### BIPB72 — Peros Banhos
**CRD page:** [Peros Banhos Coral Surveys (Apr–May 2015)](https://coralreefdiagnostics.com/peros-banhos-20150418-to-0419)

**18 Apr 2015**
**GPS:** 5.27607°S, 71.88663°E
**Setting:** protected emergent back reef pinnacles, 18 m
**Coral cover:** 72.4% (3 formal transects, D2–D4; **algal cover 22.4%**) · **published image bleaching:** 10.0%
**Live hard coral cover across methods:** formal transects 72.4% · CoralNet points 28.6% · Coralscapes masks 29.0% (76 frames — see Flag 1, this pool is contaminated with a different site's imagery). These products have different sampling footprints and denominators and should not be averaged.
**Temperature:** 30.3 °C · **Salinity:** 34.4
**Photo review:** all 34 frames of the 18 Apr survey examined; the 1 May revisit adds 42 further frames (`BiPb72b-finished`), of which a systematic sample of 5 spanning the dive sequence was reviewed (P5010949, 966, 980, P5011016, 1024) — see the revisit section and Flag 13
**Representative photos:** P4180680.jpg, P4180717.jpg, P4180718.jpg, P4180733.jpg, P4180750.jpg

BIPB72 carries the second-highest coral cover of all 109 Chagos sites in the LOF BIOT-1 survey — **72.4% living coral against just 22.4% algae**, a lead almost nothing else in the archipelago approaches. Non-coral invertebrates hold only 2.6% of the benthos and bare substrate a mere 2.7%; nearly everything that is not live coral is a thin algal fringe split between turf (6.3%), macroalgae (6.3%) and erect algae (5.7%), plus 4.1% crustose coralline. This is a protected, emergent back reef pinnacle habitat inside the Peros Banhos lagoon at 18 m — a very different setting from the exposed forereefs that carry most of the archipelago's other high-cover sites.

The assemblage is overwhelmingly a single-genus reef. ***Acropora* alone holds 59.04% of the entire benthos — 81.6% of all the coral recorded here** — one of the most extreme single-genus dominance values in the Chagos dataset, and twenty-one genera total are present but every other one is minor. *Pavona* (2.06%), *Montipora* (1.91%), *Porites* (1.80%), *Goniastrea* (1.65%), *Pocillopora* (1.35%) and *Seriatopora* (1.17%) form a modest second tier; the remaining fourteen — *Symphyllia*, *Favia*, *Lobophyllia*, *Stylophora*, *Fungia*, *Astreopora*, *Millepora*, *Gardineroseris*, *Leptastrea*, *Isopora*, *Coscinaraea*, *Cyphastrea*, *Echinopora* and *Montastrea* — each hold under 0.6%. Genus richness (21 nonzero codes) is respectable on paper, but essentially all of it is a long tail beneath one dominant taxon.

**The depth profile is a clean two-part story: an *Acropora* monoculture in the shallow-to-mid zone giving way to a lower-cover, more even community at depth.** D2 (79.4% coral, 17.6% algae) and D3 (78.1% coral, 11.6% algae) are both effectively *Acropora* reef — 70.44% and 70.67% respectively, **89% and 90% of all the coral on each transect** — with only trace amounts of *Porites*, *Favia*, *Pocillopora*, *Goniastrea*, *Pavona* and *Montipora* filling the rest. D4 breaks that pattern: cover falls to 59.5% and algae more than doubles to 38.0%, and while *Acropora* is still the leading genus at 36.00%, its share of the coral present drops to 60% as five other genera — *Pavona* (4.50%), *Goniastrea* (3.50%), *Montipora* (3.50%), *Seriatopora* (3.50%) and *Porites* (2.50%) — each cross the 2.5% mark for the first time. The site-level figure of 59.04% *Acropora* is the average of a near-total monoculture and a markedly more mixed, lower-cover deep zone; neither number describes the other.

The photographs confirm the shallow monoculture in dramatic fashion. Frame after frame — P4180680, P4180687, P4180695, P4180702–705, P4180733–740 and the wide panoramas P4180756–759 — show a near-continuous canopy of tabular and corymbose colonies, consistent with *Acropora*, tiling the reef from the foreground to the limit of visibility, exactly the structure implied by 70%+ *Acropora* on D2 and D3. P4180704, P4180720 and P4180722 include a diver's transect tape laid directly across the table field, physical documentation of the formal survey passing through this same habitat. Away from the tabular canopy, isolated massive boulder colonies interrupt the field — a cluster of brown-to-dark domes beside a giant clam in P4180686, and a taller stack of pale, weathered massive colonies hosting anthias in P4180733 — forms consistent with *Porites*- or *Favia*-type growth, though transect *Porites* (1.80%) and *Favia* (0.59%) are both minor; the imagery and the transect counts disagree on how much of the visible massive-coral cover these two genera actually represent, and that disagreement is unresolved (Flag 3). A patch of pink crustose coralline rubble studded with small purple branching colonies in P4180739 is the visible counterpart of the site's 4.06% CCA.

**Condition is the other major finding, and here the photographs and the published figure agree well.** BIPB72's published image-based bleaching is **10.0% (9.99%), ranking 29th of the 98 Chagos sites with a value** — moderately elevated, not extreme, but well above the low single digits many sites carry. This figure derives from the same 34-frame annotated point set reviewed here (34 images, 1,020 points, 38 bleached of 397 coral points = 9.57% recomputed, matching the published 10.0% within 0.43 percentage points — a confirmed match). The photographs corroborate a moderate, patchy bleaching signal rather than a mass event: scattered stark-white or pale lavender tabular colonies sit beside fully pigmented tan and olive neighbours throughout the tabular field (P4180680, P4180703, P4180704, P4180718, P4180719, P4180734–738, P4180756–759), and one large massive colony in P4180750 is bleached essentially in its entirety, tissue still clearly intact with no algal overgrowth. This is coherent, colony-scale pallor consistent with thermal stress, not growth-tip whitening or a lighting artefact, and it recurs widely enough across the 34 frames to be a genuine site-level signal — corroborating, not contradicting, the published 10.0% figure.

Fish life is abundant and diverse. A large anemone consistent with the magnificent sea anemone (*Heteractis magnifica*) hosts anemonefish beside a bleached tabular colony in P4180717. Dense, near-opaque schools of small silvery planktivores shelter directly beneath table overhangs in P4180722, P4180725, P4180726 and P4180751, and open-water schools of fusiliers and small jacks stream over the reef crest in P4180702, P4180733–737 and P4180748. Butterflyfish (raccoon butterflyfish, bannerfish) and damselfish work over and around the tabular canopy in P4180687 and P4180721. No COTS, disease lesions or predation scars were noted in the 34 frames reviewed.

<b>Assemblage composition.</b> Formal transects give 72.4% coral (2nd of 109 Chagos sites) and 22.4% algae, the coral 82% *Acropora* across 21 genera, splitting into a near-total *Acropora* monoculture at D2/D3 and a more mixed, lower-cover community at D4. The photographs document that monoculture directly — continuous tabular/corymbose fields to the horizon — plus isolated massive boulder colonies of unresolved genus, a CCA-rubble patch, a magnificent-anemone/anemonefish association and dense reef-fish schooling, against a backdrop of patchy but genuine bleaching (10.0% published) visible on tabular colonies throughout the frames and on one fully bleached massive coral.

**The 1 May 2015 revisit — the archipelago's second-best reef, thirteen days later.** BIPB72 was dived again on
1 May 2015. Bleaching stood at **9.99% on 18 April and 37.5% on 1 May** (`% bleaching (2nd training)`, Mayfield &
Dempsey 2026 supplemental S1; an earlier training run scored the same May set at 29.9%) — the largest May value of
the four Peros Banhos sites revisited that week. Everything above describes the April dive and is not restated
here; what changed in thirteen days was condition, at the site carrying the second-highest coral cover of all 109
Chagos sites.

The May photographs show a mosaic rather than a uniform whitening, and the split falls along growth form.
**Tabular *Acropora* is the worst affected.** P5010966 sets the contrast in a single frame: a bank of stark-white
tables on the right, a fully pigmented yellow-green table on the left, and olive-brown tables between them, with
pale tables receding into the distance behind. P5010949 shows a large stark-white table above substrate carrying
extensive white and lilac fine-branching colonies. **P5010980 is the most severe frame in the set — a
near-continuous field of white tabular *Acropora* running to the limit of visibility, with a diver and a yellow
survey tape crossing the upper frame**, which places a survey line over an almost wholly bleached table stand.

The *Pocillopora*-dominated slope frames show a different signature: not stark white but **widespread paling**.
In P5011016 and P5011024 the great majority of *Pocillopora* colonies are pale lilac or white-tipped rather than
fully bleached, with a minority still carrying full dark-brown pigment and the occasional intact yellow-green
*Acropora* table among them. Read together, the two signatures suggest the tables had progressed further into
bleaching by 1 May than the *Pocillopora* had — an ordering visible in the photographs, though these are
non-standardized site images and no severity index is computed from them.

**Flags:** (1) **Severe data-conflict on bleaching.** The working manifest (`chagos_work_manifest.json`) lists BIPB72's published bleaching as 37.5%, but the actual published column (`CHAGOS_SITE_REFERENCE_20260825.csv` `pct_bleaching_published`, matching `sitefacts.py`'s Mean(%bleaching) exactly) gives **9.99% for BIPB72**, independently confirmed by a recompute from the 34-image annotated point set (9.57%, delta −0.43pp, status MATCH in `chagos_pct_bleached_recomputed.csv`). The 37.5% figure instead belongs to a separate site/dive coded **BIPB72B** (`pct_bleaching_published` 37.46%, recomputed 29.4% from 42 images, no master cross-check available) with its own photo folder (`BiPb72b-finished`, 42 frames, filenames P5010933–P5011024, dated 1 May 2015) that is **not** part of this description's scope. The manifest's blended `coral_cover` (72.4, correct for BIPB72) alongside its `pct_bleached` (37.5, actually BIPB72B's) and its total frame count (76 = 34 + 42) show the two dives were merged upstream. This shard uses only the correct, sourced BIPB72 figures (9.99% bleaching, 72.35% cover) and the 34-frame `BiPb72-finished` folder; BIPB72B is a distinct site that needs its own separate, correctly-attributed description. (2) **Multimethod row is contaminated by the same mixup:** `chagos_multimethod_cover_by_site.csv` records `n_photos=76` for "BIPB72," the same pooled 34+42-frame set, so the CoralNet (28.6%) and Coralscapes (29.0%) figures quoted above are not specific to the 18 April dive and should be treated as indicative only. (3) **Massive boulder colonies are conspicuous in several frames** (P4180686, P4180733) but transect *Porites* (1.80%) and *Favia* (0.59%) are both minor; the imagery and the transect genus counts disagree on the identity/extent of this component and the disagreement is unresolved — do not caption those domes as a specific genus. (4) **Patchy colony-scale bleaching** is visible on tabular *Acropora* throughout the dive (P4180680, P4180703, P4180704, P4180718, P4180719, P4180734–738, P4180756–759) and on one fully bleached massive colony (P4180750), consistent with the published 10.0% site figure (rank 29th of 98). (5) **EXIF timestamps for all 34 reviewed frames span one continuous 51-minute window** (12:40–13:31 UTC) on 18 April 2015, confirming this is a single dive, not two visits merged under one code — resolving the "BIPB72 appeared twice in an older site list" concern for this photo set specifically (the true second entry is BIPB72B, see Flag 1). (6) All 34 frames in `BiPb72-finished` were reviewed; none were unreadable or iCloud-evicted. (7) **This shard previously stated "all 34 frames examined" and described only the 18 April dive.** The site's photographs sit in two folders — `BiPb72-finished` (34 frames, 18 Apr) and `BiPb72b-finished` (42 frames, 1 May) — and a folder lookup keyed on the site-code prefix missed the second. Corrected 2026-08-26. (8) **The May frames are a systematic sample of 5 of 42**, not a full review, and are stated as such. (9) **The multi-method cover figures pool both dates.** `n_photos` = 76 is exactly 34 April + 42 May frames, so the CoralNet (28.6%) and Coralscapes (29.0%) values blend a pre-bleaching and a bleaching-onset reef and are not interpretable as either. The 72.4% transect figure is April-only and unaffected — the apparent 44-point transect-vs-CV gap at this site is inflated by that pooling. (10) **Do not read the image-derived cover drop (56.4% April to 45.1% May) as mortality.** Bleached coral is still live coral; an 11-point loss in thirteen days is far more likely to be the classifier reassigning white colonies. (11) **A survey tape and diver are visible in P5010980**, indicating a survey line was deployed on the 1 May dive; no 1 May transect data appears in the published tables, and whether one was recorded has not been established. (12) **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 **9.99%, rank 38 of 110** (1st training: 9.99%, rank 36 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 BIPB72 benthic survey data (CSV)](../benthic_data/BIPB72_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).