High resolution mapping and analysis of shore platform morphology in Galicia, northwestern Spain

Marine Geology 436 (2021) 106471 Available online 29 March 2021 0025-3227/© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). High resolution mapping and analysis of shore platform morphology in Galicia, northwestern Spain Alejandro G´ omez-Pazo a , b , * , Augusto P´ erez-Alberti b , Alan Trenhaile c a Departamento de Xeografía, Universidade de Santiago de Compostela, Praza da Universidade, 1. 15782 Santiago de Compostela, Spain b CRETUS Institute, Universidade de Santiago de Compostela, 15782 Santiago de Compostela, Spain c School of the Environment, University of Windsor, Windsor, Ontario, Canada, N9B 3P4 ARTICLE INFO Keywords: Scale Digital surface models Equotip UAV Joints Schistosity ABSTRACT Most research on shore platforms has been at fairly large spatial scales over distances ranging from hundreds to thousands of metres. Other work at much smaller scales, ranging from one to several decimetres, often corresponds to the dimensions of micro-erosion meter stations. Few studies have been concerned with platform morphology in which the basic data are acquired at intermediate or meso-scales ranging up to a few metres. This is due, in part, to terrestrial surveying at meso-scales being time-consuming while aerial surveys using LiDAR are expensive. A meso-scale study was made on three shore platforms in western Galicia, northwestern Spain using data from an Unmanned Aerial Vehicle (UAV) to produce high resolution Digital Surface Models (DSMs) and to calculate local surface elevation, roughness, slope, and joint density at a 0.5 m pixel scale, and joint orientation and length. This was supplemented by Equotip field measurements of rock hardness. A granitic platform was higher and had rougher surfaces and steeper slopes than two platforms in schist. The relationship between platform morphology and rock hardness and joint density was complex, however, reflecting in part the role of schistosity in accounting for the formation of low, regular platforms in hard schists with low joint density. The study suggested that while tidal range, inheritance, and other environmental and evolutionary factors can be dominant in determining platform morphology at the macroor regional scale, geological, and particularly structural, factors become increasingly important in Galicia as the scale diminishes, and they are generally dominant at the local or meso-scale. 1. Introduction Shore platforms have been studied over a variety of spatial scales (Kennedy et al., 2017). Until fairly recently, most work was concerned with regional to global analyses and was dominated by debate over the genetic significance of differences in platform morphology, and their possible relationships to morphogenic and geologic factors (Bartrum, 1916; Hills, 1971, 1972; Trenhaile, 1980, 1987). The focus of this work was on mean platform gradient and other elements of their morphology, and whether platforms terminate abruptly seawards in low tide cliffs or continue below the low tidal level without any major break in slope (Everard et al., 1964; Wright, 1967, 1970; Trenhaile, 1974, 1978, 1999; Takahashi, 1977; Sunamura, 1992; Kennedy et al., 2011; Kennedy, 2016). Work has also been conducted more recently at much smaller scales (<1 m or so), often using micro-erosion meters (MEMs) and other instruments to measure progressive or temporary, shortor longterm changes in surface elevation (Stephenson et al., 2004, 2010, 2012; Porter and Trenhaile, 2007; Trenhaile and Porter, 2018; Yuan et al., 2018). There is no generally accepted classification for the spatial scales used to study the effect of various processes on rock coasts. While an ad hoc classification is proposed for the purpose of this discussion, the dimensions of each class are arbitrary and the boundaries between them are necessarily flexible, varying according to the purpose and type of investigation. Scale categories are defined as follows: a) mega-scale for regional to global studies that can extend along tens to hundreds of kilometres (Everard et al., 1964; Wright, 1967, 1970; Trenhaile, 1974, 1978; Matsumoto et al., 2017); b) macro-scale for local, site specific studies, often associated with one or more headland-bay sequences, that extend over distances from a few hundred metres up to several kilometres (Stephenson and Kirk, 2000; Kennedy and Beban, 2005; Porter et al., 2010a); * Corresponding author at: Departamento de Xeografía, Universidade de Santiago de Compostela, Santiago de Compostela, Spain. E-mail address: [email protected] (A. G´ omez-Pazo). Contents lists available at ScienceDirect Marine Geology journal homepage: www.elsevier.com/locate/margo https://doi.org/10.1016/j.margeo.2021.106471 Received 19 August 2020; Received in revised form 24 March 2021; Accepted 25 March 2021 Marine Geology 436 (2021) 106471 2 c) meso-scale for morphological and process investigations where the data are acquired over distances or intervals of up to a few metres (Robinson, 1977; Blanco-Chao et al., 2007; Naylor et al., 2016); and d) micro-scale for areas with linear dimensions of a few decimetres that are typical of micro-erosion metre stations (Yuan et al., 2018; Trenhaile and Porter, 2018; Stephenson et al., 2019). The scale at which a study is conducted determines, in part, the variables that are measured, including those concerned with surface morphology and rates of erosion. At the mega-scale, for example, one is usually concerned with the general characteristics of a coast rather than with its detailed nature. Consequently, there has been an emphasis at this scale on variables such as mean regional platform gradient and width and their relationship to tidal range, wave energy, rock type, and, at extremely large scales, climate. Conversely, these morphogenic variables are essentially constant at the micro-scale, where the concern is usually with downwearing rates, short-term changes in surface elevation, and micro-topography. Studies at the micro-scale are therefore directed towards elucidating relationships between downwearing rates and microtopography and such variables as the duration and frequency of tidal exposure and immersion, the physical and chemical characteristics of the rock and, where relevant, its cement, the ability of the rock to absorb or adsorb (from the atmosphere) water and to retain it during low tide, small-scale variations in surface topography and drainage, the occurrence of sand or other abrasive material, and the effect of weathering, bioerosion, and bioprotection (Inkpen et al., 2004; Porter et al., 2010b). Meso-scale variations in the elevation, gradient, and degree of surface roughness represent fundamental components of rocky intertidal zones. Irregular platform surfaces play central roles in creating a complex patchwork of micro-environments and geological conditions that determine the nature and efficacy of physical, chemical, and biological processes, and how they vary in space and time. For example, effective wave quarrying, involving the dislodgment of joint blocks and other rock fragments, requires steep, irregular surfaces exposed to high wave and swash impact (Trenhaile, 1987; Naylor and Stephenson, 2010; Dornbusch and Robinson, 2011; Regard et al., 2013). Local surface gradients and irregularities influence rates of wave transformation and breaking as they propagate shorewards, swash runup heights and velocity, and the removal or accumulation of sediment, and consequently its abrasive or protective efficacy. Local topography determines the size and location of any tidal ponding and rates of surface drainage, and consequently the nature and effect of weathering and, combined with variations in surface elevation, the role of any biological activity. Joint length and density also play important roles in determining the effect of wave quarrying and weathering and in directing the focus, and accounting for the efficacy, of erosional mechanisms across platform surfaces (Trenhaile, 1972; Buchanan et al., 2020). Research at the meso-scale may help to explain the evolution of shore platforms and reconcile the existing gap in information between studies conducted at megaand macro-scales and those at micro-scales (Naylor et al., 2010). Little work has been done on meso-scale variations in surface topography in the longshore and shore-normal directions, however, in part because of the time-consuming nature of the terrestrial surveying methods that have generally been employed and the expense of airborne LiDAR surveys. Consequently, these techniques have usually been used to produce a fairly small number of supposedly representative shore-normal profiles for each study area, or to determine the elevation of MEM stations, rock hardness measurement sites, or other data in relation to the tide. The purpose of the present study was to determine the degree to which meso-scale variations in surface morphology obfuscate the relationship between mean megato macro-scale morphological parameters and tidal range and rock type; to examine potential relationships between surface morphology and such factors as rock hardness and joint density and orientation; and to account for differences in the morphology of shore platforms in the study areas. Data were derived from two primary sources: high resolution images and Digital Surface Models (DSMs) obtained with an Unmanned Aerial Vehicle (UAV); and measurements of rock hardness in the field. 2. The study areas The study was made on three shore platforms in western Galicia, northwestern Spain (Fig. 1). The site at Caama˜ no, on the Barbanza Fig. 1. Location of the study areas in Galicia, northwestern Spain. In the center inset map, the four black dots in the offshore area show the location of available simulated wave data (SIMAR) by Puertos del Estado (2019) and the square symbols show the locations of Caama˜ no, Laxe Brava and Sanxenxo. The map on the right represents the main lithological categories in Galicia. A. G´ omez-Pazo et al. Marine Geology 436 (2021) 106471 3 Peninsula, is oriented towards the west and is about 600 m in length (Fig. 2a, b) (Trenhaile et al., 1999). Laxe Brava, near the tip of the Barbanza Peninsula, faces north-northwest (Fig. 2c, d) and is about 360 m long (P´ erez-Alberti et al., 2012). Sanxenxo lies between the Ria of Muros and the Ria of Pontevedra (Fig. 2e, f). The platform is oriented towards the south-southwest and extends about 750 m along the coast. Although swell waves predominate on the western coast of Galicia, the region is close to the southern boundary of the storm-wavegenerating environment to the north, resulting in this coast experiencing the most energetic wave conditions on the Iberian Peninsula. The highest waves, which are generated by Atlantic low-pressure centres in winter, approach the coast from the NW-SE (Fig. 3a). Maximum wave heights exceed 9 m off Caama˜ no and Laxe Brava, and about 6.5 m off Sanxenxo. The tides are semi-diurnal, with a maximum tidal range in this region of about 4 m and a mean range of about 2.5 m (Puertos del Estado, 2019). With regard to their orientations: Caamano is an open coast exposed to the dominant NW-SE waves; Laxe Brava faces the direction of the NW swells but is sheltered from more westerly waves; and Sanxenxo is largely sheltered from the main westward swell. Hercynian or late Hercynian granitic rocks, Precambrian schists and basic rocks dominate in western Galicia (IGME, 1981). Northwest- –southeast and northeast–southwest faults formed in the latter part of the Hercynian orogeny (Parga Peinador, 1969), followed by north-south fractures that developed due to early Mesozoic rifting in the Atlantic. Intense tectonic movements along the unstable Atlantic margin, from the Eocene until at least the early Quaternary, then formed a series of basins and uplifted blocks that produced coastal mountains ranging up to approximately 300–600 m in elevation (P´ erez Alberti, 1991). The rocks and sediments in this region were deeply weathered under tropical conditions during the Tertiary and subjected to cryogenic weathering and fluvio-nival conditions during the Weichselian. Lateglacial sediments covered and fossilized most of the Eemian cliffs during the late middle and late Weichselian glacial (P´ erez Alberti et al., 1998). The thickness and facies of these sediments, which supplied boulders and other sediments to the modern coast, vary according to the height, aspect, and gradient of the coastal mountains, and their distance from the shore. There is widespread evidence of coastal inheritance including Eemian and possibly older beaches (Trenhaile et al., 1999; Blanco Chao et al., 2003), and shore platforms that are slowly adjusting to present sea level through abrasion, wave quarrying, and weathering (Blanco-Chao et al., 2007; P´ erez-Alberti et al., 2012). Consequently, these platforms often have steep, irregular profiles with more resistant outcrops rising above the surrounding platform surface or cliff-attached ledges that are probably, given the history of this coast, vestiges of a former higher sea level. Although weathered granitic rocks are dominant along this coast most shore platforms have developed in schists, including most of the platform at Caama˜ no and all the platform at Sanxenxo. The entire platform at Laxe Brava consists of medium-grain granites, however, and there are some outcrops of fineto medium-grain granites at Caama˜ no (Fig. 4). The rocks at Caama˜ no are moderately to severely weathered, with most Schmidt Rock Test Hammer (Type L) R-values falling between 10 and 25 (Trenhaile et al., 1999). The lower R-values generally correspond to severely weathered and eroded granitic dykes that form channels and topographic depressions on the shore platform. Schmidt Hammer values are much higher in the granites at Laxe Brava, typically falling between 40 and 50 (Blanco-Chao et al., 2007). The Schmidt Hammer has not been used on the schist platform at Sanxenxo. Joint patterns in the granite at Laxe Brava promote the development of more sigmoidal than linear forms (De Vicente et al., 2011), whereas joints and Fig. 2. The three study areas a) and b) Caama˜ no, c) and d) Laxe Brava, and e) and f) Sanxenxo. Fig. 2b shows a junction between granite (light tone) and schist (darker tone). A. G´ omez-Pazo et al. Marine Geology 436 (2021) 106471 4 erosional scarps on the schist platforms are more linear and generally larger (Feal-P´ erez and Blanco-Chao, 2013). The platform at Caama˜ no lacks extensive sedimentary deposits, although there are small areas of boulders and medium-coarse sand in depressions and against topographic obstructions. The most extensive, albeit thin, deposit is of sand, which is in the south-central part of this region on a smooth and gently sloping surface at the rear of the widest section of the platform. The sediment at Laxe Brava consists of clasts of various sizes, and there is no sand. The most extensive deposit consists of a small-boulder (<0.25 m) beach in the supratidal zone, behind the low, smooth and narrow platform in the central portion of this area. The sediment at Sanxenxo consists primarily of coarse to fine sand with a few pockets of small boulders (<0.25 m) at the back of the platform; there are no large boulders at this location (Fig. 4). 3. Material and methods Low flying UAVs have been used to obtain detailed, high resolution images in a variety of rocky coastal environments (P´ erez-Alberti and Trenhaile, 2015a, 2015b; G´ omez-Pazo et al., 2019; Horacio et al., 2019; Carvalho et al., 2019; Carvalho and Woodroffe, 2019). For the present study images were obtained during low tide using an UAV (Microdrones GmbH, Model MD4–200) flying about 30 m above the shore platforms at Caama˜ no (March 2, 2015), Sanxenxo (April 30, 2015), and Laxe Brava (August 8, 2016) (Table 1). Digital surface models (DSMs) were produced for the three study areas based on the methodology described by G´ omez-Pazo et al. (2019), with statistical parameters derived from statistical software R (R Core Team, 2019). The spatial resolution was homogenized, using 0.5 m as the pixel size. To discuss and analyze spatial variations in surface topography, each study area was subdivided into three zones (Fig. 4). The DSMs were used to extract profiles, perpendicular to the shore, at 10 m intervals along the coast in each study area. Because of the irregularity of the surfaces, platform gradients were calculated from estimated lines of best fit drawn through the intertidal portion of the profiles; these lines did not extend to the high tidal level if the junction of the platform with the cliff (the cliff-platform junction) was below it, or to the low tidal level if the lower part of the platform had been under water when the area was surveyed. The DSMs were also used to calculate several important platform properties, surface roughness, maximum gradient, elevation, and joint density for each pixel. Each variable was mapped for each platform to identify spatial distributions and pixel values were used to correlate the variables within each zone to determine whether there were any relationships between them. All correlations reported in this paper were significant at the p <0.05 or lower level. Using the methodology of Riley et al. (1999), surface roughness was related to the difference in elevation between a central pixel and its neighbours, which were arrayed around Fig. 3. Wave data and joint patterns in the three study areas. a) Wave roses with annual values for wave height and frequency. The time-series started in 1958 for Caama˜ no and Laxe Brava and in 2005 for Sanxenxo (Data from Puertos del Estado, 2019). b) Joint orientations and lengths. The number of recorded joints was 609 at Caama˜ no, 825 at Laxe Brava, and 836 at Sanxenxo. c) Orientation of prominent structural elements (scarps, channels) on the platforms. A. G´ omez-Pazo et al. Marine Geology 436 (2021) 106471 5 it in a 3 ×3 pixel matrix. The local slope of the platforms, which was closely linked to the roughness, was also calculated for each pixel within a 3 ×3 matrix using ArcGIS 10.5 to compare the maximum change in elevation from a cell to its neighbours. Surface elevation was classified, relative to the Spanish datum (mean sea level at Alicante), using three categories, below 1 m, between 1 and 3 m, and above 3 m. Joints were manually traced for each study area at a 1:300 scale, using the original DSMs which had the highest spatial resolution. Joint density, which was based on the number of lines on each platform and running through each pixel, was determined by GIS photo-interpretation. Principal component analysis (PCA) was also used to analyze the data sets for the four surface variables at each of the three study areas (R Core Team, 2019). The term primary is used in this paper for deep, long joints and secondary for less prominent fractures. Primary joints were more than 2 m in length (and generally >10 m), forming troughs that were more than 0.5 m wide at the surface and at least 0.4 m in depth. Joint orientation was determined by GIS photo-interpretation, based on the lines drawn through each pixel, with the length of a joint being the length of each line in the GIS file. To investigate whether joint orientation plays an important role in controlling platform erosion and morphological development, the orientation of the principal structures on each platform, including prominent scarps and channels, were recorded from the UAV images, supplemented by field observation. Joint thickness (aperture), the distance between the joint walls, was usually <1 mm. Field observation, supplemented by regional-scale data from published geological surveys (IGME, 1981, map scales 1:50,000) demonstrated that the schistosity layers dip westerly at high angles (>75 o ) at Caama˜ no, and to the WSW with a low dip (<10 o to 15 o ) at Sanxenxo. Equotip measurements were made to determine rock hardness and to further relate platform morphology at the meso-scale to local geological Fig. 4. The types of rock, coastal zones, and location of the topographic profiles (labeled at 5 profile intervals) and Equotip measurements (triangles) at a) Caama˜ no, b) Laxe Brava, and c) Sanxenxo. The rectangular boxes in each area show the areas illustrated in detail in Figs. 9-14. Table 1 Image acquisition data. Camera Number of calibrated images Resolution (cm) Number of GCPs RMSE X (cm) RMSE Y (cm) RMSE Z (cm) Caama˜ no 17.0 MP Olympus E-PL2 563 1.4 19 1.8254 1.8775 2.0648 Laxe Brava 20.0 MP Sony ILCE6000 272 1.3 21 1.6396 1.5382 2.3916 Sanxenxo 20.0 MP Sony ILCE6000 587 1.8 27 1.6638 1.5524 2.2781 RMSE is the root-mean-square error. A. G´ omez-Pazo et al. Marine Geology 436 (2021) 106471 6 variables. This instrument consists of an impact device consisting of a 3 mm tungsten carbide ball with an electronic gauge that measures rock hardness (L, the Leeb number) according to the velocity of the ball as it rebounds from a surface (Kompatscher, 2004; Viles et al., 2011). Twenty-five measurements (henceforth referred to as the unfiltered data) were made at 30 sites at Caama˜ no, 14 at Laxe Brava, and 17 at Sanxenxo (for a total 1281 measurements) (Fig. 4). Sites were selected, where possible, to represent a variety of elevations and distances from the seaward edge, avoiding areas covered by barnacles or water. Because impact-based instruments do not operate very well on rough, uneven surfaces, Equotip measurements were restricted to fairly even, subhorizontal surfaces of about 1 m 2 in area. The lack of suitable conditions therefore prevented systematic coverage of each zone in the study areas, although the areas that were sampled appeared to be broadly representative of the areas where measurements were not possible. There have been several attempts (albeit generally for the similar Schmidt Hammer) to determine the number of rock hardness measurements that should be made at a site in order to eliminate extreme and presumably anomalous values (Kennedy and Dickson, 2006; Niedzielski et al., 2009). Following the analysis of several techniques by P´ erez Alberti et al. (2013), it was decided to use a modified data set at each site (henceforth referred to as the filtered data) consisting of 17 measurements obtained after eliminating the 8 most extreme values (relative to the mean of the original 25 values), regardless of their being negative or positive (Gupta, 2009). 4. Results 4.1. Profile-scale characteristics There were some notable differences between the profiles in the three regions (Figs. 5, 6, and 7). Granitic coasts in Galicia, including Laxe Brava, are characterized by frequent, abrupt changes in elevation and gradient associated with troughs eroded along prominent joints, residual stumps of uneroded rock, and supratidal terraces, often covered in large, rounded boulders. More regular surfaces are common in the schists at Caama˜ no (Fig. 5, for example, P29 and P33) and especially at Sanxenxo (Fig. 7, for example, P43, P49, and 64). Cliff-platform junctions are abrupt breaks of slope at the foot of steep cliffs, usually close to the high tidal level (Wright, 1970; Trenhaile, 1972). There are few well defined junctions in the 3 study areas. Platforms in granitic and occasionally schist areas frequently extend from the intertidal into the supratidal zone without any major break in slope (for example, Fig. 5, Caama˜ no P11, and Fig. 6, Laxe Brava P7 and P21), or terminate at the foot of steep scarps at about the midtidal level (for example, Fig. 5, Caama˜ no P7, and Fig. 6, Laxe Brava P27 and P29). Half the 34 profiles at Laxe Brava and more than 1/4 (14/54) of those at Caama˜ no had shore-normal gradients that were >5 o , while the corresponding figure at Sanxenxo was only 2 out of 74. Eliminating these Fig. 5. Topographic profiles derived from the DSMs for Caama˜ no. Every second profile is plotted for this area. The shaded band across each profile represents the maximum tidal range. The location of the profiles is shown in Fig. 4. A. G´ omez-Pazo et al. Marine Geology 436 (2021) 106471 7 steeper slopes, mean platform gradients at Caama˜ no, Laxe Brava, and Sanxenxo were 2.15 o , 2.71 o , and 1.95 o , respectively. The Lilliefors and Kolmogorov-Smirnov tests indicated that these modified gradient distributions did not deviate significantly from a normal distribution at each of the 3 study areas. Therefore, a series of two-tailed t-tests was run to determine whether there were any significant differences (p <0.05) in the mean gradient of the profiles (<5 o ) in the study areas. Differences in the mean gradients were statistically significant between the schist at Caama˜ no (36 profiles) and the granites at Laxe Brava (17 profiles), and between the schist at Sanxenxo (72 profiles) and the granites at Laxe Brava. The mean gradients in the schists at Caama˜ no and Sanxenxo were not significantly different. As only 4 profiles in the northern part of the study area at Caama˜ no were entirely in granite (Fig. 4a), statistically valid comparisons could not be made between these profiles and those in other types of rock in the study areas. 4.2. Sub-profile characteristics There were some important differences in the means, medians, and standard deviations of the variables representing surface characteristics in the three study areas (Table 2). Although the values for the schist platforms at Caama˜ no and Sanxenxo were quite similar, the roughest surfaces, steepest local slopes, and the densest joint systems were in the granites at Laxe Brava (Fig. 8). Although most of the platform surfaces in the three areas were at low elevations, extending from the low tide level up to about 1 m above the midtidal level, there were significant differences (p <0.05) between them. The lowest platform was at Sanxenxo, where it was dominated by surfaces at elevations of less than 1 m, and the highest at Laxe Brava, where almost a third of the exposed rock surface was in the supratidal zone, more than 3 m above the midtidal level (Table 3). There is a general increase in elevation landwards from the lowest to the highest surfaces over the entire platform at Caama˜ no, although there are small “islands” of higher elevation dispersed within the lower part of the platform in the north, where granites are dominant (Fig. 4a). The central portion of the region is generally <1 m in elevation, with higher areas in the upper intertidal and supratidal zones largely restricted to the more northerly and especially southerly parts of the region. The most extensive high areas are in the schists in the south (Fig. 5, for example, profiles P47, 49, 51, and P53 and Fig. 9) rather than in the harder granitic outcrops in the north, although this latter area (zone 1) had the highest roughness, slope, and joint density values (Table 4). These high roughness and slope values were distributed throughout the intertidal zone in the north, although high joint densities tended to be most common on the more seaward portions of the platform in the northern and southern parts of the region (Fig. 9). There are extensive surfaces at higher elevation at Laxe Brava, particularly on the two promontories at the western and eastern ends of the area, where much of the exposed bedrock is in the upper intertidal and supratidal zones (Fig. 10). The highest areas on the promontories are largely along the rear, with only a few small islands of higher ground further seawards. The central portion (Zone 2) of this area is much lower, with elevations of <1 m, with the exception of a few small areas between 1 and 3 m in height at the back of the platform (Table 4). Apart from a small area of higher platform in the supratidal zone at Sanxenxo, the granite surface at Laxe Brava was rougher and more steeply sloping at all elevations than the schist and granite surfaces in the other two areas. The higher and lower roughness and slope values were roughly evenly distributed throughout Laxe Brava, although they were a little more prominent in the western and eastern parts of the region than in the narrower central area. There was also no strong pattern in the distribution of joint density values, although they were slightly higher in the narrower central area than to the west or east (Table 4) (Fig. 10). Most of the platform at Sanxenxo is <1 m in elevation, particularly in central zone 2, with small, isolated areas of higher ground dispersed around the landward perimeter or in the form of small, elevated Fig. 6. Topographic profiles derived from the DSMs for Laxe Brava. Every second profile is plotted for this area. The shaded band across each profile represents the maximum tidal range. The location of the profiles is shown in Fig. 4. A. G´ omez-Pazo et al. Marine Geology 436 (2021) 106471 8 outcrops of resistant rock detached from the cliff on the foreshore (Fig. 7). The most pronounced areas of higher roughness and slope values are in the narrow, western part of the platform, where they are fairly uniformly distributed in the shore-normal direction. Despite local areas with higher values, the rest of the platform is generally more regular and gently sloping than in the extreme west. Joint densities are high in only a few, isolated areas, primarily in the central and eastern parts of Sanxenxo, with the lowest mean value in western zone 1 (Table 4) (Fig. 11). Despite the occurrence of some very irregular profiles, there were statistically significant (p <0.05), moderate to high correlations between surface elevation and distance from the seaward edge of the platforms in all three study areas. The only other significant correlations with shore-normal distance, however, were moderate, with surface roughness at Laxe Brava and surface slope at Caama˜ no (Table 5). With the exception of surface roughness and slope, which are closely related, correlations between other measured variables were generally low but, because of the large number of data-producing pixels and correspondingly high degrees of freedom (statistical) (>80,000, > 60,000, and >200,000 at Caama˜ no, Laxe Brava, and Sanxenxo, respectively), the Spearman Rank and Pearson correlation coefficients were consistently significant (p <0.01) (Table 6). The most notable correlations were between surface elevation and joint density at Caama˜ no (for the entire platform and for granites only) and elevation and surface roughness and slope at Sanxenxo, although these correlations only accounted for a small proportion (5–10%) of the variance. Furthermore, the direct relationship between joint density and elevation at Caama˜ no is contrary to the assumption that joints promote effective erosion and therefore surface lowering. Surface roughness and slope, representing surface characteristics, were the most important components of the first principal component (PC1), accounting for approximately half the variance in each of the Fig. 7. Topographic profiles derived from the DSMs for Sanxenxo. Every third profile is plotted for this area. The shaded band across each profile represents the maximum tidal range. The location of the profiles is shown in Fig. 4. Table 2 Indicators of central tendency for surface roughness, slope, and joint density. Location Roughness (m) Slope ( o ) Joint density (m m −2 ) Mean Std. Dev. Median Mean Std. Dev. Median Mean Std. Dev. Median Caama˜ no 0.34 0.14 0.32 13.68 11.01 10.46 0.13 0.11 0.70 Laxe Brava 0.45 0.16 0.42 22.19 14.25 19.27 0.23 0.16 0.83 Sanxenxo 0.33 0.14 0.30 13.00 11.24 9.53 0.11 0.09 0.77 A. G´ omez-Pazo et al. Marine Geology 436 (2021) 106471 9 three study areas (Table 7). All the components in PC1 were positive, consistent with the correlations between the four variables (Table 6). The first two principal components (PC1 and PC2) accounted for 75–80% of the variance and the first three components (PC1, PC2, PC3) for approximately 98% of the variance. Surface elevation was the most important variable in PC2 for Caama˜ no, and joint density for Laxe Brava and particularly Sanxenxo (Fig. 8). Joint density in PC2 had a strong, positive value in the granites at Laxe Brava but negative values at the schist-dominated areas at Caama˜ no and Sanxenxo. The eigenvector for Sanxenxo, which consists entirely of schist, was higher than at Caama˜ no, possibly due to the greater lithological diversity at Caama˜ no, where outcrops of granite reduce the effect of joint density in PC2 and increase Fig. 8. Histograms showing the percentage frequency of pixel elevation, roughness, slope, and joint density values for the three study areas. Table 3 Platform surface roughness and slope by elevation and rock type. Location Lithology Elevation (%) Mean roughness (m) Mean Slope ( o ) <1 m 1–3 m >3 m <1 m 1–3 m >3 m <1 m 1–3 m >3 m Caama˜ no All rocks 65.0 24.8 10.2 0.33 0.37 0.38 12.24 16.14 16.95 Caama˜ no Schist 64.8 24.2 11.0 0.32 0.37 0.37 11.39 15.65 16.43 Caama˜ no Granite 64.5 30.6 4.9 0.38 0.40 0.44 16.45 18.17 22.15 Laxe Brava Granite 41.1 28.5 30.4 0.41 0.47 0.47 19.06 24.08 24.30 Sanxenxo Schist 82.4 14.9 2.7 0.30 0.40 0.50 11.50 19.50 27.10 A. G´ omez-Pazo et al. Marine Geology 436 (2021) 106471 16 f) Apart from roughness and slope, which were closely related, correlations between topographic variables, and also with joint density, were generally negligible or weak. g) There was also no relationship between the hardness of the rock and surface elevation, gradient, and roughness in the three study areas. Funding AGP and APA belong to the CRETUS Institute. Alejandro G´ omezPazo was in receipt of an FPU predoctoral contract from the Spanish Ministry of Education and Innovation (reference FPU16/03050). Declaration of Competing Interest The authors state that there is no conflict of interest with the preparation or submission of this manuscript. Fig. 13. Large scale example of the relationship between the platform surface and the joint network (dashed lines) at Laxe Brava. The images show, in clockwise order beginning at the upper left: the location (as triangles) of the Equotip hardness measurements (values shown below the figure); the surface elevation; the joint density; and the surface roughness. The location of this site is represented by the rectangle in zone 2 of Fig. 4b and in Fig. 10. A. G´ omez-Pazo et al. Marine Geology 436 (2021) 106471 17 References Bartrum, J.A., 1916. High-water rock-platforms: a phase of shore-line erosion. Trans. Proc. N.Z. Inst. 48, 132–134. Blanco Chao, R., Costa Casais, M., Martínez Cortizas, A., P´ erez Alberti, A., Trenhaile, A. S., 2003. Evolution and inheritance of a rock coast: western Galicia, northwestern Spain. Earth Surf. Process. Landf. 28, 757–775. https://doi.org/10.1002/esp.496. Blanco-Chao, R., Perez Alberti, A., Trenhaile, A.S., Costa Casais, A., Valcarcel-Diaz, M., 2007. Shore platform abrasion in a Para-periglacial environment, Galicia, northwestern Spain. Geomorphology 83, 136–151. https://doi.org/10.1016/j. geomorph.2006.06.028. Buchanan, D.H., Naylor, L.A., Hurst, M.D., Stephenson, W.J., 2020. Erosion of rocky shore platforms by block detachment from layered stratigraphy. Earth Surf. Process. Landforms esp 4797. https://doi.org/10.1002/esp.4797. Carvalho, R.C., Woodroffe, C.D., 2019. Morphological exposure of rocky platforms: filling the hazard gap using UAVs. Drones 3, 42. https://doi.org/10.3390/ drones3020042. Carvalho, R.C., Kennedy, D.M., Woodroffe, C.D., 2019. A morphology-based drowning risk index for rock platform fishing: a case study from southeastern Australia. Nat. Hazards 96, 837–856. https://doi.org/10.1007/s11069-019-03572-w. Core Team, R., 2019. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna. https://www.R-project.org/. Cruslock, E.M., Naylor, L.A., Foote, Y.L., Swantesson, J.O.H., 2010. Geomorphologic equifinality: a comparison between shore platforms in H¨ oga Kusten and Får¨ o, Sweden and the Vale of Glamorgan, South Wales, UK. Geomorphology 114, 78–88. https://doi.org/10.1016/j.geomorph.2009.02.019. Fig. 14. Large scale example of the relationship between the platform surface and the joint network (dashed lines) at Sanxenxo. The images show, in clockwise order beginning at the upper left: the location (as triangles) of the Equotip hardness measurements (values shown below the figure); the surface elevation; the joint density; and the surface roughness. The location of this site is represented by the rectangle in zone 3 of Fig. 4c and in Fig. 11. Fig. 15. a) Near vertical layering in the schists at Caama˜ no and b) near horizontal layering in the schists at Sanxenxo. A. G´ omez-Pazo et al. Marine Geology 436 (2021) 106471 18 Davies, P., Sunamura, T., Takeda, I., Tsujimoto, H., Williams, A.T., 2006. Controls of shore platform width: the role of rock resistance factors at selected sites in Japan and Wales, UK. J. Coast. Res. SI 39, 160–164. De Vicente, G., Cloetingh, S., Van Wees, J.D., Cunha, P.P., 2011. Tectonic classification of Cenozoic Iberian foreland basins. Tectonophysics 502, 38–61. https://doi.org/ 10.1016/j.tecto.2011.02.007. del Estado, Puertos, 2019. Ministerio de Formento, Gobierno de Espana. Madrid, Spain. Dornbusch, U., Robinson, D.A., 2011. Block removal and step backwearing as erosion processes on rock shore platforms: a preliminary case study of the chalk shore platforms of south-East England. Earth Surf. Process. Landf. 36, 661–671. https:// doi.org/10.1002/esp.2086. Everard, C.E., Lawrence, R.H., Witherick, M.E., Wright, L.W., 1964. Raised beaches and marine geomorphology. In: Hosking, K.F.G., Shrimpton, G.J. (Eds.), Present Views on some Aspects of the Geology of Cornwall and Devon. Royal Geological Society of Cornwall, Penzance, pp. 283–310. Feal-P´ erez, A., Blanco-Chao, R., 2013. Characterization of abrasion surfaces in rock shore environments of NW Spain. Geo-Marine Lett 33, 173–181. https://doi.org/10.1007/ s00367-012-0300-4. G´ omez-Pazo, A., P´ erez-Alberti, A., Trenhaile, A.S., 2019. Recording inter-annual changes on a boulder beach in Galicia, NW Spain using an unmanned aerial vehicle. Earth Surf. Process. Landf. 44, 1004–1014. https://doi.org/10.1002/esp.4549. Gupta, V., 2009. Non-destructive testing of some higher Himalayan rocks in the Satluj Valley. Bull. Eng. Geol. Environ. 68, 409–416. https://doi.org/10.1007/s10064-0090211-4. Hills, E.S., 1971. A study of cliffy profiles based on examples in Victoria, Australia. Z. Geomorphol. 15, 137–180. Hills, E.S., 1972. Shore platforms and wave ramps. Geol. Mag. 109, 81–88. Horacio, J., Mu˜ noz-Narciso, E., Trenhaile, A.S., P´ erez-Alberti, A., 2019. Remote sensing monitoring of a coastal-valley earthflow in northwestern Galicia, Spain. Catena 178, 276–287. https://doi.org/10.1016/j.catena.2019.03.028. IGME (Instituto Geologico y Minero de Espa˜ na), 1981. 1:50,000 maps of Puebla de Carami˜ nal (map 151, Caama˜ no and Laxe Brava) and Grove (map 181, Sanxenxo). Servicio de Publicaciones. Energia, Madrid, Ministerio de Industria y. Inkpen, R., Twigg, L., Stephens˜ non, W.J., 2004. The use of multilevel modeling in evaluating controls on erosion rates on inter-tidal shore platforms, Kaikoura Peninsula, South Island, New Zealand. Geomorphology 57, 29–39. Kennedy, D.M., 2016. The subtidal morphology of microtidal shore platforms and its implication for wave dynamics on rocky coasts. Geomorphology 268, 146–158. https://doi.org/10.1016/j.geomorph.2016.06.003. Kennedy, D.M., Beban, J.G., 2005. Shore platform morphology on a rapidly uplifting coast, Wellington, New Zealand. Earth Surf. Process. Landf. 30, 823–832. https:// doi.org/10.1002/esp.1192. Kennedy, D.M., Dickson, M.E., 2006. Lithological control on the elevation of shore platforms in a microtidal setting. Earth Surf. Process. Landf. 31, 1575–1584. https:// doi.org/10.1002/esp.1358. Kennedy, D.M., Paulik, R., Dickson, M.E., 2011. Subaerial weathering versus wave processes in shore platform development: reappraising the Old Hat Island evidence. Earth Surf. Process. Landf. 36, 686–694. https://doi.org/10.1002/esp.2092. Kennedy, D.M., Coombes, M.A., Mottershead, D.N., 2017. The temporal and spatial scales of rocky coast geomorphology: a commentary. Earth Surf. Process. Landf. 42, 1597–1600. https://doi.org/10.1002/esp.4150. Kompatscher, M., 2004. Equotip-rebound hardness testing after D. Leeb. In: Proceedings, Conference on Hardness Measurements Theory and Application in Laboratories and Industries, 11–12 November, Washington, DC. Imeko, Washington, DC, 7 pp. Available online at: http://www.imeko.org/publications/tc5-2004/IMEKO-TC5-200 4-014.pdf. Matsumoto, H., Dickson, M.E., Masselink, G., 2017. Systematic analysis of rocky shore platform morphology at large spatial scale using LiDAR-derived digital elevation models. Geomorphology 286, 45–57. https://doi.org/10.1016/j.geomorph.2017.03. 011. Naylor, L.A., Stephenson, W.J., 2010. On the role of discontinuities in mediating shore platform erosion. Geomorphology 114, 89–100. https://doi.org/10.1016/j. geomorph.2008.12.024. Naylor, L.A., Stephenson, W.J., Trenhaile, A.S., 2010. Rock coast geomorphology: recent advances and future research directions. Geomorphology 114, 3–11. https://doi. org/10.1016/j.geomorph.2009.02.004. Naylor, L.A., Stephenson, W.J., Smith, H.C., Way, O., Mendelssjohn, J., Cowley, A., 2016. Geomorphological control on boulder transport and coastal erosion before, during and after an extreme extra-tropical cyclone. Earth Surf. Process. Landf. 41, 685–700. https://doi.org/10.1002/esp.3900. Niedzielski, T., Migon, P., Placek, A., 2009. A minimum sample size required from Schmidt hammer measurements. Earth Surf. Process. Landf. 34, 1713–1725. https:// doi.org/10.1002/esp.1851. Parga Peinador, J.R., 1969. Sistemas de fracturas tardihercinicas del Macizo Hesp´ erico, Trabajos del Laboratorio Geol´ ogico de Laxe, A Coru˜ na, Monograph, 37, 18 pp. P´ erez Alberti, A., 1991. La geomorfología de la Galicia sudoriental (problemas geomorfol´ ogicos de un macizo hercínico de la fachada Atl´ antica Ib´ erica: centrosudeste de Galicia). University of Santiago, Doctoral thesis, 185 pp. P´ erez Alberti, A., Costa Casais, M., Blanco Chao, R., 1998. L’importance des processus d’origine froide sur la cˆ ote Atlantique de la Galice (Nord-Ouest de la P´ eninsule Ib´ erique). Environnements p´ eriglaciaires, Association Francaise du Periglaciaire 5, 21–32. http://geoprodig.cnrs.fr/items/show/198431. P´ erez Alberti, A., Gomes, A., Trenhaile, A., Oliveira, M., Horacio, J., 2013. Correlating river terrace remnants using an Equotip hardness tester: an example from the Mi˜ no River, northwestern Iberian Peninsula. Geomorphology 191, 59–70. https://doi.org/ 10.1016/j.geomorph.2013.03.017. P´ erez-Alberti, A., Trenhaile, A.S., 2015a. An initial evaluation of drone-based monitoring of boulder beaches in Galicia, North-Western Spain. Earth Surf. Process. Landf. 40, 105–111. https://doi.org/10.1002/esp.3654. P´ erez-Alberti, A., Trenhaile, A.S., 2015b. Clast mobility within boulder beaches over two winters in Galicia, northwestern Spain. Geomorphology 248, 411–426. https://doi. org/10.1016/j.geomorph.2015.08.001. P´ erez-Alberti, A., Trenhaile, A.S., Pires, A., L´ opez-Bedoya, J., Chamine, H., Gomes, A., 2012. The effect of boulders on shore platform development and morphology in Galicia. NW. Spain. Cont. Shelf Res. 48, 122–137. https://doi.org/10.1016/j. csr.2012.07.014. Porter, N.J., Trenhaile, A.S., 2007. Short-term rock surface expansion and contraction in the intertidal zone. Earth Surf. Process. Landf. 32, 1379–1397. https://doi.org/ 10.1002/esp.1479. Porter, N.J., Trenhaile, A.S., Prestanski, K.J., Kanyaya, J.I., 2010a. Shore platform downwearing in eastern Canada: micro-tidal Gaspe, Quebec. Geomorphology 116, 77–86. https://doi.org/10.1016/j.geomorph.2009.10.010. Porter, N.J., Trenhaile, A.S., Prestanski, K.J., Kanyaya, J.I., 2010b. Patterns of surface downwearing on shore platforms in eastern Canada. Earth Surf. Process. Landf. 35, 1793–1810. Regard, V., Dewez, T., Cnudde, C., Hourizadeh, N., 2013. Coastal chalk platform erosion modulated by step erosion and debris shielding: example from Normandy and Picardy (northern France). In: Conley, D.C., Masselink, G., Russell, P.E., O’Hare, T.J. (Eds.), Proceedings 12th International Coastal Symposium (Plymouth, England). J. Coast. Res. Spec. 65, 1692–1697. Riley, S.J., DeGloria, S.D., Elliot, R., 1999. A terrain ruggedness index that quantifies topographic heterogeneity. Int. J. Therm. Sci. 5, 23–27. Robinson, L.A., 1977. Marine erosive processes at the cliff foot. Mar. Geol. 23, 257–271. Stephenson, W.J., Kirk, R.M., 2000. Development of shore platforms on Kaikoura Peninsula, south Island, New Zealand. Part II: the role of subaerial weathering. Geomorphology 32, 43–56. https://doi.org/10.1016/S0169-555X(99)00062-8. Stephenson, W.J., Taylor, A.J., Hemmingsen, M.A., Tsujimoto, H., Kirk, R.M., 2004. Short-term microscale topographic changes of coastal bedrock on shore platforms. Earth Surf. Process. Landf. 29, 1663–1673. https://doi.org/10.1002/esp.1120. Stephenson, W.J., Kirk, R.M., Hemmingsen, S.A., Hemmingsen, M.A., 2010. Decadal scale micro erosion rates on shore platforms. Geomorphology 114, 22–29. https:// doi.org/10.1016/j.geomorph.2008.10.013. Stephenson, W.J., Kirk, R.M., Kennedy, D.M., Finlayson, B.L., Chen, Z., 2012. Long term shore platform surface lowering rates: revisiting Gill and Lang after 32 years. Mar. Geol. 299–302, 90–95. https://doi.org/10.1016/j.margeo.2012.01.005. Stephenson, W.J., Kirk, R.M., Hemmingsen, M.A., 2019. Forty three years of microerosion meter monitoring of erosion rates on shore platforms at Kaik¯ oura Peninsula, South Island, New Zealand. Geomorphology 344, 1–9. https://doi.org/10.1016/j. geomorph.2019.07.012. Sunamura, T., 1992. The Geomorphology of Rocky Coasts. Wiley, Chichester. Takahashi, T., 1977. Shore Platforms in Southwestern Japan—Geomorphological Study. Coastal Landform Study Society of Southwestern Japan, Osaka. Trenhaile, A.S., 1972. The shore platforms of the Vale of Glamorgan. Wales. Trans. Inst. Br. Geogr. 56, 127–144. Trenhaile, A.S., 1974. The morphology and classification of shore platforms in England and Wales. Geogr. Ann. 56A, 103–110. Trenhaile, A.S., 1978. The shore platforms of Gasp´ e. Qu´ ebec. Ann. Assoc. Am. Geogr. 68, 95–114. Trenhaile, A.S., 1980. Shore platforms: a neglected coastal feature. Prog. Phys. Geogr. 4, 1–23. Trenhaile, A.S., 1987. The Geomorphology of Rock Coasts. Oxford University Press, Oxford. Trenhaile, A.S., 1999. The width of shore platforms in Britain, Canada, and Japan. J. Coast 15 (2), 355–364. Trenhaile, A.S., 2018. Shore platform erosion and evolution: implications for cosmogenic nuclide analysis. Mar. Geol. 403, 80–92. https://doi.org/10.1016/j. margeo.2018.05.005. Trenhaile, A.S., Porter, N.J., 2018. Shore platform downwearing in eastern Canada; a 9–14 year micro-erosion meter record. Geomorphology 311, 90–102. https://doi. org/10.1016/j.geomorph.2018.03.024. Trenhaile, A.S., P´ erez-Alberti, A., Martínez-Cortizas, A., Costa-Casais, M., BlancoChao, R., 1999. Rock coast inheritance: An example from Galicia, northwestern Spain. Earth Surf. Process. Landf. 24, 605–621. https://doi.org/10.1002/(SICI)10969837(199907)24, 68. Viles, H., Goudie, A., Grab, S., Lalley, J., 2011. The use of the Schmidt hammer and Equotip for rock hardness assessment in geomorphology and heritage science: a comparative analysis. Earth Surf. Process. Landf. 36, 320–333. https://doi.org/ 10.1002/esp.2040. Wilson, G., 1952. The influence of rock structure on coastline and cliff development. Geol. Assoc. Proc. 63, 20–48. Wright, L.W., 1967. Some characteristics of the shore platforms of the English Channel Coast and the northern part of the North Island of New Zealand. Z. Geomorph. 11, 36–46. Wright, L.W., 1970. Variation in the level of the cliff/shore platform junction along the south coast of Great Britain. Mar. Geol. 9, 347–353. Yuan, R., Kennedy, D.M., Stephenson, W.J., 2018. Hourly to daily-scale microtopographic fluctuations of supratidal sandstone. Earth Surf. Process. Landf. 43, 3142–3151. https://doi.org/10.1002/esp.4476. A. G´ omez-Pazo et al.