### BIPB74 — Peros Banhos
**CRD page:** [Peros Banhos Coral Surveys (Apr–May 2015)](https://coralreefdiagnostics.com/peros-banhos-20150418-to-0419)

**18 Apr 2015**
**GPS:** 5.26106°S, 71.95179°E
**Setting:** protected submerged lagoonal patch reef, 26 m
**Coral cover:** 62.4% (3 formal transects, D2–D4; **algal cover 32.0%**) · **published image bleaching:** 5.35%
**Live hard coral cover across methods:** formal transects 62.4% · CoralNet points 32.8% · Coralscapes masks 33.7% (26 frames — see Flag 2, that pool spans both dives). These products have different sampling footprints and denominators and should not be averaged.
**Temperature:** 30.3 °C · **Salinity:** 34.1
**Photo review:** all 10 frames of the 18 Apr transect dive examined, and all 16 stills of the 1 May revisit (`BiPb74b-finished`) examined — 26 of 26 stills across the two dives. The May folder also holds one zero-byte `.MOV` iCloud placeholder (P5010876.MOV) that carries no reviewable content.
**Representative photos:** P4180866.jpg, P4180843.jpg, P4180865.jpg, P5010867.jpg, P5010861.jpg

BIPB74 is one of the highest-cover reefs in the whole Chagos set: **62.4% living coral against 32.0% algae, ranking 7th of 109 sites for coral cover** and in the lower quartile (84th of 109) for algal cover. It is a submerged lagoonal patch reef inside Peros Banhos at 26 m under protected, leeward conditions — a sheltered mid-lagoon setting rather than an exposed forereef. Bare substrate is almost absent (2.2% of the benthos) and non-coral invertebrates hold only 3.4%, so essentially all of the non-coral space is a thin algal veneer: crustose coralline 12.0% (67th of 109), turf 8.7% (63rd of 109), erect algae 7.0% and macroalgae 4.3%. The 25 nonzero genus percentages and the substrate categories close to 100% of the benthos with nothing unaccounted for.

The assemblage is a single-genus reef with one substantial second player. **All values below are percent of total benthos, not percent of coral.** ***Acropora* alone holds 43.67% of the benthos — 70% of all the coral recorded here.** *Pavona* is a distant but real second at 5.54% (9% of the coral), then *Porites* 2.23%, *Pocillopora* 1.88%, *Montipora* 1.82%, *Stylophora* 1.73%, *Echinopora* 1.23% and *Favia* 1.16%. Those eight genera account for 94% of the coral present; the remaining seventeen are all trace — *Halomitra* 0.56%, *Gardineroseris* 0.38%, *Isopora* 0.37%, *Fungia* 0.31%, *Hydnophora* 0.22%, *Favites* and *Podabacia* 0.19%, *Montastrea*, *Alveopora* and *Cyphastrea* 0.15%, *Goniastrea* and *Pachyseris* 0.11%, *Leptoseris* and *Psammocora* 0.08%, and *Millepora*, *Astreopora* and *Platygyra* 0.04%. Genus richness (25 nonzero codes, 54th of 109) is squarely mid-pack: this site's distinction is cover, not diversity. The 25 genus values sum to 62.4074%, an exact match to the published coral cover figure.

**The depth profile turns on *Pavona*, and it is the most interesting structure in the dataset here.** D2, the shallowest transect, carries the site's *lowest* coral cover (55.0%) on its *highest* algal cover (40.3%), and it is the only transect where *Acropora* does not dominate outright: *Acropora* 26.12% is just 47% of the coral present, while ***Pavona* at 13.62% takes a full 25% of the coral on that transect** — more than twice its site-level value — with *Stylophora* 3.75%, *Pocillopora* 3.62%, *Favia* 2.38% and *Echinopora* 1.12% filling out a genuinely mixed community. D3 then flips into a near-monoculture: 71.2% coral, the profile's lowest algal cover (23.8%), and ***Acropora* at 61.22% — 86% of all the coral on that transect** — with *Pavona* collapsing to 1.00% and nothing else above 1.6%. D4 sits between the two at 61.0% coral and 32.0% algae, *Acropora* 43.67% (72% of the coral) with *Porites* 5.33% and *Montipora* 4.00% appearing as secondary taxa for the first time, plus *Halomitra* at 1.33% — the free-living fungiid component that the photographs pick up directly. **The site-level figures of 43.67% *Acropora* and 5.54% *Pavona* describe no single transect**: they average a shallower, algae-heavier, *Pavona*-substantial zone against a deeper *Acropora* monoculture, with a mixed *Porites*/*Montipora* band below that.

The 18 April photographs (14:28–14:52) document the *Acropora* canopy in detail. P4180843 is the orienting frame: a broad, low-relief patch-reef top under dim blue light with a diver and lights top-centre for scale, olive-yellow tabular and plating *Acropora* tiling the foreground, a large stand of dark maroon foliose plate colonies to the right, and one conspicuously white table plus a white fine-branching colony mid-frame. P4180860 and P4180864 show the same canopy running off a reef edge into blue water — mustard and olive tables stacked over pink crustose coralline pavement. P4180863 and P4180875 are the cleanest images of pigmented tabular *Acropora*: metre-scale tan-brown tables with fine white growing margins, tiered one above another over dark red-purple encrusting pavement. P4180862, P4180865 and P4180867 shift into a corymbose/table field where cream-pale and stark-white colonies sit beside olive and mustard ones in the same frame. **P4180866 is the definitive condition frame**: a fully white tabular *Acropora* in the centre, immediately surrounded by fully pigmented dark-maroon foliose plates, a pale grey-lilac plating colony, a mustard corymbose colony and a pink CCA-encrusted plate — within-frame contrast that rules out lighting or white balance as the explanation.

**Condition on 18 April reads as low-level but real, patchy, colony-scale bleaching, and the published figure agrees.** BIPB74's published image bleaching is **5.35%, ranking 68th of the 98 Chagos sites carrying a value — in the lower third for bleaching severity**. Across the 10 April frames a handful of individual tabular and branching colonies are stark white or cream-pale while their immediate neighbours of the same growth form are fully pigmented (P4180843, P4180861, P4180862, P4180865, P4180866, P4180867) — exactly the signature a single-digit site percentage should produce on a reef this dense with coral. No disease lesions, tissue-loss margins, *Drupella* aggregations or crown-of-thorns individuals or feeding scars were seen in any of the ten frames, and fish life is sparse (a few chromis and damselfish, no schooling).

<b>Assemblage composition.</b> Formal transects give 62.4% coral (7th of 109 Chagos sites) against 32.0% algae across 25 nonzero genera, the coral 70% *Acropora* with *Pavona* the only other genus above 2.3%. The depth profile splits that average three ways — a shallower, algae-heavier D2 where *Pavona* takes a quarter of the coral and *Acropora* under half, a D3 *Acropora* monoculture at 86% of the coral on 71.2% cover, and a mixed *Acropora*/*Porites*/*Montipora* D4 at 61.0%. The April photographs document the monoculture as a continuous field of tan and olive tabular colonies over crustose coralline pavement, punctuated by conspicuous stands of dark maroon foliose plates and by a small number of individually bleached white tables consistent with the 5.35% published figure.

**The 1 May 2015 revisit — the same reef, thirteen days later, photographed but never scored.** BIPB74 was dived again on 1 May 2015 and 16 stills were taken (`BiPb74b-finished`, P5010843–P5010893, EXIF 13:55–14:31, 1 May 2015; all dates confirmed frame by frame from EXIF `DateTimeOriginal`, not from the filename stamp alone). **There is no published bleaching value and no published cover value for that dive.** A code "BIPB74B" does not appear anywhere in the Mayfield & Dempsey supplemental table, in the site reference file, or in the CoralNet workbook — unlike BIPB70B, BIPB72B, BIPB69C and BIPB114B, which all carry their own scored rows. The May dive at this site was photographed and then left unscored, and **no substitute figure is computed or estimated here** (Flag 1). Everything above — reef type, depth zonation, generic composition, the 62.4% cover figure — describes the 18 April dive and is not restated; what follows is a qualitative photographic account of 1 May only.

**The 1 May frames show a reef in visibly worse condition than the same reef thirteen days earlier, and the pattern is a mosaic rather than a uniform whitening.** The most severe frames are unambiguous. **P5010867 is the worst in the set — a downward view in which stark-white tabular and branching *Acropora* occupy most of the visible reef, with pale yellow-green tables to one side and dark maroon plates behind.** P5010878 shows a skyline of white and cream tables running the width of the frame above a foreground littered with white branching and knobby colonies. P5010892 and P5010893 both look across a table field in which the pigmented mustard and tan colonies are in the near foreground and the mid-ground and distance are a continuous carpet of pale blue-white tables receding to the limit of visibility. **P5010877 is the cleanest paired comparison in the whole site: one enormous cream-white tabular *Acropora* sitting directly beside an equally large, fully pigmented rust-brown table with intact white growing margins**, the two colonies touching, under identical light.

**The taxonomic pattern is as informative as the severity.** The clearest close-range cases are *Pocillopora*: **P5010861 is a completely white verrucose *Pocillopora* colony filling the frame, ringed by fully pigmented dark-maroon foliose plate colonies**, and P5010879 shows a second stark-white *Pocillopora* from underneath, with a banded coral shrimp (consistent with *Stenopus hispidus*) sheltering beneath it on pink coralline pavement. Free-living fungiids are affected too: P5010845 has a stark-white disc-shaped fungiid on open substrate, and **P5010846 is a close-up of a completely white elongate free-living fungiid** — its outline is most consistent with a tongue coral, but the transects here record *Halomitra* (0.56%), *Fungia* (0.31%) and *Podabacia* (0.19%) and no *Herpolitha* at all, so it is not assigned to a genus (Flag 6). Against that, the **dark maroon foliose plating colonies are still fully pigmented in every May frame they appear in** (P5010845, P5010860, P5010861, P5010867, P5010878) — a within-site contrast between an apparently unaffected foliose component and a heavily affected tabular/branching one. Some colonies retain full colour: P5010844 is an intact olive-green corymbose *Acropora* with white-tipped branches and humbug damselfish sheltering in it, and pigmented tables persist in the foreground of P5010892 and P5010893.

Two further features of the May set need care. Several frames (P5010843, P5010845, P5010859) look across low reef flat where branching *Acropora* framework stands but is overgrown with brown turf and red macroalgae, interspersed with white branching colonies and a pedestalled massive colony consistent with *Porites*; **whether that framework is recent mortality or pre-existing dead structure cannot be determined from a photograph, and no mortality is claimed** (Flag 7). And a hawksbill turtle (*Eretmochelys imbricata* — serrated carapace margin, overlapping scutes, hawked beak) was photographed resting on the reef in P5010874 and P5010875, in the second frame lying directly on a bed of pale cream tabular *Acropora*.

**Context makes the missing May number worth chasing.** The three neighbouring Peros Banhos sites revisited the same week were scored on both visits and all three rose steeply: BIPB70 from 3.7% on 17 April to 33.3% on 1 May, BIPB72 from 10.0% on 18 April to 37.5% on 1 May, and BIPB69 from 0% on 17 April to 36.5% on 2 May (`% bleaching (2nd training)`, Mayfield & Dempsey 2026 supplemental S1). BIPB74's 1 May photographs are visually of the same character as those sites' May imagery. **That is a qualitative statement about photographs, not a number**: no value is assigned to the 1 May dive here, and the 5.35% in the header belongs to 18 April alone.

**Flags:** (1) **The 1 May 2015 revisit has NO published bleaching value and NO published cover value.** A "BIPB74B" row does not exist in `Chagos benthic data by site (n=109).csv`, in `CHAGOS_SITE_REFERENCE_20260825.csv`, or in `Chagos CoralNet images LOF vs. CoralNet coral cover.csv` — verified directly in all three. Sixteen stills exist for the dive and nothing was ever scored from them. Do not compute, interpolate or borrow a substitute figure; the correct entry for 1 May bleaching at this site is "not scored". **This is a real gap worth closing** — the 16 frames are annotatable and the three neighbouring revisits were all scored. (2) **The multi-method cover row pools both dates.** `chagos_multimethod_cover_by_site.csv` records `n_photos = 26` for BIPB74, which is exactly 10 April + 16 May frames, so the CoralNet (32.8%) and Coralscapes (33.7%) figures blend a pre-bleaching and a bleaching-onset reef and are not interpretable as either dive. The 62.4% transect figure is 18 April only and unaffected; the apparent ~29-point transect-vs-CV gap is inflated by that pooling. (3) **Two CoralNet-derived cover products disagree by 32 points here.** The CoralNet workbook (`Chagos CoralNet images LOF vs. CoralNet coral cover.csv`) gives BIPB74 **64.45%** (identical for both training runs); `chagos_multimethod_cover_by_site.csv` gives **32.8%**. Both are labelled CoralNet products; the conflict is unresolved and neither is treated as ground truth. (4) **The published bleaching figure does not reproduce from the point counts.** `chagos_pct_bleached_recomputed.csv` recomputes BIPB74 from 10 annotated images / 300 points / 11 bleached of 149 coral points = **7.38%** against the published 5.35% — delta +2.08 pp, status MISMATCH. The published column is authoritative and is what is quoted here. Those 10 annotated images are exactly the 18 April folder, confirming the published figure is April-only. (5) **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. Ranks off `Mean(%bleaching)` are approximate — that column carries at least one known bad value (BITB58 listed 22.22%, corrected 0.77%). (6) **A stark-white elongate free-living fungiid in P5010846 could not be assigned to a genus.** Its outline is most consistent with a tongue coral, but the BIPB74 transects record *Halomitra*, *Fungia* and *Podabacia* only, and no *Herpolitha*. Do not caption it beyond "free-living fungiid". (7) **Do not read the algal-overgrown branching framework in the May frames (P5010843, P5010845, P5010859) as event mortality**, and do not read the May pallor as cover loss — bleached coral is still live coral, and there is no May transect and no scored May image product for this site (Flag 1). Turf-covered standing framework cannot be dated from a photograph and may be pre-existing dead structure. (8) **Conspicuous dark maroon foliose plating stands appear in both dives** (P4180843, P4180860, P5010860, P5010861, P5010867) and stay fully pigmented on 1 May, yet the transects record *Echinopora* at only 1.23% and *Pavona* at 5.54% of the benthos. The imagery and the transect genus counts disagree on how much of the visible foliose component these genera represent; the disagreement is unresolved and those plates are not captioned to a genus. (9) **The two dives are not necessarily the same ground.** The April frames are dim and blue-lit over a low-relief patch-reef top (14:28–14:52, 10 frames); the May frames are brighter and shallower-looking (13:55–14:31, 16 frames). They share the site code and coordinates, but nothing establishes that they cover the same transect depths, so the April→May comparison here is site-level and qualitative, not colony- or transect-matched. (10) **A zero-byte `.MOV` placeholder sits in the May folder** (P5010876.MOV, 32 bytes — an iCloud stub), which is why the folder lists 17 items but only 16 reviewable stills. (11) **No COTS, disease lesions, *Drupella* or predation scarring was observed** in either dive — an explicit absence-of-evidence finding across all 26 stills, not an omission.

**Benthic data:** [Download BIPB74 benthic survey data (CSV)](../benthic_data/BIPB74_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).