### BIPB76 — Peros Banhos
**CRD page:** [Peros Banhos Coral Surveys (Apr–May 2015)](https://coralreefdiagnostics.com/peros-banhos-20150418-to-0419)

**19 Apr 2015**
**GPS:** 5.33774°S, 71.98058°E
**Setting:** exposed emergent fringing forereef, outside the lagoon, 21 m
**Coral cover:** 35.9% (3 depth transects, D2–D4; **algal cover 27.3%**) · **published image bleaching:** 22.0%
**Live hard coral cover across methods:** formal transects 35.9% · CoralNet points 17.0% · Coralscapes masks 17.1% (21 frames). These products have different sampling footprints and denominators and should not be averaged.
**Temperature:** 30.0 °C · **Salinity:** 33.7
**Photo review:** all 21 of 21 stills examined (folder `BiPb76-finished`; single dive, 11:03:46–11:49:44 local from `DateTimeOriginal` EXIF, Olympus TG-2)
**Representative photos:** P4190887.jpg, P4190901.jpg, P4190903.jpg, P4190897.jpg, P4190899.jpg

BIPB76 is an exposed, emergent fringing forereef on the outside of Peros Banhos at 21 m, and two numbers define it. The first is bleaching: **at 22.0% it is the 6th most bleached of the 98 Chagos sites carrying a published value**, and the photographs corroborate that emphatically rather than marginally. The second is a substrate figure that is easy to miss: **crustose coralline algae (PCCATS) is 2.4% of total benthos — the LOWEST of all 109 Chagos sites, 109th of 109.** Coral cover is 35.9% (65th of 109) against 27.3% algae (94th of 109, in the lower decile), but the algal budget is turf, not coralline: fleshy turf (PTTS) is **16.95%, 11th-highest of 109 sites**, with turf-plus-sediment 3.95%, erect algae 3.07% and macroalgae 0.9%. Bare substrate is 17.87% and non-coral invertebrates **18.97%** — an unusually large invertebrate fraction, and the imagery shows the likely drivers: abundant black antipatharian bushes on the upper slope and leathery soft corals among the coral. **A reef with the archipelago's lowest CCA, top-decile turf and near-19% bare substrate is one whose dead framework is being colonized by turf rather than crusted over by coralline algae** — which is what the photographs show.

Coral cover is spread across **18 nonzero genera, 84th of 109 for genus richness**, and the 18 genus percentages sum to exactly 35.90% of the benthos, matching the published cover figure. All values below are percent of total benthos, not percent of coral. *Porites* leads at 13.75% (38% of all the coral present) ahead of *Acropora* 10.20% (28%); together those two hold two-thirds of the coral. Then a steep drop to *Leptastrea* 1.85%, *Astreopora* 1.50%, *Goniastrea* 1.40%, *Turbinaria* 1.33%, *Stylophora* 1.13%, *Favia* 1.03%, *Pocillopora* 0.87% and *Montipora* 0.85%, followed by a tail of *Symphyllia* 0.47%, *Montastrea* 0.40%, *Psammocora* 0.35%, *Pavona* 0.23%, *Heliopora* 0.20%, *Platygyra* 0.13%, *Seriatopora* 0.13% and *Goniopora* 0.07%.

**The three depth transects show coral cover falling by more than half from shallow to deep, and *Acropora* disappearing completely along the way.** D2, the shallowest station, is the high-cover reef: **54.5% coral against 39.0% algae, and it is *Acropora*-dominated — that genus alone reaches 28.00% of the benthos, 51% of all the coral on the transect** — with *Porites* 18.25%, *Astreopora* 2.50%, *Pocillopora* 2.00%, *Montipora* 1.75% and *Leptastrea* 0.75%. At D3 cover falls to 32.2% against 34.8% algae and the lead changes hands: ***Porites* takes over at 13.00%** (40% of that transect's coral), with *Stylophora* 3.40%, *Acropora* down to 2.60%, *Favia* 2.60%, *Astreopora* 2.00% and *Leptastrea* 1.80%. At D4, the deepest station, cover falls again to 21.0% and **only four genera are recorded at all: *Porites* 10.00% (48% of that transect's coral), *Goniastrea* 4.00%, *Turbinaria* 4.00% and *Leptastrea* 3.00%. *Acropora* is entirely absent from the deepest transect — 28.00% of the benthos at D2, 2.60% at D3, zero at D4.** That collapse is the single most striking pattern in the site's data, and it means **the site-level figure of 10.20% *Acropora* describes only the shallow reef**: two-thirds of the profile has almost none of it. Algal cover falls even faster than coral, from 39.0% at D2 to 34.8% at D3 to just 8.0% at D4 — and at D4 coral and algae together account for only 29% of the benthos, leaving **roughly 71% unaccounted for by this depth-level summary** (against 6–7% at D2 and 33% at D3); that residual is presumably bare pavement, sand and invertebrate cover not itemized by depth, and it is flagged below.

The photographs — all 21 stills in the folder — document a reef with two distinct faces. The upper slope (**P4190879**, **P4190880**, **P4190886**, **P4190913**) is structurally varied and busy: pale carbonate slope studded with black antipatharian bushes, large crinoids with their arms fanned into the current, brown ruffled foliose plates consistent with *Turbinaria* or *Pavona*, yellow-green massive and encrusting colonies, a bright blue encrusting sponge in **P4190913**, and a constant haze of anthias and chromis in the water above. **The flatter ground below is something else entirely, and it is the site's defining image: an extensive field of large, flat, DEAD tabular *Acropora* plates**, still intact and still articulated, now uniformly coated in a tan-brown turf-and-sediment film. **P4190897**, **P4190910**, **P4190914** and **P4190915** are wide views over that field, in which the dead plates are the dominant structural element and living colonies are scattered between them; **P4190895** is a close view of a single such table entirely overgrown by a pink-red algal film. This dead tabular framework, colonized by turf rather than crusted by coralline algae, is the visual counterpart to the site's 109th-of-109 CCA figure, its top-decile turf figure, and the loss of *Acropora* with depth in the transects. Growth-form calls from photographs are given as "consistent with"; the genus percentages above are transect identifications.

**Condition is severe, and the photographic evidence supports the published figure directly rather than merely being compatible with it.** The published image-based bleaching value is **22.0%, ranking 6th of 98 Chagos sites** — in the top decile of bleaching severity — and at least seven frames show unambiguous, colony-scale bleaching with pigmented within-frame controls:

- **P4190887** — a large table *Acropora* that is **bone-white across its entire surface**, standing on a dark dead base with pigmented brown colonies immediately beside it.
- **P4190901** — a second large table, **uniformly white to cream over the whole plate**, with dascyllus damselfish sheltering above it and pigmented colonies at the frame margins.
- **P4190903** — a **stark bone-white corymbose colony** centre-frame against red-brown encrusted pavement, with pigmented tan massive colonies alongside.
- **P4190912** — a **bone-white branching *Acropora*** on red-brown pavement, with pigmented colonies above and below it in the same frame.
- **P4190911** — a **fully white verrucose *Pocillopora*-type colony** directly abutting a pigmented tan encrusting colony.
- **P4190894** — a white-to-cream corymbose *Acropora* immediately beside a fully pigmented pink-brown pocilloporid, with a leathery soft coral in the same frame; the sharpest side-by-side contrast in the set.
- **P4190898** — a pocilloporid with **white patches over most branch tips** and a wholly white branch cluster behind it, against a pigmented yellow massive colony.

Further white colonies appear in **P4190904** and **P4190915**. **P4190899 is recorded separately and as ambiguous**: it shows a pale blue-lilac encrusting colony whose margin runs into a narrow band of stark white skeleton and then into brown algae-covered dead surface — a sequence consistent with either an advancing tissue-loss front or bleaching followed by mortality and algal colonization, and not resolvable from one still. **The base of the large meandroid brain colony in P4190902 shows white patches consistent with partial basal mortality**, also recorded rather than interpreted. **No crown-of-thorns starfish, no *Drupella* aggregation, and no diagnostic disease lesion with a characteristic band or margin was identified in any of the 21 frames**; the white colonies above are recorded as bleaching because no lesion morphology is visible. These are frame-level records only — **no site bleaching percentage was recalculated from them**, and the 22.0% figure above is quoted verbatim from the published column.

Fauna is abundant: dense anthias and chromis over the upper slope in **P4190879** and **P4190880**, schooling unicornfish and surgeonfish crossing the dead-table field in **P4190900** and **P4190914**, and a single distant unidentifiable fish as the only subject of **P4190878**, otherwise an open blue-water frame. Black antipatharian bushes, large crinoids and a bright blue encrusting sponge (**P4190913**) are a plausible visual account of the site's 18.97% non-coral invertebrate cover.

<b>Assemblage composition.</b> Formal transects give 35.9% coral against 27.3% algae — an algal budget that is turf, not coralline, since PCCATS at 2.4% of benthos is the lowest of all 109 Chagos sites while PTTS at 16.95% is 11th-highest — across 18 nonzero genera led by *Porites* (13.8%) and *Acropora* (10.2%), which together hold two-thirds of the coral. The depth profile falls steeply, 54.5% coral at D2 to 32.2% at D3 to 21.0% at D4, with *Acropora* collapsing from 28.0% of benthos to 2.6% to zero and *Porites* inheriting the lead by attrition. The photographs document a varied upper slope of antipatharians, crinoids, foliose plates and massive colonies giving way to an extensive field of dead, turf-filmed tabular *Acropora* framework — and, in at least seven frames, unambiguous colony-scale bleaching consistent with this site's 6th-of-98 published bleaching figure.

**Flags:** (1) **THIS IS A TOP-DECILE BLEACHING SITE AND THE IMAGERY CONFIRMS IT: 22.0%, 6th of 98.** Seven frames (P4190887, P4190901, P4190903, P4190912, P4190911, P4190894, P4190898) carry bone-white colonies with pigmented within-frame controls, on 19 Apr 2015. Unlike sites where the published bleaching signal is driven by pale growing tips or paling sponges, the evidence here is unambiguous at colony scale. **The rank is nonetheless approximate**: the `Mean(%bleaching)` column carries at least one known bad value (BITB58 listed at 22.22% against a corrected 0.77%), and the ordering near the top of that column has not been independently audited — note that BITB58's erroneous 22.22% sits immediately above BIPB76's 22.02%, so this site's exact position in the top ten depends on a value known to be wrong. (2) **P4190899 IS AN AMBIGUOUS TISSUE-LOSS OBSERVATION AND SHOULD BE LOOKED AT BY A SECOND READER.** A pale blue-lilac encrusting colony with a narrow band of stark white skeleton at its margin, running into brown algae-covered dead surface. That sequence is consistent with an advancing tissue-loss front (a disease-like progression) or with bleaching followed by mortality and algal overgrowth. Not resolvable from a single still and not counted as either. **P4190902** separately shows white patches at the base of a large meandroid brain colony, consistent with partial basal mortality. (3) **LOWEST CRUSTOSE CORALLINE ALGAE IN THE ARCHIPELAGO: PCCATS = 2.4%, 109th of 109 sites**, paired with 11th-highest fleshy turf (16.95%), 17.87% bare substrate and 18.97% invertebrates. The imagery shows the mechanism — an extensive field of intact but DEAD tabular *Acropora* colonized by turf and sediment film rather than crusted by coralline algae (P4190897, P4190910, P4190914, P4190915, P4190895). Worth pairing with the bleaching figure in any analysis of post-mortality substrate trajectory in Chagos; this site is the extreme end of that axis. (4) **MULTI-METHOD DEVIATION IS LARGE: CoralNet points (17.0%) and Coralscapes masks (17.1%) sit ~19 points below the formal transect (35.9%; deviation range −18.9 to −18.8 pp).** Do not average the three; the transect value is ground truth for cover. Note the two CV products agree with each other to 0.1 pp while both disagree with the transect by 19 — agreement between two image methods is not evidence that either is right. (5) **`n_photos` = 21 happens to equal the folder frame-count here** — a coincidence, not a guarantee of comparability; the denominators still differ. It also means **P4190878, which is an open blue-water frame with a single distant fish and essentially no benthos, is inside the CV sample.** Frames dominated by open water are where human/model agreement is weakest corpus-wide. (6) **The CoralNet points figure is a CV product, not verified ground truth**; point-annotation data corpus-wide is roughly 76% unconfirmed machine output. (7) ***Acropora* IS ABSENT FROM THE DEEPEST TRANSECT.** 28.00% of benthos at D2, 2.60% at D3, 0.00% at D4. The site-level 10.20% figure describes the shallow station only and must not be quoted as a whole-site *Acropora* value without the depth split. Only four genera are recorded at D4 at all. (8) **D4 HAS ROUGHLY 71% OF ITS BENTHOS UNACCOUNTED FOR** by coral (21.0%) plus algae (8.0%) in the site-depth summary, against 6–7% unaccounted at D2 and 33% at D3. Presumably bare pavement, sand and invertebrate cover not itemized by depth in this dataset, but the size of the residual is large enough that D4's algal figure of 8.0% should not be used on its own. (9) **The depth codes D2–D4 carry no metre boundaries** in `Chagos benthic data by site-depth (n=336 rows).csv`; they are ordered shallow-to-deep stations, and the 21 m in the header is the site-level survey depth, not the depth of any one transect. (10) **No D1 transect exists** — the profile starts at D2, so the shallowest reef zone is unsampled (`N Rows` = 3). (11) **Exposure was revised between fields.** `exposure-original` records this site as **windward**; `exposure-revised` gives **exposed**, and the header above uses the revised value. (12) **No COTS and no diagnostic disease sign was identified across the full 21-frame review** — an explicit absence-of-evidence finding, subject to flag 2. (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, and never read it as a diver's field estimate — **at this site the photographs happen to support it, which is a per-site observation and not a validation of the column.** (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 **22.02%, rank 12 of 110** (1st training: 22.02%, rank 8 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 BIPB76 benthic survey data (CSV)](../benthic_data/BIPB76_benthic.csv) — scleractinian cover by genus with full names, plus substrate and algal categories (bare substrate, CCA, turf, turf+sediment, macroalgae, erect algae, invertebrates, cyanobacteria), from the LOF BIOT-1 survey (Mayfield & Dempsey 2026, *Oceans*, supplemental file S1).