### BISA98 — Salomon
**CRD page:** [Salomon Islands Coral Survey Sites (Apr 26–30, 2015)](https://coralreefdiagnostics.com/salomon-20150426-to-0430)

**26 Apr 2015**
**GPS:** 5.32422°S, 72.21898°E
**Setting:** intermediate-exposure emergent forereef, outside the lagoon, 25.6 m
**Coral cover:** 71.0% (**1 formal transect, D4 only** — see Flags; **algal cover 20.0%**) · **published image bleaching:** the `Mean(%bleaching)` cell is blank for this site — no value exists, and none may be inferred
**Live hard coral cover across methods:** formal transects 71.0% · CoralNet points — none · Coralscapes masks — none. **No imagery exists for BISA98, so neither computer-vision product was ever generated and no multi-method comparison is possible here.** Where those products do exist elsewhere in Chagos they carry different sampling footprints and denominators from the transects and should not be averaged with them; at this site the transect figure stands alone — and it stands on one hundred points.
**Temperature:** 30.7 °C · **Salinity:** 34.3
**Photo review:** **none — no photographs exist for this site anywhere.** **Anderson did not dive here, and the CRD Salomon page records the regret directly: "did not dive, though I wish I did because: coral cover = 71% (VERY HIGH)."** The published site table's `submitted to CoralNet?` column reads **"no photos"** for BISA98, and the supplement's `formal vs. informal comparison` sheet flags it **"did not dive"** — three independent records of the same absence. No folder resolves to site number 98 under any prefix or casing; the site returns zero photo directories and zero frames. **Everything below therefore rests on a single transect's point counts, and reef condition at BISA98 cannot be assessed** — bleaching, tissue loss, disease, predation scarring, COTS, fish and macrofauna are all simply *unavailable*, and must never be read as zero or inferred from neighbouring Salomon sites.

**BISA98 is the highest-cover reef at Salomon and third-highest-cover of the 109 Chagos sites — 3rd of 109 for coral cover — behind only BISP96 (85.98%) and BIPB72 (72.35%).** Salomon averages 43.4% across its fourteen sites; BISA98 is fifteen points clear of the next best, BISA88 at 56.74%, and the only Salomon forereef above 40% where the other comparable forereefs run 24.5–39.1%. Algal cover of 20.0% ranks **108th of 109 — the second-lowest algal cover in the archipelago** — for a coral-to-algae ratio of 3.55. Bare substrate is 9.0% and **non-coral invertebrates read exactly 0.0%**. The note on the CRD page is well founded: on the numbers available, this was the best forereef at Salomon, and it went undived and unphotographed.

Against that, **genus count is 9 — 107th of 109 sites, third-lowest in the archipelago.** All values below are **percent of total benthos, not percent of coral**, and the nine genus columns sum to exactly 71.00%, matching the published cover figure. The reef is *Acropora*-led but not *Acropora*-only: ***Acropora* 38.0% of the benthos (53.5% of live coral)**, ***Montipora* 12.0% (16.9%)**, ***Pachyseris* 7.0% (9.9%)**, *Favia* 4.0% (5.6%), *Porites* 3.0% (4.2%), then *Goniastrea*, *Lobophyllia* and *Montastrea* at 2.0% each and *Astreopora* at 1.0%. H′ = 1.52, Simpson = 0.67, evenness J′ = 0.69, effective diversity 4.58. **The *Montipora* figure is the distinctive one — 12% of the benthos and 16.9% of live coral is a far larger plating/encrusting component than most Chagos sites carry** — and with *Pachyseris* at 7%, close to a quarter of this reef's coral is plate-forming rather than branching or massive. That said, **the richness of 9 is not a trustworthy community result; see Flag 2.**

**No depth profile exists for BISA98, and this is the central limitation of the site.** The survey rests on **one transect in one depth stratum (D4), 100 points** — one of only two single-transect sites in the entire 109-site dataset (the other is BIDI11, also in this pass). The single D4 row *is* the site row: coral 71.0%, algae 20.0%, bare substrate 9.0%, non-coral invertebrates 0.0%, and the same nine genera in the same order. **That one transect is 100% of the site's value**, so there is no standard deviation, no range, no confidence interval, no shallow-to-deep handover and no test of within-site heterogeneity — none of the depth analysis that carries the finding at every other site in this group. Every figure quoted above is one 100-point read, which is why they are all whole integers: the 71.0% headline is literally 71 points out of 100. The site is marked **`Training`** in the published table (Validation = 0).

The algal and substrate breakdown is equally an artefact of that single read, but is internally coherent. Of 20.0% total algae, **crustose coralline algae is 7.0% (96th of 109 sites) and erect algae also 7.0%** — 35% each — with plain turf 5.0% (95th of 109) and macroalgae 1.0%. **Sediment-laden turf and cyanobacteria both read exactly zero**, one of a handful of sites in the survey with no sediment-turf at all, and pavement is 21.0% with **sand, rubble and non-coral invertebrates all reading zero.** Dead coral is 8.0% and recently dead 0% — above the dataset mean of about 4.6% but far below BIPB71's 15.67%. Taken at face value this is a high-cover, low-algae forereef with essentially no sediment signal; taken honestly, it is one hundred points on which several categories happened not to land.

<b>Assemblage composition.</b> A single 100-point transect at D4 gives 71.0% coral against 20.0% algae across nine nonzero genera, led by *Acropora* (38.0% of benthos, 53.5% of coral), *Montipora* (12.0%, 16.9%) and *Pachyseris* (7.0%, 9.9%) — an *Acropora*-dominated forereef with an unusually large plating component, on the second-lowest algal cover in the archipelago and no sediment-turf at all. **The cover figure is 3rd of 109 and the genus count 107th of 109 — the same one hundred points produce both, and the two numbers should not be given equal weight.** There is no photographic record whatsoever, no depth profile, and no replication, so nothing here can be corroborated and **no statement about the reef's condition — bleaching, mortality, disease, predation or fauna — is possible at BISA98.**

**Flags:** (1) **SINGLE TRANSECT, SINGLE DEPTH STRATUM, 100 POINTS — the weakest replication in the dataset alongside BIDI11.** That one transect carries **100% of the site's headline value.** No SD, no range, no depth profile, no confidence interval. **The 71.0% figure is one 100-point sample and must carry that caveat every time it is quoted**, especially as it is cited as a dataset-level extreme (3rd of 109). (2) **THE NINE-GENUS RICHNESS IS ALMOST CERTAINLY A SAMPLING ARTEFACT AND SHOULD NOT BE REPORTED AS A COMMUNITY RESULT.** Recorded richness scales with survey effort across the archipelago, and every other Salomon forereef — sampled 20–40× harder — carries 24–34 genera against an atoll mean near 27. **Nine genera on 100 points at a 71%-cover reef is a sampling floor, not an assemblage.** The central ranks file places BISA98 107th of 109 for genus count; that rank is real but it measures effort as much as ecology. **Anderson's judgement needed on whether to report richness here at all.** (3) **NO PHOTOGRAPHS EXIST AND ANDERSON DID NOT DIVE HERE.** Zero photo directories resolve to site number 98; the site table's `submitted to CoralNet?` cell reads "no photos"; the comparison sheet reads "did not dive". There are therefore no CoralNet points, no Coralscapes masks, no multi-method comparison, and **no basis for any condition assessment** — bleaching, disease, predation, COTS and fauna are unavailable, not zero. (4) **THE PUBLISHED `Mean(%bleaching)` CELL IS BLANK.** BISA98 appears in `formal vs. informal comparison` only as a placeholder row keyed under `site code 2`, carrying no cover, bleaching or image count. **% bleached is unavailable and must not be computed from anything else.** Corpus-wide the published bleaching figure is a **computer-vision product** rather than an in-situ observation, with per-image human-vs-model agreement of r = 0.40 for bleached against r = 0.72 for live hard coral (`CHAGOS_MULTIMETHOD_README.md`) — but here there is no figure at all to caveat. (5) **CORRECTION FOR THE MASTER DOCUMENT: BISA98 is not the highest-cover Chagos site — it is third**, behind BISP96 (85.98%) and BIPB72 (72.35%). **The defensible claim is that two of the six highest-cover sites in the survey went undived and unphotographed (BISA98 3rd, BIPB71 6th)** — still the most striking thing about this group. (6) **NON-CORAL INVERTEBRATES, SAND, RUBBLE, SEDIMENT-LADEN TURF AND CYANOBACTERIA ALL READING EXACTLY ZERO** on a single 100-point transect is a plausible sampling outcome, not necessarily a real absence. Do not cite any of these as confirmed absences. (7) **Integrity checks pass:** the nine genus columns sum to 71.00% against a published 71.0%; the four nonzero algal categories sum to 20.00% against a published 20.0%. (8) Per-site workbook `BiSa98-benthic data.xlsx` is present. Site depth is published as 25.6 m, but **the transect itself is recorded only as stratum D4, with no metre value**, so the 25.6 m cannot be attached to the survey depth with confidence.

**Benthic data:** [Download BISA98 benthic survey data (CSV)](../benthic_data/BISA98_benthic.csv) — scleractinian cover by genus with full names, plus substrate and algal categories (bare substrate, CCA, turf, sediment-laden turf, macroalgae, erect algae, non-coral invertebrates), from the LOF BIOT-1 survey (Mayfield & Dempsey 2026, *Oceans*, supplemental file S1).