FIDIC-GAMA 2018 CONFERENCE

ALIGNING THE CONSULTING ENGINEER’S ROLE TO THE VISION

OF AFRICA’S INFRASTRUCTURE DEVELOPMENT

Sub-theme 1: Contribution of African engineering in the realization

of intelligent and climate change resilient infrastructures (buildings,
water, energy, transport, mines)

Communication: Modeling hydrological phenomena in urban areas:

which models to choose and how to cope with data scarcity?
Feedback from a Consulting Engineer

Author : Imed Eddine Nouri – G.M. of CONCEPT Group (Tunisia)

[email protected]

1. In Africa: Urban population is and benefiting from conventional urban

growing very fast, but urban infrastructures. Large parts of urban
infrastructure does not follow! areas have particularly poor rainwater
sanitation and are exposed to periodic
Some 400 million Africans live in cities risks of runoff flooding.
today, nearly 39% of the population,
against 3% a century ago. In 2030, the
continent will have about 760 million
urban dwellers (more than the 2. Climate change impact on flood risk
population of Europe as a whole) and With climate change hitting hard sub-
1.2 billion in 2050, more than half of its Saharan Africa, the frequency of
population. Sub-Saharan Africa is catastrophic hydrological events in
ahead of all other regions of the world cities is increasing. For the Sahelian
in terms of growth of the urban zone, for example, recent academic
population: 4.1% per year in 2016 studies have attempted to apply
against 2% worldwide. Of the ten various regional climate models to the
countries with the highest urbanization prediction of rainfall field variations up
rates in the world, six are in sub- to 2050. Among the consensuses that
Saharan Africa. But south of the emerge among the used models : a
Sahara, 62% of city dwellers live in decrease in the frequency of low
precarious housing (against 36% in all rainfall (0.1 to 5 mm / day), but also an
developing countries), or 165 million increase in intensities of heavy rainfall
people. Most of the time, this (> 50 mm / day), a determining factor
urbanization explosion has in fact in the occurrence of river floods.
taken public authorities by surprise in
planning and building healthy habitats
3. Urban hydrology, poor parent and over other processes that contribute to
data scarcity the production of surface runoff.
Moreover, each part of the hydrological
In hydrology, runoff is the product of a process is explained differently
river's catchment response to rainfall. according to the authors, like the
The reliability of the estimation of flows infiltration modeled by more than one
and volumes flowing after a storm or equation: Horton, Green-Ampt, Holtan,
rainy event depends very much on the and so on. The hydrological models
reliability of the hydrologist's opt for an approach or at least for a
representation of both the rainfall and limited number of these approaches.
the watershed that converts it into flow. Each of the mathematical equations
As hydrological problems are that translate part of the hydrological
increasingly solved using computer process has more or less numerous
mathematical models, the modeling of parameters with more or less easy
rainfall and runoff events will involve determination. The more
modeling the project's rainfall and comprehensive the approach to
modeling the watershed. In either explain the hydrological process, the
case, in the African context, the greater the number of parameters. But
hydrologist will have to deal with a at the same time, the determination of
dearth of data needed for reliable these parameters becomes difficult,
modeling of hydrological phenomena in because it requires preparatory work,
urban areas. Despite the frequent or even field measurements not always
heavy rains in Sub-Saharan cities available, easy to achieve, or even
causing floods with even death of men, possible. The greater the number of
there is no record on canals and drains parameters, the greater the overall
in urban areas, neither of flow nor of error of the simulation of the
water levels. This makes it impossible hydrological process, because it is the
to use such recordings to calibrate the result of the partial errors made on the
parameters of a hydrological model. determination of each parameter. The
The values of other parameters, which questions for the hydrologist to choose
reflect certain soil properties of a Rainfall-Runoff hydrological model
urbanized watersheds, are also not are the following:
available, given the limited number of
studies that have affected this aspect • Which approaches best
in urban areas. explain the hydrological
phenomenon studied?
• Among these approaches,
4. Which models to choose to study which ones use a reasonably
urban hydrology? limited number of parameters
Computer or mathematical Rainfall- and whose determination is
Runoff models are numerous and exist made by effort and means
for more than forty years, period adequate with the constraints of
concomitant with the important the study in progress?
development of computer science. The model chosen should logically be
They are each based on a well-defined based on an approach that
approach to the hydrological appropriately translates the
phenomenon of rainfall conversion into hydrological process under study, as it
runoff. There are as many approaches should offer routines using parameters
as there are schools or hydrologists, whose determination is possible within
each focusing on physical processes the context of the study in question.
5. Feedback: study of the drainage of
runoff water in Abidjan
In 2015, was entrusted to a consortium
led by CONCEPT the study on
improving rainwater drainage systems
in the Abidjan area north of the Ebrié
lagoon. This was an edifying case
where the consulting engineer was
able to carry out a complete modeling
of hydrological phenomena in a context
of relative data scarcity. In order to do
this, the hydrologist has resorted to
several artifices and analogies allowed
by hydrological science to fill the lacks context such as that of Abidjan. The
in data coming from measurements choice among the options offered by
and field observations. this model is also explained followed
by explanation of the parameters used
It is first explained the way that was in calculations leading to flows
used to establish the Project Storm . determination at the different
For rainfall-runoff modeling, it is the computation nodes. In the absence of
rainfall intensities deduced from rainfall flow records in urban areas, the
records that are of interest, and not the calibration and validation of HEC-HMS
daily accumulations recorded at rainfall will be carried out later "indirectly"
gauges. However, it turns out that the through the HEC-RAS simulations of
only pluviograph still operating in the surface water levels produced by
Abidjan is the Airport’s one, located flows reproduced by HEC-HMS in
outside the study area. The response to recorded historical rain
heterogeneity of the rainfall in the events.
Lagunes region having been recalled,
it is justified to use rainfall records
made in Adiopoudomé, closer to the
project area. These, even of short
duration, make it possible to obtain an
IDF curve more representative of the
frequency distribution of the rainfall
intensities in the study area. But the
shape of the Project storm will be
drawn from a frequency analysis of
extreme rains recorded at the Airport
pluviograph for 30 years.
The hydrological modeling of
HEC-RAS is the tool that has served
watersheds is discussed later. After a
as a hydraulic model of the studied
justification of the choice of the model
drainage system basins. After
HEC-HMS - less "greedy" than others
describing its functionalities and
in data necessary for calibration of its
evaluating its adequacy with the issue
parameters -, this one is described
of drainage of runoff water in Abidjan,
from the point of view of its aptitude to
the procedure for calibrating the
reproduce the runoff phenomena in a
parameters of the model is described.
Calibration of the model parameters is "validate" the values of n deduced from
done by comparing, for known rainfall the calibration on M'pouto.
events, the calculated surface water The first simulations by the model thus
level with that observed via the traces calibrated were used to diagnose the
of water levels observed in the field at functioning of existing drainage
locations of well-defined control systems in the watersheds of Abidjan.
sections. The rainy episode of the Floodability maps have been prepared
beginning of June 2015, for which flood for three return periods: 10 years, 50
leashes were observed, was used for years and 100 years. For the same
this purpose. return periods were determined at
locations of each of the hydraulic
works, in particular at raods crossing,
the heights of submersion and the
regime of flow (open-channel or close-
conduit flow).

The geometry of the water courses

being supposed to be well defined by
the topographical layouts used, the
parameter which was used for the
calibration of the model is the Manning
"n" coefficient. Roughness of the
watercourse beds was defined in a first
step using tabulated values in the
specific literature, in view of the
photographs taken during wide visits of
the rivers.

However, this way of proceeding does

not give indisputable values for n and
the calibration was therefore to adjust
them by "trial and error" on the M'pouto
watercourse and to use the
observations made on other streams to
The simulations carried out at the
subsequent stage enabled to evaluate
the impact of the new structures
proposed on the improvement of
drainage for the 50-years return period.
Several options of redevelopment
and/or reinforcement of the system
were evaluated through simulations
results visualized on the water surface
profiles and cross sections outputs of
HEC-RAS.