### BITB45 — Three Brothers
**CRD page:** [Three Brothers/Central Chagos Coral Surveys BITB38-BITB49 (Mar 2015)](https://coralreefdiagnostics.com/three-brotherscentral-chagos-20150321-to-0324)

**23 Mar 2015** (date from the published supplement; an earlier review of this site recorded the CRD survey page as saying 22 Mar — see Flags)
**GPS:** 6.05076°S, 71.53503°E
**Setting:** protected submerged bank reef inside the lagoon, 26.4 m
**Coral cover:** 45.6% (15 formal transects, D3–D4; **algal cover 24.7%**) · **published image bleaching (`Mean(%bleaching)` column):** 7.4%
**Live hard coral cover across methods:** formal transects 45.6% · CoralNet points 35.6% · Coralscapes masks 35.9% (13 frames). These products have different sampling footprints and denominators and should not be averaged.
**Temperature:** 29.9 °C · **Salinity:** 34.6
**Photo review:** all 13 of 13 frames examined
**Representative photos:** P3230464.jpg, P3230476.jpg, P3230477.jpg, P3230485.jpg, P3230496.jpg

BITB45 is a protected, submerged bank reef inside the lagoon at 26.4 m, and it is the **least algal site of this Three Brothers group and one of the least algal in the archipelago: 45.6% coral against just 24.7% algae, ranking 35th of 109 Chagos sites for coral cover but 99th of 109 for algal cover** — the bottom decile. The composition of that small algal budget is as striking as its size: fleshy turf (PTTS) is only 3.35% of total benthos, **103rd of 109 sites and among the very lowest turf values recorded anywhere in the dataset**, and crustose coralline algae (PCCATS) is also low at 8.58% (90th of 109). What actually replaces algae here is **sand and bare ground**: bare substrate is 22.5% of total benthos and turf-on-sediment (PTSTS) another 10.7%, the largest single algal sub-category at the site. Non-coral invertebrates add 7.1%, macroalgae 1.5%, and encrusting and cyanobacterial categories are trace (PETS 0.37%, PCYTS 0.25%). The reef therefore reads as **coral thickets and plate stands set on open carbonate sand**, not as a continuous coral-vs-algae pavement — a picture the photographs confirm frame by frame.

Coral cover is concentrated in a small number of taxa: **23 nonzero genera, 64th of 109 Chagos sites for genus count — mid-pack, and far below the 33–34 genera of nearby BITB39 and BITB40** — and the 23 genus percentages sum to 45.64% 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 21.29% (47% of all coral present), then *Porites* 12.20% (27%), *Montipora* 4.95% (11%) and *Pocillopora* 3.28% (7%): **those four genera hold 91% of the coral at this site**, and no fifth genus reaches 1% except *Favites* (0.98%). The tail — *Favia* 0.53%, *Pavona* 0.53%, *Leptastrea* 0.28%, *Cantharellus* 0.26%, *Astreopora* 0.22%, *Cyphastrea* 0.18%, *Goniastrea* 0.17%, *Platygyra* 0.14%, *Psammocora* 0.13%, *Acanthastrea* 0.13%, *Ctenella* 0.08%, *Stylophora* 0.08%, and *Galaxea*, *Hydnophora*, *Lobophyllia*, *Millepora* and *Montastrea* at 0.04% each, with *Symphyllia* at 0.02% — contributes barely 3% of the benthos in total. The formal diversity indices computed over those 23 codes bear this out: **Shannon H′ = 1.56, evenness J′ = 0.50, effective diversity 4.76, Simpson 1−D = 0.69.** An evenness of 0.50 across 23 genera is low, and it is low for a specific reason: this is an *Acropora*-and-*Porites* reef with a long, thin tail behind it.

**The depth structure is the weakest-supported element of this site and must be read with its replication in mind.** There are only two strata, and they are wildly uneven: **D3 rests on 12 transects and D4 on just 3** — yet because the site figure is the unweighted mean of per-stratum means, **those three deep transects carry 50% of the headline 45.6%**. With that caveat stated, D4 (47.07% coral) sits slightly above D3 (44.22%), and the genus story runs the opposite way to most Chagos profiles: ***Acropora* rises sharply with depth, from 15.65% of the benthos at D3 to 26.93% at D4 — 57% of all coral on the deep stratum** — while *Porites* falls from 14.00% to 10.40%, *Montipora* from 5.50% to 4.40% and *Pocillopora* from 4.15% to 2.40%. The substrate flips just as hard: **algae halves from 32.8% at D3 to 16.7% at D4, while bare substrate nearly triples from 12.2% to 32.9%**, and non-coral invertebrates fall from 10.8% to 3.4%. So the deep stratum is an *Acropora*-dominated plate-and-thicket stand standing on almost a third open sand, and the shallow stratum a more mixed, more algal, more invertebrate-bearing ground. **The contrast is descriptive only.** Individual transects span **23% to 69% coral**, and the three D4 transects alone read 69%, 36% and 36.2% — a 33-point internal spread on the stratum that supplies half the site number. Within-patch heterogeneity at this site plainly exceeds the depth difference, and neither the D3–D4 gap nor the *Acropora* gradient should be quoted as a depth effect without that qualification. Note also that **there are no D1 or D2 strata at all**, so the site figure describes only the deeper half of the reef profile.

The photographs — all 13 frames in `BiTb45-finished`, reviewed in full — document a reef built on **broad, tiered, overlapping tabular *Acropora* plates rising above white sand**. P3230464 and P3230468 are the signature frames: layered tables stacked one over another with dark undercut shelter beneath, a large knobby cream-yellow massive colony consistent with *Porites* in the foreground, and clean white carbonate sand lanes on all sides — a direct photographic counterpart to the 22.5% bare-substrate figure. P3230465 looks down on the same architecture from above, showing how far the plate canopy extends. **P3230476 is the essential low-cover counter-frame**: a single golden table standing alone on a broad expanse of open white sand with only scattered small colonies around it, a defensible analogue of the 23% transect. P3230466 and P3230494 are intermediate — pale branching colonies, small tables, massive domes and black-coral bushes scattered across sand — and P3230467 is the most taxonomically varied frame, carrying a large hemispherical brain coral, encrusting and plating colonies, several cream-to-olive massive heads, pink-magenta coralline crust, red encrusting sponge and a tan corymbose *Acropora* in one view. **P3230477 is a full-frame close-up of a single large, perfectly hemispherical meandroid brain colony** with dense, fine, uniformly brown meanders on white sand — consistent with the Chagos-endemic *Ctenella chagius* or with *Platygyra*, both of which appear in the transect data at trace levels (0.08% and 0.14% of benthos); the frame cannot separate them and it is recorded as a hedged identification. Water clarity is very good throughout, and purple anthias (*Pseudanthias*-type) shoal over the plate stands in P3230466 and P3230478.

**Condition signals here are moderate and, importantly, denominator-sensitive.** The published image bleaching figure is **7.4% from the `Mean(%bleaching)` column, ranking 48th of 98 Chagos sites with a value — squarely mid-range, and the highest of the three Three Brothers sites in this rebuild group.** The photographic review is consistent with "some pallor, no demonstrated event." Several frames carry pale branching colonies and white branch tips — **P3230466–467, P3230478, P3230485 and P3230494–495** — and **P3230485, a close overhead view of a large tabular *Acropora*, shows widespread pale mottling across the plate surface together with small discrete white patches and a crisp white growing margin.** A bright white apical margin on an *Acropora* table is normal active growth; the pale surface mottling is not resolvable from a single still, and no frame in the set shows an advancing tissue-loss front with a clean live/dead boundary, so this is recorded as unresolved pallor rather than as bleaching or disease. **P3230495 does show genuine partial mortality**: a large golden table with an irregular dark patch of exposed, algal-colonised skeleton in the middle of an otherwise live plate — old damage with epilithic overgrowth, not an active lesion. **No COTS individuals or feeding scars, no *Drupella* aggregations and no corallivore bite scarring were seen in any of the 13 frames.**

The point-level annotation numbers behind the published bleaching figure deserve to be stated plainly, because they show how thin the evidence is. In the CoralNet point set, **7 points were labelled bleached out of 390 total points — 1.79% of all points, or 4.79% of the 146 coral points** — against a published site figure of 7.4%, while the Coralscapes masks put bleached area at roughly 0.02% of image area. Worse, **two of those seven bleached labels fall inside P3230497, the nudibranch macro**, which is a close-up of an animal on rubble and not a benthic habitat frame at all. Two of seven labels sitting in a non-habitat frame is a real sampling-sensitivity problem for a 7.4% site figure, and it was found by an earlier review of this site; it is preserved and upheld here.

Fauna in the imagery is modest but includes one standout. **P3230496 and P3230497 show a chromodorid nudibranch** — blue-white body with fine black longitudinal stripes, a broad orange-yellow mantle margin and yellow rhinophores and gills — of the *Chromodoris quadricolor* complex; *C. africana* is the Indian Ocean member of that group and is plausible here, but the animal is not resolved to species from photographs and is recorded as unconfirmed. Otherwise: butterflyfish over the plate stands and massive colonies in P3230495, a thicklip-type wrasse in P3230467, purple anthias in P3230466 and P3230478, and damselfish and chromis throughout. A diver holding a survey quadrat frame is visible in midwater at the top of P3230464 — standard field documentation of the transect work, not a data concern.

<b>Assemblage composition.</b> Formal transects give 45.6% coral against 24.7% algae, 22.5% bare substrate and 7.1% non-coral invertebrates, across 23 nonzero genera (64th of 109) in which *Acropora* (21.3%) and *Porites* (12.2%) alone hold 74% of the coral — all as percent of benthos, with H′ = 1.56 and J′ = 0.50 confirming strong dominance. The two-stratum depth summary runs 44.2% coral at D3 (12 transects) to 47.1% at D4 (3 transects), with *Acropora* climbing to 26.9% of the benthos at depth while algae halves and bare substrate nearly triples; because D4 rests on three transects reading 69/36/36.2%, that contrast is descriptive only. The photographs match the data closely — tiered tabular *Acropora* over open white sand (P3230464, P3230468), a lone table on bare sand as the low-cover end member (P3230476), a hemispherical brain colony consistent with the Chagos endemic *Ctenella chagius* (P3230477), unresolved pale mottling on a large table (P3230485), old partial mortality on another (P3230495), and a striped chromodorid nudibranch (P3230496–497). No COTS, disease front or predation scarring was seen in any of the 13 frames.

**Flags:** (1) **D4 rests on only 3 transects but supplies 50% of the headline cover figure**, because the site value is the unweighted mean of per-stratum means. Those three transects read 69%, 36% and 36.2% — a 33-point internal spread. **The 45.6% site figure is materially sensitive to a single deep transect and should be treated as such.** (2) **Only two depth strata exist (D3, D4); there are no D1 or D2 rows**, so the site figure describes the deeper half of the profile only and is not directly comparable with four-stratum sites such as BITB39. (3) **The D3→D4 contrast is descriptive, not a demonstrated depth effect** — transect-level range (23–69%) exceeds the between-stratum difference, and replication is 12 vs 3. (4) **The published 7.4% bleaching figure is not reproduced by the point data.** `chagos_pct_bleached_recomputed.csv` flags BITB45 as a **MISMATCH**: 7 bleached points of 390 recompute to 4.79% of coral points against 7.4% published, a −2.61 pp delta — one of the sites where the recomputation and the published column disagree by more than a point. Per standing rule the published column is quoted verbatim above and is not recomputed here, but **the discrepancy should be resolved centrally.** (5) **Two of the seven bleached point labels fall inside P3230497, the nudibranch macro** — a close-up of an animal on rubble, not a benthic habitat frame. Roughly 29% of the site's bleached evidence therefore comes from a frame that arguably should not be in the scored set. **This is worth acting on: the scored 13-frame set for this site includes two macro frames (P3230496–497) that are not habitat imagery.** (6) **Unresolved pallor, not scored as bleaching:** pale branching colonies and white branch tips in P3230466–467, P3230478, P3230485 and P3230494–495, and widespread pale mottling with small white patches on the large table in P3230485. White apical margins on *Acropora* tables are normal growth; the mottling cannot be separated from bleaching, sediment or recent partial mortality on a single still, and no advancing tissue-loss front is visible in any frame. (7) **Genuine old partial mortality is visible in P3230495** — an irregular patch of exposed, algal-colonised skeleton within an otherwise live tabular *Acropora*. Recorded as old damage, not an active lesion. (8) **No COTS, disease or corallivore predation was observed across the full 13-frame review** — an explicit negative. (9) **Date conflict:** the published supplement gives 23 Mar 2015; an earlier review of this site recorded the CRD survey page as stating 22 Mar. The supplement date is used above; **the one-day disagreement is unresolved and someone should check the page.** (10) **The formal data mix 100-point and 500-point transects** (D3: ten 100-point and two 500-point transects; D4: one 100-point and two 500-point), represented throughout as percentage fields — so transects contribute equally to the stratum mean regardless of point effort. (11) **Genus count is low relative to its neighbours: 23 genera, 64th of 109**, against 34 at BITB39 and 33 at BITB40 surveyed two days earlier. Combined with J′ = 0.50, this is a dominance-structured, not a genus-diverse, site. (12) **The three cover methods agree unusually well here** — transects 45.6%, CoralNet 35.6%, Coralscapes 35.9% (deviation range −10.0 to −9.7 pp, sd 5.7) — the tightest agreement of the three sites in this group. They still must not be averaged; the transect value is ground truth for cover. (13) **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, never as a diver's field estimate. `Mean(%bleaching)` is also **not a mean** — it equals the 1st CoralNet training run, while the dive-history map publishes the 2nd; the two agree here (7.44 both ways, ranks 55th and 56th of 110), but the column carries at least one known bad entry (BITB58 listed at 22.22% against a corrected 0.77%), so the 48th-of-98 rank quoted above is approximate. (14) **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 **7.44%, rank 56 of 110** (1st training: 7.44%, rank 55 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 BITB45 benthic survey data (CSV)](../benthic_data/BITB45_benthic.csv) — scleractinian cover by genus with full names, plus substrate and algal categories (bare substrate, CCA, turf, turf-on-sediment, macroalgae, non-coral invertebrates), from the LOF BIOT-1 survey (Mayfield & Dempsey 2026, *Oceans*, supplemental file S1).