You are on page 1of 2

International Conference on Applied Geology & Environment “iCAGE 2016”

19-21 May 2016, Royal El Mansour Mahdia -Tunisia


[T2-OS1]

Mapping flood risk areas IN GABES BASIN (SOUTH-EASTERN TUNISIA)

N. Dahri 1,2, M. Ellouze 1,2, A. Atoui3 and H. Abida 1,2

1 Applied Hydro-Sciences Research Unit, high Institute of Water Sciences and Techniques of Gabes,
Univ. of Gabes, 6072, Gabes, Tunisia
2
Faculty of Sciences PB 1171, Univ. of Sfax, 3000, Sfax, Tunisia
3The regional Directorate of urban hydraulics of Sfax

KEYWORDS: GIS, AHP, Vulnerability, Flood risk, Gabes Basin

Floods are one of the most severe natural disasters as they inundate large areas. Flood events
generally cause damage to agricultural crops and property, roads and rail links, and may even result in
the loss of human lives. The extension of vulnerable flooding areas is an important consideration in
the prediction of flood risk and the management of natural resources. This study is interested in the
identification and characterization of flood risk areas in the Gabes Basin, South-Eastern Tunisia.
Gabes watershed, located in Southern Tunisia on the Mediterranean Sea, covers a drainage area of
95 km2. The study area is limited by Dissa Mountain in the north, Mediterranean Sea coast in the east,
Jebel Moncef in the west. The southern limit is represented by the hill of El Manara. Gabes Catchment
is characterized by a mild topography, with altitudes varying from 0.3 m in the north east to 235 m
above sea level in the south west. This area characterized by influences of both dry/hot and humid air
masses coming from the desert and the Mediterranean Sea respectively. Rainfall is characterized by
its shortage, irregularity and erratic distribution leading to dry and intensive rainy periods with high
flooding events. Average rainfall is approximately 185 mm/year. The summer flash flood event of June
2014 registered the maximum rainfall over a period of almost 6 hrs varied between 25 and 110 mm all
over Gabe’s region, showing an erratic spatial rainfall distribution. The areas bordering the
Mediterranean Sea obtained the highest rainfall accumulation during this event (107 to 110 mm).
Gabes City is affected highly by rainfall events in both time and space, because of its geographical
location and its mild topography. The spatial distribution of rainfall data is the parameter that used to
characterize hazard map of the study event. Thematic layers of rainfall event are produced using
kriging interpolation.
Five parameters are used to elaborate the vulnerability map, including soil type, slope gradient, land
cover, hydrographic network and drainage density. The analysis starts with the digital elevation model
(DEM), created from both real elevation data and SRTM 30 m. Real elevation data were derived from
the urban management plan. The obtained elevation data were calibrated at a regional scale and
organized as grid data, corresponding to the 1:30 000-scale. In order to verify their fidelity, the DEM
contours were superimposed onto the contour maps of a 1:25 000-scale published by the office of
Topography and cartography (OTC), Tunisia. The comparison verified the good correspondence of the
contour lines.
The obtained DEM was adopted to obtain digital thematic maps. Digital slope and watercourse
network maps were developed. Different reaches of the watercourse network were categorized using
Strahler (1957) and Horton (1967) classification methods in order to identify their order of importance.
Moreover, a drainage density map is produced from stream network data using kernel density
algorithm. Drainage density is expressed in total length of hydrographic network per unit area. The
lithological basin map resulted from the digitalization of the geological maps (1:50,000) as vector files
and then rasterized. Remote sensing and GIS were considered in the treatment the Landsat TM
imagery in order to delineate different land uses in the Gabes catchment. Land cover data are
classified using the supervised classification method and field surveys. This method showed fine
details not achieved by photo-interpretation. The probability theory was used to accomplish the
classification.
Analytic Hierarchy Process (AHP) techniques were extensively utilized in mapping flood risk zones.
The AHP is a type of multi_criteria analysis subdivided into three steps: (1) standardization (2) weight
assignment, and (3) weight linear combination. The spatial standardization of criteria is associated
with a common scale of measurement. Standardization is the process of translation and normalization

177
International Conference on Applied Geology & Environment “iCAGE 2016”
19-21 May 2016, Royal El Mansour Mahdia -Tunisia
[T2-OS1]

of the various input data to a common scale, to allow comparison and analysis. As mentioned in the
beginning, this approach is necessary to obtain the criteria attribute and then to rank their order. The
standardization applied in this study attributes the value 1 to very low contract, 2 to low, 3 medium, 4
high and 5 very high contracts. The degree of criteria was assigned depending on the importance of
its effect. Factor weights are developed to assign weights to each criterion based on its rank. In
determining criteria weights for any factor, we used continuous rating scale for pair wise comparison of
Saaty's method. In this study, weight is defined as a value assigned to an evaluation, which indicates
the importance of one factor associated with another. This methodology includes two principal factors.
They are hazard and vulnerability which have the same weight based on its contribution to serious
flooding risk. All described factors were developed in raster format with a 30 m grid and the weighted
linear combination (WLC) is adopted to derive the flood risk map.
The obtained vulnerability map shows that 11.5 % of the catchment area is characterized by high
and very high vulnerability levels. The spatial distribution of vulnerable area depicted that vulnerability
is very high in both high and low parts of the catchment, especially Gabes City, which also presented
the highest hazard condition. 23.5 % of Gabes Basin area presented by moderate vulnerability levels,
located especially in low parts and close to watercourse networks. 26.6 % of the catchment area is
classified as high and very high hazard levels. High vulnerability and high hazard lead to very high
flooding risk. The high and very high flooding risks were shown to correspond to 16 % of the total
area. The maximum levels are shown in Fig. 1. This study showed that Gabes City, located in the low
part of the catchment, presented the extremely highest flooding risk. The area of very low to low
flooding risk corresponds to 69.2 % of the total area. The spatial distribution of flooding risk (Fig. 1)
depicted that the risk condition is very high in the low part of Gabes Watershed. These areas are
mainly represented by high density of urban areas and low topography indices. In fact, in Gabes City
many settlements are rapidly growing in flood prone areas. Some communities of the city are usually
inundating. It may be concluded, therefore, that risk conditions in the examined watershed have a SW-
NE geographical distribution. Based on these results, Gabes City is planning several adaptive
measures to protect the area against floods.

Fig. 1 Flooding risk map produced by AHP

178

You might also like