Urbanization and land use as a driver of flood risk

Abstract

Obviously, floods have natural causes – heavy rainfall in the case of flash floods
– nevertheless, their consequences depend strongly on urbanization and land use.
On both sides of the Mediterranean, regions have been subjected to fast changes
which have led to a concentration of assets in valleys and coastal areas (Plan Bleu,
2008). Changes in land use play a predominant part in the “risk production process” i.e. in the increasing exposure of human activities to flood risk.
This chapter gives an overview of these land changes and their impact on flood
risk, based on documented cases in the Mediterranean.

Full text

563
Sub-chapter 3.4.1
Urbanization
and land use
as a driver of flood risk
Freddy VINET
Montpellier-3 University-IRD, France
Mohammed El Mehdi SAIDI
Cadi Ayyad University, Morocco
Johnny DOUVINET
Avignon University, France
Noomène FEHRI
University of Manouba, Tunisie
Wafa NASRALLAH
University of Manouba, Tunisie
Wahiba MENAD
Hassiba Benbouali University, Algérie
Samira MELLAS
Montpellier-3 University-IRD, France
Introduction
Obviously, floods have natural causes – heavy rainfall in the case of flash floods
– nevertheless, their consequences depend strongly on urbanization and land use.
On both sides of the Mediterranean, regions have been subjected to fast changes
which have led to a concentration of assets in valleys and coastal areas (Plan Bleu,
2008). Changes in land use play a predominant part in the “risk production process”
i.e. in the increasing exposure of human activities to flood risk.
This chapter gives an overview of these land changes and their impact on flood
risk, based on documented cases in the Mediterranean.
IRD_SYN_HC_Q_COP22_Part_3.indd 563IRD_SYN_HC_Q_COP22_Part_3.indd 563 06/10/2016 16:27:2506/10/2016 16:27:25

The Mediterranean Region under Climate Change
564
Recent catastrophic floods:
land use in question
Among the growing damage linked to floods, urbanization and more generally
un-adapted land use is often called into question. Three examples taken from
three Mediterranean countries (Tunisia, Morocco, Algeria) illustrate the impact
of land use with its considerable human and material toll.
The evolution of the global disasters toll
Such changes can be seen in the evolution and variability of natural disaster
annual assessments (fig. 1). Since 1970, the cost of natural disasters has
quadrupled, essentially due to an increase in the number and value of
exposed assets. Since the mid-1990s, costs have been subjected to high
inter-annual variability that does not only concern floods. The years with a
heavy toll are those where highly urbanized areas were hit (e. g. Katrina in
2005). The high variability of damage is therefore more indicative of the
spatial concentration of activities and population than of increasing natural
variability.
Figure 1
Evolution of insured and non-insured losses for “natural” disasters.
Source: MunichRe.
IRD_SYN_HC_Q_COP22_Part_3.indd 564IRD_SYN_HC_Q_COP22_Part_3.indd 564 06/10/2016 16:27:2506/10/2016 16:27:25

565
Urbanization and land use as a driver of flood risk
Recent floods in the Mediterranean
showing the impact of urbanization
Tunisia regularly suffers from torrential floods with considerable human losses
(fig. 2). On 13 October 2007, heavy rainfall hit Greater Tunis (70 to 180 mm
in 24 hours). 16 people died (Fehri et al. 2009), mostly on the GP8 road to the
north of Tunis. Through careless construction, increasing numbers of obstacles
(buildings, roads) had been built on riverbeds. Hydraulic works were often
under-sized and badly maintained.
Figure 2
Human toll of main flood events in Tunisia (1962-2011).
IRD_SYN_HC_Q_COP22_Part_3.indd 565IRD_SYN_HC_Q_COP22_Part_3.indd 565 06/10/2016 16:27:2506/10/2016 16:27:25

The Mediterranean Region under Climate Change
566
Like Tunisia, Morocco has suffered from numerous lethal floods since the mid-
20th century (fig. 3). The one that hit the Ourika valley (503 km², High Atlas
of Marrakech) was the deadliest in the country’s modern history.
Figure 3
Main deadly floods in Morocco since 1990.
In the afternoon of 17 August 1995, intense rainfall hit the upstream basin of
Ourika to the south of Marrakesh. The peak discharge reached around 1,000
m3/s in the gorges. This was where the damage was the most severe. The
estimated human toll ranged from 289 to 730 people (according to sources)
among the 35,000 holidaymakers in the valley that day (Secrétariat d’Etat auprès
du Ministère de l’Energie, des Mines, de l’Eau et de l’Environnement, 2008).
This toll can be explained by the exposure of both the population and
holidaymakers on the river bed. Restaurant terraces were set into the wadi’s
minor bed. Furthermore, the overcrowding of the valley late in the summer
IRD_SYN_HC_Q_COP22_Part_3.indd 566IRD_SYN_HC_Q_COP22_Part_3.indd 566 06/10/2016 16:27:2606/10/2016 16:27:26

567
Urbanization and land use as a driver of flood risk
afternoon, combined with the narrowness of the only road, blocked several cars
with their occupants. Trapped in their vehicles, the latter did not survive.
The most emblematic case of torrential floods in an urbanized zone remains that
of Bab-el-Oued. The huge rainfall of 9-10 November 2001 (263 mm in two days,
and above all 70mm in two sequences of 30 min.) triggered both human (781
deaths and 115 missing) and material (3,721 destroyed or damaged houses) losses
which had never been observed here before. The most important damage occurred
at Bab-el-Oued, situated in the western part of greater Algiers. In two decades
(1985-2005) the basin of Wadi Koriche experienced considerable urbanization - up
to 78% of its surface was built upon. The basin presents a high predisposition to
runoffs in case of intense rain. The inner-basin networks (25 km of roads, 7.8 km
of pipe works) served as collectors and enabled the propagation of flows towards
the Bab-el-Oued area. The main river bed had been piped and the Chevalet-Triolet
highway was built on top of it. Setting up a road network instead of a hydrographic
one removed it from the landscape. Gradually, people forgot about it and tended
to use the wadi’s bed. Most of the victims were car-drivers trapped by the flood
on the highway, and people at the Bab-el-Oued market.
These emblematic events show the part played by urbanization in the increased
cost of flash floods. The catastrophic tolls invariably result from inconsiderate
occupation of riverbeds by economic activities, housing or roads.
Land use transformations,
a key factor in flood related damage
Since the early 20th century, world population growth has been intense. In 100 years
(1915-2015), the world’s population rose from 1.9 to 7.4 billion people. It doubled
between 1970 (3.7 billion) and 2015 (7.4). It ranged from an “empty” Earth to a
“fully populated world” – “full” in the sense that we have reached the limits of the
Earth’s system capacities. This increase in the population was accompanied by deep
land transformations which determine the extent of flood-related damage.
Demographic growth and urbanization
Countries on the north and south sides of the Mediterranean basin have
experienced high demographic growth (fig. 4). For example, Morocco’s
population increased from around 26 million inhabitants in 1994 to nearly
33.8 million in 2014, i.e. over 7 million more people within 20 years. This
demographic growth came with spatial redistribution. The population is
concentrated in the cities of coastal areas and valleys. The urbanization rate
went from 48.6% of the population in 1990 to 60.3% in 2014 (Haut-
Commissariat au Plan, 2015).
IRD_SYN_HC_Q_COP22_Part_3.indd 567IRD_SYN_HC_Q_COP22_Part_3.indd 567 06/10/2016 16:27:2606/10/2016 16:27:26

The Mediterranean Region under Climate Change
568
Figure 4
Urban population in Mediterranean countries.
Source: Plan Bleu, 2008.
Most major cities in North Africa experienced demographic growth boosted by
rural-urban migration during the first decades of independence. The Greater
Tunis population rose from 194,000 inhabitants in 1921 to 561,000 in 1956 and
2.4 million currently. Simultaneously, urbanized areas expanded from 3,387 in
1957 to 28,000 ha in 2012 (source: l’Agence d’Urbanisme du Grand Tunis
(AUGT).
In the south of France, the population increased in the plains and regressed
considerably on higher lands (Cevennes, Pyrenees).
At the end of the 19th century, the population of the Languedoc (fig. 5a) was
balanced between populated countryside and a network of midsize cities. At the
end of the 20th century (fig. 5b), the population was concentrated on the coast,
big cities and communication axes.
The attractiveness of coastal areas is naturally explained by touristic and
industrial activities. In Morocco, 70% of the country’s tourism is concentrated
on the Atlantic and Mediterranean coasts. The coast is also home to most
industrial activities. This trend should continue, particularly with the development
of the “Tanger Med” harbor (the biggest in the Mediterranean basin) and the
extension of the “Nador West Med” port. However, the concentration in low
areas cannot be narrowed down to the appeal of maritime activities. Indeed,
many so called “coastal” cities have no major activity linked to the sea (e.g.
Montpellier in southern France). The attractiveness of low lands is mainly
explained by the proximity to roads.
IRD_SYN_HC_Q_COP22_Part_3.indd 568IRD_SYN_HC_Q_COP22_Part_3.indd 568 06/10/2016 16:27:2606/10/2016 16:27:26

569
Urbanization and land use as a driver of flood risk
A
Population per municipality in 1881
B
Figure 5
Spatial distribution of population in Languedoc (southern France)
in 1881 (fig. 5a) and 2005 (fig. 5b).
IRD_SYN_HC_Q_COP22_Part_3.indd 569IRD_SYN_HC_Q_COP22_Part_3.indd 569 06/10/2016 16:27:2606/10/2016 16:27:26

The Mediterranean Region under Climate Change
570
Population growth and flood risk stakes
Demographic growth and urbanization brought about land and society
transformations that induced changes in the assessment of flood risk. As the
exposed assets increase and move, social vulnerability and resilience evolve.
The first consequence of these land transformations is the increase in assets in
flood-prone areas. As in North Africa, the French population has increased in
the last 50 years. It rose from 46 to 66 million inhabitants between 1962 and
2016. As a consequence, real estate increased by 8 million dwellings. Housing
demand was met according to urban planning regulations where environmental
risk questions remained marginal until the 1980s. Urbanization decisions were
based on the available land, in relation to transport networks and activities,
without really taking flood risk into account, except some local cases where
risk awareness was high.
The assessments of numbers of housing or people living in flood-prone areas
must be taken with caution as they vary considerably according to the chosen
context. In France the number of people living in flood-prone zones comprises
between 6 million and 17 million according to the Evaluation Préliminaire du
Risque Inondation body (EPRI) following the setup of the 2007 European
directive. French Mediterranean departments are among the most exposed (see
ONRN website www.onrn.fr ). The Alpes-Maritimes (+9,200 dwellings in flood-
prone zones in between 1999 and 2008), the Var (+8,300 dwellings) and the
Vaucluse (+5,800 dwellings) are among the six departments most affected by
real estate demand in France.
In Morocco, although the population is concentrated on the coast, the inner
valleys have experienced important growth of urban perimeters as in the Souss
valley, where newly built zones continue to encroach on the wadi bed, and as
in the Jorf and Tarrast neighborhoods in the south of Agadir (Rehaimi, 2013;
Saidi et al, 2013). Such encroachments expose numerous populations who are
sometimes already vulnerable due to their socioeconomic precariousness.
Changes in runoff conditions
Besides the increase in exposed assets in risk-prone areas, urbanization results
in changes in local runoff conditions. Such modifications are now well known
(e.g. Zevenbergen et al. 2010). Apart from the increase in impervious areas,
basin urbanization is characterized by changes in runoff conditions by
longitudinal and transversal obstacles. In addition, the disappearance of rivers
in the urban landscape contributes to forgetting the risk and encourages the
occupation of minor and major river beds. The watershed of the El-Ghrich
El-Greb wadis north of Tunis is a good example of the impact of expansion and
densification of urban areas on runoff coefficients (Fehri and Zahar, 2016). From
the 1950s onwards, this watershed experienced rapid urban expansion. So much
so that buildings and roads now take up over 70% of its total surface. The
exposure of the road network is also a growing danger factor (fig. 6).
IRD_SYN_HC_Q_COP22_Part_3.indd 570IRD_SYN_HC_Q_COP22_Part_3.indd 570 06/10/2016 16:27:2706/10/2016 16:27:27

571
Urbanization and land use as a driver of flood risk
Highways
Hidrography
Railway
National road
Regional road
Watercourse rank 1
Drainage Network
Drain
Crosscheck wadi/roads
Node
620000
2 Km0
624000
-4058000
To
Sfax
To
Soliman
Borj Cedria
Ech-Chaâbia
Hamman-Lif
To
Tunis
Hamman-Echatt
Realized by: NASRALLAH Wafa (August 2012).
Source: La Goulette topographic sheets NW, SE and SW at a scale of 1/25000.
-4062000
S
N
EW
-4066000
628000624000620000
4066000
Watercourse rank 2
Watercourse rank 3
Watercourse rank 4
Watercourse rank 5
G
u
l
f
o
f
T
u
n
i
s
M
o
u
n
t
B
o
u
G
a
r
n
i
n
e
Figure 6
Crosscheck between watercourses
and roads in the southern suburbs of Tunis.
How to control flood risk?
Faced with the increase in assets in flood-prone areas, the sometimes relative,
yet real failure of land use management must be noted. Land demand has been
too high and changes in land use too sizeable and rapid to be controlled in risk-
prone areas. Land use control tools have been set up, particularly in France but
their efficiency has been limited. Moreover, they do not efficiently address the
question of existing buildings in flood-prone zones.
Debates tend towards integrated flood risk management, from prevention to
post-crisis, via crisis management preparedness and warning devices. However,
structural measures (i.e. aiming at controlling flood hazard: dams, dikes, river
bed recalibration) remain often the first considered type of response to risk
whereas non-structural measures struggle to become widespread, except for
warning and crisis management which are increasingly efficient.
Structural measures are still dominant
The set of flood prevention tools has broadened considerably (Andjelkovitch,
2001), but despite international requirements for comprehensive risk management
IRD_SYN_HC_Q_COP22_Part_3.indd 571IRD_SYN_HC_Q_COP22_Part_3.indd 571 06/10/2016 16:27:2706/10/2016 16:27:27

The Mediterranean Region under Climate Change
572
plans (2007 European Directive; UNISDR, 2005, UNISDR, 2015), the temptation
to use structural measures (such as dikes) as the solution to all flood risk problems
is still high, whether in Morocco, Algeria, France or Tunisia.
After the Bab-el-Oued floods, the newspaper Le Quotidien claimed: “40 million
dinars for a protection plan. Bab-el-Oued will never experience another flood”.
The newspaper article presented projects commissioned or considered by the
State to protect Algerian cities from floods. At the same time, new buildings
were built (after the disaster of 10 November 2001) such as a primary school
in the minor bed of Wadi Koriche in Bab-el-Oued.
For several decades Morocco has relied on a dam building policy (150 great
dams storing 18 billion m3 of water), initially designed to fight drought. In these
dams, important parts were reserved for flood water reduction.
These works are expensive and do not guarantee 100% protection. They may
even be an additional risk factor during exceptional events (dam or dike
break). The only 100% efficient prevention method is to refrain from locating
vulnerable activities and buildings in flood-prone areas. This may seem
obvious but it does not appear to have been incorporated at every level of
land use planning.
The difficult control of land use
Land use control is a fundamental point of risk reduction policies and practices
(Pottier et al. 2005). Existing regulations in France (R 111-3, Submersible
Surface Plan (PSS)) were rarely applied until the setup of the Exposure to
Predictable Natural Hazard Plans (PER in French) in 1984 and the Risk
Prevention Plan (PPR in French) in 1995. When these plans are approved, they
are generally respected but local authorities sometimes use stratagems to delay
their application. The Ministry of Ecology (CGDD, 2009) showed that in 424
French cities, nearly 100,000 dwellings were built in flood-prone areas between
1999 and 2006. Demand is also high from local authorities to urbanize areas
“protected” by dikes.
Living with water?
For the millions of people living in flood-prone areas, living with water is an
undeniable fact. In this case, they must adapt existing buildings in order to
limit damage costs and ensure the safety of people (Kelman, 2007). In urban
renovation areas, there have been attempts to take the risk into account within
the redefinition of urban space. Activities are spread out according to their
vulnerability, in decreasing order, as the flood hazard increases. Thought has
been given to this question (November et al. 2011) and the French Ministry
of Ecology has initiated reflections on how to build safely in prone areas
(Bonnet & Morel, 2016).
IRD_SYN_HC_Q_COP22_Part_3.indd 572IRD_SYN_HC_Q_COP22_Part_3.indd 572 06/10/2016 16:27:2706/10/2016 16:27:27

573
Urbanization and land use as a driver of flood risk
Conclusion
The information provided by retrospective studies on past evolutions of
frequency and intensity of heavy rainfall in the Mediterranean and models
developed to anticipate these evolutions in the future is uncertain (Soubeyroux
et al. 2015; Tramblay et al. 2012). However, the socioeconomic and
demographic human evolutions that the Mediterranean basin has experienced
over the last 50 years have proved far more decisive. The land use
transformations have constituted a real “risk production process” through
the increase and the spatial concentration of exposed assets and the evolution
of vulnerabilities.
These land use changes are linked to broader changes in our societies. They
are disconnected from the risk issue but they have repercussions for the
population’s exposure. They are not only quantitative (rise of assets at risk)
but also qualitative. They affect the degree and nature of social and territorial
vulnerabilities facing floods. Our societies are not vulnerable in the same
way as they were 50 years ago, nor in the same places. The vulnerability of
traditional housing (adobe bricks, cob) facing floods is thus reduced by the
use of concrete but reappears in other forms (precarious housing) and other
places. Conversely, the vulnerability linked to migration increases with the
often anarchic densification of networks.
Demographic pressure is not expected to decrease before 2050. It has been
forecasted that at this date, there will be near 170 million inhabitants on the
European side (versus 140 in 2005) and over 300 million on the east and
south coasts of the Mediterranean basin (151 in 2005). Near 2030, around
42 million extra dwellings will be necessary, mainly in cities (source: http://
planbleu.org). This will also be followed by the aging of the population
including in the South (Coudert, 2002). Research shows that the elderly are
more vulnerable in the face of flood risk (Jonkman & Kelman, 2005; Vinet
et al. 2012).
Prevention disconnected from the land use control issue is doomed to fail.
Above all, the question of flood risk management and prevention is a
territorial issue with a heavy social and political resonance. The choices
made in this field may determine the risk of future and forthcoming
disasters.
While the potential impacts of climate change are certain for some and more
hypothetical for others, social and land use evolutions are decisive in the
process of risk production. Urbanization, population growth, social
transformations, and housing evolution are key parameters in flood risk
evolution. Beyond any consideration about climate change, it is urgent to
take disaster risk reduction measures – and preferably non-structural
measures that are more resilient in the face of climate change.
IRD_SYN_HC_Q_COP22_Part_3.indd 573IRD_SYN_HC_Q_COP22_Part_3.indd 573 06/10/2016 16:27:2706/10/2016 16:27:27

The Mediterranean Region under Climate Change
574
References
ANDJELKOVIC I., 2001
Guidelines on non-structural measures in urban
flood management. Paris, UNESCO, 81 p.
BONNET F., MOREL J.F. (DIR.), 2016
Atout risques. Des territoires exposés se
réinventent. Marseille, Parenthèses, coll.
territoires en projets, 176 p.
CGDD, 2009
Croissance du nombre de logements en zones
inondables. Le point sur. Commissariat Général
du Développement Durable, 6:1-4.
COUDERT E., 2002
Une approche régionale de la population et de
l’urbanisation en Méditerranée, rétrospective et
projections à 2025. In Carrière J.P. (ed.) : Villes
et projets urbains en Méditerranée,
PUFR, 135 p.
FEHRI N., SAMAALI S.,
ABAZA K., 2009
Les inondations catastrophiques du 13 octobre
2007 dans le secteur de Sabbelet Ben Ammar :
entre aléa climatique et responsabilité
anthropique (Grand Tunis – Tunisie). Revue
tunisienne de géographie, 40: 31-55.
FEHRI N., 2014
L’aggravation du risque d’inondation en Tunisie :
éléments de réflexion. Physio-Geo, 8: 149-175.
FEHRI N., ZAHAR Y., 2016
Etude de l’impact de l’extension et de la
densification du tissu urbain sur les coefficients
de ruissellement dans le bassin versant des oueds
El-Ghrich et El-Greb (Tunis) par l’application de
la méthode SCS aux évènements de septembre
2003. Physio-geo, 10: 61-79.
HAUT-COMMISSARIAT AU PLAN, 2015
Note sur les résultats du recensement général de
la population et de l’habitat de 2014 au Maroc.
Rabat, 57 p. Http://www.hcp.ma
JONKMAN S. N., KELMAN I., 2005
An analysis of the causes and circumstances of
flood disaster deaths. Disasters, 29 (1): 75-97.
KELMAN I., 2007.
Decision-making for flood-threatened properties.
In Begum S et al. (eds.). Flood risk management
in Europe, Springer, Dordrecht, 534 p.
EL MEHDI SAIDI M.,
BOULOUMOU Y., ED-DAOUDI S.,
ARESMOUK M. EL HASSANE, 2013
Les crues de l’oued Issil en amont de Marrakech
(Maroc), un risque naturel récurrent. European
scientific journal, 9(23): 189-208.
MENAD W., 2012
Risques de crue et de ruissellement superficiel en
métropole méditerranéenne : cas de la partie
ouest du grand Alger. Thèse de doctorat en
géographie. Université Paris Diderot, 330 p.
NOVEMBER V., PENELAS M.,
VIOT P. ( DIR.), 2011
Habiter les territoires à risque. Lausanne :
presses polytechniques et universitaires romandes,
collection « espaces et sociétés », 252 p.
ONRN OBSERVATOIRE NATIONAL DES RISQUES
NATURELS
http://www.onrn.fr/
PLAN BLEU. 2008
Les perspectives du Plan Bleu sur le
développement durable en Méditerranée. PNUE/
PAM, Sophia-Antipolis, 27 p.
REHAIMI H., 2013
le système de transport et la ségrégation sociale
de l’espace urbain : cas du grand Agadir. Thèse
de l’université Ibnouzohr, Agadir, 354 p.
SECRÉTARIAT D’ETAT AUPRÈS DU MINISTÈRE
DE L’ENERGIE, DES MINES, DE L’EAU
ET DE L’ENVIRONNEMENT, 2008
Etude pour la réalisation d’une cartographie et
d’un système d’information géographique sur les
risques majeurs au Maroc. Mission 1 :
identification des risques d’inondation. Etudes et
mesures, Maroc, 57 p.
SOUBEYROUX J. M., NEPPEL L.,
VEYSSEIRE J. M., TRAMBLAY Y.,
CARREAU J., GOUGET V., 2015
Evolution des précipitations extrêmes en France
en contexte de changement climatique. La
Houille Blanche, 1: 27-33.
TRAMBLAY Y., NEPPEL L.,
CARREAU J., SANCHEZ GOMEZ E., 2012
Extreme value modelling of daily areal rainfall
over Mediterranean catchments in a changing
climate. Hydrological processes, 26: 3934–3944.
IRD_SYN_HC_Q_COP22_Part_3.indd 574IRD_SYN_HC_Q_COP22_Part_3.indd 574 06/10/2016 16:27:2706/10/2016 16:27:27

575
Urbanization and land use as a driver of flood risk
UNISDR, 2005
Hyogo framework for action 2005-2015.
Building the resilience of nations and
communities to disasters. World conference on
disaster reduction, 18-22 January 2005, Kobé,
Hyogo, Japan.
UNISDR, 2015
Sendai framework for disaster risk reduction
2015-2030. 35 p.
VINET F., LUMBROSO D.,
DEFOSSEZ S., BOISSIER L., 2012
A comparative analysis of the loss of life during
two recent floods in France: the sea surge caused
by the storm Xynthia and the flash flood in Var.
Natural hazards, 61(3): 1179-1201.
ZEVENBERGEN C., CASHMAN A.,
EVELPIDOU N., PASCHE E.,
GARVIN S., ASHLEY R., 2010
Urban flood management. CRC Press Routledge,
340 p.
IRD_SYN_HC_Q_COP22_Part_3.indd 575IRD_SYN_HC_Q_COP22_Part_3.indd 575 06/10/2016 16:27:2706/10/2016 16:27:27