### BITB43 — Three Brothers
**CRD page:** [Three Brothers/Central Chagos Coral Surveys BITB38-BITB49 (Mar 2015)](https://coralreefdiagnostics.com/three-brotherscentral-chagos-20150321-to-0324)

**22 Mar 2015**
**GPS:** 6.10527°S, 71.49875°E
**Setting:** intermediate-exposure submerged bank forereef, outside the lagoon, 29 m
**Coral cover:** 28.7% (13 formal transects, D3–D4; **algal cover 26.8%**) · **published image bleaching:** 0.0% (`Mean(%bleaching)` column)
**Live hard coral cover across methods:** formal transects 28.7% · CoralNet points 22.2% · Coralscapes masks 22.9% (17 frames). These products have different sampling footprints and denominators and should not be averaged.
**Temperature:** 29.7 °C · **Salinity:** 34.7
**Photo review:** all 22 of 22 frames in `BiTb43-finished` examined (the folder holds only these 22 jpg stills; no subfolder, no video)
**Representative photos:** P3220398.jpg, P3220413.jpg, P3220408.jpg, P3220430.jpg, P3220387.jpg, P3220425.jpg

BITB43 is a submerged bank forereef off Three Brothers at 29 m under intermediate wave exposure, and its defining feature is not its coral: **non-coral invertebrates cover 32.2% of the benthos — more than the 28.7% coral cover.** The site sits at 80th of 109 Chagos sites for coral cover, well into the lower third, but it is also one of the *least* algae-dominated sites in the archipelago, **95th of 109 for algal cover at 26.8%**, with both algal components low in the corpus context — crustose coralline algae (PCCATS) 9.4%, 82nd of 109, and fleshy turf (PTTS) 6.1%, 88th of 109. Bare substrate is 12.3%. Coral, algae, invertebrates and bare rock sum to exactly 100% of the benthos, so nothing at this site is unaccounted for. What that budget describes is a moderate-cover, low-algae reef whose free space has largely been taken by sponges, tunicates and gorgonians rather than by weed — and the photographs bear that out directly, with large gorgonian sea fans a conspicuous element in nine of the 22 frames (P3220408, P3220409, P3220411–415, P3220425, P3220427–429, P3220431).

Coral cover is spread across **28 nonzero genera, 30th-highest of the 109 Chagos sites**, and the 28 genus percentages sum to 28.72% of the benthos — an exact match to the published cover figure. All values below are percent of total benthos, not percent of coral. The assemblage is unusual for Chagos in being led by two non-*Acropora* genera: *Porites* at 7.33% of the benthos (26% of the coral present) and *Montipora* at 6.18% (22%), which together account for nearly half of all coral at the site. *Pocillopora* follows at 2.79% (10%), and **only then *Acropora*, at 2.46% (9%)** — a strikingly small share for a Chagos forereef. *Favites* (2.04%), *Favia* (1.37%), *Leptastrea* (0.98%), *Stylophora* (0.86%), *Pavona* (0.77%) and *Goniastrea* (0.73%) fill out the upper ranks, followed by a tail of 18 further genera — *Symphyllia*, *Turbinaria*, *Pachyseris*, *Platygyra*, *Cyphastrea*, *Astreopora*, *Echinophyllia*, *Montastrea*, *Acanthastrea*, *Ctenella*, *Lobophyllia*, *Oxypora*, *Fungia*, *Gardineroseris*, *Leptoseris*, *Psammocora*, *Coscinaraea* and *Cantharellus* — each under 0.6% of the benthos.

**The two depth transects record an *Acropora* collapse, not a cover collapse.** Coral cover barely moves between strata — 30.4% at D3 (10 transects) and 27.1% at D4 (3 transects) — and algal cover is flat as well (26.2% and 27.3%). What changes is who holds the reef. At D3 the leaders are *Montipora* 6.50% of the benthos (21% of coral there), *Porites* 5.72%, ***Acropora* 4.72%** (16%) and *Pocillopora* 3.98%, across 25 genera. At D4, **the deeper station, *Acropora* has fallen to 0.20% — a twenty-four-fold drop and just 0.7% of the coral present** — while *Porites* climbs to 8.93% (33% of coral), *Montipora* holds at 5.87%, and *Favites* more than doubles to 2.80%; *Pocillopora* and *Stylophora* both fall away (3.98→1.60% and 1.66→0.07%). Genus count drops with it, 25 to 21. The substrate budget shifts in the same direction: **bare substrate triples from 5.9% at D3 to 18.7% at D4**, while non-coral invertebrates fall from 37.5% to 26.9%. The D3 figure is the striking one — **at the shallower station, invertebrate cover (37.5%) exceeds coral cover (30.4%)**. The site-level *Acropora* value of 2.46% describes no transect at either depth; it is the midpoint of a genus that is a real, if modest, component shallow and effectively absent deep. Replication is uneven (10 transects at D3 against 3 at D4), so the deep numbers are the weaker half of the comparison, and individual transects span 17–52% coral — within-depth patchiness is larger than the 3.3-point difference between the depth means.

The 22 photographs document exactly that patchiness. Broad, low-relief sand and rubble plains with only scattered small colonies (P3220386, P3220387, P3220409, P3220410) alternate with dense, high-relief gardens (P3220398, P3220399, P3220411–415, P3220425, P3220431, P3220432) where coral cover in-frame approaches continuous. Growth-form composition in the imagery matches the transect ranking rather than the Chagos norm: **foliose, plate and encrusting colonies dominate almost every high-cover frame**, with large ridged and vase-like plate stands consistent with *Pachyseris*, *Merulina*, *Echinopora* or *Montipora* forming the single biggest structural elements at P3220398 and P3220413, and broad mustard-green encrusting sheets covering the substrate at P3220432. Massive and sub-massive boulder colonies consistent with *Porites* and faviid genera are frequent, brain-patterned colonies appear at P3220414, and branching/digitate colonies are present but never dominant — consistent with *Acropora* holding only 9% of the coral here. Purple-pink crustose coralline algae is visible on dead framework in most close frames. Water clarity varies within the set: P3220398, P3220399 and P3220413–415 are clear with deep-blue water columns, while P3220386, P3220409, P3220427 and P3220429 carry visible sediment haze.

**Condition is mostly good, but the published 0.0% bleaching figure is contradicted by the photographs.** The `Mean(%bleaching)` column records **0.0% for BITB43 — 97th of 98 Chagos sites with a value, effectively the archipelago floor** — and both CoralNet training runs agree at 0.0 (108th of 110), so this is not a run-selection artifact. The imagery nonetheless carries three unambiguous instances of white coral tissue. **P3220430 is the clearest: a brown, verrucose pocilloporid-type colony with a sharply demarcated bone-white sector across its upper branches, tissue structure still intact** — the classic appearance of partial bleaching rather than old dead skeleton, with several smaller pale colonies beside it. P3220424, a macro of a small branching colony on sand, shows bone-white branch tips against otherwise tan-pigmented tissue at the point of contact with adjacent rock; bleaching, contact aggression and recent tissue loss cannot be separated from a single still. P3220387 shows an isolated bommie carrying a branching colony whose upper and lower-right sectors are fully white while its centre remains pigmented. A pale cream plate colony at the left of P3220415 is ambiguous and is not counted. None of this amounts to a coherent reef-wide bleaching field — the overwhelming majority of colonies in all 22 frames are normally pigmented — but a site-level figure of exactly zero is not what the photographs show, and the discrepancy is flagged below. **No COTS individual, feeding scar, disease lesion, tissue-loss margin, *Drupella* aggregation or corallivore bite scarring was seen anywhere in the 22-frame set**; a prior review noted a 0.18% COTS mask signal in P3220408 and judged it false, and the present frame-by-frame pass agrees — P3220408 shows gorgonians and foliose plates, no acanthasterid.

Fauna is abundant and includes several large-bodied animals. **Grey reef sharks (*Carcharhinus amblyrhynchos*-type, identification kept conservative) cross P3220427, P3220428 and P3220429** in three separate frames over the deeper, hazier ground. A large dark grouper with pale blotching, attended by a cleaner wrasse, is photographed twice on the open pavement (P3220409 resting, P3220410 cruising). P3220425 and P3220426 document a dense mixed aggregation of pennant bannerfish (*Heniochus*-type) with moorish idols and butterflyfish over a low bommie. Dark schooling surgeonfish/unicornfish form a curtain over the reef in P3220411 and P3220412, damselfish and chromis are near-constant, and humbug damsels shelter in bommies at P3220428 and P3220431. P3220397 is a macro of small solitary/free-living polyps with green-brown oral discs on sand beside encrusting *Porites*, tunicates and pink CCA — consistent with *Fungia*/*Cantharellus*-type juveniles or corallimorphs and left hedged; a small red spot in the frame is a camera scaling laser, not biota. A white metal camera/survey rig and diver are visible at the top of P3220410 — standard field documentation, not a data concern.

<b>Assemblage composition.</b> Formal transects give 28.7% coral, 26.8% algae, 12.3% bare substrate and 32.2% non-coral invertebrates across 28 nonzero genera, led by *Porites* (7.3%) and *Montipora* (6.2%) with *Acropora* a distant fourth (2.5%). The two-stratum depth profile holds cover almost flat (30.4% at D3, 27.1% at D4) while *Acropora* collapses from 4.72% to 0.20% and *Porites* rises to a third of all deep coral, and bare substrate triples with depth. The photographs corroborate a foliose-and-encrusting reef of alternating sand plains and dense plate gardens, with gorgonian sea fans as a signature element, three grey reef shark encounters, and localized but genuine white-tissue coral in three frames against a published bleaching figure of zero.

**Flags:** (1) **The published bleaching figure of 0.0% conflicts with the photographs.** `Mean(%bleaching)` records 0.0% and both CoralNet training runs agree, yet P3220430 shows a sharply demarcated bone-white sector on an otherwise pigmented pocilloporid colony, with corroborating white tissue in P3220424 and P3220387. Someone should decide whether the scored 17-frame subset simply never sampled these colonies or whether the annotation missed them. (2) **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 far weaker for bleaching than for cover: r = 0.40 for bleached against r = 0.72 for live hard coral (`CHAGOS_MULTIMETHOD_README.md`). Treat it as the weaker of the two published numbers and never as a diver's field estimate. (3) **Three different frame denominators are in play at this site.** The photo folder holds 22 frames; `chagos_multimethod_cover_by_site.csv` scores `n_photos=17`; `chagos_pct_bleached_recomputed.csv` used 12 images / 360 points. Never print the 22-frame folder count beside a CoralNet or Coralscapes figure. (4) **Filename collision with BITB42.** Frames P3220426, P3220427, P3220428, P3220429 and P3220430 exist in *both* `BiTb43-finished` and `BiTb42-finished` at different byte sizes (two cameras on 22 Mar). Never key on the bare stem — always carry the folder path. (5) **Multi-method agreement is unusually tight here** (transects 28.7%, CoralNet 22.2%, Coralscapes 22.9%; deviation −6.5 to −5.8 pp, `sd_all` 3.6) against a corpus mean between-method SD of 10.2 pp. That is consistent with the documented pattern that the photo methods track transects best at mid cover and diverge most at 40%+ cover. Still do not average them; the transect value is ground truth for cover. (6) **`coralnet_points_pct` is a human 30-point read, not a model output** (`CHAGOS_MULTIMETHOD_README.md`) — do not describe it as a CV product. (7) **Depth replication is uneven:** 10 transects at D3 against 3 at D4, so the D4 stratum carries a third of the site headline on three transects. Site cover is the unweighted mean of the two stratum means. (8) **Formal data mix 100- and 500-point transects** (10 at 100 points, 3 at 500) and are represented here by percentage fields only. (9) **No D1 or D2 stratum exists at this site** and no per-frame depths are recorded, so no photograph can be tied to a depth transect. (10) **Non-coral invertebrates (32.2%) exceed coral cover (28.7%) at site level, and at D3 do so by seven points (37.5% vs 30.4%)** — a real structural finding, not a data artifact, and consistent with the sea fans and sponges throughout the imagery. (11) **No COTS, disease or predation scarring was observed** across the full 22-frame review — an explicit absence-of-evidence finding. The 0.18% COTS mask signal previously reported in P3220408 is confirmed false. (12) **Fish identifications are kept conservative:** the sharks in P3220427–429 are grey-reef-type, the grouper in P3220409–410 and the bannerfish aggregation in P3220425–426 are identified to type only, and the solitary polyps in P3220397 are not resolved to genus. (13) **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 **0.00%, rank 108 of 110** (1st training: 0.00%, rank 108 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 BITB43 benthic survey data (CSV)](../benthic_data/BITB43_benthic.csv) — scleractinian cover by genus with full names, plus substrate and algal categories (bare substrate, CCA, turf, turf+sediment, macroalgae, erect algae, cyanobacteria, invertebrates), from the LOF BIOT-1 survey (Mayfield & Dempsey 2026, *Oceans*, supplemental file S1).