Rock Platform Morphodynamics: Examples from the Southern NSW Coast Free

Eroding rocky coasts usually consist of a cliff, often with a rock (or shore) platform at its subaerial base, with the seaward edge of the platform marked by a drop-off which may be vertical to sloping and a subaqueous or marine rock base or reef beyond. These four morphological components—cliff, platform, edge, and rock base—constitute a morphodynamic system in which the emphasis within each component is on understanding the interaction of process, material, and form, all of which are continually adjusting over time in response to external (weathering, wave climate, and sea level) and internal (rock type and structure) factors (Wright and Thom, 1977). Kennedy and Dickson (2006) attributed platform morphology to the feedback of local geology, tidal range, wave climate, and weathering environment, whereas Dickson et al. (2013) see platform evolution modifying through time in response to changing boundary conditions. However, Dickson and Stephenson (2014, p.233) concluded that ‘a detailed description of the morphodynamics on rock coast development is yet to be provided’. The aim of this paper is to attempt to address this gap and examine the linkages among rock type, subaerial and marine processes, and rocky coast morphology that contribute to rocky coast morphodynamics. In doing so, it also considers the respective contribution of subaerial weathering and wave action to rock coast erosion under current sea-level conditions based largely on field investigations along the central–southern New South Wales (NSW) coast. In this context, subaerial weathering refers to normal atmospheric weathering processes combined in this environment with wetting and drying of the rock surface by wave processes and salt spray. Wave action refers to wave shoaling, breaking and swash, and associated currents.

As noted later, this landform has been the subject of extensive study going back to the 19^th century. However, although platforms are a lower-energy part of rock coasts, they are both highly visible and usually readily accessible, at least at low tide and in calm seas, making them a considerably easier option to study than the cliffs that back them and the subaqueous rocky surfaces seaward of the platform edge. Moses (2006, p. 52) noted that in the British Isles, with ‘few exceptions, most researchers of shore platforms … have focused on the intertidal platform and the cliff–platform junction and there is no consensus on the behaviour of the outer edge of the platform because it is not very accessible’. However, it is the outer edge of the platform, its drop-off and immediate subaqueous base, that is most exposed to marine attack and the highest-energy part of the system. Wave and swash inundation of the platform decreases rapidly away from the edge; the cliff is mainly impacted during high swell and storm wave inundation. However, the platform and cliff are continuously exposed to subaerial weathering and saltwater spray, which lead to a range of secondary processes. In essence, a high-energy platform edge and base is backed by a lower-energy platform and subaerial cliff. It is the rate of retreat of these two subsystems that controls the platform presence, elevation, width, and overall rate of rock coast retreat. The present morphodynamic system must also be viewed as the end product of a time-transgressive system involving the longer-term planation history of a rocky coast subject to subaerial and marine processes during periods of sea-level oscillations and tectonic effects throughout the Cenozoic era (Thom et al., 2010).

Most studies of rock platforms have largely been concerned with contemporary processes across the platform or cliff, and at most the Holocene evolution of the platform. Extensive debate has occurred about their formative processes, basically divided into subaerial weathering versus wave erosion or wave cut. The subaerial weathering approach relies on water layer (level) weathering involving normal weathering processes coupled with wetting and drying by salt spray and swash which leads to rock slackening, salt crystallisation, and solution. This results in accelerated geochemical decomposition and disintegration of the rock fabric above the water level or saturated rock. Gravity and waves remove the products of weathering, gradually eroding the cliff to the level where the rock is regularly saturated, leaving the saturated rock as the platform surface. In contrast, the wave erosion approach invokes the dominate role of physical wave attack removing rock from the cliff and sweeping the platform surface of debris. Many authors acknowledge the roles of both processes.

Investigation of rock platforms in SE Australia dates to Dana (1875, 1880). He described platforms in the Sydney region, attributing them to contemporary processes. Although Andrews (1916) argued that these platforms were formed at higher sea level, subsequent workers have stressed contemporary processes, involving divergence between the wave cut and the level of saturation or subaerial weathering schools of platform formation (see Matsumoto, Dickson, and Kench, 2018, Table 1, for a selected summary of these contrasting views). During the 20^th century, studies were undertaken along the southern NSW and Victorian coasts by Jutson (1939, 1954), Edwards (1941, 1951), Fairbridge (1949), Hills (1949, 1971), Bird and Dent (1966), Gill (1967), Saunders (1968), Davies (1980), and Gill and Lang (1983). Elsewhere studies were undertaken by Bartrum (1916), Johnson (1933), Wentworth (1938a,b, 1944), Guilcher (1958), Cotton (1963), and McLean and Davidson (1968). In more recent decades, Kennedy, Paulik, and Dickson (2010); Kennedy (2014); and Stephenson and Thornton (2013) have also conducted research in this region, whereas elsewhere, Trenhaile (1980, 1987); Semeniuk and Johnson (1985); Sunamura (1992); Stephenson and Kirk (2000a,b); Dickson et al. (2013); Stephenson, Dickson, and Denys (2017); and Gómez-Pazo, Pérez-Alberti, and Trenhaile (2021) have all investigated rock platforms. A global assessment of coastal rock cliff distribution is provided by Young and Carilli (2018); an overview of global rock coast research is provided by Stephenson (2000) and Kennedy, Stephenson, and Naylor (2014); and one of Australian research is provided by Stephenson and Thornton (2013).

The term water level weathering was first proposed by Wentworth (1938b), with the term water layer weathering suggested by Hills (1949). However, the process was first presented by Bartrum (1916), Bartrum and Turner (1928), and Bartrum (1935). It was also supported subsequently by Cotton (1951), Russell (1971), Bird and Dent (1966), McLean and Davidson (1968), Davies (1980), Short (1982a), Stephenson and Kirk (2000b), Dickson and Stephenson (2014), and Brodie and Cohen (2021). Basically, the weathering operates where the rock is both exposed to subaerial weathering and alternately wetted and dried during tidal cycles and by wave spray, onshore winds, and storm wave inundation, allowing a constant coating of salt on the rock surface. This provides an environment for continuous chemical reactions between the salt and the rock minerals. This process extends across the whole swash–spray zone, including the cliff face. For the chemical reaction to occur, the rock needs to be above the level of saturated rock. In this zone, oxygen combined with continual wetting, drying, and salt spray provides an environment for accelerated geochemical decomposition and disintegration of the rock fabric. This action is clearly visible, with flaking, swelling, and lose rock grains and fragments observable on the cliff face and accumulating along the cliff base, where they are gradually removed by wave action. Bird and Dent (1966) fully supported water-layer weathering, also stating that although wave action removes the products of the decomposition, strong wave action destroys (through wave quarrying), rather than forming, high-tide platforms.

This paper accepts that these processes operate based on physical evidence; however, no studies have examined in detail the geochemical processes assumed to be taking place. In addition, retreat of the subaerial cliff is assisted by gravity and mass movement. This is expressed by episodic rock falls and slides from the cliff, the accumulation of and then slow removal of this debris by waves both offshore and longshore, with some debris deposited downdrift as boulder beaches.

Below the saturated rock surface, the rock is only exposed to physical wave attack, abrasion, and biological process, the latter both protecting and eroding the rock. The degree of saturation depends on the elevation of the rock and its permeability, which is a function of lithological type, grain size, and structure (bedding and jointing), together with tide range, breaker wave height, and climate (Bartrum, 1938). Surface rock saturation generally increases in height with greater wave exposure and higher tides. It is assumed that subsurface saturation occurs in less permeable rocks and finely joined and bedded rocks and is absent from impermeable rocks, such as granite, where platforms are also absent. The seaward edge of the platform is attacked and eroded by wave action, usually through the process of wave quarrying together with minor abrasion, and solution in limestone platforms. Biological processes also play a role both erosive, as with burrowing sea urchins, and protective, by blanketing the rock to maintain moisture and saturation. However, it is assumed their role in overall rock coast erosion is minor and is not considered further.

McLean and Davidson (1968) concluded that cliff retreat on the Gisborne, New Zealand (NZ) coast (39° S) is driven by mass movement, which is accelerated by wetting and drying in a salt-rich environment which enhances weathering of the cliff base, combined with removal of debris by tides and infragravity waves. This conclusion is also supported by Dickson et al. (2013). Stephenson and Kirk (2000b) estimated that the Kaikoura, NZ, platform (42° S) is wetted and dried between 100 and 380 times a year, with the maximum occurring 0.6 to 0.9 m above mean sea level (MSL) on the more landward margins of the platform where erosion is also highest. Stephenson and Kirk (2000b) also observed surface swelling (expansion and cracking of the rock surface) of the platform bedrock because of this process. Stephenson (2001) found that at this site, although the cliff was retreating between 0.5 and 0.91 m/y, it was not possible to detect horizontal erosion of the seaward platform edge, concluding it was probably not eroding. Davies (1980) found that erosion to the level of saturation was favoured by permeable, closely bedded rocks with low dips, high evaporation, and mixed or diurnal tides to facilitate drying, high temperatures to speed chemical rates and low tide range. He added that these processes are more dominant in warmer and lower latitudes, whereas cooler and wetter climates have lower evaporation rates and less drying. The removal of material above the level of saturated rock results in the formation of the platform.

Where wave erosion has been favoured as the dominant formative process of platforms, the term wave cut becomes the operative word associated with platform development. Wave cutting as the dominant process has been preferred by Edwards (1951), Hills (1949), Juston (1939), Bradley (1958), So (1965), Sunamura (1978, 1983, 1991, 1992), and Trenhaile (1972, 2000). In this process, waves break on the seaward platform edge and physically attack the backing cliff, causing it to retreat faster than the platform edge. Gómez-Pazo, Pérez-Alberti, and Trenhaile (2021) attribute macroregional platform morphology to tide range, inheritance from previous sea-level highstands, and other environmental factors, with geology and rock structure important at the local scale. They conclude that platforms are slowly adjusting to present sea level through abrasion, wave quarrying, and weathering, without specifying their relative roles.

Questions that both weathering and wave processes need to address relate to the elevation and morphology of the platforms and their seaward edge, where they drop off to deeper water and rock base. In SE Australia, horizontal platforms are generally formed above MSL, usually in the supratidal region. They tend to increase in height with increasing exposure both longshore and cross-shore, the latter resulting in a rampart at the seaward edge. On Oahu, Wentworth (1938a) observed ramparts up to 3 m above the platform surface and 3.6 m above MSL. Bezore, Kennedy, and Ierodiaconou (2023) present several platform cross-sections along the high-energy, microtidal Victorian coast, most (15 of 21) of which have a seaward rampart. In other words, platforms are often highest at their seaward edge, where wave attack is greatest, whereas they tend to the lowest and eroding most rapidly at the rear of the platform and base of the cliff, where wave energy is lowest.

Several studies of wave transformation across platforms document the rapid decrease in both inundation and energy across the platform, the rear being the least inundated and having the lowest wave energy. Stephenson and Kirk (2000a) found at Kaikoura that there was an 80% reduction in wave energy across the platform. Dickson and Pentney (2012) made microseismic measurements across a platform and cliff and found ‘most wave energy was delivered to the outside edge of the shore platform, not the cliff toe’. Likewise, Poate et al. (2016, p. 27) measured wave dissipation and transformation across four platforms, finding substantial reduction in wave height and energy across the platforms, with the greatest dissipation on rough platforms. Power et al. (2018) conducted a 6-month study of platform inundation, just south of Sydney, using cameras and pressure sensors. The well-exposed platform lies above the highest tides and drops off to a 5-m depth and then steeply to 8 m. They found that the seaward platform edge was always inundated, with waves greater than 0.8 m initiating platform inundation. Inundation decreased to 40% immediately landward of the edge and continued to decrease towards the cliff.

Dickson and Stephenson (2014) argue for the duality of marine and subaerial processes, with subaerial weathering often necessary to reduce rock strength and wave erosion necessary to remove debris or jointed rocks, implying the weathering is the dominant formative processes and waves are a secondary transport mechanism. Likewise, Trenhaile (2002) states that although weathering reduced the rock resistance in the middle latitudes, wave erosion is required to remove the rock weakened by weathering. That waves remove and transport cliff and platform debris is not questioned and evidenced by masses of rock debris often downdrift of the platform, forming boulder beaches, or lying seaward of the platform drop-off. What is questioned is the ability of waves to physically erode rocks along a near-horizontal plane without the assistance of prior subaerial weathering within the intertidal to supratidal zone independent of rock structure.

This paper forms part of a broader study on the evolution of the SE Australia rocky coast and shelf extending from the subaerial cliffs out across the continental shelf (Figure 1), including the marine abrasion surface described by Thom et al. (2010). Companion papers under preparation examine the following: (1) the morphology of the subaqueous rock base of the cliff or platform (Thom et al., in prep a) and (2) the polygenetic and polycyclic processes that have operated since perhaps the Cretaceous period along this passive continental margin across a rocky inner continental shelf (Thom et al., in prep b).

Five field sites were investigated on the southern NSW coast between Werri Point (34.7° S) and Bermagui (36.4° S). This is a 360-km-long rocky coast with 234 embayed beaches occupying 55% of the coast, all usually bordered by prominent headlands and most with rock platforms. The coast faces E-SE into the Tasman Sea and receives a moderate to high wave environment, with H_s averaging 1.6 m and T_z equal to 10 s. Extreme waves generated by tropical, east coast, and midlatitude cyclones can reach several metres (Shand et al., 2010; Short and Trenaman, 1992). Waves predominately arrive from the S and SE, with a lesser percentage from the E and NE. Tides are microtides, with a 1.3- to 1.9-m average spring range. The coast has a rich history of studies into its Quaternary depositional history and morphodynamics (e.g., Oliver et al., 2020; Short, 2020; Thom and Roy, 1985; Thom et al., 1981; for rock coast morphology, see Bird and Dent, 1966).

The coast south of Sydney is in the southern section of the Sydney Basin geological province and the Lachlan Fold Belt (Roy and Thom, 1981). Field investigations were undertaken on platforms at five sites, three in the Sydney Basin and two in the Lachlan Fold Belt (Table 1 and Figure 2). Rock types vary in lithology and include sedimentary sandstone and siltstone, metasedimentary rocks, and intrusive and extrusive igneous rocks. There is also considerable diversity in rock structure and strength, as listed in Table 1.

Platforms were surveyed using a theodolite and staff, usually along the base of the cliff and across the platform to its seaward edge. All elevations are relative to the Australian Height Datum (AHD), which is approximately MSL. One site (Mosquito Bay) was surveyed immediately offshore of the platform using a Raytheon fathometer and scuba gear. Offshore bathymetry obtained from other sites were derived from marine LIDAR (DCCEEW, 2018). All LIDAR visualisation and geoprocessing (platform cross-section generation) were completed in ESRI ArcMap using its 3D Analyst extension.

The following summarises key observations made at each of the five sites. Three sites occur within the northern part of the Sydney Basin, which is dominated by near-horizontal sedimentary sequences of Permian age along with local lava beds. To the south, the Lachlan Fold Belt has some of the oldest rocks in eastern Australia, the Wagonga beds of Lower Palaeozoic age (Packham, 1969). The rocks consist of steeply dipping metasediments, as well as some intrusive volcanics. They are represented in this study by two sites. Bird and Dent (1966) have also described 10 sites in the southern coast section in both the Sydney Basin and the Lachlan Fold Belt.

Werri Point has a well-developed supratidal rock platform cut in heavily jointed, horizontally bedded Permian sandstone overlain by a weathered Permian basalt (latite). The cliff face, both sandstone and basalt, is friable and actively disintegrating, supplying generally small debris and boulders to the platform. The near-horizontal platform rises abruptly along its seaward edge to a rampart (Figure 3). The rear of the platform has an average elevation of 1.2 m (AHD). It increases in elevation longshore along the base of the cliff (0.83–1.9 m) and cross-shore towards a rampart that reaches heights between 2.25 and 4.5 m and averages 2.9 m, 1.7 m higher than the rear platform (Figure 4a). Seaward of the platform, the marine LIDAR shows that the depth to the rock base increases from ∼5 m along the more sheltered northern side of the platform to ∼7 m off the more exposed eastern side (Figure 4b,c).

Penguin (Culburra) Head is surrounded by a well-developed rock platform formed in horizontally bedded sandstone, with some ramparts along most of its seaward edge and boulder debris at the rear. It is backed by a 5-m-high cliff, which is stabilised by vegetation along its northern side (Figure 5). The higher-energy southern side also has considerable cliff debris in the form of large sandstone boulders (1–2 m in diameter; Figure 6c). The boulders are scattered the length of the platform, generally along the cliff base, with some deposited in a rough boulder beach at its junction with the sand beach (Figure 6d). Some boulders are encrusted with oyster bases, indicating an intertidal (platform) source (Figure 6c). The platform surface is relatively horizontal, with considerable surface pitting (Figure 6a,b) and some circular potholes with raised rims. Along its southern and eastern sides, it increases in elevation seaward along the base of the cliff from 2.1 to 4.61 m, averaging 2.4 m. It also increases across the platform from 2.44 to 4 m, averaging 2.85 m, but reaches 6.6 m at the most exposed eastern tip and between 3 to 5 m along the northern side (Figure 7a). Marine LIDAR shows the platform drop-off steepening and increasing in depth towards the most exposed eastern tip, where it reaches 10 m (Figure 7b,c).

Dolphin Point is formed of massive horizontally bedded, jointed sandstone. It has a broad (∼100 m wide) horizontal platform flanked by a prominent rampart (Figure 8a). Its surface possesses considerable surface pitting where saturated (Figure 8b), with some honeycomb weathering of slightly raised areas. Just two points were surveyed towards its exposed eastern tip. Here, the platform rises from 2 m against the cliff to a rampart reaching 5 m (Figure 9a). Profiles using marine LIDAR show the depth of the edge of the platform increases from 1.5 m on the sheltered northern side to more than 5 m on the more exposed eastern face (Figure 9b,c).

Mosquito Bay is located 240 km south of Sydney in the Lachlan Fold Belt (Figure 2). Its southern headland is cut in vertically dipping metasedimentary rock. The platform extends 200 m eastward and then turns and trends 200 m to the south. Short (1986, in Thom et al., 1986) mapped the northern 200 m and easternmost point of this partially sheltered headland and supratidal platform. It ranges from 25 to 35 m in width and has a prominent rampart (Figure 10a,b). The base of the cliff at the rear of the platform increases in elevation in a seaward direction from 2.1 to 3.7 m (averaging 2.4 m, AHD) and cross-platform from a low of 1.8 m against the cliff to a maximum rampart crest of 4.86 m (averaging 3.04 m; Figure 11a). Immediately seaward of the platform, the depth to the rock base increased from 2 m inside the bay to 8 m at the most exposed seaward tip. Marine LIDAR shows a similar trend, with depth off the platform increasing from 2 to 3 m on the more sheltered northern side to ∼5 m off the NE tip and 10 m off the most exposed eastern side (Figure 11b,c). The LIDAR profiles also show the platform increasing in height, both longshore towards the most exposed areas and seaward, where it contains a rampart.

The Bermagui headland is partly sheltered by Dickinson Point 700 m to the east and is composed of vertically dipping metasedimentary rocks. It has a well-developed rock platform and rampart along its more exposed eastern and northern side (Figure 12a,b). The base of the cliff increases in height seaward from 2.08 to 3.5 m (averaging 2.44 m, AHD) and cross-shore towards the rampart, with elevations rising from 2.32 m to reach 3.48 m in the east and 5.23 m at the northern tip (averaging 3.5 m, AHD; Figure 13a). The marine LIDAR shows the dramatic increase in depth off the platform as wave exposure increases. Along the partly sheltered inner Bermagui headland, the depth increases from 3 m at B4 to 5 m at the more exposed B1. On the outer, well-exposed Dickinson Point, the depth to the base reaches more than 10 m (Di1 and Di2; Figure 13b,c), the deepest observed on the south coast (Table 2).

Table 2 summarises the range of platform and rampart elevations and drop-off or rock base depths obtained from the ground surveys and marine LIDAR. In all cases, the platforms are inter- to supratidal, reaching an average height of 2.6 m (AHD) at Mosquito Bay but generally lying between 1 and 2 m (AHD). Ramparts increase in height with exposure and range from 2 to 6 m (AHD), that is, 1 to 4 m higher than the backing platforms. The drop-off or rock base depth is also positively related to wave exposure, ranging from ∼1.5 m in sheltered locations to 10 to 15 m in exposed locations.

Two unpublished studies on the northern Sydney coast (Figure 2) documented cliff, platform, and seaward edge morphologies in sedimentary sequences of Triassic age (Crozier, 1988; Reffell, 1978). Here, near-horizontal bedded rocks provided ideal locations to study both rock coast processes and responses to driving forces under present sea-level conditions. The studies conducted detailed field investigations of 10 headlands, including Barrenjoey, Little Head, Bangalley, South Avalon, Bilgola, Turimetta, Dee Why, Fairy Bower, and North Head (Figure 14). Each headland is formed in interbedded sandstone, siltstone, and shale, with some dominated by claystones and shales and others by massive sandstone. Their field work consisted of observations both of cliff and platform morphology and of the subaqueous rock surfaces extending up to 500 m seaward of the headlands. Of interest to this paper are their observations of platform morphology and its seaward drop-off. Reffell (1978) proposed a model of platform formation, together with a variation based on the predominate SE wave climate in the Sydney region (Figure 15). Crozier (1988) also developed a morphological model in which the platform increases in width and decreases in elevation in a shore-normal (lower energy) direction, with platform elevation increasing with wave exposure. He found that the highest platforms occur in less permeable rocks (sandstone higher than shale), whereas wave quarrying dominates the platform edge and subaqueous area. Both Reffell and Crozier observed that as both the platform elevation and the depth to the rock base increased, the seaward edge drop-off increased in steepness and the subaqueous sand–rock boundary increased in depth. Reffell also found that platforms in softer claystone are lower than those in harder sandstones and that they have a more inclined subaqueous cliff and smaller debris and boulders.

Crozier (1988) also measured the variation in depth to the platform drop-off or rock base at each headland. The base is usually marked by a pronounced change in gradient from steep to sloping. Crozier found consistent increases in depth with exposure, as shown in Table 3. On average, the depth increases from 5.5 m in more sheltered locations to 14.7 m in the most exposed location, which usually face SE into the dominant swell. The maximum depth of 20 m occurs off the exposed North Head.

This paper focuses on the results of field studies conducted along the NSW south coast, complemented by two previously unpublished studies from the Sydney area. Several other studies, reviewed later, have also focused on these issues.

In the study area, Bird and Dent (1966) examined 10 sites on the NSW south coast and found broad horizontal platforms were developed at or slightly above mean high tide. They noted that they are best developed in fine-grained rocks, including shales, siltstones, mudstones, sandstones, schists, phyllites, and homogenous basalt, and often truncate the local dip. As mentioned earlier, they attribute the platform development to accelerated erosion above the saturated platform, with the debris washed away during spring tides and occasional storms. No platforms were found in the impermeable intrusive igneous rocks.

Elsewhere, Kennedy, Paulik, and Dicon (2011) surveyed several of Bartrum’s (1916) sites specifically to test the subaerial versus weathering erosion theories. However, their findings are inconclusive. Matsumoto, Dickson, and Kench (2018) modelled the relative dominance of wave versus weathering erosion of intertidal platforms, finding wave domination increased where the seaward edge was retreating, whereas weathering favoured stable seaward edges.

Acharya-Chowdhury et al. (2024) used a range of techniques to measure cliff erosion in thick beds of heavily jointed sandstone and thinly bedded siltstone. The cliff is in a relatively sheltered area fronted by a 140-m-wide platform, with waves only reaching the cliff during storms and elevated water levels. They found that cliff erosion rates ranged between 3.5 and 180 mm/y over annual to Holocene timescales, with an average of 41 mm/y. If a platform has widened during the Holocene stillstand, as assumed by Sunamura (1992), then it is expected that cliff retreat would slow as the platform widens. This is a feedback effect, with the present rate of retreat becoming slower over time given no change in relative sea level. However, Acharya-Chowdhury et al. (2024) found the opposite, with cliff retreat accelerating as the platform widened. This finding lends support to the resilient platform backed by a disintegrating cliff, with the cliff retreating at a considerably faster rate than the seaward edge of the platform. In contrast, Dickson and Pentney (2012) suggest that because they found most wave energy is delivered to the seaward edge of the platform, the edge is possibly being destroyed faster than the cliff retreats.

One of the few studies to examine both platforms and depth seaward of the platforms (the drop-off or rock base) was undertaken by Dickson and Woodroffe (2005) and Dickson (2006) around the lithologically diverse Lord Howe Island. They found that platforms in both basalt and calcarenite were wider in exposed locations. Dickson (2006) also found that more porous calcarenite developed wider and lower platforms close to MSL, whereas the more resilient breccias, tuffs, and basalts developed narrow, higher platforms (4–6 m, AHD), with some seaward ramparts occurring on all rock types. These researchers undertook bathymetric surveys of ∼100 sites around the island and found a usually vertical low tide cliff (the drop-off) off the platforms, with depth positively related to wave exposure and ranging from less than 5 m in sheltered locations to more than 15 m in more exposed locations. This is in close agreement with the Crozier (1988) findings (Table 3).

The five field sites investigated for this study all have well-developed, near-horizontal platforms cutting across two major rock types: resistant sedimentary and dipping metasedimentary. Werri Point has the lowest platform, with an average elevation of 1.2 m (AHD) cut into sandstone. The highest average, at Mosquito Bay (2.6 m), is formed in vertically dipping shales and schists, followed by Bermagui (2.4 m) with a similar rock type and Culburra (2 m) in massive sandstone. These findings suggest that the more resilient and massive rock types favour higher platforms. Likewise, the seaward edge or ramparts are lowest at Werri Point (averaging 2.76 m), followed by Mosquito Bay (3.04 m), Penguin Head (3.16 m), and Bermagui (3.46 m), whereas the highest ramparts reach 4.86 m at Mosquito Bay and 6.63 m at Penguin Head, the latter a difference of 4.68 m between the lowest and the highest measured parts of the platform. All increases in height occur in the direction of greater wave exposure and energy. Seaward of the platforms, there is also a consistent trend, with the depth to the rock base increasing with exposure. There is a general increase from less than 2 m in the more sheltered locations to several metres on most platforms’ exposed eastern faces to a maximum of 15 m off the exposed Dickinson Point (Table 2). These observations are broadly consistent with the findings of Crozier (1988, as shown here in Table 3), Dickson and Woodroffe (2005), and Dickson (2006).

Field data from this study, as well as the work of Reffell (1978) and Crozier (1988), found four consistent morphological characteristics common to all sites. First, all platforms increase in elevation towards their more exposed section. They also increase in elevation normal to the cliff face, often terminating in a prominent rampart up to a few metres higher than the platform. Third, platform elevation is influenced by rock type, and softer more finely bedded or jointed shales have lower platforms than harder and/or more massively bedded rock. Finally, the depth to the base of wave erosion increases with exposure.

The implications of these NSW studies are as follows. First, the seaward increase in platform elevation both normal and perpendicular to the cliff favours the Wentworth level of saturation (subaerial weathering) mode of formation, discounting the effect of wave cutting. This is especially the case in the near-horizontal bedded sedimentary rocks that are so common along the central–southern NSW coast. Assuming the platform surface represents the upper level of regular rock saturation, then as the degree of saturation increases with exposure both seaward and across the platform, so too does the level of planation and level of the platform.

Second, waves are most energetic at the platform edge. Here the platform is highest, with wave energy and occurrence of inundation decreasing rapidly shoreward towards the cliff base, where the studied platforms were lowest. If platform erosion by waves is significant today at present sea level, the reverse morphology would be expected, that is, platform elevation would decrease seaward. Although waves are critical in the removal of cliff and weathered debris, they make a lesser contribution to direct erosion of the platform surface and sheltered cliff face.

Third, platform topography is exposed to several secondary physical and chemical processes. The process of abrasion within potholes by toolstones is a relevant erosive process. Where there are folded metasediments, the ongoing effects of differential erosion with toolstones in operation becomes more apparent, generating a serrated surface. But even here, the relative relief exhibited in the serrated microterrain of the platform surface (Figures 10 and 12) is smoother than what has been observed below low water (Thom et al., in prep a). In addition, the platform surface may contain negative and positive relief features and chemical concretions in both joints and rimming potholes and even honeycomb weathering of slightly raised sections (e.g., Bartrum, 1936; Wentworth, 1944). All such microtopography results in minor etching and lowering of the platform surface but is secondary to the initial lowering of the overburden. The presence of this range of secondary microfeatures also indicates the platform surface has remained relatively stable and been exposed for some time, sufficient to permit the formation of these features.

Part of the divergence of the subaerial weathering versus wave erosion approaches is that the former tends to be favoured by researchers based in warmer subtropical to temperate locations (21–46° latitude). In the mid- to lower latitudes, particularly with microtides (e.g., Hawaii, southern Australia, and NZ), supratidal horizontal platform surfaces dominate. In the cooler climates and with the higher tide ranges of Europe and Japan (31–58° latitude), sloping platforms are more common. One would expect subaerial weathering to be more prominent in warmer climates and decrease in its contribution into cooler and in particular cold climates, a point mentioned by Trenhaile (2002).

Figure 16 summarises key components of platform morphology and dynamics of a typical central–southern NSW coast cliff and platform. Figure 16a shows the inter- to supratidal platform, increasing in height towards its seaward edge (rampart), with a vertical drop-off and some its other key components. Geochemical processes related to subaerial weathering and marine processes dominate above the saturated platform surface, whereas physical processes primarily related to wave action dominate below. In addition, the platform has potholes, boulder debris, and encrusting organisms. Figure 16b is a conceptual model of platform evolution in which the elevation depends on the degree of saturation and the width depends on the rate of cliff (dC) versus platform edge (dE) retreat. The inset indicates platforms are highest and narrowest where waves are highest and decrease in height into more sheltered locations, with width optimal between the two wave extremes. If the edge retreats at the same rate or faster than the cliff (dE ≥ dC), no platform will form, with dC > dE required for platform formation. The platform may widen to an equilibrium position point where dC = dE. Measurements of the relative rates of dC are few and essentially nonexistent for dE. What evidence there is (e.g., Acharya-Chowdhury et al., 2024) indicates that dC is generally substantially more rapid than dE, hence the formation of platforms. Many of this study’s platforms maintained well-developed elevated ramparts, backed by platforms tens of metres in width, which suggests that dC > dE, and in some locations dE may be minimal, because sea level reached around its present position or higher 6000 to 7000 years ago (the Holocene stillstand on this coast). This interpretation is supported by the range of secondary platform surface features mentioned earlier. Assuming an average platform width of ∼50 m (Table 2) and minimal width if there is any edge retreat, this would represent a Holocene cliff retreat on the order of 8 mm/y in the study area’s more resilient shale, sandstone, and metasedimentary rocks. Elsewhere in southern Australia. cliff retreat has been estimated in softer calcarenite to be up to 30 mm/y (Fotheringham, 2009) and in limestone to be more than 30 mm/y (James et al., 2006; Short, 2020, p. 978).

The present paper assumes that wave climate, sea level, and tidal range have not significantly changed on the open coast over the last 6000 to 7000 years to affect platform development. Difficulties in attributing details of present platform morphology, for instance, to sea-level change have long been discussed on this coast (e.g., Johnson, 1938; Jutson, 1939; Hails, 1965; Langford-Smith and Thom, 1969; Hopley and Thom, 1983). This recalls the statement attributed to E.S. Hills of Victoria: ‘Unless we understand the ways in which shore platforms are formed it is obviously fruitless to draw conclusions from them about relative movements of land and sea’ (cited by Hails, 1965, p. 74). This approach is to accept the possibility that such movements may have occurred within a narrow stillstand embracing the mid- to late Holocene (Hopley, 1983; Thom and Chappell, 1975; Thom and Roy, 1985). However, more recent work by Sloss, Murray-Wallace, and Jones (2007) and others has identified a higher Holocene level; for purposes of this study, consider that there is sufficient uncertainty about its occurrence and effects to discount its impact in line with the statement by Hills.

A further complication related to sea-level change involves uncertainties associated with impacts of previous interglacial highstands. Convincing evidence exists to conclude that along this coast, the Last Interglacial Maximum (MIS 5e) sea level stood 2 to 4 m above present (Murray-Wallace and Belperio, 1991). This period, along with possible earlier Pleistocene highstands, would have contributed to marine erosion of the studied sites, leaving cliff, platforms, drop-offs, and rock bases that were subsequently all exposed to subaerial denudation and weathering during the low stands. Elevated Pleistocene platforms containing distinctive potholes have been described along the southern Australian coast (e.g., Buckley et al., 1987; Maze, 1933; Short and Fotheringham, 1986). Brooke et al. (1994) found that along parts of the studied coast, present-day cliffs and platforms may represent reactivated Pleistocene surfaces. Likewise, in Spain, Gómez-Pazo, Pérez-Alberti, and Trenhaile (2021) found that inheritance, from previous sea-level highstands, played an important role in platform formation. Their contribution to present platform morphology must be at least considered even if much of the evidence has been eroded. However, the studied platforms in this paper show no clear remnants of such erosion apart from the northern side of the Turimetta headland (Figure 14e), described by Short (1982b). Rather, the present cliff, platform, drop-off, and rock base morphology, especially with the prominent ramparts, appears to be adjusted to an active suite of processes operating in the Holocene around present sea level. The extent to which these features have modified any preexisting Pleistocene inherited platform surface is unclear.

An idealised model of Holocene evolution of platforms in this area is depicted in Figure 16. Local differences in geological rock resistance are not shown. As sea level rises on a permeable rocky shore, so too does the degree of rock saturation, with the highest level being at the wave-washed, saturated seaward edge of the cliff or platform. Relative sea-level stability enables erosion to focus on the weathered rock above this saturation surface. If the weathered rock retreats faster than physical wave erosion of the lower saturated rock, then a platform begins to form (or reform) towards an equilibrium level. Assuming an initial exposed cliff face and little if any platform, then the greater wave attacks of the cliff may lead to faster cliff retreat following stillstand, the rate decreasing as the cliff retreats and platform widens. As the cliff retreats landward and wave inundation decreases shoreward, the degree of saturation also decreases in frequency and elevation. This feedback effect causes the level of the platform to lower towards the rear of the platform and towards lower-energy downdrift sections of the cliff headland. However, physical wave attack on the edge will slowly erode it, primarily by wave quarrying. All studied platforms had varying degrees of ramparts, which suggests that relatively little platform edge erosion has occurred compared with the still-retreating cliffs. As the platform widens or is covered by sand or boulders, it affords increasing protection of the subaerial cliff, even to the point of it becoming vegetated and no longer retreating. In this way, a steady-state relationship emerges whereby the cliff retreat depends partly on platform width.

Rock type, strength, and structure (bedding and jointing) also play roles in platform morphology, to some extent independent of time. This is evident in both the differential erosion of various rock types and a tendency for higher platforms in more resilient rocks. Rock permeability is clearly important, as is the degree of jointing and differences in hardness of bedding within sedimentary and metasedimentary sequences. The greater the permeability and more finely jointed and bedded the rocks, the lower the degree of rock saturation and lower the platform about the mean sea-level position. This effect was shown in Puerto Rico in limestones (Kaye, 1959). Exposure to high waves and less permeable rocks will induce higher degrees of saturation and higher platforms. Therefore, on a headland, the platform will be higher in more exposed locations and decrease in height as wave energy decreases into embayments, as well as decreasing normal to the platform towards the cliff where wave energy declines. At the same time, the depth to the rock base immediately seaward of the platform will decrease into a lower-energy embayment (Figure 16b). Once formed, the platform remains exposed to subaerial weathering and wave incursions and a range of physical, biological, and chemical processes which can result in positive and negative surface relief features (e.g., Gómez-Pazo, Pérez-Alberti, and Trenhaile, 2021; Short, 1982a).

Rock platforms are continuing to evolve along this coast at present sea level. They represent the interaction of several processes with rock surfaces that are differentially subjected to varying levels of rock saturation. This leads to a distinctive adjustment of form to process around headlands, highlighting the role of exposure to wave energy and weathering in the more permeable and jointed rock types. However, this present mid- to late Holocene state is but a minor subset of antecedent states that have predisposed the present coast to marine forces. These antecedent states have formed over millions of years as the continental margin has subsided, leading to the formation of the rocky central–southern NSW coast and shelf (Thom et al., 2010; Thom et al., in prep a).

Field investigations at five sites along the central–southern NSW coast, together with studies in the same region across a range of rock lithologies, provide insights into rock platform morphodynamics along this coast. All platforms were located above MSL and tend to be nearly horizontal. Platform elevation tends to be lowest against the cliff and into more sheltered parts of embayments, with elevations increasing both longshore and shore normal as exposure to wave energy increases. The cross-shore increase can result in a resilient seaward rampart up to a few metres higher than the platform. On the NSW coast, platforms form in permeable sedimentary, metasedimentary, and igneous rocks and truncate horizontally to vertically dipping beds. They are nonexistent in impermeable intrusive igneous rocks. Rock permeability is negatively related to platform elevation, with permeability also increasing in more prominently jointed rocks.

The dominant ongoing process in platform surface formation is subaerial weathering together with consistent wetting and drying and salt precipitation, all combining to generate accelerated geochemical disintegration (expansion, swelling, solution, flaking, etc.) of the exposed rock eventually down to the permanently saturated rock surface. This provides a control on the elevation of cliff erosion and height of the platform, hence the overall macroplatform morphology. Gravity and occasional wave attack remove the fallen and weaken debris from the cliff and its base.

Wave erosion controls the retreat of the platform edge and base of rock erosion, with the depth of erosion increasing in more exposed area and reaching a 15- to 20-m depth in exposed sites. Waves are the major source of physical energy to erode rocky coasts; however, the focus of erosion is not the platform surface but rather the seaward platform edge.

Since the Holocene stillstand (∼6.5 ka), sufficient time appears to have elapsed for a quasi-equilibrium adjustment to have occurred between the two recession rates and accompanying morphologies around exposed headlands, with apparent minimal retreat of the seaward platform edge, whereas cliff retreat is evidenced by the platforms many tens of metres in width. Wide platforms, persistence of the ramparts, and development of platform surface features suggest that although the cliffs have retreated on the order of less than 1 cm/y, the seaward edge and ramparts have shown minimal retreat.

The model of Holocene platform formation outlined in this paper could apply more generally for those in micro- to mesotidal tropical through midlatitude locations. Here, subaerial weathering dominates platform surface formation. This contrasts those in cooler latitudes, where subaerial weathering has been suggested to be not as effective.

In morphodynamic terms, these rock cliffs and platforms are operating within a subset of spatial and temporal changes that extend laterally out onto the shoreface–inner continental shelf and have operated both subaerially and subaqueously throughout the Cenozoic sea-level history. Under conditions of present sea level, two dominant morphodynamic processes are influencing ongoing evolution of rock platforms: subaerial weathering and gravity- or wave attack–expressing morphodynamic adjustment of process, form, and material.

We acknowledge and thank Dr Michael Kinsela for his assistance with accessing the seamless terrestrial–marine LIDAR dataset and derived digital elevation model. In addition, we thank Peter Crozier for providing a copy of his master’s thesis and Peter Johnson for drafting some of the figures. We also acknowledge the five anonymous reviewers whose comments and suggestions substantially improved the manuscript.