### BITB56 — Three Brothers
**CRD page:** [Three Brothers Coral Surveys](https://coralreefdiagnostics.com/three-brotherscentral-chagos-20150325-to-0328)

**27 Mar 2015**
**GPS:** 6.17943°S, 71.56844°E
**Setting:** protected submerged bank reef (lagoon, leeward), depth not reported (3 formal transects, D2–D4)
**Coral cover:** 21.3% (3 formal transects, D2–D4; **algal cover 48.9%**) · **published image bleaching:** 16.0%
**Live hard coral cover across methods:** formal transects 21.3% · CoralNet points 15.6% · Coralscapes masks 14.5% (13 frames). These products have different sampling footprints and denominators and should not be averaged.
**Temperature:** 29.8 °C · **Salinity:** 34.2
**Photo review:** all 13 frames examined
**Representative photos:** P3270386.jpg, P3270401.jpg, P3270419.jpg, P3270415.jpg, P3270394.jpg

BITB56 is an algae-dominated reef: at 21.3% transect coral cover against 48.9% algae, non-coral cover here is more than double the coral itself. Substrate scoring backs this up — bare pavement is 21.8% of the benthos and crustose coralline algae (CCA) 17.6%, with turf-plus-sediment a further 16.7%, erect algae 7.5% and macroalgae 3.0% (these four algal categories sum to exactly the published 48.9%). Non-coral invertebrate cover is 8.0%. This is open, low-relief hardground more than a continuous coral framework.

The genus list is long — 22 nonzero genera — but heavily lopsided. **Porites alone holds 9.78% of the benthos, 46% of all recorded coral**, roughly three-and-a-half times the second-place genus. Acropora is next at 2.85%, then Montipora 1.60%, Pocillopora 1.59%, Favites 1.43% and Favia 0.92%. A long tail of sixteen further genera — Platygyra, Echinopora, Leptastrea, Stylophora, Astreopora, Goniastrea, Acanthastrea, Montastrea, Goniopora, Hydnophora, Pavona, Millepora, Echinophyllia, Psammocora, Coscinaraea and Symphyllia — each contribute under 0.6%, most under 0.2%. The genus values sum to 21.29%, matching the published 21.3% cover exactly, so nothing is missing from the accounting.

**The three depth transects trace a non-monotonic and genuinely unusual profile.** D3 (mid-depth) is both the coral-richest and highest-cover transect: 32.4% coral against the lowest algal cover of the three (41.5%), and the widest genus list — Porites 12.69% (39% of the transect's coral), Favites 3.34%, Montipora 3.29%, Pocillopora 2.97%, Acropora 2.46% and Favia 1.71%, plus roughly 5.9 more percentage points spread across genera outside the top six. D2 (shallow) carries only 20.0% coral against a striking 60.2% algae — the highest algal cover recorded at this site — with Porites still dominant (10.45%, 52% of the coral) and Pocillopora, Acropora, Montipora, Favia and Echinopora each in the 1.0–1.7% range. D4 (deep) collapses to just 11.5% coral, the lowest of the three, and to only four nonzero genera: Porites 6.20% and Acropora 4.90% together account for 96.5% of all coral present, with Montipora reduced to 0.30% and Pocillopora to a trace 0.10%. **Acropora's relative share of the coral assemblage rises steadily with depth — 6% of coral at D2, 7.6% at D3, 42.6% at D4** — even as total coral cover falls by nearly two-thirds between D3 and D4. The site-level figures of 9.78% Porites and 2.85% Acropora describe no single transect: they average a moderately diverse mid-depth assemblage against a near-monoculture shallow zone and a two-genus deep zone.

The photographs are consistent with a low-relief, rubble-dominated hardground carrying scattered coral colonies rather than a continuous coral wall. P3270386 is the clearest single frame: an extensive field of bare, algae-turfed pavement runs to the limit of visibility with a school of dark surgeonfish/fusiliers over it, and only in the foreground does a single large, olive-green massive lobate colony break the pattern — a growth form consistent with *Porites*, matching the genus's outsized share of the transect data. P3270407, P3270415, P3270417 and P3270419 show the same pattern repeated across the dive: isolated massive or corymbose colonies (a flattened brain-coral-type dome in P3270415, tan corymbose/staghorn heads consistent with *Acropora* in P3270417 and P3270419) standing on otherwise open rubble and sand, with fish moving through open water above. P3270393 is a tight bottom shot of the same substrate at close range — turf, dark-red CCA crusts and small yellow-green digitate colonies, illustrating the fine-scale texture behind the 17.6% CCA and 16.7% turf-plus-sediment figures. P3270394 is a macro of a single small coral head with a distinct green-yellow corallite pattern — the growth form is too close-cropped to assign with confidence, but is plausibly a *Favia*- or small massive *Porites*-type colony.

A tiered ledge of large plating/foliose colonies (P3270401–P3270406) is the one structurally complex feature in the set — stacked shelves of a ridged, bumpy-surfaced coral, olive-brown and fully pigmented, sheltering a spiny lobster (*Panulirus* sp.) across four consecutive frames, with anthias and other reef fish schooling above. The plate morphology is consistent with a foliose growth form of *Montipora* or *Porites*, but cannot be assigned to genus from the imagery alone.

**Condition assessment is the point of tension at this site.** The published image-based bleaching figure is **16.0%, notably high for this survey batch**, yet none of the 13 examined frames shows the stark white, intact-tissue pallor that would corroborate it — every coral colony visible, from the massive *Porites*-type heads to the corymbose *Acropora*-type thickets to the plating ledge, is fully pigmented in natural olive, tan or brown tones. This is the reverse of the usual mismatch direction (elsewhere in this batch photos have shown visible bleaching against a low published figure); here the published figure runs high while the imagery shows none. The site value is taken as published per protocol, but the disagreement is real and unresolved — it may reflect bleaching concentrated in frames or reef sectors not represented among these 13 stills, or an artifact of the image-bleaching algorithm on this site's pale, sediment-dusted pavement background.

No crown-of-thorns predation scars, disease lesions, or other conspicuous condition anomalies were observed in any frame. The one clear faunal record is the spiny lobster occupying the plating-coral ledge.

<b>Assemblage composition.</b> Formal transects give 21.3% coral against 48.9% algae — an algae-dominated hardground — carried overwhelmingly by *Porites* (9.78% of benthos, 46% of the coral) with *Acropora*, *Montipora*, *Pocillopora*, *Favites* and *Favia* trailing well behind and sixteen further genera in the tail. Depth structure is non-monotonic: coral cover and genus richness peak at the mid-depth transect (D3, 32.4%), are lower and algae-dominated in the shallows (D2, 20.0% coral, 60.2% algae), and collapse to a two-genus *Porites*-*Acropora* assemblage at depth (D4, 11.5% coral). The photographs document extensive bare/turf/CCA pavement punctuated by isolated massive and corymbose colonies, plus one structurally complex plating-coral ledge sheltering a spiny lobster — fully consistent with the low-to-moderate cover and heavy algal dominance in the transect data, but showing no visible bleaching despite the high 16.0% published figure.

**Flags:** (1) **Bleaching mismatch, high-published/no-visible direction:** the published image-bleaching figure is 16.0%, notably high for this batch, but none of the 13 examined photographs shows white or pale coral tissue — every colony is fully pigmented. The published figure is retained as-is (it is a sourced column, not recomputed here), but the disagreement is unresolved. (2) **D4 genus collapse:** only four genera are nonzero at the deepest transect (Porites 6.20%, Acropora 4.90%, Montipora 0.30%, Pocillopora 0.10%), together summing exactly to the D4 coral cover of 11.5% — no other genus is present there. Site-level genus percentages (9.78% Porites, 2.85% Acropora) should never be quoted as describing this or any single transect. (3) **Acropora's relative dominance increases with depth** (6% of coral at D2 to 42.6% at D4) even as absolute coral cover falls — an unusual pattern worth flagging for anyone building depth-response models from this dataset. (4) **Annotation provenance:** the CoralNet point-annotation set behind the multi-method comparison is approximately 76% unconfirmed machine output archipelago-wide (only ~600 of 2,500 Chagos images are Anderson-confirmed); treat the 15.6% CoralNet-points figure as a robot-majority product, not a human-verified one. (5) **Multi-method comparison, no frame-count mismatch:** all three methods here draw on the full complement of 13 frames in the site folder (transects 21.3%, CoralNet points 15.6%, Coralscapes masks 14.5% — both CV methods read roughly 6–7 percentage points below the transect figure). Unlike some other sites in this batch, there is no discrepancy between the photo folder's frame count and the multi-method audit's n_photos. (6) **Plating-ledge colony genus is unresolved.** The tiered plating colonies in P3270401–P3270406 could not be assigned to genus from the imagery (plausibly *Montipora* or a foliose *Porites*); do not caption them beyond "plating/foliose coral." (7) **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.

**Benthic data:** [Download BITB56 benthic survey data (CSV)](../benthic_data/BITB56_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).