### BISA93 — Salomon
**CRD page:** [Salomon Islands Coral Survey April 2015](https://coralreefdiagnostics.com/salomon-201504)

**25 Apr 2015**
**GPS:** 5.29926°S, 72.25788°E
**Setting:** exposed (published as windward) emergent forereef, outside the lagoon, 30.4 m
**Coral cover:** 39.1% (19 transects across 4 depth strata, D1–D4; **algal cover 34.9%**) · **published image bleaching:** 11.9% (`Mean(%bleaching)` column)
**Live hard coral cover across methods:** formal transects 39.1% · CoralNet points 18.3% · Coralscapes masks 17.1% (n_photos = 28). These products have different sampling footprints and denominators and should not be averaged.
**Temperature:** 30.0 °C · **Salinity:** 34.2
**Photo review:** all 28 of 28 jpg stills examined. The folder also holds five video files that are outside the scope of this still review, two of which are 32-byte iCloud placeholder stubs with no content (see Flag 8). The folder label "great site" is retained as a field note, not a score.
**Representative photos:** P4250948.jpg, P4250949.jpg, P4250954.jpg, P4250971.jpg, P4251002.jpg

BISA93 is an exposed, windward outer-slope forereef on the seaward side of Salomon, surveyed at 30.4 m on 25 Apr 2015. **Coral cover of 39.1% ranks 56th of 109 Chagos sites and algal cover of 34.9% ranks 71st of 109** — squarely mid-pack on both. What sets the site apart is not its cover but its diversity and its condition, which pull in opposite directions. Bare substrate is very low at 5.5%, and the two large non-coral categories are crustose coralline algae at 19.2% of the benthos (30th of 109) and **non-coral invertebrates at 20.5%** — the latter unusually high and, in the imagery, plainly attributable in part to the conspicuous gorgonian sea fans, whip corals and encrusting sponges that clothe this slope. Fleshy turf holds 10.2% (51st of 109), with macroalgae 3.2%, sediment-laden turf 1.3%, erect algae 0.7% and cyanobacteria 0.3%. Coral, algae, bare substrate and invertebrates sum to exactly 100.0% of the benthos, so nothing is unallocated.

**Genus richness is the site's headline strength: 34 non-zero genera, 5th-highest of the 109 Chagos sites.** All values below are **percent of total benthos, not percent of coral**, and the 34 genus percentages sum to 39.06%, an exact match to the published cover figure. ***Porites* leads decisively at 18.99% of the benthos — 49% of every coral point recorded** — followed by *Acropora* 5.87% (15%), *Pocillopora* 3.39%, *Montipora* 2.42%, *Stylophora* 1.50%, *Favia* 1.45% and *Pavona* 1.09%. Below 1% the tail is long and genuinely diverse: *Montastrea* 0.68%, *Platygyra* 0.61%, *Leptastrea* 0.59%, *Favites* 0.42%, *Isopora* 0.40%, *Heliopora* 0.18%, *Goniastrea* 0.17%, *Fungia* 0.16%, *Symphyllia* 0.14%, *Pachyseris* 0.14%, *Seriatopora* 0.12%, *Alveopora* 0.11%, *Cantharellus* 0.09%, *Cyphastrea* 0.08%, *Turbinaria* 0.08%, *Astreopora* 0.08%, *Lobophyllia* 0.05%, *Hydnophora* 0.04%, ***Ctenella* 0.04%** — a Chagos/Indian Ocean endemic worth noting — *Echinopora* 0.04%, *Stylocoeniella* 0.04%, *Leptoseris* 0.03%, *Mycedium* 0.03%, *Galaxea* 0.01%, *Gardineroseris* 0.01%, *Physogyra* 0.01% and *Psammocora* 0.01%. A site that is half *Porites* by cover and still carries 34 genera is doing two things at once.

**The depth profile is a sharp peak followed by a collapse, and it is the real finding here.** **D1 holds 38.0% coral against the site's highest algal cover, 47.0%, across 12 genera** — *Porites* 13.40%, *Acropora* 8.60%, *Pocillopora* 6.00%, *Montipora* 2.90%, *Stylophora* 2.70%, *Favia* 2.10%. **D2 is the best band on the reef by a wide margin: 58.6% coral against only 26.2% algae, with *Porites* alone at 36.70% — 63% of the coral on that stratum** — plus *Acropora* 9.57%, *Pocillopora* 3.70% and a 20-genus assemblage behind them. **D3 falls back to 39.4% coral on 25.6% algae but is the most genus-diverse stratum on the reef, 26 genera**, still *Porites*-led at 19.37% with *Acropora* 4.86%, *Pocillopora* 2.83% and *Pavona* 1.69%; this is the band where *Heliopora*, *Alveopora*, *Ctenella*, *Stylocoeniella* and *Seriatopora* all appear. **D4 then collapses to 20.25% coral — barely a third of the D2 figure — against 40.8% algae and 17.1% bare substrate, and the assemblage reorganises entirely: *Porites* drops to 6.50%, ***Acropora* falls to 0.45%, essentially out of the reef***, and *Montipora* becomes the second-ranked genus at 3.65%**, followed by *Montastrea* 1.55%, *Pavona* 1.35%, *Platygyra* 1.05% and *Pocillopora* 1.05% across 21 genera. The site-level figures of 18.99% *Porites* and 5.87% *Acropora* describe no single stratum: they average a *Porites*-dominated mid-depth peak against a deep, algae-dominated, *Acropora*-free band.

**The headline number is mildly weighting-sensitive.** Site cover is the unweighted mean of the four per-stratum means, 39.06%; the same 19 transects (D1 n = 2, **D2 n = 6, D3 n = 7, D4 n = 4**) re-computed as a transect-weighted mean give **41.28%**, a 2.2 pp difference, because the two best-replicated strata are the two better ones. **The one thin stratum is D1, resting on two transects** — worth noting, since D1 also carries the site's highest algal cover.

The 28 photographs document a steep, structurally complex outer slope dropping into deep blue water, and they are consistent with the transect picture at every point. **P4250935 opens the dive on the wall itself**: a near-vertical face carrying large gorgonian sea fans and black whip corals, grey tube and barrel sponges, foliose plate colonies and maroon-pink coralline encrustation, worked over by a dense aggregation of dark surgeonfish. Gorgonians recur along the drop-off in **P4250947, P4250953, P4250954 and P4251002**, adding vertical habitat above the plate and massive colonies — and, together with the sponges, they are the visible face of the 20.5% non-coral-invertebrate figure that the transect protocol records but does not resolve to taxon. Away from the wall, the reef is a mosaic: extensive corymbose and digitate *Acropora* fields (P4250952, P4250954), stacked plating and tabular colonies down-slope (P4250947, P4250951, P4251005), large lobate and dome-shaped massive colonies consistent with *Porites* throughout (P4250946, P4250956, P4251002, P4251003), foliose plates, and purple-pink *Pocillopora*-type colonies scattered across the pavement. **P4250955 carries a bright yellow transect tape running diagonally across the frame** — direct documentation of the formal survey in progress. Water clarity is excellent throughout and the reef reads as structurally intact, with high relief and no evidence of physical damage.

**Condition is where this site is in trouble, and the photographic evidence is far more conspicuous than the published figure implies.** The published image bleaching value is **11.9%, which ranks 20th of 98 Chagos sites — the worst quintile in the archipelago**, and the second CoralNet training run agrees closely (11.903%, 30th of 110). **The imagery shows bleaching in a substantial majority of the reviewed frames, and at a visual scale that reads well above 12%.** The clearest single records: **P4250949**, a close view in which two entirely bone-white fine-branched colonies and a white meandroid massive colony sit directly beside a fully pigmented brown lobate colony under identical lighting — which eliminates a strobe or white-balance explanation. **P4250948** shows a broad field in which a large fraction of the massive and plating colonies are white or pale lilac. **P4250954** shows a foreground of pale lilac-white corymbose *Acropora* extending across the frame with a starkly white tabular plate at left-centre, an emperor angelfish passing a gorgonian above it. **P4250956** contains a massive dome colony that is white across one half and normally pigmented across the other — textbook partial bleaching of a single colony. **P4250969, P4250988, P4250990, P4251003, P4251004 and P4251005** all carry extensive white plating and massive colonies, and even the fish frames (P4250972, P4250975, P4250977) show bone-white plate colonies on the reef edge below the schools. Two interpretive cautions apply and are important: **plating *Acropora* is naturally light-toned**, so some of the pale cream plates may simply be lightly pigmented rather than bleached; and **some of the flattest white surfaces are old exposed skeleton**, distinguishable by eroded, sediment-veneered texture. Neither caution accounts for the intact-surfaced, colony-scale white colonies in P4250949, P4250956 and P4251004. **No diagnostic disease band, tissue-sloughing front, *Drupella* aggregation or discrete fresh feeding scar with a visible predator was identified in any of the 28 frames, and no crown-of-thorns starfish was seen.** The whitening is therefore recorded as bleaching or very recent tissue loss, not assigned to disease or predation. Older broken edges and encrusted dead surfaces occur locally but cannot be dated from a single visit.

**The fish assemblage is the other defining feature of the record, and it is exceptional.** A continuous sequence, **P4250971 through P4250978 plus P4250990 and P4250992**, documents a sustained predation event: **schools of bluefin trevally (*Caranx melampygus* — silver flanks stippled with electric-blue spots, blue-violet fins)** hunting through immense, dense schools of small elongate baitfish and fusiliers over the reef edge, with dark surgeonfish and unicornfish moving through the same water. P4250971 is the clearest close view of the trevally; P4250978 is a pure school frame in which the baitfish fill the entire water column. Adjacent frames certainly repeat individuals, so no abundance estimate is inferred, but the repeated frames establish a substantial predator-and-baitfish aggregation rather than a single chance encounter. Separately, **P4251001 and P4251002 record a large grazing aggregation of parrotfish** — thirty or more blue-green individuals with pink-orange cheek and dental markings, moving as a single front across the reef and feeding on the substrate. On a reef carrying 34.9% algae and 19.2% CCA, a grazing school of that size is a functionally significant observation. Orange anthias swarm the reef edge in many frames, and butterflyfish, Moorish idols, powder-blue surgeonfish, damselfish and an emperor angelfish appear throughout.

<b>Assemblage composition.</b> Formal transects give 39.1% coral against 34.9% algae, 20.5% non-coral invertebrates and only 5.5% bare substrate, across **34 non-zero genera — 5th of 109 Chagos sites** — of which *Porites* alone is 49% of the coral. The four-stratum depth profile peaks at D2 (58.6% coral, *Porites* 36.70%) and collapses at D4 (20.25% coral, 40.8% algae, *Acropora* essentially gone at 0.45%), with genus richness highest at D3. All 28 photographs corroborate that structure — a gorgonian- and sponge-clothed outer wall, plating and massive *Porites*-type colonies, corymbose *Acropora* fields — and add two things the transects cannot: **widespread, colony-scale bleaching across most reviewed frames**, and **large aggregations of bluefin trevally hunting baitfish and of parrotfish grazing the reef**.

**Flags:** (1) **BLEACHING — the photographic evidence is more extensive than the published 11.9% figure suggests, and this site belongs in the running COTS/bleaching/disease log for 25 Apr 2015.** Colony-scale whitening appears in a substantial majority of the 28 frames, with unambiguous cases in P4250949 (white colonies beside pigmented neighbours under identical lighting), P4250956 (a massive dome bleached across one half only) and P4251004. The published figure ranks **20th of 98** — already the worst quintile — but the visual impression is of considerably more affected coral. **This site is a good candidate for a targeted re-scoring**, bearing in mind that pale plating *Acropora* and old exposed skeleton can both inflate a visual impression. (2) **The multi-method deviation is −22.0 to −20.8 pp** (39.1% transect vs 18.3% CoralNet vs 17.1% Coralscapes). Do not average them; the in-situ transect value is ground truth for cover. **The CoralNet points figure is a CV product, not verified ground truth** — point-annotation data corpus-wide is roughly 76% unconfirmed machine output. (3) **The previous version of this shard misreported the replication.** It stated that the `N Rows` column "records 2 source rows, so both site-wide formal figures have limited transect replication." The published site table records **N Rows = 4** (four depth strata) and the transect file carries **19 transects** for BISA93. Replication at this site is above average, not limited. Its bleaching-source attribution (28 rows for the second training comparison) matches `n_photos` = 28 and is retained. It also described the site as "coral-rich," which conflicts with the standing terminology rule that "rich" refers only to species richness. (4) **Genus richness is exceptionally high — 34 non-zero genera, 5th of 109** — a genuine site strength distinct from its mid-pack cover, and the reason this site merits attention beyond its bleaching. ***Ctenella*, an Indian Ocean endemic, is present** at 0.04% of the benthos. (5) **D4 is a near-total *Acropora* loss** — 0.45% of the benthos against 8.60% at D1 — on 40.8% algal cover and 17.1% bare substrate. Whether that is a depth limit or a legacy of past mortality cannot be told from a single visit, but the site-level *Acropora* figure of 5.87% must never be quoted without it. (6) **D1 rests on two transects** and carries the site's highest algal cover (47.0%), so the shallow end of the depth story is the least replicated. The site figure moves from 39.06% (unweighted stratum mean, the published value) to 41.28% (transect-weighted) — a 2.2 pp swing. (7) **Non-coral invertebrates are 20.5% of the benthos and are not resolved to taxon** in the published table. The imagery shows abundant gorgonian sea fans, black whip/wire corals and tube and barrel sponges along the drop-off (P4250935, P4250947, P4250953, P4250954, P4251002) — **none of which appear in the transect genus list**, which records scleractinians plus *Millepora* and *Heliopora* only. The octocoral and antipatharian component of this reef is real, visually dominant on the wall, and unquantified. (8) **Five video files in the folder were not reviewed**, and **two of them are 32-byte iCloud placeholder stubs with no content** — `P4250989.MOV` and `P4250994.MOV` — sitting beside full-size duplicates named `P4250989 (1).MOV` and `P4250994 (1).MOV`. The stubs are evicted-file artefacts, not data; they should be replaced or removed rather than treated as missing footage. `P4250970.MOV` is a full 102 MB file, also unreviewed. Note that a bare-stem file listing counts these as five extra "frames" and would report 33 rather than 28 — never print that count beside a CV figure. (9) **Large fauna aggregations recorded:** bluefin trevally (*Caranx melampygus*) hunting baitfish across P4250971–978, P4250990 and P4250992; a 30-plus parrotfish grazing school in P4251001–1002. Species-level calls on the schooling baitfish and the parrotfish are avoided; body markings and fin detail vary too much with angle and distance. (10) **No COTS, disease band or fresh predation scar was identified** across the full 28-frame review — an explicit absence-of-evidence finding, given how much whitening is present and how easily it could be mistaken for disease. (11) **The published bleaching figure is a COMPUTER-VISION product, not an in-situ observation.** It comes from CoralNet annotation of the informal survey images (Mayfield & Dempsey 2026, *Oceans*). Per-image human-vs-model agreement is substantially weaker for bleaching than for cover: **r = 0.40 for bleached against r = 0.72 for live hard coral** (`CHAGOS_MULTIMETHOD_README.md`) — which is exactly the weakness that matters at a site where the human reading of the frames and the model's 11.9% disagree in impression. Note also that the `Mean(%bleaching)` column is **not** a mean — it equals the first CoralNet training run, while the dive-history map publishes the second. At BISA93 the two runs agree (11.90% and 11.903%, ranking 26th and 30th of 110), so the known divergence on bleaching-onset dives does not bite here; but the column carries at least one known bad value elsewhere (BITB58 listed at 22.22% where both runs give ~0.77%), so any rank taken off it is approximate. (12) **NON-CANONICAL BLEACHING RANK — added 2026-08-26.** Any "of 98" bleaching rank above was computed off the site table's `Mean(%bleaching)` column, which equals the **1st** CoralNet training run. Anderson's decision of 2026-08-26 (HOW_TO rule 4e) makes **`% bleaching (2nd training)` canonical**, The file carries TWO pools for it, and they are not interchangeable: a **110-row pool** ranking every dive (revisit dives included, so BIPB70B/72B/69C compete) and a **99-site pool** ranking one value per site. The rank quoted below is on the 110-row dive pool. For this site the canonical value is **11.90%, rank 30 of 110** (1st training: 11.90%, rank 26 of 110). The two runs agree to <0.05 pp on 100 of 110 rows and diverge only on the bleaching-onset dives, so for most sites the distinction changes the rank but not the number.

**Benthic data:** [Download BISA93 benthic survey data (CSV)](../benthic_data/BISA93_benthic.csv) — scleractinian cover by genus with full names, plus substrate and algal categories (bare substrate, CCA, turf, sediment-laden turf, macroalgae, erect algae, cyanobacteria, non-coral invertebrates), from the LOF BIOT-1 survey (Mayfield & Dempsey 2026, *Oceans*, supplemental file S1).