### BIVB109 — Victory Bank
**CRD page:** [Blenheim, Nelson Island & Victory Bank Coral Surveys (Apr 2015)](https://coralreefdiagnostics.com/new-page-4)

**30 Apr 2015**
**GPS:** 5.54759°S, 72.21877°E
**Setting:** protected submerged forereef outside the lagoon, 31.2 m (the `reef type` and `exposure-original` cells are blank in the published table — see Flags)
**Coral cover:** 34.7% (4 depth transects, D2–D5; **algal cover 42.0%**) · **published image bleaching:** 8.7%
**Live hard coral cover across methods:** formal transects 34.7% · CoralNet points 24.5% · Coralscapes masks 32.7% (33 frames). These products have different sampling footprints and denominators and should not be averaged.
**Temperature:** 30.0 °C · **Salinity:** 34.2
**Photo review:** 17 of 33 jpg stills examined — a systematic every-second-frame sample spanning the whole dive sequence from P4300936 to P4300993 (P4300936, 939, 941, 943, 950, 958, 960, 962, 971, 973, 975, 978, 985, 987, 989, 991, 993). The folder holds 34 files; the 34th is a video clip (see Flags).
**Representative photos:** P4300978.jpg, P4300950.jpg, P4300987.jpg, P4300975.jpg, P4300962.jpg

BIVB109 is a submerged forereef on Victory Bank, an isolated offshore bank at the northern end of the archipelago, surveyed at 31.2 m under protected exposure. Its headline numbers are close to the Chagos median — **34.7% coral against 42.0% algae, ranking 66th of 109 sites for coral cover and 45th of 109 for algal cover** — but the substrate composition is not median at all. Crustose coralline algae (PCCATS) holds **20.7% of total benthos, 20th-highest of the 109 Chagos sites**, against only 7.3% fleshy turf (PTTS, 77th of 109). Bare substrate runs 16.6%, macroalgae 4.7%, encrusting algae 5.4%, turf-and-sediment 3.9% and non-coral invertebrates 6.8%. This is a hard, coralline-cemented offshore bank where "algae" means pink pavement rather than weedy turf — the opposite balance to the lagoonal sites, and visible in the imagery as pink-and-mauve crust between colonies rather than an olive turf mat.

Coral cover is spread across **26 nonzero genera, 47th of 109 Chagos sites for genus richness — mid-pack**, and the 26 genus percentages sum to 34.7300% of the benthos, an exact match to the published cover figure. All values below are percent of total benthos, not percent of coral. **Unusually for a Chagos site, the top three genera are near-equal and one of them is a hydrocoral:** *Acropora* 7.19% (21% of the coral present), *Porites* 6.90% (20%) and ***Millepora* 6.48% (19%)** — together 59% of all coral recorded. Behind them come *Platygyra* 2.48% (7%), *Goniastrea* 1.71%, *Montipora* 1.70%, *Pocillopora* 1.66%, *Stylophora* 1.39% and *Favia* 1.35%, then *Cyphastrea* 0.95%, *Leptastrea* 0.89%, *Pavona* 0.60%, *Favites* 0.44%, *Astreopora* 0.24% and *Goniopora* 0.20%. A tail of eleven further genera each sits below 0.15%: *Turbinaria*, *Echinopora*, *Montastrea*, *Symphyllia*, *Acanthastrea*, *Fungia*, ***Ctenella***, *Galaxea*, *Psammocora*, *Gardineroseris* and *Leptoseris*. **The presence of *Ctenella* (0.05% of the benthos) is worth noting on its own account: *Ctenella chagius* is endemic to the Chagos Archipelago** and appears at only a minority of sites.

**The four depth transects are the real story at this site, and they are among the most depth-structured in the dataset: the entire *Millepora* signal comes from one transect, and it is enormous there.** D2, the shallowest station, carries only 20.5% coral against the profile's highest algal cover, 70.3%; the assemblage is small-colony and *Porites*-led — *Porites* 5.50%, *Acropora* 5.30%, *Stylophora* 4.40%, *Pocillopora* 2.50%, *Montipora* 1.00%, *Pavona* 0.60%, *Favia* and *Favites* 0.40% each, *Millepora* 0.20%, *Fungia* and *Goniastrea* 0.10% each (11 genera). At D3 the reef transforms: **coral cover nearly triples to 59.22%, algae falls to 31.66%, genus count leaps to 24 of the site's 26 genera, and *Millepora* alone reaches 25.70% of the benthos — 43% of all the coral on that transect and four times its own site-level value.** *Acropora* (9.94%) and *Porites* (9.70%) are a distant second and third, with *Pocillopora* 2.64%, *Favia* 2.28%, *Montipora* 1.98%, *Stylophora* 1.16%, *Pavona* 1.00%, *Astreopora* 0.94% and *Platygyra* 0.90% following, above a tail of fourteen minor taxa. **Seven genera occur on this transect and nowhere else at the site: *Astreopora*, ***Ctenella***, *Echinopora*, *Galaxea*, *Gardineroseris*, *Leptoseris* and *Psammocora*.** **At D4 the reef collapses again: 21.2% coral, 33.3% bare substrate, and *Millepora* disappears entirely (0.00%)**, leaving *Acropora* 8.50% and *Porites* 4.40% to lead a 15-genus assemblage with *Favia* 2.70%, *Favites* 1.10% and *Goniopora* 0.80%. D5, the deepest station, recovers to 38.0% coral against 29.0% algae but on only 8 genera, and **the lead changes hands to the massive faviids: *Platygyra* 9.00% (24% of that transect's coral) and *Goniastrea* 6.00%**, with *Porites* 8.00%, *Acropora* 5.00%, *Cyphastrea*, *Leptastrea* and *Montipora* 3.00% each and *Pocillopora* 1.00%. **The site-level figures of 6.48% *Millepora* and 2.48% *Platygyra* describe no transect whatsoever** — each is a four-station mean of one large value and three near-zeros. All four depth rows account for 100% of the benthos with no shortfall.

The photographs corroborate the *Millepora* finding directly and forcefully. **Roughly half of the sampled frames are dominated by extensive, near-monospecific thickets of pale cream-to-tan branching and bladed colonies whose growth form — smooth, blunt-tipped, anastomosing branches and upright fan-like plates — is characteristic of *Millepora*** rather than of any scleractinian in the genus list. P4300978 is the single clearest frame: yellow-green upright bladed colonies consistent with *Millepora platyphylla* beside cream palmate colonies, above a floor of pale branching thicket. P4300950, P4300985, P4300989 and P4300991 each show this thicket carpeting the entire field of view to the limit of visibility, and P4300936 and P4300939 show the same habitat at closer range. P4300987 adds the other conspicuous element: a broad tan corymbose *Acropora*-type field interleaved with yellow-green *Millepora* blades. The frames that break this pattern show the mixed, massive-colony habitat that matches D2 and D5 — P4300941, P4300943 and P4300962 record open pavement and sand carrying lobate and helmet-shaped massive colonies consistent with *Porites*, brain-patterned colonies consistent with *Platygyra* or *Goniastrea*, encrusting faviids, tabular and corymbose *Acropora* and pink crustose coralline crust. P4300960 shows a stack of overlapping tabular colonies at the reef edge. Image-based genus calls remain tentative and no coverage figure is derived from them, but the correspondence between a *Millepora*-dominated photographic set and a transect in which *Millepora* is the leading taxon is unusually clean.

**Condition reads as moderate.** The published image-based bleaching figure is **8.7%, ranking 39th of 98 Chagos sites carrying a value — above the archipelago median but far from the extreme end.** Colour interpretation is genuinely difficult in this set, because *Millepora* is naturally pale cream-to-yellow and a *Millepora* thicket photographed under ambient blue light reads as near-white regardless of its condition. For that reason **no bleaching call is made from the wide *Millepora* thicket frames**, which is precisely where a colour-based CV annotator would be most likely to over-read pallor. In the mixed-habitat frames where naturally pigmented scleractinians are visible against each other, colour looks normal: the tan, olive, brown, purple and yellow pigments in P4300941, P4300943, P4300962, P4300971 and P4300987 are what healthy tissue looks like, and no colony there shows the stark within-frame white-against-pigmented contrast that marks unambiguous bleaching. A pale table with a hole in it appears at the left edge of P4300971; this is read as physical damage or an old dead area rather than bleaching, and is not resolvable from a single still. **No disease lesions, tissue-loss bands, *Drupella* aggregations, crown-of-thorns starfish or predation scarring were seen in any of the 17 frames reviewed.**

Fauna is conspicuous. **Three separate whitetip reef shark encounters are documented — P4300973, P4300975 and P4300993** — each an individual cruising low over the thicket, which together with the site's isolation reads as an intact predator presence. Dense clouds of dark damselfishes and chromis hang over the reef in P4300936 and P4300987, and the sampled frames also carry moorish idols, butterflyfishes, parrotfishes and small anthias. A diver silhouette is visible at the surface edge of P4300960 — standard field documentation.

<b>Assemblage composition.</b> Formal transects give 34.7% coral against 42.0% algae across 26 nonzero genera, led by three near-equal taxa — *Acropora* (7.19%), *Porites* (6.90%) and the hydrocoral *Millepora* (6.48%) — which together hold 59% of the coral, with the Chagos-endemic *Ctenella* present in trace amounts. The depth profile is extreme: cover runs 20.5% at D2, 59.22% at D3, 21.2% at D4 and 38.0% at D5, and the entire *Millepora* signal sits at D3, where it reaches 25.70% of the benthos and 43% of that transect's coral before vanishing to zero at both deeper stations; D5 hands the lead instead to *Platygyra* (9.00%) and *Goniastrea* (6.00%). The photographs match this closely, with roughly half the sampled frames showing near-monospecific *Millepora* thicket and the remainder showing mixed massive-colony pavement over pink coralline crust, plus three whitetip reef shark encounters and no sign of disease, predation or unambiguous bleaching.

**Flags:** (1) **Two published reef-classification cells are BLANK for this site** — `reef type` and `reef type-simplied` are empty in `Chagos benthic data by site (n=109).csv`, and `exposure-original` is empty as well. The header above therefore reports only what exists: `reef location` = forereef, `lagoon` = outside, `reef emergence` = submerged, `exposure-revised` = protected. **No reef class was invented to fill the gap.** (2) **The site-level *Millepora* figure (6.48%) is an artefact of averaging and must not be read as a description of the reef.** *Millepora* is 0.20% at D2, **25.70% at D3**, 0.00% at D4 and 0.00% at D5. Any site-level analysis that treats 6.48% as the typical *Millepora* cover here will be wrong at every depth actually surveyed. The same applies to *Platygyra* (0, 0.90, 0, 9.00 → 2.48%) and *Goniastrea* (0.10, 0.42, 0.30, 6.00 → 1.71%). (3) **Genus richness is almost entirely a D3 phenomenon:** D3 carries 24 of the site's 26 genera, against 11 at D2, 15 at D4 and only 8 at D5 (the two genera absent from D3, *Goniopora* and *Symphyllia*, occur at D4 only). **Seven genera — *Astreopora*, the endemic *Ctenella*, *Echinopora*, *Galaxea*, *Gardineroseris*, *Leptoseris* and *Psammocora* — occur on the D3 transect and nowhere else at this site**, so the site's genus count leans heavily on a single transect. (4) ***Ctenella chagius* is a Chagos endemic and is recorded here at 0.05% of the benthos (D3 only).** Flagged for anyone assembling an endemic-species occurrence layer. (5) **Bleaching cannot be scored from the photographs at this site, and the published figure deserves extra scepticism here.** *Millepora* is naturally pale cream-to-yellow, and roughly half the imagery is wide-field *Millepora* thicket that reads near-white under ambient blue light. A colour-based annotator is at elevated risk of over-reading pallor in exactly this habitat. No bleaching call was made from those frames. (6) **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. Any rank taken off that column is approximate: it carries at least one known bad value (BITB58 listed at 22.22% against a corrected 0.77%). (7) **Multi-method agreement is unusually GOOD here and is worth recording as such.** Coralscapes masks (32.7%) sit only 2.0 pp below the formal-transect figure (34.7%), with CoralNet points at 24.5% (published deviation range −10.2 to −2.0 pp) — far tighter than the 20–38 pp gaps seen at several other Chagos sites. This is not licence to average them: the three products still have different sampling footprints and denominators, and the transect value remains ground truth for cover. (8) **`n_photos` = 33 for the CV comparison, which equals the jpg count in the folder.** That coincidence does not make the CV figures comparable with a frame-by-frame photo review; the denominators still differ, and only 17 of those 33 frames were examined here. (9) **One video file in the folder is 32 bytes long** — `P4300977.MOV` is effectively empty and appears truncated or corrupt. It was not reviewed (this is a still-photo review), but the zero-content file is worth checking against the original media. (10) **No COTS, no disease, no *Drupella* and no predation scarring were observed** across the 17-frame sample — an explicit absence-of-evidence finding, bounded by the fact that 16 frames were not examined. (11) **Three whitetip reef shark encounters (P4300973, P4300975, P4300993)** are logged for the fauna record. (12) **Formal replication is `N Rows` = 4**, and there is no D1 transect, so the shallowest reef zone is unsampled. The depth codes D2–D5 carry no metre boundaries in `Chagos benthic data by site-depth (n=336 rows).csv`; they are ordered shallow-to-deep stations, and the 31.2 m in the header is the site-level survey depth, not the depth of any one transect. (13) **Bare substrate swings widely with depth (PB 2.8% at D2, 6.1% at D3, 33.3% at D4, 24.0% at D5)**; a third of the D4 transect is open uncolonised ground. (14) **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 **8.73%, rank 47 of 110** (1st training: 8.73%, rank 46 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 BIVB109 benthic survey data (CSV)](../benthic_data/BIVB109_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).