### BIPB70 — Peros Banhos
**CRD page:** [Peros Banhos Coral Surveys (16–17 Apr 2015)](https://coralreefdiagnostics.com/peros-banhos-20150416-to-0417)

**17 Apr 2015**
**GPS:** 5.32515°S, 71.85739°E
**Setting:** protected submerged lagoonal patch reef, 22.3 m
**Coral cover:** 60.3% (4 formal transects, D2–D5; **algal cover 32.9%**) · **published image bleaching:** 3.7%
**Live hard coral cover across methods:** formal transects 60.3% · CoralNet points 28.8% · Coralscapes masks 26.5% (51 frames — see Flag 2). These products have different sampling footprints and denominators and should not be averaged.
**Temperature:** not reported · **Salinity:** not reported
**Photo review:** all 28 frames of the 17 Apr survey examined; the 1 May revisit adds 28 further frames (`BiPb70b-bleaching-finished`), of which a systematic sample of 7 spanning the dive sequence was reviewed (P5010837, 846, 854, 886, 898, 911, 925) — see the revisit section and Flag 13
**Representative photos:** P4170794.jpg, P4170821.jpg, P4170867.jpg, P4170791.jpg, P4170875.jpg

BIPB70 is one of the highest-cover reefs in the Chagos set: at **60.3% transect coral cover it ranks 9th of 109 sites**, against 32.9% algae. Twenty-eight genera are non-zero, but the assemblage is overwhelmingly a single-genus story — ***Acropora* holds 42.92% of the benthos, 71% of all coral recorded here.** A long tail follows well below it: *Porites* 3.06%, *Echinopora* 2.17%, *Lobophyllia* 2.10%, *Pocillopora* 1.86%, *Montipora* 1.68%, *Fungia* 1.03%, *Seriatopora* 0.81%, *Goniastrea* 0.70%, *Favites* 0.67%, then *Heliopora* at 0.50% (the HELI code — a blue coral, not the free-living mushroom *Heliofungia* the LOF codebook label suggests, per the corrected labelling in the benthic CSV), *Stylophora* 0.40%, *Favia* 0.33%, *Herpolitha* and *Platygyra* 0.30% each, *Pavona* 0.28%, *Acanthastrea* 0.25%, *Montastrea* 0.15%, and a scatter below 0.11% including *Ctenella* (CNEL — the Chagos-endemic monotypic genus, not *Cynarina*) at 0.11%. Substrate is led by erect algae at 12.18%, well ahead of CCA (7.34%), turf (7.21%), macroalgae (5.12%), bare substrate (4.33%) and invertebrates (2.48%) — a reef that is mostly coral or fleshy/erect algae, with little bare space.

**The depth structure here is the most important feature of the dataset and it splits the site into two very different reefs.** D2 carries 69.3% coral of which *Acropora* is 56.44% — 81% of the coral on that transect. D3 is higher still at 75.3% coral, with *Acropora* at 68.44% — 91% of the coral, an effective monoculture and the site's single highest-cover, highest-*Acropora* row. Then the pattern breaks: D4 drops to 41.4% coral, the site's lowest, on the site's highest algal cover of 48.8%, and *Acropora* falls to 19.80% — only 48% of a much more diverse coral component that now includes *Lobophyllia* 4.40%, *Fungia* 3.00%, *Favites* 2.00%, *Pocillopora* 2.00% and *Porites* 1.80%. D5 recovers somewhat to 55.0% coral (39.0% algae), with *Acropora* at 27.00% — still under half the coral present — and *Echinopora* 8.00%, *Lobophyllia* 4.00%, *Montipora* 3.00%, *Seriatopora* 3.00% and *Goniastrea* 2.00% filling out a genuinely mixed assemblage. **The site-level figures of 60.3% cover and 42.92% *Acropora* describe neither zone well**: they average a near-monoculture tabular *Acropora* reef (D2–D3, 69–75% cover) against a lower-cover, algae-heavier, genus-diverse zone (D4–D5, 41–55% cover) where *Acropora* holds under half the coral present.

The photographs document the tabular *Acropora* zone in striking depth. P4170794 is the defining frame: pedestalled tables tiling the reef to the limit of visibility with a diver visible top-left, the visual match to the 91%-*Acropora* D3 transect. P4170817, P4170818, P4170830–836, P4170842–843 and P4170850–851 repeat the same structure — dense, overlapping plate and table colonies stacked several deep over a hard pavement, punctuated by massive lobate and mounding heads (consistent with *Porites* or *Favites*-type growth, not confidently assignable from the images alone) tucked beneath the canopy. P4170867 is the most colorful frame in the set: a dense corymbose/digitate *Acropora* thicket in mixed yellow-green, brown-purple and cream color morphs, a visual counterpart to the genus-rich D4/D5 transects even though this particular frame is still Acropora-dominated. Large schools of small silvery planktivorous fish (likely fusiliers) stream over the reef in P4170791–794, P4170830, P4170850 and P4170861, with trevally-type jacks working the school in P4170793 and a diver with camera and strobe visible resting on a massive coral head in P4170866, giving useful scale.

**Condition is where the photographs and the published number pull apart.** The published image-based bleaching figure of 3.7% is low for the archipelago — it ranks 78th of 98 Chagos sites with a value, placing BIPB70 among the less-bleached sites. Yet a large fraction of the frames — P4170818, P4170819, P4170821, P4170830, P4170834, P4170835, P4170836, P4170842, P4170850, P4170866, and P4170874 through P4170878 — show individual fully white or pale tabular *Acropora* colonies sitting directly beside pigmented tan, olive or brown tables of the same growth form. This pattern recurs from the earliest to the latest frames of the dive (14:21 to 15:17), so it is not confined to one moment or patch of reef; it reads as a real, colony-scale, patchy bleaching signal rather than a lighting or exposure artefact. The likely reconciliation is that a small number of visually conspicuous white colonies contribute little to a cover-weighted or tissue-area-weighted bleaching average on a reef this dense with coral, but that has not been tested here and the tension between the two lines of evidence is flagged rather than resolved.

<b>Assemblage composition.</b> Formal transects give 60.3% coral and 32.9% algae, with *Acropora* at 71% of all coral recorded — but that average conceals a sharp depth split between a near-monoculture tabular *Acropora* zone at 69–75% cover (D2–D3) and a lower-cover, more diverse, algae-heavier zone at 41–55% cover (D4–D5) where *Acropora* drops to under half the coral present. The photographs document the tabular zone in detail — stacked plate and table colonies, massive lobate heads beneath the canopy, and a colorful corymbose/digitate *Acropora* thicket — plus large fish schools and scattered, colony-scale pale/white *Acropora* tables inconsistent at face value with the low published bleaching figure.

**The 1 May 2015 revisit — the same reef, fourteen days later.** BIPB70 was dived again on 1 May 2015, and the
published record separates the two visits: bleaching stood at **3.7% on 17 April and 33.3% on 1 May**, a nine-fold
rise in a fortnight (`% bleaching (2nd training)`, Mayfield & Dempsey 2026 supplemental S1; an earlier training run
scored the same May set at 26.3%). Everything above — reef type, depth zonation, generic composition, the
*Acropora*-dominated assemblage — describes the April dive and is not restated here; what changed in fourteen days
was condition.

The May photographs show that change unevenly, and the unevenness is the finding. Early in the sequence the reef
still reads as largely pigmented: P5010854 is dominated by intact yellow-green *Acropora* tables, olive foliose
colonies and brightly pigmented faviids, with only scattered pale branches. P5010837 is a gorgonian sea-fan garden
on the slope whose fans are naturally pale and carry no bleaching signal. By mid-sequence the signal is
unmistakable — P5010886 shows a stark white tabular *Acropora* mid-frame beside cream-white massive colonies and
numerous pale branching colonies, and **P5010911 is an unambiguous, fully white *Pocillopora* colony photographed
directly above a normally pigmented dark-brown *Pocillopora***, the within-frame contrast that rules out flash
artefact or naturally pale tissue. The most severe frames come late: **P5010898 shows extensive stark-white
tabular and plating *Acropora* along a reef edge**, and **P5010925 is a landscape-scale view in which large white
tables occupy the mid-ground and pale tables recede across the whole visible flat**. Some table surfaces in
P5010898 are brown-olive rather than white, which is consistent with either recently dead, turf-colonised skeleton
or pre-existing dead framework; fourteen days is fast for that transition and the photographs cannot separate the
two, so no mortality is claimed here.

**The taxonomic pattern matters at this site in particular.** The visibly bleached colonies in the May set are
overwhelmingly tabular and branching *Acropora*, with *Pocillopora* the clearest close-range case — and BIPB70's
April transects record ***Acropora* at 42.92% of the benthos, 71% of all coral at the site**. A reef whose coral
is nearly three-quarters one thermally sensitive genus is maximally exposed to exactly this event, and the
photographs are consistent with that exposure. This is an observed association between the assemblage recorded in
April and the condition photographed in May; it is not a tested causal claim.

**Flags:** (1) **A second, unincorporated photo folder exists for this site code.** `BiPb70b-bleaching-finished` (28 frames, filenames P5010837–P5010928, EXIF-dated **1 May 2015**, alongside a `ChSe121-meta data.jmp` file) sits in the same parent directory as `BiPb70-finished` (the 17 Apr 2015 folder used for this shard). This is direct photographic confirmation of the known BIPB70 code-duplication issue: the current single-entry manifest maps BIPB70 to the 17 Apr dive documented here, but a second dive roughly two weeks later was also once filed under a BIPB70-derived name (and an internal reference "ChSe121" that does not appear in the current benthic CSV). That folder's imagery and any associated data are NOT part of this shard and have not been reconciled against the current single-entry benthic record; do not merge the two photo sets or assume they share transect data without independently verifying which dive the D2–D5 rows used here actually belong to. (2) **Photo-count mismatch, reversed direction.** `chagos_multimethod_cover_by_site.csv` records `n_photos = 51` for BIPB70, but the `BiPb70-finished` folder mapped to this site holds only 28 stills — fewer than the CoralNet/Coralscapes analysis apparently drew on. The CoralNet (28.8%) and Coralscapes (26.5%) cover figures cannot be fully traced to the 28 frames reviewed here; the extra ~23 frames may originate from the `BiPb70b` folder (Flag 1) or from unedited source imagery not present in the finished/edited set. (3) **Bleaching tension:** published bleaching is 3.7% (78th of 98 valued sites, i.e. comparatively low), yet numerous frames spanning the whole dive show conspicuous pale/white tabular *Acropora* beside pigmented neighbors (see frame list above). Unresolved. (4) **D2 and D3 are near-monocultures** (81% and 91% *Acropora* of the coral present, respectively); never quote the site-level 42.92% *Acropora* as representative of either shallow transect. (5) **Massive lobate/mounding heads visible beneath the *Acropora* canopy** in several frames (P4170791–794, P4170830–832, P4170850, P4170861) could not be confidently assigned to a genus from the imagery; do not caption them beyond "consistent with *Porites* or *Favites*-type" growth. (6) Actual meter depths for transect rows D2–D5 are not present in the source data — only the site mean of 22.3 m is reported; do not invent per-row depths. (7) **This shard previously stated "all 28 frames examined" and described only the 17 April dive.** The site's photographs sit in two folders — `BiPb70-finished` (28 frames, 17 Apr) and `BiPb70b-bleaching-finished` (28 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 7 of 28**, not a full review, and are stated as such. (9) **The multi-method cover figures pool both dates.** `n_photos` = 51 against 28 April + 28 May frames, so the CoralNet (28.8%) and Coralscapes (26.5%) values blend a pre-bleaching and a bleaching-onset reef and are not interpretable as either. The 60.3% transect figure is April-only and unaffected. (10) **Do not read the image-derived cover drop (46.2% April to 38.9% May) as mortality.** Bleached coral is still live coral; a 7-point loss in fourteen days is far more likely to be the classifier reassigning white colonies than real tissue death. (11) **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 **3.66%, rank 87 of 110** (1st training: 3.66%, rank 86 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 BIPB70 benthic survey data (CSV)](../benthic_data/BIPB70_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).