### BISP100 — Speaker's Bank
**CRD page:** [Speaker's Bank Coral Survey Sites](https://coralreefdiagnostics.com/speakers-bank)

**28 Apr 2015**
**GPS:** 4.78432°S, 72.34454°E
**Setting:** intermediate-exposure submerged forereef, outside the lagoon, 24.0 m (the detailed reef-type and original-exposure cells are **blank** in the published table — see Flag 4)
**Coral cover:** 53.9% (13 transects across 3 depth strata, D3–D5; **algal cover 24.5%**) · **published image bleaching:** 4.8% (`Mean(%bleaching)` column)
**Live hard coral cover across methods:** formal transects 53.9% · CoralNet points 16.7% · Coralscapes masks 18.9% (n_photos = 4). These products have different sampling footprints and denominators and should not be averaged.
**Temperature:** 30.0 °C · **Salinity:** 34.3
**Photo review:** all 7 of 7 frames examined
**Representative photos:** P4280852.jpg, P4280853.jpg, P4280880.jpg, P4280869.jpg, P4280879.jpg

BISP100 is a submerged forereef on the outer edge of Speaker's Bank, the northernmost of the Chagos survey areas, and on the published numbers it is a high-cover, low-algae site: **53.9% coral against 24.5% algae, ranking 19th of 109 Chagos sites for coral cover and 100th of 109 for algal cover** — that is, among the ten least algae-dominated reefs in the archipelago. The algal budget is unusual in its composition as well as its size. Crustose coralline algae hold only 5.8% of the benthos, **99th of 109 sites — close to the lowest CCA share recorded anywhere in the dataset** — while fleshy turf at 11.8% (40th of 109) is the largest single algal category, backed by sediment-laden turf 4.4%, macroalgae 2.0%, erect algae 0.5% and cyanobacteria 0.1%. What replaces the coralline pavement is bare substrate: **PB = 19.4% of the benthos**, and the photographs make that figure the most visible thing about the site. Non-coral invertebrates are minor at 2.3%. Coral, algae, bare substrate and invertebrates together account for 100.0% of the benthos, so nothing is unallocated at site level.

Coral cover here is concentrated rather than spread. Only **16 non-zero genera were recorded, 93rd of 109 Chagos sites for genus count** — a low-diversity assemblage by archipelago standards, and a striking contrast with its own high cover. All values below are **percent of total benthos, not percent of coral**, and the 16 genus percentages sum to 53.88%, an exact match to the published cover figure. ***Acropora* alone holds 41.44% of the benthos — 77% of every coral point recorded at the site.** Everything else is a remainder: *Porites* 3.45%, *Montipora* 2.81%, *Favia* 1.33%, *Pocillopora* 1.21%, *Astreopora* 0.95%, *Leptastrea* 0.68%, *Goniastrea* 0.58% and *Pavona* 0.55%, then a seven-genus tail below 0.4% — *Psammocora*, *Montastrea*, *Favites*, *Hydnophora*, *Platygyra*, *Acanthastrea* and *Cyphastrea*. This is as close to a single-genus reef as a 50%-cover site gets in this dataset.

**The depth profile is the real finding here, and it is also the reason the headline number is fragile.** Coral cover doubles from the shallowest surveyed stratum to the deepest while the assemblage collapses to three genera. **D3 carries 36.7% coral against 36.4% algae and 22.1% bare substrate, across 15 genera** — a genuinely mixed band in which *Acropora* leads at 17.33% but *Montipora* (4.44%), *Porites* (4.35%), *Astreopora* (2.85%), *Leptastrea* (2.04%) and *Pocillopora* (1.64%) all hold real ground. **D4 rises to 53.0% coral on 30.0% algae across only six genera, with *Acropora* at 43.00% — 81% of the coral on that transect** — trailed by *Favia* 4.00%, *Pocillopora* 2.00%, *Porites* 2.00%, *Goniastrea* 1.00% and *Pavona* 1.00%. **D5 reaches 72.0% coral against just 7.0% algae and zero recorded non-coral invertebrates, across three genera only: *Acropora* 64.00%, *Montipora* 4.00%, *Porites* 4.00% — 89% of the deepest transect's coral is *Acropora*.** The site-level figure of 41.44% *Acropora* therefore averages a moderately diverse mid-depth band onto what is effectively a deep *Acropora* monoculture, and should never be quoted alone.

**BISP100 is the most weighting-sensitive site in the archipelago and this must travel with the number.** Site cover is the unweighted mean of the three per-stratum means — (36.65 + 53.0 + 72.0) / 3 = **53.88%** — but the three strata are not remotely equally sampled: **D3 rests on 11 transects, D4 on one, and D5 on one.** Re-computed as a transect-weighted mean, the same 13 transects give **40.63%**, a swing of 13.3 percentage points, and the two single-transect strata that carry two-thirds of the published figure between them contribute 2 of 13 transects to it. No other Chagos site moves this far under a change of weighting. The 53.9% value and its 19th-of-109 rank are the published, citable figures and are used throughout this shard, but any comparison of BISP100 against other sites should carry the caveat that its deep, high-cover, *Acropora*-dominated end rests on two transects.

The seven photographs document a reef whose structure alternates between locally dense coral and wide open pavement, which is exactly what a 53.9% site figure with a 19.4% bare-substrate figure should look like. **P4280852 is the highest-relief frame**: overlapping tabular and digitate *Acropora* colonies set at several heights over broken pavement, producing shaded cavities, with a fluted giant clam (*Tridacna* sp.) embedded in the rock left of centre. **P4280880** shows the same architecture at wider scale — tabular plates and corymbose colonies spaced by pale pavement — and carries a diver and a laid transect line in the upper-left background, direct documentation of the formal survey in progress. **P4280869** shows two large cream-coloured *Acropora* tables standing on an otherwise open pavement plain scattered with small encrusting and foliose colonies, and **P4280879 is the most open frame in the set**: a broad pale pavement-and-rubble plain carrying only sparse small colonies and a single rounded massive bommie. P4280860 sits between the two extremes, with well-spaced tables over clean pavement. P4280839 shows mixed rubble ground with scattered corymbose *Acropora*, one large rounded massive colony consistent with *Porites*, and an olive-green turf band along the mid-ground ridge — a photographic counterpart to the 11.8% turf figure. **The retained sequence supports a patchy rather than a uniformly continuous canopy**, and reinforces the point that no single frame should be attached to the 53.9% site figure.

**Coral condition reads as broadly pigmented with one clear exception.** The published image-based bleaching figure is **4.8%, 72nd of 98 Chagos sites with a value — in the low-bleaching quartile** — and the two CoralNet training runs agree exactly here (both 4.76%, 79th of 110), so this is one of the more trustworthy bleaching numbers in the set. Against that, **P4280852 contains a large, bone-white branching *Acropora* colony occupying the lower-right quarter of the frame**: whitening is colony-scale rather than tip-restricted, the branch surfaces appear intact rather than eroded, and a normally pigmented tan digitate colony sits immediately above it in the same lighting — which argues against a strobe or white-balance artefact and makes recent bleaching or very recent tissue loss the more likely reading, at medium confidence from a single still. **P4280853** shows two corymbose *Acropora* colonies with numerous white branch tips; most of that is consistent with normal growth margins, but several larger white patches mid-colony are broader than growth tips and are noted rather than resolved. An interpretive caution applies to the open frames: **much of the pale ground at this site is old skeletal material, rubble and sediment rather than bleached tissue**, distinguishable by its irregular eroded texture rather than an intact coral surface. No photographic cover or bleaching estimate was made from these contrasts.

Fauna is modest but consistent. Humbug damselfish (*Dascyllus* sp.) are the most numerous fish in the set, sheltering in corymbose *Acropora* thickets in P4280839, P4280853 and P4280880 — a dozen or more per frame. Parrotfish, surgeonfish and blue chromis appear in the mid-field of P4280860 and P4280880, and a goatfish or wrasse crosses the right edge of P4280880. **The large echinoderm at the bottom centre of P4280853 — a broad, dark, reticulate net of black lines over tan cushioned plates — is a cushion star (*Culcita* sp.), not a holothurian**; this shard corrects the earlier reading of that animal as a sea cucumber. *Culcita* is a facultative corallivore that feeds on small branching corals, so the record is worth carrying, but it is a single individual with no associated feeding scar visible and is not a crown-of-thorns starfish. No COTS, no discrete fresh feeding scar, no advancing disease band and no tissue-sloughing front was identified in any of the seven frames.

<b>Assemblage composition.</b> Formal transects give 53.9% coral against 24.5% algae, 19.4% bare substrate and 2.3% non-coral invertebrates, across only 16 non-zero genera (93rd of 109) of which ***Acropora* is 77%**. The three-stratum depth profile runs from a mixed 36.7%, 15-genus band at D3 to a 72.0%, three-genus, 89%-*Acropora* deep transect at D5, and the site figure is the unweighted mean of strata sampled 11 : 1 : 1. The seven photographs document that patchiness directly — dense overlapping *Acropora* tables in some frames, wide bare pavement plains in others — plus a giant clam, a cushion star, abundant humbug damselfish, a diver-and-transect-line survey frame, and one colony-scale white *Acropora* colony that is the site's only clear bleaching candidate.

**Flags:** (1) **BISP100 is the most weighting-sensitive site in the archipelago.** The published 53.88% is the unweighted mean of three per-stratum means, but the strata carry 11, 1 and 1 transects respectively; a transect-weighted mean of the same 13 transects gives **40.63%, a 13.3 pp swing**. The two single-transect strata (D4 53.0%, D5 72.0%) supply two-thirds of the headline figure from 2 of 13 transects. Any cross-site comparison, and the 19th-of-109 cover rank, should carry this caveat. (2) **The multi-method spread is roughly 37 percentage points** (53.9% transect vs 16.7% CoralNet vs 18.9% Coralscapes, deviation range −37.2 to −35.0 pp) — among the widest in the dataset — **and the two photographic products rest on `n_photos` = 4**, against 7 frames in the folder. That is the thinnest photographic basis of any site reviewed in this batch; treat the CoralNet and Coralscapes values as weakly supported and do not pool them with the transect figure. (3) **Seven frames is a very short set.** Each frame carries substantial interpretive weight and none is a random areal sample; the photo-based statements above are qualitative only. (4) **Both the detailed reef-type and the original-exposure cells are blank** in the published 109-site table. The Setting line is built from `reef location` = forereef, `lagoon` = outside, `reef emergence` = submerged and `exposure-revised` = intermediate. **No narrower reef class has been invented.** (5) **D5 rests on a single three-genus transect and D4 on a single six-genus transect.** Do not read the site genus list as uniform with depth, and do not treat the 72.0% deep-cover figure as replicated. (6) **P4280852 carries a large bone-white branching *Acropora* colony** — colony-scale, intact-surfaced, adjacent to normally pigmented coral under the same lighting. Recent bleaching or very recent tissue loss at medium confidence; not resolvable from one still. (7) **A cushion star (*Culcita* sp.) is present in P4280853** — a facultative corallivore, single individual, no feeding scar visible. This corrects an earlier description of the animal as a sea cucumber. **No crown-of-thorns starfish, disease band or fresh predation scar was identified in any of the seven frames** — an explicit absence-of-evidence finding across a small set, not a clean bill of health. (8) **A giant clam (*Tridacna* sp.) appears in P4280852** and gorgonian/octocoral cover is not itemised by the transect protocol, which records scleractinians plus *Millepora* and *Heliopora* only. (9) **Crustose coralline algae are near the archipelago minimum here (5.8%, 99th of 109)** while bare substrate is high (19.4%) — an unusual combination for a high-cover site, and worth checking against a second method. (10) **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. At this site the two CoralNet training runs agree exactly (both 4.76%), which is the favourable case; note also that the `Mean(%bleaching)` column is **not** a mean — it equals the first training run — and that it carries at least one known bad value elsewhere (BITB58 listed at 22.22% where both runs give ~0.77%), so any rank taken off that column is approximate. (11) **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 **4.76%, rank 79 of 110** (1st training: 4.76%, rank 79 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 BISP100 benthic survey data (CSV)](../benthic_data/BISP100_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).