The oldest known stromatolites include those of the Isua Group (3.7 Ga), Greenland (Nutman et al., 2016); the Strelley Pool occurrence (3.43 Ga), Australia (Allwood et al., 2006); the Barberton Mountain Land (3.22 Ga), South Africa (Gamper et al., 2011); and the Pongola Group (2.9 Ga), South Africa (Mason and von Bruun, 1977; Bolhar et al., 2015). Comparison of the environment in which extant and sub-fossil stromatolites occur with ancient examples may advance our knowledge concerning the conditions under which life developed and in what environment it began. Stromatolite-building organisms probably dominated the Earth during the Archean and Proterozoic eons, but under contemporary conditions only thrive in extreme environments that limit metazoan competition. Such environments include geothermal springs (Jones et al., 2000; Berelson et al., 2011), peritidal marine environments (Logan et al., 1964; Reid et al., 2000; Smith and Uken, 2003; Smith et al., 2005, 2011; Perissinotto et al., 2014; Rishworth et al 2016; Edwards et al., 2017), and salt lakes (Martin and Wilczewski, 1972). Prokaryotes are also recorded from the Earth's upper crust to depths of (at least) 7 km (Sankaran, 1997) and within the troposphere (DeLeon-Rodriguez et al., 2012).

Stromatolites are biosedimentary structures produced by sediment trapping and binding and/or mineral precipitation as a result of the growth and metabolic activity of a microbial community (Awramik and Marguilis, 1976; Burne and Moore, 1987). The best known extant stromatolite models are based on the trapped and bound stromatolites of Shark Bay, Western Australia and several sites in the Caribbean (Logan et al., 1964; Reid and Browne, 1991). Precambrian stromatolites, in contrast, are of the mineral precipitation variety (Awramik and Grey, 2005; Reid et al., 2011; Smith et al., 2011; Perissinotto et al., 2014; Rishworth et al., 2016; Edwards et al., 2017) and offer a plausible Precambrian stromatolite analog, which deserves scrutiny.

Cape Morgan (Fig. 1) in the Eastern Cape province, South Africa, was the first location where SPS were found. SPS form on shore platforms within the peritidal zone where suitable (carbonate-rich) terrestrial runoff is present (Smith et al., 2011; Perissinotto et al., 2014; Rishworth et al., 2016). These supratidal/ high intertidal zones experience extreme environmental changes, which partially exclude metazoans (Perissinotto et al., 2014; Rishworth et al., 2016) and enable prokaryotes to dominate. The Cape Morgan locality was discovered by Mountain (1937), although its significance was only realised much later (Smith and Uken, 2003). Smith et al. (2011) proposed the extant Cape Morgan (Fig. 1) stromatolites as a partial analogue for the 3.43 Ga stromatolites from Strelley Pool Australia. Smith et al. (2011) furthermore documented extant and subfossil SPS from Cape Morgan and indicated that they were found in patches from Tofo (Mozambique) to Port Elizabeth (Eastern Cape, South Africa)(Fig. 1). Since then colonies have been found in Northern Ireland (Cooper et al., 2013), additional localities in the Eastern Cape province have been documented, (Perissinotto et al, 2014; Rishworth et al., 2016; Edwards et al., 2017),and further colonies elsewhere in southern Africa have been uncovered (Fig. 1). More recently, new SPS discoveries have also been made on the west coast of Harris, in the Scottish Hebrides.

This work is heavily reliant on fieldwork which has taken advantage of many serendipitous trips. No attempt has been made to institute a scientific survey due to the geographical distances involved and limited manpower. Stromatolite (SPS in this paper) morphology, microstructure, and ecology is discussed in detail elsewhere (Smith et al., 2011; Perissinotto et al., 2014; Rishworth et al., 2016, 2017; Edwards et al., 2017). New localities are discussed and compared with known; such review is required to establish similarities and differences. This paper will consequently focus on the synthesis of a facies association using tabulated geomorphological, lithological, oceanographic, and climatological elements. This is a relatively new branch of stromatolite science and is heavily reliant on description and comparisons.

We document all known SPS localities (Table 1) with the aim of distilling of a generalised global SPS facies association. To this end we also describe several new SPS localities on the southern African eastern seaboard and on the Atlantic side of Harris Island in the Scottish Hebrides, UK. We then compile and review the physical, oceanographic, climatological, lithological, and geomorphological settings in which extant SPS are forming in the Scottish Hebrides, the northeast coast of Northern Ireland and the southern African seaboard. All known SPS location coordinates and source data are given in Table 1. On the basis of this review, we present a facies association model for SPS.

We then compare the SPS facies association to extant stromatolites from contrasting peritidal environments in Hamelin Pool, Western Australia (Logan et al., 1964), Kuwait (Alshuaibi et al 2015), the Bahamas (Reid et al., 2000), and the Dutch North Sea (Kremer et al., 2008). Further, on the basis of this comparison, extant SPS are assessed as potential analogues for Precambrian stromatolites. We also comment briefly on potential target location regarding the search for SPS on Mars.

All locations discussed are open coast except for Luskentyre Bay and Kuwait Bay (Table 2). The geomorphology of each shore platform is controlled by lithology, jointing, and bedding style. Shore platforms comprised a variety of competent lithologies (Table 2). Shore platforms vary from 5–60 m wide and are generally backed by a boulder beach or boulder ridge (Fig. 6). The boulder ridges contain angular blocks or megaclasts (up to 80 t), as opposed to the smaller (> 50 cm diameter) rounded boulders found in boulder beaches and gullies in, and adjacent to, the shore platform (Table 2).

Where the coast comprises incompetent lithologies no shore platforms can form. Competent sandstones form wide shore platforms (Fig. 6; Table 2). The SPS bearing shore platforms at Mtentu and Luphatana (Fig. 6c; Table 2) are formed in well-indurated Lower Devonian Msikaba Formation sandstone. These are the widest at up to 60 m wide. The Msikaba Formation is well-bedded (more-or-less horizontal) and vertically jointed. The shore platform has been formed by wave quarrying of large blocks, which break along bedding and joint planes. The eroded boulders have accumulated in a boulder ridge at the rear of the platform (Fig. 6a).

At Cape Point SPS occurrences are rare but massive tufa deposits are present inland. Tufa is the main lithological component at Tofo, in comparison, and the shore platform has been excavated into this lithology (Fig. 6b), with thin (1–2 cm thick) sub-fossil stromatolites interbedded (Table 2). Sub-fossil mineral precipitated stromatolites are also found as upper intertidal and supra tidal active pothole linings. No growing tufa or stromatolites were observed, but the presence of SPS in potholes within an active shore platform indicates them to be recent. It is possible that this tufa deposit is related to a lower sea level stillstand and has been reworked into a shore platform where SPS has developed at a later stage.

Dolerite sill shore platforms are present at Tinley Manor, Ballito, and Cape Morgan (Fig. 1); the latter (Fig. 7) being the best example (Table 2). The dolerite is strongly jointed and forms a rugged shore platform, which tends to undulate and shows minor sea cliffs and pools. The sea cliffs are produced by wave bore quarrying of blocks along joint surfaces. Pools, in comparison, are formed by mechanical joint widening, pot holing, and chemical weathering, while barrage pools are impounded by stromatolite growth (Fig. 4).

At Port Edward (Fig. 1) granite forms a poorly developed shore platform littered with megaclasts. Minor SPS and tufa were noted near the landward shore platform boundary. The Luskentyre Bay, Harris Island, Scottish Hebrides (Fig. 1) granitic shore platform is backed by an extensive but thin (±30 cm) peat bog. Seaward of this shore platform is a very extensive intertidal fine-grained low-end macrotidal flat (Fig. 6).

At Richards Bay (Fig. 1) the south-east African coastline is marked by the semi-indurated Pleistocene Port Durnford beds (Fig. 1), which comprise incompetent muds, silts and fine sands. The shore platform is poorly developed and backed by a retreating coastal cliff made up of the poorly consolidated Port Durnford Beds sediments. Only tufa is present growing on the cliff face but will not be preserved due to ongoing marine action.

Storm swash terrace deposits (McKenna et al., 2012; Dixon et al., 2015) occur at the rear of the Port Edward, Ballito, Tinley Manor, Mtentu, Luphatana, and Cape Morgan shore platforms. These are associated with boulder ridges, boulder beaches, and coastal dunes. Storm swash terrace deposits (McKenna et al., 2012) may partially bury older beach ridges. Boulder beaches are present at the Luskentyre Bay, Harris, Scottish Hebrides (UK), SPS sites but here the hinterland is characterised by peat bog overlying bedrock which projects through at high points (Fig. 6). At Tofo (Mozambique) the tufa is backed by a very extensive (kilometres wide) coastal dune cordon, which is strongly impacted by farming and urbanisation. Several SPS localities have a coastal dune cordon hinterland and some are associated with bogs (Table 3). The hinterland at Giant's Causeway comprises high basalt cliffs (Fig. 6). A model of the SPS facies association is shown in Fig. 7.

SPS occur in a variety of climatic and oceanographic settings (Table 4) and although no systematic survey of SPS distribution has yet been undertaken, we expect that they are globally widespread. Shore platforms, associated with boulder ridges and boulder beaches, are indicative of high wave activity (Hall et al., 2006; McKenna et al., 2012; Dixon et al., 2015). In the case of the southern African eastern seaboard this is confirmed by modally high wave conditions (Guastella and Rossouw, 2012). Extreme waves, however, exceed modal conditions by several metres. On the southern African eastern seaboard, extreme waves may originate from extratropical low pressure systems, tropical storms/ cyclones (Smith et al., 2010), and tsunamis can also not be ruled out.

A high swell event on 18–20 March 2007 (impacts ranged from Port Elizabeth to Maputo) (Fig. 1), was produced by a cut-off low pressure system. This event produced swells up to 14 m high (Hs=8.5 m) with run-ups of 7–11 m (a.m.s.l.). (Smith et al., 2007). A further high swell event between 31 August and the 1st September 2008 impacted the southern and south-east African coast, from Cape Point to Cape Morgan (Fig. 1). This was generated by an extremely deep low pressure system and produced swells of Hs=10.7 m (Guastella and Rossouw, 2012) with a 7–8 m run-up at Cape Morgan (Smith et al., 2014).

Field inspection following both the March (2007) and September (2008) high swell events indicated changes to the Cape Morgan colonies (the other colonies listed in Table 1 were unknown prior to 2007). In the case of both storms, growing and unconsolidated microbial mat and large blocks of tufa were largely removed by wave action, but lithified stromatolites remained on the shore platform. Following the March 2007 event, surface stromatolite growth at Cape Morgan had been largely restored by January 2008.

At Ballito (Fig. 1), sub-fossil peritidal stromatolites that had been buried under coastal reclamation were exposed by erosion during the same 2007 high swell event. This event deposited a boulder beach (boulders > 1 m diameter) over the dolerite sill shore platform. At one point a thin crust of sub-fossil stromatolites was observed, which extended landward under the boulder beach and storm swash terrace. This shows that SPS can be interbedded with storm deposits. At Cape Morgan, boulders on the boulder beach are often coated by stromatolite growth indicating periods of disaggregation and stability. Several rounded boulders were found with multiple rims of stromatolite, indicating they had experienced several cycles of stability and movement.

At Luphatana an extant SPS colony is located just in front of a boulder ridge which contains 80 t boulders, indicative of extreme waves (date unknown), the parameters of which are as yet unquantified (Fig. 5a). Similarly the St Johns Point and Giant's Causeway stromatolites (Cooper et al., 2013) are both associated with very large boulders. Wave spray deposits have been found at levels of 60 m (a.m.s.l.) at Cape Morgan, South Africa (Smith et al., 2014), whilst at Aird Uig (Harris, Scotland) wave scouring has taken place on a cliff top at a height of 20–30 m OD (Hall et al., 2006), proving that SPS can withstand the present extremes in wave climate.

The Irish and Scottish coasts are known for strong wave action (Hall et al., 2006; O'Brien et al., 2013) as the 2013–2014 storm season proved (Wadey et al., 2015). Both Giant's Causeway (on the west coast) and Luskentyre Bay are afforded some protection, but the boulder storm beaches at both localities (Table 2) have shown themselves to be vulnerable to wave attack.

It must be restated that this sample is far from complete, representing only coastlines that have been investigated. From Table 4, however, the following about extant SPS can be noted:

SPS have been recorded in passive plate margins and epeirogenic settings.

SPS grow in shore platform depressions on high-energy coasts, such as shallow rock pools (Smith et al., 2011; Perissinotto et al., 2014; Rishworth et al., 2016), potholes, and SPS dammed barrage pools (Forbes et al., 2010; Perissinotto et al., 2014). SPS comprise thin micritic crusts, with only rare examples of trapped and bound stromatolite varieties being associated. Not all localities show the complete tripartite stromatolite morphology (Fig. 4a). The pustular variety may be absent but the subaqueous laminar and columnar and the colloform variety are ubiquitous. Where vertical surfaces are present tufa grows.

Any SPS model must take into account that they form at the interface of freshwater seeps and high energy rocky peritidal zones (Smith et al., 2011; Perissinotto et al., 2014). The SPS are calcium carbonate mineralised due to the high pH regime (Smith et al., 2011) and grow on older siliceous rocks (Tofo may be an exception but the base is not visible) within a siliciclastic contemporary setting. Mud and sand pass through this system but are not deposited in significant amounts. Microbialites are ubiquitous in the peritidal setting but only become stromatolites if a carbonate rich groundwater plume is present (Smith et al., 2011; Rishworth et al., 2016).

The following points also need to be considered, although they may, or may not, be necessary for SPS growth. Sea level rise is ongoing. SPS possess seasonal to sub-seasonal (due to storm abrasion) laminae (Smith et al., 2005). The lack of climax lamination (Reid et al., 2000), characterised by diatoms (Smith et al., 2005) in the calcified form may be due to storm activity or non-calcification of this lamina type. Warm thermal groundwater may be present indicating a groundwater source. The stromatolites and shore platform contact is an unconformity.

The diversity of shore platform substrates suggests that lithology is not important for SPS growth; however, the competency of the substrate is vital for growth and probably for potential preservation. Tufa deposits and pustular stromatolite deposits are very unlikely to survive transgression or regression as there is no accommodation space. However, SPS in rock pools and barrage pool build-ups can form lenses or layers of stromatolite which could then be overstepped during marine transgression (possibly similar to the multi-metre steps which post-dated the last glacial maximum) and preserved in a future stratigraphy. Marine processes may break up the stromatolites and free stromatolite encrusted boulders from barrage and rock pools, such as is seen at Cape Morgan (Fig. 2a), and transport them as littoral drift for deposition elsewhere as conglomerates. Sea level rise (SLR) is taking place globally, thus SPS deposits could form as part of a global transgressive coastal sequence.

The best opportunity for preservation is provided by rock pools, especially potholes, in competent shore platform rock as this provides accommodation space. SPS growth itself could seal the SPS deposit, especially in flat bedded competent sandstone as at Mtentu and Luphatana (Fig. 5). Thus SPS environments could be preserved as lenses on palaeo-shore platforms. The extant stromatolite regrowth seen at Ballito (Fig. 1) suggests that preservation can take place. An investigation of subaqueous post-last glacial maximum coastlines might resolve this issue.

Extant SPS are present on high energy rocky coasts, on passive margins, and in epeirogenic settings; however this may be a function of sampling. They are associated with well-indurated shore platforms, boulder ridges, and boulder beaches. The hinterland is frequently boggy or marshy. SPS appear to be unrelated to climate and tidal regime, but they may be better developed in warmer climates but the sample size is currently too small to confirm this. If vertical surfaces are present on the shore platform, tufa forms. In shallow shore platform pools SPS develop. There is a gradation from SPS to tufa. SPS develop within transgressive settings, as was the case with many Precambrian stromatolites. The preservability of these micrite stromatolites is probably low, but may improve with stromatolite cementation to competent substrate rock. Contemporary rocky and Quaternary coastlines should be investigated globally for the SPS environment. The SPS setting is a valid analogue for the Archean Strelley Pool stromatolite and possibly the base of the Gunflint stromatolite occurrence; however, most Precambrian stromatolites appear to have formed on soft coastlines. The association with trapped and bound stromatolites and the Luskentyre Bay environment (shore platform and tidal flat) hints at a possible SPS soft coastline link. Finally, shore platform settings should be targets in the search for Martian stromatolites.