Pleistocene reefs of the Egyptian Red Sea: environmental change and community persistence

Wadi Wizr

I made 130 identifications of taxa or sediment type on the Wadi Wizr Middle Pleistocene terrace transect, including nine coral species. Coral cover was 60%, coralline algae 12.3%, marine sand 9%, and carbonate mud was 8.5% (see Table 2 for a summary of all terraces). The most abundant coral were massive Porites spp. (65.4% of coral identifications). Goniastrea stelligera (20.5%) was the second most abundant, followed by Stylocoeniella guentheri (6.4%). Cyphastrea microphthalma, Gardineroseris planulata, Astrea curta, Platygyra sinensis, Dipsastraea sp., and Goniastrea sp., all accounted for 1.8% (see Table 3 for species abundances on all terraces). The Middle Pleistocene Wadi Wizr terrace had the second lowest Shannon–Wiener index (H) (1.57) and the fourth lowest Simpson’s diversity index (0.732).

Wadi Gawasis

One hundred seventy-nine identifications were made of taxa or sediment type on the Wadi Gawasis Middle Pleistocene terrace. These included 24 coral species. Coral cover was 78%, non-marine sand was 12.9% of the transect points, and marine sand 9% (Table 2). The most abundant coral were primarily branching Porites spp. (42.9% of coral identifications). Galaxea fascicularis was the second most abundant at 9.3%, and Echinopora lamellosa was the third most abundant at 8.6%. Favites micropentagonus and Goniopora sp. were 7.1% each, and Pocillopora damicornis, Stylocoeniella guentheri, and an unidentified Faviidae were 2.9%. Goniastrea retiformis and Astrea curta accounted for 2.1% each. Goniastrea stelligera, Leptastrea bottae, and Pavona c.f. bipartita were 1.4% each, while Acropora sp., Cyphastrea seralia, Dipsastraea pallida, Dipsastraea speciosa, Echinopora gemmacea, Gardineroseris planulata, Gyrosmilia interrupta, Pavona decussata, Siderastrea savignyana, Favites sp., and Goniastrea sp., all accounted for fewer than 1% of coral identifications (Table 3). Wadi Gawasis had the highest diversity of the three Middle Pleistocene sites and the third highest diversity of all the terraces with an H of 2.42 and a Simpson’s diversity index of 0.844.

Sharm el Arab

I made 110 identifications of taxa or sediment type on the Sharm el Arab Middle Pleistocene terrace transect, including seven coral species. Coral cover was 68.2%, marine sand was 28.2%, and coralline algae was 3.6% (Table 2). The most abundant coral were massive Porites spp. (86.7% of coral identifications). Echinopora lamellosa, Astrea curta, Stylocoeniella guentheri, and an unidentifiable Agariciidae were 2.7% each, and Leptoria phrygia and Pocillopora damicornis accounted for 1.3% each (Table 3). The Middle Pleistocene terrace at Sharm el Arab had the lowest diversity of all eight terraces with an H of 1.16 and a Simpson’s diversity index of 0.569.

Wadi Wizr—upper terrace

One hundred twenty-nine identifications were made of taxa or sediment on the Wadi Wizr Late Pleistocene upper terrace transect, including 20 coral species. Coral cover was 97.7%, marine sand was 2%, and coralline algae 1% (Table 2). The most abundant coral were primarily branched Porites spp. (50% of coral identifications). Galaxea fascicularis (19%) was the second most abundant, followed by Pocillopora damicornis (5.6%). Goniastrea stelligera, Dipsastraea speciosa, and Favites pentagona were 3.2% each, Acropora sp. and Dipsastraea pallida were 2.4% each, and Favites vasta and Hydnophora microconos were 1.6% each. Dipsastraea favus, Echinopora gemmacea, Echinopora lamellosa, Erythrastrea flabellate, Favites abdita, Leptastrea bottae, Lobophyllia hemprichii, Pavona minuta, Platygyra acuta, and Platygyra sinensis all accounted for less than 1% of coral identifications (Table 3). The upper terrace at Wadi Wizr had the fourth highest H (1.92) and the third lowest Simpson’s diversity index (0.718).

Wadi Wizr—lower terrace

I made 142 identifications on the Wadi Wizr Late Pleistocene lower terrace transect, including 11 coral species. Coral cover was 82.4%, coralline algae was 12%, and marine sand was 5.6% (Table 2). The most abundant coral were a mix of massive and branched Porites spp. (57.3% of coral identifications). Galaxea fascicularis (18.8%) was the second most abundant, followed by Pocillopora damicornis (7.7%). Goniastrea stelligera and Favites pentagona were 4.3% each, and Platygyra daedalea was 2.6%. Lobophyllia hemprichii, Cyphastrea microphthalma, Echinopora gemmacea, Dipsastraea speciosa, and Favites paraflexuosus all accounted for 0.9% of coral identifications (Table 3). The lower Wadi Wizr terrace had the fourth lowest H and Simpson’s diversity index (1.75 and 0.728, respectively).

Wadi Gawasis—upper terrace

One hundred thirty-five identifications of taxa or sediment type were made on the Wadi Gawasis Late Pleistocene upper terrace transect, including 16 species. Coral cover was 94.8%, marine sand was 4.4%, and coralline algae was 0.7% (Table 2). The most abundant coral were primarily branched Porites spp. (64% of coral identifications). Pocillopora damicornis (7.2%) was the second most abundant, followed by Platygyra acuta (5.6%) and Platygyra daedalea (4%). Dipsastraea speciosa and Leptoria phrygia were 3.2% each, Goniastrea pectinate was 2.4%, and Cyphastrea microphthalma, Echinopora gemmacea, Favites pentagona, and Leptastrea bottae were 1.6% each. Favites paraflexuosus, Galaxea fascicularis, Hydnophora microconos, Platygyra lamellina, and Acropora sp. all accounted for less than 1% of coral identifications (Table 3). The upper terrace of Wadi Gawasis had the third lowest H (1.74) and the second lowest Simpson’s diversity index (0.634).

Wadi Gawasis—lower terrace

I made 146 identifications on the Wadi Gawasis Late Pleistocene lower terrace transect, including 28 species. Coral cover was 93.8%, coralline algae was 4.8%, and marine sand was 1.4% (Table 2). The most abundant coral were primarily branched Porites spp. (22.6% of coral identifications). Galaxea fascicularis (13.2%) was the second most abundant, followed by Platygyra daedalea and Goniastrea stelligera at 6.6% each. Leptastrea bottae and Pocillopora damicornis were 5.1% each, followed by Dipsastraea speciosa (4.4%), Lobophyllia hemprichii (3.6%), and Echinopora lamellosa (3.7%). Acanthastrea echinata, Favites pentagona, and Goniastrea pectinata were 2.9% each, and Favites paraflexuosus, Hydnophora microconos, Leptoria phrygia, and Platygyra acuta were all 2.2% each. Cyphastrea microphthalma, Dipsastraea favus, Favites flexuosa, and Platygyra lamellina were 1.5% each while Acanthastrea rotundoflora, Dipsastraea matthaii, Favites abdita, Favites vasta, Leptastrea pruinosa, Pavona frondifera, Platygyra sinensis, and Acropora sp. all accounted for less than 1% of coral identifications (Table 3). The lower terrace at Wadi Gawasis had the highest diversity of all the terraces with an H of 2.95 and a Simpson’s diversity index of 0.919.

Sharm el Arab—lower terrace

One hundred fifty-eight identifications were made on the Wadi Gawasis Late Pleistocene lower terrace transect, including 31 species. Coral cover was 86%, marine sand was 10.2%, and coralline algae was 3.8%. The most abundant coral was Galaxea fascicularis (22.1% of coral identifications). The second most abundant were primarily branching Porites spp. (15.4%), followed by Favites pentagona (7.4%), Echinopora gemmacea, and Pocillopora damicornis at 6% each, and Acropora sp. at 5.1%. Cyphastrea microphthalma was 4%, Dipsastraea pallida and Platygyra daedalea were 3.7% each, and Dipsastraea speciosa, Goniastrea stelligera, and Goniastrea pectinata were 2.2% each. Dipsastraea favus, Favites abdita, Favites flexuosa, Favites paraflexuosus, Leptastrea bottae, Pavona cactus, and Platygyra acuta were each 1.5%, while Acanthastrea echinata, Caulastrea tumida, Cyphastrea seralia, Gyrosmilia interrupta, Hydnophora microconos, Leptastrea pruinosa, Leptoria phrygia, Lobophyllia hemprichii, Platygyra sinensis, Pavona maldivensis, and Pavona venosa all accounted for less than 1% of coral identifications. The lower terrace at Sharm Al Arab had the second highest diversity of all the terraces, with an H of 2.87 and a Simpson’s diversity index of 0.915.

Middle Pleistocene terraces

The Middle Pleistocene terraces at Wadi Wizr and Sharm Al Arab have the lowest species richness (Fig. 21), lowest coral cover, and highest sedimentation of all the examined Pleistocene terraces (Table 2). Low diversity and high sedimentation indicates a reef flat or shallow back reef zone (Chappell, 1980). Loya (1972) recorded an average of 8.37 (±3.27) species on 108 m transects on Stylophora pistillata-dominated back reef/reef flats on fringing reefs at Eilat in the Red Sea; Sharm Al Arab is within one standard deviation of this average with seven species found on a 110 m transect (Table 2). Likewise Wadi Wizr is within one standard deviation of 9.37 (±3.49) species found on 135 m transects of back reef/reef flat (see Table 2 in Loya, 1972), with nine species found on a 130 m transect (Table 2). Riegl & Velimirov (1994) characterized coral communities of offshore and onshore fringing reefs with varying degrees of hydrodynamic exposure near Hurghada, Egypt. They did not specifically characterize back reef or reef flat zones, however they found Porites species only accounted for more than 40% of transect intercepts in sheltered reef environments (Riegl & Velimirov, 1994). Like nearly all of the terraces examined here, Middle Pleistocene Wadi Wizr and Sharm Al Arab are dominated by Porites (Table 3), which may indicate a sheltered environment. The presence of fine-grained carbonate mud at Wadi Wizr further supports the idea that it was a low energy, sheltered environment, because very fine-grained sediments only accumulate in low-energy environments.

Riegl & Pillar (1999) described different coral frameworks from the Red Sea. They reported high abundances of massive Porites from sheltered reef edges and reef slopes (68% ± 30% and 85% ± 29%, respectively), however reef edges and slopes do not usually have large amounts of sediment because they have higher wave action (Chappell, 1980). Riegl & Pillar (1999) also describe Porites “carpets,” however these occur in environments without a pronounced slope and are dominated by columnar forms; Wadi Wizr and Sharm Al Arab are fringing reefs, which are sloped, and the Porites at both locations are usually massive (see Figs. 4A and 4B).

The most extensive analysis of coral communities of the modern Red Sea was provided by Sheppard & Sheppard (1991). They surveyed 200 sites, estimating percent coral cover of all species by eye. As in this study, Sheppard & Sheppard (1991) found similar species of Acropora and Porites (as well as Montipora, Goniopora, and Alveopora) difficult to distinguish, so they treated them as single taxon in ecological analyses. Using cluster analysis, they identified 13 distinct types of coral community located along the east coast of the Red Sea. These community descriptions include key features, such as dominant and characteristic taxa, degree of sedimentation, and hydrodynamic exposure. These features can be compared to data from this study to see if the same communities are found on Pleistocene terraces.

Of the communities described by Sheppard & Sheppard (1991), Wadi Wizr and Sharm Al Arab most closely resemble community 4. It is dominated by Porites lutea (47%–85% cover) and is described thus:

The very sheltered bays and back reef slopes of patch reefs in the north and central regions support this community of high Porites cover. Cover values provided by the enormous colonies of Porites range up to 85%, amongst which are the equally characteristic Pavona variens, Pavona cactus, the Pocilloporidae and soft coral genus Xenia. These habitats are always fairly sedimented.

Middle Pleistocene Wadi Wizr and Sharm Al Arab are dominated by Porites (65.4% and 86.7%, respectively) with a co-occurrence of Pocillopora damicornis; the main difference is that Pavona are absent from both sites.

The low species richness, dominance of massive Porites, and high percentage of sediment cover all point to a shallow back reef or reef flat environment for the Middle Pleistocene terraces at Wadi Wizr and Sharm Al Arab. The presence of fine-grained sediments at Wadi Wizr supports sheltered, low-energy conditions, as does the high occurrence of rhodolithic algae (see Fig. 3A). Likewise, large colonies of Goniastrea stelligera account for 20.5% of taxa at Wadi Wizr, and in the Red Sea they are only found in shallow, sheltered environments (Sheppard & Sheppard, 1991). Sharm Al Arab may have been a more exposed site, but the presence of Leptoria Phrygia, which is rarely found deeper than 4 m in the modern Red Sea (Sheppard & Sheppard, 1991), supports a shallow reef environment.

The Middle Pleistocene terrace at Wadi Gawasis has the highest diversity of all the Middle Pleistocene terraces, and the third highest species richness of all terraces (Fig. 21). It also has the third highest sediment cover; more than 12% of it non-marine sand deposited from land (Table 2). High sedimentation is typical of back reef slopes (Chappell, 1980), however this could also be a shallow fore-reef slope on a fringing reef with a very narrow back reef and reef flat. Loya (1972) found an average of 26.67% ± 3.87 species on 90 m transects on fore-reef slopes between 20 and 30 m, which is a similar to the 25 species found on 140 m in this study. This paleodepth could be feasible if Lambeck et al.’s (2011) hypothesis of tectonic uplift in the northern and central Red Sea is correct, but the 9.3% cover of Galaxea fascicularis may indicate Wadi Gawasis was at the shallower end of this range, as the modern depth range of Galaxea fascicularis on clear-water reef slopes is only 5–20 m (Sheppard & Sheppard, 1991). Fossil reefs also have higher diversity than living reefs because they are a combination of the living and death assemblages (Edinger, Pandolfi & Kelley, 2001), so Wadi Gawasis may actually represent a reef slope that had lower live diversity, perhaps indicating a depth shallower than 20–30 m.

A mix of branching and massive forms of Porites spp. account for 42.9% of identified coral on the Middle Pleistocene Wadi Gawasis terrace (21.2% of total). Riegl & Velimirov (1994) only found Porites cover in this range on sheltered platform reefs at 12 m, and none of the reef environments they surveyed had such high species richness (see Figs. 6 and 8 in Riegl & Velimirov, 1994). The Porites reef framework described by Riegl & Pillar (1999) is dominated by massive Porites in shallow water and dominated by branching forms at depth, though they do not report an exact depth for the transition. Although not dominant, Echinopora lamellosa (8.6% of taxa) is the third most abundant species after Porites and Galaxea fascicularis. In the modern Red Sea this species is common on reef slopes deeper than 15 m (Sheppard & Sheppard, 1991).

The high species richness at Wadi Gawasis points to a back or fore-reef slope, and the modern depth range of Galaxea fascicularis, Echinopora lamellosa, and the mix of massive and branching Porites seem to indicate a depth range between 15 and 20 m.

Late Pleistocene terraces

The Late Pleistocene lower terraces at Wadi Gawasis and Sharm Al Arab have the highest species richness of all the surveyed terraces (Fig. 21). Their very high diversity indicates a fore-reef slope environment according to Chappell (1980). The highest diversity found by Loya (1972) is an average of 26.67 (±3.87) species on 90 m transects for reef slopes at a depth of 20–30 m. Wadi Gawasis had 28 species on a 146 m transect and Sharm Al Arab had 30 on 158 m. Although both terraces have a high occurrence of Porites (22.6% at Wadi Gawasis and 15.4% of identified corals at Sharm Al Arab) neither has the very high percent cover seen on sheltered reefs by Riegl & Velimirov (1994). Exposed and semi-exposed reefs surveyed by the same authors are dominated by Acropora and Millipora, respectively, and these species only occur in very low percentages, if at all, on the lower terraces at Wadi Gawasis and Sharm Al Arab (Table 3). Likewise, these terraces do not clearly fit into any of the defined coral carpet or framework reef types described from the modern Red Sea by Riegl & Pillar (1999), although they are closest to the Porites framework reef for which they report a Porites cover of 67% ± 30%.

These two terraces do bear some resemblance to community 5 described thus by Sheppard & Sheppard (1991):

A widespread community of high diversity is found in the north and central region of the Red Sea where there is moderate exposure. They do not differ greatly from community 3 [found between 5 and 15 m], except that they may be dominated by one or two species. Three sub-groups appear according to whether one of the component stony corals attains dominance or not: the most common is one dominated by the hydrozoan Millepora, another small group is dominated by Goniopora … while the third group lacks any strongly dominant species.

Porites are the most abundant taxa at Wadi Gawasis (22.6%) followed by Galaxea fascicularis (11.7%), while Galaxea fascicularis accounts for 22.1% of all taxa at Sharm Al Arab, and Porites 15.4%. Although neither of these taxa are identified by Sheppard and Sheppard as being one of the three sub-groups, these terraces do have high diversity, and they are “dominated by one or two species.” A shallow slope environment is also supported by the relatively high abundance of shallow water species (occurring shallower than 25 m) on both terraces. These include Goniastrea stelligera, Pocillopora damicornis, Lobophyllia hemprichii, Acanthastrea echinata, and Hydnophora microconos (see Table 2, Sheppard & Sheppard, 1991). Galaxea fascicularis is not reported as a dominant or even common taxon on most modern Red Sea reefs (Heiss et al., 2005; Riegl & Velimirov, 1994; Riegl & Piller, 2000; Sheppard & Sheppard, 1991), however it is reported on some slopes between 5 and 20 m, and as very abundant (up to 50% cover) in very shallow, turbid environments (Sheppard & Sheppard, 1991). It could be that’s its relatively high abundance on the Wadi Gawasis and Sharm Al Arab terraces indicates a turbid environment, or it may be that this taxon was more common on Pleistocene reefs than it is today.

The case for a relatively shallow fore-reef slope is strengthened by the poorly sorted coarse-grained marine sand found on both terraces, which is typical on reef slopes below the reef crest where rubble dominates sediments, and the deep fore-slope, where sediments are usually much finer (Dudley, 1996). The high occurrence of marine sand (10.2% of transect points) at Sharm Al Arab also strengthens the argument for a relatively turbid environment, because the finer grains in this poorly sorted sediment would have been available for resuspension if hydrodynamic conditions allowed. The very high diversity, high coral cover, and abundance of shallow water taxa indicate the Late Pleistocene lower terraces at Wadi Gawasis and Sharm Al Arab represent the shallow fore-reef slopes of fringing reefs. Comparisons to Sheppard & Sheppard’s (1991) community 5 suggest it may be a moderately exposed site and shallower than 15 m. The high abundance of Galaxea fascicularis may mean it was also relatively turbid.

The upper and lower Late Pleistocene terraces along the Red Sea coast have both been dated to the MIS 5e, and the platform morphology was determined to be a result of erosion (Plaziat et al., 2008), therefore the upper terrace and lower terrace represent two depths on the same fringing reef. The upper terrace is approximately 2.5 m higher (shallower) than the lower terrace. At Wadi Gawasis, the upper terrace has lower species richness than the lower terrace, which meets the expectation of increasing diversity with depth on shallow fore-reef slopes (Chappell, 1980). Sixteen species were identified from 135 m on the upper terrace at Wadi Gawasis, which is within range of the 12.33 ± 4.89 species reported from 120 m transects at depths of 3–13 m by Loya (1972).

Branching forms of Porites spp. account for 64% of taxa on the upper terrace. This abundance is consistent with reef edge or shallow reef slope environments on sheltered reefs reported by Riegl & Pillar (1999), however they note that massive forms of Porites dominate at the reef edge and upper slope, and columnar forms only become more common at depth. The general trend in coral growth form (independent of taxon) is that growth becomes more massive with increasing hydrodynamic stress, but more branching with increasing sedimentation (Chappell, 1980; Perrin, Bosence & Rosen, 1995). If the Wadi Gawasis site was relatively turbid, it may account for the branching growth form of Porites at shallower than expected depths.

Based on the high abundance of Porites, the upper terrace at Wadi Gawasis most closely resembles community 4 described by Sheppard & Sheppard (1991), however the terrace lacks the high sediment cover characteristic of this community (Table 2). Even without an exact corollary in the literature on modern reefs of the Red Sea, the moderate diversity and proximity to the lower terrace strongly suggests this was a shallow upper slope environment.

The Late Pleistocene lower terrace at Wadi Wizr has the lowest diversity of all the Late Pleistocene terraces (Fig. 21). This is consistent with a reef flat or shallow back reef zone (Chappell, 1980). Loya (1972) found an average of 9.37 (±3.49) species on 135 m transects in back reef and reef flat environments, and the lower terrace at Wadi Wizr had 11 species on 142 m transects, which is within one standard deviation of Loya’s average.

A mix of massive and branched Porites accounted for 57.3% of taxa, which is near to the values found on sheltered platform and fringing reefs by Riegl & Velimirov (1994), and within range of sheltered reef edges (68% ± 30%), and very nearly in range of reef slopes (85% ± 29%) found by Riegl & Pillar (1999) on Porites framework reefs. This terrace doesn’t fit well with any communities described by Sheppard & Sheppard (1991), however their analysis excluded reef flats.

Theoretically, diversity on reef flats first increases from reef crest toward the back reef as wave energy decreases, then begins to decrease again as subaerial exposure increases (Chappell, 1980). In this framework, the lower terrace at Wadi Wizr could be a reef flat close to the reef edge, or a reef edge. This is further supported by the relatively high abundance of coralline algae (12%), which is characteristic of reef crests (edges) (Perrin, Bosence & Rosen, 1995; Riegl & Pillar, 1999).

In this interpretation the higher diversity of the Late Pleistocene upper terrace at Wadi Wizr can be explained as the higher diversity expected on mid reef flats relative to reef edges, where wave energy is minimal and subaerial exposure is not yet a factor (Chappell, 1980). The high abundance of shallow water taxa on both terraces, including Pocillopora damicornis, Goniastrea stelligera, and Lobophyllia hemprichii supports the idea of a shallow water environment (Table 3), and the high abundance of Galaxea fascicularis on both the lower (18.8%) and upper (19%) terraces (Table 3) suggests a turbid environment, which is consistent with a wave-washed reef flat. The upper terrace at Wadi Wizr is dominated by branching Porites (50%), which may also be expected in a turbid environment, and the mix of branching and massive Porites on the lower terrace might be expected on a turbid reef edge.

Middle and Late Pleistocene assemblages

In this study, only 30.7% of identified taxa occurred on both Middle and Late Pleistocene terraces. A total of 26.9% were only found on Middle Pleistocene terraces and 42% were only found on Late Pleistocene terraces. However, this is most likely an artifact of sampling bias: five transects were carried out on Late Pleistocene reefs, and only three were carried out on Middle Pleistocene reefs. This bias is further exacerbated by the paleoenvironments preserved, because two of the three Middle Pleistocene terraces preserve low-diversity, back reef environments. When species data are compiled for all the published studies on Red Sea fossil reefs (Table 1) only one species, Favites micropentagonus, is found exclusively on Middle Pleistocene terraces.

Favites micropentagonus was first described in 2000 (Veron, 2000). Originally recognized from locations within the Coral Triangle, a marine biodiversity hotspot centered on Indonesia and Papua New Guinea, it has since been identified on reefs in the Seychelles Archipelago in the western Indian Ocean (Sheppard & Obura, 2005), the Gulf of Aden (DeVantier et al., 2004), the Gulf of Oman, eastern Africa, Madagascar, Chagos Archipelago, Thailand, Southeast Asia, Vietnam, southern Japan, Papua New Guinea, southern coast of Solomon Islands, and Micronesia (DeVantier et al., 2008). Fossil specimens have been reported from the Miocene of Fiji, the Quaternary of Indonesia (Bromfield, 2013), and the Quaternary of Japan (Humblet, Iryu & Nakamori, 2009). The coral reefs of the Red Sea are understudied relative to those found in the Caribbean or on Australia’s Great Barrier Reef (Berumen et al., 2013), so it’s possible Favites micropentagonus’s current range extends into the Red Sea but it hasn’t been recognized yet. It may also have lived there during the Middle Pleistocene, suffered local extinction, and failed to recolonize the Red Sea during later interglacial periods.

Pavona decusassata and Stylocoeniella guentheri are the only other species identified from the Middle Pleistocene in this study that have not been reported from the Late Pleistocene anywhere else in the Red Sea, although they are reported from the modern Red Sea. This could be another artifact of inadequate sampling of fossil reefs at all time periods in this region, however Stylocoeniella guentheri at least is reported as a very common coral in the modern Red Sea, occurring from 5 to 50 m, in every environment from lagoonal back reefs to deep fore slopes, in clear and turbid water, and in any degree of hydrodynamic exposure from sheltered to exposed (Sheppard & Sheppard, 1991). It’s relatively high abundance on all three Middle Pleistocene terraces in this study (Table 3) suggests that it was equally common during the Middle Pleistocene. If it lived in the Red Sea during the Late Pleistocene, it seems highly probably that it would have been found already.

The individual-based rarefaction curves in Fig. 21 provide a reliable comparison of species richness between Middle and Late Pleistocene coral terraces despite differences in sampling intensity (Gotelli & Colwell, 2001). On living reefs, the expectation is that diversity varies with zonation (Chappell, 1980; Perrin, Bosence & Rosen, 1995). That’s why in his study of fossil coral reefs in Papau New Guinea, Pandolfi (1996) only compared communities belonging to the same reef zone to see if diversity had change through time. In this study, there are six reef zones represented on eight reef terraces, so it isn’t possible to perform a rigorous, statistical comparison of zone diversity across time. However, Pandolfi’s study (1996) and a similar study by Pandolfi & Jackson (2006) in the Caribbean, both found diversity of zones stable across Pleistocene glaciation events. My results meet the expectation that diversity varies by reef zone, and there is no strong temporal pattern that can’t be explained by differences in zonation, which is consistent with the previous studies (Pandolfi, 1996; Pandolfi & Jackson, 2006).

The diversity found on each terrace is consistent with the theoretical expectations for each reef zone identified for it. For example, the lower terraces of the Late Pleistocene at Wadi Gawasis (G1) and Sharm Al Arab (A1) are both likely shallow fore-reef slopes; they have the highest diversity of all the studied terraces, which is consistent with fore-reef slopes being the most diverse zone on a fringing reef (Chappell, 1980). The Middle Pleistocene terraces at Wadi Wizr (W3) and Sharm Al Arab (A3) are both likely sheltered back reef environments, and they have the lowest diversity of all the terraces, which is expected for shallow back reefs (Chappell, 1980). Relative diversity on the remaining terraces is also consistent with the theoretical expectations for diversity in different reef zones. For example, the Late Pleistocene reef crest at Wadi Wizr (W1) has low diversity similar to the Middle Pleistocene back reefs (W3 and A3), while the Middle Pleistocene shallow slope (back or fore-reef) at Wadi Gawasis (G3) has moderately high diversity, similar to a reef flat with limited exposure and limited wave stress found on the upper terrace of Wadi Wizr (W2) of the Late Pleistocene.

Although species richness is determined by reef zone, a cluster analysis using the Chao dissimilarity index shows terraces grouping by age rather than environment (Fig. 22). When Pandolfi (1996) and Pandolfi & Jackson (2006) performed similar analysis, they found that reef communities clustered by geographical location rather than time period, indicating no significant change in reef species across Pleistocene glaciations. The small number of transects used in this analysis means that any interpretation needs to be tested with more data, specifically with more terraces of the same reef zones. However, one simple explanation for the difference between my cluster results and those of Pandolfi (1996) is in our different approach to rare taxa. The Chao dissimilarity index weights shared rare taxa, which I included in my analysis, while Pandolfi (1996) excluded rare taxa from his analysis. It makes sense that rare taxa would be more vulnerable to environmental shifts, and therefore more likely to change through time. They are also less likely to be sampled, and therefore less reliable for characterizing communities.

Pleistocene and modern assemblages

All but two taxa reported from the Pleistocene of the Red Sea are reported from the modern Red Sea: Favites micropentagonus (see Discussion), and Pavona minuta. Pavona minuta has been reported from the Late Pleistocene of the Red Sea by Dullo (1990), and from the modern Gulf of Oman (DiBattista et al., 2015). It’s possible that Pavona minuta occurs so rarely in the modern Red Sea that it has yet to be identified, or it may have gone locally extinct during the last glacial low stand. Pavona minuta also resembles Pavona duerdeni, a species that is found in the modern Red Sea but which has larger corallites and more exert septo-costae. Thus another possibility is that these Pleistocene Pavona minuta are morphological variants of Pavona duerdeni, or that Pavona minuta in the modern Red Sea are routinely misidentified as Pavona duerdeni. Based on preservation (see Fig. 19C) the one Pavona minuta specimen identified in this study looks younger than Late Pleistocene, and may actually be Holocene-aged.

DiBattista et al. (2015) report twenty-one endemic coral species from the modern Red Sea, and none of them have been reported from the Pleistocene of the Red Sea. Many Red Sea endemics have only been described recently: 13 of 21 have been described since 2000 (DiBattista et al., 2015). It is possible that they have been found by earlier fossil workers but misidentified; however, DeVantier et al. (2000) indicate most of these newly described taxa are distinctive and unlikely to be confused with another species. It is more likely that they haven’t been found because (1) they are rare, and (2) Pleistocene Red Sea reefs are under-sampled. The recently described Red Sea endemics occur in relatively low abundances (DeVantier et al., 2000), and currently 121 coral species have been identified from the Pleistocene of the Red Sea (Table 1), while 346 have been reported from the Modern Red Sea (DiBattista et al., 2015). Even if taxa were significantly different between the Pleistocene and the present, there is no reason to think diversity would be lower in the past, which suggests two-thirds of Pleistocene coral species have yet to be identified.

Both Erythrastrea flabellata and Favites vasta were once thought to be endemic to the Red Sea (DeVantier et al., 2000), but recently they have also been reported from the Gulf of Aden and Socotra outside the Bab al Mandab (DiBattista et al., 2015). To the best of my knowledge, this study is the first report of these species from the Pleistocene: a single colony of Erythrastrea flabellata was found on the Late Pleistocene upper terrace at Wadi Wizr where Favites vasta was also found. The latter was also found on the Late Pleistocene lower terrace at Wadi Gawasis. As more fossil terraces are studied it seems highly probably that more rare and endemic species will be found.