### BIPB114 — Peros Banhos
**CRD page:** [Peros Banhos Coral Surveys (30 Apr – 3 May 2015)](https://coralreefdiagnostics.com/peros-banhos-)

**2 May 2015**
**GPS:** 5.25724°S, 71.97576°E
**Setting:** exposed emergent fringing forereef; **survey depth is blank in the published table** (see Flag 3)
**Coral cover:** 39.6% (3 formal transects, D2–D4; **algal cover 33.8%**) · **published image bleaching:** 22.6% on 2 May and 22.7% on 3 May (`% bleaching (2nd training)`; see Flag 1 — a second published column gives 2.5% for the same dive)
**Live hard coral cover across methods:** formal transects 39.6% · CoralNet points 19.6% · Coralscapes masks 18.9% (41 frames — see Flag 2, that pool spans both dives). These products have different sampling footprints and denominators and should not be averaged.
**Temperature:** blank in the site table; 29.9 °C in `CHAGOS_SITE_REFERENCE_20260825.csv` (Flag 4) · **Salinity:** not reported
**Photo review:** all 21 stills of the 2 May dive and all 20 stills of the 3 May dive examined — 41 of 41 across the two dives. The 2 May folder also holds one zero-byte `.MOV` iCloud placeholder (P5020842.MOV) with no reviewable content. Every date was confirmed frame by frame from EXIF `DateTimeOriginal`, not from the filename stamp alone: 2 May 16:50–17:30 and 3 May 07:28–08:23.
**Representative photos:** P5020884.jpg, P5020886.jpg, P5020876.jpg, P5030837.jpg, P5030884.jpg

BIPB114 is a windward, emergent fringing forereef on the outside of Peros Banhos, and it is the archipelago's mid-pack in cover but near its top in diversity: **39.6% coral against 33.8% algae — 54th of 109 sites for coral cover, 76th of 109 for algal cover — across 31 nonzero genera, the 14th-highest genus count of any of the 109 Chagos sites.** The substrate budget separates it sharply from the sheltered lagoon reefs nearby: **bare substrate holds 12.8% of the benthos and non-coral invertebrates 13.9%**, so more than a quarter of the reef is neither coral nor algae — consistent with the extensive open pavement, rubble and conspicuous sponge cover the photographs show. The algal fraction itself is spread across crustose coralline 12.2% (65th of 109), turf 8.3% (68th of 109), thin turf on sand 4.8%, erect algae 5.2% and macroalgae 3.2%. Coral, substrate and algal categories together close to 100.0% of the benthos with nothing unaccounted for.

**All genus values below are percent of total benthos, not percent of coral.** This is a *Porites* reef with a long tail: ***Porites* alone holds 24.28% of the benthos — 61% of all the coral recorded here.** *Pocillopora* (3.26%), *Montipora* (2.87%) and *Acropora* (2.04%) follow at 8%, 7% and 5% of the coral respectively, and those four genera account for 82% of the coral present. Below them sit *Astreopora* 1.14%, *Platygyra* 1.09%, *Leptastrea* 0.86%, *Favia* 0.75%, *Pavona* 0.51% and *Psammocora* 0.47%, then twenty-one further genera each under 0.31%: *Montastrea*, *Favites*, *Goniastrea*, *Heliopora* (the HELI code — blue coral, not the free-living mushroom *Heliofungia* the LOF codebook label suggests), *Symphyllia*, *Turbinaria*, *Ctenella* (CNEL — the Chagos-endemic monotypic genus, not *Cynarina*), *Stylophora*, *Pachyseris*, *Oulophyllia*, *Cyphastrea*, *Millepora*, *Heliofungia*, *Plesiastrea*, *Fungia*, *Cantharellus*, *Acanthastrea*, *Blastomussa*, *Leptoseris*, *Lobophyllia* and *Gardineroseris*. **The unusually low *Acropora* share is the site's defining structural feature** — on a Chagos forereef where *Acropora* routinely holds half to four-fifths of the coral, here it is fourth-ranked at one-twentieth. The 31 genus values sum to 39.5533%, an exact match to the published coral cover figure.

**The depth profile is the cleanest monotonic decline of any site described in this batch, and *Porites* carries the reef all the way down.** D2 holds 56.7% coral on 29.9% algae, with ***Porites* at 41.77% — 74% of all the coral on that transect** — and *Pocillopora* 3.53%, *Acropora* 2.95%, *Astreopora* 1.95%, *Montipora* 1.35% and *Psammocora* 1.25% behind it. D3 falls to 38.8% coral on 36.2% algae: *Porites* drops to 17.71% and, more tellingly, to 46% of the coral present, while *Pocillopora* (5.02%) and *Montipora* (4.18%) rise into a genuinely mixed second tier alongside *Acropora* 2.46%, *Platygyra* 1.88% and *Astreopora* 0.92%. D4, the deepest, **collapses to 23.16% coral — a 33-point fall from D2** — on 35.3% algae, with *Porites* 13.36% (58% of the coral), *Montipora* 3.08%, *Leptastrea* 1.36%, *Pocillopora* 1.24%, *Pavona* 0.84% and *Acropora* down to 0.72%. **The site-level figure of 39.6% cover describes only D3**; it averages a shallow zone carrying nearly 57% coral against a deep zone carrying under a quarter, and *Acropora*'s already-small share shrinks by a further factor of four across that gradient.

The 2 May photographs split into two distinct kinds of frame, and this matters for what can be read from them. **Roughly the first half of the dive (P5020837–P5020848, 16:50–16:55) is ambient blue-water photography** — silhouetted schools and divers with essentially no colour information on the benthos. Those frames are excellent fauna documentation: dense schools of unicornfish and surgeonfish (*Naso*-type) with two divers in P5020837 and P5020838; fusiliers, bluefin trevally and a barracuda working the drop-off in P5020840 and P5020841; a sea turtle silhouetted above a diver in P5020845 and swimming through a cloud of planktivores in P5020846; and **a grey reef shark (consistent with *Carcharhinus amblyrhynchos*) cruising the slope in P5020848**. They are close to useless for colour-based condition scoring, and they are part of the pool the published image products drew on (Flag 5).

**The strobe-lit second half of the 2 May dive (P5020861–P5020888, 17:06–17:30) is where the reef and its condition are actually visible, and it shows a *Porites*-dominated pavement carrying widespread colony-scale bleaching.** P5020861 is the survey frame: two divers on the slope, one with a survey pole and tape and one carrying a large white quadrat frame, over a benthos of tan encrusting and massive heads, magenta crustose coralline, dark red macroalgae, black featherstars — and several conspicuously white branching colonies in the foreground. P5020874 and P5020875 are close views of the massive component: metre-scale mustard and pale-olive lobate heads consistent with *Porites*, swarmed by orange anthias, with a pale lilac leather-coral (*Sarcophyton*-type) and scattered white branching colonies between them. P5020876, P5020887 and P5020888 are the informative wide frames — a reef top of tan and mustard massive *Porites*-type heads and encrusting plates over pink coralline pavement, **with numerous small stark-white and pale-lilac branching and corymbose colonies scattered right across each frame** beside fully pigmented neighbours of the same growth form.

**Two subjects on 2 May are unambiguous and worth recording as such.** P5020883 and P5020884 photograph a giant clam (*Tridacna* sp.) set into a mound of **completely white, coarsely vesicular coral tissue** — the close frame P5020884 shows expanded, wholly bleach-white polyps at high magnification, with a dead clam valve beside them. The colony's growth form does not resolve to a genus in the imagery and is not assigned one (Flag 7). **P5020885 and P5020886 document a bleached anemone**: a cluster of long-tentacled anemones whose tentacles are white with pale lime tips and whose oral discs and columns show through as bare pink-orange, still hosting anemonefish. Bleaching in a host anemone is a different signal from coral bleaching — it is not counted in any coral cover or bleaching figure — and it is direct evidence that the thermal stress on 2 May extended beyond the scleractinians.

**The 3 May 2015 revisit — the same forereef, one day later, and separately scored.** BIPB114 was dived again the next morning (`BiPb114b-finished`, 20 stills, P5030837–P5030885, EXIF 07:28–08:23, 3 May 2015). Unlike the BIPB74 revisit, this one carries **its own published row**: the site code **BIPB114B** appears in the CoralNet workbook and in `CHAGOS_SITE_REFERENCE_20260825.csv` with its own bleaching value of **22.7%**, against 22.6% for 2 May. **The two consecutive days are statistically indistinguishable — a 0.1-percentage-point difference** — which is what a real, already-established bleaching state should look like across 15 hours, and is a useful internal consistency check on the scoring. BIPB114B has **no coral cover value of its own**: its LOF transect cell is blank and the 39.55% and 33.77% that appear beside it are copied down from the 2 May row (Flag 6). Everything above — reef type, depth zonation, generic composition, the 39.6% cover figure — describes the 2 May dive and is not restated.

**The 3 May photographs are the better benthic record of the two dives, and they confirm the same picture at close range.** **P5030837 is the definitive frame of this site: a fully stark-white corymbose *Acropora* colony sitting immediately beside a cream-pale colony of the identical growth form and a pigmented blue-grey meandroid brain colony, under one strobe** — within-frame contrast that rules out lighting or white balance. P5030876 is the most severe wide frame: a slope on which several stark-white massive and encrusting colonies and white digitate colonies are interleaved with fully pigmented tan domes and pink coralline pavement. **P5030884 places a survey line directly across a large stark-white plate colony** while a diver works above it and a shark passes in the distance — documentation that the formal survey ran over bleached ground. P5030862 shows a large pale-cream corymbose *Acropora* and a smaller fully white one on open sand beside pigmented magenta-brown *Pocillopora*, and P5030852 and P5030838 show the same mosaic at slope scale. Against that, the wide frames P5030883 and P5030885 read as a mostly pigmented reef — extensive brown and mustard massive and encrusting cover with white colonies as a conspicuous minority — which is the honest overall impression: **a serious, broadly distributed but minority bleaching signal on a reef whose dominant *Porites* component was largely still pigmented.**

The 3 May set also carries the site's best assemblage and fauna imagery. P5030859 and P5030860 are close views of a diverse patch — a cream-white massive colony, pigmented pink-brown meandroid colonies consistent with *Lobophyllia* or *Symphyllia*, a tan brain colony consistent with *Platygyra* or *Oulophyllia*, lime-yellow digitate colonies and a giant clam mantle — the visible counterpart of a 31-genus site. P5030877, P5030878 and P5030879 are macro frames of fine-polyped encrusting colonies (consistent with *Montipora*) growing against dark barrel and bright red encrusting sponges, the photographic counterpart of the site's 13.9% non-coral invertebrate cover. Fauna includes a moray with a bright yellow jaw in P5030845 (consistent with a yellow-edged or yellowmouth *Gymnothorax*, not resolved further), **two camouflaged ambush predators consistent with scorpionfish on sand in P5030881 and P5030882**, a sea cucumber in P5030882, and divers with survey gear in P5030861, P5030883 and P5030884.

<b>Assemblage composition.</b> Formal transects give 39.6% coral against 33.8% algae across 31 nonzero genera (14th of 109 for genus count), the coral 61% *Porites* with *Acropora* only fourth at 5% — an unusual structure for a Chagos forereef. Coral cover falls monotonically with depth, 56.7% at D2 to 38.8% at D3 to 23.2% at D4, with *Porites* leading every transect and *Acropora* thinning to 0.72% at the deepest. Both dives' photographs show the same reef: a *Porites*-dominated pavement with extensive bare and sponge-covered substrate, carrying a widespread but minority bleaching signal concentrated in small branching and corymbose colonies plus a few white massives and plates, alongside a bleached host anemone, a giant clam, a grey reef shark, large *Naso* and trevally schools, and two scorpionfish.

**Flags:** (1) **Severe internal conflict in the published bleaching columns for this site.** The CoralNet workbook (`Chagos CoralNet images LOF vs. CoralNet coral cover.csv`) gives BIPB114 `% bleaching (1st training)` = **2.47%** and `% bleaching (2nd training)` = **22.62%**, and BIPB114B 10.11% and **22.72%**. The `Mean(%bleaching)` column — the one `Chagos benthic data by site (n=109).csv` and `sitefacts.py` carry, and therefore the one the map and the central rank file use — is **2.469% for BIPB114**, which is the *first-training* value, not a mean of the two (their mean would be 12.5%); BIPB114B's `Mean(%bleaching)` cell is a fill-down copy of the same 2.469%. This shard quotes the 2nd-training values (22.6% / 22.7%), matching how the sister revisit sites BIPB70B and BIPB72B were handled, **but the conflict is unresolved and someone should decide which column is canonical for Chagos before these numbers reach the map.** (2) **The multi-method cover row pools both dates.** `chagos_multimethod_cover_by_site.csv` records `n_photos = 41` for BIPB114 — exactly 21 (2 May) + 20 (3 May) — so the CoralNet (19.6%) and Coralscapes (18.9%) figures blend the two dives. The 39.6% transect figure is 2 May only. Note the workbook's own per-dive CoralNet cover runs the other way, 20.4% for 2 May against 30.3% for 3 May, which no cover interpretation here relies on. (3) **The survey depth cell is blank in the published table**, as is salinity; neither is invented here. The D2–D4 depth codes carry no meter values either, so do not assign per-transect depths. (4) **Temperature conflict.** `Chagos benthic data by site (n=109).csv` leaves temperature blank for BIPB114, while `CHAGOS_SITE_REFERENCE_20260825.csv` lists 29.9 °C for both BIPB114 and BIPB114B. Provenance of the 29.9 is not established; it is reported as a secondary source. (5) **Roughly half the 2 May frames are ambient blue-water images with no usable colour** (P5020837–P5020848: schooling fish, divers, a turtle, a shark, all silhouetted), yet they are part of the pool the published bleaching and CoralNet cover figures drew on, and colour-based scoring on such frames is unreliable. Related: `chagos_pct_bleached_recomputed.csv` shows only **18 images annotated for BIPB114** against 21 stills in the folder — three frames were dropped upstream and which three is not recorded. (6) **BIPB114B has no cover value of its own.** Its `coral cover (%; LOF)` cell is blank; the 39.55%/33.77% shown beside it are copied from the 2 May row, and there is no 3 May transect. Never quote a 3 May cover figure from that row. (7) **A stark-white, coarsely vesicular colony in P5020883/P5020884 could not be assigned to a genus**, and none of the site's 31 recorded genera is an obvious match for that polyp morphology. Do not caption it beyond "bleached colony, genus unresolved". (8) **A bleached host anemone with resident anemonefish is documented in P5020885 and P5020886.** Anemone bleaching is captured by no coral cover or bleaching figure and is recorded here as an independent observation of thermal stress on 2 May. (9) **The published bleaching values do not reproduce from the point counts.** `chagos_pct_bleached_recomputed.csv` gives **3.85%** for BIPB114 (18 images, 540 points, 3 bleached of 78 coral points) and **14.29%** for BIPB114B (20 images, 600 points, 23 bleached of 161 coral points), against the published 22.6% and 22.7%. Both are marked `no_value_in_master`, so no automated status was assigned; the recomputes sit closer to the first-training column than the second (Flag 1). Unresolved. (10) **The published bleaching figure is a COMPUTER-VISION product, not an in-situ observation** — CoralNet annotation of the informal survey images (Mayfield & Dempsey 2026, *Oceans*). Per-image human-vs-model agreement is much weaker for bleaching than for cover, r = 0.40 bleached against r = 0.72 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. (11) **The central bleaching rank does not apply to the value quoted here.** `CHAGOS_CENTRAL_RANKS.json` ranks BIPB114 85th of 98 for bleaching, but computed from `Mean(%bleaching)` = 2.469% (Flag 1), not from 22.6%. **No bleaching rank is claimed in this shard.** The cover (54/109), algae (76/109), genus-count (14/109), CCA (65/109) and turf (68/109) ranks come off columns with no such conflict. Ranks off `Mean(%bleaching)` are approximate in general — that column carries at least one known bad value (BITB58 listed 22.22%, corrected 0.77%). (12) **The CRD page slug looks truncated.** `chagos_site_to_page.json` gives BIPB114 `https://coralreefdiagnostics.com/peros-banhos-` with no date range, where the same-window site BIPB112 carries `peros-banhos-20150430-to-0503`. Copied verbatim as the authoritative file requires, but **verify it before publishing — as it stands it is very likely a dead link.** (13) **A zero-byte `.MOV` placeholder sits in the 2 May folder** (P5020842.MOV, 32 bytes — an iCloud stub), which is why that folder lists 22 items but only 21 reviewable stills. (14) **No COTS, disease lesions, *Drupella* or predation scarring was observed** across either dive — an explicit absence-of-evidence finding across all 41 stills. (15) **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 **22.62%, rank 11 of 110** (1st training: 2.47%, rank 94 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 BIPB114 benthic survey data (CSV)](../benthic_data/BIPB114_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).