour Dr. University of Naama, Algeria
Internal Examiner Latéfa Sail Dr. University of Tlemcen, Algeria
DECLARATION

I, N’guessan Aimé KONAN, hereby declare that this thesis represents my personal work,
realized to the best of my knowledge. I declare that all information, material and results from
other works presented here, have been fully cited and referenced in accordance with the
academic rules and ethics.

Signed: Date of submission ___ 06/08/2018________

N’guessan Aimé KONAN

ii
CERTIFICATION

This thesis has been submitted with my approval as the supervisor

Signed: Date ___06/08/2018______________

Prof. Dr. Eng. Yasser Ahmed Eldeberky

Helwan University, Cairo, Egypt - Faculty of Engineering at Mataria

Senior Marine Projects Specialist at Environment Department of Dubai Municipality, UAE

iii
ABSTRACT

Since 2009, Côte D’Ivoire has been experiencing floods events each rainy season in general
and particularly the capital city Abidjan. According to the Office for the Coordination of
Humanitarian Affairs (OCHA, 2014), from 2009 to 2014, an average of 13 flood-related deaths
per year were recorded in Abidjan. The same report has shown that 26% of the Abidjan District
is at risk of flooding (Danumah, 2016).

Flood modelling and simulation assist in the prediction of the hazard for better flood
preparedness and thus reduce flood damages. The study has simulated flood occurrence at the
watershed of Bonoumin – Rivièra Palmeraie which is an urban area in the capital city Abidjan,
south Côte D’Ivoire. Digital elevation model (DEM) for this area was processed in the ArcGIS
10.3 environment (HEC-GeoHMS) using terrain pre-processing tools to delineate the basin,
sub-basins, and stream network. And then results from the terrain pre-processing were used to
extract the hydrologic parameters of the watershed. These hydrologic parameters were used in
the estimation of runoff in HEC-HMS. The peak discharges generated by HEC-HMS were used
in HEC-RAS for steady hydraulic simulations whose purpose was to show the conveyance of
stormwater through the canal of “Rue minister” and also flood wage propagation for further
floodplain delineation in RAS Mapper/HEC-GeoRAS and ArcGIS. After the calibration of the
hydrologic model and computations of discharges for specific return periods, HEC-RAS model
was performed with return periods of 5, 10, 25, 50 and 100 years and flood extents associated
to them were determined. Thus, flood map of the watershed was created for visualization and
decision making, along with strategies for flood preparedness.

Key words: Cocody, Bonoumin watershed, Flood modelling, HEC-HMS, HEC-RAS,

ArcGIS

iv
RESUME

Depuis 2009, la Côte D’Ivoire fait face à des phénomènes d’inondation chaque saison des pluies
dans plusieurs localités du pays en général et en particulier à Abidjan, la capitale économique.
Selon l’Office pour la Coordination des Actions Humanitaires (OCAH, 2014), de 2009 à 2014
13 personnes en moyenne ont perdu la vie chaque année. Le même rapport a mentionné que
26% du territoire de la ville d’Abidjan est à risque d’inondation (Danumah, 2016).

La modélisation et la simulation des inondations aident à la prédiction des dangers pour une
meilleure préparation contre les aléas et ainsi à réduire les dégâts liés aux inondations. Cette
étude a porté sur la simulation de l’occurrence du phénomène d’inondation dans le bassin
versant de Bonoumin – Rivièra Palmeraie, une zone urbaine de la capitale économique Abidjan,
au sud de la Côte D’Ivoire. Le model numérique de terrain (MNT) a été traité sous ArcGIS 10.3
par le biais de HEC-GeoHMS en utilisant les outils de prétraitement de terrain pour la
délimitation du bassin versant et sous bassin et l’extraction du réseau hydrographique. Ensuite
les résultats du prétraitement de terrain ont été utilisés pour extraire les paramètres
hydrologiques du bassin. Ces paramètres ont constitué les entrées du modèle HEC-HMS pour
la transformation de la pluie en débit. Les pointes des débits générés par HEC-HMS ont été
entrés dans HEC-RAS pour des simulations en régime permanent pour voir les éventuels
débordements du canal de la rue ministre afin de délimiter et cartographier plus loin la zone
inondable en utilisant RAS Mapper et HEC-GeoRAS dans ArcGIS. Après le calage indirect du
model et calculs des débits de certaines périodes de retour, le logiciel HEC-RAS a été exécuté
avec les différentes périodes de retour 5, 10, 25, 50 et 100 ans et l’étendue de l’inondation de
chaque période a été déterminée. Ainsi, la carte d’inondation a été créée pour une meilleure
visualisation et des prises de décision et de ce fait des stratégies de préparation contre les
inondations.

Mots clés : ArcGIS, Bassin de Bonoumin, Cocody, Modélisation d’inondation, HEC-

HMS, HEC-RAS

v
ACKNOWLEDGEMENT

I hereby acknowledge the effort of people who helped me to achieve this study and reach this
level of master thesis report.

My thanks go, first at all to my advisor Professor Yasser Eldeberky for accepting to supervise
my work with helpful advice and constant support since the proposal. My thanks go also to Dr.
Abdessamed Derdour who was very helpful at the beginning of the current work.

Moreover, I would like to thank the African Union for the scholarship they awarded me and the
research grant and also PAUWES administration, the Director, Prof. Abdellatif Zerga, his
Deputy Director, Prof. Chewki Ziani and their team.

I can’t ignore the internship host company: SODEXAM. My thanks go also to Dr. Kouakou
Bernard Djê the head of the department of Climatology and Applied Meteorology and his team
who not only provided the required meteorological data but also helped in the choice of the
study area. I really appreciated the criticism and advice from Mrs. Kangah, Hydrology Engineer
and Soumahoro, Meteorology Engineer.

I express my deepest gratitude to the Head of Gourou Watershed Project, Professor Gougou,
who allowed me to have one day practice in measuring flowrate of my canal with a velocity
meter and deduct the flow. I would like to thank also his team member, Mr. Angoua Marc
Kouadio, Civil Engineer at the Gourou Project for his useful advice.

I also want to thank my team members from the Modelling and Applied Meteorology Group at
SODEXAM for the great collaboration. Adou Bérenger Kouakou for sharing his knowledge of
remote sensing and ArcGIS. Pesson, a PhD student for his analysis and support, and all the
other members of the documentation office.

I would like to thank my senior Hydraulics Engineer at TERRABO Consulting Engineer, Mr.
Abdoulaye Traoré who did a lot within a short period of time.

Lastly and very important, I would like to thank my family, Christelle Appali my best friend
for her encouragement all times. My family and friends for giving me strength. Gnamien
Constant Kouakou, Freddy Houndekindo and Gisele ICYIMPAYE for their collaboration,
support and guidance; all my friends and brothers who helped me in this work but whose names
are not mentioned here.

vi
Contents

DECLARATION ...................................................................................................................... II

CERTIFICATION .................................................................................................................. III

ABSTRACT ............................................................................................................................. IV

RESUME ................................................................................................................................... V

ACKNOWLEDGEMENT ...................................................................................................... VI

LIST OF ABREVIATIONS ................................................................................................... XI

LIST OF FIGURES ............................................................................................................... XII

LIST OF TABLES ................................................................................................................ XIII

CHAPTER 1: PURPOSE AND SIGNIFICANCE OF THE STUDY ......................................... 1

1.1 Background .......................................................................................................................... 1

1.2 Problem statement and Justification ..................................................................................... 2

1.3 Research objectives: ............................................................................................................. 3

1.4 Research questions ............................................................................................................... 4

1.5 Hypothesis ............................................................................................................................ 4

1.6 Impacts of urbanization on floods in Abidjan city ............................................................... 4

CHAPTER 2: LITERATURE REVIEW ....................................................................................... 8

2.1 Urban Flooding .................................................................................................................... 8

2.2 Urban Flood Models............................................................................................................. 9

vii
2.3 Runoff Curve Number Method ............................................................................................ 9

2.4 GIS techniques in Hydrologic and Hydraulic Modelling................................................... 11

2.5 Model selection .................................................................................................................. 13

2.6 HEC-RAS ........................................................................................................................... 13

2.7 HEC-HMS .......................................................................................................................... 16

2.8 Data requirements .............................................................................................................. 16

2.8.1 HEC-RAS .................................................................................................................... 16
2.8.2 HEC-HMS ................................................................................................................... 17

2.9 Stormwater management .................................................................................................... 18

2.9.1 Historical evolution of Abidjan drainage network ...................................................... 18
2.9.2 Characteristics of catchments in Abidjan .................................................................... 19
2.9.3 Hydrometeorological data and observations ............................................................... 19
2.9.4 Recent and under way development ............................................................................ 20
2.9.5 Hydrological methods and approaches ........................................................................ 20
2.10 Previous studies on flood events in Abidjan ................................................................ 21

CHAPTER 3: METHODOLOGY ............................................................................................... 22

3.1 Study area ........................................................................................................................... 22

3.1.1 Geographic Location and Boundaries ......................................................................... 23
3.1.2 Socio-Economic Background ...................................................................................... 24
3.1.3 Urbanization Process ................................................................................................... 27
3.1.4 Geomorphology ........................................................................................................... 27
3.1.5 Hydro-Climatology...................................................................................................... 28
3.1.6 Geology ....................................................................................................................... 32
3.1.7 Soil ............................................................................................................................... 33
3.1.8 Hydrogeology .............................................................................................................. 34
3.1.9 Hydrology and Drainage ............................................................................................. 34
3.1.10 Characteristics of Bonoumin-Rivièra Palmeraie watershed ...................................... 35

viii
 3.1.11 Conclusion ................................................................................................................. 43

3.2 Data Collection ................................................................................................................... 43

3.2.1 Topographic Data ........................................................................................................ 43
3.2.2 Soil Data ...................................................................................................................... 44
3.2.3 Vegetation Data ........................................................................................................... 44
3.2.4 Hydro-meteorological Data ......................................................................................... 45

3.3 Models and Softwares description ..................................................................................... 45

3.3.1 Overall Methodology................................................................................................... 45
3.3.2 Rainfall-Runoff Model: HEC-HMS ............................................................................ 48
3.3.3 Hydraulic Model: HEC-RAS....................................................................................... 55

3.4 Models development .......................................................................................................... 65

3.4.1 HEC-HMS ................................................................................................................... 65
3.4.2 HEC-RAS .................................................................................................................... 75

CHAPTER 4: FINDINGS AND DISCUSSION ........................................................................ 81

4.1 Discussion of Results ......................................................................................................... 81

4.1.1 Hydrologic Modelling Results..................................................................................... 81
4.1.2 Hydraulic Modelling Results ....................................................................................... 87
4.1.3 Inundation Mapping .................................................................................................... 92

4.2 Comparison of this study’s results with other did on Cocody ........................................... 93

4.3 Flood Preparedness Strategy .............................................................................................. 95

CHAPTER 5: GENERAL CONCLUSION AND RECOMMENDATION ............................. 96

5.1 General Conclusion ............................................................................................................ 96

5.2 Recommendation ................................................................................................................ 97

REFERENCES:............................................................................................................................. 98

ix
(I) APPENDIX A: RESULTS OF PRE-PROCESSING IN ARCGIS .............................. 105

(II) APPENDIX B: HYDROLOGIC MODELLING RESULTS ....................................... 106

(III) APPENDIX B: HYDRAULIC MODELLING RESULTS .......................................... 108

x
LIST OF ABREVIATIONS

BNETD Bureau National d’Etudes Techniques et de Développement

BRP Bonoumin – Rivièra Palmeraie

BCET Bureau Central d’Etudes Techniques

CURAT Centre Universitaire de Recherche Appliquée à la Télédétection

CN Curve Number

DEM Digital Elevation Model

GIS Geographic Information System

HEC-HMS Hydrologic Engineering Center’s - Hydrologic Modelling System

HEC-RAS Hydrologic Engineering Center’s - River Analysis System

IDF Intensity-Duration-Frequency

IPCC Intergovernmental Panel on Climate Change

NRCS Natural Resource Conservation Service

ONAD Office National de l’Assainissement et du Drainage

PUIUR Progamme d’Urgence d’Infrastructures Urbaines

SCS Soil Conservation Service

xi
LIST OF FIGURES

Figure 3.1: Location of Bonoumin and Palmeraie Watershed in Abidjan/Côte D'Ivoire ........ 23
Figure 3.2: Location of Bonoumin and Rivièra Palmeraie Watershed on Google Earth ......... 24
Figure 3.3: Neighbourhoods of Cocody municipality.............................................................. 25
Figure 3.4: DEM of Abidjan .................................................................................................... 28
Figure 3.5: Average monthly rainfall at Abidjan airport from 1961 to 2017 ........................... 30
Figure 3.6: Annual rainfall trend from 1961 to 2017 ............................................................... 30
Figure 3.7: Average monthly temperature from 1961 to 2017 at Abidjan airport ................... 31
Figure 3.8: Ombrotherlic diagram of Abidjan from 1961 to 2017........................................... 31
Figure 9: Monthly evaporation at Abidjan airport from 2000 to 2009 .................................... 32
Figure 3.10: Geological map of Abidjan (Danumah, 2016)..................................................... 33
Figure 3.11: Digital Elevation Model of BRP watershed ........................................................ 37
Figure 3.12: Hypsometric map of BRP watershed ................................................................... 39
Figure 3.13: Hypsometric curve of BRP watershed ................................................................. 40
Figure 3.14: Slope map of BRP watershed .............................................................................. 42
Figure 3.15: The student into the field for collecting elevations and geometric data of the
canal on 11 May 2018 .............................................................................................................. 44
Figure 3.16: Chart showing the process involved in the methodology .................................... 47
Figure 3.17: Runoff process in HEC-HMS (HEC, 2000) ........................................................ 48
Figure 3.18: HEC-HMS components (Heimhuber, 2013) ....................................................... 49
Figure 3.19: Representation of Terms in the Energy Equation (HEC, 2016) .......................... 56
Figure 3.20: HEC-RAS Default Conveyance Subdivision Method (HEC,2016) .................... 58
Figure 3.21: Alternative Conveyance Subdivision Method (HEC, 2016) ............................... 59
Figure 3.22: Application of the Momentum Principle ............................................................. 61
Figure 237: HEC-GeoHMS tools bar within ArcGIS and terrain pre-processing menu ......... 71

xii
LIST OF TABLES

Table 1.1 : Comparison of Abidjan Ubanization rate with other cities ...................................... 5
Table 1.2 : Comparison of Abidjan Urbanization level with other cities and categories........... 6
Table 3.3: Elevation data of BRP for hypsometric curve ........................................................ 38
Table 3.4: Characteristics of the BRP watershed ..................................................................... 40
Table 3.5 : Hydrologic model selection and categorization ..................................................... 52
Table 3.6: Soil texture (Kouamé et al., 2011) .......................................................................... 67
Table 4.7: Peak discharges derived from the figure 4.35 ......................................................... 82
Table 4.8: Peak discharges for various return periods at J37 ................................................... 82
Table 4.9: Sub-basins characteristics ....................................................................................... 84
Table 4.10: Sub-basins characteristics with increased imperviousness ................................... 84
Table 4.11:Comparison of 5 years flows by 10% decrease in imperviousness ....................... 85
Table 4.12: Characteristics of Angré 3 watershed (CONCEPT, 2015) ................................... 93
Table 4.13: Peak discharges computed at the Angré watershed (CONCEPT, 2015) .............. 94
Table 4.14: Peak discharges for various return periods at the BRP watershed ........................ 94

xiii
Chapter 1: PURPOSE AND SIGNIFICANCE OF THE STUDY

1.1 Background

Flooding occurs when water accumulates in places that are not normally submerged. In urban
areas, they are usually the consequence of extreme rainfall, which creates an excess of runoff
that is above the capacity of the drainage systems (Adetunji & Oyeleye, 2013) or it occurs when
an extreme event coincides with a vulnerable physical and socio-economic environment,
surpassing society’s ability to control or survive the consequences (Baldassarre et al., 2013).
With more than half of the global population living in urban areas, these phenomena are
becoming an increasing public concern (Nasiri & Shahmohammadi-kalalagh, 2013). Their
occurrences and amplitudes will increase due to climate change and the increasing urban
population (IPCC, 2007).

Urban flooding is a serious and growing development challenge. Against the backdrop of
demographic growth, urbanization trends and climate changes, the causes of floods are shifting
and their impacts are accelerating. Urban floods are one of the most common and widely
distributed natural risks for life and property worldwide. It has been reported that in the last
decade, urban floods have impacted most parts of the world including the USA, Europe, Asia,
and Africa. Flooding occurs in the least developed nations as well as in the most developed
countries. In the developed world, urban floods are often related to hazards, such as climate
change, storm surge, flash floods, and consecutive heavy precipitations. However, in addition
to the highlighted causes that prevail in the developed world, flooding in developing countries
is also due to the precariousness of the drainage system, the lack of maintenance of the
infrastructures and the mismanagement of household wastes.

In Sub-Saharan Africa where about 72% of urban inhabitants live in slums, the consequences
of urban flooding could be worsened by the lack of adequate protection infrastructures.
Unplanned growth and development in such areas usually results in flooding when the flood
protection structures fail under extreme hydrological weather conditions. Studies have reported
that West Africa is particularly subject to urban floods. The region was struck in 2007 by heavy
precipitation causing widespread regional flooding that affected a total of 792,676 people and
caused 210 deaths (Danumah, 2016).

1
Abidjan, the economic capital of Côte d’Ivoire, offers a prime example of this phenomenon of
urban flooding. In that city, every rainy season is characterized by its number of deaths and the
extent of its damages. According to the Office for the Coordination of Humanitarian Affairs
(OCHA, 2014), from 2009 to 2014, an average of 13 flood-related deaths per year were
recorded in Abidjan (Ketcha, Kablan, Dongo, & Coulibaly, 2017). In these conditions, flood
vulnerability assessment and forecasting are important not only because flooding inflicts harm
to humans, but also for proper urban planning and adaptation to climate change. A good
knowledge of flood-prone areas, the level of vulnerability of an urban area to flooding and the
socio-economic, environmental and physical factors that play a major role in shaping the
hazard, could be an important step toward improving the resilience of growing African cities
like Abidjan, Côte d’Ivoire. Although flood hazards are natural phenomena, the vulnerability
of an area to flooding is a combination of socio-economic and environmental factors that vary
spatially from one place to another.

While the primary cause of flooding is normally heavy rainfall, it is also due to human activities
such as land degradation, deforestation of catchment areas, sprawl and increased population
density along river banks, poor land use planning, zoning and control of floodplains
development, inadequate drainage system particularly in big cities and inadequate management
of discharges from rivers’ reservoirs (Danumah, 2016). Assessing extreme rainfall, predicting
flood risks, mapping flood prone areas and forecast floods are more and more essential to
mitigate floods and enable a good living environment to people.

1.2 Problem statement and Justification

Since 2009, Flood has been occurring in Côte D’Ivoire each rainy season in general and in
Abidjan in particular. The Office for the Coordination of Humanitarian Affairs (OCHA, 2014)
mentioned that, from 2009 to 2014, an average of 13 flood-related deaths per year were
recorded in Abidjan. The same report has shown that 26% of the Abidjan District is at risk of
flooding. Indeed, 80,000 people are threatened by flooding in Abidjan with 40,000 at risk in
Cocody, 12,500 in Abobo, 10,000 in Adjamé, 9,500 in Yopougon and 8,000 in Attécoubé
municipalities. These statistics indicate the extent of damage resulting from flooding in the
country and especially in Abidjan. To these lives’ losses one should add destruction of
infrastructures such as drainage systems, roads and buildings, water bodies’ pollution by the
conveyance of chemicals.

2
Abidjan is situated in the south of the country under the tropics and is characterized by
important recurrent flooding events. Floods risk in the city is higher in slums due to violent
storms (Dongo, Cissé, & Biémi, 2008). Erosion phenomena resulted in landslides and
especially flooding occurs in slums, generally settled in high risk areas and protected areas from
the urbanization master plan of Abidjan (Dongo et al., 2008).

Information regarding the flooding characteristics and its effect are essential for management
of water bodies for decision making in flood management strategies such as construction of
flood protection structures, development of flood emergency plan and human settlement
planning.

Thus, floods management in slums in particular and in the whole city has become a major
concern for researchers interested in urban issues (Dongo et al., 2008).

However, recent scientific work undertaken in the District of Abidjan concentrated on certain
factors controlling flood risk such as rainfall risks and uncontrolled urban growth in two
municipalities of Abidjan: Attécoubé and Abobo (Hauhouot, 2008). This is piecemeal approach
and will not provide solution to the problem of flooding in the District. Other studies (Danumah,
2016; Dongo et al., 2008) focused on flood analysis in some slums in Yopougon and flood risks
assessment and mapping in Abidjan. These studies don’t give any idea of the depth and
propagation of floods. Ivorian decision makers are still lacking knowledge for appropriate
decisions to save lives, infrastructures and economic activities.

With advances in hydrodynamic modelling these days it is possible to model flood extent,
depth, duration and even flood propagation in the temporal and spatial dimension. The
limitation of the required topographic data for representing the topography of river and
floodplain in the modelling is a major problem.

Thus, this research is necessary to fill the gap by developing models for floods modelling and
mapping useful in managing floods.

1.3 Research objectives:

From the background information and problem statement the following general and specific
research objectives are formulated below.

General objective

To estimate floods in Bonoumin – Rivièra Palmeraie watershed and identify flood prone areas.

3
Specific objectives

✓ Foster a better understanding of floods in that part of the city

✓ To study the flooding characteristics in the catchment
✓ To produce maps of floods depth, level and extent for various return periods
✓ To come up with floods preparedness strategies

1.4 Research questions

The relevant research questions that will be addressed in this research are:

✓ How accurate can the observed inundated flood area be simulated by the selected model
approach?
✓ What are the causes of frequent occurrence of flood in BRP watershed?
✓ How does the inundation pattern vary along the canal?
✓ How to get prepared against flood damages?

1.5 Hypothesis

To address the research questions above, the following hypothesises need to be tested:

✓ The Government has to prioritize the maintenance and cleaning of open channels to
avoid their obstruction
✓ The canal of “rue minister” needs to be widened for it to be able to convey much water
✓ Uncontrolled urbanization contributes a lot in the frequent occurrence of flood

1.6 Impacts of urbanization on floods in Abidjan city

Urbanization and growth go hand in hand, and no one can deny that urbanization is essential
for socioeconomic transformation, wealth generation, prosperity and development.

Urbanization levels reflect the degree of economic development of a region, but it also changes
water cycle in many aspects. It modifies hydrological process by affecting runoff generation
and concentration which has been research focus currently(Chen et al., 2015; Mukherjee, 2016).
This phenomenon is partly caused by the massive departure of people from rural areas to the
big cities seeking for a better life. This rapid population growth in large cities inevitably leads
to the rapid urbanization of urban neighbourhoods and, as a consequence, an increase in
impervious surfaces. Indeed, building roofs, parking lots, streets and sidewalks, all
infrastructures in concrete, roads limit the infiltration of water into the soil. This results in an
4
 increase increased runoff as a result of decreased natural infiltration of stormwater which
contributes to devastating urban floods(Moglen & Beighley, 2002).

Since 1995, Abidjan has been growing very fast and this can’t take place without impacting the
natural water paths of the watershed. It is not only the biggest city of the country but also the
city where the government is and where the economic activities of the country are done. Table
… gives some facts to better understand or better perceive what’s happening in terms of
urbanization in this city of West Africa.

Table 1.1 : Comparison of Abidjan Ubanization rate with other cities (UN HABITAT, 2016)
Population of urban Share in national urban
agglomeration ('000) Annual rate of change (%) population (%)
1995- 2005- 1995-
Country City 1995 2005 2015 2025 2005 2015 2015 1995 2005 2015 2025

Algeria El Djazaïr (Algiers) 1,973 2,282 2,594 3,149 1.45 1.28 1.37 12.0 10.5 9.0 9.0

Algeria Wahran (Oran) 705 783 858 1,035 1.05 0.91 0.98 4.3 3.6 3.0 2.9

Angola Luanda 1,899 3,533 5,506 8,567 6.21 4.44 5.32 54.3 59.0 54.8 54.9

Angola Huambo 444 751 1,269 2,078 5.25 5.25 5.25 12.7 12.5 12.6 13.3

Argentina Buenos Aires 11,390 13,330 15,180 16,479 1.57 1.30 1.44 37.1 38.3 39.2 39.0
San Miguel de
Argentina Tucumán 666 781 910 1,024 1.60 1.53 1.56 2.2 2.2 2.4 2.4

Argentina La Plata 656 723 846 955 0.96 1.58 1.27 2.1 2.1 2.2 2.3
Burkina
Faso Ouagadougou 667 1,328 2,741 4,732 6.89 7.25 7.07 43.7 45.9 51.2 53.6

China Beijing 8,305 12,813 20,384 26,494 4.34 4.64 4.49 2.2 2.3 2.6 2.8
Côte
d'Ivoire Abidjan 2,535 3,545 4,860 6,729 3.35 3.15 3.25 43.3 43.5 42.1 42.1
Côte
d'Ivoire Bouake 427 572 762 1,048 2.91 2.87 2.89 7.3 7.0 6.6 6.6

France Paris 9,510 10,092 10,843 11,565 0.59 0.72 0.66 21.9 21.3 21.0 20.8

France Toulouse 714 825 938 1,048 1.45 1.28 1.37 1.6 1.7 1.8 1.9

Germany Berlin 3,471 3,391 3,563 3,654 -0.23 0.49 0.13 5.7 5.5 5.7 5.8

Germany Köln (Cologne) 965 976 1,037 1,080 0.12 0.60 0.36 1.6 1.6 1.7 1.7

Ghana Kumasi 909 1,544 2,599 3,707 5.30 5.20 5.25 13.5 15.3 17.8 19.0

Ghana Accra 1,415 1,854 2,277 2,870 2.70 2.06 2.38 21.0 18.3 15.6 14.7

Niger Niamey 552 816 1,090 1,744 3.91 2.89 3.40 38.2 37.0 30.2 27.5

Nigeria Lagos 5,983 8,859 13,123 20,030 3.93 3.93 3.93 17.1 16.2 15.0 15.1

Switzerland Zürich (Zurich) 1,048 1,130 1,246 1,406 0.75 0.98 0.86 20.3 20.8 20.5 20.6

5
 It’s true that cities like Luanda in Angola have an urbanization rate of 5% from 1995 to 2015
and Abidjan has 3%; nevertheless, urbanization should go together with a good sanitation or
drainage master plan for a good management of stormwater.

Table 1.2 : Comparison of Abidjan Urbanization level with other cities and categories (UN
HABITAT, 2016)

Urban population ('000) Level of urbanization (%)

1995 2005 2015 2025 1995 2005 2015 2025
WORLD 2,568,063 3,199,013 3,957,285 4,705,774 44.7 49.1 54.0 58.2
More developed regions 860,171 920,702 985,831 1,034,150 73.3 75.8 78.3 80.4
Less developed regions 1,707,892 2,278,311 2,971,454 3,671,623 37.4 43.0 49.0 54.0
Least developed countries 133,757 198,147 295,178 427,084 22.9 26.5 31.4 36.6
Less developed regions, excluding least developed countries 1,574,134 2,080,164 2,676,276 3,244,540 39.5 45.7 52.2 57.6
Less developed regions, excluding China 1,303,727 1,693,998 2,166,067 2,696,694 39.5 42.9 46.8 50.7
High-income countries 873,730 954,869 1,042,669 1,106,576 75.5 78.0 80.4 82.3
Middle-income countries 1,544,557 2,033,716 2,615,346 3,180,233 38.7 44.8 51.3 56.8
Upper-middle-income countries 928,664 1,229,547 1,574,772 1,857,018 44.6 53.9 63.5 70.7
Lower-middle-income countries 615,893 804,168 1,040,574 1,323,215 32.3 35.7 39.8 44.5
Low-income countries 133,543 191,782 278,657 397,055 23.2 26.4 30.8 35.7
Sub-Saharan Africa 163,172 240,036 359,534 522,530 29.1 33.0 37.9 42.9
AFRICA 236,904 330,742 471,602 658,814 33.1 36.3 40.4 44.9
Ethiopia 7,885 11,958 19,266 30,190 13.8 15.7 19.5 24.2
Kenya 5,007 7,757 11,978 17,973 18.3 21.7 25.6 30.3
Gabon 814 1,151 1,526 1,916 75.4 83.4 87.2 88.5
Algeria 16,416 21,677 28,739 35,145 56.0 63.8 70.7 75.6
Egypt 26,188 30,884 36,538 43,610 42.8 43.0 43.1 45.0
Botswana 776 1,033 1,181 1,357 49.0 55.1 57.4 60.5
South Africa 22,572 28,717 34,663 39,313 54.5 59.5 64.8 69.4
Côte d'Ivoire 5,859 8,147 11,538 15,968 41.2 46.8 54.2 60.5
Ghana 6,728 10,116 14,583 19,506 40.1 47.3 54.0 60.0
Niger 1,446 2,204 3,609 6,332 15.8 16.7 18.7 22.2
Nigeria 34,919 54,541 87,681 132,547 32.2 39.1 47.8 55.3
Senegal 3,452 4,634 6,544 9,283 39.6 41.1 43.7 47.8
China 383,156 560,518 779,479 947,540 31.0 42.5 55.6 65.4
Japan 97,117 109,174 118,572 118,715 78.0 86.0 93.5 96.3
Sweden 7,399 7,614 8,319 9,056 83.8 84.3 85.8 87.3
France 43,456 47,393 51,674 55,548 74.9 77.1 79.5 81.7
Germany 60,936 61,498 62,170 62,654 73.3 73.4 75.3 77.5

From this table, one can see that from 2005 to 2015, Côte D’Ivoire got respectively 46.8% and
54.2% level of urbanization which is an increase of 7.4% in ten years. While, the middle-income
countries achieved an increase of 6.5% over the same period of time. And Côte D’Ivoire Is a

6
middle-income country; so, its urbanization rate is above the average of this category of
countries. Egypt’s urbanization is almost constant between 2005 and 2015. In addition, Cocody
is growing very fast because it has more non-built-up area (Kablan, 2014).

What makes this comparative study interesting is that frequent flooding in Abidjan, the most
urbanized city, has been occurring since 2007. This is also supported by the intense rainfall at
Abidjan airport on June 20, 1983 with a water depth of 311 mm (SODEXAM) but did not flood
any place in the city, while we got 302 mm on June 19, 2018 which resulted in a very severe
flooding.

Therefore, urbanization is clearly intensifying flooding in the city of Abidjan.

7
Chapter 2: LITERATURE REVIEW

2.1 Urban Flooding

As a result of uncontrolled human activities, urbanization and unpredictable tropical weather

conditions, most of tropical developing countries are facing a risk of urban flood. This is
becoming catastrophic during tropical cyclone and flash flood situations. Such events cause
loss of lives, damage properties, road and drainage systems, and whole environment.

There are many studies which emphasize that, with the threat of climate change; such natural
disasters are likely to amplify this trend in years to come. In many situations it is the population
living in low lying urban areas that are most prone to these disasters. From this it follows that
the scientific and professional communities have responsibility to constantly evolve with better
stormwater management approaches to minimize hazard risks of urban flooding while
addressing different climate conditions.

The water management practices can vary based on climate conditions, geographic location,
availability of resources and culture. Especially for tropical islands the sudden and heavy
rainfall can unexpectedly occur, since typically they have mountainous geographical character,
low lands and basins are suddenly flooded with high flows. Therefore, in order to minimize the
frequent urban flash flood damages in tropical Island, the structural measures, effective non-
structural measures such as better hazard mitigation and prevention, improved preparedness
and warning systems, well organized pre-emptive action and emergency response have to be
effectively integrated. Therefore, the management of urban flooding is a multi- disciplinary
process. The urban planners, economists; lawyers, emergency services and other professionals
should be involved along with the engineers in this multi- disciplinary process to develop
strategic plans for hazard reduction.

National Research Council (NRC), 2008; Gautam, Acharya and Stone, 2010, have presented
the short, medium and long-term objectives of storm water management strategies. In the short-
term, the priorities are runoff control flood protection and pollution mitigation strategies, which
in many developing countries have yet to be addressed effectively. The medium-term objectives
focus on the development and implementation of water quality improvement, water
conservation and strategies to preserve the hydrology and natural catchments. The long-term
objectives place greater emphasis on preservation of natural resources and the amenity value of
water in the urban environment for recreational activities and to promote an increased

8
awareness of environmental issues. Although these objectives may initially appear to be
somewhat idealistic goals especially considering the existing situation in many developing
countries, it is important that planners and designers of urban drainage systems aim to satisfy
the need of future generations while keeping with the objectives of sustainable development as
defined by World Commission on Environment and Development in 1987.

2.2 Urban Flood Models

Flooding is a natural and variable phenomenon, it can occur on any land surface either in rural
or urban. Flooding results in damage to lives, property, crops and negative impacts on human
welfare. Flood Plain Management aims to minimize damages and reduce the threat to human
life and welfare when major flood events occur. Numerical simulation provides good
information of physical process. Expansion process and the distribution of the water depth by
numerical simulation results are helpful for discussion and consideration comparing with field
survey(Singh & Frevert, 2002).

From the review of application concerning modelling of urban flooding it can be concluded that
it is important to have a hydrodynamic model based on full dynamic equations in order to
describe the flood sufficiently. In this section, a model of such specification has been reviewed.
The one-dimensional numerical models are based on the cross-sectional averaged Saint-Venant
equations, describing the development of the water depth h and the discharge Q or the mean
flow speed U. There are 4-point and 6- point numerical schemes. Examples of such model can
be found in commercial software such as MIKE-11, MOUSE, SOBEK, Infoworks, SWMM,
MIKE-21, etc. Briefly, the numerical solution is obtained from a finite difference formulation
of the equations, using a scheme, which is based on alternating Q and h points (Michas &
Lazaridis, 2005; Singh & Frevert, 2002; Yener, Sorman, & Gezgin, 2007).

2.3 Runoff Curve Number Method

The SCS Runoff Curve Number method is used in this study to compute the runoff. It was
developed by the United States Department of Agriculture (USDA) Soil Conservation Service
(SCS) and is a method of estimating rainfall excess from rainfall (Hjelmfelt, 1991; Kamal,
2015). The method is described in detail in National Engineering Handbook (USDA-NRCS,
2010). The chapter was prepared originally by Mockus (1964) and was revised by Hjelmfelt

9
with assistance from the NRCS Curve Number work group and H.F. Moody. Despite the wide
use of the curve number procedure, documentation of its origin and derivation are incomplete
(Hjelmfelt, 1991). The conceptual basis of the curve number method has been the object of both
support and criticism (Ponce & Hawkins, 1996). The major disadvantages of the method are
sensitivity of the method to Curve Number (CN) values, fixing the initial abstraction ratio, and
lack of clear guidance on how to vary Antecedent Moisture Conditions (AMC)(USDA-NRCS,
1986). However, the method is used widely and is accepted in numerous hydrologic studies.
The SCS method originally was developed for agricultural watersheds in the mid-western
United States; however, it has been used throughout the world far beyond its original developers
would have imagined. The SCS Curve Number method is one of the most popular methods for
computing the volume of surface runoff for a given rainfall event from small watersheds. This
method’s usefulness is mainly dependent on its convenience, authoritative origins, simplicity
and responsiveness to four distinct catchment properties: soil type, land use/treatment, surface
condition and antecedent condition (Kamal, 2015). The CN method was used to perform
hydrologic modelling of the Cedar Creek Watershed for the construction of a rainfall-runoff
model and gave satisfactory results (Kamal, 2015).

The basis of the curve number method is the empirical relationship between the retention
(rainfall not converted into runoff) and runoff properties of the watershed and the rainfall.
Mockus found equation 2.1 appropriate to describe the curves of the field measured runoff and
rainfall values (National Engineering Handbook, 2004). Equation 2.1 describes the conditions
in which no initial abstraction occurs.
𝐹 𝑄
=𝑃 Eq. 2.1
𝑆

where F = P – Q = actual retention after runoff begins;

Q = actual runoff
S = potential maximum retention after runoff begins (S ≥ F)
P = potential maximum runoff (i.e., total rainfall if no initial abstraction).
For most applications, a certain amount of rainfall is abstracted. The three important
abstractions for any single storm event are rainfall interception (Meteorological rainfall minus
throughfall, stem flow and water drip), depression storage (topographic undulations), and
infiltration into the soil. The curve number method lumps all three abstractions into one term,
the Initial abstraction (𝐼𝑎), and subtracts this calculated value from the rainfall total volume.
The total rainfall must exceed this initial abstraction before any runoff is generated. This gives

10
the potential maximum runoff (rainfall available for runoff) as P – 𝐼𝑎. Substituting this value
in equation 2.1 yields following equation
𝑃−𝐼𝑎−𝑄 𝑄
= 𝑃−𝐼𝑎 Eq. 2.2
𝑆

Rearranging terms in Equation 2.2 for Q gives

(𝑃−𝐼𝑎)2
Q = (𝑃−𝐼𝑎)+𝑆 Eq. 2.3

The SCS provided the following empirical Equation 4 based on the assumption 𝐼𝑎 was a
function of the potential maximum retention S.
𝐼𝑎 = 0.2𝑆 Eq. 2.4
The potential maximum retention S is related to the dimensionless parameter CN in the
range of 0 ≤ CN ≤ 100 by Equation 2.5.
1000
S= − 10 Eq. 2.5
𝐶𝑁

Substituting Equation 2.4 into Equation 2.3 yields,

(𝑃−0.2𝑆)2
Q= Eq. 2.6
𝑃+0.8𝑆

Equation 2.6 has only one parameter that needs to be evaluated (i.e., S) which can be
determined by using Equation 2.5 and curve number tables published by the SCS.

2.4 GIS techniques in Hydrologic and Hydraulic Modelling

The increasing availability of spatial data (terrain and rainfall), GIS software to manage spatial
data, faster processors, and the availability of interfaces to connect simulation models with GIS,
have increased use of GIS in watershed modelling (Ne & Sw, 2004; Singh & Frevert, 2002).
HEC-GeoRAS is the geospatial tool used in this study, which serves as the interface between
GIS and the simulation model HEC-RAS. HEC-GeoRAS allows engineers to concentrate on
hydraulic model development and analysis rather than GIS mechanics (C. T. Ackerman, 2009;
C. T. Ackerman, Jensen, & Brunner, 2010). The user environment provides engineers an
opportunity to view real-world systems of interest, which in turn assists them to rectify errors
and make informed decisions in the model development (A. S. Ackerman et al., 2000). Tate
and Maidment, (1999) applied HEC-GeoRAS successfully to create a terrain model for
floodplain mapping. A widely used approach is watershed modelling that divides the drainage
basin into discrete units possessing similar rainfall-runoff and physical characteristics. This
approach reduces model complexity and spatially distributed data requirements in basin-scale
models (Beighley, Dunne, & Melack, 2005).

11
As a result of rapid urbanization and climate changes urban flooding has become an increasing
and continuous threat all over the world. Therefore, better analytical understanding and
visualization of this disaster is essential to develop strategies that will minimize the risk of
urban flooding. At present, 1D River models, digital elevation models and other GIS data sets
for hydraulic modelling and floodplain mapping are often collectively used to predict areas at
at risk of flooding. Hydraulic and hydrological modelling are an obvious choice for predicting
those areas of the floodplain most at risk to flooding and for providing information for use in
the evaluation of the associated economic damage.

Geographic Information System (GIS) has evolved over the last couple of decades into a
powerful tool for storing, managing, analysing and displaying spatial data (Burrough, and
McDonnell, 1998). Generally, the integration of hydraulic models and GIS for floodplain
mapping aims to provide

✓ Functions to extract information describing the channel system from a terrain model to
provide a network description (e.g. topographic data of channel network and adjacent
area)
✓ Tools that are capable of manipulation results from hydraulic models and displaying
and automating mapping of floodplain in GIS (e.g. water surface profiles) (Jones, Mike, &
Young, 1997).

Approaches for integrating hydraulic model to GIS have resulted in many different tools for
flood prediction and floodplain mapping, e.g. HEC-GeoHMS, MIKE 11 GIS (Muller &
Rungoe, 1995), HEC-GeoRAS (C. T. Ackerman, 2009).

Hydraulic models are complex tools, requiring large amounts of input data for their
specification to a particular application and produce a vast amount of output data. The data
requirements for distributed hydraulic models are grouped into topographic and hydrologic data
(Cunge, Holly, & Verwey, 1980).

Topographic data: describing the channel geometry of the river system and adjacent areas
(channel widths, cross-sectional areas) and elevations of the flood plain.

12
Hydrologic data: model boundary conditions (e.g. inflow hydrographs) and discharge and water
level data for the calibration of model parameters (e.g. bed roughness and weir coefficients).

2.5 Model selection

For selecting potentially an appropriate modelling tool in any research, there are various criteria
which can be applied to choose the most suitable model. According to Cunderlik and
Simonovic, (2004), the choice mainly depends on the requirements and needs of the research
or project under interest. Cunderlik and Simonovic, (2004) put the following as criteria:

✓ Required outputs of the model

✓ Availability of input data
✓ Prices and availability of the model and
✓ The model structures

2.6 HEC-RAS

The Hydrologic Engineering Center-River Analysis System (HEC-RAS) is an integrated

system of software, designed for interactive use in a multi-tasking environment. The system is
comprised of a graphical user interface (GUI), separate analysis components, data storage and
management capabilities, graphics and reporting facilities (Hydrologic Engineering Center
(HEC), 2016b)(Hydrologic Engineering Center (HEC), 2016b). The HEC-RAS system
contains the following river analysis components for: (1) steady flow water surface profile
computations; (2) one dimensional and/or two-dimensional unsteady flow simulation; (3) Quasi
unsteady or fully unsteady flow movable boundary sediment transport computations; and (4)
water quality analysis. A key element is that all four components use a common geometric data
representation and common geometric and hydraulic computation routines. In addition to the
four river analysis components, the system contains several hydraulic design features that can
be invoked once the water surface profiles are computed.

User Interface
The user interacts with HEC-RAS through a graphical user interface (GUI). The main focus in
the design of the interface was to make it easy to use the software, while still maintaining a high
level of efficiency for the user. The interface provides for the following functions (HEC, 2016):

13
 ✓ File management
✓ Data entry/editing and GIS data interfaces
✓ River analyses
✓ Tabulation and graphical displays of input and output data
✓ Inundation mapping and animations of water propagation
✓ Reporting facilities
✓ On-line help

River Analysis Components

Steady Flow Water Surface Profiles. This component of the modelling system is intended for
calculating water surface profiles for steady gradually varied flow. The system can handle a full
network of channels, a dendritic system, or a single river reach. The steady flow component is
capable of modelling subcritical, supercritical, and mixed flow regime water surface profiles.
The basic computational procedure is based on the solution of the one-dimensional energy
equation. Energy losses are evaluated by friction (Manning's equation) and
contraction/expansion (coefficient multiplied by the change in velocity head). The momentum
equation is utilized in situations where the water surface profile is rapidly varied. These
situations include mixed flow regime calculations (i.e., hydraulic jumps), hydraulics of bridges,
and evaluating profiles at river confluences (stream junctions).

The effects of various obstructions such as bridges, culverts, dams, weirs, and other structures
in the flood plain may be considered in the computations. The steady flow system is designed
for application in flood plain management and flood insurance studies to evaluate floodway
encroachments. Also, capabilities are available for assessing the change in water surface
profiles due to channel modifications, and levees.

Special features of the steady flow component include: multiple plan analyses; multiple profile
computations; multiple bridge and/or culvert opening analysis; bridge scour analysis; split flow
optimization; and stable channel design and analysis.

Unsteady Flow Simulation. This component of the HEC-RAS modelling system is capable of
simulating one-dimensional; two-dimensional; and combined one/two-dimensional unsteady
flow through a full network of open channels, floodplains, and alluvial fans. The unsteady flow
component can be used to performed subcritical, supercritical, and mixed flow regime

14
(subcritical, supercritical, hydraulic jumps, and drawdowns) calculations in the unsteady flow
computations module.

The hydraulic calculations for cross-sections, bridges, culverts, and other hydraulic structures
that were developed for the steady flow component were incorporated into the unsteady flow
module.

Special features of the unsteady flow component include: extensive hydraulic structure
capabilities Dam break analysis; levee breaching and overtopping; Pumping stations;
navigation dam operations; pressurized pipe systems; automated calibration features; User
defined rules; and combined one and two-dimensional unsteady flow modelling.

Sediment Transport/Movable Boundary Computations. This component of the modelling

system is intended for the simulation of one-dimensional sediment transport/movable boundary
calculations resulting from scour and deposition over moderate to long time periods.

The sediment transport potential is computed by grain size fraction, thereby allowing the
simulation of hydraulic sorting and armouring. Major features include the ability to model a
full network of streams, channel dredging, various levee and encroachment alternatives, and
the use of several different equations for the computation of sediment transport.

The model is designed to simulate long-term trends of scour and deposition in a stream channel
that might result from modifying the frequency and duration of the water discharge and stage,
or modifying the channel geometry. This system can be used to evaluate deposition in
reservoirs, design channel contractions required to maintain navigation depths, predict the
influence of dredging on the rate of deposition, estimate maximum possible scour during large
flood events, and evaluate sedimentation in fixed channels.

Water Quality Analysis. This component of the modelling system is intended to allow the user
to perform riverine water quality analyses.
The current version of HEC-RAS can perform detailed temperature analysis and transport of a
limited number of water quality constituents (Algae, Dissolved Oxygen, Carbonaceous
Biological Oxygen Demand, Dissolved Orthophosphate, Dissolved Organic Phosphorus,

15
Dissolved Ammonium Nitrate, Dissolved Nitrite Nitrogen, Dissolved Nitrate Nitrogen, and
Dissolved Organic Nitrogen).

2.7 HEC-HMS

The Hydrologic Engineering Center-Hydrologic Modelling System (HEC-HMS) is designed to

simulate the precipitation-runoff processes of dendritic watershed systems. It is designed to be
applicable in a wide range of geographic areas for solving the widest possible range of
problems. This includes large river basin water supply and flood hydrology, and small urban or
natural watershed runoff. Hydrographs produced by the program are used directly or in
conjunction with other software for studies of water availability, urban drainage, flow
forecasting, future urbanization impact, reservoir spillway design, flood damage reduction,
floodplain regulation, and systems operation (Hydrologic Engineering Center (HEC), 2016a).

The program is a generalized modelling system capable of representing many different

watersheds. A model of the watershed is constructed by separating the hydrologic cycle into
manageable pieces and constructing boundaries around the watershed of interest. Any mass or
energy flux in the cycle can then be represented with a mathematical model. In most cases,
several model choices are available for representing each flux. Each mathematical model
included in the program is suitable in different environments and under different conditions.
Making the correct choice requires knowledge of the watershed, the goals of the hydrologic
study, and engineering judgment.

The program features a completely integrated work environment including a database, data
entry utilities, computation engine, and results reporting tools. A graphical user interface allows
the seamless movement between the different parts of the program. Program functionality and
appearance are the same across all supported platforms.

2.8 Data requirements

2.8.1 HEC-RAS

Geometric data consist of establishing the connectivity of the river system (River System
Schematic), entering cross-section data, defining all the necessary junction information, adding

16
hydraulic structure data (bridges, culverts, dams, levees, weirs, etc…), pump stations, storage
areas, and two-dimensional flow areas. The geometric data is entered by selecting Geometric
Data from the Edit menu on the HEC-RAS main window. The drawing area will be blank on
your screen, until you have either drawn in your own river system schematic or imported data
from a GIS.

Once the geometric data are entered, the modeller can then enter either steady flow or unsteady
flow data. The type of flow data entered depends upon the type of analyses to be performed.
For the discussion in this chapter, it is assumed that a steady flow hydraulic analysis will be
performed. The data entry form for steady flow data is available under the Edit menu bar option
on the HEC-RAS main window.
Steady flow data consist of: the number of profiles to be computed; the flow data; and the river
system boundary conditions. At least one flow must be entered for every reach within the
system. Additionally, flow can be changed at any location within the river system. Flow values
must be entered for all profiles.
Boundary conditions are required in order to perform the calculations. If a subcritical flow
analysis is going to be performed, then only the downstream boundary conditions are required.
If a supercritical flow analysis is going to be performed, then only the upstream boundary
conditions are required. If the modeller is going to perform a mixed flow regime calculation,
then both upstream and downstream boundary conditions are required. The Boundary
Conditions data entry form can be brought up by pressing the Reach Boundary Conditions
button from the Steady Flow Data entry form.
HEC-RAS has the ability to import geometric data in several different formats. These formats
include: a GIS format (developed at HEC); the USACE Standard Surveyor format; HEC-2 data
format; HEC-RAS data format; UNET geometric data format; and the MIKE11 cross section
data format. Data can be imported into an existing HEC-RAS geometry file or for a completely
new geometry file. Multiple data files can be imported into the same geometric data file on a
reach-by-reach basis.

2.8.2 HEC-HMS

The physical representation of a watershed is accomplished with a basin model.

17
Hydrologic elements are connected in a dendritic network to simulate runoff processes.
Available elements are: sub-basin, reach, junction, reservoir, diversion, source, and sink.
Computation proceeds from upstream elements in a downstream direction.
Meteorological data analysis is performed by the meteorological model and includes shortwave
radiation, precipitation, evapotranspiration, and snowmelt. Not all of these components are
required for all simulations. Simple event simulations require only precipitation, while
continuous simulation additionally requires evapotranspiration. Generally, snowmelt is only
required when working with watersheds in cold climates.
A geographic information system (GIS) can use elevation data and geometric algorithms to
perform the same task much more quickly. A GIS companion product has been developed to
aid in the creation of basin models for such projects. It is called the Geospatial Hydrologic
Modelling Extension (HEC-GeoHMS) and can be used to create basin and meteorologic models
for use with the program.

2.9 Stormwater management

2.9.1 Historical evolution of Abidjan drainage network

One finds in Bouvier, (1989), BCET, (1984), Colcanap and Dufour, (1982) and Bouvier and
Desbordes, (1990), elements which help in relating the evolution of the urban drainage network
of Abidjan. This one was closely related to the sewerage in general, the technical choice in the
first times consisting in treating wastewater and stormwater in the same unit network (combined
system). The technical mission of the Consulting Engineers Colcanap and Dufour, appointed
by the French Ministry of the Environment in 1981, produced a report of 300 pages on the
sanitation of Abidjan (Colcanap Michel and Dufour Philippe, 1982); that report focused on the
treatment of wastewater but also dealt with the drainage of stormwater. That report makes
understand that the construction of the drainage network of Abidjan goes back to the middle of
the years 1970, the first investments being initiated within the framework of the Program
FNA/BIRD: drainage networks were constructed in the neighbourhoods Adjamé, Attécoubé,
Treichville, Koumassi, Vridi, Bouet Port, Williamsville and Zone 4 for a sum of 8.9 billion
FCFA (an equivalent of 16.2 million USD) and in the “new” neighbourhood of Abobo and
Banco for a sum of 5 billion FCFA (9.1 million USD). Early 1980, the choice of the separate
system prevails and funds are allocated for development of Gouro catchment (5.5 million USD)
and the implementation of the drainage network of Riviera (11 million USD). It was at the same

18
time that the BCET of the Ministry of Transport, former BNETD, studied the development of
the Creek of Danga in the municipality of Cocody (BCET, 1984). Bouvier (1989) and
Desbordes (1990) studied in 1989-1990 the urban drainage in West Africa and reported also
the evolution of the drainage network of Abidjan. Bouvier chose Yopougon as the case study
of his PhD work and described the open canal system, the canal of UNIWAX1 included, which
was constructed during the decade 1975-1985.

2.9.2 Characteristics of catchments in Abidjan

Several documents from technical study reports, doctorates thesis, scientific research papers
give the geomorphological characteristics of the catchments of Abidjan (Bouvier Christophe,
1989, Cazenave A. et al, 1984, Desbordes M. and Bouvier C, 1990, Hauhouot, 2008, Jourda et
al., 2006, Jourda et al., 2003 and Sighmonou, 1986). The PhD theses of Bouvier and of
Sighomnou (1986) give in particular experimental estimates of the permeability of soil and
runoff coefficients. Jourda et al. (2003) studied the groundwater of the Continental Terminal,
prevailing aquifer in the south of the country, and their works made us understand the low
relative permeability of the grounds of the studied catchment. The research works within the
framework of the experimental watershed of the ORSTOM at Adiopoudomé in the south of
Yopougon - emphasized in Bouvier Christophe (1989), Cazenave A et al, (1984), Desbordes M
and Bouvier C (1990) and Sighmonou (1986) - also made us apprehend the aptitudes of the
grounds infiltration capacity and then their ability to generate runoff. Hauhouot (2008) gives in
his paper an insight on the vulnerability of Abidjan’s riverbanks and their exposure to the risk
of landslide, with an estimation of the damage caused by rainfall events in the precarious
neighbourhood of Attécoubé set up on riverbanks.

2.9.3 Hydrometeorological data and observations

If rainfall data are sufficiently available at SODEXAM, it is not the same case for the
observations of flows in channels and/or flowrates in urban waterways. The theses of Bouvier
and of Sighomnou, mentioned above, particularly related the simultaneous measurements of
hyetographs and hydrographs at the ORSTOM gage of Adiopoudomé. Desbordes and Bouvier
report written on the behalf of CIEH2 made the synthesis of it whereas before them, Cazenave

1
One of the Factories neighbourhoods of Abidjan
2
West and Central Africa’s hydraulics comity

19
et al. and Sighmonou (1983) reported partial results of field measurements. CONCEPT and ICI,
(2012) for their part talked about thirty pluviographic years of recordings carried out by
SODEXAM at Abidjan-Airport between 1970 and 2000, in the study of the Gourou catchment
(CONCEPT-ICI, 2012) realized in 2012. In that last document, is described the frequency
analysis of storms that made possible the establishment of design storm in an original way at
this precipitation gage.

2.9.4 Recent and under way development

The urban area of Abidjan which has been undergoing for several decades a rapid growth, is
subject to big urban development and infrastructures projects. On one hand, the projects that
are already implemented or under way have direct impact on land use and thus on their
imperviousness. On the other hand, the project of improvement of the drainage network must
consider future projects to come. The master plan of Abidjan planning (Oriental Consultants
Co. Ltd et al., 2014), completed in 2014, is obviously the collection of references concerning
the synthesis of the urban evolution of the city and of the orientation given by the public
authority to its future development. That document gave areas of future urban growth and
infrastructures projects.

2.9.5 Hydrological methods and approaches

As one of the main topics is the conversion or transformation of rainfall into runoff, a great part
of the literature review focused on the hydrological approaches. It was not about a pragmatic
review tending, on one hand, to select the most suitable methods to treat the topic, on the the
other hand to get estimated parameters and geomorphological factors whose direct
measurements were not possible due to lack of time and equipment. The already mentioned
works of Bouvier (1989 and 2004), of Sighomnou and of Desbordes and Bouvier were
obviously the first references as such. They offer the specific advantage indeed to estimate and
quantify runoff in the context of the city subject of this study; equations and values of the
hydrological parameters these authors resulted in are for “Abidjan” or at least West African.
On another side, recent reports of BRLi, (2015) and previous report of CONCEPT-ICI (2012),
report of study of Gourou catchment were made profitable considering they are another way of
dealing with runoff generation issues compared to the traditional approaches of Caquot or other

20
empirical methods. It was extracted from the last document original calculation of design storm
whereas the first gave useful comparative tools so that waste of time was avoided for this work.

2.10 Previous studies on flood events in Abidjan

Though flooding is a common problem in the floodplain of the Gourou catchment, researches
regarding this are limited. The current flood problems and the lack of studies in the area show
the importance of an in-depth study. Researches by Danumah (2016) and Kablan et al., (2017)
were on assessing flood risks under changing climate and land use in the District of Abidjan.
They focused on developing maps of flood hazard and risk levels for the whole town including
this catchment using remote sensing. He concluded that 34% of the city is at flood risk and that
extreme rainfall are more frequent in the city nowadays. Other studies, Célestin (2008) and
Savane et al., (2003) treated certain factors controlling flood risk such as rainfall risks and
uncontrolled urban growth in two municipalities which don’t include this study area. Kouamé
et al., (2013), Jourda et al., 2003 and Ahoussi et al., (2013) raised the issues of inefficiency of
the drainage network and impervious areas as main divers of flooding. Dongo et al., (2008)
developed a hydraulic model to prevent flooding events in some slums of Yopougon the biggest
municipality of Abidjan in the north of the town. Kangah and Della, (2015) used digital
elevation models (DEM) and geographic information systems (GIS) to determine flood areas
and identify the type of flood, and the risk factors in the Bonoumin-Palmeraie watershed, one
of the most impacted neighbourhoods in Abidjan. In that work, they use multicriteria to identify
floodprone areas in the watershed of Bonoumin – Rivièra Palmeraie and then did not include
any hydrologic nor hydraulic modelling to estimate flood that is likely to occur.
The lack of researches on the catchments of the country in general and of Abidjan in particular
because of the scarcity and quality of long-term hydrological data (Chu & Steinman, 2009) is
a major problem in Ivorian water issues management.

21
Chapter 3: Methodology

3.1 Study area

The District of Abidjan is located in the south of Côte D’Ivoire between latitude 5° 10’ and 5°
38’ North and longitude 3° 4’ and 5° 21’ West. It encompasses thirteen (13) municipalities
since 2001 and has an estimated population of 4, 739,752 inhabitants (Institut National de la
Statistique (INS), 2014) which represents 20.3% of the national population. Its area is about 2,
119 Km2 of which we have 566 Km2 of lagoons (Christophe Bouvier, 2004). Abidjan is the
economic capital of the country and the most developed city of the country. It is one of the main
trade centres of West Africa and even Africa and is the place where the African Development
Bank’s headquarter is. Abidjan shares border with the following towns and ocean:
• The Atlantic Ocean in the south;
• The city of Dabou in the southwest;
• The city of Grand-Lahou in the west;
• Agboville city in the north;
• The city of Grand-Bassam in the south-east;
• And Alépé city in the east.
In terms of geomorphology, the municipality of Cocody is on a large plateau. This relief is
separated from the municipalities of Adjamé and Plateau by a large neckline which starts in the
south of Abobo and ends at the bay of Cocody. With an average altitude around 40-50 m, it has
a north – south flow direction. Indeed, altitudes vary between 80 and 100 m in the north and
between 20 and 30 m in the south. This plateau ends on the Ebrié lagoon but the transition is a
clift on the side of the neighbourhood of Cocody. This cliff is detached from the lagoonal costal
line at Riviera where it behaves like a real bank from, separating the plateau from a small plain
of 2 to 9 m. With a west-east direction, this bank, just like the cliff, is made of a series of valleys.
These valleys and their tributaries are used as natural channels for the drainage of stormwater
of the municipality of Cocody (Kangah and Alla Della, 2015).

Because of the insufficiency and lack of maintenance of the drainage network on the catchment
as well as lawless settlement in the sites planned for infrastructures, residential areas are flooded
recurrently each rainy season, causing extensive damages on the existing infrastructures as well
as disturbances of the traffic, economic activities and loss of lives. Yearly floods have been

22
Figure 3.1: Location of Bonoumin and Palmeraie Watershed in Abidjan/Côte D'Ivoire

taking place in several places in the city: Gourou watershed (roundabout of Indénié), Angré,
Koumassi, Abobo, Yopougon, and Palmeraie.

3.1.1 Geographic Location and Boundaries

From Figure 3.1 one can see that the municipality of Cocody shares border with:
• The municipality of Adjamé in the west;
• The municipality of Abobo in the north;
• The municipality of Bingerville in the east
• And in the south with the municipalities of Marcory and Koumassi; but the Ebrié lagoon
is between Cocody and these two municipalities.

Cocody has an area of 132 km2 and is still growing even very fast because some years ago it
was among the smallest municipalities of Abidjan.

23
The watershed of Bonoumin-Rivièra Palmeraie, subject of this study, (figure 3.2) is located
entirely in the municipality of Cocody and between longitudes 3° 50' and 4° 10' West and
latitudes 5° 10' and 5° 30' North. This watershed is itself a sub-basin of the Bonoumin - Rivièra
Golf watershed and is the largest watershed of the municipality of Cocody and covers more
than ten neigbourhoods of which Bonoumin, Palmeraie, Allabra, Riviera 2. It remains a
beautiful residential area and is recognized for the quality and the architecture of its buildings.

Figure 3.2: Location of Bonoumin and Rivièra Palmeraie Watershed on Google Earth

It is a place where people who have high income live including diplomats.

Here is the entire view of Abidjan. In green is the watershed, area of this study and in red the
outlet of the watershed. The Ebrié lagoon is in the south of the watershed and starts from the
left of the map and ends almost at the right where it meets the Aghien lagoon. In yellow is the
modelled canal through HEC-RAS and the blue lines are the hydrographic network on the
catchment.

3.1.2 Socio-Economic Background

The municipality of Cocody is one of the best residential area of Abidjan in terms of comfort
and buildings, even for the infrastructures. That is noticed by the quasi permanent presence of

24
habitats of middle and high standing and the presence of big hotels of such as Hotel Ivorire,
Golf hotel, Pulman, Ibis and recently Radisson Blue as well as many embassies such as that of

Figure 3.3: Neighbourhoods of Cocody municipality

the United States of America, Russia, United Kingdom, and so on. Neighbourhoods (Figure
below) like Angré, Bonoumin, Rivièra 3 and 4, Les Rossiers and even Cocody are very
esteemed are known for their comfort.

From primary school to higher education including professional, many education centres are
located in Cocody: we have the famous University of Cocody now Felix Houphouet Boigny
University, from the name of the first president, the West African Catholic University. As
professional institutes there are the National Administration School, the National School of

25
Statistics and Applied Economy and the National Institute for Art and Culture (higher
education) that make this municipality an undeniable place for education and culture.
The national television is also located there and also some serious hospital centres like the
university hospital (CHU3) and the Marie Thérèse Polyclinic (PIZAM).
However, despite its rapid growth, there are no industry units in this municipality. It seems to
be planned for only habitations. Absence of disturbance noticed in many neighbourhoods make
Cocody a good place to live in.

The municipality of Cocody is still keeping its tradition of the past with its beautiful colonial
villas of the Old Cocody (Danga district) and still still has not urbanized areas where remain
traditional houses and even surrounding villages. Thus, M’pouto and M’Badon are Ebrié people
villages which seem to resist to the pressures of a modern city, that is a contrast. The district of
Blockhaus, just close to the Hotel Ivoire is a perfect illustration (KONATE, 2007). The
population of Cocody can be described by using two features: demography and welfare.
Concerning the demography, since 1998 and owing to crisis in the country, no national

census has been carried out to know exactly the population of Côte d’Ivoire. Thus, the
population of Cocody after 1998 was based on estimation. This information is confirmed by the
estimated population of Cocody released by PUIUR4 (2011) for the years 2005, 2010 and 2015
(respectively 242,208; 280,786 and 4,748,052 inhabitants). Nevertheless, according to the
Institut National de Statistique (INS), the population of Abidjan doubles every 10 years.

Cocody has a population of about 447,055 inhabitants (INS, 2014) and its population is still
growing with a growth rate of 4.8% against 7% in 1998. However, this rate is still greater than
the national growth rate which is 2.65% (INS, 2014).

Natural or non-built-up spaces occupy an important surface (56%) of the total area of Cocody
and are primarily constituted by vegetations (DEH, 2005). The habitats occupy 20% of the
Municipality area. The 73.4% of these habitats consist of individual houses of economic
standing and especially of middle and high standing. The Ebrié villages are the slums (8.3% of
the habitats) of the municipality and also some precarious places (6.1%). These precarious
living areas lie in the creek (Gobellet) of “2 Plateaux”, at Rivièra and at Cocody centre. But
those at “2 Plateaux” and Rivièra were destroyed two years ago.

3
Hospital
4
Development project initiated for the construction of urban infrastructures

26
3.1.3 Urbanization Process

The urban policy of the state officers was based on constructing the city of Abidjan in a modern
style with a western architecture and technology (Dongo et al., 2008). From 1928 to 2000, seven
master plans (1928, 1952, 1960, 1969, 1974, 1985 and 2000) were set to build and design
Abidjan as a great and well-known city; however, these plans did not really succeed leading to
two types of landscapes: on one side, there is a well-planned city with modern infrastructures
(Plateau, Cocody, Bietry, Yopougon-Millionnaire) and, on the other side, there is a not well-
planned sections (Abobo, Adjame Macaci and Boribana, Port-bouet, Koumassi campement,
Marcory Sans Fil) (Kouamé, 2003).

The description of the urbanization in Cocody district derives from the final report of
PUIUR (2011) entitled “Etude Stratégique pour la Gestion des Déchets Solides dans le District
d’Abidjan” (Strategic study for the Management of Solid Wastes in Abidjan District).
The district is stretched on 7476 hectares representing 20.8 % of the overall Abidjan district
area. The built area represents 16.8% of the total area, while the natural environment accounts
for about 56% of the total area of Cocody and is mainly dominated by the bush (PUIUR, 2011).
This data shows the potentiality of Cocody in terms of extension. It should be noticed that many
real estate companies contributed to boost urban development and mitigate housing shortage
and sanitation issues. The most known in Cocody are SIPIM, Batim-CI, and Laurriers. All types
of buildings exist in Cocody. There are high standing buildings, low-cost housings and slums.
The presence of these slums in this district is explained by its proximity to the working areas
that are the residential areas (PUIUR, 2011). However, these slums seem to disappear due the
new policy to prevent catastrophe in Abidjan and unplanned settlements. According to PUIUR,
2011, the land use is mostly dominated by habitats, current extensions, equipments activities,
natural environment (green spaces), and roads (Kablan, 2014).

Concerning the welfare, Cocody is among the well-planned and luxurious districts of Abidjan.
Accommodations are mostly expensive. That justifies its beautiful landscape.

3.1.4 Geomorphology
The District of Abidjan, in which our area of study is, presents globally a little contrasted and
monotonous relief. Indeed, we have three (3) units geomorphologically individualized (Tastet,
Martin, & Aka, 1979) which are:

27
in the north, high plateau on two (2) levels: 40 to 50 m and 100 to 120 m. These high plateaux
of the tertiary sector are notched by deep valleys resulting from the northern centre from area.
These valleys drain, like the different creeks, all water of the northern part towards the lowest
parts or outlet. These high plateaux of the north make the plateau of Cocody. In the South, the
low plates of Quaternary whose altitude is lower, variable generally between 8 and 12 m. Ce
are the outcrops of the offshore bar. And at side of these two types of units, we can quote sands
of the coast of Quaternary whose altitude varies between 6 and 8 m.

Figure 3.4: DEM of Abidjan

3.1.5 Hydro-Climatology

3.1.5.1 Climate

The movement of the Inter Tropical Convergent Zone (ITCZ) determines the climate in

Côte d’Ivoire (Saley, 2003). The study area is characterized by an equatorial

28
climate with four seasons (the description below is derived from Le front intertropical en
Afrique occidentale, (1989):

• The first rainy season (March to July) which reaches the first peak in June and

corresponds to the upward of the ITCZ and along which two third of the

annual rainfall is recorded;

• A short dry season (August-September). It is the coldest season of the year;

• A second peak of the rainy season (October to November), which corresponds to the

downward of the ITCZ; and

• A long dry season from December to February with few rainy days. It is the hottest

period of the year, especially February and March.

The subdivision of annual season is not respected currently mainly due to climate
variability/change that the planet witnesses.

3.1.5.2 Climate and Rain Data

The hydro-meteorological data used in this work has been provided by the Ivorian
Meteorological Agency (SODEXAM). Rain data from the station located at Abidjan airport
have been used for a while. But, since 2012, SODEXAM has been installing pluviometers in
the municipalities of Abidjan. The one at Cocody was installed in 2013.

a) Rain Data

The climate of Cocody is characterized by two rainy seasons (March to July

and October to November) with the first peak reached in June (figure 3.5), and the second peak
reached in October (Figure 3.4). The first rainy season is preceded by a long dry season
(December to February) and followed by a small dry season (August and September).

29
 600
518
500
Rainfall Depth (mm)

400

287.1
300

200 158.6 174.7 162.1 155.2

98.8
100 68.8 74.6
46.5 32.0
20.2
0

Months

Figure 3.5: Average monthly rainfall at Abidjan airport from 1961 to 2017

Below also is the trend in annual rainfall from 1961 to 2017. Years 2002 and 2003 don’t have
data; they have not been measured because of the political crisis started on September 18, 2002.
The highest rainfall depth was observed in 1963 (2755.7 mm) and the lowest in 1990 (1059.0).
Globally, rainfall decreased from 1961 to 2009 and has gone back up again since 2010. It is
expected intense rainfall this year as it was seen in May and June.

3000

2500
Rainfall depth (mm)

2000

1500

1000

500

0
1960 1970 1980 1990 2000 2010 2020
Year

Figure 3.6: Annual rainfall trend from 1961 to 2017

30
 a) Temperature

The data show that the average monthly temperatures from 1961 to 2016 vary from 24.2 ᵒC to
27.4 ᵒC. An increase temperature is observed from January to May, corresponding to the long
dry season. This is followed by a progressive decrease in the temperature from June to August,
corresponding to the period of the first peak of the rainy season (Figure 3.7). August is the
coolest month of Abidjan. The variations of temperatures are very important in this study,
because they affect human perception of warm days
29.0
28.1
27.8 27.7 27.7 27.6
28.0 27.4
27.0
27.0 26.4 26.3
Temperature

26.0
25.1
24.8
25.0 24.2
24.0

23.0

22.0

Months

Figure 3.7: Average monthly temperature from 1961 to 2017 at Abidjan airport

From the below ombrothermic diagram it can be concluded that temperatures are low during
the months of heavy rainfall and just after the longest rainy season. Therefore, it seems obvious
that the rainy season has effect on it even if temperatures are relatively high during rainy season.

600 29.0
Rainfall Depth (mm)

500 28.0
Temperature

400 27.0
26.0
300
25.0
200 24.0
100 23.0
0 22.0

Months

Rainfal Depth Temperature

Figure 3.8: Ombrotherlic diagram of Abidjan from 1961 to 2017

31
 b) Evaporation
The evaporation measured at the station of Abidjan-Airport is that of Pichet. It is very important
in March during the longest dry season (December-March) with a peak (Figure 09). In August,
evaporation drops significantly and can reach 25.70 mm. January, April and February are the
months during which evaporation varies considerably with coefficients of variation respectively

of 0,40,0,37 and 0.33.

Figure 9: Monthly evaporation at Abidjan airport from 2000 to 2009

3.1.6 Geology

Geological context of the town of Abidjan is that of the sedimentary basin of Côte D’Ivoire.
The sedimentary basin in Abidjan is characterized by a length of 350 km from the east to the
west and a north-south width, very reduced, which lies between 10 and 40 km. The successive
layers are in position monocline, and the dip in direction of the ocean.

One notes, within this sedimentary basin, a major accident of east-west direction presenting a
layout which corresponds appreciably to the lagoons. This break of lagoons determines two
sedimentary sequences; one not very thick (approximately 100 meters) located in the north, and
the other, very thick (3 000 meters) in the South. The sedimentary formations are of a large
variety: sands, clays, ferruginous sandstone and silts.

The stratigraphic log is made up from the top to the bottom, by (Aghui & Biémi, 1984):
• The formations of Quaternary which outcrop at the South of the lagoons break and in
the fluvio-lagoonal lowlands. They are primarily made up of sands, coarse sands, silts
or clays, muddy sands and sandy or silty;

32
 • The formations of the Continental Tertiary sector, which are made of coarse sands,
variegated clays, ferruginous sandstones and iron ores. All these formations are of age
Mio-Pliocene and result from the disintegration of the base;
• the formations of the Secondary, superior Jurassic of the superior Cretaceous and of the
marine Tertiary sector, are made up mainly of sands, the conglomerates, variegated
clays, broken into leaf clays with intercalations of marnes and sandstone, of the
sandstones, fluviatile sands and sometimes dolomitic sandy limestones. The Paleocene
one and the Eocene, on the other hand, are made by glauconitic clays, of sands and
small deposits limestone.

Figure 3.10: Geological map of Abidjan (Danumah, 2016)

3.1.7 Soil

The soils of Abidjan belong to the class of desaturated lateritic soils, depleted-modal, on tertiary
sand (continental terminal) according to Aghui and Biémi, (1984). They are ferralitic soils,
hydromorphic and recent.

➢ The ferralitic soils, met on low and high plateaux, have a structure in which the
deterioration of minerals is complete. The settlement of this pedological texture comes
from the process of ferralisation developed under the influence of the paleo-climatic
factors and the very old types of vegetation. The abundance of the rains and the high
temperatures involve the constitution of a profile staged with:
• a first not very thick horizon, low in humus and rich in organic matter;

33
 • a second horizon, very thick, with prevalence of red or brown colour and alumina and
iron abundance;
• a third argillaceous, compact and somewhat permeable horizon;
• a horizon basic, very thick, of variable colour related to the nature of the bed rock.
➢ the hydromorphic soils constitute the second important pedological element of the
sector of Abidjan. This hydromorphic aspect was caused by a pedological evolution
dominated by an excess water.
➢ the recent soils and very little evolved, although spatially more reduced than the two
others, developed in the littoral sector, and present coarser facies where sands are
dominant. They are the littoral dunes.

3.1.8 Hydrogeology

Abidjan District’s hydrogeology is made of homogeneous and highly permeable aquifers. The
lithological column of the mainland sedimentary basin highlights sands, sandy-clays and clays.
The aquifers of the coastal sedimentary basin are quite homogeneous and very permeable. The
sedimentary basin reservoirs which have an important role in the hydrogeological network three
(Aghui and Biemi, 1984).

• Quaternary aquifer;
• Continental Terminal aquifer, it is also called Abidjan’s;
• Aquifer Maastrichtian.

Only the Continental Terminal aquifer is used for water supply in Abidjan, the two others are
not used at the moment. It has very interesting chemical and hydrodynamic properties.

Except lenticular argillaceous benches, the aquifer of Abidjan is made of fluviatile coarse sands
to pass from variegated clays, sandy clays and clayey sands. It profits from a natural protection,
in particular its slope of north-south direction and the existence of the major break of the
lagoons and this protects it from a salt water intrusion in boreholes.

However, it is intensely exploited and, especially, is at high risks pollution that may originate
from industrial and domestic wastes dumped in urban environment.

3.1.9 Hydrology and Drainage

34
Abidjan is a seaside town. It is bordered by the Ebrié lagoon. This lagoon is a big open surface
water and has a length of 150 km and a surface area of 556 km2. It is supplied in fresh water by
three catchments (Agnéby, Comoé, Mé) and by some small inland (Bété, Djibi) and coastal
rivers (Angédédou, Gougbo, Bank value), (Kamagaté et al., 2016). It is connected to the sea by
the artificial channel of vridi. The Ebrié lagoon of east-west direction surrounds Cocody
municipality at the south. This lagoon is the outlet of all the creeks that are draining of the
Northern part of Abidjan city.

The Ebrié lagoon represents clearly the most important surface water of the South of the country
(Yao, Kouamé, Kouassi, Koffi, & Goula, 2015). It has a complex configuration with lakes more
or less closed, deep bays and very cut out banks. At the urban part of Abidjan, the lagoon is dug
by many pits and channels where depths are higher than twenty (20) meters (Bays of Cocody,
and of the Banco). These bays occupy close to 20% of the lagoonal plan (VARLET, 1978).

The Djibi (78 Km2) and the Bete (206 Km2) watersheds are partly urbanized while the Mé basin
(4,000 Km2) is mainly rural. The Mé flows into a channel downstream to the Aghien lagoon
(Kamagaté et al., 2016).

3.1.10 Characteristics of Bonoumin-Rivièra Palmeraie watershed

DEM of the study area was used to extract the physical characteristics of the watershed. DEM
of 30 m resolution was processed in ArcGIS 10.3 using the geospatial extension of HEC-
HMS (HEC-GeoHMS). In the next chapters, those parameters are described in details.

3.1.10.1 Shape

The main parameters that characterize a watershed are the perimeter (P), the area (A) that will
influence the resulting discharge, the length (L) related to the time of concentration (TC)and
the Gravelius coefficient (KG). this last one gives information on the form of the watershed
and is defined as the ratio of the watershed perimeter to that of circle having the same surface.
The area of the watershed the perimeter and the lag time (TLag) are directly computed in
ArcGIS during the terrain processing. The lag time is related to the time of concentration by
the equation 𝑇𝐿𝑎𝑔 = 0.6𝑇𝐶 . More details are provided in chapter 3.1 on models’ development.
About the Gravelius coefficient, it is mathematically given by the following formula:
𝑃 𝑃
𝐾𝐺 = ≈ 0.28
2√𝜋𝐴 √𝐴
Where: 𝑃 = Perimeter of the watershed (Km)
𝐴 = Area (Km2)

35
For this study area 𝑃 = 54.9 𝐾𝑚 and 𝐴 = 37.47 𝐾𝑚2 then
𝐾𝐺 = 2.53
As 𝐾𝐺 is not close to 1, it can be concluded that the watershed of Bonoumin-Rivièra
Palmeraie is elongated.
3.1.10.2 Spatial Distribution of elevations

DEM of Cocody was pre-processed into ArcGIS using HEC-GeoHMS for the extraction of
the physical parameters of the watershed. These GIS tools are very useful because they
provide satisfactory results and are time saving. These operations don’t only aim at extracting
watershed characteristics but also help describe the topography and understand the spatial
distribution of altitudes.
From the DEM of BRP watershed in Figure 3-10, it is obvious that altitudes of the watershed
are between 5 m and 121 m. They are very low compared to other countries. This is to testify
that the study area is seaside place. It also reveals that the southern part of the watershed is
lower than the northern part. This confirms the global topography of the city of Abidjan
(southern Abidjan is lower than northern Abidjan). People are at the moment settled in
lowlands: at the extreme south and centre. But the west and the north-western part also are
residential areas; these are the neighbourhoods of Angré, Les Rausiers and 2 Plateaux.
The most flooded (roundabout of Rivièra 3 and Allabra) area is located at that lower part.
This distribution is satisfactory and represent the actual situation of altitudes. The canal of
“Rue minister” that will be modelled starts at the centre part and is of south direction; it
means it flows in the outlet direction. It also obeys the law of hydraulics engineering: after the
construction of drainage infrastructures, natural flow should be established and then preserve
environment. Therefore, in this exercise they try to follow natural waterways. The Ebrié
lagoon lies at the south of the watershed; all these watercourses flow into it.
The next paragraphs will give more details on the BRP watershed as the hypsometric curve
and the slope distribution will give ideas about the way water flows on the watershed.

36
Figure 3.11: Digital Elevation Model of BRP watershed

37
3.1.10.3 Hypsometric Curve

The hypsometric curve provides an overall view on the watershed’s slope. This curve
represents the distribution of watershed total surface area against altitudes. It is generated
applying the following steps:
Classification and reclassification of elevations by using ArcGIS (Arc Toolbox). Five
(05) classes of elevation were defined based on grid codes corresponding to elevation
intervals: [5 - 28], [28 - 51]; [51 - 75]; [539 - 585]; [75 - 78].
Calculation of areas of altitude classes
Determination of partial area of each elevation class surface in percentage by using the
following simple formula:
𝐸𝑙𝑒𝑣𝑎𝑡𝑖𝑜𝑛 𝑐𝑙𝑎𝑠𝑠 𝑎𝑟𝑒𝑎
𝑃𝑎𝑟𝑡𝑖𝑎𝑙 𝑎𝑟𝑒𝑎(%) = × 100
𝑇𝑜𝑡𝑎𝑙 𝑎𝑟𝑒𝑎 𝑜𝑓 𝑡ℎ𝑒 𝑤𝑎𝑡𝑒𝑟𝑠ℎ𝑒𝑑
After partial altitude classes area have been determined, cumulative percentages were
determined from calculated partial areas. The hypsometric curve is then obtained by
plotting elevations (m) versus cumulative percentages (%) in Excel.
Table 3.3: Elevation data of BRP for hypsometric curve
Percentage
Percentage of
Partial Area Partial Area of partial cumulative
(m2) (Km2) Elevation (m) areas (%) areas
0 0 121 0.00% 0.00

5 - 28 1923666.23 1.92 98 5.13% 5.13%

28 - 51 9149688.872 9.15 75 24.42% 29.55%

51 - 75 11050252.32 11.05 51 29.50% 59.05%

75 - 98 14621836.39 14.62 28 39.03% 98.08%

98 - 121 718056.8606 0.72 5 1.92% 100.00%

38
All these elements were computed in ArcGIS and then copied into excel for plotting.
However, the hypsometric map is generated directly in ArcGIS. The hypsometric map of BRP
watershed is shown on the next figure:

Figure 3.12: Hypsometric map of BRP watershed

39
From these computations, the elevation data for the hypsometric curve were calculated and
plotted.

130
120
110
100
90
Elevation (m)

80
70
60
50
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100
Percentage of cumulative area (%)

Figure 3.13: Hypsometric curve of BRP watershed

This curve shows that the average slope of this watershed is steep; it means that the time of
concentration will be low. From the curve:
39% of the watershed area has elevation between 75 m and 98 m,
Almost 29% is of low altitude, less than 51 m and
Almost 2% of the watershed total area is above 98 m, between 98 m and 121 m.
Other parameters such as the global slope index Ig, the length of the equivalent rectangle, the
average elevation and the median elevation also provide useful information about the
watershed response. The average elevation is computed by applying the following equation:
∑ 𝐴𝑖 × 𝐻𝑖
𝐻𝑎𝑣𝑔 =
𝐴

Where 𝐴𝑖 = Partial area of elevation class,

𝐻𝑖 = Mean altitude of elevation class and
𝐴 = Total area of the watershed.

Table 3.4: Characteristics of the BRP watershed

40
 Parameter Elevation (m)
Highest elevation 121

Lowest elevation 5

Elevation > 5% 98

Elevation < 95% 28.5

Theoretical difference elevation ΔH 69.5

Average elevation Havg 64.9

𝐸𝑙𝑒𝑣5% − 𝐸𝑙𝑒𝑣95%
𝐼𝑔 =
𝐿
Where: 𝐼𝑔 = Global slope index

𝐸𝑙𝑒𝑣5% = Elevation > 5%

𝐸𝑙𝑒𝑣95% = Elevation < 95% and
𝑃+√𝑃2 −16𝐴
𝐿= , is the length of the equivalent rectangle.
4

54.9+√54.92 −16×37.47
𝐿= = 26.01 𝑚. Then
4
98−28.5
𝐼𝑔 = = 2.67 𝑚/𝐾𝑚
26.01

The average altitude is approximately 64.92 m while median altitude which corresponds
to elevation of 50% of the total watershed (H50%) is 59 m. These two parameters are close;
one can thus state that BRP watershed is of regular slope.
3.1.10.4 Spatial Distribution of Slope

The slope map is obtained through terrain processing while extracting the physical
characteristics of the watershed.

41
 Figure 3.14: Slope map of BRP watershed

The slope map shows areas with different slopes on the watershed. Indeed, from the map, it
can be seen that the light green and the yellow are dominant; specifically, slopes on the
watershed are likely between 8% and 20%. These slopes are still high, they will generate low
time of concentration. This is the confirmation of the previous part on the hypsometric curve.

42
3.1.11 Conclusion

The climate of Abidjan is wet; accordingly, the watershed enjoys this cool weather. From the
chapter 3.0, it rains a lot on the watershed with an average of annual rainfall being 1796.6 mm
and the highest more than 2000 mm. All this quantity of water flowing on a watershed that has
steep slope needs to be seriously addressed. All these factors are probably impacting floods in
this area.

3.2 Data Collection

Data used in this work are from various sources as they are of different types. Because the study
wanted to combine three software such as ArcGIS, HEC-HMS and HEC-RAS, the required
data needed to be in compatible format of the three.

3.2.1 Topographic Data

Topographic data are mandatory for both hydrologic and hydraulic studies. In this study it was
about hydrologic modelling first and further the hydraulic one. One should differentiate survey
data from elevations data from satellites. Satellite topographic data are quite acceptable
nowadays for hydrologic modelling, but should be of high resolution for hydraulic modelling.

For the first one, DEM of the study area was processed to extract the input to HEC-HMS. It
was for the whole country (Côte D’Ivoire) and for Cocody was extracted from it. This data was
directly downloaded from the USGS website just by creating an account. The DEM is available
in 30 m resolution but was resampled to 15 m.

Concerning the hydraulic part, the modeller tried to get field data through survey using a
theodolite. The canal’s length is more than 2 Km (the length the modeller was interested in).
Therefore, surveying this length was not an easy task. It is time consuming and costly. So, some
were obtained through survey and others in reports of Consulting Engineers provided by ONAD
(the Ivorian Agency for Sanitation and Drainage) created recently (2011). Fields measurements
were more about the dimensions of the canal because it is variable from the most upstream
section to the outlet. But elevations also were measured in one day. The picture below was
taken during that day of field measure with the help of a Consulting Engineer:

43
 Figure 3.15: The student into the field for collecting elevations and geometric data of
the canal on 11 May 2018

3.2.2 Soil Data

Soil data were required to generate the curve numbers (CN). As mentioned earlier, soil data
were combined with the DEM of BRP and land use to compute the CN grid in ArcGIS. These
data were provided by BNETD (National Bureau for Technical Study and Development) in
shapefile format and were processed with the other data.

3.2.3 Vegetation Data

Vegetation data was necessary in the computation of the curve number as it was merge to soil
data and DEM to generate the CN. For this study, it was downloaded on the website of the
USGS. It is Landsat image Oli 2 downloaded directly on the website and was processed in

44
ENVI 5.1 for reclassification before exporting it to ArcGIS. Image of a larger area (Cocody,
Adjamé, Abobo) was obtained and clipped later to meet our need.

3.2.4 Hydro-meteorological Data

In hydro-meteorological data, we have hydrological data and meteorological data. SODEXAM,

the place where I did the research internship is in charge of collecting meteorological data:
temperature, evaporation, rainfall, humidity, wind speed, sunshine. They gave the required data
for the hydrologic simulations: rainfall depth for specific events (flood event of May 11 and
June 19, 2018), IDF, rainfall depth for various return periods (5, 10, 25, 50, 100 and 200 years),
evaporation and temperature.

The Directorate of hydraulics of the Ministry of Economic Infrastructures (Ministry of

Hydraulics since June 2018) is the one responsible of the collection of discharge/flow data.
They collect or measure flow in the main rivers of the country (Comoé River, Sassandra River,
Cavaly River and Bandama River). The locations of flow measurement are provided in
appendix.

However, there is not any flow measurement in the southern part of the country, accordingly,
flows in canal are not measured neither in Abidjan or other cities by the Directorate of
hydraulics. The first flows measured in the city of Abidjan started with the Gourou watershed
Project, a very big project funded by the AFDB, West African Development Bank and the
government of Côte D’Ivoire. Today, they have got equipment and has been taking flow
measurement (flow in canal) for more than two (2) years. The modeller also initiated one day
flow measurement with the people of Gourou project on June 19, 2018.

3.3 Models and Softwares description

3.3.1 Overall Methodology

This chapter provides the theoretical background for the understanding of the data processing
and modeling procedures used in this study. The three software solutions and the mathematical
models used in this study are presented in detail. ArcMap is used for all GIS related tasks,
HEC-HMS for hydrologic- and HEC-RAS for hydraulic modelling. HEC-GeoHMS and HEC-
GeoRAS serve as the interface between GIS and the hydraulic and hydrologic modelling. At

45
the end of this section a brief literature review gives an overview about the applicability and
limitations of the applied models.

The forcing condition for any hydrologic model is the rainfall. Several rainfall sources are
available in variety of formats and can be used as is or in combination. HEC-HMS simulates
rainfall-runoff process (Scharffenberg, Ely, Daly, Fleming, & Pak, 2010). From the given
precipitation, it deducts losses and convolutes the excess rainfall with specified unit hydrograph
and routes it through the channel to generate a runoff hydrograph. This flow information is used
then in HEC-RAS to estimate water surface elevations. HEC-RAS simulates one and two-
dimensional flow and generates water surface profiles for given flow conditions. It is capable
of modelling both steady and unsteady flow conditions. The geometry file necessary for HEC-
RAS simulation can be created in a GIS environment using HEC-GeoRAS toolbar and can be
imported to a RAS environment (Khattak et al., 2016). However, with the new version 5.0.4 of
HEC-RAS this is possible directly in HEC-RAS using RAS mapper. In this case the output
from HEC-RAS simulation can be exported into a GIS environment for floodplain delineation
using either HEC-GeoRAS or RAS mapper.

The models chosen for hydrologic and hydraulic simulations are the United Army Corps of
Engineers Hydrologic Engineering Center’s HEC HMS and HECRAS models. The models
were used together with the GIS tool HEC-GeoRAS/RAS Mapper and HEC-GeoHMS coupled
with AutoCAD and the ArcGIS data model; HEC-GeoHMS provide an interface with GIS.
Drainage features and Hydrologic Response Units were delineated using the Digital Elevation
Model (DEM) and HEC-GeoHMS. Stage-Discharge rating curves were generated at each HRU
outlet using the simulation model HEC RAS. The necessary geometry files for HEC-RAS
simulation were developed using topographic data (AutoCAD) collected from ONAD and also
field measurement. Fields measurements were carried out to get observed discharged at a
particular outlet (J37). HEC-HMS was used to simulate the watershed response to rainfall and
sensitivity analysis was performed to have the effect of any over/under estimation of parameters
on the resulting hydrographs. After the indirect calibration and validation of the hydrologic
model for BRP watershed, the NRCS statistical storm events for 5, 10, 25, 50 and 100-year
recurrence intervals were run to determine the flows through various hydrologic elements of
the model. The flows for the 5, 10, 25, 50 and 100-year storm event were input in HEC-RAS
to generate water surface profiles. The water surface profiles from HEC-RAS were exported to

46
 the GIS environment and the floodplain was delineated using the RAS Mapper/HEC-GeoRAS
interface. The resulting floodplain was compared against the observed flood level.

Data collection

Pre-processing of Pre-processing of Preparation of

Spatial Data using RAS Spatial Data in thematic layers using
Mapper/HEC-GeoRAS HEC- GeoHMS ArcGIS

Hydraulic Modeling in Hydrologic Modeling

HEC-RAS using HEC-HMS

Post Processing of Data

Using HEC-GeoRAS

Creation of Inundation/Depth Maps

in ArcGIS

Figure 3.16: Chart showing the process involved in the methodology

47
3.3.2 Rainfall-Runoff Model: HEC-HMS

3.3.2.1 Fundamentals

HEC-HMS is an open source software for the modeling of the rainfall-runoff process developed
by the U.S. Army Corps of Engineering´s Hydrologic Engineering Center. The software
includes a graphical user interface for the management and analysis of the model data. It is
important to mention that HEC-HMS itself is not an actual hydrological model rather than a
software that enables the user to perform hydrological modeling based on a wide selection of
common mathematical models used in hydrology. In HEC-HMS, the rainfall-runoff process in
a watershed is represented in a simplified manner as shown in Figure 3.16

Figure 3.17: Runoff process in HEC-HMS (HEC, 2000)

This simplified representation of the runoff process does not account for the storage and
movement of water vertically within the soil layer. It is however sufficient to model a flood
hydrograph as the result of a storm (HEC, 2000). For modelling purposes, this simplified
hydrologic cycle is further divided into four components, which are modelled separately. The
models included in the software can thus be categorized as follows:
Loss Method: A model to compute the runoff volume is often referred to as the loss method
since it accounts for the losses that occur during a rainfall event as a result of infiltration and

48
evapotranspiration. For each time interval in the modelling process, the loss method calculates
the amount of water that contributes to the runoff in the river (effective rainfall).

Transform Method: Models of direct runoff are also called transform method, since they
convert the effective rainfall over a watershed into a hydrograph at the outlet of the watershed.
These models account for the surface roughness and geometry of the watershed.

Baseflow Method: Baseflow models are used to simulate the fraction of the runoff contributed
by groundwater.

Routing Method: If the analysed watershed is divided into sub-watersheds, the flow at the
outlet of a certain upstream watershed has to be routed through the river channel in the
downstream watershed. The models used to simulate this routing process are therefore called
routing methods. They account for the geometry and roughness of the relevant river channel.

Software Components
Figure 3.17 below shows a schematic overview of the HEC-HMS software
environment. In the control specifications, the computational time step and the date of the run
are defined. The meteorological model is the representation of the rainfall event that is intended
to be modelled. The physical basin model is essentially a simplified physical representation of
the watershed which is prepared with HEC-GeoHMS in this study. The main features of the
basin model are sub-basins, reaches and junctions. The modelling results comprise runoff
hydrographs for each sub-basin as well as graphical and numerical representations of rainfall,
losses and direct runoff for each sub-basin(Khattatı et al., 2016; Maidment & Mays, 1988).

Figure 3.18: HEC-HMS components (Heimhuber, 2013)

49
3.3.2.2 Hydrologic Model Selection and Description

Depending on the situation that is being modelled and the available data, an adequate
mathematical model for each of the previously defined four components of the rainfall-runoff
process needs to be chosen. In this study, the hydrologic modelling is performed primarily to
generate flood hydrographs with certain statistical return periods resulting from single design
storm events with the same statistical return periods. Since baseflow does not occur in the
concrete canal, it can be neglected in the modelling process. Furthermore, the rivers or canal
and the watersheds are ungauged and due to lack of equipment and time, complex field surveys
were not possible. As the object of this study is a drainage canal which must convey stormwater
from upstream at Saint Viateur (a sub neighbourhood in the northern part of Rivièra) of Abidjan,
we should first look at the hydrological issue of converting rain into runoff. Several traditional
methods (statistics) of rainfall-runoff conversion exist, but nowadays, with the development of
technologies, this task is carried out by the means of computer tools which model the
hydrological phenomenon of generation of runoff starting from the rains.

Computer-based or mathematical models are many and have been in use for more than 40 years
(CONCEPT and ICI, 2012 and CONCEPT and ROCHE, 2016), this period coincides with the
significant development of data processing, and then computers. They are each one based on a
well-defined approach of the hydrological phenomenon of conversion of precipitations into
runoff. There are as many approaches as of schools or hydrologists, each one privileging
physical process rather than others among the set of processes or methods which generate
runoff; some approaches account for, for example, the retention of precipitations by vegetation
for losses computation, contrary to other approaches which neglect this part of the hydrological
process. In addition, each part of the process hydrological is explained differently according to
the authors, following the example of the infiltration which is modelled by more than one
equation: Horton, Green-Ampt, Holtan. Hydrological models choose one approach or at least a
limited number among them.

In addition, each one of the mathematical equations which describes part of the hydrological
process comprises parameters in more or less great number and may be calculated easily. More
the approach adopted to explain the hydrological process is exhaustive, the more the number of
parameters is large. But at the same time, the determination of these parameters becomes

50
challenging, because it requires preliminary works, and measurements are not always possible,
neither easy to realize. The larger the number of parameters is, the more the total error of the
hydrological process simulation of the is high, because it is the result of the partial errors made
on the determination of each parameter(Choudhari, Panigrahi, & Paul, 2014; Liew, 2001).

Hydrologists specialist of models for rainfall-runoff are always challenged by the following
questions:
• Which approaches explain the best the studied hydrological phenomenon?
• Among those approaches, which are those which require less parameters and whose
determination is carried out easily in the allocated time?

The selected model should in all logic be based on an approach describing suitably the
hydrological process of the study regarding the objectives.

3.3.2.3 Suitability of HEC-HMS for a data scarce region like Abidjan

As it can be seen from the presentation of HEC-HMS, it computes runoff and determine the
hydrograph at the outlet of any given watershed, by accounting for the most significant parts of
the hydrological process of generation of runoff from a given rain. For the estimation of those
various parts, HEC-HMS offers the hydrologist/engineer a large variety of calculation methods
that touch the discussed cases. Each sub-model corresponding to those methods of calculation
requires more or less parameters and whose determination is more or less easy.

There exists a certain number of hydrologic and hydraulic models - sometimes coupled -
designed to better adapt to urban areas. We can quote SWMM among most classical or MIKE
URBAN among most recent. But these models require particularly large inputs to describe the
watershed and render the transform and transfer method equations. For SWMM for example,
not less than 13 parameters are necessary and if the hydrologist/engineer can to determine 8
from literature, it will remain 5 parameters which only be got from field measurements or by
calibration with observed and recorded events. As mentioned above, more the number of
parameters is large, the more the error of the hydrological process simulation is high, because
it is the result of partial errors made on the determination of each parameter (CONCEPT, 2015).

However, in Abidjan and in spite of the frequent rain causing floods with loss of lives, neither
flows nor water depth through drainage canals are recorded exempt for those of the Gourou

51
catchment project located in the Gourou catchment. And this makes the hydrological model
calibration difficult even impossible. Besides, soil parameters’ values are not either available,
as works interested in this field about Abidjan region are limited.

HEC-HMS, at the same time complete from the explanation point of view regarding the process
of transforming rainfall into runoff, offers alternatives of sub-models that don’t require an
excessive number of parameters. It is thus better appropriate for a situation of data scarce
region; and these data serve as basis to calculate some other parameters. Indeed, when there are
not available the use of other more sophisticated models will make them not very reliable and
trivial.
Based on this background, the models shown in Table 2.1 were chosen for each of the four
components of the runoff process.

Table 3.5 : Hydrologic model selection and categorization

Component Chosen Model Categorization

Loss Method SCS Curve Number event, semi-distributed, fitted parameter
Transform Method SCS Unit Hydrograph event, lumped, empirical, fitted parameter
Routing Method Muskingum event, lumped, empirical, fitted parameter

All three chosen models are designed to model single storm events rather than continuous

precipitation data (HEC, 2000). Furthermore, they are lumped models, meaning that spatial
variations of processes and characteristics are not considered explicitly rather than averaged for
each sub-watershed. The SCS Curve Number (CN) and Unit Hydrograph (UH) models

are both of empirical nature meaning that they are based on observations of the in- and output

of a certain system without trying to represent the actual conversion processes as done in
conceptual models. The Muskingum model is quasi conceptual since it is based on

simplified equations of shallow water flow. Each of the three chosen models and the
underlying mathematical equations are described in detail in the following sections.

Muskingum Method
Calculation of time of concentration (TC)
Kirpich method can be used for Calculation of TC as below (Alizadeh, 2001):
Tc = 0.0195L0.77S-0.385 Eq. 3.1

52
Where: TC is time of concentration (hour), S is mean slop of main river (m/m) and L is length
of main river (m).
Flow calculation in reaches
In Muskingum method, for flow modelling X and K parameters must be evaluated.
Theoretically, K is time of passing of a wave in reach length. They can be calculated
respectively by below equation (Alizadeh, 2001):
0.6𝐿
𝐾= Eq. 3.2
𝑉

Where: L is length of reach and V is velocity (m/s)

X parameter presented by Manning equation as below:
𝐼 0.5
𝑋 = 𝑛𝑝2/3 Eq. 3.3

Where: 𝐼 is the river slope, n is the roughness coefficient of Manning and P is the wet ted
perimeter (m) (Asif & Urarande, 2016).

Loss Method: SCS Curve Number Method

The U.S. Natural Resource Conservation Service (NRCS) (formerly the Soil Conservation
Service (SCS)) Curve Number method used in this study estimates the effective rainfall as a
function of the cumulative rainfall, the land use, the soil type and the antecedent moisture
condition of the soil. The model is described in detail in the National Engineering Handbook
(NEH) (NRCS 2004). It was created based on the analysis of a large number of small and
gauged agricultural watersheds throughout the US. Apart from the input precipitation, the
method uses a single parameter, the CN to characterize the watershed. The CN quantifies the
infiltration capacity and theoretically ranges between 0 (100% of the total rainfall infiltrate) to
100 (0% of the total rainfall infiltrate). The basic runoff equation of the CN method is shown
in Eq. 3.4. as described previously in chapter 2.
(𝑃−𝐼𝑎)2
𝑄 = (𝑃−𝐼𝑎)+𝑆 Eq. 3.4

where Q = runoff (mm)

P = rainfall (mm)
S = potential maximum retention after runoff begins (mm)
Ia = initial abstraction

Transform Method: NRCS Unit Hydrograph

The transformation of excess precipitation into runoff is commonly done using a unit
hydrograph (UH). A UH is defined as the hydrograph of surface runoff resulting from effective

53
rainfall in a unit of time (1 min or 1 hour) produced uniformly in space and time over the
watershed (Sherman, 1942). Every watershed has a unique UH which highly depends on the
topography and shape of the area and is usually generated through rainfall and runoff
measurements. Based on the UH, the flood hydrograph of any given rainfall event, can be
generated based on the principle of superpositioning(Hawkins, 1993; Maidment & Mays,
1988).

Model Representation of Rainfall

In order to address different hydrological modelling requirements, HEC-HMS includes a
variety of different ways to model precipitation. For the development of the flood hazard
zoning, the impact of floods with different statistical return periods is estimated. This approach
is commonly used to design flood protection infrastructure in a way that it can handle a
threshold flood with a specified return period. The return period is used to define the likelihood
of flood or storm events. A flood with a return period of 100 years can be expected to occur
once in a time span of 100 years. Since such a 100-year flood has a 1 % chance to occur in any
given year, its annual exceedance probability (AEP) is 1 %. For instance, in the USA and
Germany, major drainage-system elements are designed to resist a 100-year flood (Chin, 2006).
If the river of interest is gaged, flood discharges for different AEPs can be determined using
statistical analysis methods. For ungagged watercourses like the ones analysed in this study, it
is a common approach to estimate the flood discharge for a specified AEP by modelling a design
rainfall with the same AEP (HEC, 2000). In this approach, the definition of the storm duration
and the distribution of the rainfall intensity within the specified duration are crucial since both
parameters significantly influence the shape and magnitude of the resulting flood wave. Since
intensity distribution patterns for different rainfall durations were not found for the area of
investigation, the frequency-based hypothetical storm method included in HEC-HMS was used
to create design storms with specified return period. This method is based on the alternating-
block method which is described in detail by Chow et al. (1988). Hereby, design storms are
generated based on IDF curves in a way that the amounts of rainfall for any time interval within
the storm, centered around the peak intensity, have a consistent return period.
The input rain data for a frequency-based design storm with a return period of 100 years (100
years storm) are the precipitation depths for various rainfall durations with a 100-year return
period which can be derived from IDF curves. IDF curves are created based on continuous
precipitation data from a rain gauging station and represent the long-term precipitation
characteristics of the area around the gauging station.

54
3.3.3 Hydraulic Model: HEC-RAS

HEC-RAS is a hydraulic modelling software developed by the U.S. Army Corps of Engineer´s
Hydrologic Engineering Center. In this study, version 5.0.4 of HEC-RAS was used. The
software is capable of performing one, two and one/two-dimensional (1D/2D) steady and
unsteady-flow simulations(Claman, 2014). It comprises a graphical user interface, separate
hydraulic analysis components, data storage and management capabilities as well as graphics
and reporting facilities (HEC, 2016). The HEC-RAS Technical Reference Manual (HEC, 2016)
is a detailed and complete documentation of the model and the underlying equations. The
following section however explains the basics of 1-D unsteady-flow routing as well as the most
important features of the model that are necessary for the understanding of the applied
methodology.

3.3.3.1 Basics of One-Dimensional Flow Routing

In inundation analysis, flow modelling is used to simulate the flow of a flood wave through a
river reach and its floodplains. In hydraulics, the flow of water in a river is referred to as open
channel or free-surface flow since the water surface is exposed to the atmosphere. For
modelling and design purposes, continuity, momentum and energy equations have been
developed in the past to represent open channel flow in a mathematical way. These equations
are based on three basic laws of physics, which are the continuity of mass, the continuity of
energy and the continuity of momentum. Flow models simulate the flow through an open
channel in a way that satisfies these basic equations for open channel flow or simplified
versions of them.

In one-dimensional flow routing, flow through the river channel and the floodplains is treated
only in the longitudinal direction parallel to the conduit. Even though in reality, the flow in a
natural channel is never truly 1-D, these flow models were found to deliver acceptable results
for predicted hydraulic parameters in many applications (Arizona Department of Water
Resources, 2002). In the 1-D HEC-RAS flow model, the geometry of the channel and the
floodplains are represented by a series of cross sections along the reach.

In general, 1-D models are subdivided into steady and unsteady-flow models. In steady-flow

55
simulations, a constant inflow is modelled so that the depth of flow at any specified location
does not change over time. In comparison to that, a discharge hydrograph is applied as the
inflow into the modelled reach in unsteady-flow simulations resulting in changes in depth at
specified locations over time. The appropriate choice between these two basic approaches
highly depends on the situation that is intended to be modelled. Since the purpose of this study
is to identify flood prone areas and then delineate inundated areas and not to design detention
pools upstream, steady-flow modelling was chosen in order to account for the area reached by
water (Hydrologic Engineering Center (HEC), 2016b; Zainalfikry & Ghani, 2018).

Figure 3.19: Representation of Terms in the Energy Equation (HEC, 2016)

Another relevant aspect that is important for the modelling of open channel flow is the
distinction between subcritical and supercritical flow. This distinction is due to the fact, that
for a fixed discharge, there is a critical flow depth, for which the specific energy of the flow is
at its minimum. The specific energy of a flow is defined as the energy of the flow with reference
to the channel bed as the datum and thus only depends on the depth and the velocity of the flow.
If the specific energy is not at its minimum, there are two possible flow depths for a fixed
discharge. If the flow depth is higher than the critical depth, the flow is subcritical, if it is lower
the flow is supercritical. A practical distinction can be made based on the fact that in
supercritical flow, a wave resulting from any type of disturbance cannot propagate upstream
compared to subcritical flow. In consequence, water surface computations from one cross
section to another are done in downstream direction for supercritical flow and upstream
direction for subcritical flow. Situations in which both flow types occur in the modelled reach

56
are referred to as a mixed-flow regime.

3.3.3.2 1D Steady Flow Water Surface Profiles

HEC-RAS is currently able of performing 1D water surface profile calculations for steady
gradually varied flow in natural and constructed channels. Subcritical, supercritical and mixed-
flow regime water surface profile can be calculated (HEC, 2016). Topics discussed in this
section include equations for basic profile calculations; cross section subdivision for
conveyance calculations; contraction and expansion losses; application of the momentum
equation and limitations of the steady flow model (Hydrologic Engineering Center (HEC),
2016b).

a. Equations for Basic Profiles Calculations

Water surface profiles are computed from one cross section to the next by solving the energy
equation with an iterative procedure called the standard step method. The energy equation is
written as follows (from the above figure):
𝑎2𝑉2 𝑎1𝑉1
Z2 + Y2 + = Z1 + Y1 + + he Eq. 3.5
2𝑔 2𝑔
𝑎2𝑉2 𝑎1𝑉1
𝑍2 + 𝑌2 + = 𝑍1 + 𝑌1 + + ℎ𝑒 Eq. 3.6
2𝑔 2𝑔

Where Z1, Z2 = elevation of the main channel inverts,

Y1, Y2 = depth of water at cross sections,
V1, V2 = average velocities (total discharge/ total flow area)
a1, a2 = velocity weighting coefficients,
g = gravitational acceleration,
he = energy head loss
The energy head loss (he) between two cross sections is comprised of friction losses and
contraction or expansion losses. The equation for the energy head loss is as follows:

𝑎2𝑉22 𝑎1𝑉12
he = L𝑆̅f + C| − | Eq. 3.7
2𝑔 2𝑔

where: L = discharge weighted reach length

𝑆̅f = representative friction slope between two sections
C = expansion or contraction loss coefficient
57
b. Cross Section Subdivision for conveyance Calculations

The determination of total conveyance and velocity coefficient for a cross section requires that
flow be subdivided into units for the velocity is uniformly distributed. The approach used in
HEC-RAS is to subdivide flow in the overbank areas using the input cross section n-value break
points (locations where n-values change) as the basis for subdivision (Figure 3.19).
Conveyance is calculated within each subdivision from the following form of manning’s
equation:
𝑄 = 𝐾𝑆𝑓 1/2 Eq. 3.8

1
𝐾 = 𝑛 𝐴𝑅 2/3 Eq. 3.9

Where: K = conveyance for subdivision

n = Manning’s roughness coefficient for subdivision
A = flow area for subdivision
R = hydraulic radius for subdivision (wetted perimeter)
Sf = slope of the energy gradeline

The program sums up all the incremental conveyances in the overbanks to obtain a conveyance
for the left overbank and the right overbank. The main channel conveyance is normally
computed as a single conveyance element. The total conveyance for the cross section is
obtained by summing the three subdivision conveyances (left, channel, and right) (HEC, 2016).

Figure 3.20: HEC-RAS Default Conveyance Subdivision Method

(HEC,2016)

58
An alternative method available in HEC-RAS is to calculate conveyance between every
coordinate point in the overbanks (Figure 3.20) The conveyance is then summed to get the total
left overbank and right overbank values. This method is used in Corps HEC-2 program. The
method has been retained as an option within HEC-RAS in order to reproduce studies that were
originally developed with HEC-2.

Figure 3.21: Alternative Conveyance Subdivision Method (HEC, 2016)

The two methods for computing conveyance will produce different answers whenever portions
on the overbank have ground sections with significant vertical slopes. In general, the HEC-RAS
default approach will provide o lower total conveyance for the same water surface elevation.

Contraction and Expansion Losses Evaluation

Contraction and expansion losses in HEC-RAS are evaluated by the following equation:

𝑎2𝑉22 𝑎1𝑉12
hce = C| − | Eq. 3.10
2𝑔 2𝑔

where: C = contraction or expansion coefficient

The program assumes that a contraction is occurring whenever the velocity head is downstream
is greater than the velocity upstream. Likewise, when the velocity head upstream is greater than
the velocity head downstream, the program assumes that a flow expansion is occurring (HEC,
2016).
Typical C value can be found in chapter 2, “Basic Data Requirements.”

59
c. Application of the Momentum Equation

Whenever the water passes through critical depth, the energy equation is not considered to be
applicable. The energy equation is only applicable to gradually varied flow situations, and the
transition from subcritical to supercritical or supercritical to subcritical is a rapidly varying flow
situation. There are several instances when the transition from subcritical to supercritical and
supercritical to subcritical flow can occur. These include significant changes in channel slope,
bridge constrictions, drop structures and weirs, and stream junctions (HEC, 2016). In this work,
the momentum equation has been chosen to compute the energy losses at the junction at the
confluence of flow between the canals of” Rue minister” and “C2”. In some of these instances
empirical equation can be used (such as drop structures and weirs), while at others it is
necessary to apply the momentum equation in order to obtain an answer.

Within HEC-RAS, the momentum equation can be applied for the following specific problems:
the occurrence of a hydraulic jump, low flow hydraulics at bridges, and stream junctions. In
order to understand how the momentum equation is being used to solve each of the three
problems, a derivation of the momentum equation is shown here.
The momentum equation is derived from Newton’s second law of motion:

Force = Mass x acceleration (change in momentum)

∑ 𝐹𝑥 = 𝑚𝑎 Eq. 3.11

Applying Newton’s second law of motion to a body of water enclosed by two cross sections at
locations 1 and 2 (Figure 3.21) the following expression for the change in momentum over a
unit time can be written:

𝑃2 − 𝑃2 + 𝑊𝑥 − 𝐹𝑓 = 𝑄𝜌∆𝑉𝑥 Eq. 3.12

Where: P = Hydrologic pressure force at locations 1 and 2.

𝑊𝑥 = Force due to the weight of water in the X direction.
𝐹𝑓 = Force due to external friction losses from 2 and 1.
𝑄 = Discharge.
𝜌 = Density of water.
60
 𝑉𝑥 = Change on velocity from 2 to 1, in the X direction.

Figure 3.22: Application of the Momentum Principle

Hydrostatic pressure Force

The force in the X direction due to hydrostatic pressure is:
𝑃 = 𝛾𝐴𝑌̅ cos 𝜃 Eq. 3.13

The assumption of a hydrostatic pressure distribution is only valid for slopes less than 1:10.
The cos 𝜃 for a slope of 1:10 (approximately 6 degrees) is equal to 0.995. Because the ordinary
channels slope is far less than 1:10, the cos 𝜃 correction for depth can be set to equal to 1.0
(Chow, 1959). Therefore, the equations for the hydrostatic pressure force at section 1 and 2 are
as follows:

𝑃1 = 𝛾𝐴1 𝑌̅1 Eq. 3.14

𝑃2 = 𝛾𝐴2 𝑌̅2 Eq. 3.15

Where: 𝛾 = Unit weight of water

𝐴𝑖 = Wetted area of the cross section at location 1 and 2
𝑌i = Depth measured from water surface to the centroid of the cross sectional

61
 area at locations 1 and 2.

Weight of Water Force

Weight of water = (unit weight of water) x (volume of water)
𝐴1 +𝐴2
𝑊 = 𝛾( )𝐿 Eq. 3.16
2

𝑊𝑥 = 𝑊 × sin 𝜃 Eq. 3.17

𝑍2 +𝑍1
sin 𝜃 = = 𝑆0 Eq. 3.18
𝐿

𝐴1 +𝐴2
𝑊𝑥 = 𝛾( )𝐿 × 𝑆0 Eq. 3.19
2

Where: 𝐿 = Distance between section 1 and 2 along the X axis.

𝑆0 = Slope of the channel, based on mean bed elevation
𝑍𝑖 = Mean bed elevation at locations 1 and 2

Force of External Friction

𝐹𝑓 = 𝜏𝑃̅𝐿 Eq.3.20
Where: 𝜏 = Shear stress
𝑃̅ = Average wetted perimeter between sections 1 and 2

𝜏 = 𝛾𝑅̅ 𝑆𝑓̅ Eq. 3.21

Where: 𝑅̅ = Average hydraulic radius (R=A/P)

𝑆𝑓̅ = Slope of the energy grade line (friction slope)

𝐴̅
𝐹𝑓 = 𝛾 𝑃̅ 𝑆𝑓̅ 𝑃̅𝐿 Eq. 3.22

𝐴1 +𝐴2
𝐹𝑓 = 𝛾( )𝑆𝑓̅ 𝐿 Eq. 3.23
2

62
Mass time acceleration
𝑚𝑎 = 𝑄𝜌∆𝑉𝑥 Eq. 3.24

𝛾
𝜌 = 𝑔 and ∆𝑉𝑥 = (𝛽1 𝑉1 − 𝛽2 𝑉2) Eq. 3.25

𝛾
𝑚𝑎 = 𝑄 𝑔 (𝛽1 𝑉1 − 𝛽2 𝑉2 ) Eq. 3.26

Where: 𝛽 = Momentum coefficient that accounts for a varying velocity

Distribution in irregular channels.

Substituting Back into Equation 3.12 and assuming Q can vary from 2 to 1

𝐴 +𝐴 𝐴 +𝐴 𝑄 𝛾 𝑄 𝛾
𝛾𝐴2 𝑌̅2 − 𝛾𝐴1 𝑌̅1 + 𝛾 ( 1 2 2 ) 𝐿𝑆0 − 𝛾 ( 1 2 2 ) 𝑆𝑓̅ 𝐿 = 𝑔1 𝛽1 𝑉1 − 𝑔2 𝛽2 𝑉2 Eq. 3.27

𝑄2 𝛽2 𝑉2 𝐴1 +𝐴2 𝐴1 +𝐴2 𝑄1 𝛽1 𝑉1
+ 𝐴2 𝑌̅2 + ( ) 𝐿𝑆0 − ( ) 𝐿𝑆𝑓̅ = + 𝐴1 𝑌̅1 Eq. 3.28
𝑔 2 2 𝑔

𝑄2 2 𝛽2 𝐴1 +𝐴2 𝐴1 +𝐴2 𝑄1 2 𝛽1
+ 𝐴2 𝑌̅2 + ( ) 𝐿𝑆0 − ( ) 𝐿𝑆𝑓̅ = + 𝐴1 𝑌̅1 Eq. 3.28
𝑔𝐴2 2 2 𝑔𝐴1

Equation 3.28 is the functional form of the momentum equation that is used in HEC-RAS. All
applications of the momentum equation within HEC-RAS are derived from equation 3.28

1D Steady Flow Program Limitations

The following assumptions are implicit in the analytical expressions used in the current version
of the program:
1. Flow is steady
2. Flow is gradually varied. (Except at hydraulic structures such as bridges, culverts, and
weirs. At these locations, where the flow can be rapidly varied, the momentum equation
or other empirical equations are used.)
3. Flow is one dimensional (i.e, velocity components in directions other than the direction
of flow are not accounted for.)
4. River channels have “small” slopes, say less than 1:10.
Flow is assumed to be steady because time dependent terms are not included in the energy
equation (equation 3.5). Flow is assumed to be gradually varied because Equation 3.5 is based

63
on the premise that a hydrostatic pressure distribution exists at each cross section. At locations
where the flow is rapidly varied, the program switches to the momentum equation or other
empirical equations. Flow is assumed to be one-dimensional because the total energy head
𝑎𝑉 2
equation of a cross section: 𝐻 = 𝑊𝑆 + is based on the fact that the total energy head is the
2𝑔

same for all points in a cross section.

The limit on slopes as being less than 1:10 is based on fact that the true derivation of the energy
equation computes the vertical pressure head as:

𝐻𝑝 = 𝑑 cos 𝜃
Where: 𝐻𝑝 = vertical pressure head
𝑑 = depth of the water measured perpendicular to the channel bottom.
𝜃 = channel bottom slope expressed in degrees.

For a channel bottom slope of 1:10 (5.71 degrees) or less, the cos(𝜃) is 0.995. So instead of
using 𝑑 cos(𝜃), the vertical pressure head is approximated as d and is used as the vertical depth
of water. As you can see for a slope of 1:10 or less, this is a very small error in estimating the
vertical depth (0.5%).

If HEC-RAS is used on steeper slopes, you must be aware of the error in the depth computation
introduced by the magnitude of the slope. Below is the table of slopes and the cos(𝜃):

Figure 3.22: Slopes and 𝑐𝑜𝑠(𝜃) (HEC, 2016)

If you use HEC-RAS to perform the computations on slopes steeper than 1:10, you would need
to divide the computed depth of water by the cos(𝜃) in order to get the correct depth of water.
Also, be aware that very steep slopes can introduce air entrainment into the flow, as well as

64
other possible factors that may not be considered within HEC-RAS (Silva, Bonumá, & Uda,
2010).

3.3.3.3 HEC-RAS justification in the context of Abidjan

Abidjan is a city where the system of waterways or creeks and channels are particularly dense
and long of 137 km (the canal we are working on is about 2.4 km). These waterways are large
in some places and present steep slopes which facilitate the drainage of stormwater towards
various points of the lagoon and its bays. The buried drainage network (conduits) is not very
developed and all its tributaries or ramifications end in open channels.
In Abidjan, the most known historical events of flooding that caused more losses of lives and
damages from stormwater were the result of drainage canal and drains overflow. In comparison,
the damage which can be caused by the buried network remains relatively “minor” and are
especially due to an inversion of the hydraulic gradient because of the overflow of the major
open canal network. Dealing with the issues of stormwater drainage in Abidjan results mainly
recalibrating and redesigning the major network of waterways which is supposed to cope with
it. This can only be done by a modelling of this network, agreeing with the reality and allowing
to simulate flowing conditions that are taking place in canals or creeks with a high reliability.
HEC-RAS is indicated perfectly to achieve these goals. The stormwater management model
(SWMM), from its part simulates also open channels but for conduits with gravity flow and
requires a lot of data.

3.4 Models development

3.4.1 HEC-HMS

In this study, as mentioned initially, the ArcGIS extension of HEC-HMS, HEC-GeoHMS was
used to pre-process the hydrologic model data. The DEM of Cocody of 30 m resolution
resampled to 15 m has been applied in order to extract the physical characteristics of the
watershed.
Before the extraction of the watershed characteristics, a land use map prepared on Envie 5.1
was combined with a soil map to generate the curve numbers of the sub-basins of the watershed.

65
Therefore, by lack of observed discharges, the calibration and validation were done indirectly
by comparing the flood extents from HEC-RAS and RAS Mapper/HEC-GeoRAS to the May
11, 2018 and June 19, 2018 flood events observed by the modeller.

3.4.1.1 HEC-HMS inputs preparation

For a watershed to respond to a rainfall event on HEC-HMS model, this should include a basin
model, a meteorological model and a control specification.
This software thus allows the modeller to use different types of methods for the modelling
basins and rain. The control specification allows to specify for each case the start date and end
date of the simulation and the time step of the (Derdour, Bouanani, & Babahamed, 2017).

3.4.1.2 Pre-processing and watershed physical extraction

Kouamé et al. provided soil characteristics and detailed explanation on their granulometry. and
this put together with the NRCS hydrologic soil group allowed the computation of the CN in
ArcGIS. The NRCS gave four hydrologic soil groups for the computation of the CN. The groups
are shown below in the table:

Figure 3.23: Hydrologic Soil Group (NRCS, 1986)

In the TR-55, these four hydrologic soil groups are described regarding their infiltration rate
and main characteristics as follows (NRCS, 1986), (Fleming & Scharffenberg, 2000; Yener et
al., 2007):

66
Table 3.6: Soil texture (Kouamé et al., 2011)

% %
Type of Soil % Silt Texture
Sand Clay

Sandy-
1 Hydromorphic Soil 33.0 28.9 38.1
Clay
Ferralitic soil highly
unsaturated in high Loamy-
2 87.1 6.9 6.0
pluviometric zone; Tropical Sand
ferruginous soil
Soil over magmatic rocks and Clay-
3 20.6 48.0 31.4
cuirass zones Loam
Ferralitic soil highly
unsaturated in low pluviometric Sandy-
4 75.4 9.4 12.2
zone; Ferralitic soil middling Loam
saturated

Group A soils have the highest infiltration rates and consist chiefly of deep, well to
excessively drained sand or gravel.
Group B soils have moderate infiltration rates and consist of moderately well to well drained
soils with moderately fine to moderately coarse textures.
Group C soils have low infiltration rates and consist of soils with a layer that impedes
downward movement of water and soils with moderately fine to fine texture.
Group D soils have very low infiltration rates and consist chiefly of clay soils with a high
swelling potential.

Considering the above information and added to the soil file we got from the BNETD, the soil
of the watershed was characterized and further provided all the necessary information to
generate the CN.
We have the four groups present in the watershed. Abidjan is located in the south of the country
as said earlier, and this part is mainly sandy but mixed to some clay and loam. This has been
mentioned in the description of the geology of the study area.

67
The step was the processing of the land use, soil data and the DEM to generate the CN. The
land use was reclassified and converted into polygon so that it can be merged with the soil data.

Figure 3.24: Reclassified land use (left) and soil data (right) of BRP
The land use was reclassified in four groups before merging it with the soil data. Ones they are
merged, they are combined with the DEM to generate the CN numbers.

Figure 3.25: CN generation in HEC-GeoHMS

68
 The CN lie between 82 and 86. This is understandable as we are in an urban area. It means that
this watershed is likely to generate much runoff.

Figure 3.26: Curve Number grid generated using HEC-GeoHMS

Watershed physical characteristics extraction is known as terrain pre-processing. To do it, the

following steps are performed successively using HEC-GeoHMS in ArcGIS. This also can be
done using ArcHydro tools in ArcGIS (Oleyiblo & Li, 2010).

The basin model was created by applying the HEC-geoHMS software functionality within the
ArcMap GIS environment. The first major step in creating the basin model was to delineate
the stream network and the watershed boundaries of the area of interest. This process is
commonly referred to as terrain pre-processing as mentioned earlier and is entirely based on
the input DEM. The following grid files were derived from the DEM by following the step by
step functionality of HEC-HMS.

69
 ➢ Fill Sinks grid: This function creates a depressionless or hydrologically corrected
DEM based on the input DEM. Therefore, the software automatically increases the
elevation value of any pit cell to the level of the surrounding terrain.

➢ Flow Direction grid: This grid is delineated from the Fill Sinks grid. In the grid
processing, the direction of the steepest descent to a neighbour cell is defined for each
grid cell.

➢ Flow Accumulation grid: This grid is delineated from the flow direction grid
and defines the number of upstream cells draining into any given cell in the grid.

➢ Stream Definition grid: In this step, the cells that form the stream network are
defined based on a threshold number of cells that drain into a given cell. In this analysis
the threshold for the definition of streams was set to 1 km2
respectively. The result is a grid, in which the stream network is represented by lines
of connected grid cells that all fulfils the threshold criteria.

➢ Stream Segmentation grid: This grid is created by splitting the streams as defined
in the stream definition grid at any junction.

➢ Catchment grid: For every stream segment defined by the stream segmentation grid,
the corresponding watershed is delineated and stored in a grid file.

To complete the terrain processing, the vector layers were created based on the outcomes of the
previous computational steps:

➢ Catchment Polygons: This function uses the catchment grid to delineate the
boundaries of each sub-basin in the form of a vector layer.

➢ Drainage Line: The stream segments defined by the stream segmentation grid are
transformed into a vector stream layer by this function.

➢ Adjoint Catchment: In this step, the upstream sub-basins are aggregated at any stream

70
 confluence. This step is not hydrologically relevant but enhances the computational
performance in subsequent steps.

Figure 237: HEC-GeoHMS tools bar within ArcGIS and terrain pre-processing menu
A project point has to be defined after the completion of the terrain pre-processing. The project
point defines the outlet of the watershed that is intended to be modelled and thus has to be
placed on a drainage line. Based on the outcomes of the terrain pre-processing and the definition
of the project point, HEC-HMS delineates the project area and creates all necessary layer files
for this area. All the created data is stored in a new geodatabase.

The resulting watershed has a total area of 37.47 km2, a perimeter of 54.9 Km and includes 8
sub-basins with a minimum area of 0.011 km2.

For each of the resulting stream segments and the related sub-basins, a serious of physically
based characteristics were computed based on the depressionless DEM by using the HEC-
GeoHMS functions. These characteristics include the lengths and slopes of each river segment
as well as the average basin slope and the longest flow-path of each sub-basin. The resulting
data is automatically stored in the attribute table of the river and sub-basin layer.

Figure 3.28 on the following page shows the extend of the outlet and the 8 sub-basins as a result
of the previously described methodology in detail. For each sub-basin, the flow channel as well
as the longest flow path is illustrated.

71
As mentioned in Chapter 2, the hydrologic modelling was based on the NRCS Curve Number
loss method, the NRCS Unit Hydrograph transform method and the Muskingum routing
method. The computation of the CN has been described in this chapter, and the review of
literature gave some details on the method of unit hydrograph. For the Muskingum routing
method part, it is based on some work done in Abidjan and also based on the knowledge of the

Figure 3.28: Lag times computed in ArcGIS

study area on one hand and on the other hand on the results of the terrain processing. The lag
times of the sub-basins were computed during the process and are low. The figure below is the
illustration:

The NRCS shew that 𝑇𝐿𝑎𝑔 = 0.6𝑇𝑐 , where 𝑇𝐿𝑎𝑔 is the lag time and 𝑇𝑐 is the time of
concentration. It means that the Tc will also be low, this show that water flows rapidly on the
watershed and then from the knowledge of the watershed, Muskingum K and X were estimated.
The terrain is hilly and the watershed is urbanized. The shape, the size and soil type of the
watershed impact these parameters. As seen in chapter 2, the Muskingum X accounts for the
storage or retention through the canal. It lies between 0 and 0.5(Skhakhfa & Ouerdachi, 2016).

72
Figure 3.29: Basins model of Bonoumin and Rivièra watershed

73
3.4.1.3 Import of the GeoHMS results to HEC-HMS

After extracting all the characteristics of the watershed and computing some HEC-HMS input
parameters, the next step is to import this project to HEC-HMS for it to be finalised and
prepared for simulations. The main input that is to be entered now is the precipitation. Whatever
it is just a water depth or rainfall intensity. So, the GeoHMS results were imported to HEC-
HMS, the figure below shows details:

Figure 3.30: Basin Model in HEC-HMS

74
3.4.2 HEC-RAS

As it was not possible for the canal, subject of this study, to appear on the TIN of the watershed
because the precision was not so good, AutoCAD data and field measurements were combined
to establish the canal geometric data. Indeed, HEC-GeoRAS needs a topographic data of good
precision. Therefore, the DTM or TIN should not have a resolution less than 10 m. but one can
work with DTM of 30 m resolution for natural rivers. Topographic data constitute the basis on
which all hydraulic modelling works are based. It is thus very essential and crucial to pay
particularly attention to the method of procuring those data. Several works highlight the
importance of precision of topographic data: Casas et al. (2006) quoted by Geoffroy (2007) put
ahead topographic precision like the “most critical factor” when hydraulic modelling is being
carried out. In this article, they tested various topographies on one section of a river of 2 km
using HEC-RAS software by varying the different parameters (coefficient of Manning-Strickler
and flow). They compare the results (extent or inundated areas and average water surface
elevation) starting from a topography of reference defined thanks to a ground survey using a
GPS and a bathymetric survey. Sensitivity analysis carried out on the coefficients of Manning
confirms the idea according to which more the flow is important less the model is sensitive to
the variations of these coefficients.

But in this study, due to urbanization, the natural waterway has been modified and then has
been built in a canal with variable dimensions (cross sections). A field measurement campaign
was initiated in order to get the profiles of the canal.

3.4.2.1 Model Geometry creation

One of the objectives of this work is estimation of flood extent and depth. To achieve this, the
HEC-RAS project should be georeferenced so that visualization can be performed. Generally,
this task is performed using HEC-GeoRAS or now RAS Mapper with the new version of HEC-
RAS. Unfortunately, as mentioned above, it was not directly the case for this study for the
reasons mentioned above (section 3.2.1).
An AutoCAD map was used to identify the canal centreline and also elevations. The canal
centreline was extracted by superimposing a web imagery (Google earth) and the DEM of
Cocody. Ones the canal was digitized, it was called or brought in the HEC-RAS geometric data
editor for its completion.

75
The field measurements gave the dimensions of cross sections. Knowledge of the site was also
very helpful in this project.

3.4.2.2 Model Completion in HEC-RAS

After the extraction of the canal centreline, it was brought within HEC-RAS geometry editor
for its completion. The first thing after the centreline is drawn is to draw the cross sections
perpendicularly to the centreline. They are drawn from upstream to downstream and from left
to right looking in downstream direction. The expansion and contraction coefficients are
defined. Figure 3.31 shows the river schematic with cross sections and the other elements of
the geometry. The modelling was carried out on two canals: C2 and the canal of “rue minister”
and they come together at the junction “carrefour commissariat” which is the most critical point
of the modelled length. The grey block lines are culverts. Information about culverts were
obtained through the survey did in the month of May (Alaghmand, Abdullah, Abustan, &
Eslamian, 2012).

Figure 3.31: Canal network developed in HEC-RAS

76
Figure 3.32: Manning's n values (Chow, 1959)

a. Manning’s n-Value

Selection of an appropriate value for Manning’s n is very significant to the accuracy of the
computed water surface elevations. The Manning’s n is highly variable and depends on a
number of factors including: surface roughness; vegetation, channel irregularities, channel

alignment; scour and deposition; obstruction; size and shape of the channel; stage and
discharge; seasonal changes; temperature; and suspended material and bedload (HEC, 2016).

There are several references a user can access that show Manning’s n values for typical
channels. An extensive compilation of n values for streams and floodplains can be found in
Chow’s “Open-Channel Hydraulics” (Chow, 1959).

77
 In this work, the canal is in concrete and then the Manning’s n values are known (𝐾𝑠 = 70; then
1
𝑛 = 70 = 0.014). Except the floodplains which are of variable constitutes: asphalt road (𝑛 =

0.016 and earth (𝑛 = 0018).

b. Expansion and Contraction Coefficient

These loss coefficients are applied in the hydraulic computations in order to account for energy
losses resulting from contraction and expansion of flow due to changes in cross section
geometry along the reach. The energy loss caused by a transition in channel geometry is
calculated by multiplying these coefficients by the absolute difference in velocity head between
one cross section and the next downstream cross section (HEC, 2016). For gradual transitions
and supercritical flow, HEC (2016) suggests to use values of 0.01 for contraction and 0.03 for
expansion. When the change in river cross section is small, and the flow is subcritical,
coefficients of contraction and expansion are typically on the order of 0.1 and 0.3 respectively.
When the change in effective cross section area is abrupt such as at bridges, contraction and
expansion coefficients of 0.3 and 0.5 are often used. After a first simulation the flow in the
canal has been found to be a mixed-flow regime with supercritical flow prevailing. At the
junction (Carrefour commissariat), the flow is supercritical and there is a culvert just after the
junction. Then, values of 0.01 for contraction and 0.03 for expansion were used for the
computation of the losses.

Figure 3.33: Junction at Carrefour Commissariat

78
In HEC-RAS, for steady flow hydraulic computations, a junction can be modelled by either the
energy equation or the momentum equation. The energy equation does not consider the angle
of any tributary coming in or leaving the main stream, while the momentum equation does
(HEC, 2016). In most cases, the amount of energy loss due to the angle of the tributary flow is
not significant and using the energy equation to model the junction is more than adequate.
However, there are situations where the angle of the tributary can cause significant energy
losses. In these situations, it would be more appropriate to use the momentum approach.
Therefore, it is safer to use the momentum approach here as not only the different angles are
known but also serous hydraulic jumps take place at the junction during rain events.

c. Flow Data and Boundary Condition

Steady flow data are required for performing a steady water surface profile calculation. Steady
flow data consist of: flow regime; boundary conditions; and discharge information (peak flow
or flow data from a specific instance location).

The flow regime (subcritical, supercritical, or mixed flow regime) is specified on the steady
flow analysis window of the user interface. In cases where the flow will pass from subcritical
to supercritical, or supercritical to subcritical, the program should be run in a mixed flow
regime.

Here, the computations were run under steady flow conditions because what we are interested
in in this study is flood extent and depth and any damage that could be caused. However, in
case of retention ponds design at the upstream side, one should consider the unsteady flow
simulation because volume of floods becomes important.

Peak discharges were collected from HEC-HMS simulations and were entered in HEC-RAS
for the steady flow analysis. Floods events of May 11 and June 19, 2018 were used to calibrate
and validate indirectly the hydraulic model. And then discharges for various return periods were
run. Below is a figure showing how the boundary conditions were set. The average slope of the
canal is 0.008, and this is the requirement when the normal depth condition is selected. Here
the normal depth was selected at the downstream and critical depth at the upstream. The flow
change locations are known based on the watershed delineation did in ArcGIS (HEC-

79
GeoHMS). After the comparison of the extent reached by these two storms in HEC-RAS,
discharges for 5, 10, 25, 50 and 100 years return period were simulated. Based on the results

Figure 3.34: Steady Flow Conditions

from these simulations, inundated areas were delineated and then characterized so that policy
makers can find it useful for any future purpose.

when the critical depth condition is selected, there is nothing else to enter as parameter; the
program itself will compute the critical depth in the simulation (Abbas, Khaghan, & Mojaradi,
2016; Aksoy, Ozgur Kirca, Burgan, & Kellecioglu, 2016; Astite et al., 2015).

80
Chapter 4: FINDINGS AND DISCUSSION

The hydrologic model was used for runoff computation to be entered in the hydraulic model.
The hydrologic and hydraulic modelling results as well as the flooded area for various return
periods are presented in this chapter. The hydrologic model is calibrated indirectly by
comparing recent known floods marks to what HEC-RAS simulated as flow data were not
available. The validity of the hydrologic modelling results is evaluated, by analysing the models
loss and transform calculations as well as by discussing the appropriateness of the applied
design storms. The accuracy of the hydraulic modelling results is validated in the frame of a
sensitivity analysis. The development of the flood hazard zoning is based on the results of the
hydraulic modelling and a detailed analysis of the hydrogeomorphology of the study area.

4.1 Discussion of Results

4.1.1 Hydrologic Modelling Results

Two storms were run for indirect calibration and validation of the model. They were floods
events of 11 May 2018 and 18 to 19 June 2018. The June’s storm was intense and heavier by
comparing generated discharges and also damages.
Based on this first step’s results, various design storms for 5, 10, 25, 50 and 100 years return
period were run for specific analysis. This part is more significant in the industry and also for
decision makings. The figure above gives hydrographs generated at the outlet.

700

600

500
Discharge (m3/s)

400

300

200

100

0
00:00 04:48 09:36 14:24 19:12 00:00 04:48
-100
Time (hr)

Q5 (m3/s) Q10 (m3/s) Q25 (m3/s) Q50 (m3/s) Q100 (m3/s)

Figure 4.35: Flood hydrographs for BRP with various return periods

81
The hydrographs are abrupt and the peaks are high. This is the explanation of storms occurring
in the south of the country: they are short and intense most of the time (less than 6 hours). The
storms used for the simulations are of 3 hours. The table 4.8 represents the peaks from the
hydrographs simulated at the outlet of the watershed and the second table (table…) the peaks
at junction J37 for the hydrologic modelling and “carrefour commissariat” for the hydraulic
modelling.
This work is more interested in J37’s discharges because it is closed to the modelled canal’s
outlet (9 Kilo or roundabout of Rivièra 3) and is also the most critical place among flooded
areas in the catchment.
One can see how the discharges are high both at the outlet and J37. At the point of interest
(J37), the peak for 5-year storm is 254.2 m3/s. This is too high in comparison with the capacity
of the canal at the same point; a rectangular canal of 4×2.5 is able to accommodate only 60.88
𝐴
m3/s when using Manning-Strickler’s equation: 𝑄 = 𝐾𝑠 (𝑃)2/3 × 𝐴 × 𝑆 1/2

Where: 𝑄 = Discharge in m3/s

1
𝐾𝑠 = 𝑛, Manning-Strickler coefficient

𝐴 = Cross sectional area

𝑃 = Wetted perimeter
𝑆 = Canal bed slope (Here 𝑆 is 0.8%, average slope).
A value of 70 for Ks (concrete canal) gives n = 0.014.
This result shows that this canal can never convey those discharges even a 5-year storm while.
Most of the time drainage systems and sewers are designed for 10 years return period. The fact
that floods occur yearly, each rainy season seems to be understandable.

Table 4.7: Peak discharges derived from the figure 4.35

Return period 5 yr 10 yr 25 yr 50 yr 100 yr

Peak Discharge
320.8 410.6 492.8 549 640.7
m3/s)

Table 4.8: Peak discharges for various return periods at J37

Return period 5 yr 10 yr 25 yr 50 yr 100 yr

Peak Discharge
254.2 333.9 400.5 445.8 490.7
m3/s)

82
The 11 May 2018 storm event is presented below to show the seriousness of what is occurring
in this part of Abidjan.

Figure 4.36: Hydrograph generated at J37 on 11 May 2018

The peak discharge is 94 m3/s but caused flood that day whereas it is less than 1/5 of the 5-year
storm discharge. The situation is more serious than what is thought. Given, the small size of the
analysed watershed (37.47 km2), discharges of more than 600 m3/s seem very high. The shape
of the hydrographs however is typical for flash flood events as the result of short and intense
rainfalls on hilly watersheds. The high rainfall intensities that the study area experiences in
combination with the short time of concentration of the analysed watershed leads to the steep
and short rising limb of the flood hydrographs as well as a short overall duration of the flood.
Nevertheless, the presented hydrographs are also a function of the percentage of impervious
areas of the watershed as it is an urban catchment, the method used to calculate losses, design
storms and the models that were used in the transformation of rainfall into runoff. Therefore,
the effect of impervious areas estimation, the loss and transform calculations are analysed and
discussed in detail in the following sections.

83
4.1.1.1 Sensitivity Analysis

The CN may be calibrated since it is a calibrated parameter but the impervious is not. In this
study, the CN are more reliable than the impervious in terms of accuracy because the last one
was estimated based on a study done by a Consultant (CONCEPT) and the knowledge of the
study area by the modeller. This consultant worked on two watersheds in the municipality of
Cocody; those watersheds on both sides of the one of this study. Table 4.10 provides sub-basins
areas, the associated CN and estimated imperviousness.

a) Percentage of Impervious Areas

In first place, a decrease of the impervious by 10% was done to see how the peaks would be. A
decrease in impervious means an increase in the infiltration; this imply a decrease in runoff or
computed discharge.
Table 4.9: Sub-basins characteristics

Sub-basin W110 W120 W130 W140 W150 W160 W170 W180

Area 4.16 6.83 9.6 10 1.76 3.52 1.57 0.01

CN 84.6 85.02 85.81 82.67 86.65 85.05 86.24 86.32

impervious (%) 35.88 35.4 55.37 21.29 66.89 42.89 68.27 0.0

The imperviousness may have been overestimated or underestimated. But the overestimation is
more likely because of the values of peak discharges for return periods. Table … is the
repetition of the table above but the impervious has been increased by 10% this time.
Table 4.10: Sub-basins characteristics with increased imperviousness

Sub-basin W110 W120 W130 W140 W150 W160 W170 W180

Area 4.16 6.83 9.6 10 1.76 3.52 1.57 0.01

CN 84.6 85.02 85.81 82.67 86.65 85.05 86.24 86.32

impervious (%) 32.29 31.86 49.83 19.16 60.20 38.60 61.44 0.00

The results of the simulation for 5 year-storm are shown below:

84
One can see that the effect on the generated runoff of the percentage impervious is very low.
After a simple calculation of the decrease rate, it has been found to be about 1.19% for an
increase of 10% of the infiltration.
Table 4.11:Comparison of 5 years flows by 10% decrease in imperviousness

Element J37 Outlet

Peak before increase

(m3/s) 254.2 320.8

Peak after increase (m3/s) 251.2 316.4

This result is in support of a reliable hydrologic model set up for the BRP watershed.

b) The Lag Time

It is defined as the length of time between the centroid of precipitation mass and the peak flow
of the resulting hydrograph (HEC, 2016). It is the only required parameter while selecting the
SCS Unit Hydrograph. It is related to the time of concentration as discussed earlier. The time
of concentration of each sub watershed defines the steepness of the resulting hydrograph.
Another sensitivity analysis was performed in order to get the impression about the effect that
a possible over- or underestimation of the time of concentration has on the modelling results.
Figure 4.37 shows the resulting hydrographs at the outlet of BRP watershed for a 5-year storm.

initial Tlag (Qpeak=320.8 m3/s)

20% reduced Tlag (Qpeak=326 m3/s)
20% increased Tlag (Qpeak=304.2 m3/s)
400
350
discharge (m3/s)

300
250
200
150
100
50
0
00:00
00:50
01:40
02:30
03:20
04:10
05:00
05:50
06:40
07:30
08:20
09:10
10:00
10:50
11:40
12:30

Time

Figure 4.37: Comparison of flood hydrographs resulting from 20% in/de/crease of the lag
times

85
The sensitivity analysis was performed for a 5-year storm. Then; the 5-year 3-hour storm was
simulated with lag times that were 20% shorter and longer. And it is obvious from the figure
above that a decrease in lag time leads to an increase in the peak discharge; and an increase in
lag time yields a slight decrease in the peak discharge. Thus, a decrease of 20% in lag time
produces an increase of 1.6% in the peak discharge and an increase in the lag time of 20%
produces a decrease in the peak discharge of 5.5%. These error margins are acceptable. If the
error on the time of concentration due to an increase or decrease of slope because of
urbanization is lying between ±20%, then the actual peak discharge will be the simulated
discharge +1.6% or -5.5%.

4.1.1.2 Verification of the Loss Calculation

This section will discuss the computation of the losses for various return periods as well as for
the 11 May and 19 June 2018 flood event. The sub-basin W160 is chosen for this discussion.
Figure 4.29 and 4.30 show the total amount of precipitation, loss along with the total infiltration
losses that resulted in the presented hydrographs.

200 190.26
174.72
180 164.7 158.9
160 149.55
134.15135.73
140
120 111.76114.41
(mm)

100 91.39
80.68 77.68
80 59.87 57.09
60
40 25.55 25.17 24.76 23.98 23.02 20.81 20.57
20
0
100 yr 50 yr 25 yr 10 yr 5 yr 2 yr 1 yr
Statiscal return periods

Precipitation Loss Discharge

Figure 4.38: Total precipitation, losses and discharge for various return
periods

The total duration is 3 hours for all the return periods. The modeller was provided with the IDF
of Abidjan airport and the precipitation derived from it. The total duration of the raw data is
4hours but as HEC-HMS doesn’t offer this possibility it is the precipitation up to 3 hours that
was considered.

86
 350
301.98
300 274.78

250

200
(mm)

150

100 64.62
45.36
50 19.24 27.22

0
01-May-18 01-Jun-18
Date

Precipitation Loss Discharge

Figure 4.39: Total precipitation, losses and discharge for May and June 2018 storm events

These graphs show the total precipitation, losses and runoff for the storms used in this work.
Total losses are very low compared to the total precipitation. These results make sense since
the watershed is highly urbanized. However, when considering the 11 May 2018 storm (about
3 hours duration), the loss is high and for the same duration and higher precipitation depths
(100, 50, 25, 10, 5, 2, 1-year return period) the loss is not proportional. The reason for this is
that the CN model does not directly account for the intensity and duration of the rainfall (see
Chapter 2.3). The absolute loss of a certain event is only a function of the Curve Number and
the absolute rainfall depth regardless of the intensity distribution. Nevertheless, a time
component is introduced in the model when it is applied for the estimation of runoff from
successive intervals in a storm as done in this study. Hereby, HEC-HMS first calculated the
accumulated discharge 𝑄 from the accumulated precipitation 𝑃 of each time step and then
derived the runoff for each time step as the difference between the accumulated 𝑄 at the
beginning and end of each time interval.

4.1.2 Hydraulic Modelling Results

Based on the peak discharges resulting from the rainfall runoff modelling seen in the previous
chapter, steady flow simulations were first performed for 11 May 2018 and 19 June 2018 flood
events and later for return periods of 5, 10, 25, 50 and 100 years. This section provides details

87
on the hydraulic modelling results in the form of cross sections plots, profiles plots 3D view,
summary tables and inundated maps. Figure 4.32 is a 3D view of the modelled canal.
Master_Thesis Plan: Plan: May and June 2018 flood events 16/08/2018
Legend

WS PF 1_jun2018
2.063
1.9855* Ground
1.9079* Bank Sta
1.8202
1.7535* Ground
1.74354
1.6303*
1.5531*
1.4759*
1.3987*
1.3022*
1.2250*
0.99874*
0.99621*
0.8500
0.84036*
0.7500
0.3206 0.72120*
0.26060* 0.57839*
0.20060* 0.54936*
0.15360* 0.3880
0.09360* 0.38400
0.03360* 0.3387
0.2989
0.06345*0.20345*
0.00848*
0
0.00623

Figure 4.40: Perspective plot of the modelled canal for May and June storms
The modelled canal is made of two canal that come together at the junction “Commissariat”.
They are “Canal Rue Ministre” and canal C2. The canal “Rue Ministre” is the longest and starts
at river station 2.063 when the C2 starts at river station 0.320. The junction is at the downstream
part. These results show that this place was the most flooded area following the May and June
storms.

These canals are existing and then its operation is analysed by the means of simulations for
flows of 5, 10, 25, 50 and 100 years return period. The level reached by water in the canals was
the main simulated parameter; it was compared to the Banks’ level of the canals to detect the
places of possible banks overflow. Profiles given by HEC-RAS allowed to visualize these
places. A special attention was paid to the simulated depths of water at the level of the hydraulic
structures (culverts) located across canals, because on one hand, those structures generally
cause a raising of the water surface and, in addition, roads submersion may occur. And also,
cross sections plot from HEC-RAS allowed to appreciate any submersion and to estimate the

88
value of this submersion, for the five return periods. Based on the results from the simulations,
Inundation mapping was performed using firstly HEC-GeoRAS and secondly RAS Mapper

Master_Thesis Plan: Plan: Design Storm 16/08/2018

Legend

WS 100 yr
Ground
Bank Sta
Ground

0.58419*
0.57258*
0.56097*
0.54936*
0.3900
0.3880
0.38622*

0.38400

0.3447
0.3387
0.33373
0.31873*

0.2919
0.20345* 0.25345*
0.04205
0.10345* 0.15345*
0.05345*

Figure 4.41: 3D view of the 100-year storm profile of canal “Rue Ministre” upper reach
within HEC-RAS since geospatial processing is now available directly in HEC-RAS newer
version. The inundation is discussed in details in the next section.
HEC-RAS has the capability to compute the velocity in the canal and on the floodplain, the
surface area, the volume of water and the water depth in graphics and tables before any
visualization in RAS Mapper or HEC-GeoRAS.

The flows modelled in this study are high, the lowest being 172.3 m3/s for the 2-year storm.
The parameters influencing the flow are the Manning’s n-values, the cross sections, contraction
and expansion coefficients, while running an unsteady flow simulation (The Manning’s n may
do). But they are not in the case of a steady flow computation. In addition, when simulated
flows are high, the model is less sensitive to Manning coefficients variations, Casas et al (2006).
For this reason, no sensitivity analysis was performed.

89
On the figure below, we have the simulation for 5 and 10-year storm and one can see how water
depths are important. The river bank stations are totally submerged. Although, in most
developing countries, drainage networks are designed with 10-year storm. The question one is
Master_Thesis Plan: Plan: Design Storm 16/08/2018
The most upstream XS
.018 .014 .016
51 Legend

EG 10 yr
EG 5 yr
50
WS 10 yr
Crit 10 yr
WS 5 yr
49
Elevation (m)

Crit 5 yr
Ground
Bank Sta
48

47

46
-4 -2 0 2 4 6 8 10 12
Station (m)

Figure 4.42: The most upstream cross section for 5 and 10 year-storm
persuaded to ask is how can engineers design a canal which capacity is not only under the
standard of 10 years but very far from it (under 2 years)? During the phase of literature review,
it was discovered in reports of consultants (BNETD, 2012; TERRABO, 2012) that it was
planned the construction of retention ponds over the watershed. Unfortunately, they were not.
These ponds should have to retain water during storm events and release it regularly according
to the convey capacity of the downstream canals (“Rue Ministre”, C1, C2). Not only they have
not been constructed their sites have been transformed in residential areas where high standing
houses have been built (neighbourhood of Saint Viateur). Based on this fact, it is almost
impossible today to construct those ponds. That’s why BNETD proposed a widening of the
existing canal (“Rue Ministre”) from a distance before river station 2.063 to the main outlet
(Ebrié lagoon). To widen it will not be an easy task since the area is highly urbanized; then now
it is more a question of cost/benefit analysis to choose the best option.

A discharge of 108 m3/s is the capacity of the widened canal at the junction (“commissariat du
18e arrondissement”). Even this one also is still a problem because the peak discharge of the 2-
90
year storm is 172.3 m3/s as seen in the section above. Unless they planned a flow diversion to
another canal or stream (which will be difficult) or construction of small retention ponds on
other sites to retain some amount.
The figure below is the profile of the main reach (Canal Rue Ministre upper reach).

Master_Thesis Plan: Plan: Design Storm 16/08/2018

Rue Ministre Upper Reach
55 Le ge nd

EG 10 yr
50
EG 5 yr
WS 5 yr
45
Crit 5 yr
Elevation (m)

Crit 10 yr
40
WS 10 yr
Ground
35

30

25

20
0 500 1000 1500 2000 2500
SA Total (1000 m2), SA Left (1000 m2), SA Chan (1000 m2), SA Right (1000 m2)

Main Channel Distance (m)

Figure 4.43: Water surface profile in the upper reach for return period of 5 and 10 years

Master_Thesis Plan: Plan: Design Storm 16/08/2018

Rue Ministre Upper Reach
70 Legend

60 SA Total 5 yr
SA Total 10 yr
50
SA Left 5 yr

40 SA Left 10 yr
SA Right 5 yr
30
SA Right 10 yr

20 SA Chan 10 yr
SA Chan 5 yr
10

0
0 500 1000 1500 2000 2500
Main Channel Distance (m)
Figure 4.44: Total flood extent along the upper reach for 10 years

91
The total surface area (flood extent) is about 62,000 m2 = 6.2 Ha. The 5 and the 10 years have
the total extent because of the defined flow path by the modeller. Ones, water reaches the
defined limits, it can only go up so that the water depth will be higher. Water depth plot, flow
along the canal and floodplains, Froude number variations, velocities are also available (see
appendix C).

4.1.3 Inundation Mapping

From the hydraulic modelling results, one can view directly geospatial data in RAS Mapper or
export to HEC-GeoRAS for visualization. Both methods were used in this study. In RAS
Mapper, the results are automatically available and they can be visualized by opening RAS
Mapper while in HEC-GeoRAS export to GIS is necessary and also some steps before the flood
plain be delineated. The figure below is the water depth map for the 100-year storm from RAS
Mapper.

Figure 4.45: Water depth map for the 100-year storm

The visualization gives a very good impression of the flood event. One can see easily the areas
touched by water also view the water depth and velocity. Besides, the flood extent is estimated
numerically. For the 100-year, the inundated area is 9.35 Ha and water depth varying from 0.3

92
to 15 m above the ground. The depth is high and not real but it is because water is maintained
in the flow path as discussed in the previous section.

4.2 Comparison of this study’s results with other did on Cocody

Due to lack of observed data, the hydrologic model was not calibrated directly. Given this fact,
it is good to have a look at similar works carried out on Cocody and more, in watersheds sharing
borders with the one of this study.

The government of Côte D’Ivoire asked two foreign consultants in 2015, Roche and Concept,
to analyse the existing drainage system in 5 municipalities within Abidjan and propose solution
for a better drainage of stormwater. Thus, they used HEC-HMS for rainfall runoff modelling
and HEC-RAS for hydraulic modelling. It was the first time these softwares are used in the
country. In the municipality of Cocody, they worked on the watershed of Angré located on the
right side of BRP watershed and also on M’Pouto watershed on the left when looking the south
direction, but they didn’t work on the BRP watershed because BNETD and TERRABO, two
Ivorian consulting engineers, studied it using other methods.

They got a watershed Angré 3) area of 598.4 ha = 5.98 km2 and a peak discharge of 862.5 m3/s
for 100-year storm while in the case of this study, the watershed area is 37.47 km 2 and a peak
discharge of 604.7 m3/s at the outlet for 100-year storm. The BRP watershed is 6.26 times the
Angré 3. However, their peak discharge for a 100-year storm is 1.43 times the one of BRP. It
is true that the area is not the only parameter accounting for the runoff generation, however one
can comfortably believe that the discharges computed by the hydrologic model are in acceptable
range of values. They also calibrated their model indirectly by comparing water marks for
Table 4.12: Characteristics of Angré 3 watershed (CONCEPT,
2015)

93
specific flood events (June 2015) to what was simulated by HEC-RAS and used the initial and
constant loss for the loss method in HEC-HMS while for this work it is the SCS curve number
method that was used. The table below contains the characteristics of the Angré 3 watershed.
To obtain the total area of the watershed one should sum the areas of the sub-watersheds.
The next table gives the peak discharges calculated with HEC-HMS.

Table 4.13: Peak discharges computed at the Angré watershed (CONCEPT, 2015)

Table 4.14: Peak discharges for various return periods at the BRP watershed

2 yrs 5 yrs 10 yrs 25 yrs 50 yrs 100 yrs

Subwatershed/element Q Q Q Q Q % Q %
(m3/s) (m3/s) (m3/s) (m3/s) (m3/s) Q10yrs (m3/s) Q10yrs
W110 50.0 76.4 95.6 114.5 128.1 1.34 142.1 1.49
W120 66.7 97.6 124.7 148.3 164.3 1.32 180.4 1.45
W130 90.8 126.1 163 191.5 210.7 1.29 230.1 1.41
W140 72.1 106.5 144.6 174.2 194.3 1.34 214.3 1.48
W150 28.4 40.6 49.5 58.2 64.3 1.30 70.7 1.43
W160 41.3 60.1 75.3 89.3 99 1.31 108.8 1.44
W170 27.1 38.1 46.2 54.1 60 1.30 66.6 1.44
W180 0.2 0.4 0.5 0.6 0.7 1.40 0.8 1.60
J37 172.3 254.2 333.9 400.5 445.8 1.34 490.7 1.47
Outlet 215.3 320.8 410.6 492.8 549 1.34 604.7 1.47

The ratio of the Q100yrs by Q10yrs is about 1.28% and Q50yrs by Q10yrs is 1.20% while for this study
the first ratio gives 1.47% and the second 1.33%. These results are not far from one another. As
said before the results of this work seem acceptable.

94
4.3 Flood Preparedness Strategy

Considering the 10-year peak discharge from table 4.15 above, this discharge is very important
(410.6 m3/s at the outlet or 254.2 m3/s at J37) while the capacity of the canal at J37 is 60.88m3/s.
Then this canal can never convey this quantity without overflowing. As people cannot move
from the sites the only thing that can be done is to construct the retention ponds planned initially.
It will be costly for sure but it is more efficient technically. Water should be stopped upstream
before getting in the canal, so the only to do it is to construct storage areas. Then if those sites
are occupied now, probably a tunnel (underground culvert) may be helpful to support the
existing canal since there is no more space for any widening. This solution requires specific
skills and technics.

These approaches are long term solution. Because the amount of precipitation cannot be
reduced, on the ground what should be done right now is just a diversion canal to convey certain
amount somewhere else to reduce the quantity to be carried by the existing canal. By doing so,
lives will be saved downstream (World Meteorological Organization, 2011).

95
Chapter 5: GENERAL CONCLUSION AND RECOMMENDATION

5.1 General Conclusion

The main objective of this work was to estimate flood extent and create a flood map for the
Bonoumin-Rivièra Palmeraie watershed. With the topographic data got from ONAD through
reports and field visit and measurements, and also the results from the hydrologic modelling
(HEC-HMS), it was possible to create inundation map of the study area. Thus, the main goal
has been fulfilled satisfactory. Peak discharges from the hydrographs generated by HEC-HMS
were used for the hydraulic simulations since this work was interested in the canal bank
overflow just to estimate the flood extent. If it was to design a retention pond upstream, an
unsteady flow simulation would have been performed because the volume of flood is an
important parameter in this case. A sensitivity analysis was carried out for the hydrologic model
and shew that the results are representative then reliable as the possible deviation is very small.
From the results of this study, one now has a clear idea of where and how flood occurs in the
watershed ((BCET), 1984; Heimhuber, 2013; C. G. Patel & Gundaliya, 2016).

However, many issues were encountered during the study. The first issue is the time spent to
acquire data. Task that can be achieved in two or three days can take weeks even a month. In
addition, the acquisition of topographic data was uneasy since it was done without the support
of the internship host company. Some topographic information was collected from an
AutoCAD file provided by ONAD (TERRABO, 2012). Unfortunately, the canal centreline
drawing was not exactly where it should be, but this was quickly solved by gathering other
sources of data and also fields observation and measurement. The use of Google satellite images
was useful when dealing with the issue of topographic data. One other challenge was the
absence of observed flow data for the calibration of the hydrologic model. Here again a flow
measurement campaign was initiated but the data were not sufficient to calibrate and validate
the model. Thereby, an indirect calibration by comparing flood marks to the simulated water
level was done and also a comparison with other studies to measure the reliability of the work.

Since March 2018, a project has been launched to cope with flood issue in BRP watershed.
They want to install equipment all over the watershed for recording discharges and precipitation
in order to require data of quality. This is a very good initiative to address flood issue.

96
5.2 Recommendation

As seen in chapter 1, flood has been occurring frequently in the watershed of BRP causing a lot
of damages including losses of lives. This year again according to the government, about 10
people died at Rivièra Palmeraie and Rivièra 3 due to floods. Thereby, urgent actions should
be taken in order to stop suffering of people living in that area.

In an engineering point of view, the widening of the canal proposed by BNETD (8.00×1.60)
seem very difficult because the area is highly urbanized, then there is no space for this type of
work unless they destroy houses. Therefore, the Government (ONAD) should designate a
consultant to carefully study the issue through hydrologic and hydraulic modelling as the one
of this study for an optimum solution of course with reliable and sufficient data as financial
support will be provided (Colcanap & Dufour, 1982; Peter M. Bach et al., 2013; Rao & Ramana,
2015)(Colcanap & Dufour, 1982; Peter M. Bach et al., 2013; Rao & Ramana, 2015). Surely, an
unsteady flow simulation should be performed in order to design the capacity of the retention
pond again (analysis of what was done by the BNETD), (Hydrologic Engineering Center
(HEC), 2016b; K. P. Patel, 2009). Then, this study should be a basis for future work to be
carried out since the tools used are up to date and are widely used with satisfactory results.

97
References:

(BCET), B. C. d’Etudes T. du M. des T. P. (1984). Thalweg du Danga, Note de calcul.

Abbas, A., Khaghan, M., & Mojaradi, B. (2016). The Integrate of HEC-HMS and HEC-
RAS Models in GIS Integration Models to Simulate Flood ( Case study : the area of
Karaj ). Journal Current World Environment, 11(1), 1–5.
https://doi.org/http://dx.doi.org/10.12944/CWE.11.Special-Issue1.01

Ackerman, C. T. (2009). HEC-GeoRAS GIS Tools for Support of HEC-RAS using ArcGIS
User ’ s Manual.

Ackerman, C. T., Jensen, M. R., & Brunner, G. W. (2010). Geospatial Capabilities of

HEC-RAS for Model Development and Mapping. In 2nd Joint Federal Interagency
Conference (pp. 1–6).

Adetunji, M., & Oyeleye, O. (2013). Evaluation of the Causes and Effects of Flood in Apete,
Ido Local Government Area, Oyo State, Nigeria. Civil and Engineering Research, 3, 8.

Aghui, N., & Biémi, J. (1984). Geology and Hydrogeology of Groundwater in the region of
Abidjan and Contamination risks (C No. 20).

Aksoy, H., Ozgur Kirca, V. S., Burgan, H. I., & Kellecioglu, D. (2016). Hydrological and
hydraulic models for determination of flood-prone and flood inundation areas. IAHS-
AISH Proceedings and Reports, 373, 137–141. https://doi.org/10.5194/piahs-373-137-
2016

Alaghmand, S., Abdullah, R. Bin, Abustan, I., & Eslamian, S. (2012). Comparison
between capabilities of HEC-RAS and MIKE11 hydraulic models in river flood risk
modelling (a case study of Sungai Kayu Ara River basin, Malaysia). International
Journal of Hydrology Science and Technology, 2(3), 270.
https://doi.org/10.1504/IJHST.2012.049187

Asif, M., & Urarande, S. K. (2016). Calculating the parameters of Muskingum Methods for
a rectangular channel.pdf. International Journal of Recent Advances in Engineering &
Technology, 4(3), 5.

Astite, S. W., Medjerab, A., Belabid, N.-E., Mahmouhi, E., Wartiti, E., & Kemmou, S.
(2015). Cartography of flood hazard by overflowing rivers using hydraulic modeling and
geographic information system: Oued El Harrach case (North of Algeria). 2015, 13.
https://doi.org/https://doi.org/10.4995/raet.2015.3985

98
Baldassarre, G. Di, Viglione, A., Carr, G., Kuil, L., Salinas, J. L., & Bl, G. (2013). Socio-
hydrology : conceptualising human-flood interactions, 3295–3303.
https://doi.org/10.5194/hess-17-3295-2013

Beighley, R. E., Dunne, T., & Melack, J. M. (2005). Understanding and modeling basin
hydrol...rpreting the hydrogeological signature.pdf. Journal of Hydrological Processes,
19, 1333–1353.

BNETD. (2012). Aménagement du bassin versant de Bonoumin-Rivièra Golf, Etudes

techniques; Rapport d’APD. Abidjan.

Bouvier, C. (1989). Contribution à l’étude du ruissellement...écoulements sur petits bassins

urbains.pdf.

Bouvier, C. (2004). De la pluie à l’inondation : contribution à la compréhension et à la

prévision des événements extrêmes sur petits bassins tropicaux et méditerranéens.
Sciences-New York, 1–93.

Bouvier, C., & Desbordes, M. (1990). Un modèle de ruissellement urbain pour les villes d ’
Afrique de l ’ Ouest. Hydrol Continent.

BRLi. (2015). Projet de gestion intégrée du bassin versant de Gourou, synthèse du rapport
de mission A,. Abidjan.

Burrough, P. A., & McDonnell, R. A. (1998). Principles of Geographical Information

Systems.

Chen, X., Tian, C., Meng, X., Xu, Q., Cui, G., Zhang, Q., & Xiang, L. (2015). Analyzing
the effect of urbanization on flood characteristics at catchment levels. IAHS-AISH
Proceedings and Reports. https://doi.org/10.5194/piahs-370-33-2015

Chin, C. (2006). Classroom interaction in ESL classroom .pdf. International Journal of

Science Education, 1315–1346.

Choudhari, K., Panigrahi, B., & Paul, J. C. (2014). Simulation of rainfall-runoff process
using HEC-HMS model for Balijore. International Journal of Geomatics and
Geosciences, 5(2), 253–265. https://doi.org/Code : EIJGGS5021

Chu, X., & Steinman, A. (2009). Event and Continuous Hydrologic Modeling with HEC-
HMS. Journal of Irrigation and Drainage Engineering, 135(1), 119–124.
https://doi.org/10.1061/(ASCE)0733-9437(2009)135:1(119)

Claman, D. (2014). Comparison between 1D and 2D modeling. In 52nd ASCE

99
 Environmental Water Resource Conference (p. 65).

Colcanap, M., & Dufour, P. (1982). L’assainissement de la ville d’ab.idjan (Vol. 16).

CONCEPT, & ICI. (2012). Etudes hydrauliques et géotechniques de l’aménagement du

carrefour de l’Indénié, Rapport d’APD.

CONCEPT, & ROCHE. (2016). Aménagement du réseau de drainage des bassins versants
d’Abidjan Est et Abidjan Ouest, étude technique et étude d’impact environnemental et
social; Rapport d’APD. Abidjan.

Cunderlik, J., & Simonovic, S. P. (2004). Calibration, Verification and Sensitivi...alysis of

the HEC-HMS Hydrologic Model.pdf. ONTARIO.

Cunge, J. A., Holly, F. M., & Verwey, A. (1980). Practical aspects of computational river
hydraulics. In Water Resources Engineering.

Danumah, J. H. (2016). Assessing Urban Flood Risks Under Changing Climate and Land
Use in Abidjan District, South Côte D’Ivoire. Kwame Nkrumah University of Science
and Technology.

Derdour, A., Bouanani, A., & Babahamed, K. (2017). Hydrological Modelling in Semi-
Arid Region Using HEC-HMS model: Case Study in Ain Sefra Watershed, Ksour
Mountains. Journal of Fundamental and Applied Sciences, 2(6), 419–424.
https://doi.org/10.5185/amlett.2011.2228

Dongo, K., Cissé, G., & Biémi, J. (2008). Étude Des Inondations Par Débordement De
Réseau Dans Les Milieux Défavorisés D’Abidjan; Côte D’Ivoire. 13th IWRA World
Water Congress, Global Changes and Water Reseources, 1(i), 1–12.

Fleming, M., & Scharffenberg, W. (2000). Hydrologic Modeling System (HEC-HMS): New
Features for Urban Hydrology. https://doi.org/CDP-74B

Gautam, M. R., Acharya, K., & Stone, M. (2010). Best Management Practices for
Stormwater Management in the Desert Southwest.pdf. Journal of Contemporary Water
Research & Education, (146), 39–49.

Hauhouot, C. (2008). Analyse du risque pluvial dans les quartiers précaires d’Abidjan.pdf.
Geo-Eco-Trop, 32, 75–82.

Hawkins, R. H. (1993). Asymptotic Determination of Runoff Curve Numbers from Data.

Journal of Irrigation and Drainage Engineering. https://doi.org/10.1061/(ASCE)0733-
9437(1993)119:2(334)

100
Heimhuber, V. (2013). GIS Based Flood Modeling as Part of an Integrated Development
Strategy for Informal Settlements.

Hjelmfelt, A. T. (1991). Investigation of Curve Number Procedure. Journal of Hydraulic

Engineering. https://doi.org/10.1061/(ASCE)0733-9429(1991)117:6(725)

Hydrologic Engineering Center (HEC). (2016a). HEC-HMS User’s Manual, Version

4.2.pdf.

Hydrologic Engineering Center (HEC). (2016b). HEC-RAS River Analysis System User ’ s
Manual.

Institut National de la Statistique (INS). (2014). Recensement Général de la Population et

de l’Habitat. Abidjan.

IPCC. (2007). Climate Change 2007: Impacts, Adaptation and Vulnerability.

Jones, J., Mike, P., & Young, J. (1997). The VDGIF Project on the Distributio and
Abundance of Freshwater Mussels, Snails and Crayfish of the Virginia Section of the
New River_Jones et al 1998.pdf.

Jourda, J. P., Saley, M. B., Djagoua, E. V., Kouamé, J. K., Biémi, J., & Razack, M.
(2006). Utilisation des données ETM+ de LandSat et d’un SIG pour l’évaluation du
potentiel en eau souterraine dans le milieu fissuré précambrien de la région de
Korhogo.pdf, 339–357.

Jourda, J. P., Saley, M. B., Kouamé, J. K., & Kouame, K. F. (2003). Gestion et protection
des eaux souterraines: Apport d’un SIG à la réalisation de la carte de vulnérabilité du CT
au niveau de l’agglomération d’Abidjan. In Conférence francophone ESRI “SIG 2003.”

Kablan, M. K. A. (2014). The Dynamics of Urban Land Cover and Resident’s Vulnerability
to Climate Change: A Case Study of Cocody District, Abidjan, Côte D’Ivoire.

Kamagaté, B., Effebi, R. K., Albert, T., Bi, G., Goné, D. L., & Noufé, D. (2016). Aghien
lagoon : a sustainable resource of fresh water for the city of Abidjan ( Ivory Coast )?
Description of the project and preliminary results, 18, 8548.

Kamal, N. (2015). A Project Proposal on Creating SCS curve number of Painesville

Watershed Using ArcGIS.

Kangah, A., & Della, A. A. (2015). Détermination des zones à risque d ’ inondation à partir
du modèle numérique de terrain ( MNT ) et du système d ’ information géographique (
SIG ) : Cas du bassin-versant de Bonoumin-Palmeraie ( commune de Cocody , Côte d ’

101
 ivoire ). Geo-Eco-Trop, 39, 297–308.

Ketcha, M., Kablan, A., Dongo, K., & Coulibaly, M. (2017). Assessment of Social
Vulnerability to Flood in Urban Côte D’Ivoire Using the MOVE Framework. Water, 1–
19. https://doi.org/10.3390/w9040292

Khattak, M. S., Anwar, F., Saeed, T. U., Sharif, M., Sheraz, K., & Ahmed, A. (2016).
Floodplain Mapping Using HEC-RAS and ArcGIS: A Case Study of Kabul River.
Arabian Journal for Science and Engineering, 41(4), 1375–1390.
https://doi.org/10.1007/s13369-015-1915-3

Khattatı, M., Serroukh, M., Rafık, I., Mesmoudı, H., Brırhet, H., Bouslıhım, Y., & Hara,
F. (2016). Hydrological Modeling of Aguibat Ezziar Watershed (Morocco), Comparative
Study of Two Different Hydrological Models. Journal of Geographic Information
System, 08(01), 50–56. https://doi.org/10.4236/jgis.2016.81005

Kouamé, I. K., Douagui, A. G., Koffi, K., Dibi, B., Kouassi, L. K., & Savane, I. (2013).
Modeling of Quaternary Groundwater Pollution Risk by GIS and Multicriteria Analysis
in the Southern Part of Abidjan District (Côte d’Ivoire). Journal of Environment
Protection, 4, 1213–1223.

Le front intertropical en Afrique occidentale.pdf. (1989). Montpellier.

Liew, M. W. Van. (2001). Sensitivity of Hydrologic Response of an Experimental Watershed

to Changes in Annual Precipitation Amounts. Soil and Water.

Maidment, D. R., & Mays, L. W. (1988). APPLIED HYDROLOGY. (B. . Clark & J.
Morriss, Eds.) (McGRAW-HIL). New York.

Michas, S. N., & Lazaridis, L. S. (2005). Hydrologic Modeling for the Determination of
Design Discharges in Ungauged Basins. Global NEST Journal, 7(3), 296–305.

Moglen, G. E., & Beighley, R. E. (2002). Spatially Explicit Hydrologic Modeling of Land
Use Change. Journal of the American Water Resources Association.
https://doi.org/10.1111/j.1752-1688.2002.tb01548.x

Mukherjee, D. (2016). Effect of Urbanization on Flood - a Review With Recent Flood in

Chennai (India). International Journal of Engineering Sciences & Research Technology.
https://doi.org/10.5281/zenodo.57002

Muller, H. G., & Rungoe, M. (1995). Integrating Floodplain Management and Numerical
Modelling, Using ArcVIEW.

102
Nasiri, H., & Shahmohammadi-kalalagh, S. (2013). Flood Vulnerability Index as a
Knowledge Base for Flood Risk Assessment in Urban Area. Journal of Novel Applied
Sciences.

National Research Council (NRC). (2008). Urban Stormwater Management in the United
States. Washington, D.C.

Ne, N. W., & Sw, S. E. (2004). Principles of Geographic Information Systems.

OCHA. (2014). OCHA Annual Report 2014. Annual Report 2014 Royal Consun.

Oleyiblo, J. O., & Li, Z. (2010). Application of HEC-HMS for flood forecasting in Misai and
Wan’an catchments in China. Water Science and Engineering, 3(1), 14–22.
https://doi.org/http://dx.doi.org/10.3882/j.issn.1674-2370.2010.01.002

Oriental Consultants Co. Ltd, HYDRO-CO, ArcIngénierie, Medi, & Roche. (2014).
Schéma Directeur d’urbanisme du Grand Abidjan, Rapport final, Volume II. Abidjan.

Patel, C. G., & Gundaliya, P. J. (2016). Floodplain Delineation Using HECRAS Model —
A Case Study of Surat City. Open Journal of Modern Hydrology, 6(January), 34–42.

Patel, K. P. (2009). WATERSHED MODELING USING HEC-RAS, HEC-HMS, AND GIS

MODELS – A CASE STUDY OF THE WRECK POND BROOK WATERSHED IN
MONMOUTH COUNTY, NEW JERSEY.

Peter M. Bach et al. (2013). A critical review of integrated urban w... modelling - Urban
drainage and beyond.pdf. Environmental Modelling & Software, 54, 88–107.

Ponce, V. M., & Hawkins, R. H. (1996). Runoff Curve Number: Has It Reached Maturity?
Journal of Hydrologic Engineering. https://doi.org/10.1061/(ASCE)1084-
0699(1996)1:1(11)

Rao, Y. R. S., & Ramana, R. V. (2015). STORM WATER FLOOD MODELING IN

URBAN AREAS, 2319–2322.

Scharffenberg, W., Ely, P., Daly, S., Fleming, M., & Pak, J. (2010). Hydrologic Modeling
System (Hec-Hms): Physically-Based Simulation Components. 2nd Joint Federal
Interagency Conference, 8.

Sighmonou, D. (1986). Assainissement pluvial en zone urbaine en Afrique tropicale, cas de

Yopougon (Côte D’Ivoire), Thèse de Doctorat de 3ème cycle. Université des Sciences,
Montpellier.

Silva, F. V, Bonumá, N. B., & Uda, P. K. (2010). Flood Mapping in Urban Area Using Hec-
103
 Ras Model Supported By Gis.

Singh, V. P., & Frevert, D. K. (2002). Mathematical modeling of watershed hydrology. J

Hydrol. Eng.

Skhakhfa, I. D., & Ouerdachi, L. (2016). Hydrological modelling of Wadi ressoul

watershed, Algeria, by HEC-HMS model. Journal of Water and Land Development,
31(1), 139–147. https://doi.org/10.1515/jwld-2016-0045

Tastet, J.-P., Martin, L., & Aka, K. (1979). Géologie et environnements sédimentaire...la
marge continentale de Côte d’Ivoire.pdf.

Tate, E., & Maidment, D. R. (1999). Floodplain Mapping Using HEC-RAS and ArcView
GIS.

TERRABO. (2012). Etudes techniques de protection des champs captants dans le district
d’Abidjan: cas des champs captants de zone Niangon-Nord, zone Nord Adjamé et Nord
Rivièra; Rapport d’APD. Abidjan.

USDA-NRCS. (1986). Urban Hydrology for Small Watersheds - TR-55.

USDA-NRCS. (2010). National Engineering Handbook Chapter 15, Time of Concentration,

1–15.

VARLET, F. (1978). Le régime de la lagune Ebrié. Traveaux et Documents de l’ORSTOM

N°83.

World Meteorological Organization. (2011). Manual on flood forecasting and warning.

Yao, A. B., Kouamé, K. I., Kouassi, K. A., Koffi, K., & Goula. (2015). Estimation de la
recharge d’une nappe côtière en zone tropicale humide_cas de la nappe du Continental
Terminal d’Abidjan.pdf. International Journal of Innovation and Applied Studies, 12,
888–898.

Yener, M. K., Sorman, a U., & Gezgin, T. (2007). Modeling studies with HEC-HMS and
runoff scenarious in Yuvacik Basin, Turkiye. International Congress on River Basin
Management, 621–634.

Zainalfikry, M. K., & Ghani, A. A. (2018). HEC-RAS One-Dimensional Hydrodynamic

Modelling for Recent Major Flood Events in Pahang River Corporation Engineers
Hydraulic Engineering Centre to, 1–10.

104
(i) APPENDIX A: Results of pre-processing in ArcGIS

105
(ii) APPENDIX B: Hydrologic Modelling Results

106
107
 (iii) APPENDIX B: Hydraulic Modelling Results

Master_Thesis Plan: Plan: Design Storm 17/08/2018

The most upstream XS
.018 .014 .016
52 Le ge nd

EG 100 yr
EG 50 yr
51
EG 25 yr
EG 10 yr
WS 100 yr
50
Crit 100 yr
Elevation (m)

EG 5 yr
WS 50 yr
49
Crit 50 yr
WS 25 yr
Crit 25 yr
48
WS 10 yr
Crit 10 yr
WS 5 yr
47 Crit 5 yr
Ground
Bank Sta
46
-4 -2 0 2 4 6 8 10 12
Station (m)

Master_Thesis Plan: Plan: Design Storm 17/08/2018

Canal C2 Tributary
31 Legend

EG 100 yr

EG 50 yr

30 EG 25 yr

EG 10 yr
Crit 25 yr

29 EG 5 yr
Crit 10 yr
WS 100 yr
Crit 100 yr
Elevation (m)

28 Crit 5 yr
WS 25 yr
WS 10 yr
27 WS 50 yr
Crit 50 yr
WS 5 yr

26 Ground

25

24

23
0 50 100 150 200 250 300 350
Main Channel Distance (m)

108
 Master_Thesis Plan: Plan: Design Storm 17/08/2018
Canal C2 Tributary
10
SA Total (1000 m2), SA Left (1000 m2), SA Chan (1000 m2), SA Right (1000 m2) Vel Left (m/s), Vel Chnl (m/s), Vel Right (m/s)

Legend

Vel Chnl 100 yr

Vel Chnl 50 yr
8 Vel Chnl 25 yr
Vel Right 100 yr
Vel Right 50 yr

6 Vel Chnl 10 yr
Vel Right 25 yr

Vel Left 100 yr

Vel Left 50 yr
4
Vel Chnl 5 yr
Vel Right 10 yr
Vel Left 25 yr
2
Vel Right 5 yr
Vel Left 10 yr
Vel Left 5 yr
0
0 50 100 150 200 250 300 350
Main Channel Distance (m)

Master_Thesis Plan: Plan: Design Storm 17/08/2018

Canal C2 Tributary
18 Legend

SA T ot al 5 yr

16 SA T ot al 10 yr

SA T ot al 25 yr

SA T ot al 50 yr
14 SA T ot al 100 yr
SA Right 10 yr
12 SA Right 25 yr

SA Right 5 yr

10 SA Right 50 yr

SA Right 100 yr

SA Left 10 yr
8 SA Left 50 yr

SA Left 5 yr
6 SA Left 100 yr

SA Left 25 yr

4 SA Chan 5 yr
SA Chan 25 yr

SA Chan 10 yr
2 SA Chan 100 yr

SA Chan 50 yr
0
0 50 100 150 200 250 300 350
Main Channel Distance (m)

109