### BISA84 — Salomon
**CRD page:** [Salomon Islands Coral Survey April 2015](https://coralreefdiagnostics.com/salomon-201504)

**22 April 2015**
**GPS:** 5.31730°S, 72.22408°E
**Setting:** windward emergent fringing forereef, 30.0 m
**Coral cover:** 48.8% (3 depth transects, D2–D4; **algal cover 41.0%**) · **published image bleaching:** 34.0%
**Live hard coral cover across methods:** formal transects 48.9% · CoralNet points 11.0% · Coralscapes masks 16.7% (10 frames). These products have different sampling footprints and denominators and should not be averaged.
**Temperature:** 29.7 °C · **Salinity:** 33.9
**Photo review:** all 10 stills examined
**Representative photos:** P4220837.jpg, P4220838.jpg, P4220841.jpg, P4220869.jpg, P4220890.jpg

BISA84 is an emergent fringing forereef on the outside of Salomon atoll, and it carries **above-average coral cover for the archipelago — 48.8% coral against 41.0% algae, ranking 28th of 109 Chagos sites for coral cover and mid-pack (52nd of 109) for algal cover.** Crustose coralline algae (PCCATS) accounts for 16.5% of total benthos (40th of 109) against 14.1% fleshy turf (PTTS, 21st of 109), with bare substrate at 5.9%, PTSTS 5.8%, PITS 4.3%, PMTS 2.6% and PETS 1.8%. **The site's defining number is not its cover but its bleaching: at 34.0% it is the highest value in the `Mean(%bleaching)` column, 1st of 98** — though that ranking depends on which CoralNet training run is used, and BIPB112 exceeds it under the other (see Flags).

Coral cover is concentrated in remarkably few taxa: **25 nonzero genera (55th of 109 for genus richness — middling), but the two leading genera alone hold three-quarters of all coral present.** The 25 genus percentages sum to 48.8484% of the benthos, an exact match to the published cover figure. All values below are percent of total benthos, not percent of coral. *Acropora* leads at 24.41% of the benthos — **50% of all the coral at the site** — followed by *Porites* at 12.30% (25%), then a steep drop to *Montipora* 2.55% (5%), *Pocillopora* 1.60% (3%) and *Goniastrea* 1.35% (3%); those five account for 86% of the coral recorded. *Pachyseris* (1.08%), *Favia* (0.88%), *Seriatopora* (0.75%), *Psammocora* (0.73%) and *Isopora* (0.69%) follow, then a tail of 15 further genera — *Fungia*, *Favites*, *Platygyra*, *Pavona*, *Echinopora*, *Gardineroseris*, *Cyphastrea*, *Montastrea*, *Leptastrea*, *Acanthastrea*, *Lobophyllia*, *Stylophora*, *Symphyllia*, *Astreopora* and *Galaxea* — each under 0.35% of the benthos. **An assemblage this concentrated in *Acropora* is the relevant context for the site's archipelago-high bleaching figure**, *Acropora* being among the most thermally sensitive of the genera present; the transect and bleaching data are independent measurements and this shard does not claim a tested causal link.

**The three depth transects show coral cover nearly doubling from the shallowest station to mid-depth, driven almost entirely by *Acropora*, then handing over to *Porites* at depth.** D2, the shallowest station sampled, carries 34.6% coral against the profile's highest algal cover, 56.9%; *Acropora* is already dominant at 17.37% of the benthos (50% of that transect's coral), ahead of *Porites* 7.70%, *Montipora* 2.00%, *Psammocora* 1.97%, *Pocillopora* 1.67% and *Favia* 1.43%. At D3 cover jumps to 57.2% and algae falls to 37.5%, and **the transect becomes overwhelmingly monogeneric: *Acropora* alone reaches 40.20% of the benthos — 70% of all the coral on that transect** — with *Porites* 6.94%, *Montipora* 3.91%, *Pocillopora* 2.03%, *Isopora* 1.06% and *Platygyra* 0.71% trailing far behind. At D4 cover stays high at 54.8% and algae drops further to 28.5%, but **the lead changes hands: *Porites* takes over at 22.25%** (41% of that transect's coral) ahead of *Acropora* 15.65% (29%), with *Goniastrea* 3.75%, *Pachyseris* 3.25%, *Seriatopora* 2.25% and *Montipora* 1.75%. **The site-level figures of 24.41% *Acropora* and 12.30% *Porites* describe no single transect**: they average a shallow station at half that *Acropora* value, a mid station at nearly double it, and a deep station where *Porites* leads instead.

The photographs — all 10 stills in the folder — document a table- and branch-dominated windward reef consistent with the *Acropora* signal in the transects. P4220837 and P4220839 show broad overlapping plates and tables above encrusted framework, while P4220838 records a survey quadrat among the same habitat — standard field documentation rather than a data concern. P4220854 opens onto a sand-and-rubble patch with a discrete coral head, demonstrating local breaks in the otherwise coral-abundant architecture, and P4220842 shows more of the same open ground where crust-covered framework, turf-coated surfaces, rubble and sand are most visible. The final frames return to mixed plates, massive colonies and branches. The visual dominance of tables and branching colonies is compatible with the strong *Acropora* signal, but image-based genus calls remain tentative.

**Condition is the striking finding at this site, and the photographs corroborate it directly.** The published image-based bleaching figure is **34.0% — the highest of the 98 Chagos sites under the `Mean(%bleaching)` column, ranking 1st of 98 — and bleaching is unmistakable in the stills rather than marginal. P4220841 is a close view of a compact branching colony that is uniformly white against pigmented and crust-covered surroundings. P4220869 and P4220870 show white branching colonies beside a darker massive colony, providing strong within-frame contrast that exposure or flash artefact cannot explain. P4220890 contains additional pale plates and branches. These observations support the published high bleaching signal but do not generate a new site percentage, and they were logged as frame-specific bleaching records rather than as a recalculated figure. No crown-of-thorns starfish or diagnostic disease sign was identified in the 10 stills; the white colonies are recorded as bleaching rather than disease because no characteristic lesion margin or tissue-loss band is visible.

Small reef fishes occur around the tables and branching colonies, but the set is not standardized for fish abundance and no unusual large-animal event is inferred.

<b>Assemblage composition.</b> Formal transects give 48.8% coral against 41.0% algae across 25 nonzero genera, but the assemblage is unusually concentrated: *Acropora* (24.4%) and *Porites* (12.3%) together hold three-quarters of all coral, with *Montipora* (2.6%), *Pocillopora* (1.6%) and *Goniastrea* (1.3%) far behind. The depth profile shows cover rising from 34.6% at D2 to 57.2% at D3 — where *Acropora* alone reaches 70% of the coral present — then 54.8% at D4 where *Porites* replaces *Acropora* as the leading genus. The photographs document the matching table- and branch-dominated architecture with local sand-and-rubble breaks, and provide direct frame-level evidence of the bleaching behind this site's archipelago-high published figure of 34.0%.

**Flags:** (1) **Exposure classification conflicts between sources.** `BIOT-1-MASTER-WkBk.xlsx`, sheet `Master`, records this site as **windward**; `CHAGOS_SITE_REFERENCE_20260825.csv` gives `exposure_published` = **intermediate**. The master workbook is taken as authoritative here and the header reads windward, but the conflict is unresolved and both values are recorded. (2) **Multi-method deviation is the largest in this batch.** CoralNet points (11.0%) and Coralscapes masks (16.7%) sit 32–38 percentage points below the formal-transect figure (48.9%, deviation range −37.9 to −32.2 pp). Do not average the three figures; the transect value is treated as ground truth for cover. (3) **The CV figures rest on a 10-frame scored subset** which happens to equal the folder count here — that coincidence does not make them comparable, and the denominators still differ. (4) **The CoralNet points figure is a CV product, not verified ground truth**; point-annotation data corpus-wide is roughly 76% unconfirmed machine output. (5) **"Most bleached in the archipelago" is column-dependent, and this site is 1st under only one of the two published columns.** Supplemental S1 carries two CoralNet runs. They agree to within 0.05 pp on 100 of 110 rows, and they agree exactly here (34.03 both ways) — but they diverge sharply on the bleaching-onset dives, always with the second run higher. Under `% bleaching (2nd training)` **BIPB112 reaches 38.22% and takes 1st place**, with BIPB72B (37.46) and BIPB69C (36.47) also above this site; BISA84 falls to 4th. The site table's `Mean(%bleaching)` column is **not a mean** — it equals the 1st-training value — while the dive-history map publishes the 2nd. Always name the column. Two entries in that column match neither run: BITB58 (22.22 against 0.77 both ways — a data-entry error) and BIPB61 (12.86 against 12.0 both ways). The value itself is reported as published and was not recalculated from these photographs or from CoralNet labels. (6) **Formal replication is thin: `N Rows` = 3** for the site-level figures. The `N Rows` column of the published `formal vs. informal comparison` sheet records **7** rows for the bleaching comparison — a different and smaller basis than the 10 website stills reviewed here. (7) **No D1 transect exists for this site** — the profile starts at D2, so the shallowest reef zone is unsampled. (8) **The depth codes D2–D4 carry no metre boundaries** in `Chagos benthic data by site-depth (n=336 rows).csv`; they are ordered shallow-to-deep stations, and the 30.0 m in the header is the site-level survey depth, not the depth of any one transect. (9) **No COTS, disease, or predation scarring was observed** across the full 10-frame review — an explicit absence-of-evidence finding. (10) **A survey quadrat and field gear are visible in P4220838** — standard documentation, not a data issue. (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`). Treat the site bleaching percentage as the weaker of the two published numbers, and never read it as a diver's field estimate. (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 **34.03%, rank 4 of 110** (1st training: 34.03%, rank 1 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 BISA84 benthic survey data (CSV)](../benthic_data/BISA84_benthic.csv) — scleractinian cover by genus with full names, plus substrate and algal categories, from the LOF BIOT-1 survey (Mayfield & Dempsey 2026, *Oceans*, supplemental file S1).