Regional Studies in Marine Science 55 (2022) 102564

Contents lists available at ScienceDirect

Regional Studies in Marine Science

journal homepage: www.elsevier.com/locate/rsma

Spatio-temporal analysis of shoreline changes and erosion risk

assessment along Jerba island (Tunisia) based on remote-sensing data
and geospatial tools
∗
Amina Boussetta a , , Simona Niculescu b , Soumia Bengoufa b,c , Mohamed Faouzi Zagrarni a
a
U.R Applied Hydrosciences UR13ES81, Institut Supérieur des Sciences et Techniques des Eaux de Gabès (ISSTEG), University of Gabes, Tunisia
b
CNRS, LETG Brest UMR 6554 CNRS, Technopôle Brest-Iroise, University of Western Brittany, France
c
Ecole Nationale Supérieure des Sciences de la Mer et de l’Aménagement du Littoral, (ENSSMAL), University Campus of Dely Ibrahim Bois des
Cars, Algeria

article info a b s t r a c t

Article history: Jerba, the largest island in North Africa, is among the most exposed islands to the combined effects
Received 25 January 2022 of climate change and human activities in the Mediterranean. The eastern coast of Jerba island has
Received in revised form 14 July 2022 long been subject to intense tourist development due to its attractive natural potential. The main
Accepted 15 July 2022
issue addressed in this study concerns the evolutionary trends of Jerba shoreline and the assessment
Available online 25 July 2022
of coastal erosion risk. This work focuses on the historical study of the shoreline positions based on
Keywords: satellite data processing. The use of geographic information system (GIS) and the Digital Shoreline
Coastal erosion Analysis System (DSAS) geostatistical calculation tool allowed for the identification of erosion and
Sandy spit accretion rates and the description of evolutionary trends across 32 years covering three periods
Band ratioing (1989–1999, 1999–2009, and 2009–2021). The results show an overall erosion trend (89% of the
Satellite imagery coastline). The statistics revealed that the study area, with the exception of the sandy spit, has
Sandy beaches experienced an intense and alarming level of erosion between 1989 and 1999, with an end point
DSAS
rate (EPR) equal to −4.12 m/year. This net rate continued to be negative in the period between 1999
and 2009, with an EPR of −0.5 m/year. Between 2009 and 2021, the net rate of shoreline change
is marked by slight accretion, at an average of +1.4 m/year. The overall long-term trend shows the
shoreline is eroding, with an average linear regression rate of −1.38 m/year for the whole period. This
trend is due to a combination of the cumulative effects of sea-level rise and anthropogenic actions.
Therefore, sustainable management strategies need to be implemented urgently by coastal managers
to restore the sectors that are susceptible to further erosion.
© 2022 Elsevier B.V. All rights reserved.

1. Introduction
Coastal areas are a valuable shelter for natural heritage but
also the most developed land zones where most of the megaci-
ties are situated, attracting people and simulating the migration
process (Barragán and de Andrés, 2015). Therefore, any negative
physical impact on these areas is certain to create massive natural
and socioeconomic damages at a global scale (Brown et al., 2016;
Hallegatte et al., 2013; Ranasinghe, 2016). Due to the threat of
climate change and significant contributions from local and re-
gional causes, coastal risks have increased (Quadrado et al., 2021).
In fact, environmental and climate change are leading to sea-
level rise (SLR), which is one of the most obvious consequences
of global warming. Over the 1993–2014 period, altimetry data
∗ Corresponding author.
E-mail addresses: amina.boussetta434@gmail.com (A. Boussetta),
simona.niculescu@univ-brest.fr (S. Niculescu), soumia.bengoufa@univ-brest.fr
(S. Bengoufa), zagrarni_m_f@yahoo.fr (M.F. Zagrarni).
suggested that the global sea level has risen by an average of
3.2+-0.4 mm/yr (Cazenave and Cozannet, 2014). The International
Panel on Climate Change (IPCC, 2021) predicts that SLR will be
0.63–1.01 m by 2100, which would be twice as fast as that which
has occurred over the last 100 years (Masson-Delmotte et al.,
2021). The rate of coastal erosion is over twice the rate of SLR.
Thus, even small changes in the sea level result in significant
losses for low-lying zones (Leatherman et al., 2000), and the
existence of some insular environments and deltaic coasts is
threatened. A further threat to economically valuable real estate
comes from an exacerbation of sandy beach erosion. As the beach
disappears, nearby fixed structures are increasingly exposed to
direct wave impacts and will eventually be damaged or destroyed
(Zhang et al., 2004).
Coastal erosion is a global phenomenon exacerbated by natu-
ral and anthropogenic factors. The coastal concentration of pop-
ulation and economic activities makes coastal erosion a global
concern (Bera and Maiti, 2019). According to Awange et al. (2018),

https://doi.org/10.1016/j.rsma.2022.102564
2352-4855/© 2022 Elsevier B.V. All rights reserved.
A. Boussetta, S. Niculescu, S. Bengoufa et al.
83% of sandy beaches worldwide are retreating. The Mediter-
ranean basin has about one third of its population concentrated
along coastal regions (Hřebíček et al., 2015) and does not es-
cape this evolution trend. The Tunisian coastline is one of the
areas most affected by SLR, being affected much faster than the
world average and is estimated locally at 5.7 mm/year (Pirazzoli,
2021). The sandy coasts are marked by an intense retreat of the
shoreline (Amrouni et al., 2019). Shoreline retreat along island
coasts is particularly alarming because of their limited beach di-
mensions and insufficient sediment supply (Peduzzi et al., 2013).
According to the national report of the Minister of Environ-
ment and Land Management of Tunisia, the coastal areas of Jerba
island are ranked among the most affected coasts of Tunisia
(Bouchahma and Yan, 2012), where there is no input of sand from
the mainland (Paskoff, 2004). This situation on Jerba island has
been exacerbated by the development of tourist infrastructures
located near the shoreline, which generate significant environ-
mental stress in the drift zones (Anfuso and Martínez del Pozo,
2005).
The impact of the erosion phenomenon, the tourism infras-
tructure, and the harbor on the evolution of the sandy beaches of
Jerba have been considered the major interests of regional stud-
ies. Indeed, there is a limited number of recent studies regarding
shoreline retreat along Jerba Island. Miossec and Paskoff (1979)
focused on the evolution of sandy coasts based on aerial pho-
tographic analysis. Bourgou (2005), Oueslati (2004), and Paskoff
(2004) focused on the morphological evolution, the imbalance
of the coastal ecosystem of Jerba and human intervention. More
recently, Bouchahma and Yan (2014) and Bouchahma et al. (2017)
studied the shoreline evolution of Jerba Island based on satel-
lite images analysis. These studies have focused on the coastal
dynamic and evolutions that took place in Jerba, but no study
has calculated the erosion and accretion rates and estimated the
shoreline evolution trends in detail. An erosion measurement
specific to the site is required for any management process (Bera
and Maiti, 2019). The erosion and accretion rates are commonly
estimated based on shoreline indicator detection and monitoring,
regarding which Boak and Turner (2005) have reviewed the im-
portance of the choice of an adequate indicator depending on the
coastal geomorphological type.
Furthermore, the detection and extraction of the shoreline
indicator are tedious tasks when using traditional techniques,
such as in-situ surveys or manual digitization, which are costly
and time consuming. The availability of remotely-sensed data
is emerging as one of the best means for monitoring of the
earth (Konko et al., 2020) and provides synoptic coverage of
the coastal zone. Satellite image processing is therefore pre-
ferred for detailing and monitoring geographical features (At-
sri et al., 2018; Konko et al., 2016). Moreover, recent advances
in image-processing methods overcome the limits of traditional
techniques. The use of remote sensing offers new technologi-
cal advantages for research in the field of coastal monitoring
and allows researchers to obtain reliable information regarding
shorelines (Konko et al., 2020). Furthermore, shoreline positions
are detected based on several image-processing techniques, such
as thresholding (Cui and Li, 2011; Yamano et al., 2006), super-
vised (Bengoufa et al., 2021; Garcia Rubio et al., 2014; Chang
et al., 2005), and band-ratio calculations (Islam et al., 2021; Bera
and Maiti, 2019; Louati et al., 2015; Cui and Li, 2011; Kuleli,
2010). These techniques allow for the automatic extraction of
the shoreline and could minimize the subjectivity of manual
digitizing. Furthermore, many publications that have focused on
shoreline extraction and evolution either simply present the re-
sults without providing sufficient detail on the chosen methods
(Toure et al., 2019) or do not make any explicit validation of
their extracted shoreline with reference data (Dada et al., 2016;
2
Regional Studies in Marine Science 55 (2022) 102564
Hegde and Akshaya, 2015) even though a separate dataset is
essential for understanding and validating the performance of any
remote-sensing product (Bengoufa et al., 2021; Lillesand et al.,
2015).
Overall, the present paper contributes to the understanding of
the sandy coast’s evolution along Jerba island. The paper proposes
a fully automatic and validated method to extract the shoreline
indicator position of sandy coasts and implements the use of
a geographic information system (GIS) and the processing of
satellite images to characterize the risk of coastal erosion. Us-
ing the Digital Shoreline Assessment System (DSAS), statistical
parameters were derived to estimate erosion rates. Based on
spatio-temporal analysis of the shoreline position over a period
of 32 years, specific information and precise cell-by-cell values
for every 30 meters are provided for the eastern coast of Jerba
island.
2. Study area
Jerba, the largest Island of North-Africa, is 514-km2 island
with 150 km of shoreline located in south-eastern Tunisia at a
northern latitude between 33◦ 57’0.56" and 33◦ 37’46.82" and an
eastern longitude between 10◦ 45’38.43" and 11◦ 3’53.802" (Fig. 1).
It is characterized by a semi-arid Mediterranean climate with
an average annual precipitation of 247.47 mm/year (1984–2014)
according to global meteorological data for Soil and Water As-
sessment Tool (SWAT) and an average annual temperature of
20.61 ◦ C. The wind annual distribution, according to data from
the synoptic station at Jerba airport, indicates a high average of
easterly winds (47%), followed by westerly winds (28%). Only 19%
of periods are calm in autumn; 15%, in winter; 10%, in spring; and
16%, in summer. The tidal type is semi-diurnal and regular, with
the most frequent range of 0.2 m (IHEE, 2015).
Jerba is characterized by a low average altitude of 20 m with
average slope values between 0% and 4%. The coastal fringe is
narrow and flat, which explains the extension of the sabkha and
chotts. This topography is prolonged in the submarine domain
by an extended and shallow continental platform. The isobath
of −10 m is located at an average of 3 km from the shore.
The north-eastern shoreline, which stretches for about 30 km
between Rass Rmel and Aghir, is the most dynamic sector of
the island, affected by the degradation of the components of the
natural environment.
Since the 1960s, Jerba has started to attract tourists and has
become one of the most popular destinations in the region. Be-
fore the elaboration of the tourist development plan (ITALCON-
SULT, 1975), the hotel space was already being exploited along
the eastern coast. Sixteen hotel establishments occupied 140.95
hectares there. This first development led to a large variety of
initiatives, which resulted in the construction of buildings estab-
lished without any reference to an organized spatial framework.
The result was ‘‘bad littoralization’’ (Chapoutot, 2011), with a
scattering of hotels along the coastline and a privatization of the
waterfront.
3. Materials and methods
To assess and characterize the coastal erosion phenomenon
across Jerba’s shoreline, the present study’s approach is based on
an estimation of the shoreline evolution trends (retreat, stability,
or advance) for the period from 1989 to 2021. The erosion rates
are calculated based on a spatio-temporal analysis of the shore-
line position extracted automatically from remote-sensing data
based on the band ratioing method.
A systemic approach was developed highlighting different
steps, from the analysis of the characteristics and geomorpho-
logical type of the coastal area and data collection, to the data
A. Boussetta, S. Niculescu, S. Bengoufa et al.
Fig. 1. Geographic location of the study area.
processing, extraction of the shoreline, and the erosion rate
estimation. Proper implementation of these steps is essential
to accurately detect and monitor the shoreline indicator using
appropriate tools and techniques.
3.1. Environmental analysis of the studied area
The first step of our approach aims to analyze the geographical
and geomorphological context of the study area via an envi-
ronmental analysis. Several field surveys were conducted over
different periods of 2018, which gave us the opportunity to ob-
serve the different signs of coastal erosion and to visualize the
morphological changes along the coastline.
We therefore focused on the morphology of the coastline with
the aim of locating and describing its types (i.e. sandy beaches,
3
Regional Studies in Marine Science 55 (2022) 102564
rocky coastline, spit) before analyzing the shoreline’s evolution.
This geomorphological analysis allowed us to identify the possi-
ble choices for a shoreline indicator to monitor the shoreline’s
kinematics. Indeed in addition to the preliminary analysis via
field investigations, visual inspections of very high-resolution
satellite images were conducted. We note that the study area is
characterized by a homogeneous geomorphology, with its near
totality being marked by the presence of linear sandy beaches
in the south (touristic and Aghir zones), sandy spit (Rass Rmel)
in the north, and the absence of rocky coasts or river and wadi
mouths. Furthermore, it is remarkable that a high percentage of
the coastline is artificial since most of the touristic infrastruc-
ture is located in the Public Maritime Domain. Indeed several
protection structures have been implemented to prevent coastal
erosion.
A. Boussetta, S. Niculescu, S. Bengoufa et al.
Fig. 2. Workflow of the global study approach.
3.2. Data set
Three types of Landsat images of medium resolution were
used, namely Thematic Mapper (TM), Enhanced Thematic Mapper
Plus (ETM+), and Operational Land Imager (OLI) images. These
data were downloaded freely in GeoTIFF format via the United
States Geological Survey (USGS) Earth Explorer portal (http://
www.earthexplorer.usgs.org) (Table 1), with Path = 190 and Row
= 37 from the Worldwide Reference System WRS 2 indexing
grid. The period covered, the duration of the evaluations and the
indicator used, in relation to the season in which the satellite data
was obtained, would generate variations in the rate of change
results (Daramola et al., 2022). Thus the Landsat images selected
4
Regional Studies in Marine Science 55 (2022) 102564
complied with the conditions of 0% cloud cover and no flooding
period. These selection conditions (Fig. 2) resulted in nine images
covering a period of three decades (1989–2021).
The downloaded scenes are accompanied by metadata indicat-
ing the sensor type, the shooting parameters, and the acquisition
date. To prepare and pre-process the Landsat scenes, the suitable
tools of QGIS 3.10 and IDRISI-selva were used.
3.3. Image pre-processing
Image pre-processing allows one to adapt the data to a partic-
ular objective. Landsat images must be radiometrically coherent
A. Boussetta, S. Niculescu, S. Bengoufa et al.
Table 1
Metadata of the used satellite images.
Date acquired Acquisition Satellite images
time and sensor
07 January 1989 09:19:53 Landsat
26 October 1999 09:42:27 Landsat
28 June 2002 09:38:01 Landsat
30 May 2003 09:38:17 Landsat
07 August 2009 09:35:44 Landsat
05 August 2013 09:51:35 Landsat
13 February 2014 09:50:26 Landsat
16 August 2017 09:49:38 Landsat
15 January 2021 09:49:45 Landsat
(pixel values refer to the same physical units) and geometrically
matched (pixels to be compared are in the same area) (Louati
et al., 2015). If these corrections are not made, the user may
misinterpret the ratio values. This misinterpretation could be
related to the differences in the locations and the reflectance of
the land cover types (Singh, 1989; Zhan et al., 2002; Song et al.,
2014).
The pre-processing of Landsat scenes starts with radiometric
correction to minimize the effects of the alteration of radiometric
values generated by the interpolation during the georeferencing
phase. The dark object subtraction model (Chavez, 1996) can
make a rough correction of the atmospheric effect based on the
same satellite image. The necessary additional parameters (solar
elevation and viewing angles, offset/gain) are mentioned in the
metadata. QGIS 3.10 offers, as a preprocessing tool, the Semi-
Automatic Classification plugin, which ensures the attenuation
of the atmospheric effect in one step, the compensation of the
differences in the shooting conditions, and the calibration of the
sensors. Furthermore, the topographic effects are negligible for
low-relief areas (Gu and Gillespie, 1998; Hantson and Chuvieco,
2011), which is the case with Jerba island. In addition, the ratios
method has a reduced sensitivity to the variability of topographic
conditions (Lillesand et al., 2015).
For the geometric correction, the Landsat 8-OLI image of 5
August 2013 was georeferenced using the world geodetic system
(WGS 84) datum and Universal Transverse Mercator zone 32
projection. The georeferencing process was performed based on
topographic maps of 1/50.000 of 1963. The Landsat 8-OLI image
was then used to georeference the remaining Landsat scenes
based on ground control points (GCPs). During each georeferenc-
ing process, more than 30 well-distributed GCPs were carefully
identified in situ from permanent landmarks such as road in-
tersections and coastal constructions using a differential global
positioning system. A first-order polynomial transformation and
nearest-neighbor image resampling were also applied to better
preserve the original radiometric values of the pixels.
3.4. Shoreline detection and extraction
Before undertaking shoreline detection or extraction, it is es-
sential to define a suitable shoreline indicator (Boak and Turner,
2005). For the present study, the limit of the high water level from
satellite images was detected.
The detection method is based on the processing of a series
of multi-date Landsat images covering a period of 32 years. As
mentioned, nine Landsat images covering our study area were
used. The techniques performed are based on a calculation of visi-
ble and mid-infrared band ratios, since the near and mid-infrared
wavelengths are strongly absorbed by water and reflected by
soil (Alesheikh et al., 2007; Cui and Li, 2011). Therefore, the
shoreline can be extracted automatically after constituting the
binary images (sea/land). The band-ratio method is simple, quick,
and does not require operator intervention (Chand and Acharya,
Regional Studies in Marine Science 55 (2022) 102564
Resolution Bands used (µm)
(m)
5 TM
B1: 0,45 - 0,52 (Bleu)
7 ETM+
B2: 0,52 - 0,6 (Green)
7 ETM+
B3: 0,63 - 0,69 (Rouge)
7 ETM+
B5: 1,55 - 1,75 (Mid-Infrared)
5 TM
30*30
8 OLI B2: 0,45 - 0,515 (Bleu)
8 OLI B3: 0,525 - 0,6 (Green)
8 OLI B4: 0,63 - 0,68 (Rouge)
8 OLI B6: 1,56 - 1,66 (Mid-Infrared)
2010). This eliminates the strong level of subjectivity and the
amount of time required for the most used method, manual
digitalization.
In the first step, the ratios between the green and mid-infrared
bands, which are equivalent to Band 5/Band 2 in the case of Land-
sat TM and ETM+ and to Band 6/Band 3 in the case of Landsat-OLI,
were calculated (see (1)). In the second step, a reclassification
was performed to construct a binary image. To extract the target
shoreline, which is the boundary between the sea and land, the
ratios with values less than one (<1) were reclassified as water
(water pixels); otherwise, the values greater than one (>1) were
classified as land (land pixels). The overlay and reclass tools of
the IDRISI system were used respectively to build the selected
ratios and to complete the binarization (Fig. 3).
Green/Mid-infrared (1)
The third step consisted of a vectorization of the binary images,
that is, the conversion of the raster format of the results to poly-
line format. The resulting vector data were exported in Shapefile
format for processing in ArcGIS, where the shorelines were specif-
ically processed. These shorelines have a zigzag pattern, giving a
non-real representation of their geometric nature; a smoothing
process was performed using the Smooth Line tool in ArcGIS.
Each shoreline automatically extracted from the ratio method
was compared to the corresponding reference shoreline that was
manually digitized from the RGB image. This comparison enabled
us to assess the accuracy of the automatic shoreline extraction.
Thus, the distances between the extracted and reference shore-
lines were estimated based on the DSAS and its net shoreline
movement (NSM) parameter.
3.5. Erosion rate estimation
After extraction of the multi-date shoreline, a diachronic anal-
ysis was performed using the DSAS v5.0. This tool is an extension
of ArcGIS software, was developed by the USGS (Himmelstoss
et al., 2018), and is based on the principle of measuring distances
between shorelines and calculating rates of change (in m/year).
The advantage of the DSAS is that it provides the graphical and
statistical evolution of the shoreline while analyzing past, present,
and future change. Thus, the DSAS offers a series of simple tools
to automatically estimate the erosion rates (Nicu, 2021).
The use of the DSAS requires the proper preparation of shore-
lines in vector format within a geodatabase in ArcCatalog. Then,
the analysis consisted of three main steps: (i) the generation
of the baseline, (ii) the creation of the transects, and (iii) an
estimation of the uncertainty and the choice of the statistics for
the calculation of the evolution rates.
The baseline was generated following a buffer zone at an
offshore distance of 200 m from the source shoreline, which
remains parallel and offset to the reference shoreline orientation.
Then, transects were generated with an equidistance of 30 m.
5
A. Boussetta, S. Niculescu, S. Bengoufa et al.
Fig. 3. Shoreline extraction method; (A) band ratioing, (B) reclassification, (C) vectorization, (D) smoothing.
Therefore, calculations of the distances between the points of
intersection of the transects and the shorelines were used for the
calculation of statistical parameters (Fig. 4).
For the actual shoreline evolution estimation, two statistical
parameters were used, the end point rate method (EPR) and the
linear regression rate (LRR). The LRR was chosen to estimate long-
term rates of change over a 32-year period (1989–2021). The
EPR was used to analyze short-term shoreline changes between
1989–1999, 1999–2009, and 2009–2021. The EPR is calculated by
dividing the distance of shoreline movement by the time elapsed
between the oldest and most recent shoreline (Himmelstoss et al.,
2018).
The main advantages of the EPR are the simplicity of the
calculation and the need for only two shoreline dates. On the
other hand, the disadvantage is that, with more than two shore-
lines, additional information is lost (Himmelstoss et al., 2018).
Furthermore, the reliability of the calculated rates of change
is generally proportional to the amount of historical shoreline
data used (Appeaning Addo et al., 2012). For this reason, the
LRR was used, which is the most statistically robust method for
estimating long-term trends in shoreline evolution. The values of
the shoreline rates change were then classified with respect to
the standard interval of Himmelstoss et al. (2018) and displayed
as color nuances projected on hazard maps.
3.6. Errors and uncertainty analysis
Several error sources could be considered for uncertainty anal-
ysis of our approach. The Landsat images used were taken during
the summer period under calm sea conditions, which could ne-
glect the error related to seasonal variations and meteorological
tide. Furthermore, the Mediterranean basin exhibits a micro-tidal
regime (0.2 m) and, given the spatial resolution of the Landsat
images of 30 m, as adopted in previous works (Guariglia et al.,
6
Regional Studies in Marine Science 55 (2022) 102564
2006; Kuleli, 2010; Louati et al., 2015), it is possible to assess
shoreline change from Landsat images, without the interference
of the tidal factor. Considering all of the above, we estimate that
the uncertainty of the used method results from the conjunction
of three types of errors:
(1) The error related to the automatic shoreline extraction
(ASEe) was estimated based on the calculated average distance
between a reference shoreline and the extracted one using the
NSM parameter of the DSAS.
(2) The error related to the image-processing method could be
the result of the georeferencing process, and could be expressed
by the root mean square error (RMSe), which is the distance
(in pixels) between the corrected coordinates and the reference
coordinates. This distance does not exceed 1/2 pixels.
(3) The margin of error related to the calculations of the
shoreline evolution rate, called the uncertainty of the end point
rate (EPRunc), was calculated automatically using the DSAS tool
according to the following equation (Himmelstoss et al., 2018)
over three periods, 1989–1999, 1999–2009, and 2009–2021:
√
(uncyA)2 + (uncyB)2
EPRunc = ± , (2)
dateA − dateB
Where:
uncyA is uncertainty from the attribute field of shoreline A,
uncyB is uncertainty from the attribute field of shoreline B, dateA
is the date of shoreline A (most recent), and dateB is the date of
shoreline B (oldest).
Finally, the total error is given by the square root of the sum of
the squares of the values of the error sources according to Eq. (3)
(Fletcher et al., 2003):
√
E α= RMSe2 + ASEe2 + EPRunc 2 , (3)
A. Boussetta, S. Niculescu, S. Bengoufa et al.
Fig. 4. Distances calculation between the intersection points of transects and shorelines using DSAS tool.
Where:
RMSe = the georeferencing error (root mean square);
ASEe = the automatic shoreline extraction error;
EPRunc = the uncertainty of the end point rate calculation; and
E α = the total error.
4. Results
The results of the present study were derived from a spatio-
temporal analysis of shoreline positions that were automatically
extracted from satellite data processing. This analysis was con-
ducted over a 32-year period subdivided into intervals of around
10 years (1989–2021). Below, the accuracy of the shoreline ex-
traction method is first evaluated. Then, using the EPR and LRR
statistical parameters, the rate of shoreline evolution along the
eastern coast of Jerba island is detailed to interpret the erosion
phenomenon. Finally, the uncertainty analysis of the results is
presented.
7
Regional Studies in Marine Science 55 (2022) 102564
4.1. Assessment of shoreline extraction accuracy
Nine shorelines were extracted automatically from Landsat
images based on the band-ratio method. The accuracy of this
extraction method has been evaluated using a reference shoreline
derived from manual digitization. Table 2 shows the distances
between the nine extracted shorelines and the corresponding
reference shoreline. These distances have been calculated au-
tomatically using the DSAS and the NSM parameter (Fig. 5).
The average distances were used as indicators for analyzing the
extraction accuracy.
Overall, the ratio method generated accurate shoreline detec-
tion (Table 2). We noticed that, for the touristic zone – which is a
straight and homogeneous shore – the method provided the most
accurate shoreline extraction, with average distances of 1.9 m,
1.56 m, 1.97 m, and 1.83 m, respectively, for 1989, 1999, 2009,
and 2021. Over the same period, the average distances recorded
in the Aghir zone were 2.85 m, 2.44 m, 3.4 m, and 3.75 m, respec-
tively. In contrast, along the sandy spit, high average distances of
11.57 m, 6.6 m, 12.2 m, and 11.4 m were observed, respectively,
A. Boussetta, S. Niculescu, S. Bengoufa et al. Regional Studies in Marine Science 55 (2022) 102564

Table 2
NSM statistics basing on DSAS calculation.
Years Sectors Cells Min Mean Max Mean
distance distance distance distance/sector
(m) (m) (m) (m)
Rass Rmel C1 1.37 11.57 23.78 11.57
Touristic zone C2 0.5 2.5 8.9 1.9
1989 C3 0.2 1.9 8.6
C4 0.3 1.41 7.5
Aghir zone C5 0.8 3.1 13.7 2.85
C6 0.9 2.6 6.6
Rass Rmel C1 2.5 6.6 32.1 6.6
Touristic zone C2 0.3 1.2 5.2 1.56
1999 C3 0.19 1.9 7.9
C4 0.5 1.6 7.9
Aghir zone C5 0.81 3.39 9.5 2.44
C6 0.9 3.5 8.1
Rass Rmel C1 0.7 12.2 30.9 12.2
Touristic zone C2 0.6 1.53 8.8 1.97
2009 C3 0.99 2.3 9.2
C4 0.6 2.1 8.3
Aghir zone C5 0.31 3.3 7.4 3.4
C6 0.7 3.5 9
Rass Rmel C1 0.95 11.4 31 11.4
Touristic zone C2 0.5 1.6 9 1.83
2021 C3 0.85 2 8.9
C4 0.6 1.9 7.9
Aghir zone C5 0.4 3.7 6.9 3.75
C6 0.9 3.8 8.2

Fig. 5. Assessment of shoreline extraction accuracy: distances calculation between the extracted and the digitized shoreline using Net Shoreline Movement (NSM)
parameter.

revealing that the ratio method is less accurate for the shoreline
extraction of complex coast types.
4.2. Analysis of shoreline evolution between 1989 and 2021
A spatio-temporal analysis was conducted to estimate the
shoreline change rates and to characterize the erosion of Jerba’s
eastern shoreline. Intervals of roughly the same length were
fixed to estimate the erosion within almost three mid-periods
(1989–1999, 1999–2009, 2009–2021) using the EPR parameter.
Furthermore, to minimize information loss, all the shorelines
8
extracted, which correspond to nine dates (Fig. 7), were used for
the LRR calculation.
The studied coastline was divided into six coastal cells (Fig. 6)
to simplify the estimation of the shoreline evolution and to fa-
cilitate their interpretation, analysis, and presentation. Using the
EPR and LRR values, six classes were defined: high erosion (−2
m/yr), low erosion (−2 to −1 m/yr), stabilized (−1 to 1 m/yr),
low accretion (1 to 2 m/yr), moderate accretion (2 to 3 m/yr), and
high accretion >3 m/yr. We present the results of the EPR values
classification along each zone and cell for the three periods and
A. Boussetta, S. Niculescu, S. Bengoufa et al. Regional Studies in Marine Science 55 (2022) 102564

Fig. 6. The eastern coasts of Jerba island (A) Rass Rmel, (B) Touristic Zone (C) and Aghir zone.

the LRR values classification for the entire period in the following
sections.

4.2.1. Shoreline evolution along Rass Rmel sandy spit

The analysis of the multi-temporal series of Landsat images
allows us to characterize the evolution along the Rass Rmel spit
(Fig. 7). As early as 1989, the spit underwent a very rapid shift to
the northwest. Between 1989 and 1999, it has extended by about
760 m at an average speed of about 76.24 m/year (Fig. 8).
The spit further evolved between 1999 and 2009 and reached
an extent in its point over a distance of approximately 250 m
(Fig. 8). The minimum and maximum evolution speeds were
15.4 m/yr and 27 m/yr, respectively, which translates into an
average annual speed of approximately 25.87 m/yr. Between 2009
and 2021, the average and maximum shoreline evolutions were
200 m and 300 m, respectively. The evolution of the sandy spit
is three times faster between 1989 and 1999 than between 1999
and 2009 and 2009 and 2021 (Fig. 8).
It is evident that, over the entire period studied (from 1989 to Fig. 7. Shorelines extraction results along the Rass Rmel sandy spit.
2021), the Rass Rmel sandy spit recorded a net positive evolution
(accretion of sediment). The LRR clearly indicates this evolution,
giving an annual average speed of about 38.74 m/year (Fig. 8).
9
A. Boussetta, S. Niculescu, S. Bengoufa et al.
Fig. 8. Evolution of the shoreline along the Rass Rmel spit.
4.2.2. Shoreline evolution along the touristic zone
The touristic zone of the north-east coast of Jerba island ex-
tends over 9 km and includes three cells. This zone had an erosion
rate of −3.3 m/year between 1989 and 1999 (Fig. 9 a), was rel-
atively stable between 1999 and 2009 (Fig. 9b), and experienced
slight accretion between 2009 and 2021 (Fig. 9c).
The spatio-temporal analysis of the shoreline evolution along
the first cell (cell 2) of the touristic zone, which extends over
three kilometers between two tourist structures, revealed a very
pronounced erosive trend. Between 1989 and 1999 (Fig. 9a), the
average shoreline retreat was around −3.59 m/year, and the
highest retreat rate (−6.29 m/year) was measured in front of the
Hotel Les Sirènes. Between 1999 and 2009 (Fig. 9b), the erosive
trend continued with a relatively low average rate. The erosion
mainly affected the sandy beaches in front of the Radisson Hotel,
with a maximum shoreline retreat rate of −3.38 m/year. After
the installation of two Geotubes in 2009, these beaches were in
a situation of upstream-drift sedimentation (accretion), with a
maximum rate of +2.75 m/year, and downstream-drift erosion,
with an average rate of −0.7 m/year. Therefore, between 2009
and 2021 (Fig. 9c), the shoreline along this cell was stable, with a
small retreat in the north of the Hotel Les Sirènes (−1.04 m/year)
due to the hotel wall (Fig. 12), which acts as a barrier to the
longshore drift.
The erosive trend has characterized the second cell (cell 3)
of the touristic zone, mainly during the periods of 1989–1999
(Fig. 9a) and 1999–2009 (Fig. 9 b), where the average and min-
imum speeds of the shoreline retreat were −3.59 m and −6.14
m/year and −0.14 m and −1.49 m/year, respectively. Slight ac-
cretion has characterized the 2009–2021 period (Fig. 9c), with a
maximum and average speed of 2.67 m and 1.74 m/year, respec-
tively.
The third cell (Cell 4) was also characterized by an erosive
trend reflected by the retreat of the shoreline, with average and
minimum rates of −2.74 m and −4.87 m/year and −1.27 m
and −4.25 m/year between 1989 and 1999 (Fig. 9a) and 1999
and 2009 (Fig. 9b), respectively. An accretion of sediment was
noticeable between 2009 and 2021 (Fig. 9c) on the northern side
of this zone, with a maximum speed of 4 m/year, and a slight
retreat was found in front of the hotel Dar Jerba, with an average
speed of −2 m/year.
Over the 32-year period, the touristic zone has been signif-
icantly subjected to erosion, with an average LRR indicating an
average rate of shoreline retreat of −1.05 m/yr and a maximum
of −1.06 m/year and a minimum of −2.87 m/year (Fig. 10).
10
Regional Studies in Marine Science 55 (2022) 102564
4.2.3. Shoreline evolution along Aghir zone
This zone extends over 6.5 km and includes two cells (cell
5 and cell 6). Between 1989 and 1999 (Fig. 9a), the retreat of
the shoreline was general for both cells 5 and 6. The average
and minimum rates have reached −4.96 m and −18.27 m/year,
respectively.
For the 1999–2009 period (Fig. 9b), sediment accumulation
was noted between Rass Rouga and the Aghir harbor, with a
maximum shoreline accretion rate of +1.86 m/year. The main
reason for this accretion could be the further implementation of
dykes (Fig. 13 g). There were about twenty dykes for a shoreline
of about 5 km in 1999. However, we noticed that this has caused
an intense shoreline retreat in front of the hostel of Aghir, with
average and maximum rates reaching −2.7 m and −5.4 m/year,
respectively. In fact, the installation of dyke with a 70 m wide
channel has led to a sediment transit toward the south and thus
to an accretion of sediment at the level of the beach located
immediately downstream of the harbor jetty, with a maximum
shoreline evolution rate of +9 m/year. However, this has led to a
retreat of the shoreline at the level of the beaches adjacent to the
Jerba Paradise hotel, with a maximal rate of about −9.38 m/year.
Between 2009 and 2021 (Fig. 9c), an accretion was notice-
able along the whole zone, with a maximum shoreline evolution
rate of +8.58 m/year; accretion was not found in the sector
that extends between Jerba la Douce and Golf Beach, which is
characterized by relative stability.
According to the LRR values for the entire period (1989–2021)
(Fig. 10), a retreat of shoreline position has occurred all along the
southeast coast of the island, with average and minimum rates
reaching −1.7 m and −8.42 m/year, respectively.
4.3. Margin of total error
The erosion rate cannot be estimated without accounting for
the possible sources of error, in other words, the error related due
to the remotely-sensed data processing (RMSe), the error of the
automatic shoreline extraction from the Landsat data based on
ratio methods (ASEe), and the uncertainty of the EPR calculations
(EPRunc).
The RMSe was ±3 m/yr for the 1989–1999 and 1999–2009
periods and ±2.5 m/yr for the 2009–2021 period (Table 3).
It is worth mentioning that the shorelines extracted from
scenes of 2002, 2003, 2013, 2014, and 2017 were included in
the LRR calculations over the study period but excluded from the
A. Boussetta, S. Niculescu, S. Bengoufa et al. Regional Studies in Marine Science 55 (2022) 102564

Fig. 9. Statistical synthesis of shoreline change rate according to EPR parameter and coastal erosion mapping within the mid-periods (a) 1989–1999; (b) 1999–2009
and (c) 2009–2021.

intermediate calculations of the EPR. Therefore, the EPRunc was
automatically estimated from the DSAS (Table 3). The results of
the total uncertainty for each period are presented in Table 3.
Overall, the total error ranges between 3.5 m and 4.6 m except
11
along the sandy spit of Rass Rmel, where the total error ranges
between 9.6 m and 12 m. These variations are especially due
the variability of the error related to the automatic shoreline
extraction.
A. Boussetta, S. Niculescu, S. Bengoufa et al. Regional Studies in Marine Science 55 (2022) 102564

Fig. 10. Long-term statistic synthesis of shoreline change rate according to LRR parameter and coastal erosion mapping (1989–2021).

Table 3
Estimated errors and total uncertainty.
Periods 1989–1999 1999–2009 2009–2021
Rass Touristic Aghir Rass Touristic Aghir Rass Touristic Aghir
Rmel zone zone Rmel zone zone Rmel zone zone
SAEe 9.08 1.73 2.64 9.4 1.76 2.92 11.8 1.9 3.57
RMSe 3 3 3 3 3 3 2.5 2.5 2.5
EPRunc 0.93 0.93 0.93 1.45 1.45 1.45 1.46 1.46 1.46
Eα 9.6 3.58 4.1 9.97 3.76 4.4 12.14 3.46 4.59

5. Discussion which should enable the development of a shared database used

One of the objectives of this paper was to develop an ef-
ficient method for shoreline detection and extraction. Optical
satellite imagery provides regular and consistent observations of
the earth with minimal requirements for individual operators
(Spinosa et al., 2021) and offers the most practical approaches
for detecting shoreline positions at different spatial and tempo-
ral scales (Almonacid-Caballer et al., 2016; Hegde and Akshaya,
2015). This data source has been widely used in various shoreline
studies. Most recently, Bengoufa et al. (2021) used remotely-
sensed data and tested the ability of new machine-learning meth-
ods for shoreline extraction. Even though these methods have
been found to be highly accurate, they are still semi-automatic
and can only be used for a specific coastal type, either a rocky
coast or a sandy coast. In the absence of studies on fully automatic
shoreline extraction, the present work is an original study, as it
evaluates a form of image processing based on the band-ratio
method. This evaluation was undertaken by comparing the results
to a reference shoreline extracted using a digitization method.
This method is widely used for shoreline extraction, however, it
is time consuming and relatively subjective (Liu and Jezek, 2004),
and expertise in the coastal and marine field is required to iden-
tify an adequate shoreline indicator. The band-ratio method was
found to be acceptably accurate and compatible, especially along
homogeneous and straight shorelines such as sandy beaches, but
had some limits when used to extract the shoreline of the sandy
spit. The band-ratio method is also simple and quick to use and
requires no operator intervention. A large panel of users could
reproduce this method in an automatic and standardized way,
12
for efficient coastal monitoring.
Remote-sensing satellite data in combination with GIS tools
contribute to the continuous short- and long-term monitoring of
the highly dynamic features of the shoreline (Baral et al., 2018)
and accurately characterize coastal erosion (Kaliraj et al., 2015).
Using these tools, the coastal erosion along Jerba island was
characterized via a diachronic analysis of the shoreline evolution.
Given the absence of updated recent studies and the importance
of the results obtained, this study could remedy the serious lack
of knowledge regarding the basic elements (statistics) of coastal
erosion and accretion, which we detailed for the most vulnerable
areas of Jerba island (eastern coast) for the first time. This kind of
element together with field data are valuable for coastal zone and
island-environment managers in developing appropriate sustain-
able management plans and mitigating the vulnerability of sandy
coasts (Natarajan et al., 2021).
Our study area demonstrated a clear erosive trend, with the
exception of the Rass Rmel spit, which has experienced signifi-
cant sedimentary accretion (+38.74 m/year between 1989–2021)
(Fig. 8). These results cohere with stereoscopic observations of
aerial photographs, which indicate that the length of the spit
increased by 950 m between 1988 and 1999 (Bourgou, 2005).
This increase entails an accretion rate equal to 86.36 m/year.
According to Oueslati (2004), the spit gained 1,500 m between
1985 and 1991, which thus increased its length to about 7.360
km. These findings are similar to Khlifi (2005), who indicated that
the annual rate of extension is about 90 m/year (Gzam et al.,
2009). The sandy beaches have been subjected to a high rate
of shoreline retreat over the last three decades (1989–2021).
A. Boussetta, S. Niculescu, S. Bengoufa et al. Regional Studies in Marine Science 55 (2022) 102564

Strong erosion of the sandy beaches along the touristic and Aghir
zones was recorded during the 1989–1999 period, which explains
why the sandy spit extended three times faster over this same
period than during the periods of 1999–2009 and 2009–2021
(Fig. 8). This can be demonstrated by the sedimentary transit
provided by the northward longshore drift, which is estimated at
40,000 m3 /year according to the Central Company for Territorial
Equipment (CCTE, 1999). Measurements made between 1999 and
2000 in the touristic zone show that the coastal transit is approx-
imately 23,677 m3 /year (Roberge, 2005). The most recent study
estimated the annual input provided by the longshore drift to be
30,000 m3 (Gzam et al., 2009).
The DSAS results sufficiently conform to the real-world situ-
ation. The results provide a detailed overview of the shoreline
evolution and account for local erosion trends. These results can
support researchers in many investigations of coastal and espe-
cially insular environments. We noticed alarming signs of marine
erosion through our field surveys. Among the observed signs, we
found the roots of the coastal palm trees were denuding (Fig. 13b
and 13d). These roots used to be a natural stabilizer of the dunes
and the soil, preventing erosion and protecting the old coastal
road (Figs. 13c and 13f). Furthermore, the significant erosion of
the upper beach has led to the denudation of the upper walls,
which are now beaten by sea waves at high tide; this has led to
cracks and deformation by subsidence and even traces of basal
scouring (Fig. 13a).
In addition, coastal erosion threatens historical monuments,
such as the monument of Sidi Garous, of which only ruins remain
(Fig. 13e). Moreover, with the disappearance of the beach in front
of the very famous marabout monument of Sidi Bakour (Fig. 11a),
the monument is now entirely underwater. The Sidi Bakour mon-
ument was slightly destroyed by waves in 1979 (Oueslati, 1986),
still standing in 1991, and collapsed into the sea in 1993 (Fig. 11b)
(Oueslati, 2004). Today, it has completely disappeared. A memo-
rial monument was installed in 2005 further away from the
shoreline (on the top of the beach; Fig. 11c). Field observations
in 2018 and 2019, show that this memorial monument is subject
to swash processes and located in the intertidal zone (Fig. 11c). Fig. 11. Monument of Sidi bakour ; (a) the existence of the monument in 1976
(Paskoff, 2004), (b) the affected monument in 1991 (Oueslati, 2004), (c) The
Furthermore, the degradation of the sandy beaches is leading disappearance of the monument and the implementation of memorial of Sidi
to the destruction of the hotel infrastructures. Several hotels are Bakour (photo taken on 09/08/2018).
threatened or have already closed and have lost their beach. This
is the case of the Les Sirènes (Fig. 12). Riprap and a protector
wall have been installed near the Les Sirènes hotel to absorb tourist activities have stopped the transport of continental sand
the impact of the marine forcing. However, these constructions that naturally nourishes the beaches. Hoteliers have continued to
accelerated the erosion of the adjacent beaches. clear the Posidonia leaves of these beaches, which has led to a
The erosion in general could be explained by the accelerated reduction in the braking effect of the waves.
rise in sea level, which is an evident consequence of climate The shoreline of Jerba Iisland is therefore constantly altered
change. The relationship between SLR and sandy beach erosion due to the combined influence of the natural environment and
is very strong. In fact, this link is highly multiplicative. According human activities. The decline of these beaches is very serious and
to studies of Leatherman et al. (2000) coastal erosion of sandy worrying. The situation is becoming a critical environmental crisis
beaches is expected to be 150 times the rate of sea level rise. In requiring rapid intervention.
our case, a sustained sea level rise of 5.7 m (Pirazzoli, 2021) could
result in 0.85 m of shoreline erosion. While the recorded erosion 6. Conclusions
in our study area is less than 1 m. Therefore, coastal erosion rates
are expected to be more than 150 times the rate of sea level The rapid retreat of the shoreline, which demonstrates the
rise for disturbed sandy beaches. Future coastal recession is then hazard of coastal erosion, represents a real problem in most
expected to be largely contributed by sea level rise (Athanasiou coastal zones. Island environments are particularly affected, as
et al., 2020; Ranasinghe, 2016). is the case for the sandy beaches of the eastern coast of Jerba
In addition to the effect of the SLR, erosion could be accentu- island in Tunisia. It is now necessary to monitor the shoreline to
ated by the subsidence of the Gulf of Gabes (Miossec and Paskoff, define the exposed areas and to anticipate the erosion process.
1979). Moreover, these beaches are made up of residual stocks The present study, which is based on remote-sensing techniques,
inherited from ancient times that are not renewed and therefore falls within this perspective and addresses the lack of updated
exhibit a critical sedimentary deficit (Oueslati, 2004). Neverthe- studies.
less, it is anthropogenic actions that have seriously exacerbated A patio-temporal analysis of the shoreline position has been
the situation since the 1960s. The destruction of coastal dunes, performed to characterize the coastal erosion along eastern coast
the establishment of hotel units on the seafront, and balneal and of Jerba island. The shoreline was extracted automatically from
13
A. Boussetta, S. Niculescu, S. Bengoufa et al.
Fig. 12. The total disappearance of the beach in front les Sirenes Hotel, (a) photo taken in July 1992 and the beach measurements were established by Oueslati
(2004) (b) Hotel Les Sirènes closed (Photo taken on 09/08/2018).
multi-temporal, multi-sensor, Landsat satellite images covering a
32-year period (1989–2021). After extensive radiometric and geo-
metric corrections, and the application of the band-ratio method,
nine shorelines from different times were extracted (1989, 1999,
2002, 2003, 2009, 2013, 2014, 2017, and 2021). Three short-
term trends and one long-term trend were studied. The analysis
revealed that the study area is undergoing a massive erosion
process (89%) in the long term (1989–2021). The transect-based
analysis of the shoreline evolution revealed spatio-temporal vari-
ability among the different sites studied. This analysis allowed us
to distinguish two types of situations: certain sites are experi-
encing generalized erosion (the tourist area and Aghir) and other
sites are experiencing progradation (the Rass Rmel sandy spit).
In several spots, the erosion has become almost irreversible, and
reestablishing stability will not be easy.
The Landsat image processing based on the band-ratio method
was found to be an accurate method for fully extracting the
shoreline automatically. Therefore, a large panel of users could
reproduce this method for the updating of an eventual shared
database of coastal monitoring. The monitoring of the shoreline
changes, after the estimation of the global error, produced results
in conformity with the real-world situation. The results provide
a detailed overview of the shoreline’s evolution, account for local
erosion trends, and can thus help many investigations of coastal
and especially insular areas. These data would then facilitate fu-
ture analyses of the vulnerability of hotel infrastructure to climate
change and sea level rise. Due to the attractive nature of Jerba
island from a touristic point of view and the related socioeco-
nomic issues this attractiveness entails, the results of this study
should be seriously considered with respect to the development
of much-needed coastal management and risk-mitigation plans.
14
Regional Studies in Marine Science 55 (2022) 102564
CRediT authorship contribution statement
Amina Boussetta: Conceptualization, Acquisition of data,
Methodology, Software, Writing – original draft, Writing,
Writing – review & editing. Simona Niculescu: Resources,
Visualization, Funding acquisition, Writing – review & editing.
Soumia Bengoufa: Validation, Writing – original draft, Writing
– review & editing. Mohamed Faouzi Zagrarni: Supervision,
Project administration, Writing – review & editing.
Declaration of competing interest
The authors declare the following financial interests/personal
relationships which may be considered as potential competing
interests: Simona NICULESCU reports financial support was
provided by ISblue. Simona NICULESCU reports a relationship
with ISblue that includes: non-financial support.
Data availability
No data was used for the research described in the article.
Acknowledgments
The authors are very grateful to the Jlij Association for the Ma-
rine Environment AJEM, for their assistance in the field surveys.
The authors are very grateful to the program of ERASMUS+ Chair
Jean Monnet European Spatial Studies of Sea and Coastal zones —
599967-EPP-1-2018-1-FR-EPPJMO-CHAIR.
A. Boussetta, S. Niculescu, S. Bengoufa et al. Regional Studies in Marine Science 55 (2022) 102564

Fig. 13. Coastal erosion along the eastern coast of Jerba (a) (b) (c) signs of erosion (d) commemorative monument of Sid Garous (e) the roots of the palm trees (f)
The old coastal road in direct contact with the sea (g) multiplication of protection dykes.

Formatting of funding sources References

Alesheikh, A.A., Ghorbanali, A., Nouri, N., 2007. Coastline change detection using
This work was supported by ISblue project, Interdisciplinary remote sensing. Int. J. Environ. Sci. Technol. 4, 61–66. http://dx.doi.org/10.
1007/BF03325962.
graduate school for the blue planet (ANR-17-EURE-0015) and
Almonacid-Caballer, J., Sánchez-García, E., Pardo-Pascual, J.E., Balaguer-
co-funded by a grant from the French government under the Beser, A.A., Palomar-Vázquez, J., 2016. Evaluation of annual mean shoreline
program ‘‘Investissements d’Avenir’’. position deduced from Landsat imagery as a mid-term coastal evolution

15
A. Boussetta, S. Niculescu, S. Bengoufa et al.
indicator. Mar. Geol. 372, 79–88. http://dx.doi.org/10.1016/j.margeo.2015.12.
015.
Amrouni, O., Hzami, A., Heggy, E., 2019. Photogrammetric assessment of shore-
line retreat in North Africa: Anthropogenic and natural drivers. ISPRS J.
Photogramm. Remote Sens. 157, 73–92. http://dx.doi.org/10.1016/j.isprsjprs.
2019.09.001.
Anfuso, G., Martínez del Pozo, J.A., 2005. Towards management of coastal erosion
problems and human structure impacts using GIS tools: case study in Ragusa
Province, Southern Sicily, Italy. Environ. Geol. 48, 646–659. http://dx.doi.org/
10.1007/s00254-005-1322-2.
Appeaning Addo, K., Jayson-Quashigah, P.N., Kufogbe, K.S., 2012. Quantitative
analysis of shoreline change using medium resolution satellite imagery in
Keta, Ghana. MS 1, 1–9. http://dx.doi.org/10.5923/j.ms.20110101.01.
Athanasiou, P., van Dongeren, A., Giardino, A., Vousdoukas, M.I., Ranasinghe, R.,
Kwadijk, J., 2020. Uncertainties in projections of sandy beach erosion due
to sea level rise: an analysis at the European scale. Sci. Rep. 10, 11895.
http://dx.doi.org/10.1038/s41598-020-68576-0.
Atsri, H.K., Konko, Y., Cuni-Sanchez, A., Abotsi, K.E., Kokou, K., 2018. Changes
in the West African forest-savanna mosaic, insights from central Togo. PLoS
One 13, e0203999. http://dx.doi.org/10.1371/journal.pone.0203999.
Awange, J.L., Saleem, A., Konneh, S.S., Goncalves, R.M., Kiema, J.B.K., Hu, K.X.,
2018. Liberia’s coastal erosion vulnerability and LULC change analysis: Post-
civil war and Ebola epidemic. Appl. Geogr. 101, 56–67. http://dx.doi.org/10.
1016/j.apgeog.2018.10.007.
Baral, R., Pradhan, S., Samal, R.N., Mishra, S.K., 2018. Shoreline change analysis
at Chilika Lagoon Coast, India using digital shoreline analysis system. J.
Indian Soc. Remote Sens. 46, 1637–1644. http://dx.doi.org/10.1007/s12524-
018-0818-7.
Barragán, J.M., de Andrés, M., 2015. Analysis and trends of the world’s coastal
cities and agglomerations. Ocean Coast. Manag. 114, 11–20. http://dx.doi.org/
10.1016/j.ocecoaman.2015.06.004.
Bengoufa, S., Niculescu, S., Mihoubi, M.K., Belkessa, R., Rami, A., Rabehi, W.,
Abbad, K., 2021. Machine learning and shoreline monitoring using optical
satellite images: case study of the Mostaganem shoreline, Algeria. J. Appl.
Rem. Sens. 15, http://dx.doi.org/10.1117/1.JRS.15.026509.
Bera, R., Maiti, R., 2019. Quantitative analysis of erosion and accretion (1975–
2017) using DSAS — A study on Indian Sundarbans. Reg. Stud. Mar. Sci. 28,
100583. http://dx.doi.org/10.1016/j.rsma.2019.100583.
Boak, E.H., Turner, I.L., 2005. Shoreline definition and detection: A review. J.
Coast. Res. 214, 688–703. http://dx.doi.org/10.2112/03-0071.1.
Bouchahma, M., Barhoumi, W., Yan, W., Wardi, H.Al., 2017. Optical-flow-based
approach for the detection of shoreline changes using remote sensing data.
In: 2017 IEEE/ACS 14th International Conference on Computer Systems
and Applications (AICCSA). Presented At the 2017 IEEE/ACS 14th Interna-
tional Conference on Computer Systems and Applications (AICCSA). IEEE,
Hammamet, pp. 184–189. http://dx.doi.org/10.1109/AICCSA.2017.173.
Bouchahma, M., Yan, Wanglin, 2012. Long-term coastal changes detection system
based on remote sensing and image processing around an island. In: 2012
20th International Conference on Geoinformatics. Presented At the 2012 20th
International Conference on Geoinformatics. IEEE, Hong Kong, China, pp. 1–5.
http://dx.doi.org/10.1109/Geoinformatics.2012.6270334.
Bouchahma, M., Yan, W., 2014. Monitoring shoreline change on Djerba Island
using GIS and multi-temporal satellite data. Arab. J. Geosci. 7, 3705–3713.
http://dx.doi.org/10.1007/s12517-013-1052-9.
Bourgou, M., 2005. Les plages: impact des aménagements touristiques et portu-
aires sur leur évolution récente : exemples tunisiens. Répablique tunisienne,
Ministère de la recherche scientifique et de la technologie et du développe-
ment de compétences, Centre d’etudes et de recherches economiques et
sociales.
Brown, S., Nicholls, R.J., Lowe, J.A., Hinkel, J., 2016. Spatial variations of sea-
level rise and impacts: An application of DIVA. Clim. Change 134, 403–416.
http://dx.doi.org/10.1007/s10584-013-0925-y.
Cazenave, A., Cozannet, G.L., 2014. Sea level rise and its coastal impacts. Earth’s
Future 2, 15–34. http://dx.doi.org/10.1002/2013EF000188.
Chand, P., Acharya, P., 2010. Shoreline change and sea level rise along coast of
bhitarkanika wildlife sanctuary. In: Orissa: An Analytical Approach of Remote
Sensing and Statistical Techniques, vol. 1. p. 21.
Chang, Jun, Liu, Gaohuan, Huang, Chong, Xu, Lirong, 2005. Remote sensing
monitoring on coastline evolution in the Yellow River Delta since 1976.
In: Proceedings. 2005 IEEE International Geoscience and Remote Sensing
Symposium, 2005. IGARSS ’05. Presented At the 2005 IEEE International
Geoscience and Remote Sensing Symposium, 2005. IGARSS ’05. IEEE, Seoul,
Korea, pp. 2161–2164. http://dx.doi.org/10.1109/IGARSS.2005.1526446.
Chapoutot, Jean Mohamed Medhi, 2011. Profil de durabilité dans quelques
destinations touristiques méditerranéennes (Rapport final).
Chavez, P.S., 1996. Image-based atmospheric corrections - revisited and
improved. 12.
Cui, B.-L., Li, X.-Y., 2011. Coastline change of the Yellow River estuary and
its response to the sediment and runoff (1976–2005). Geomorphology 127,
32–40. http://dx.doi.org/10.1016/j.geomorph.2010.12.001.
16
Regional Studies in Marine Science 55 (2022) 102564
Dada, O.A., Li, G., Qiao, L., Ding, D., Ma, Y., Xu, J., 2016. Seasonal shoreline
behaviours along the arcuate Niger Delta coast: Complex interaction between
fluvial and marine processes. Cont. Shelf Res. 122, 51–67. http://dx.doi.org/
10.1016/j.csr.2016.03.002.
Daramola, S., Li, H., Otoo, E., Idowu, T., Gong, Z., 2022. Coastal evolution
assessment and prediction using remotely sensed front vegetation line along
the Nigerian Transgressive Mahin mud coast. Reg. Stud. Mar. Sci. 102167.
http://dx.doi.org/10.1016/j.rsma.2022.102167.
Fletcher, Charles, et al., 2003. Mapping shoreline change using digital
orthophotogrammetry on Maui, Hawaii. J. Coastal Res..
Garcia Rubio, G., Huntley, D., Russell, P., 2014. Evaluating shoreline identification
using optical satellite images. Mar. Geol. 359, http://dx.doi.org/10.1016/j.
margeo.2014.11.002.
Gu, D., Gillespie, A., 1998. Topographic normalization of landsat TM images
of forest based on subpixel sun–canopy–sensor geometry. Remote Sens.
Environ. 64, 166–175. http://dx.doi.org/10.1016/S0034-4257(97)00177-6.
Guariglia, A., Buonamassa, A., Losurdo, A., Saladino, R., Trivigno, M.L., Za-
ccagnino, A., Colangelo, A., 2006. A multisource approach for coastline
mapping and identification of shoreline changes 10.
Gzam, M., Souissi, R., Choura, M., Jedoui, Y., 2009. Caractérisation mor-
phologique et sédimentologique de la côte orientale de l’île de Jerba (Sud-Est
tunisien). In: Edition 1, Hammamet, Tunisie. Presented At the Conférence
Méditerranéenne Côtière Et Maritime - Coastal and Maritime Mediterranean
Conference, Editions Paralia. pp. 99–102. http://dx.doi.org/10.5150/cmcm.
2009.025-2.
Hallegatte, S., Green, C., Nicholls, R.J., Corfee-Morlot, J., 2013. Future flood losses
in major coastal cities. Nature Clim. Change 3, 802–806. http://dx.doi.org/10.
1038/nclimate1979.
Hantson, S., Chuvieco, E., 2011. Evaluation of different topographic correction
methods for Landsat imagery. Int. J. Appl. Earth Obs. Geoinf. 13, 691–700.
http://dx.doi.org/10.1016/j.jag.2011.05.001.
Hegde, A.V., Akshaya, B.J., 2015. Shoreline transformation study of Karnataka
Coast: Geospatial approach. Aquat. Proc. 4, 151–156. http://dx.doi.org/10.
1016/j.aqpro.2015.02.021.
Himmelstoss, E.A., Henderson, R.E., Kratzmann, M.G., Farris, A.S., 2018. Digital
Shoreline Analysis System (DSAS) Version 5.0 User Guide (USGS Numbered
Series No. 2018–1179), Digital Shoreline Analysis System (DSAS) Version
5.0 User Guide. Open-File Report. U.S. Geological Survey, Reston, VA, http:
//dx.doi.org/10.3133/ofr20181179.
Hřebíček, J., Jensen, S., Steenmans, C., 2015. The framework for environmental
software systems of the European environment agency. In: Denzer, R., Ar-
gent, R.M., Schimak, G., Hřebíček, J. (Eds.), Environmental Software Systems.
Infrastructures, Services and Applications, IFIP Advances in Information and
Communication Technology. Springer International Publishing, Cham, pp.
44–55. http://dx.doi.org/10.1007/978-3-319-15994-2_4.
IHEE, 2015. ETUDE D’IMPACT sur L’ENVIRONNEMENT (EIE) DU PROJET de
REALISATION de la STATION de DESSALEMENT D’EAU de MER a DJERBA
(Rapport Définitif). SONED, Tunis, Tunisie.
Islam, Md.S., Uddin, Md. A., Hossain, M.A., 2021. Assessing the dynamics of
land cover and shoreline changes of Nijhum Dwip (Island) of Bangladesh
using remote sensing and GIS techniques. Reg. Stud. Mar. Sci. 41, 101578.
http://dx.doi.org/10.1016/j.rsma.2020.101578.
ITALCONSULT, 1975. Programme D’Infrastructure Pour Le Développement Du
Tourisme En Tunisie. Ministère du tourisme et de l’Aménagement du
territoire, Tunis.
Kaliraj, S., Chandrasekar, N., Magesh, N.S., 2015. Evaluation of coastal erosion
and accretion processes along the southwest coast of Kanyakumari, Tamil
Nadu using geospatial techniques. Arab J. Geosci. 8, 239–253. http://dx.doi.
org/10.1007/s12517-013-1216-7.
Khlifi, I., 2005. Etude d’un modéle de sédimentation littorale active: cas de la
fléche littorale Ras R’Mel-Jerba. Université de Sfax. Master.
Konko, Y., Okhimambe, A., Nimon, P., Asaana, J., Rudant, J.P., Kokou, K., 2020.
Coastline change modelling induced by climate change using geospatial
techniques in Togo (West Africa). ARS 09, 85–100. http://dx.doi.org/10.4236/
ars.2020.92005.
Konko, Y., Rudant, J.-P., Kokou, K., 2016. Contribution of Remote Sensing and
GIS to the Integrated Management of Community Forest Resources in the
Bas-Mono Valley (South-East Togo). http://dx.doi.org/10.13140/RG.2.2.20280.
70403.
Kuleli, T., 2010. Quantitative analysis of shoreline changes at the Mediterranean
Coast in Turkey. Environ. Monit. Assess. 167, 387–397. http://dx.doi.org/10.
1007/s10661-009-1057-8.
Leatherman, S.P., Zhang, K., Douglas, B.C., 2000. Sea level rise shown to drive
coastal erosion. EOS Trans. Am. Geophys. Union 81, 55–57. http://dx.doi.org/
10.1029/00EO00034.
Lillesand, T., Kiefer, R.W., Chipman, J., 2015. Remote Sensing and Image
Interpretation. John Wiley & Sons.
Liu, H., Jezek, K.C., 2004. Automated extraction of coastline from satellite
imagery by integrating Canny edge detection and locally adaptive thresh-
olding methods. Int. J. Remote Sens. 25, 937–958. http://dx.doi.org/10.1080/
0143116031000139890.
A. Boussetta, S. Niculescu, S. Bengoufa et al.
Louati, M., Saïdi, H., Zargouni, F., 2015. Shoreline change assessment using
remote sensing and GIS techniques: a case study of the medjerda delta coast,
Tunisia. Arab J. Geosci. 8, 4239–4255. http://dx.doi.org/10.1007/s12517-014-
1472-1.
Masson-Delmotte, Valérie, Zhai, Panmao, Pirani, Anna, Connors, Sarah L.,
Péan, Clotilde, Chen, Yang, Goldfarb, Leah, Gomis, Melissa I.,
Robin Matthews, J.B., Berger, Sophie, Huang, Mengtian, Yelekçi, Ozge,
Yu, Rong, Zhou, Baiquan, Lonnoy, Elisabeth, Maycock, Thomas K.,
Waterfield, Tim, Leitzell, Katherine, Caud, Nada, 2021. Climate Change
2021 the Physical Science Basis. Front cover artwork, Switzerland.
Miossec, J.-M., Paskoff, R., 1979. Evolution des plages et aménagements touris-
tiques à Jerba (Tunisie): Le cas du littoral nord-est de l’île. Medit 35, 99–106.
http://dx.doi.org/10.3406/medit.1979.1903.
Natarajan, L., Sivagnanam, N., Usha, T., Chokkalingam, L., Sundar, S., Gowrap-
pan, M., Roy, P.D., 2021. Shoreline changes over last five decades and
predictions for 2030 and 2040: a case study from cuddalore, southeast coast
of India. Earth Sci. Inform. 14, 1315–1325. http://dx.doi.org/10.1007/s12145-
021-00668-5.
Nicu, I.C., 2021. Is digital shoreline analysis system fit for gully erosion
assessment? CATENA 203, 105307. http://dx.doi.org/10.1016/j.catena.2021.
105307.
Oueslati, A., 1986. Jerba Et Kerkna, îLes de la Côte Orientale de la Tunisie:
Leur évolution GéOmorphologique Au Cours Du Quaternaire. Publications de
l’Université de Tunis, Tunis.
Oueslati, A., 2004. Littoral et aménagement en Tunisie: des enseignements
de l’expérience du vingtième siècle et de l’approche géoarchéologique à
l’enquête prospective. Publications de la Faculté des Sciences Humaines et
Sociales. Tunis, Tunisie.
Paskoff, R., 2004. L’érosion des côtes: le cas des plages de l’île de Jerba (Tunisie).
La Houille Blanche 4, 8–51. http://dx.doi.org/10.1051/lhb:200401006.
Peduzzi, P., Velegrakis, A.F., Estrella, M., Chatenoux, B., 2013. Integrating the
Role of Ecosytems in Disaster Risk and Vulnerability Assessments: Lessons
from the Risk and Vulnerability Assessment Methodology Development
Project (RiVAMP) in Negril Jamaica. The role of Ecosystems in Disaster Risk
Reduction 109–139.
17
Regional Studies in Marine Science 55 (2022) 102564
Pirazzoli, P.A., 2021. Secular trends of relative sea-level (RSL) changes indicated
by tide-gauge records. J. Coast. Res. 27.
Quadrado, G.P., Dillenburg, S.R., Goulart, E.S., Barboza, E.G., 2021. Historical
and geological assessment of shoreline changes at an urbanized embayed
sandy system in Garopaba, Southern Brazil. Reg. Stud. Mar. Sci. 42, 101622.
http://dx.doi.org/10.1016/j.rsma.2021.101622.
Ranasinghe, R., 2016. Assessing climate change impacts on open sandy coasts:
A review. Earth-Sci. Rev. 160, 320–332. http://dx.doi.org/10.1016/j.earscirev.
2016.07.011.
Roberge, C., 2005. Estimation Du Facteur de Frottement Et Du Transport Sédi-
mentaire Par la Houle sur Le Littoral Nord-Est de Jerba. Université du Québec,
Institut national de la recherche scientifique.
Singh, A., 1989. Review Article Digital change detection techniques using
remotely-sensed data. Int. J. Remote Sens. 10, 989–1003. http://dx.doi.org/
10.1080/01431168908903939.
Song, C., Huang, B., Ke, L., Richards, K.S., 2014. Remote sensing of alpine lake
water environment changes on the Tibetan Plateau and surroundings: A
review. ISPRS J. Photogramm. Remote Sens. 92, 26–37. http://dx.doi.org/10.
1016/j.isprsjprs.2014.03.001.
Spinosa, A., Ziemba, A., Saponieri, A., Damiani, L., El Serafy, G., 2021. Remote
sensing-based automatic detection of shoreline position: A case study in
apulia region. JMSE 9 (575), http://dx.doi.org/10.3390/jmse9060575.
Toure, S., Diop, O., Kpalma, K., Maiga, A., 2019. Shoreline detection using optical
remote sensing: A review. IJGI 8 (75), http://dx.doi.org/10.3390/ijgi8020075.
Yamano, H., Shimazaki, H., Matsunaga, T., Ishoda, A., McClennen, C., Yokoki, H.,
Fujita, K., Osawa, Y., Kayanne, H., 2006. Evaluation of various satellite sensors
for waterline extraction in a coral reef environment: Majuro Atoll, Marshall
Islands. Geomorphology 82, 398–411. http://dx.doi.org/10.1016/j.geomorph.
2006.06.003.
Zhan, X., Sohlberg, R.A., Townshend, J.R.G., DiMiceli, C., Carroll, M.L., Eastman, J.C.,
Hansen, M.C., DeFries, R.S., 2002. Detection of land cover changes using
MODIS 250 m data. Remote Sens. Environ. 83, 336–350. http://dx.doi.org/
10.1016/S0034-4257(02)00081-0.
Zhang, K., Douglas, B.C., Leatherman, S.P., 2004. Global warming and coastal
erosion. Clim. Change 64, 41. http://dx.doi.org/10.1023/B:CLIM.0000024690.
32682.48.