### BIPB63 — Peros Banhos
**CRD page:** [Peros Banhos Coral Surveys (Apr–May 2015)](https://coralreefdiagnostics.com/peros-banhos-)

**15 Apr 2015**
**GPS:** 5.41518°S, 71.77499°E
**Setting:** protected submerged lagoonal patch reef, 29 m
**Coral cover:** 56.0% (4 formal transects, D1–D4; **algal cover 28.2%**) · **published image bleaching:** 2.0%
**Live hard coral cover across methods:** formal transects 56.0% · CoralNet points 31.7% · Coralscapes masks 32.8% (16 frames). These products have different sampling footprints and denominators and should not be averaged.
**Temperature:** 30 °C · **Salinity:** 34.3
**Photo review:** all 31 frames examined — 2 at the top level of the folder (P4150563, P4150634) plus 29 nested one level down in the `BiPb63-CoralNet` subfolder, reviewed together as one site (a `P4150633.MOV` video at the top level was noted but not reviewed as a photo)
**Representative photos:** P4150602.jpg, P4150625.jpg, P4150637.jpg, P4150629.jpg, P4150647.jpg

BIPB63 posts the highest transect coral cover of any site reviewed so far in this batch — **56.0% live coral against 28.2% algae** — and it is overwhelmingly a one-genus story. Of 22 non-zero genera in the site summary, ***Acropora* alone holds 48.57% of the benthos, 87% of all coral recorded here**. A distant second tier follows: *Porites* 2.27%, *Pocillopora* 1.08%, *Favites* 1.03%, *Favia* 1.01%. Everything else — *Fungia*, *Turbinaria*, *Montipora*, *Lobophyllia*, *Platygyra*, *Herpolitha*, *Goniastrea*, *Mycedium*, *Pavona*, *Symphyllia*, *Galaxea*, *Echinopora*, *Leptastrea*, *Millepora*, *Oxypora*, *Seriatopora*, *Stylophora* — sits below 0.4% of the benthos individually. The raw benthic CSV also carries *Sandalolitha* at 0.1%, a 23rd genus omitted from the site's genus summary entirely; this is noted rather than resolved (Flag 4).

**The depth structure is the most interesting single finding at this site, and it documents a near-monoculture collapsing with depth.** D1, the shallowest transect, carries the site's highest cover at 83.2%, of which *Acropora* is 78.00% — 94% of the coral on that transect — with algae at just 15.8%. D2 holds 66.2% cover, *Acropora* 62.53% (still 94% of the coral), algae rising to 20.9%. D3 holds 52.7% cover, *Acropora* 51.33% — **97% of the coral on that transect, an even higher fraction than the two shallower stations despite total cover falling**, algae 21.2%. Then D4 breaks the pattern entirely: cover collapses to 22.1%, algae rises to 55.0% — more algae than coral — and no genus dominates: *Favites* 4.10%, *Favia* 4.00%, *Pocillopora* 3.40%, *Acropora* down to only 2.40%, *Fungia* 1.50%, *Turbinaria* 1.50%. Three stations of essentially wall-to-wall *Acropora*, declining steadily in total cover (83→66→53%) while becoming proportionally more *Acropora*-dominated, give way at the fourth to a sparse, algae-dominated, genuinely multi-genus assemblage. **The site-level figure of 48.57% *Acropora* is a weighted average of this gradient: it understates how close to monospecific D1–D3 actually are, and overstates how much *Acropora* survives at the D4 depth.**

The photographs document the shallow *Acropora* system in depth. P4150602 and P4150623 show dense staghorn and corymbose thickets running to the frame edge, tan-brown and fully pigmented, consistent with the 94–97%-*Acropora* transects. P4150601, P4150603, P4150622, P4150624, P4150625, P4150628, P4150630 and P4150646 show extensive tabular *Acropora* plate fields forming shelf-like structures over a standing dead-coral framework — the classic Chagos "table reef" architecture — with large massive, boulder-form colonies (consistent with *Porites* or *Favia/Favites*, hedged, since none of these are individually confirmed from a photograph) filling the gaps beneath and diver pairs surveying above. P4150631, P4150636, P4150638, P4150639, P4150640, P4150641 and P4150644 show a deeper community of large boulder colonies hosting schooling anthias, damselfish, and a resident grouper (P4150641; spotted body pattern consistent with a large serranid, not identified to species). Dense baitfish schools recur through the set (P4150563, P4150632, P4150634, P4150635, P4150637, P4150639) — a strong pelagic presence typical of an exposed lagoon patch reef in blue water off the reef edge.

**Photographic bleaching is clearly visible despite an unusually low published figure.** The site's mean image-based bleaching is just **2.0%**, the lowest of any Chagos site compared in this batch so far, yet several frames show unambiguous, colony-scale white pallor. P4150626 shows a sharply white corymbose colony standing directly beside fully pigmented tan-brown neighbors. P4150629 shows a single pale blue-white tabular colony amid otherwise tan and olive plate corals. P4150645 shows a large bright white patch on or beside a massive boulder colony — hedged, since it is not resolvable from the image alone whether this is recently bleached tissue or exposed bare skeleton. P4150647 is the starkest example in the set: a conspicuous, large stark-white patch sitting among otherwise olive-tan tabular colonies, intact in form and clearly colony-scale rather than a lighting artifact. This is the same tension documented at neighboring BIPB62: coherent, colony-scale pallor in the imagery against a very low transect-based bleaching statistic — here the disparity is even sharper given how low the published value is. The published value is taken as authoritative; the visual discrepancy is flagged rather than explained.

Non-scleractinian fauna is present but sparse in the photo record. P4150577 shows a purple-red foliose colony with strongly radiating growth (consistent with *Lobophyllia*, hedged), a cream-tan massive boulder, a brain-textured colony (consistent with *Platygyra* or *Favia*, hedged), small white branching remnants, an orange brittle-star arm draped over the substrate, and encrusting coralline algae. P4150650 is the set's only clear gorgonian: a tan-yellow sea fan with red encrusting octocoral or sponge polyps growing over reef rock at the reef's structural margin — a fauna type not otherwise recorded at this site.

<b>Assemblage composition.</b> Formal transects give 56.0% coral, 28.2% algae, and an assemblage that is 87% *Acropora* by benthos cover across the four transects, with *Porites*, *Pocillopora*, *Favites* and *Favia* the only other genera above 1%. The photographs confirm a shallow, near-monospecific *Acropora* table-and-staghorn system giving way with depth to sparser boulder-coral ground, matching the transect-documented collapse from 83% cover and 94–97% *Acropora* at D1–D3 down to 22% cover and a coral-to-algae ratio below parity at D4.

**Flags:** (1) **D1–D3 are 94–97% *Acropora* by coral cover; D4 collapses to 22.1% total cover with 55% algae and no dominant genus.** The site-level 48.57% *Acropora* figure describes no single transect — it averages three near-monocultures against one algae-dominated station. (2) **Photographic bleaching (P4150626, P4150629, P4150645, P4150647) is unambiguous** — sharply white, intact colonies beside fully pigmented neighbors, including one large stark-white patch (P4150647) — despite a published mean image bleaching of only 2.0%, the lowest of any site compared so far. The site value is taken as published; the discrepancy is unresolved and may reflect the transects and the photographer sampling different parts of the reef. (3) **Multi-method frame-count mismatch:** `chagos_multimethod_cover_by_site.csv` records `n_photos = 16` for the CoralNet-point (31.7%) and Coralscapes-mask (32.8%) figures, while the photo folder holds 31 usable frames — those cover estimates rest on roughly half the available imagery. (4) **Genus-count discrepancy:** the raw benthic CSV lists 23 genera including *Sandalolitha* at 0.1%, but the site's genus summary carries only 22 non-zero genera and omits *Sandalolitha* entirely. Flagged, not resolved. (5) **CoralNet point annotations underlying the 31.7% figure are majority-unconfirmed machine output** (project-wide provenance: ~600 of 2,500 site images across the corpus are Anderson-confirmed, the remainder are robot predictions) — treat the CoralNet percentage as a point-annotation estimate, not a human-verified count. (6) **All 29 CoralNet-subfolder frames plus the video sit under the nested `BiPb63-CoralNet` directory; only 2 stills (P4150563, P4150634) sit at the parent folder level** — any tool that only scans the top level of `BiPb63-great dive-finished` will silently miss 94% of this site's imagery. (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 BIPB63 benthic survey data (CSV)](../benthic_data/BIPB63_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).