Beach erosion and recovery during consecutive
storms at a steep-sloping, meso-tidal beach
Michalis I. Vousdoukas,
1,2
*
Luis Pedro M. Almeida
2
and Óscar Ferreira
2
1
Forschungszentrum Küste. Hannover, Germany
2
CIMA, University of Algarve, Campus de Gambelas, Faro, Portugal
Received 20 July 2011; Revised 24 October 2011; Accepted 7 November 2011
*Correspondence to: Michalis I. Vousdoukas, Forschungszentrum Küste. Merkurstraße, 11, 30419, Hannover, Germany. E-mail: vousdoukas@fzk-nth.de
ABSTRACT: This study analyses beach morphological change during six consecutive storms acting on the meso-tidal Faro Beach
(south Portugal) between 15 December 2009 and 7 January 2010. Morphological change of the sub-aerial beach profile was mon-
itored through frequent topographic surveys across 11 transects. Measurements of the surf/swash zone dimensions, nearshore bar dy-
namics, and wave run-up were extracted from time averaged and timestack coastal images, and wave and tidal data were obtained
from offshore stations. All the information combined suggests that during consecutive storm events, the antecedent morphological
state can initially be the dominant controlling factor of beach response; while the hydrodynamic forcing, and especially the tide
and surge levels, become more important during the later stages of a storm period. The dataset also reveals the dynamic nature of
steep-sloping beaches, since sub-aerial beach volume reductions up to 30 m
3
/m were followed by intertidal area recovery (–2 < z
3 m) with rates reaching ~10 m
3
/m. However, the observed cumulative dune erosion and profile pivoting imply that storms, even
of regular intensity, can have a dramatic impact when they occur in groups. Nearshore bars seemed to respond to temporal scales
more related to storm sequences than to individual events. The formation of a prominent crescentic offshore bar at ~200 m from
the shoreline appeared to reverse the previous offshore migration trend of the inner bar, which was gradually shifted close to the sea-
ward swash zone boundary. The partially understood nearshore bar processes appeared to be critical for storm wave attenuation in
the surf zone; and were considered mainly responsible for the poor interpretation of the observed beach behaviour on the grounds of
standard, non-dimensional, morphological parameters. Copyright © 2011 John Wiley & Sons, Ltd.
KEYWORDS: coastal storms; coastal erosion; nearshore bars; beach recovery; storm groups
Introduction
Storms constitute a significant hazard in coastal regions, trig-
gering geomorphological change and threatening harbour
facilities, coastal tourism infrastructure, houses, and even
human lives, through storm-surge flooding and wave attack
(e.g. Forbes et al., 2004; Lantuit and Pollard, 2008; Seymour
et al ., 2005). As a r esult, understanding coastal response to
environmental forcing has become an urgent issue, espe-
cially given sea-level rise projections and the increasing
occupation of coastal areas. In addition, reduced terrestrial
sediment supply (e.g. Velegrakis et al., 200 8) and i nfrastruc-
tural development of the backshore has reduced the capacity
of many beaches and of their associated dune systems to
absorb storm energy. For the above reasons, there is a great
interest in better understanding the morphological response
of coastal areas to storm events.
Storm driven morphological change can be more significant
when consecutive storms (storm groups) occur, as they can
have a cumulative effect on coastal morphology (Callaghan
et al., 2008; Ferreira, 2005). Storm groups can be considered
as one ‘event’ when the recovery time between storms is short
and their quick succession can have a large impact on
morphology (Lee et al., 1998). For example, Van Enckevort
and Ruessink (2003) reported that the temporal scales of
nearshore bar position fluctuations were more related to storm
sequences than to individual events. However, the probability
of storm group occurrence is significantly lower than the one
of individual storms, limiting the number of related studies
and leading to important knowledge gaps.
Swash represents the main coastal hazard agent, and its
intensity is related to the amount of wave energy attenuated
during shoaling and breaking that is controlled by beach
morphology (e.g. Komar, 1998). Steeper beach slopes tend
to allow higher energy levels at t he shoreline due to reduced
dissipation (e.g. Vousdoukas et al., 2009; Wright and Sho rt,
1984) an d to focus wave power at narrower profile sections
rather than at gently sloping beaches. As a result, steep
beaches are more likely to undergo rapid morp hological
changes resulting from variations in offshore wave energy
level and direction (e.g. Qi et al., 2010). However, most
existing studies on morphological impacts of storms are
related to dissipative/intermediate beaches (after Wright and
Short, 1984) and very few (e.g. Backstrom et al., 2008)
involve steep, sandy beaches with beach face slopes exceed-
ing tan(b) > 010. The number of previous stud ies (if any) on
the i mpact of storm groups on such environments is even
more limited.
EARTH SURFACE PROCESSES AND LANDFORMS
Earth Surf. Process. Landforms 37, 583–593 (2012)
Copyright © 2011 John Wiley & Sons, Ltd.
Published online 7 December 2011 in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/esp.2264
Given the foregoing context, this paper analyses beach ero-
sion and recovery during consecutive storms acting on the
steep-sloping, meso-tidal Faro Beach (south Portugal) between
15 December 2009 and 7 January 2010. The impact of the
storm group was monitored through topographic surveys and
with a coastal video system which provided information on
wave run-up and nearshore bar dynamics. All the available
information is combined and thoroughly discussed in an effort
to expand our understanding of (i) erosion and recovery
processes under storm groups on steep-sloping, dynamic
beaches, and of (ii) the relative importance of acting forces,
such as offshore wave characteristics, tidal elevation, the
grouping of storms, as well as of the morphological feedback
on wave attenuation.
Study Area
Faro Beach is located in the central and eastern areas of the
Ancão Peninsula, in the westernmost sector of the Ria Formosa
barrier island system (Algarve, south Portugal; see Figure 1).
Tides in the area are semi-diurnal, with average ranges of
28 m for spring tides and 13 m during neap tides, although a
maximum range of 35 m can be reached. Wave climate in
the area is moderate, with an average annual significant off-
shore wave height H
s
=092 m and average peak wave period
T
p
=82 seconds (Costa et al., 2001). Waves approach from
the west–southwest (W-SW) and east –southeast (E-SE) for 71%
and 23% of the time, respectively (Costa et al., 2001). Storms
are considered to occur when the significant wave height (H
s
)
exceeds 3 m (Almeida et al., 2011b).
While the eastern sector of Faro Beach is accreting and veg-
etated foredune development is evident (Almeida et al., 2011a;
Ferreira et al., 2006), the central and western parts tend
toward erosion, with much of their natural dune ridge having
being destroyed and replaced by urban development (Figure 1).
As a result, a part of the ocean front has been artificially stabi-
lized with sea walls, which together with roads and walls are
often overwashed during spring tides or under stormy condi-
tions. The area monitored in this study is a coastal stretch of
~100 m alongshore length, located in the central, steep-sloping
part of Faro Beach (Figure 1). Although settlements and various
coastal facilities (e.g. car parking, restaurants) can be found
behind the beach, a natural dune exists at this area with an
average crest elevation ~6 m above mean sea level (MSL).
Faro Beach is characterized by a steep beach-face, with aver-
age slope around 10%, varying from 6% to 15% (Vousdoukas
et al., 2011c), with a tendency to decrease eastwards along the
beach (Ferreira et al., 2006). The western part discussed in the
present study has been classified as ‘reflective’ (see classification
of Wright and Short, 1984) by Martins et al. (1996); even though
the ‘intermediate’ class is more appropriate during the energetic
winter period (Vousdoukas et al., 2011c). Sediments are medium
to very coarse, moderately well sorted sands (see classification of
Folk, 1980) with d
50
~05mmand d
90
~ 2 mm.
Materials and Methods
Wave and tide measurements
Wave data were provided by a wave buoy deployed offshore
from Faro Beach at 93 m depth (Figure 1a) by the Portuguese
Hydrographic Institute (http://www.hidrografico.pt). Tidal data
were available from a pressure transducer (Infinity PT, Figure 1b),
deployed offshore Faro Beach at ~14 m depth MSL. Wave spec-
tra and statistical wave parameters (e.g. significant wave height
and peak wave period) were obtained using standard pressure
attenuation correction and zero-crossing techniques. Infragravity
and incident frequency band energy contributions were esti-
mated by integrating the spectra energy density considering a
cutoff frequency of 005 Hz.
Figure 1. (a) Map of the Algarve region (south Portugal) showing the locations of the study area (arrow) and the IH buoy. (b) Map of Faro Beach, containing
LiDAR topographic and bathymetric information and showing the bathymetric survey lines, as well as the location of the pressure transducer (InfinityPT).
(c) Zoom-in on the sub-aerial profile section showing the topographic survey grid, the location of the video station, and the crossshore transect along which
the swash motion was monitored (Timestack transect). This figure is available in colour on line at wileyonlinelibrary.com/journal/espl
584 M. I. VOUSDOUKAS, L. P. M. ALMEIDA AND Ó. FERREIRA
Copyright © 2011 John Wiley & Sons, Ltd. Earth Surf. Process. Landforms, Vol. 37, 583–593 (2012)
Topographic and bathymetric measurements
Topographic data were obtained along 11 cross-shore transects
with 10 m alongshore spacing (Figure 1c) using a real-time-
kinematic global positioning system (RTK-GPS). Seventeen sur-
veys took place during low tide over a period of 20 days, with
an estimated accuracy of ~5 cm for both vertical and horizontal
dimensions. Pre- and post-storm bathymetric surveys took
place on 15 December 2009 and 8 January 2010, respectively,
using two RTK-GPS for geo-location and an ecosounder
(ODOM CV100) for depth measurements. The estimated accu-
racy is around 1 m for horizontal positioning and 20 cm for ver-
tical levelling, and the surveys covered four cross-shore
transects with 30 m alongshore spacing (Figure 1b).
All topo-bathymetric data were initially acquired in Datum
73 (EPSG:27493) coordinate system and were subsequently ro-
tated so that the x- and y-axes corresponded to cross-shore and
alongshore dimensions, respectively. The rotation angle was
39
anti-clockwise around the origin of the new coordinate sys-
tem: [x,y] = [12255, –295575] in Datum 73. Grids were gener-
ated from each topographic survey; from which alongshore
mean and standard deviation elevation profiles were estimated.
The alongshore averaged elevation profile was considered as
the representative one; while the alongshore standard deviation
elevation profile was used as a proxy of alongshore morpholog-
ical variability (e.g. the presence of beach cusps). The bathy-
metric data were also alongshore averaged and, after being
combined with the topographic data, provided full pre- and
post-storm beach profiles. Volumetric changes were estimated
for the sub-aerial profile, considering sediments lying above
MSL. The minimum elevations observed at the alongshore av-
eraged profile during the monitoring period were used as the
baseline for sediment volume estimation.
Wave run-up measurements
A video monitoring station installed on the roof of a building
facing Faro Beach was acquiring coastal zone imagery during
the monitoring period for 10 minutes every hour during day-
light, with an acquisition frequency of 1 Hz. The system con-
sisted of two Mobotix M22, 3.1 megapixel (2048 1536
pixel resolution), Internet Protocol (IP) cameras, connected to
a personal computer (PC) with permanent internet access.
The elevation of the centre of view (COV) was around 20 m
above MSL. The two cameras provided a ~90
view of the coast
westward of the cameras, covering an alongshore length of
500m, including the area monitored through topographic sur-
veys (Vousdoukas et al., 2011a).
Transformation from image to world coordinates (image geo-
rectification) took place applying standard lens distortion cor-
rection (Bouguet, 2007) and perspective transformation theory
(Hartley and Zisserman, 2006). Timestack images were gener-
ated for the hourly sets acquired during those days on which to-
pographic surveys took place (Figure 2). The cross-shore
resolution of the processed timestack images was 02 m, equal
to the minimum pixel footprint along the monitored transect
(Vousdoukas et al., 2011c). On three occasions, when no major
morphological changes were detected between days, image
data from the day following a topographic survey were also
used. Therefore, although run-up was not measured every
day, the approach used served to diminish geo-locational arte-
facts on the obtained wave run-up measurements.
All timestack images were processed in an open-access
Graphical User Interface (GUI) software (https://sourceforge.
net/projects/guitimestack), specially developed on MATLAB©
to extract and process time-series of the cross-shore position
of the swash extrema (Vousdoukas et al., 2011c). The software
allows the extraction of 2% exceedence ( R
2
) and maximum
wave run-up height values (R
max
), as well as the total run-up
elevation
total
(t) series and the corresponding spectra. The
boundary between the surf and swash zones for each image
set was defined as the location of the minimum run-up eleva-
tion, while the maximum value was defined as the upper swash
zone limit (maximum
total
). All the earlier boundaries were
estimated in terms of both elevation and cross-shore distance
as measured from the topographic data.
Surf zone and nearshore bars
Plan-view TIMEX images were generated with a horizontal reso-
lution of 025m, projected onto a horizontal plane with elevation
equal to the tidal water level
tide
(Vousdoukas et al., 2011a,
Vousdoukas et al., 2011b), and were used to extract nearshore
bar locations. To reduce geo-locational errors, images acquired
during low tide were considered and with sea levels belonging to
the first (lower) quartile of each tidal cycle. Pixel intensity profiles
were extracted from each image from cross-shore topographic
monitoring transects with 5 m spacing and values were averaged
and normalized, to cope with longshore variability. Analysis of
the alongshore-averaged and normalized pixel intensity profiles
showed that the positions of the nearshore bars were characterized
by the presence of ‘bell-shaped’ peaks (Figure 3b), and could be
represented as a function of the sum of a second-order polynomial
and N
g
Gaussian functions (e.g. Aarninkhof and Ruessink, 2004):
IðxÞ¼p
1
þ p
2
x þ p
3
x
2
þ
X
N
g
i¼1
g
i
1
exp
x g
i
2
g
i
3
2
!
(1)
where x is cross-shore distance, N
g
is the number of Gaussians/
nearshore-bars (two for t he present case), p
1–3
are coefficients
Figure 2. Example of the transformation from world to distorted im-
age coordinates in order to generate timestack images from the raw
images (a), considering the irregular beach morphology (b). The dotted
line indicates the pixels used for the timestack generation, while colours
denote vertical elevation (in metres). This figure is available in colour
online at wileyonlinelibrary.com/journal/espl
585BEACH EROSION/RECOVERY DURING CONSECUTIVE STORMS
Copyright © 2011 John Wiley & Sons, Ltd. Earth Surf. Process. Landforms, Vol. 37, 583–593 (2012)
to param eterize the quadratic function, g
i
j
are the Gaussian
parameters (j ~ [1 2 3]), for amplitude, centre, and width, re-
spectively), and I is alongshore averaged pixel intensities.
An iterative solver (Lagarias et al., 1998) was applied to esti-
mate the optimal p and g parameters in Equation 1, and the
centres of the Gaussians were taken as the locations of the
nearshore bars (see Figure 3). Areas where bore-generated tur-
bulence took place (considered to be the surf zone) were iden-
tified from the alongshore averaged and normalized pixel
intensity profiles when a threshold value of 02 was exceeded.
Even though the procedure is automatic, user supervision and
manual corrections were applied when necessary, so as to
identify poor quality images, increase the amount of extracted
information, and enhance accuracy.
Results
Offshore wave conditions
Five storm events occurred during the monitoring period,
according to the significant wave height (H
s
= 3 m) storm
threshold value for the Algarve coast. A two-day period with
moderate conditions (H
s
reaching 24 m), starting on 16
December 2009, resulted in significant morphological change
and is also considered as an event (Figure 4b, Event A) making
up a total of six events labelled from A to F (Figure 4b). The first
storm (Event B) occurred on 18 December 2009 with a duration
of 12 hours, coinciding with a tidal range of around 2 2m
(Figure 4a), followed by ~40 hours of milder conditions. The values
of H
s
varied from 25 to 3 m for two days, until 23 December
2009 06:00 when it reached the peak value of H
s
=43m
(Figure 4b, Event C), during almost neap tide with a range
15 m. The value of H
s
remained above 4 m for six hours and then
fluctuated between 3 and 35 m for two days; this was followed
by two periods, one of 44 hours duration with H
s
=25 m and
one of 34 hours duration with H
s
around 14m.
Wave conditions started to build up again from H
s
=15m on
28 December 2009 06:00, to H
s
=34 m on 29 December 2009
17:00 and remained at that level for five hours (Event D), during
which tidal range was ~2m. The peak wave period reached also
the maximum, for the monitoring period, value of T
p
=155 sec-
onds. Another high H
s
value took place from 30 December 2009
12:00 until 31 December 2009 18:00, during almost spring tide
(range 3 m), when the H
s
reached 37 m (Event E). A mild period
of ~35 days followed and a final event occurred on 5 January
2010 with a duration of approximately six hours and a maximum
H
s
= 4 m (Event F), during which tidal range was 25m.
During the analysed period, wave direction varied from 210
to 270
and was consistently between 230
–240
during the
wave height peaks (Figure 4d); slightly deviating from shore-
normal conditions (~220
). Similarly, the peak wave period
was around 10 seconds, but fluctuated from 5 to 16 seconds
(Figure 4c), with the highest values for the monitoring period
being observed during Event D (T
p
~ 15 seconds) and the low-
est ones during Events B and C (T
p
~ 11 seconds).
Swash zone measurements
Wave run-up measurements were acquired during daylight
with the exception of 21– 22 December and 1 January, when
−300
−250
−200
−150
−100
−50
0
50
100
150
y (m)
a
−100 0 100 200
0
1
x (m)
b
−0.5
Figure 3. Time averaged (TIMEX) image (a) and resulting identification
of the surf limits (vertical lines, b) and the nearshore bar locations (solid
line peaks, b); based on the average pixel intensities (y-axis, b) along
cross-shore transects in the study area (horizontal lines, a; only five
shown for better display). A limited number of transects is shown for bet-
ter display. This figure is available in colour online at wileyonlinelibrary.
com/journal/espl
14/12
16/12 18/12 20/12 22/12 24/12 26/12 28/12 30/12 01/01 03/01 05/01
08/01
−1
0
1
η
t
ide
(m)
AB C DE F
a
1
2
3
4
H
s
(m)
b
5
10
15
T
p
(s)
c
14/12 16/12 18/12 20/12 22/12 24/12 26/12 28/12 30/12 01/01 03/01 05/01 08/01
220
230
240
250
260
270
Mwd (
o
)
d
Figure 4. Hydrodynamic conditions during the December 2009–
January 2010 storms: (a) tidal elevation; (b) significant wave height; (c)
peak wave period; (d) mean wave direction.
586 M. I. VOUSDOUKAS, L. P. M. ALMEIDA AND Ó. FERREIRA
Copyright © 2011 John Wiley & Sons, Ltd. Earth Surf. Process. Landforms, Vol. 37, 583–593 (2012)
the storms resulted in blackouts at the barrier island and
stopped temporarily the operation of the video station (see
missing data in Figure 5). The 2% exceedence wave run-up
height values R
2
varied from 017 to 225 m, around an average
value of R
2
~111 m (Figure 5a). The average
total,2
value was
26 m MSL, while the minimum and maximum values were
04 and 51 m MSL, respectively (Figure 5b), with the pre-storm
dune foot elevation along the study area measuring ~4 m MSL
and the dune crest elevation ranging from ~57to65 m MSL.
Below-storm-threshold conditions during Event A resulted in
increased values of R
2
and especially
total,2
,(>4 m MSL; Fig-
ure 5b). Event B was more energetic than Event A, but resulted
in R
2
and
total,2
values which were lower than those in Event
A, possibly because of the slightly lower tides, but also because
of the adapting morphology. Event C resulted in higher R
2
values than the previous events (Figure 5a), but lower
total,2
as it coincided with neap tides (Figure 5b). The highest R
2
and
total,2
values, exceeding 2 m and 5 m MSL, respectively,
were obtained during Events D and E, with Event D coinciding
with the highest period measured during the monitoring period.
The observed R
2
and
total,2
values during Event F were the low-
est ones among the discussed storm events (Figure 5).
The swash zone length ranged from 38 m to 226 m and av-
eraged 128 m for the monitoring period, while it appeared to
increase after 18 December and fluctuated less during neap
tidal cycles (24–29 December). As expected, the horizontal
swash excursion extrema, x
swash
, followed the trend of the
beach-face water level values
total,2
, with the exception of 3
January (Event F) when x
swash
reached its peak value (x
swash
220 m), which was not the case for
total,2
. The swash zone
appeared to become wider (> 17 m) when intense wave condi-
tions coincided with rising tides (e.g. H
s
> 25 and
tide
> 2 m),
and the most intense regime was observed on the 29 December
and 31 December (Events D and E) when the two parameters
peaked together.
Wave spectra
Offshore wave spectra revealed two dominant frequencies at
the initiation stage of each storm event, one at f
p
< 01 1/s
and another at f
p
~023 1/s, both gradually shifting towards a
final peak value of f
p
~007–01 1/s, depending on the particular
event. Out of all the events, Event C had the highest total energy,
but had lower infra-gravity energy compared to Events D, E, and
F (Figure 6a). Infra-gravity band contributions to the total offshore
wave energy ranged from 029% to 59%, around an average
value of 17%; while the swash spectra were dominated by
low frequency contributions, especially during storm conditions,
with energy contributions varying from 1904% to 911%,
around a mean of 5676% (Figure 6c). Interestingly, while off-
shore energy was gradually reduced during the period between
Events C and D (Figure 6a), swash energy levels remained rela-
tively high (Figure 6b). The effect of the tidal water-level fluctua-
tions is discerned by the swash spectra energy, and the incident
frequency band energy is in phase with
tide
. Infra-gravity band
contributions in the swash energy were shown to also vary with
tidal level, and tended to gradually increase from Event A to
Event E (Figure 6c).
Surf zone/nearshore bars
Analysis of the TIMEX images showed that the outer bar was
not well defined during the first several days (16–21 December;
Events A and B) and it seemed more like a submerged, terraced
feature (Figures 7a and 7c). However, the inner bar was more
prominent and detached from the beach (Figures , 7a–7e) and
appeared to gradually move offshore until 19 December from
x ~ 140 m to x ~ 180 m (Figure 8b). There was a two-days gap
in image acquisition until 23 December, when the inner bar
was almost attached to the beach and started to gradually move
shoreward, eventually to become more stable, fluctuating
around a position x = 170 m (Figure 8b).
A well defined, crescentic outer bar was visible after Event C
(Figures , 7g–7j), especially during low tides and H
s
values
exceeding 2 m and appeared to move offshore during the 23
December Event C and again on 4 January (Event F; Figure 8b).
The peak wave conditions during Event C were also related to
0.5
1
1.5
2
2.5
3
R
2
(m)
AB C DE F
a
14/12 16/12 18/12 20/12 22/12 24/12 26/12 28/12 30/12 01/01 03/01 05/01 08/01
0
1
2
3
4
5
6
η
total,2
(m)
b
Figure 5. (a) The 2% exceedence run-up height and (b) 2% exceedence
values of the total run-up elevation (MSL) during the monitoring period. This
figure is available in colour online at wileyonlinelibrary .com/journal/espl
14/12 16/12 18/12 20/12 22/12 24/12 26/12 28/12 30/12 01/01 03/01 05/01 08/01
0
0.008
0.016
0.024
0.032
0.04
E
infra
(m
2
)
AB C DE F
a
0
0.2
0.4
0.6
0.8
1
E (m
2
)
b
14/12 16/12 18/12 20/12 22/12 24/12 26/12 28/12 30/12 01/01 03/01 05/01
08/01
0
2
4
6
8
10
%−Infragravity, offshore
c
0
0.2
0.4
0.6
0.8
1
E
inci
(m
2
)
Infragravity
Incident
0
20
40
60
80
100
%−Infragravity, swash
Offshore
Swash zone
−2
−1
0
1
2
3
[:]η
tide
(m)
Figure 6. Infra-gravity (solid line, left axis) versus incident band energy
(dashed line, right axis) from the offshore wave (a) and swash zone (b)
spectra, as well as comparisons of the contribution of the infra-gravity
band to the total energy (c). The tidal elevation is also shown on the right
axis by a black dashed line in (b). This figure is available in colour online
at wileyonlinelibrary.com/journal/espl
587BEACH EROSION/RECOVERY DURING CONSECUTIVE STORMS
Copyright © 2011 John Wiley & Sons, Ltd. Earth Surf. Process. Landforms, Vol. 37, 583–593 (2012)
