LONG-TERM COASTLINE EVOLUTION OF FIGUEIRA DA FOZ – NAZARÉ SECTOR (PORTUGAL)

: Some sectors of the west Portuguese coast are particularly endangered by erosion and flooding. Regional to local scale information, on coastline evolutionary trend, is particularly valuable in sectors that includes areas with relevant erosion. A continuous, high-resolution, dataset on coastal evolution, from 1947 and 2015, between Figueira da Foz and Nazaré , was achieved within the Programme “Geological and Coastal Hazard Mapping at a 1:3000 resolution scale” at the National Laboratory of Energy and Geology (LNEG). This work, due to the detailed scale of analysis in a wide geographic context, allows to have both, a general overview of the coastal evolution and, at the same time, when zooming in up to 1:3000 scale, to observe the local behaviour and to quantify the occurred changes. Also, the well time-spaced aerial photograph dataset allows to compare the resultant coastline movement between the oldest and the youngest coastline (NSM index), with the total coastline oscillation (SCE index), bringing new insights on the coastline stability at a local scale. The evolution trend shows an overall erosional behaviour, if considering the entire sector. Erosion occurs predominantly in the north, as the south shows more stability and progradation. Quantification of the land-lost and land-gain due to the coastline shift in a 68-year period shows that 1 164 888 m 2 of land were lost along 30 470 m of the coastal fringe, and 462 330 m 2 were gained along an extension of 21 010 m.


INTRODUCTION
The coastline of the Portuguese mainland extends approximately along 900 km, having a wide morphological diversity and comprising a multiple range of environments, being the beaches, wetlands, hardened and cliffed coasts the main typologies.
As ¾ of the population live in the coastal areas, generating around 80% of the gross national product, this sector has a considerable economic importance (Santos and Miranda, 2006). Moreover, worldwide, but particularly in Portugal, coastal environments are critically endangered by flooding and erosion, which will be expectedly intensified by the climate change effects, consequently placing more emphasis on the need of an urgent holistic approach to a successful adaptive coastal governance (Taveira-Pinto et. al., 2021, Vousdoukas et al., 2020Schmidt et al., 2013, Coelho et. al., 2009 ).
The achievement of updated and reinterpreted regional to localscale information, on long-term coastline evolutionary trend, is particularly valuable in areas with relevant erosion as the safety measures to apply in those sectors should be strongly supported and backed by scientific information. That is the case of the centre of Portugal western coast, as shown in previous studies, which is significantly endangered by erosion, with average values of rate of change, varying between -0.19 and -0.91 m/year (Lira et al., 2016).
In order to refine and contribute to the knowledge of this peculiar coastal sector, a high-resolution coastal evolution assessment of the sandy coastline sectors from Figueira da Foz to Nazaré, for a 68-year period and with an approximately 10-year interval, has been determined within the Programme "Geological and Coastal Hazard Mapping at a 1:3000 resolution scale" of the National Laboratory of Energy and Geology (LNEG) (Nave and Rebêlo, 2018). It is expected that the produced dataset constitutes a valuable tool and a contribute to support coastal managers and users of littoral regions.

REGIONAL SETTINGS
The western Portuguese coastline is fully exposed to the strong energetic marine wave climate, which, in combination with an NNW-SSE orientation, is powered by a high magnitude of potential of longshore drift (of the order of 10 6 m 3 year -1 ) (Santos et al., 2014).
According to the geomorphological characteristics and sedimentary dynamics cells that characterize the continental Portuguese coastal classification (Andrade and Freitas, 2002;Santos et al., 2014), which divides the coast in 8 cells, the study area (figure 1) is located at the western side within the designated first cell (which extends from the mouth of the river Minho to Nazaré). Cell 1 was, in turn, divided into 3 subcells: from Minho to Douro (1a), from Douro to Cape Mondego (1b) and Cape Mondego to Nazaré (1c) (Santos et al., 2014), being the last sub-cell extent, the focus of the current work. The intense and growing human activity on the coast, and also on hydrographic basins, has led to a sharp reduction in the sedimentary supply (Veloso-Gomes et. al., 2004, Oliveira et al., 1982), to which was associated an erosion tendency of specific segments of this cell, namely Espinho -Furadouro, Costa Nova -Mira and Cova Gala -Leirosa. The Figueira da Foz -Nazaré littoral sector comprises, mainly, three types of environments: low-lying beaches, where 72 % of them are backed by dunes (Lira et al., 2016), hardened coasts and cliffed shores, bordered, or not, by narrow beaches.
Although showing an apparent continuity, a diverse coastal geology plays an important role in coastal processes and hence, in the different existent coastal morphologies.
At Cape Mondego, the coast is rocky, with cliffs and a marine abrasion platform carved mostly in limestones, passing progressively to a sandy beach, which become extremely wide, north of the Mondego River, due to the sand retention on the northern breakwater of the Figueira da Foz harbour (Santos et al., 2014).
From the Mondego River mouth to São Pedro de Moel, a straight low-lying sandy coast prevails, with wind-blown sand migrating inland. A foredune or, more rarely, a sequence of foredunes, backed by an extensive transgressive dunefield is the most common geological setting. Rivers play a local role in shaping the coastline and its change. In some sectors a morphological depression between these two different dune systems exists, which favours the marine intrusion when the foredune ridge is eroded. Some foredunes ridges are related with fences, deployed in the recent past as sand traps. Those structures are presently visible in the actual dune erosion scarps and are a good indicator of coastline retreat.
São Pedro de Moel plays nowadays an important role in the coastline evolution for several reasons: (1) it acts as a natural groin, anchoring the shore position and promoting updrift accumulation; (2) south of this locality the coast is dominated by a rocky shore; (3) south of this point the higher altitude of the bedrock, if compared with the northern sector, gives rise to a cliff coast bordered by a narrow beach and (4) the effect of a slightly northward orientation of the coast, induces a small increase in the potential of the coastal drift (Santos et al., 2014), which may contribute, in some sections, to the presence of a small beach, limited by cliffs and promontories. Although the transgressive dunes continue to exist as a top cliff dunefield, the continuity between beach and the dunefield is generally interrupted as a result of a smaller sediment availability. Aeolian ramps and climbing dunes, which allows the inland sand migration, are only visible at the south of the study area, north of the Nazaré promontory (an approximately 3,7 km coastal stretch), as a result of sand accumulation against the natural cliff, even in scarce sedimentary periods, as presently.
The two main sources of sediments to this sector (sub-cell 1c) are related with the Cape Mondego sediment bypassing and the Mondego River sediment discharge. The sediments that cross the Mondego Cape are transported southward along Buarcos beach to Figueira da Foz beach. Here, its natural course is conditioned by the northern breakwater of the Mondego River, which promotes a sedimentary retention northward that structure, causing the observed exceptional growth of Figueira da Foz beach from 1960 to 1980 (Dias et al., 1994in Santos et al., 2014. These promoted coastal morphology and river flow changes, alongside with the updrift sediment retention, induced a noticeable retreat of the coastline southward, on the Cova Gala -Pedrogão section, which led, as a safety measure, to the construction of a set of rigid coastal protection structures. As result of a new 400 m expansion of the north jetty, that occurred between 2008 and 2010, there was a new increase in the northern beach width, exceeding 500 m. This increased retention reinforced, once again, the previously installed erosive trend on the southward sectors, especially at Cova Gala, Lavos and Leirosa, which likely extends even further south (André and Cordeiro, 2013).
As a result of the observed erosive process, mitigated by the dredging promoted by the port with sediment deposition near shore of Cova Gala, it is estimated that the sedimentary transposition is slightly more than half of the potential drift (6 x 10 5 m 3 year -1 ) (Santos et al., 2014). However, following the saturation of the north jetty, it is expected that the volume of sediments that (naturally and artificially) cross the inlet tends to increase, reducing the observed severe erosion process.
The southern limit of the study area, which coincides with southern limit of the cell 1c, is characterized by two major morphologic features that have a major influence in the littoral drift: the Nazaré promontory and the Nazaré canyon.
The Nazaré promontory acts like a natural groin and is responsible for an updrift sand accumulation. The Nazaré canyon, that extends from a water depth of 50m near the Portuguese coast to 5000 m at the edge of the Iberian Abyssal Plain. Even though the canyon constitutes a giant sedimentary sink, its influence on sediment transport is up to the beach (Beach to Canyon (2013). Actually, the canyon system is not draining sediments from a modern river obtaining, instead, its present-day sediment input by capturing of along-shelf sediment transport and depositing much of it in the middle canyon, between about 2700 and 3800 m (Masson et al., 2011). Thus, most of the sand transported southward in the coastal drift (11 x 10 5 m 3 year -1 ) is conducted, and therefore totally lost, to the deep ocean by the Nazaré canyon system (Santos et al., 2014).

METHODOLOGY
As the produced data intends to be comparable to the previous systematic works carried within the "Geological and Coastal Hazard Mapping at a 1:3000 resolution scale" programme of the National Laboratory of Energy and Geology, the used methodology is similar to the one described at Nave and Rebêlo (2021).

Coastline indicator
As the objective of the current work is to provide high-resolution quantitative data on the long-term (nearly 70 years) coastline evolution, the foredune shoreward edge was used as a marker of the coastline to exclude short-term (tidal) variations or mediumterm (seasonal) variations, as described at Nave and Rebêlo (2018) and references therein. In cliff areas, the position of the coastline was marked using the boundary of the cliff base (Rebêlo and Nave, 2019).
The continuous coastline position was smoothly depicted at the maximum scale of 1: 800 to undoubtedly identify the outer limit of the dune zone. Interruptions in the vegetation line of less than about 150 m were not considered, maintaining in this case the continuity of the line between adjacent segments. In the case of settlements or large infrastructures that hindered the expansion of the dune field, the continuity of the line between adjacent segments of the interruption was also considered (Rebêlo and Nave, 2019).
In order to determine the error inherent to the operator, for each year, the coastline was marked by two different operators using the same criteria. The coastline used for each year thus corresponds to the average position of the lines calculated between the two operators (Rebêlo and Nave, 2019).

Data sources
Aerial photographs and ortophotomaps from nine different years (1947, 1958, 1973, 1980, 1982, 1988, 2004, 2010 and 2015) were used in order to achieve a near-decadal analysis. The data was provided by Direcção Geral do Território and by CiGeoE (Army Geospatial Information Center). The non-digital aerial photos were scanned at 2400 dpi and then, mosaics were made, using the Adobe Photoshop CS6 image software. This was done with the aim to expedite the georeferencing process, since it allows to avoid marking duplicated links in the overlap area between two consecutive aerial photos (Rebêlo and Nave, 2019).
Additionally, given that the link creation process on the scanned image leads to uncertainties of magnitude similar to the accuracy of the images, the use of mosaics also reduces georeferencing errors. The georeferencing process was accomplished using ESRI software, ArcMap, from ArcGIS Desktop 10.6.1, based on the 2015 orthophotomap as reference map (the most recent one available at the date this work was done). Since the objective was the digitization of the coastline, this process had a greater focus on the area next to the coastline, with increased control points density close to it. Those points were marked, using the Spline interpolation method, preferably with a zoom, at the scale of the order of 1: 400, 1: 600 or, at most, 1: 800 (Rebêlo and Nave, 2019).

Mapping procedures and DSAS calculation of coastline evolution
After determination of the average coastline position for each of the nine historical aerial photographs, coastline evolution was calculated using the Digital Shoreline Analysis System (DSAS) Software, that is an add-in within the Environmental System Research Institute (ESRI) ArcGIS© ArcGIS desktop v. 10.4-10.6. It, thus, enables a user to calculate a range of statistical change measures derived within DSAS, based on the comparison of coastline positions through time (Himmelstoss et al., 2018); (https://www.usgs.gov/centers/whcmsc/science/ digital-shoreline-analysis-system-dsas?qt-science_center_ objects=0#qt-science_center_objects).
For this work, DSAS allowed us to calculate coastline rate-ofchange statistics based on the measured differences between nine coastline positions associated with analysed time periods (1947, 1958, 1973, 1980, 1982, 1988-90, 2004, 2010, 2015), producing three parameters: (i) Shoreline Change Envelope (SCE), which is the measure of the total change in coastline movement considering all available coastline positions and reporting their distances, without reference to their specific dates; (ii) Net Shoreline Movement (NSM), corresponding to the distance between the oldest (in this case, 1947) and the youngest coastline (2015). For the present work the NSM refers to an overall time-frame of 68 years; (iii) Linear Regression Rate (LRR), which is a statistical parameter, that determines a rate-of-change by fitting a least square regression to all coastlines at a specific transect.
Calculations were done using 10 m spaced transects, generated perpendicular to the coast, and the results are represented as a graphic and plotted as a histogram. The coloured histogram, represents two DSAS parameters simultaneously, allowing a fast perception of the coastline evolution over time. The bar length indicates how much the coastline has shifted during the period of analysis (SCE parameter), and the bar colour indicates the NSM and LLR, with the colour (green, yellow, and red) being associated with the evolutionary trend of the coastline. The green scale stands for sectors in accretion and the yellow and red ones, for the sectors in erosion. The absolute value of this variation is shown in the legend of each figure. Due to the scale of the figures presented in the manuscript, the bar length variations are difficult to distinguish. However, they are very well visible at online LNEG GeoPortal, where zoom favours both, the bar length and colour variation perception.
Combining both variables visually, the user may rapidly deduce if, although the coast today is in the same or similar position than it was in the past, it has changed over time, and how much it has changed, in absolute value.
The calculation of the uncertainty of the coastline position, for each year, was determined according to Fletcher et al. (2003), where the Uncertainty (U) is the result of the square root of the sum of the squares of the errors associated with 3 parameters (Resolution of the aerial photo image (E ir ), Image rectification (E r ) and Digitization of the coastline (E d ): The component associated with the coastline digitization (E d ) procedure, for each year, was determined with DSAS using the two coastline lines marked by the two different operators for each analysed year. The image resolution, E ir parameter, corresponds to the pixel resolution of aerial / orthophotographs. Given that for the rectification procedure (E r parameter), the Spline interpolation method, whose associated RMS error is null, was used, the georeferencing error was estimated by using, at least, 35 control points randomly distributed across the mosaics of each flight. Thus, the rectification error (E r ) was assumed to be the average value of the error measured at all control points, for each flight, as listed in table 1.
The uncertainty associated with the overall coastline evolution (U ce ) during the time frame of 68 years (t), was determined using the uncertainty values (U) calculated for each year according to Fletcher et al. (2012): For this sector, the overall uncertainty (Uce) associated with the variation of the coastline along 68 years is 0.06 m / year.

Coastline oscillation index (CO)
The ratio between the absolute value of the Net Shoreline Movement (NSM), which represents the resultant movement of the coastline during the time-period of analysis, and the Shoreline Change Envelope (SCE) value, which gives the maximum length of coastline displacement during the same period of time (Absolute value of [NSM/SCE]), is an index that, when used in comparison with the NSM, may indicate the importance of the coastline oscillation in a particular place, or sector, or, read in a different direction, the relative importance of the NSM in terms of the total displacement during the timeperiod of analysis. The CO index value should always be used and read together in comparison with the NSM. As previously mentioned, the colour and the length of the bar displayed on both hazard maps (figures 2 and 3) give, simultaneously, information on two DSAS parameters (SCE and LLR on figure 2 and SCE and NSM on figure 3), allowing a fast perception of the historical coastline evolution trend during the last, nearly, 7 decades. Thus, the bar length indicates how much the coastline has shifted (SCE), and the colour of the bars, NSM and LLR, stand for sectors in accretion (green scale) or sectors in erosion (yellow and red colour scale). The LRR map, represents the rate-of-change by fitting a least square regression  to all coastlines at a specific transect (figure 2) whereas the NSM map indicates the absolute distance between the oldest (1947) and the youngest (2015) coastline position (figure 3). All the data is available as polylines with 10 m spaced resolution and openly accessible at LNEG GeoPortal [https://geoportal. lneg.pt/mapa/?escala=4000000&mapa=geologiacosteira#].

Historical lines,
Here the produced digital data may be seen in a greater detail and combined with other available layers for this coastal sector (beach profile, sedimentary grain size distribution, photos from the beach area, geology, and others).
Coastline movement analysis, based on the NSM parameter, reveals that 2/3 of the sandy coast between Figueira da Foz and Nazaré has suffered erosion since 1947 (30.4 km of the 51.5 km total extension). The average retreat, taking only in account the sectors in erosion, is approximately 38 m, with a maximum retreat of 145 m, and the progradation, taking only in account the sectors in accretion (20.1 km in length), is approximately 22 m, with a maximum displacement of 322m. Solely one kilometre of the coastal fringe is in the same position, when compared to 1947. If the overall coastline extension is considered (51.5 Km), an average erosion of 13.5 m is obtained.
The coastline movement rate, which is expressed in this work by the LRR parameter, shows an average retreat of 0.51 my -1 , with a maximum value of 2.46 my -1 , and an average progradation of 0.43 my -1 , with a maximum value of 5 my -1 (taking only in account the respective sectors in erosion or accretion, respectively). Taking into account the total studied coastal fringe (51.5 km), the obtained results indicate an average erosion rate of 0.12 my -1 .
To a better reading of the DSAS parameters, the Figueira da Foz -Nazaré coastal stretch, fully shown in figures 2 and 3, was divided and zoomed in 9 sequential figures, and hence the detailed results, will be presented in the following sub-chapters, which division reflects merely the pragmatic circumstance of map division.
The oscillation index for each sector (table 2) was also calculated to weight the NSM value in a maximum coastline displacement context.

Figueira da Foz -Costa de Lavos coastal sector
This first sector, with a total length of 5860 m (figure 4), has a significant diversity in terms of coastal environments, and hence, the evolution of the coast reflects these differences.
The first subsector, with a 380 m length, is located between two groins of the

Costa de Lavos -Osso da Baleia (N) coastal sector
This sector (figure 5) may be divided in two parts, regarding coastline retreat: The Costa de Lavos -Leirosa subsector and the subsector south of Leirosa. The first sub-sector is located between two groins and the strong erosional behaviour, southward the last Costa de Lavos groin, diminishes gradually and shifts to accretion as we approach the Leirosa groin, to the  (table  2) shows a small decrease in the ratio, however, like in the previous sectors, the NSM values may be considered similar to the maximum range of displacement (given by the SCE index).

Pedrogão coastal sector
In opposition with the preceding sector, accretion prevails in this coastal stretch ( In this sector, SCE and the absolute value of NSM are significantly The CO index, [Abs (NSM)/SCE], for this sector is 0.5 (table 2) as a consequence of SCE being 1.9 times bigger than NSM.

Samouco coastal sector
The erosional trend detected southwards Praia da Vieira continues to the initial part of this sector, with erosion prevailing over the accretion (figure 9) changing, then, gradually towards an accretion tendency, which dominates throughout the south part of the sector. Therefore, in the first 2900 m, where erosion prevails, the average NSM is -11.15 m, while in the last 4260 m, where progradation prevails, the average NSM has a value of 9.87 m. In this sector, average SCE (21.9 m) is 2.1 times larger (table 2) than the absolute value of NSM (10.4 m).

São Pedro de Moel coastal sector
The São Pedro de Moel sector was not fully covered by DSAS analysis, as may be seen in figure 10, because sandy beaches backed by foredunes are not present in some parts of the sector.

Paredes da Vitória coastal sector
As in the previous sector, the Paredes da Vitória coastal sector (figure 11) was not fully covered by our analysis since, although beaches are always present, they are baked up by cliffs and not by dunes, hampering these sectors to fulfil the defined criteria for coastline delimitation.
The analysed parts of the sector, with a total length of 2390 m, show mainly a stable coastline position with a small accretion tendency, occupying 1990 m of the coastal stretch.
NSM ranges from -9.4 to 36.5 m, with an average of 7.3 m and LRR ranges from -0.06 to 0.61 my -1 , with an average value of 0.15 my -1 .

Vale do Pardo -Nazaré coastal sector
The

The importance of coastline evolution trend estimates, at a high-resolution scale
The cartographic assessment of the European coasts exposure to coastal erosion based on spatial data and GIS analysis has been, since the last decades, a mutual international goal as all European coastal states are, to some extent, affected by coastal erosion (Commission, 2004;Ferreira and Matias, 2013). The need of acquired dataset and its analyses, first at a European scale (Ferreira and Matias, 2013), and thereafter at national scale (Lira et al., 2016) has been, since then, successfully achieved for the Portuguese scenarios.
However, even though a small-scale dataset is vital for addressing mitigation measures for coastal erosion, the further achievement of updated and reinterpreted regional to local- scale information, on long-term coastline evolutionary trend, is particularly valuable in areas with strong human occupation and relevant beach erosion, as the mitigation and defence measures to apply in those sectors should be robustly supported and backed by detailed scientific information. This would strongly contribute for the planning and implementation of effective adaptive measures on the climate change scenarios particularly in regions significantly endangered by erosion.
When comparing to previously published results produced at a national scale (Lira et al., 2016), the erosion and accretion trends are similar, although the mean coastline rates of change are smaller in the present work and differences are noticed locally with maximum retreat or accretion rates identified in distinct areas.
The differences on local coastal Linear Regression Rates detected at this systematic and high-resolution study, in comparison to the previous national-scale dataset (Lira et al., 2016), put in evidence the advantage of the achievement of continuous updated and reinterpreted high-resolution data at a local scale, as a zoomed analysis likely favour to apply the mitigation measures (as sand nourishment and others) at the exact key target locations.

Coastline evolution trend of the last, nearly, 7 decades at the western sector from Figueira da Foz to Nazaré
The sedimentary flux, already diminished by the reduced fluvial source at the northward limit of cell 1 (i.e., sub-cell 1a), has long been the central issue that explains the generalized retreat of the western coastal shoreline (Santos et al., 2014). The average rate of coastal retreat, NSM, of the last 68 years, calculated in this work (-13.6 m) for the sandy coastline, backed up by dunes, from Figueira da Foz and Nazaré sector, confirms the previously reported readjustments of the coastal system due to It is also important to stress that the NSM data, which in this paper is mostly used to quantify the coastline movement, between the oldest and the most recent coastline position, should be complemented by the SCE to fully understand the coastal behaviour. In the analysed sector, the calculation of the total area associated to the coastline movement resulting from the NSM determination, which is 1 627 265 m 2 , is largely exceeded by the area obtained by the SCE value (2 466 377 m 2 ). Thus, the SCE area, being 52% larger than the NSM area, means that the overall coastline oscillation during the period of analysis was significantly larger than the resultant displacement given by the NSM. This relevant information should be considered when managing local coastal activities, as the SCE index denotes a greater oscillation of the coastline position, in comparison to the NSM index.

Figueira da Foz -Costa de Lavos coastal sector
At the northern limit of this sector there is a small beach anchored between the two southern jetties of Mondego River mouth. The evolution of this sector is strongly influenced by the southern jetties of the Figueira da Foz inlet. In 1947, a jetty, in the left side of the river, was already in place and downdrift erosion was already present, being visible by the liner aspect of the foredune/beach contact. In 1958 the erosion was even more evident, with foredune destruction and occurrence of overwash at backdune deposits, northward of Cova Gala. Between 1958 and 1973 a new and longer jetty was built approximately 500 m further south and, in the right margin of the river, a new jetty was also built (the jetty responsible for the Figueira da Foz beach accumulation). The installation of these two long jetties is the major responsible for the beach accretion.
Between 1982 and 1990 the first groin was expanded, together with the placement of a new one in the right margin, in order to straighten the inlet channel. When analysing the coastlines dataset, the influence of these coastal structures ( South of this beach, this sector presents a worrisome erosion, along a 5400 m coastal strip, that is only reversed at Costa de Lavos area, where a groin, placed between 1973 and 1980, helped to stabilize the coastline position. The reduction of the longshore drift led to a continuous inland coastline shift and to the need to deploy coastal defences in front of Cova Gala. The five groins, complemented by seawalls, were built between 1973 and 1980 and manage to reduce the coastline retreat, although the protected coastal stretch continued to be vulnerable to overwash, during storms. The 720 m protected seafront suffered As previously mentioned, a small accretion area is observed at the southern limit of this sector. Along this 500 m end-stretch, a tenuous average accretion (NSM) of 10.6 m, and an average progadation rate of 0.12 my -1 is measured. This accretion is related with the Costa de Lavos groin, built between 1973 and 1980 with the aim to stop the prevailing erosion.
Human interference induced in coastal dynamics through the introduction of rigid constructions was not able to stop erosion completely and, as a consequence, beach and dune nourishment was carried out. Between 1973 and 1976, 294 020 m 3 of sand were deposited on Cabedelo beach (table 3). The reinforcement, with a later nourishment, in 2014, at Cova Gala and Lavos beach (table 3) (table 2), the value of 0.9 for this sector indicates that NSM gives a good indication of the coastal evolution behaviour during the period of analysis (figure 13), since both lines are almost coincident.

Costa de Lavos -Osso da Baleia (N) coastal sector
This segment (figure 5) is a good example of the effect of perpendicular coastal defences deployment in a longshore drift coastal environment exposed to a negative sedimentary budget. As N-S longshore drift prevails in the study area, groins deployed perpendicular to the coast promote accumulation of sand updrift (in the present case, towards the north of the structures) and erosion downdrift (to the south of the structures).
Therefore, southwards of Costa de Lavos groin, erosion is intense (higher value of NSM of -117.8 m, with a LRR of -1.68 my -1 ), diminishing gradually in the direction of the Leirosa groin, where progradation occurs (higher value of NSM of 77.6 m, with a LRR of 0.9 my -1 ). South of this groin, erosion occurs again (higher value of NSM of -103.7 m, with a LRR of -1.37 my -1 ) (figure 5).
It is also important to stress that, despite the deployment of coastal defences, erosion continued to occur. To mitigate the erosion effect, an artificial nourishment of sand was done in 2014 at Leirosa beach (table 3), but it was not enough to overcome the erosional trend, as the 2015 coastline position is still the one placed at a more inland position (figure 5).
Regarding the CO index, [Abs (NSM)/SCE], (table 2), the value of 0.9 for this sector indicates, like in the previous sector, that NSM is a good indicator of the coastal evolution behaviour during the time period analysed ( figure 14).

Osso da Baleia coastal sector
Erosion prevails in the Osso da Baleia sector (approximately in 94 % of its length), extending the trend described in the southern part of the preceding sector. Although having an average coastline retreat of -47 m (and an average LRR of -0.51 my -1 ),  no intervention has been made on the coast, since there is no human occupation to protect. Accretion is solely present in the last 400 m of the sector.
The observed erosion is likely the result of the diminishing of the longshore sediment drift (Santos et al., 2014). In fact, if we consider a larger area than this sector, from Leirosa (in the precedent sector) and Pedrogão (a natural rocky outcrop, in the next sector), the process is similar as it happens between groins: Erosion in the updrift side (north), changing gradually to accretion (which happens in this sector), to a more notorious accretion environment in the southern part, near the natural sand trap obstacle, the Pedrogão promontory.
Regarding the CO index, [Abs (NSM)/SCE], (table 2), the 0.8 value indicates, like in the previous sector, that the NSM gives a good indication of the coastal evolution behaviour during the period of analysis. However, a decoupling occurs at the end of this segment ( figure 15).

Pedrogão coastal sector
The accretion tendency observed from the southward area of Osso da Baleia beach continues to the Pedrogão sector ( figure 7), being present in 73 % of its length.
From the comparison of SCE and the absolute value of NSM ( figure 16), it is noticeable that the two variables differ substantially, being the distance from the oldest (1947) and the most recent (2015) coastline (average of 12.3 m), significantly smaller than the shoreline envelope, SCE (average of 50.0 m). Analysing the coastline positions, an offshore displacement tendency is observed from 1947 until the 70's and 80's, followed by a more recent onshore displacement.
Although the Pedrogão promontory, acting like a natural groin, may be seen as responsible for the observed updrift sand accumulation, the coastline instability near that natural structure is significant and erosion is dominant. A road and  housing were built on top of the foredune and, due to the dune erosion and associated overwash events, a seawall had to be built to protect those structures. Human activity and hard coastal stabilization may have contributed to the unrecovered state of the dune in that particular area.
Southward of Pedrogão erosion prevails. This may be interpreted as a result of the littoral drift diminishment during the analysis period.
Regarding the CO index, [Abs (NSM)/SCE] (table 2), the 0.2 value indicates that the NSM has to be used carefully when analysing the coastal evolution behaviour during the analysed time period. SCE is consistently higher than NSM which means that coastline oscillation is significant, playing an important role in this sector (figure 16).

Praia da Vieira coastal sector
The most notorious changes in this sector are related with the Liz River mouth, which has been artificially fixed, with two jetties, between 1947 and 1958. These structures induced coastal changes, being that the intensity of the variations diminish with the distance to the inlet.
As previously mentioned, this sector reverts gradually from erosion (related with the Pedrogão promontory) to accretion, as we move south in the direction of the Liz River north jetty. The sand accumulation against the jetty starts to be noticeable approximately 1700 m north of the structure. Near the north jetty, accretion reaches 154.2 m in contrast with 26,3 m average of the progradation values observed in the sub-sector, showing the importance of the structure in the longshore drift retention.
Southward of the jetties, as often happens, the erosion reappears, occurring in 3130 of the 3440 m remaining sector length. Immediately south of the second jetty, the erosion reaches -84.2 m in contrast with the sub-sector average, which is -17.1 m, illustrating, once again, the erosional effect of these structures in a longshore drift coastal sand transport.
Erosion values southward of the jetty could be even higher than the ones obtained with the DSAS analysis. Praia da Vieira is protected by a seawall and artificial nourishment was already applied to minimize and prevent the coastline inland displacement.
As previously mentioned, the CO index, [Abs (NSM)/SCE], for this sector is 0.5 (table 2). This is a consequence of SCE being 1.9 times higher than the Abs (NSM). However, in the present case, oscillation shouldn't be regarded with great concern once that the CO index values are significantly smaller. Although the ratio is high, the SCE average is only 37.7 m and the average for the absolute value of NSM is 19.8 m, for the last 68 years. Looking to figure 17, it may be concluded that this is a relatively stable coastline, and the oscillation is only more significant in the last 300 m of the sector.

Samouco coastal sector
The Samouco littoral sector (figure 9), with slight erosion prevailing in the northern part and progradation in the southern part, is the steadiest sector of the study area, with mean LLR of 0.06 my -1. This sector is influenced by two features that lies outside its boundaries: to the north, the Liz River jetties and, to the south, the São Pedro de Moel promontory. As seen in similar northern sectors, also influenced by hard structures, erosion tends to occur downdrift and sand accumulation tends to occur updrift those structures. This sector, being located between those two hard structures, is in the transition zone, what may explain its stability. In this segment, the coastline oscillation follows the same pattern as the one of the southern limit of the predecessor sector, with a slightly decoupling of the SCE and Abs (NSM) indexes (figure 18). With a CO ratio of 2.1, caution should be taken in the analysis of coastline displacement, based solely in the NSM values. However, as coastal changes absolute values are small, with an average NSM of 10.4 m and an average SCE of 21.9 m, the ratio shouldn't be regarded with great concern.

São Pedro de Moel coastal sector
An SCE value approximately 2.5 times larger than the NSM (table  2) is an indication that we are in presence of a coastal stretch where oscillation plays, through time, an important role in the characterization of coastal evolution. The São Pedro de Moel rocky shore, already eroded and aligned with the general coastal orientation (figure 10), act as a natural groin, resulting in the updrift accumulation of sand, and establishing a maximum seaward coastline position prior to bypassing occur.
In a finer analysis, the S. Pedro stream also plays an important role in local coastal behaviour. To the north of the stream mouth, nearby Praia Velha (located around the 161 (m x 10), in figure 19), the mean SCE value is 19.1 m, while southwards of this beach, the SCE average value is 57.7 m. This higher value is likely related with the observed shift of the mouth stream from 1947 to 2015.
Although the observed coastline oscillation is important in terms of coastal evolution analysis for this sector [Abs (NSM)/ SCE ratio of 0.4, table 2], the LRR value of 0.29 my -1 , with shifts between eroding and accretion coastal stretches, indicates that this sector is relatively stable, and the occurred changes may be regarded as part of the natural coastal evolution process in a longshore drift, high energetic coast.
South of São Pedro de Moel, the Valeiras beach, with 570 m length, has an average NSM value of -3.7 m, showing a slight erosional trend (LRR of 0.08 my -1 ). However, coastline oscillation has also to be considered, since the average SCE is 16.2 m.

Paredes da Vitória coastal sector
With an SCE value approximately 3 times larger than the NSM, oscillation plays, as in the preceding sector, an important role in the characterization of coastal evolution. Progradation values (NSM average of 7.3 m and LRR average of 0.15 my -1 ) should, thereby, be valued within this context.

Vale do Pardo -Nazaré coastal sector
Although the slightly northward orientation of the coast at these sectors may induce a small increase in the potential of the coastal drift (Santos et al., 2014), sand accumulation occurs, confirmed by the presence of wide beaches and well developed dunes (especially in the southern part), where, due to the existence of an aeolian ramp, a sand continuity between the beach/dune system and top-cliff dunes are observed.
Like in the several existent rocky promontories in the coast northward this sector, the Nazaré headland has an important role in sand accumulation due to the littoral drift, being the North beach a good example. However, in this sector, the coastline has a peculiar shape (figure 12), drawing a curve and extending seaward, approximately 1 km to the north of the promontory, surpassing what was expected to be the natural coastline orientation. This singular shape of the coastline is due to the effect of the Nazaré canyon which tends to rotate the waves orientation anticlockwise (Forone, 2014), promoting a higher accumulation in that area.
This sector, where accretion prevails, may be divided in two subsectors in terms of evolutionary trend. In the first 2470 m of coast, NSM values reflects an oscillation between erosion and accretion. Also in this first subsector, the CO index, [Abs (NSM)/ SCE], has a value of 0.3, which reflects a very strong oscillation/ instability of the coastal position. In the following 3090 m, accretion is dominant and the Abs (NSM)/SCE ratio has a value of 0.7. However, although ratios are very different, the average difference between the SCE and Abs (NSM) is approximately the  , showing the precaution that has to be taken when using this ratio alone for coastal displacement analysis: depending on the NSM and SCE absolute values involved, very different ratios may be obtained.

Land loss by coastal erosion
The obtained data may also be used to quantify the lost and gain of territory due to the coastline shift during the 68-year period of analysis. Considering the positive and negative coastline shifts from 1947 to 2015, data shows that 1 164 888 m 2 of land were lost, along 30 470 m of coast, and 462 330 m 2 were gained, along 21 010 m of coast. The overall result may be expressed as a loss of territory of 702 558 m 2 . The analysis may also be made by sectors, as shown in Table  4. As a result, 4 of the 9 sectors show erosion, while 5 show accretion. Table 4 also displays a clear predominance of erosion in the northern sectors, despite the artificial beach nourishment performed in the northern part, versus a clear tendency for accretion in the southern sectors, during the time-period of analysis.

CONCLUSIONS
The current work brings new and detailed insights on the historical evolution trend of the western Portuguese littoral, occupied by beaches with dunes, from Figueira da Foz to Nazaré. Being part of a broader national programme at LNEG [https://geoportal. lneg.pt/mapa/?escala=4000000&mapa=geologiacosteira#], this study achieves, for the first time, a long-term (nearly 7 decades, from 1947 to 2015), high-resolution (10 m spaced), systematic (with an approximately 10-year analysis interval) and continuous coastline evolution analyses, using different sources of aerial imagery and USGS DSAS software to calculate coastline displacement.
The applied DSAS methodology, widely used and recognized to be appropriated for the current propose, makes the presented dataset easily comparable to other previously acquired datasets, with the advantage of, being a continuous and a high-resolution dataset, becoming more adequate to support the planning and efficacious implementation of adaptive measures, not only in a regional, but also on a local scale. Thus, the current results are valuable, in a scientific perspective, but also, and perhaps most importantly, in an applied context, as the understanding of long-term and mesoscale shore evolution of the Portuguese western coast, is needed for the planning and implementation of effective adaptive measures on the climate change scenarios, in particular in this region, significantly endangered by erosion.
Coastline evolution results for the 68 years' time window, obtained in the coastal stretch comprising beach and dunes (51 480 m of the total 63 408 m coastline extension from Figueira da Foz to Nazaré), reveals an erosional trend, with an average retreat of -13.6 m, despite the artificial beach nourishments deployed in the area as presented in table 3 (759 650 m3). This erosional trend is also shown by the 702 558 m 2 net land loss area during the analysed period. However, and looking only for the coastal sectors where erosion occurred, a total of 1 164 888 m 2 of land loss was observed.
Erosion, that is more severe in the northern part, reaching a maximum coastal retreat of -145 m and an erosion rate of 2.46 my -1 south of Cova Gala, is due to a negative sedimentary budget that affects this sector (Santos et al., 2014). The erosional pattern seems to be induced by human interference in coastal dynamics, namely by the introduction and enlargement of the original rigid constructions at Figueira da Foz harbour and by the Cova Gala, Costa de Lavos and Leirosa groins installation. In these later, the slight updrift accumulation is followed by a much expressive downdrift erosion, the typical pattern of coastal areas subjected to a negative sediment budget littoral drift, with maximum retreats of, respectively, -145, -118 and -103 m.
Seaward displacement of the coastline, occupying approximately 40.8% of the studied area, are mainly present in the southern part, reaching a maximum of 81 m of progradation at Praia do Norte, near Nazaré, although they are also associated with updrift accumulation against the installed protection groins in the north and central part of the study area (maximum values of 322 m between the Mondego jetties, 18 m at Costa de Lavos, 78 m at Leirosa and 154 m at Praia da Vieira).
The importance of coastline evolution trend evaluation, at a high-resolution scale, was somehow evidenced by the current work. The advantage of the achievement of continuous updated and reinterpreted high-resolution data at a regional scale, likely favour successful application of the needed mitigation measures (as sand nourishment, and others) at the exact key target locations.