### BIEA22 — Eagle Island
**Date:** 2015-03-16 (dive window 08:25–09:11, from frame EXIF — first dive of 16 March, ahead of BIEA23; 20 of the 23 frames fall in one 12-minute burst, 08:58–09:11)
**GPS:** 6.27659°S, 71.27627°E
**Frames reviewed:** 23 of 23 (100% of the site folder, `BiEa22-finished`)
**Coral cover:** 21.6% (formal transect surveys) · **% bleached:** 6.9% (CoralNet analysis of informal survey images; Mayfield & Dempsey 2026, Oceans)
**Reef type:** bank, intermediate · **Temperature:** 29.1 °C
**Live hard coral cover across methods** (n=23 photos): transect 21.6% · CoralNet points 42.8% · Coralscapes 46.0%. Deviation vs transect **+21.2 to +24.4 pp** — **the largest method divergence in the redo, both photo methods reading roughly twice the transect.** **Mean of 3: 36.8% ± 13.3.** The frames explain it: **only 4 of 23 are reef-scale wide views; the other 19 are close-focus frames built around one octopus, 11 of them filled edge-to-edge by a single coral head.** On the four wide frames the methods land almost on the transect — **25.0% and 23.2% against 21.6%** — while the 19 close frames read 46.5% and 50.7%. See the flags.
**Coral genera identified:** 7 scleractinian groups — ***Porites*** in four growth forms (columnar-nodular heads, massive/domal boulders, encrusting-plating sheets and pale cream digitate colonies), tabular and laminar plates compatible with ***Acropora***, corymbose *Acropora*-type colonies, bushy verrucose colonies compatible with ***Pocillopora***, a vermiform meandroid brain compatible with ***Platygyra***/*Leptoria*, smooth brown encrusting plates compatible with ***Montipora***, and ridged foliose plates compatible with ***Echinopora***/*Merulina*/*Pachyseris*. Non-scleractinian cover is unusually conspicuous: **red, orange and pale-pink encrusting sponges**, **ascidians compatible with *Didemnum molle***, **hydroids compatible with *Aglaophenia*/*Lytocarpia***, and ***Halimeda***. Growth-form calls from photographs, given as "compatible with".
**Representative photos:** P3160314.jpg, P3160323.jpg, P3160353.jpg, P3160356.jpg, P3160365.jpg, P3160372.jpg

**BIEA22 is the octopus site, and that fact governs everything the numbers say about it.** A large day octopus compatible with ***Octopus cyanea*** occupies **16 of the 23 frames** — P3160344 through P3160372 — and the camera stays on it for eleven straight minutes, moving from arm's length to near-macro. The animal cycles through its repertoire: flat near-black over pavement in **P3160344** and **P3160353**, deep maroon with the pale funnel showing in **P3160351**, **P3160357**, **P3160358** and **P3160359**, rust-red with a raised ocular papilla in **P3160349**, **P3160361** and **P3160369**, wedged into a fissure in a huge salmon-pink *Porites* head in **P3160366**, and fully papillate in **P3160356** and **P3160372**. **P3160359 is the best portrait**; **P3160356 the best camouflage frame**, the octopus almost invisible against a *Porites* head that fills the whole picture.

**The reef itself is best read from the four wide frames, and it is modest.** **P3160342** and **P3160343** look down over a **low-relief pavement-and-rubble plain** with coral as scattered small colonies — laminar and tabular plates compatible with *Acropora*, small corymbose colonies, pale massive *Porites* domes, mauve encrusting sheets, pink crustose coralline algae, turf and fine green algae between them. **P3160323** is the same plain at distance in poor light, coral visibly thin. **P3160314** is the highest-cover and the only frame with real structure: cream *Porites* plates stepping down a slope, a large tan massive *Porites* right, brown ridged plating coral in the foreground, white sand channels through it. Across these four the point count gives 25.0% live coral — **entirely consistent with the 21.6% transect**, and the reason to trust the transect over the folder mean.

**The macrofauna is thin but holds the site's other headline.** **P3160323 shows at least two, probably three, requiem sharks** over the plain in mid-water; outline and caudal shape fit **grey reef shark (*Carcharhinus amblyrhynchos*)**, but the frame is dim and distant and, following the BIEG04 lesson, **the species call should not be made from this image**. The same frame carries a **giant clam (*Tridacna*)** at lower right, mantle open. **P3160365 is the best fish frame**: a small **bright red serranid, densely white-spotted with pale-yellow blotches and an orange-ringed eye — compatible with a *Cephalopholis* (coral hind, *C. miniata*-type)** — among bushy *Pocillopora*-type colonies. Red soldierfish- or squirrelfish-type fish sit in the shadows of **P3160356**, and yellow butterflyfish are just resolvable behind **P3160314**.

**Mortality and bleaching are the weakest part of the record, and the imagery does not support the point data.** No frame shows a chalky, tissue-free colony of the kind BIEA20's P3150281 does. What they show is **pale-but-pigmented coral**: the cream-tan digitate and nodular *Porites* dominating the octopus sequence is naturally light-coloured, and in **P3160367** and **P3160368** — extreme macro of one branching colony — the pale look is **white growth tips and fully expanded white polyps over healthy tan coenosarc**, not bleached tissue. Those two frames carry 9 of the site's 24 bleached points; **P3160353** carries 6 more, again on pale cream *Porites*. Coralscapes puts bleaching at **0.06%** of image area site-wide. The transect-derived 6.9% figure is retained above as the formal value, but the photographic record neither confirms nor illustrates it.

<b>Assemblage composition.</b> Across 23 analysed frames the Coralscapes masks put live hard coral at 46% of image area, of which massive and meandroid is the largest component at 42.3%; the rest breaks down as other hard coral 39.6%, tabular <i>Acropora</i> 11.3% of live cover. The independent 30-point human annotation of the same frames (690 points) gives live hard coral as 42.8% of all points and apportions it <i>Porites</i> 46.4%, other hard coral 41.4%, <i>Acropora</i> 10.8%, <i>Pocillopora</i> 1.4% of live coral. Non-coral cover is dominated by crustose coralline algae 4.3%, turf 12.2%, macroalgae 4.2% of all points. Structural loss registers as dead coral 1% of image area and rubble 1.5%. Bleaching stands at 3.5% of annotated points — **24 points in total** — concentrated in other hard coral 2.9%, <i>Acropora</i> 0.6% of points. Frame-to-frame live cover runs 6.7% to 90.0% by point count. Conspicuous fauna and sessile invertebrates appear as sponges in 15 frames, anemones in 9 frames, giant clams in 2 frames.

**Flags:** (1) **THE +21 pp DIVERGENCE IS A FRAMING ARTEFACT, NOT DILUTION.** The dilution mechanism flagged at BIEA18 and BIEA20 cannot apply: **not one BIEA22 frame reaches 40% water — the maximum is 33.3%, the site mean 5.7%.** With no blue-water frames to pull the estimate down, the folder mean simply reports what the photographer pointed at, and **splitting the frames resolves the bias cleanly (wide 25.0%/23.2% vs close 46.5%/50.7%).** BIEA22 is **not** evidence that photo methods over-read at high cover — it is evidence that **a 23-frame folder built around one animal is not a cover sample.** **n is effectively 4, not 23** — sixteen frames contain the same octopus, eleven are filled by a single coral head — so **any site-level cover statistic from this folder should be flagged non-random and recomputed on the wide frames only.** (2) **THE BLEACHING POINTS ARE NOT SUPPORTED BY THE IMAGERY.** All 24 bleached points sit on 8 close-focus frames; **zero fall on the four wide reef frames.** **P3160368 (4) and P3160367 (5) score expanded white polyps and white growth tips on a healthy tan branching colony as bleached; P3160353 (6), P3160355 (3), P3160352 (2) and P3160354 (2) score naturally pale cream *Porites*.** Coralscapes independently reads 0.06% bleached area site-wide. **Recommend targeted re-annotation of these 8 frames before the 6.9% figure is reused in a photo-based context.** (3) **P3160372 IS THE WORST SINGLE-FRAME DISAGREEMENT — human 43.3% live coral vs Coralscapes 3.2%.** The frame is red and orange encrusting sponge, pavement and the camouflaged octopus, with almost no live scleractinian. **The human value is wrong.** (4) **P3160372 IS NOT A SCORPIONFISH.** The old description flagged "a well-camouflaged scorpionfish- or stonefish-like animal"; the horizontal slit pupil in a raised red ocular papilla, the papillate skin and the cream funnel below it are **cephalopod characters matching the octopus photographed continuously for the preceding eleven minutes**. **Worth one full-resolution check before the correction is published.** (5) **P3160314 CARRIES A THIN DARK LINE RUNNING VERTICALLY THROUGH THE FRAME** from the top edge onto the reef, with a fainter second at upper left. **If it is a shot line or transect tape it is the site's only methodology imagery** — the sole BIEA22 candidate for the BIDI08/BIEG04/BIEA18/BIEA19/BIEA20 series. **Verify at full resolution; it could equally be discarded monofilament.** (6) **P3160323: two, probably three, requiem sharks** — grey reef (*Carcharhinus amblyrhynchos*) fits best but **the frame is too dim and distant to call**; do not make a confident species claim. Same frame holds a **giant clam (*Tridacna*)**. (7) **P3160365: small red serranid compatible with *Cephalopholis miniata*** — white rather than blue spotting is atypical; **confirm before using the name.** (8) **Ascidians compatible with *Didemnum molle* appear in six frames**, scored variously as CCA, "other" or Unknown — **the point scheme has no tunicate class**, as with BIEA18's octocorals. **Unknown runs at 23.8% of all points site-wide**, peaking at 56.7% (P3160346). (9) No COTS, no disease lesions, no predation scars and no chalky recent-mortality colonies in any frame. (10) No per-frame depth data.

**Benthic data:** [Download BIEA22 benthic survey data (CSV)](../benthic_data/BIEA22_benthic.csv) — scleractinian cover by genus with full names, plus substrate and algal categories (bare substrate, CCA, turf, macroalgae, invertebrates), from the LOF BIOT-1 survey (Mayfield & Dempsey 2026, *Oceans*, supplemental file S1).