The material that was manually moved during the excavation of the trench was apparent because of its light color in 2013 (Fig. 2a and b). The light color was a consequence of burial that had inhibited weathering by surface processes such as sea spray or lithobiontic weathering (Spiske and Halley, 2014). In 2015 these manually moved clasts were no longer distinguishable (Fig. 2c) because weathering had already darkened their surfaces.

Colonization by epi- and endolithic cyanobacteria (blue-green algae) is an inevitable process that affects coastal sites (Golubic et al., 1980). The rubble ridge at Anegada is positioned in the supralittoral zone, often being wetted by sea spray, as well as by wave splash during storms. This ecological niche is a preferred living environment of cyanobacteria (Radkte et al., 1997; Spencer and Viles, 2002; Gómez-Pujol et al., 2006). The rate of colonization is a function of the temperature, and of the exposure to insolation, waves, spray or splash water (e.g. Folk et al., 1973; Radkte et al., 1997; Schneider and Le Campion-Alsumard, 1999). The activity of the cyanobacteria causes a dark grey to black coating (bio-patina) on the clast surfaces and a few micrometers to millimeters deep (e.g. Folk et al., 1973). Porous, bored or karstified rocks, such as the coral and limestone clasts on Anegada, promote the colonization by endolithics (Hoppert et al., 2004). However, endolithic colonization rates are much slower compared to epilithics (e.g. Viles, 1987; Hoppert et al., 2004; McNamara et al., 2006). Under ideal conditions, colonization rates can be extremely rapid with initial occupation of fresh surfaces within 8–9 days of exposure and a complete overgrowth in three weeks to less than four months (Spencer (1988); and references therein). Spencer (1988) notes that colonization leaves the initially fresh surface visually indistinguishable from the surrounding surfaces. This indeed applies for the surfaces of the clasts on Anegada that were manually moved and put onto the ridge in 2013 that were no longer distinguishable 23 months later (Fig. 2a–c). Warscheid and Braams (2000) reported that freshly quarried stone could be covered by biofilms within months. On the Aldabra Atoll (Indian Ocean), a tropical setting similar to Anegada, Whitton and Potts (1979) documented that supralittoral beach rock which was newly exposed by coastal erosion already had a light steel-blue color caused by blue-green algae within two weeks after exposure. This report is underlined by Folk et al. (1973) who state that algae coatings grow fastest and cover is more dense and abundant in tropical humid climates with high temperatures and in settings with constant wetting by sea spray or splash. Thus, it is not surprising after all, that algae would have darkened the light-colored clasts removed from the trench the next two years. In that case, the surface color of calcareous clasts in tropical climate settings, like Anegada, quickly saturates as a guide to the duration of clasts exposure, i.e. relative dating of ridge emplacement.

In close-up view, the infilled material and the ridge surface in 2015 show a mixture of clasts with light and dark gray (bioweathered) surfaces (Fig. 2d). This distribution results from (i) mixing of weathered and fresh rubble and (ii) overturning of the ridge-derived clasts during transport. The ridge-derived clasts exhibit at least one darkened side, previously being exposed at the ridge surface, and brighter sides that were not directly exposed to spray, splash and insolation. Assuming that hurricane Gonzalo emplaced the clasts in mid October 2014, four months passed until the survey in February 2015. At the time of the survey the clast surfaces of the moved material were still light-colored and did not yet show signs of darkening by biofilms. Consequently, at the given conditions on Anegada, the minimum time for algal colonization is > 4 months. Contrariwise, the maximum colonization time is less than 23 months because clasts that were artificially moved when the trench was dug in March 2013 were no longer distinguishable from longer exposed clasts in February 2015.

Most obviously the trench dug in 2013 was nearly completely infilled (Fig. 2c) and the minor depression as the only visible remain of the previous trench (Fig. 2d) was hard to detect in 2015. Detailed analyses of the trench infill document an average clast size of 15 × 10 × 5 cm, which is very similar to the average clast size in 2013 (16 × 11 × 4 cm). Likewise, the degree of reworking and encrustation (up to 96 %) of the clasts is very high both in 2013 and 2015. The main components that constitute the ridge in 2013 and its infill in 2015 are corals, conch shells, Pleistocene limestone, serpulite rock, and minor portions of reef rock and beach rock (Fig. 4a). The same applies for the main coral species (Acropora, Diploria, Porites, Montastrea; Fig. 4b). The percentages of each component and coral species shifted, but the overall composition did not change significantly, even though some minor constituents like Gorgonia, Siderastrea and gastropods were only counted during one of the surveys. The deviation in percentage and the occurrence of some species in just one of the surveys may be the result of spatial variations in ridge composition. Spiske and Halley (2014) presented these variations when comparing transects I and II.

The composition of the ridge and the infilling of the trench do not significantly differ. This implies that (i) either the source of the material remained the same and still provided sufficient material for the infilling, or (ii) the ridge itself provided (portions of) the material as it was partly reworked by swell, most probably related to the passage of hurricane Gonzalo in October 2014.

Infilling of the trench was only possible during a storm-induced surge that allowed waves to overtop the ∼ 1–1.5 m high coastal platform along Anegada's central north coast (for a detailed view of the platform see Figs. 2a and 3a in Spiske and Halley, 2014). Even during rough sea conditions, with offshore wave heights > 2.5 m, as observed during the survey in 2015, the fringing reef and the shallow lagoon effectively decrease wave heights, not allowing for an inundation of the coastal platform. Consequently, only a storm surge, i.e. an elevated wedge of water pushed towards the coast, on top of which the storm waves can propagate, enables waves to cross the shoreline and the small limestone platform.

The clast-supported framework of the ridge with interlocking components is quite stable. However, storm wave action will at least rework the loose clasts on the surface of the seaward flank of the ridge. At the transect site, material from the uppermost parts of the ridge was eroded and either transported downslope or slumped into the excavated trench. Slumping of clasts from the upper trench walls into the trench may vaguely be observed in Fig. 2d. Nevertheless, it remains unknown which portions of the ∼ 0.74 m3 of infilled material were newly entrained in the shallow marine environment and which derive from reworking of the ridge. If a significant number of clasts was entrained and transported onshore, clasts with all surfaces being unweathered should be found all along the ridge. However, as nearly all clasts have one dark side, it is assumed that reworking of ridge material is the dominant process.

Spiske and Halley (2014) assumed that the rubble ridge was emplaced during a hurricane-related inundation event which was most probably more energetic in terms of surge height, wave speed and wave amplitude, than hurricanes Earl and Donna. However, less energetic storm conditions account for the infilling of the trench that occurred between March 2013 and February 2015. The largest storm at Anegada in that interval was hurricane Gonzalo, which passed about 60 km north of Anegada in October 2014 (category 1–2). Nevertheless, since most of the infill seems to represent eroded ridge material and only a small portion of fresh material was added, it can be referred that such a weak event may not create a ridge, due to the lack of energy to transport clasts onshore, but at least modify a preexisting ridge. Consequently, coarse-clast ridges are vulnerable to even small storms and thus ridge systems are highly dynamic. This conclusion is underlined by a recent study of Xu et al. (2015) who sampled the coarse-clast ridge on Anegada to create a 20th century coral calibration. They counted the growth bands and used uranium-series dating to determine the duration of coral growth represented by two coral clast samples. The end of growth, i.e. timing of the entrainment of the coral and its subsequent death, were dated as AD 1953.3 ± 0.72 and AD 1989.5 ± 0.58, respectively. Of course, these are only two samples, however they show that clasts were most probably added to the preexisting ridge at least for decades. No notable hurricanes are documented to have passed Anegada at the respective times, again pointing to the fact that even minor storm events are capable of adding material to the ridge.

Future storm impact on Anegada should be monitored in greater detail to get information on parameters that can govern the characteristics of hurricane-related inundations, amongst others, the role of the hurricane intensity, distance to the island, alignment of the storm track relative to the island, or direction of swell (Gardener et al., 2005). These factors can influence the surge height and the clast transport capacity of the waves, and under particular conditions even tropical storms or category 1–2 hurricanes may cause severe coastal damage, whereas in turn the effectiveness of category > 2 storm may be attenuated by factors such as swell direction or alignment of the storm track relative to the island.