Content uploaded by Nguyen Viet Thanh
Author content
All content in this area was uploaded by Nguyen Viet Thanh on May 26, 2020
Content may be subject to copyright.
Indian Journal of Geo Marine Sciences
Vol. 49 (02), February 2020, pp. 207-217
Prediction of shoreline changes in Almanarre beach using geospatial techniques
Minh Tuan Vu1,2, Yves Lacroix1,*, Van Van Than3, & Viet Thanh Nguyen4
1SEATECH, University of Toulon, La Valette du Var, 83162, France
2National University of Civil Engineering, Hanoi, 100 000, Vietnam
3Thuyloi University, Hanoi, 100 000, Vietnam
4University of Transport and Communications, Hanoi, 100 000, Vietnam
*[E-mail: yves.lacroix@univ-tlf.fr]
Received 17 February 2017; revised 25 April 2017
In this study, the remote sensing and Geographic Information System (GIS) techniques coupled with the Digital
Shoreline Analysis System (DSAS) is applied to detect the historical shoreline changes as well as to predict the future
shoreline position along Almanarre beach which is being threatened by severe erosion. The results show that Almanarre
beach suffered erosion with an average annual change rate of about -0.24 m/year over the period of 1973-2015. The most
severe erosion was observed near Landmark B17 with the maximum erosion rate of -0.86 m/year. Moreover, the shoreline
change in 2020 and 2050 are predicted at approximately -0.05 m/year and -0.22 m/year, respectively. The areas around
Landmarks B06-08 and Landmarks B16-18 will be eroded with the maximum recession rates of -0.89 m/year and -0.94
m/year, respectively. This research proves that the combination of geospatial techniques and numerical model can be a
reliable approach for investigating the shoreline change trend.
[Keywords: Accretion; DSAS; Erosion; Remote sensing; Shoreline extraction]
Introduction
The shoreline change is mostly controlled by natural
causes or anthropogenic intervention, or both of them.
The primary natural factors include waves, winds,
currents, storm surge, sea level rise, and
geomorphologic changes resulting from man-made
factors through coastal construction, mining of beach
sand, dredging of seabed sand, dam construction, or
deforestation. The shoreline change results from
coastal erosion or accretion which greatly affects
human life along the coastal zone. Correspondingly,
the detection and prediction of shoreline change is an
essential task for protection of infrastructure and
coastal zone management1.
Until now, researchers have developed some
approaches to investigate shoreline changes which can
be divided into five categories. Firstly, historical map
charts can reveal a historic information that is
unavailable from other data sources, but many potential
errors associated with these historical coastal records
occur. Secondly, conventional field surveying can
achieve high accuracy of measurement, but is labour
intensive, time consuming and involves high cost1.
Thirdly, aerial photographs can provide sufficient
pictorial information. However, the frequency of data
acquisition is not enough, temporal coverage is limited
by depending on the flight path of the fixed-wing
airplane; the photogrammetric procedure is costly and
time consuming. Additionally, errors in shoreline
interpretation may be introduced by the minimal
spectral range of these sources2. At present, with rapid
development of computational technology, the
numerical models have become increasingly popular
and have been successfully used to detect and predict
the past and future shoreline evolution. They can be
validated easily by measuring and observing the actual
behaviour of the natural systems under the impacts of
the wave climate, mean sea level, coastal defence
works, etc. For instance, the project effects on the tidal
current, tidal volume, sediment concentration and
morphological evolution in Xiaomiaohong tidal
channel, Jiangsu Coast, China were investigated by a
two-dimensional numerical model and the result
demonstrated that the change of morph-dynamic
processes due to the reclamation project has only
occurred near the project area and overall channel
evolution was not significantly affected by this
project3. Furthermore, Zheng et al.4 developed a new
model including interactions between waves and
undertow and an empirical time-dependent turbulent
eddy viscosity formulation that accounts for the phase
dependency of turbulence on flow velocity and
INDIAN J. MAR. SCI., VOL. 49, NO. 02, FEBRUARY 2020
208
acceleration. Gu et al.5 introduced the Near CoM
model coupling a Simulating Waves Nearshore model
(SWAN), a nearshore circulation model
(SHORECIRC), and a sediment transport model in a
fully parallelized computational environment. The
simulation results indicated that two factors contribute
to the final double-sandbar morphology in different
ways. Waves determine the final sandbar morphology,
regardless of the antecedent bathymetry for energetic
waves with high angles of wave incidence, whereas the
pronounced morphological variability (crescentic
pattern of sandbars) will control the evolution, which
remarkably enhances the existing morphological
patterns regardless of the changed wave condition for
moderate waves with small angles of wave incidence.
Nevertheless, using numerical models can be difficult
for the establishment of boundary conditions,
calibration coefficients and parameters regarding
variables representing the reality of the system, that
may generate the fail results6. In addition, the
numerical modelling calculations are more time
consuming than analytical and statistical calculations as
well as the area of study domain are limited by the
capacity of computer. Over the recent decades, remote
sensing techniques are widely used and are more
attractive as they have large ground coverage, are less
time-consuming, inexpensive to implement, and also
satisfactory acquisition repetition. Therefore, this
technology becomes an effective solution for
monitoring shoreline changes7.
The shoreline of Almanarre beach lying on the
western part of Giens tombolo, South France
(Fig. 1(a)) has suffered both accretion and erosion
processes caused by natural factors such as waves,
winds and storm surge, or by human activities. By
studying the submarine sediment logy, Blanc8 showed
that the waves caused the rapid degradation of the
sandy spit in the western branch. Moreover, he
uncovered the erosion at the sea bottom and the
regression of Posidonia seagrass in the Giens gulf.
Grissac9 presented the research about dynamic
sedimentology of Giens and Hyères bay. Through the
in situ granulometric experiments as well as the
analysis of wind and wave fields, he showed that there
are two main longshore currents in the Gulf of Giens.
The northern currents direct from west to east and from
north to south, whilst the southern currents direct from
south to north, then they follow the sandy spit to reach
its centre. In addition, from the visual comparison of
aerial photographs between 1955 and 1972, he also
highlighted the significant decline of the western
branch (down from 50 to 80 m in the center and 75-90
m in the south part) associated mainly with the
degradation of Posidonia seagrass. Courtaud10 used the
aerial photographs and some field surveys to study the
shoreline evolution of Giens tombolo. The study results
showed that the zone from Landmark B01 to B23 (the
northern part of Almanarre beach) was eroded with an
average retreat of -6.4 m/ml corresponding to an
erosion area of -13,900 m2 during the period from 1950
to 1998, whilst the zone from Landmark B23 to B46
was accreted with an average advance of +19.8 m/ml
corresponding to an accretion area of +41,100 m2. In
addition, Than11 applied digital imagery processing
techniques to find the average erosion annual rate of
(-0.01 to -0.63) ± (0.27 to 1.82) m/year in the northern
shoreline and the average accretion annual rate of (0.02
to 2.01) ± (0.14 to 5.1) m/year in the central and
southern parts of the Almanarre beach over the period
from 1920 to 2012.
In this research, the methodology of remote sensing
and GIS technology along with DSAS was used to
attain the main objectives: (1) quantify shoreline
changes as well as accretion and erosion in Almanarre
beach over the period from 1973 to 2015; (2) determine
the main factor influencing the shoreline evolution of
this area; and (3) predict the movement trends of
shoreline in the future.
Materials and Methods
Study Area
The shoreline of Almanarre beach extends from
north to south with the length of approximately 4.5 km
through the Salt Road. According to the measurement
data, the presence of deep cross-shore troughs and
submerged shoals causes the complex bathymetry in
the study area10. An average slope in the direction
perpendicular to the shore is estimated from 1-1.5 %
over the entire beach.
In the study area, the west and southwest winds play
a decisive role in the coastal geomorphology and the
wave agitation in Giens gulf. According to the wind
data recorded at Hyères station (Fig. 1(a)), they blow
with the total frequency of 25.66 % during the
observation time. The winds coming from the west
sector has the highest velocity 21.47 m/s with lower
frequency, whereas the southwest winds maintain the
highest frequency of 13.56 % with an average speed of
5.48 m/s. In addition to winds, waves have strong
impact on the shoreline evolution of Almanarre beach.
The west and southwest waves always attack the west
VU et al.: PREDICTION OF SHORELINE CHANGES IN ALMANARRE
209
coast of Giens tombolo with frequencies of 36.92 %
and 28.84 %, respectively. The southwest waves
generally have low energy with heights from 0.5 to
1.25 m and periods of less than 6 seconds occupying
about 77 % of cases, whilst the waves coming from
the west sector have medium energy with heights
from 0.5 to 2.5 m occupying about 75 % of observed
cases. Due to the largest tidal variation of less than
0.3 m, waves are the main factor acting on the
hydrodynamic and sediment transport processes in the
study area12.
The oblique waves striking the coast of Giens
tombolo generate the long shore current drift which
redistributes sediment. The average current speed of
long shore current varies from 3 to 7 cm/s in normal
sea conditions and from 15 to 25 cm/s in stormy
conditions at 4 m isobaths in the north zone of
Almanarre beach. The dominant directions of long
shore current on an average are from South to East
and West to South, then they meet and mix together to
generate cross-shore flow seaward at central zone of
Almanarre beach13. Although some small streams
flow in to the Giens Gulfs, between the points of
Carqueiranne and Almanarre, the volume of sediment
transported from them is not significant.
The principal sediment transport in the Giens Gulf
starts from Almanarre in the north to Madrague of
Giens in the south. Sediments are finer and finer with
a decrease in the percentage of pebbles (from 55 % to
1 %) and well sorted (1.2 to 0.8 ) southbound.
However, the sector between Landmark B08 and B23
has a coarse grain size of 0.6 mm to 0.8 mm and a
quasi-permanent and homogeneous sorting index due
to the beach nourishment10.
From the different characteristics of the
geomorphology and the limitations of landmarks, the
study area is separated by three main zones, namely,
North zone, Central zone, and South zone (Fig. 1 (b)).
The North zone of 1.5 km length is located between
Landmark B01 and Landmark B14. The Central zone,
1.35 km long, starts from Landmark B15 up to
Landmark B28 whereas the South zone with the
length of 1.525 km spreads from Landmark B29 to
Landmark B42.
Data Source
A series of satellite images such as Landsat 1 MSS,
Landsat 4 TM, Landsat 7 ETM, and Landsat 8 OLI
were acquired at non-equidistant intervals between
1973 and 2015. All the images have been collected
almost at the same time in summer with good quality
in order to exclude the effects of storm surge and
waves. The details with respect to satellite images are
listed in Table 1. However, the raw satellite images
must be pre-processed by image enhancement
(radiometric calibration, atmospheric correction, gap
filling, pan-sharpening) and geometric rectification
steps before being used as map base; because many
defects, like radiometric distortion, wedge-shaped
Fig. 1 — The study area
INDIAN J. MAR. SCI., VOL. 49, NO. 02, FEBRUARY 2020
210
gaps, geometric distortion, presence of noise, etc., due
to variations in the altitude, attitude, and velocity of
the sensor platform usually occur in these images14.
Analysis Methods
In order to extract shoreline from satellite images,
several methods have been generated and developed.
Firstly, a single band method can be used to map and
extract shoreline from optical imagery. This method
has some advantages, viz. the reflectance of water is
almost equal to zero in reflective infrared bands, and
the reflectance of absolute majority of land covers is
greater than the water’s2. However, the main
disadvantage of this method is how to define the
appropriate threshold value. Alesheikh et al.2 proposed
a new procedure for shoreline detection using a
combination of band ratio and histogram thresholding
techniques. Nonetheless, the shoreline moves toward
water in some of the coastal zones and the procedure of
extracting shoreline is quite time consuming. Another
approach of extraction is by automation of shoreline
detection which can be easily to apply and execute15.
The main purpose of edge detection is to reduce the
quantity of data and remove the irrelevant information
in an image, but retain its structural properties. In this
research, the shoreline position was mapped and
extracted by using a nonlinear edge-enhancement
technique with the Canny edge detector. The Canny
edge detection is the most common edge detection
method that performs well optimizing detection
localization and number of responses criteria16. Due to
the edge of the image corresponding to the
discontinuity of the image grey value, the Canny
algorithm is used to determine the pixels in the land-
water boundary if their grey values have relatively
large changes17. This technique gives an outstanding
delimitation of the land-water boundary, and is time
saving. The best colour composites of RGB (Red
Green Blue) 567 (for Landsat MSS images), 543 (for
Landsat TM and ETM+ images), and 652 (for Landsat
OLI images) were be utilized for extracting the
shorelines. These colour composites are improved to
distinguish the objects clearly, such as between soil,
vegetated land, and water, and hence they are easily
digitized. The extracted shorelines were imported in
DSAS module running in ArcGIS environment.
Although many methods were available in the
DSAS, the most commonly used, LRR (Linear
Regression Rate-of-change) statistic and EPR (End
Point Rate) calculations, were hired to quantify the
shoreline changes in Almanarre beach. Specifically,
EPR was utilized for short term change analyses (1973-
1988; 1988-2000; 2000-2008; 2008-2015), whereas
LRR was applied for long term change analysis (1973-
2015). Based on these settings, a total of 176 transects
along the western tombolo were generated each 200
meters perpendicular to the baseline, at every 25 meters
alongshore. Furthermore, intersection point coordinates
between transect lines and shorelines as well as other
statistical results were also computed by DSAS.
Finally, the distances between multiple historic
shorelines and the baseline at each transect computed
by DSAS was input into the code which the authors
created to predict the positions of shorelines in 2020
and 2050. This code uses linear regression equation
and runs in Matlab. The linear regression method
which is used to define shoreline position change rate
eliminates short-term variability and potential random
error by using a statistical approach18. This method
assumed that the observed periodical rate of change of
shoreline position is the best determinant for
prediction of the future shoreline. Its main
shortcoming is that the sediment transport19 and wave
action are not taken into account because the
cumulative impact of all the underlying processes is
presumed to be captured in the position history20. To
predict the future shoreline position, the baseline was
first determined as the buffer of the shoreline in 2015.
Next, transects casting perpendicular to the baseline at
a user-specific spacing alongshore were generated by
DSAS. Next, the intersection points between multi-
temporal shorelines and transects were generated to
input into the linear regression formula to determine
the position of future shoreline at each transect.
Finally, these positions were connected together to
create the future shoreline21.
bxay .
… (1)
Where:
y: Predicted distance from baseline,
x: The shoreline date,
a: Slope (the rate of change) and computed as follows:
Table 1 — List of satellite images used in the study
SI
no.
Satellite and
sensor
Acquisition date
(dd/mm/yyyy)
Local
time
Spatial
Resolution (m)
1
Landsat 1 MSS
01/03/1973
09:51
79
2
Landsat 4 TM
27/08/1988
09:47
30
3
Landsat 7 ETM
28/08/2000
10:08
15/30
4
Landsat 7 ETM
18/08/2008
10:06
15/30
5
Landsat 8 OLI
30/08/2015
10:17
15/30
VU et al.: PREDICTION OF SHORELINE CHANGES IN ALMANARRE
211
n
i
n
iii
n
ii
n
ii
n
iii
xxn
yxyxn
a
1
2
1
2
111 .
… (2)
yi: The distance from baseline to shoreline at date of xi,
n: The number of shorelines,
b: y-intercept (where the line crosses the y-axis) and
computed as follows:
n
ii
n
iixay
n
b11 .
1
… (3)
The accuracy and model quality were defined by
using the cross-validation of the determined historical
shoreline positions22. Particularly, the positional shift in
the estimated shoreline of western Giens tombolo of
2015 was compared and validated with the extracted
shoreline of 2015 from the satellite image. The results
of validation are shown in Fig. 2. It is easily seen that
the predicted shoreline is close to the actual one. The
root mean square error (RMSE) for the entire shoreline
of Almanarre beach was about 3.32 m. The value of
this error is acceptable and reasonable; hence this
method can be applied for predicting the position of
future shorelines.
The accuracy of shoreline position as well as
shoreline change rates can be influenced by several
error sources. There are two kinds of uncertainties
comprising positional uncertainty and measurement
uncertainty. Positional uncertainties are related to the
features and phenomenon that reduce the precision and
accuracy of defining a shoreline position in a given
year, viz. seasonal error Es, and tidal fluctuation error
Etd23. Seasonal error, Es, is induced by the movements
in shoreline position under the action of the waves and
storms. Based on the measurement report of E.O.L24,
seasonal shoreline position differences between the
spring and fall were estimated about ±5 m. The tidal
Fig. 2 — Actual shoreline position (2015) and predicted shoreline position (2015) along Almanarre beach with 5 m linear space transect.
INDIAN J. MAR. SCI., VOL. 49, NO. 02, FEBRUARY 2020
212
fluctuation error, Etd, comes from horizontal movement
in shoreline position along a beach profile due to
vertical tides. The base water level used to define the
shoreline is the High Water Level (HWL). The study
area is in a micro-tidal region with the tidal range less
than 0.3 m, so this error can be neglected25. Regarding
measurement uncertainties, they are associated with the
skill and approach including digitizing error Ed,
rectification error Er and pixel error Ep23. Before
digitization, the satellite images of 1973 and 1988 were
re-sampled from 79 m to 15 m and from 30 m to 15 m,
respectively without adding any spatial information.
Hence, the digitizing errors were estimated about ±12
m for 1973, ±6 m for 1988 and ±3 m for remainders26.
Finally, rectification error, Er, is calculated from the
orthorectification process.
According to Fletcher et al.23, these errors are
random and uncorrelated and can be represented by a
single measure calculated by summing in quadrature.
The total positional uncertainty, Ut, is:
22222 rpdtdst EEEEEU
… (4)
The annualized uncertainty of shoreline change rate
at any given transect was calculated as follows27:
T
UUUUU
Uttttt
a
2
5
2
4
2
3
2
2
2
1
… (5)
where
2
1t
U
,
2
2t
U
,…,
2
5t
U
are the total shoreline
position error for the various year and T is the 42
years period of analysis.
The maximum annualized uncertainty evaluated for
individual transects is about ±0.67 m/year (Table 2).
Results and Discussion
Historical Shoreline Changes over Period of 1973-2015
Based on the short-term analysis for North Zone,
the rate of change was measured along 1.5 km, from
Transect 01 to Transect 61 corresponding to
Landmark B01-B14, and both erosion and accretion
were observed, but erosion is dominant (Fig. 3 (a)). In
the period of 1973-1988, most of transects was
deposited with the maximum accretion rate of about
1.51 m/year observed near Landmark B06. However,
from 1988 to 2000, all transects in this zone suffered a
severe retreat with the highest erosion rate of -2.13
m/year. The main reason of this phenomenon came
from the high frequency of storms. There were 79
heavy seas and storms observed from 1992 to 199910.
This comment is valid for the other zones. The retreat
trend was maintained in the periods of 2000-2008 and
2008-2015 with the highest erosion rates of -1.77
m/year and -2.89 m/year, respectively (Table 3). On
the other side, the long-term analysis of 1973-2015
demonstrates that 90.16 % of transects were subjected
to erosion, whereas only 9.84 % of those were
prograded (Fig. 4). The high accretion rates are
concentrated around Landmark B01 and B03 because
the main long shore sediment transport oriented West-
East is blocked by cross-shore submerged ridge.
Inversely, the high erosion rates are recorded near
Landmark B06 and B08 because this area is directly
exposed to the strong waves from the southwest, with
frequency of 28.84 %.
The central zone is composed of 54 transects over a
total distance of 1.35 km, from Transect 62-115 in
proportion to Landmark B15-B28. The short-term
analysis indicates that this zone has been undergoing
both accretion and erosion, as shown in Fig. 3 (b). In
the 1973-1988 period, only area from Landmark B16
to 18 was eroded with the highest rate of -0.7 m/year.
The remainders were prograded with the maximum
accretion rate of 1.75 m/year. As the first zone from
1988 to 2000, most of transects in Central Zone was
retreated at the mean rate of -1.27 m/year.
Nevertheless, in the period of 2000-2008, 100 % of
transects advanced seaward with the maximum
accretion rate of 1.7 m/year (Table 3). The sudden
positive shoreline change came from the beach
replenishment in this period. According to the report
of CERAMA28, Giens tombolo suffered 12 storms
from December 1999 to December 2008 in which the
strongest storm approached Almanarre beach on 24th,
January 2007. These storms caused severe erosion
and breaching the salt road, especially between
Landmark B13 to B18. Hence, a total of 10,000 tons
of aggregates and nearly 10,400 m3 of sand were used
to nourish this area11,12. However, the beach
nourishment only maintained the balance of this
beach in the short time. The negative trend reappeared
in the period 2008-2015 with the mean rate of -0.76
Table 2 — Calculation of errors for different shorelines.
1973
1988
2000
2008
2015
Seasonal error (Es)
5
5
5
5
5
Tidal fluctuation (Etd)
0
0
0
0
0
Digitizing error (Ed)
12
6
3
3
3
Rectification error (Er)
12
9.9
7.35
10.8
6.75
Pixel error (Ep)
0.5
0.5
0.5
0.5
0.5
Total error (Ut)
17.70
12.62
9.40
12.28
8.93
Annualized error (Ua)
0.67 m/year
VU et al.: PREDICTION OF SHORELINE CHANGES IN ALMANARRE
213
m/year. With regard to the overall shoreline changes,
61 % of transects record erosion and 39 % of transects
record accretion. Erosion frequently occurred from
Landmark B16 to B18 where deep trough nearly
reaches the beach and the slope of the near shore zone
is steeper. As a result, the waves approach the
shoreline with high energy and take away sand from
the beach. Moreover, this area is immediately
influenced by the action combination of the southwest
and west waves with total frequency of 65.76 %.
Otherwise, the area from Landmark B19 to B29 is
advancing with the highest accretion rate of 0.46 m
per year induced by the cross-shore sediment
transport channel9.
The South Zone stretches over 61 transects (No.
116-176), corresponding to Landmark B29 to B42.
Fig. 3 (c) represents the shoreline change of South
Zone between 1973 and 2015. Generally, the erosion
and accretion phenomenon happen alternatively, but
erosion predominantly dominates in most of the
transects. During the period of 1973-1988, the
maximum erosion and accretion rates are -1 and 2.14
m per year, respectively. About 72 % of transects in
this period were deposited with the mean rate of 0.51
m per year. Nonetheless, the long shore pattern was
completely changed to erosion with mean rate of -
0.74 m per year between 1988 and 2000. The negative
trend was kept over the periods of 2000-2008 and
2008-2015 with the average erosion rates of -0.61 and
-0.06 m/year, respectively (Table 3). Additionally, the
long-term analysis reveals a slight accretion with the
maximum rate of 0.46 m/year from Landmark B30 to
B36 and erosion at the maximum rate of -0.71 m/year
from Landmark B36 to B42 over the period of
1973-2015 (Fig. 4). According to Grissac9, the cross-
shore sediment transport in the Giens channel results
in deposition in the area from Landmark B30 to B36.
The eroded area may be provoked by the direct effect
of northwest waves accompanied with the Mistral
wind from the Rhone Valley.
The shoreline of Almanarre beach can be estimated in
the short period of 2015-2020 and the long period of
2015-2050 by using linear regression analysis. In
these periods, the shoreline change rates are predicted
based on historical observations, but the special
climate events, viz. sea level rise caused by global
warming and tropical storms, have not been taken into
account. Fig. 5 represents the positions of future
shorelines and the shoreline change rates, whereas
Table 4 summarizes statistical results of shoreline
a. North Zone from Landmark B01-B14 (Transects 1-61)
b. Central Zone from Landmark B15-B28 (Transects 62-115)
c. South Zone from Landmark B29-B42 (Transects 116-176)
Fig. 3 — Positions of shorelines and transect lines as well as
shoreline change rates using EPR method along Almanarre beach
over a period of 1973-2015
INDIAN J. MAR. SCI., VOL. 49, NO. 02, FEBRUARY 2020
214
change assessment in Almanarre beach; they consist
of the maximum, minimum and mean shoreline
changes as well as % of accreted and eroded transects
for periods: 2015-2020, 2020-2050, and 2015-2050.
Future Shoreline changes over period of 2015-2050
In the first region, North Zone, the shoreline
recession is observed during all periods of 2015-2020,
2020-2050 and 2015-2050 (Fig. 5(a)) with the
average erosion rate of -0.36 m/year, -0.32 m/year,
and -0.33 m/year, respectively. All transects, except
transect no 1-5 that are subjected to accretion, exhibit
erosion in the period of 2015-2020 with the maximum
retreat rate of -1.77 m/year around Landmarks B13
and B14. However, in the next periods of 2020-2050
and 2015-2050, the maximum recession rates are seen
in the vicinity around Landmark B07-08 with the
values of -0.84 m/year and -0.89 m/year (Table 4).
This is the most vulnerable area to erosion in North
Zone. Moreover, the % of eroded transects drastically
increased over time, from 67.21 % in the period of
2015-2020 to 90.16 % in the period of 2020-2050.
In the next region, Central Zone, the results of
statistical analysis obtained for 54 transects illustrate
alternating areas of erosion and accretion, but
majority of transects show erosion (Fig. 5(b)). During
all the periods, the maximum erosion rates are mainly
concentrated in the area between Landmarks B16 and
Table 3 — The statistical summary of shoreline change rate for Almanarre beach over a period of 1973-2015
Zone
Period
Total no.
of transects
Coast
length (m)
Min rate
(m/yr)
Max rate
(m/yr)
Mean rate
(m/yr)
No. of eroded
transects
No. of accreted
transects
% of eroded
transects
% of accreted
transects
North
2015-2020
-1.77
2.59
-0.36
41
20
67.21
32.79
-1.77
2.59
2020-2050
-0.84
0.39
-0.32
55
6
90.16
9.84
-0.84
0.39
2015-2050
-0.89
0.66
-0.33
58
3
95.1
4.9
-0.89
0.66
Central
2015-2020
-1.8
2.32
0.26
20
34
37
63
-1.8
2.32
2020-2050
-0.87
0.46
-0.16
33
21
61.11
38.89
-0.87
0.46
2015-2050
-0.94
0.67
-0.11
31
23
57.4
42.6
-0.94
0.67
South
2015-2020
-2.5
2.62
0
29
32
47.54
52.46
-2.5
2.62
2020-2050
-0.72
0.47
-0.23
45
16
73.77
26.23
-0.72
0.47
2015-2050
-0.91
0.54
-0.2
42
19
68.85
31.15
-0.91
0.54
Table 4 — Statistical summary of shoreline change rate for Almanarre beach over a period of 2015-2050
Zone
Period
Total no.
of transects
Coast
length(m)
Min. rate
(m/yr)
Max. rate
(m/yr)
Mean rate
(m/yr)
No. of
eroded
transects
No. of
accreted
transects
% of
eroded
transects
% of
accreted
transects
North
1973-1988
61
1500
-0.79
1.51
0.53
19
42
30.65
69.35
1988-2000
-2.13
-0.03
-1.27
61
0
100
0
2000-2008
-1.77
2.55
-0.19
36
25
59
41
2008-2015
-2.89
2.36
-0.1
42
19
68.85
31.15
1973-2015
-0.83
0.38
-0.32
55
6
90.16
9.84
Central
1973-1988
54
1350
-0.7
1.75
0.66
6
48
11.11
88.89
1988-2000
-2.36
0.03
-1.27
52
2
96.3
3.7
2000-2008
0.02
1.7
0.74
0
54
0
100
2008-2015
-1.61
-0.06
-0.76
54
0
100
0
1973-2015
-0.86
0.46
-0.16
33
21
61.11
38.89
South
1973-1988
61
1525
-1
2.14
0.51
17
44
27.87
72.13
1988-2000
-1.93
0.71
-0.74
52
9
85.25
14.75
2000-2008
-2.41
1.1
-0.61
44
17
72.13
27.87
2008-2015
-2.62
1.68
-0.06
31
30
50.82
49.18
1973-2015
-0.71
0.46
-0.22
45
16
73.77
26.23
Fig. 4 — The variation of shoreline change rates using LRR and
EPR methods along Almanarre beach over a period of 1973-2015.
Future Shoreline Changes over Period of 2015-2050
VU et al.: PREDICTION OF SHORELINE CHANGES IN ALMANARRE
215
B17, while the maximum progradation rates are
observed between Landmarks B20 and B22 that
coincide with the position of the cross-shore sediment
transport channel. During 2015-2020, 62 % of the
transects were accreted with the highest rate of 2.32 m
per year at Landmark B20 and B22 that coincide with
the position of the cross-shore sediment transport
channel. During 2015-2020, 62 % of the transects
were accreted with the highest rate of 2.32 m per year
at Landmark B20 (Table 4). Nonetheless, the positive
trend will be suddenly changed to the negative trend
in the next periods. Particularly, 61.11 % of eroded
transects are recorded with the highest rate of -0.87 m
per year during the period 2020-2050, whereas the
eroded transects are predicted at about 57.4 % with
the highest erosion rate of -0.94 m per year between
2015 and 2050 (Table 4). Additionally, the prediction
of future shoreline positions also revealed that the
shoreline between Landmark B16 and B18 would be
the most seriously eroded area.
For the South zone of Almanarre beach, a complex
pattern of shoreline evolution is predicted with areas
in erosion alternating with areas in accretion
(Fig. 5(c)). During the period of 2015-2020, the
eroded and accreted transects are almost the same
with 48 % and 52 %, respectively. The maximum
recession and accumulation rates are predicted about -
2.5 m/year and 2.62 m/year (Table 4). The accretion
areas are also forecasted in the vicinity of Landmarks
B31-33 and B35-37. In the next periods of 2020-2050
and 2015-2050, erosion is dominant with the average
rates of -0.23 m/year and -0.2 m/year, respectively.
The maximum accretion rates are predicted to take
place around Landmark B38-39, while the maximum
erosion rates can be observed near Landmark B35-36.
Finally, (Fig. 5 (c) also shows that the area between
Landmark B37 and B42 could be subjected to severe
erosion in future.
The study on the shoreline changes in Giens
tombolo from 1973 to 2015 reveals that alternate
accretion and erosion occurred along the shoreline of
Almanarre beach, but erosion is dominant. Especially,
most of transects along this zone were subjected to
erosion from 2008 onwards. The main reasons of this
recession are the action of waves on the near-shore
bathymetry. Additionally, there is no sediment supply
feeding the beach continuously, viz. no river mouths
in the Giens gulf. The prediction of shoreline
positions in 2020 and 2050 reveals that the erosive
tendency will continue, particularly and severely in
a. North Zone from Landmark B01-B14 (Transects 1-61)
b. Central Zone from Landmark B15-B28 (Transects 62-115)
c. South Zone from Landmark B29-B42 (Transects 116-176)
Fig. 5 — Positions of shorelines and transect lines as well as
shoreline change rates using EPR method along Almanarre beach
over a period of 2015-2050.
INDIAN J. MAR. SCI., VOL. 49, NO. 02, FEBRUARY 2020
216
the areas of Landmarks B07-08 and Landmarks
B16-18. Finally, this study also demonstrates that the
beach nourishment method only helps in the summer
time and that too for the short term.
Conclusion
The study on the shoreline changes in Giens tombolo
from 1973 to 2015 reveals that alternate accretion and
erosion occurred along the shoreline of Almanarre
beach, but erosion is dominant. Especially, most of
transects along this zone were subjected to erosion from
2008 onwards. The main reasons of this recession are
the action of waves on the near-shore bathymetry.
Additionally, there is no sediment supply feeding the
beach continuously, viz. no river mouths in the Giens
gulf. The prediction of shoreline positions in 2020 and
2050 reveals that the erosive tendency will continue,
particularly and severely in the areas of Landmarks B07-
08 and Landmarks B16-18. Finally, this study also
demonstrates that the beach nourishment method only
helps in the summer time and that too for the short term.
Acknowledgment
This search was financially supported by the 911
Program of Vietnamese Ministry of Education and
Training. The authors thank EOL, CETMEF,
CEREMA, SHOM, and REFMAR for providing the
bathymetric and hydrodynamic data. The USGS were
appreciated for sharing the satellite images.
References
1 Tran, T.V. & Tran, T.B., Application of remote sensing for
shoreline change detection in Cuu Long Estuary. VNU J
Earth Environ Sci, 2009. 25.
2 Alesheikh, A., Ghorbanali, A., & Nouri, N., Coastline
change detection using remote sensing. Int J Environ Sci
Technol, 2007. 4(1): p. 61-66.
3 Zhang, C., Zheng, J., Dong, X., Cao, K., & Zhang, J.,
Morphodynamic response of Xiaomiaohong tidal channel to
a coastal reclamation project in Jiangsu Coast, China. J
Coastal Res, 2013. 65: p. 630-635.
4 Jinhai, Z., Chi, Z., Zeki, D., & Lihwa, L., Numerical Study
of Sandbar Migration under Wave-Undertow Interaction. J
Waterw, Port, Coastal, and Ocean Eng, 2014. 140(2): p.
146-159.
5 Gu, Z., Zhang, C., & Zheng, J., Influences of wave forcing
and morphological variability on the evolution of a double-
sandbar system. Proc Inst Mech Eng, Part M, 2015. 230(3):
p. 467-480.
6 Pereira, C., Coelho, C., Ribeiro, A., Fortunato, A.B., Lopes,
C.L., & Dias, J.M., Numerical modelling of shoreline
evolution in the Aveiro coast, Portugal – climate change
scenarios. J Coastal Res, 2013: p. 2161-2166.
7 Winarso, G., janto, J., & Budhiman, S., The potential
application of remote sensing data for coastal study, in 22nd
Asian Conference on Remote Sensing. 2001: Singapore.
8 Blanc, J.J., Recherches de sédimentologie littorale et sous-
marine en Provence occidentale. 1958, Université de Paris:
Masson et Cie. p. 140.
9 Grissac, A.J.D., Sédimentologie dynamique des rades
d'Hyères et de Giens (Var). Problèmes d'Aménagements.
1975, Université d'Aix-Marseille II: Marseille. p. 86 +
annexes.
10 Courtaud, J., Dynamiques geomorphologiques et risques
littoraux cas du tombolo de giens (Var, France méridionale).
2000, Université Aix-Marseille I. p. 263.
11 Than, V.V., Modélisation d'érosion côtière : application à la
partie Ouest du tombolo de Giens, in LATP. 2015, Aix
Marseille Université: Marseille. p. 400.
12 SOGREAH, Etudes sédimentologiques de la rade d Hyères.
Littoral de port Pothuau à la Badine. 1988. p. 68 + annexes et
cartes.
13 Lacroix, Y., Than, V.V., Leandri, D., & Liardet, P., Analysis
of a Coupled Hydro-Sedimentological Numerical Model for
the Tombolo of GIENS. Int J Environ Ecol Geol Geophys
Eng, 2015. 9(3): p. 117 - 124.
14 Lillesand, T.M., Kiefer, R.W., & Chipman, J.W., Remote
sensing and image interpretation. 2008: John Wiley & Sons.
15 Loos, E.A. & Niemann, K.O. Shoreline feature extraction
from remotely-sensed imagery. in IEEE International
Geoscience and Remote Sensing Symposium. 2002.
16 Canny, J., A Computational Approach to Edge Detection.
IEEE Trans Pattern Anal Mach Intell, 1986. PAMI-8(6):
p. 679-698.
17 Liu, H. & Jezek, K.C., Automated extraction of coastline
from satellite imagery by integrating Canny edge detection
and locally adaptive thresholding methods. Int J Remote
Sens, 2004. 25(5): p. 937-958.
18 Douglas, B.C. & Mark, C., Long-Term Shoreline Position
Prediction and Error Propagation. J Coastal Res, 2000.
16(1): p. 145-152.
19 Michael, S.F., Dolan, R., & Elder, J.F., A New Method for
Predicting Shoreline Positions from Historical Data. J
Coastal Res, 1993. 9(1): p. 147-171.
20 Rongxing Li, J.-K.L.Y.F., Spatial Modeling and Analysis for
Shoreline Change Detection and Coastal Erosion Monitoring.
Mar Geod, 2001. 24(1): p. 1-12.
21 Nguyen, L.D., Minh, N.T., Thy, P.T.M., Phung, H.P., &
Huan, H.V. Analysis of changes in the riverbanks of Mekong
river-Vietnam by using multi-temporal remote sensing data.
in International Archives of the Photogrammetry, Remote
Sensing and Spatial Information Science. 2010. Kyoto,
Japan.
22 Maiti, S. & Bhattacharya, A.K., Shoreline change analysis
and its application to prediction: A remote sensing and
statistics based approach. Mar Geol, 2009. 257(1–4):
p. 11-23.
23 Fletcher, C.H., Romine, B.M., Genz, A.S., Barbee, M.M.,
Dyer, M., Anderson, T.R., Lim, S.C., Vitousek, S.,
Bochicchio, C., & Richmond, B.M., National assessment of
shoreline change: Historical shoreline change in the
Hawaiian Islands, in Open-File Report. 2012: Reston, VA. p.
i-55.
24 E.O.L, Suivi de l évolution des plages de la commune
Hyères-les-palmiers. 2000-2010: Commune de Heres-Les-
Palmiers.
VU et al.: PREDICTION OF SHORELINE CHANGES IN ALMANARRE
217
25 Rajasree, B.R., Deo, M.C., & Sheela Nair, L., Effect of climate
change on shoreline shifts at a straight and continuous coast.
Estuar Coast Shelf Sci, 2016. 183, Part A: p. 221-234.
26 Jayson-Quashigah, P.-N., Addo, K.A., & Kodzo, K.S., Medium
resolution satellite imagery as a tool for monitoring shoreline
change. Case study of the Eastern coast of Ghana. J Coastal Res,
2013: p. 511-516.
27 Hapke, C.J., Himmelstoss, E.A., Kratzmann, M.G., List, J.H.,
& Thieler, E.R., National Assessment of Shoreline Change:
Historical Shoreline Change along the New England and
Mid-Atlantic Coasts. 2011, U.S. Geological Survey.
p. 57.
28 CERAMA, Details of the campaign 08301 and 08302-
Porquerolles. 2014.
