Evolution of the Beach-Dune Systems in Mediterranean Andalusia (Spain) Using Two Different Proxies

1. Introduction

Coastal environments are complex systems that are influenced by many interrelated physical, chemical and biological processes [1,2,3]. In developed coastal areas, human pressure represents an added threat to valuable natural environments [4,5,6,7] and, at many places, the risk of coastal erosion/flooding represents a severe problem for human settlements [8,9,10].

Scientific studies on coastal erosion/accretion processes have proliferated during recent decades as a result of the increased human interest related to coastal developments and infrastructure [11,12,13,14] and the effects of climatic change-related processes [15,16,17] such as sea level rise, increased height of extreme waves or changes in the frequency and intensity of storms [6,7,18,19,20,21,22,23,24,25,26].

Many different proxies are used for shoreline change analysis depending on the particular coast study and its environment, the availability of data sources and the aim and expected outcomes of the study [27,28], and several studies have compared different proxies in coastal evolution studies (e.g. [27,29,30,31]. The various shoreline proxies used to determine sandy beach evolution [32,33] only capture the dynamics of a certain portion of the coastal area [34]. The use of more than one indicator is recommended by many authors [28,35,36,37] as each proxy captures different processes in different areas of the coastal system, providing a more nuanced picture of the morphodynamic behavior and processes in a given area.

An appropriate time scale approach is also necessary to avoid associated errors, e.g. errors in the definition and extraction procedures [27,28]. Depending on the purpose of the study, different time scales can be selected, e.g., if the aim of the study is the observation of interannual changes, the analysis of shorter time periods can be sufficient, but if the aim is to predict future shoreline trends, larger time periods (>60 years) are needed [38,39].

The dry beach represents a buffer zone that absorbs, reflects and dissipates energy delivered by waves to the shore specially during storm events, protecting in this way the areas behind it from the impact of erosion and flooding [40] and, therefore, dry beach characteristics have to be taken into consideration in coastal sensitivity determination [41]. In addition, foredunes are one of the most relevant coastal ecosystems that work as natural defenses able to reduce flood sensitivity/vulnerability [42,43,44], as they often protect large sections of low-lying coasts against flooding during extreme storms [45,46,47]. Therefore their maintenance/emplacement has been considered as an effective coastal protection measure that is included among possible “Disaster Risk Reduction” (DRR) strategies in several European directives [43,48,49,50,51]. Many authors agree that the temporal natural variations of sediment supply and wind regime are among the most important factors controlling the natural beach-dune system relationship [52,53,54,55,56]. Human impacts on the beach-dune system arise from urbanization and decreases in sediment supplies (by, for example, construction of dam and coastal protection structures [53,57,58,59,60,61,62].

Irrespective of tidal range, beach and dune responses to erosion processes are different [27]: dune erosion can be very fast and episodic, and recovery can occur over several years; beach erosion is more dynamic and variable (interannual cycles of erosion/accretion) and recovery can be fast, i.e. days/weeks [63,64].

This paper investigates the evolution of the beach-dune systems of the Mediterranean coast of Andalusia (Spain) in the medium-term (between 10 and 60 years, [38]), using two different proxies: the dune toe was used to determine foredune evolution and the high-water line was used to assess shoreline evolution. The results obtained for each proxy where then combined to obtain different beach-dune systems states, i.e. from “Erosive” and “Mixed” to “Accretion” states. The results obtained provide enhanced information for coastal managers and administrations to properly understand and deal with contemporary erosion processes. The method used in this investigation can be applied at other areas around the world where a similar data base is available.

2. Geographical Setting

The 564 km-long Mediterranean coast of Andalusia is located in southern Spain and administratively belongs to Cádiz, Málaga, Granada and Almería provinces (Figure 1). It has a rectilinear E-W orientation, with two NE-SW facing sectors located at its western and eastern ends.

The coast is micro-tidal (tidal range <0.2 m) and is mainly composed of beaches of medium to coarse dark sand and/or pebbles. Foredunes are especially well developed in Cádiz and Almería provinces and extend along ca. 76 km of coastline [58,66,67,68].

The Betic Range, a tectonically active mountain chain that, at places, reaches to >2000 m elevation close to the coast, determines coastal orography and morphology, forming cliffs, embayments and promontories. Several small coastal plains are present at the mouth of short rivers and seasonal streams.

Large coastal towns include Málaga (>500,000 inhabitants), Marbella (150,000 inhabitants), Fuengirola (80,000) and Torremolinos (70,000). Málaga has the most densest coastal occupation in Andalusia due to the development of national and international tourism [69,70]. Along the coast there are several marinas and the main commercial port is located at Málaga [53,60,71].

The coast of Málaga province is one of the most heavily developed coastal stretches in Europe. Tourist infrastructure development during the 1960s, without any rational spatial planning, resulted in a large population increase and transformation of the coastal landscape. Although urban developments have produced economic benefits, the environmental impacts in some cases are irreversible and the coastal sediment budget has been altered in most of the sandy sectors [70,71,72].

Cádiz, Málaga and Granada have a Mediterranean climate with “Humid and Subhumid” and “Tempered Dry-subhumid” areas, with average annual temperatures from 15 ºC to 19 ºC and annual rainfall between 600 and 1000 mm [73]. Almería Province has a semi-arid Mediterranean climate with sparse episodes of rain (<200 mm/year in some places, [73]), and average annual temperature of 21 ºC, reaching 26 ºC in summer [74].

The coast is generally exposed to winds blowing from E to W and from NNE to SW in the easternmost part of Andalusia, with minimum and maximum velocities ranging from 0.4 to 9.0 m/s [75]. The wave climate and storm energy are very variable as the coast of Málaga, Granada and (partially) Almería provinces are exposed both to western and eastern storms, and the easternmost area of Almería province is primarily exposed to eastern storms [65,75].

Waves show a clear seasonal behavior with storm conditions being recorded during winter (November–March) [72,75,76], and significant wave heights reach 4.73 m during extreme storms [75]. A storm characterization for the study area [75] using the Energy Flux parameter, classified storm events into five classes, from weak (Class I) to extreme (Class V). The most energetic coast is between Málaga and Almería provinces, [75].

Shoreline orientation, predominant easterly winds and associated storm waves give rise to a prevailing westward littoral drift [76]. An opposing drift is present in some coastal sectors and/or periods [72,77].

3. Materials and Methods

Material and methods used in this work are summarized in Figure 2.

3.1. GIS Project Creation

In this paper, aerial orthophotos from 1977, 2001 and 2019 were used to digitize and extract shoreline and foredune proxies. The orthophotos were obtained from the Web Map Services (WMS) (https://www.juntadeandalucia.es/medioambiente/portal/acceso-rediam/geoportal/servicios-ogc/web-map-service-wms, accessed in March 2023) developed by the Regional Government according to Open Geospatial Consortium interoperability standards. All information was presented in projected coordinate system ETRS89 UTM zone 30N (Reference System EPSG: 25830). The spatial resolution of the orthophotos was 0.5 m for 1977 and 2001 (black and white) orthophotos and 0.35 m for 2019 orthophoto (color). In order to reduce uncertainty related to storm impacts and recovery sequences which could affect the decision of using the selected orthophotos, displacement distances and rates were also calculated for intermediate periods between the selected photographs [77]. Results showed a constant trend confirming the validity of data obtained using the three aforementioned orthophotos. Further, according to [38] the influences of short-term fluctuations, e.g. the effects of high energy storm events, are minimized in the medium-term coastal studies and this is the case of this paper that investigates 24 and 18 year intervals.

The two proxies selected in this study were the high-water line (HWL) for the shoreline and the dune toe line for foredunes [27,78]. The HWL, also called wet/dry line, is identified by a change in the color of the sand. This line corresponds to the last tide mark that does not change very much in the micro-tidal Mediterranean Andalusia coast. The dune toe line was selected as it represents the line that separates the backshore from the shoreward limit of foredunes and can be determined in the orthophotos by a change in color between bare sand and vegetation. The dune proxy was digitized for foredunes > 100 meters in length.

3.2. Change Rate Calculations

Corrections of the shoreline position were carried out according to wave run-up and tidal conditions using the relation of the total uncertainty (σ_T):

$${\mathsf{\sigma }}_{\mathrm{T}}=\sqrt{{\sigma }_d^2+{\sigma }_p^2+{\sigma }_r^2+{\sigma }_{co}^2+{\sigma }_{wr}^2+{\sigma }_{td}^2}$$

(1)

Such total uncertainty combines digitization errors and the photo characteristics (Moore, 2000), i.e. the digitalizing error (σ_d), accuracy linked to pixel size (σ_p), ortho-rectification error (σ_r), image co-registration error (σ_co), and onshore definition and position determination, i.e. wave run-up (σ_wr) and tidal conditions (σ_td).

The uncertainty of the position of dune toe was calculated according to the first four parameters of equation (1).

The coastline was divided into different coastal sectors each of which corresponded to a single beach-dune system. Change rates of the two proxies were then calculated using the same baselines and transects to be able to pair the data at each point.

The Shoreline Change Envelope (SCE), the Net Shoreline Movement (NSM) and the End Point Rate (EPR) were calculated for two periods, i.e. 1977 – 2001 and 2001 – 2019, using the DSAS extension of ArcGIS 10.6 that takes into account the uncertainty of each proxy using equation (1), [79]. The SCE method deals with shoreline variability at each transect taking into account the maximum spatial recorded displacement, regardless of the time span over which it was recorded; The NSM is associated with the dates of only two shorelines and it reports the distance between the oldest and youngest shorelines, i.e. 1977 and 2019, for each transect although this movement may be not the maximum shoreline displacement; the EPR is calculated by dividing the distance of shoreline movement by the time elapsed between the oldest and the most recent shoreline. Baselines were digitized parallel to the shoreline and the spacing of transects (25 m) was determined in accordance with the regional scope of this paper.

A total of 53 beach-dune systems were observed for the 1977 – 2001 period and 38 for the 2001 – 2019 period, i.e. 15 systems disappeared during the second study period essentially because of urban development [67]. Thus, in this paper, erosion/accretion rates were calculated for the 38 beach-dune systems that were observed during the two study periods. The evolution of these systems was not affected by beach nourishment programs that were only carried out at urban beaches and these lack dunes.

3.3. Statistical Analysis

The same categories of change were selected for both shoreline and dune proxies (Table 1). Values between +0.2 and -0.2 m/year were considered to indicate “stability conditions” within the expected seasonal shoreline variability of the investigated shoreline [77].

A total of 3234 transects were measured among the 38 beach-dune systems. During the study period, foredunes eroded, accreted or/and migrated laterally, and the number of transects varied in each study period. As the aim of this work is to determine the behavior of foredunes and the shoreline in front of them and associated relationships, only transects that intersected both the shoreline and the dune toe, were considered. This resulted in 2731 transects for the 1977 – 2001 period and 2654 transects for the 2001 – 2019 period.

The evolution classes obtained for each proxy were then combined, obtaining 9 combinations of evolution states of the beach-dune systems:

• Accretion/Accretion (AA): accretion classes were observed at the same transects for both shoreline and dune proxies.
• Accretion/Erosion (AE): accretion was obtained for the dune proxy and erosion for the shoreline.
• Accretion/Stability (AS): accretion was obtained for the dune proxy and stability for the shoreline.
• Erosion/Erosion (EE): erosion classes were obtained for both shoreline and dune proxies.
• Erosion/Accretion (EA): erosion was obtained of the dune proxy and accretion for the shoreline.
• Erosion/Stability (ES): erosion was obtained for the dune proxy and stability for the shoreline.
• Stability/Stability (SS): stability class was obtained for both dune and shoreline proxies.
• Stability/Accretion (SA): stability was obtained for the dune proxy and accretion for the shoreline in the same transect.
• Stability/Erosion (SE): stability was obtained for the dune proxy and erosion for the shoreline.

4. Results

Trends of transects used to determine shoreline and foredune evolution are summarized in Figure 3. Regarding foredune behavior in the first period (1977 – 2001) 24% (644 transects) of transects showed accretion, 22% (603 transects) stability and 54% (1484 transects) erosion. In the second period, i.e. 2001 – 2019, 28% (755 transects) of transects showed accretion, 30% (786 transects) showed stability, and 42% (1113 transects) showed erosion.

Concerning shoreline evolution for the first period (1977 – 2001), 40% (1082 transects) of transects showed accretion, 26% (715 transects) stability and 34% (934 transects) erosion (Figure 3A). In the second period (i.e. 2001 – 2019), 34% (902 transects) of the studied transects showed accretion, 26% (702 transects) showed stability, and 40% (1050 transects) showed erosion (Figure 3B).

Evolution rates for foredunes and the shoreline were paired, and combinations of evolution classes of each proxy were determined to analyze their behavior in the 1977-2001 and 2001-2019 periods (Figure 4).

As shown in Figure 4, there is a weak positive correlation between data: the behavior of the two proxies was similar but the intrinsic higher variability of the shoreline compared to the dune toe (Figure 3) makes the relation statistically weak. The bar graph shows that the EE combination of classes clearly dominates (Figure 4).

A general trend of the distribution of the different types of combinations of foredunes and shoreline evolution is presented in Table 2.

The dominant combinations of evolution classes were EE and AA for both periods (Figure 4;Table 3). Specifically, there were two areas in Almeria province, i.e. Punta Entinas-El Sabinar and Cabo de Gata, where this category occurred in association with other “minority” combinations (see Discussion).

Slight differences were evident between the 1977-2001 and 2001-2019 periods. EE and AA trends were very similar for the two periods while the other combinations showed more variability. Erosion of foredunes coincident with accretion (EA) or stability (ES) of the shoreline were quite frequent classes in the first period, but in the second period stability of the foredunes was more commonly associated with shoreline accretion (SA) or stability (SS) (Figure 4).

These results suggest that dunes are better indicators of coastal erosion/accretion trend because of their lower magnitude of variability compared to the shoreline, as also observed by several authors, e.g. Pollard et al. (2020). With this in mind a classification of the evolution state of each beach-dune system was developed based on the combined evolution of the shoreline and dune proxies (Table 3).

Approximately half of the beach-dune systems recorded an “Erosion” state for both periods and, from the first to the second studied period, there was a slight reduction in extent of the “Erosion” state and a slight increase of the “Mixed” and “Accretion” states (Figure 5).

5. Discussion

5.1. Spatial and Temporal Distribution of Evolution Classes

Generally, dune and shoreline proxies recorded an eroding trend (Figure 3), a behavior that was also reported by several authors and mainly attributed to the emplacement of dams that reduced sediment input to beach-dune systems [62,80,81,82,83] and the effect of coastal urbanization, i.e. land occupation and the implementation of coastal protection structures [58,60,67,77]. Accretion was usually observed up-drift of structures and natural promontories, as well as at pocket beaches and in areas directly affected by coastal protection structures [84,85,86,87,88]. At few places, accretion was also observed at the mouth of seasonal streams. According to the definition used in this paper, Stability did not represent no change in dune toe or the shoreline, but morphological changes within the range of seasonal changes and it was usually observed between erosion and accretion areas [89].

Dune erosion was the most frequent class during both periods but showed lower values during the second period during which accretion and stability slightly increased in frequency (Figure 3), possibly influenced by changes in management policies in Spain after implementation of the Coastal Act in 1988. In the 60s and 70s, prior to the Spanish Coastal Act, rapid coastal development prioritized tourism and urbanization of the coastal area leading to development of the back-beach [60,71] and a decrease and/or destruction of foredunes, especially in Malaga province [57,58,60,67,71,83,90,91,92].

The increase of accretion and stability classes of foredunes in the 2001-2019 period might reflect the implementation of management and restoration measures [58,90,93], e.g. the establishment of protected areas such as Punta-Entinas and Laguna de Adra in Almería [94] or Cabo de Gata [95] that restrict public access to particular areas, or the development of specific management plans for dune conservation [96].

The distribution of shoreline evolution classes showed no important changes during the period investigated, although erosion was slightly more widespread in the 2001-2019 period compared to 1977-2001 (Figure 3).

5.2. Beach-Dune System Behavior

This large spatial scale study obtained a great amount of data that indicates marked spatial variability. This reflects the heterogeneity of the Mediterranean coast of Andalusia and the distribution of areas influenced by human activities. Furthermore, shoreline and foredunes have different behavior and erosional/accretional processes affect them in different ways:

• Beach erosion nis normally associated with winter storm events or groups of them and its recovery takes place over weeks to months during fair weather conditions, especially in summer [63,68].
• Dune erosion or disappearance may be linked to natural processes and/or human activities [59,67,68,97] and usually occurs very quickly. Foredune recovery requires months to years depending on sediment availability, the accommodation space, the colonization and growth of appropriate vegetation and wind conditions [55,56,59,68,98,99,100,101].

In general, the shoreline and foredunes showed the same trend. For example, accretion of both proxies was recorded after the enlargement of the port of Algeciras (Cádiz province) up-drift of the structure (Figure 6A). The opposite behavior, i.e. erosion recorded by both proxies, was observed at the Guadalhorce river mouth (Málaga, Málaga province) (Figure 6B) mainly due to the critical reduction of the river sediment supplies [62,82,91]. The areas accreted/eroded in the 2001-2019 period at the mentioned study sites is presented in Figure 7: in the first example, foredunes recorded a mean accretion rate of 3.40 m/year, creating an area of over 18600 m² in front of the existing foredunes, and the shoreline advanced at a mean rate of 4.23 m/year, forming over 21000 m² of new beach area in front of the remnant foredunes (Figure 6A); in the second example, a mean value of -1.30 m/year was recorded for the evolution of the dunes, i.e. the loss of over 19600 m² in front of the previous dune toe line and the shoreline registered a mean rate of -1.78 m/year, i.e. the loss of over 26000 m² in front of the remnant foredunes (Figure 6B).

The most frequent combination of classes were EE and AA in both periods followed by EA and ES in the 1977 – 2001 period and SA and SS in the 2001 – 2019 period (Figure 4). In places a clear spatial trend was observed alongshore with EA and ES areas located between EE and AA areas, e.g. at Punta Entinas-El Sabinar (Figure 7). This is a large protected natural area where the system is accommodating to a new equilibrium because the natural coastal processes that favor erosion of sandy cuspate forelands and sedimentation in adjacent areas [8]. In this case an EE combination was observed on the shoreline salient and AA at both sides of it where the eroded sediment eroded had accumulated, i.e. the salient constitutes a divergent fixed limit [84,86]. The areas in between showed dune erosion and beach accretion (EA) or stability (ES) reflecting a changing trend.

At several places a change was noted between the first and second study periods. In some instances, this was due to new coastal protection structures. For example, at Playa del Perdigal and Garrucha in Almería province EA and ES combinations in the 1977-2001 period were replaced by an EE combination in the 2001-2019 period. The change was caused by the emplacement of 5 groins that caused down-drift erosion (Figure 8A). In the second period, the behavior of this sector changed after enlargement of the port of Garrucha. Up-drift of the port, EA combination recorded during the first studied period, switched to the AA combination in the second period (Figure 8B).

Changes attributed to natural processes were also recorded [4,5,18,22]. For example, the Artola foredunes (Málaga province) have a complex dynamics behavior [102]. The area is characterized by an alternation of easterly and westerly winds and high-energy storm events that approach from the east generating an eastward directed littoral transport. The trend described by previous authors was reflected in the findings reported here: in the first period, EE and ES combinations were generally observed and AA combination was recorded close to the port, which had been enlarged in 1980 (Figure 9A). In the second period, SA and SS combinations prevailed (Figure 9B). The welding of nearshore bar at this sites [102], reflects the changes observed in this paper.

5.3. Considerations for Coastal Management

A combination of different shoreline proxies provides more information about the whole coastal system evolution as each proxy is able to capture certain, specific processes [28,35,36]. The use of a combination of shoreline and dunes proxies provides an opportunity for better understanding of beach-dune system behavior and the establishment of sound management measures and plans.

The combination of evolution classes calculated for the shoreline and foredunes reflects the state of each beach-dune system (Table 3) and prompts a range of potential management responses:

• Erosion states represent systems that present different level of degradation and need management measures;
• Mixed states represent systems with diverse levels of variability or stability conditions, which can shift to an erosion state in the short/medium term. These systems need to be monitored to full comprehend their present and future behavior;
• Accretion states represent systems in a good state of health, and they do not need management measures in the short/medium term.

The methodology developed in this paper represents a useful tool to determine areas that need more attention. at both local and regional scale. As an example, because the dominance of the “Erosion” category in the 2001-2019 interval indicates that great proportion of the beach-dune systems in the Mediterranean coast of Andalusia need some kind of management measures to avoid future erosion problems. It also identifies the particular areas where the problem exists.

Pranzini et al. [103] presented a review of the protection strategies carried out in Europe, concluding that there is no single solution to coastal problems but rather a range of practical possibilities. Many methods exist for beach and dunes restoration (e.g. [104,105]).

As an example, the area of the Gualdalhorce river delta (Málaga province), which presents an “Erosion” state, is affected by several groins that stop longshore transport and cause up-drift accretion and down-drift erosion. A solution could be a sediment bypass from the accreting to the erosion areas and the emplacement of fences and planting of endemic vegetation for foredune stabilization.

6. Conclusions

In this paper the evolution of 38 beach-dune systems was investigated using two different proxies: the dune toe line for foredune evolution and the high-water line for the shoreline position changes. Evolution rates were calculated for the 1977 – 2001 and 2001 – 2019 periods. Results were compared obtaining a general erosional behavior of both proxies along the studied coast and a relevant variability for the shoreline versus the foredunes.

Evolution classes of foredunes and beach changes were paired obtaining 9 combinations of classes: Erosion/Erosion (EE) was the most frequent and was recorded by 25% and 27% of the transects in the first and second periods respectively, and Accretion/Accretion (AA) that was represented 17% of the total cases observed in both periods. Changes in the evolution trend of the beach-dune systems were also found especially in areas where human interventions occurred between the first and the second periods.

The use of different proxies to monitor coastal change is frequently recommended, as each proxy provides different information about the system and can capture different processes. In this case, the shoreline proxy reflected very well medium-term changes due to its great variability and the dune proxy better reflected long-term changes, e.g. the impact of high energy storms. The dune proxy did not capture the effects of short fair-weather conditions due to the slow dune recovery capacity since they need several months/years of fair weather conditions to recover.

The classification used to reflect the state of the beach-dune systems resulted in a useful application tool for coastal management purposes since it makes easy to recognize areas that need more attention; furthermore, its periodic updating could help to verify whether the measures are working or not at a short-medium term.

Future research could be devoted to fully understand the behavior and/or evolution of the beach-dune systems investigated. The enlargement of the temporal resolution used in this paper would allow a better understanding of beach-dune system variability at different spatial scales, e.g. from months and seasons to years.