### BIEA20 — Eagle Island
**Date:** 2015-03-15 (dive window 14:04–14:55, from frame EXIF — the third and last Eagle Island dive of the day, after BIEA18 at 08:11 and BIEA19 at 10:41)
**GPS:** 6.1899°S, 71.3633°E
**Frames reviewed:** 41 of 41 (100% of the site folder, `BiEa20-finished`)
**Coral cover:** 10.8% (formal transect surveys) · **% bleached:** 4.6% (CoralNet analysis of informal survey images; Mayfield & Dempsey 2026, Oceans)
**Reef type:** bank, exposed · **Temperature:** 29.2 °C
**Live hard coral cover across methods** (n=41 photos): transect 10.8% · CoralNet points 11.5% · Coralscapes 9.3%. Deviation vs transect −1.5 to +0.7 pp. **Mean of 3: 10.5% ± 1.1** — **the closest three-method agreement of any site in the redo.** But that agreement is partly luck: **restricted to the 25 benthic frames (<40% water), the photo methods read 16.0% and 11.9%, i.e. +5.2 pp and +1.1 pp ABOVE the transect.** See the flags.
**Coral genera identified:** 7 scleractinian groups + two octocoral orders — *Acropora* (corymbose, digitate and small tabular), *Porites* (massive/domal, columnar-digitate **and a large yellow-tan foliose form**), *Pocillopora*, ridged foliose plates compatible with *Echinopora*/*Merulina*/*Pachyseris*, meandroid brains compatible with *Platygyra*, encrusting sheets compatible with *Montipora*, and assorted small submassives not resolvable to genus; plus **black corals (antipatharian whips and bushes) in at least five frames**, bushy soft corals compatible with *Sinularia*/*Dendronephthya*, and the calcified alga *Halimeda*. Growth-form calls from photographs, given as "compatible with".
**Representative photos:** P3150248.jpg, P3150250.jpg, P3150279.jpg, P3150281.jpg, P3150311.jpg

**BIEA20 is the fish site.** More than fifteen of its forty-one frames are given over to **enormous schools of dark surgeonfish and unicornfish** — P3150283 through P3150289, P3150302 through P3150313 — banked against the blue, funnelling down onto the reef, and in P3150310 grazing right on the substrate. Set against that biomass, the reef beneath is thin: a **low-relief plain of pale consolidated carbonate, rubble and sand** across which coral sits as small, well-spaced colonies. P3150299, P3150313 and P3150262 show how far that plain runs. The 10.8% transect figure is entirely believable from the imagery, and the frames explain the site's character better than the number does — **this is a heavily grazed, structurally simple bank carrying a very large herbivore population.**

Where coral does concentrate it is worth looking at. P3150276, P3150280 and P3150292 carry **large yellow-tan foliose and columnar *Porites***; P3150290, P3150297 and P3150298 show **brown ridged foliose plates** compatible with *Echinopora* or *Merulina* spreading over the pavement; P3150291 has a tan meandroid brain compatible with *Platygyra*; and P3150279 — the site's highest-cover benthic frame at 46.7% live by point count — is a stand of **pale tan digitate *Porites*** with a **spotted moray eel (*Gymnothorax* sp.)** looking out from between the branches. **Black corals are a quiet but repeated feature**: antipatharian whips in P3150248 and P3150272, and bushes in P3150262 and P3150274.

**The macrofauna is the best of the redo so far.** **P3150248 is the finest shark frame in the collection** — a **whitetip reef shark (*Triaenodon obesus*)**, close and sharply lit, cruising over white sand beneath a maroon-crusted overhang, its white dorsal and caudal tips unmistakable. A whitetip appears again on the slope in P3150262, and **P3150259 holds a second, different requiem shark in open blue water** — dark caudal margin but a pale first-dorsal tip, so **grey reef (*Carcharhinus amblyrhynchos*) or silvertip (*C. albimarginatus*); it is not callable from the frame**, and the BIEG04 lesson (Chagos "whitetips" that proved to be silvertips) argues for leaving it hedged. Beyond the sharks: a **giant clam (*Tridacna*)** with its mantle open and boldly patterned in P3150250, an adult **emperor angelfish (*Pomacanthus imperator*)** in P3150274, a **two-spot red snapper (*Lutjanus bohar*)** in P3150308, a **coral hind (*Cephalopholis miniata*-type)** in P3150307, **powder-blue surgeonfish (*Acanthurus leucosternon*)** in P3150301, a ***Diadema*-type urchin** in P3150290, blue-and-yellow damselfish swarming P3150300, and black triggerfish, hogfish, cleaner wrasse and anthias throughout.

**This is also the mission's best-documented methodology site, and that is worth recording deliberately.** Eight frames show the survey being made: **P3150273** — a camera housing and strobe arm in the foreground aimed squarely at a massive *Porites* dome, a diver with a slate behind it (the frame the old description meant by "two divers photographing a coral colony"); **P3150275** — a diver with a camera and, lying on the reef beside them, **a second twin-cylinder set**; **P3150303 and P3150305** — divers with housings and twin strobes filming the fish schools; **P3150277 and P3150278** — divers working over the rubble plain; and two frames that may be the most useful of all — **P3150311 carries a distinct white rectangular open frame lying on the substrate at lower right, with a thin pale line running away from it across the reef**, and **P3150301 shows a similar rectangular frame silhouetted in the background**. If those are a **photoquadrat and a transect tape**, this site holds a direct image of the apparatus behind the 10.8% figure. **Both need checking at full resolution before the claim is made.**

**Mortality is localised but real.** **P3150281 is the single most degraded frame at the site**: a branching colony gone **chalky grey-white**, corallite texture still sharp but no live tissue colour, ringed by dark maroon crusts and with green algal film beginning to take it. Coralscapes reads **10.85% dead** here, by far its highest at BIEA20, and the human annotation puts two bleached points on it. Whether this is terminal bleaching or coral already dead a few weeks is not resolvable from the photograph, but it is not a healthy pale morph. P3150292 holds a second, much smaller white patch.

<b>Assemblage composition.</b> Across 41 analysed frames the Coralscapes masks put live hard coral at 9.3% of image area, of which massive and meandroid is the largest component at 37.1%; the rest breaks down as other hard coral 25.9%, tabular <i>Acropora</i> 14.5%, <i>Pocillopora</i> 9.8%, <i>Acropora</i> (non-tabular) 8.1% of live cover. The independent 30-point human annotation of the same frames (1230 points) gives live hard coral as 11.5% of all points and apportions it <i>Porites</i> 36.6%, <i>Acropora</i> 35.9%, other hard coral 22.5%, <i>Pocillopora</i> 4.9% of live coral. Non-coral cover is dominated by crustose coralline algae 3.2%, turf 15.6% of all points. Structural loss registers as dead coral 0.8% of image area and rubble 4.6%. Bleaching stands at 1.1% of annotated points — **13 points in total** — concentrated in <i>Acropora</i> 0.6%, other hard coral 0.3%, <i>Pocillopora</i> 0.2% of points. Frame-to-frame live cover runs 0% to 46.7%. Conspicuous fauna and sessile invertebrates appear as sponges in 16 frames, urchins in 2 frames, giant clams in 1 frame, sea cucumbers in 1 frame, anemones in 1 frame.

**Flags:** (1) **THE HEADLINE THREE-METHOD AGREEMENT IS PARTLY AN ARTEFACT.** ±1.1 pp is the tightest in the redo, but 16 of 41 frames are ≥40% water. **Drop them and the photo methods rise to 16.0% (CoralNet) and 11.9% (Coralscapes) against a 10.8% transect — both now reading HIGH, CoralNet by +5.2 pp.** This is the same sign-flip found at BIEA18 and it is larger here. **At low-cover sites blue-water dilution is the dominant bias, and it happens to cancel against the transect at BIEA20 rather than agreeing with it.** Do not cite this site as evidence that the methods converge. (2) **3 of 13 bleached points (23%) sit on frames ≥40% water.** The clearest problem is **P3150302: 73.3% water, 3.3% live coral, a distant hazy rubble ridge — and 2 bleached-*Acropora* points.** Nothing in that frame supports a bleaching call. **P3150308** (2 points) and **P3150301** (1 point) are weaker versions of the same. (3) **WATER SCORING IS INTERNALLY INCONSISTENT AT THIS SITE.** **P3150283 is roughly 60% open blue water yet is scored WATER 0.0% with 18 points as "Unknown"**, while P3150259 and P3150289 are correctly scored at 96.7% and 93.3%. **P3150306** is mostly blue but scored 16.7%. **Any water-based frame filter will behave unpredictably here** — fix the water scoring before filtering. (4) **METHODOLOGY IMAGERY, EIGHT FRAMES** — P3150273, P3150275, P3150277, P3150278, P3150301, P3150303, P3150305, P3150311. **P3150311 and P3150301 appear to show a photoquadrat frame (and possibly a transect tape); verify at full resolution.** This is the densest methodology record of any site in the redo, and it belongs with BIDI08/P3120298, BIEG04/P3100401, BIEA18/P3150341 and BIEA19/P3150369. (5) **P3150281: chalky grey-white branching colony, Coralscapes 10.85% dead** — the site's most degraded frame; terminal bleaching vs recent death is unresolved. (6) **P3150248 is the best shark image in the collection** (whitetip reef shark, close and sharp) — flagged for illustrative use. **P3150259's shark is NOT callable** (grey reef vs silvertip); do not repeat the old description's confident "a whitetip reef shark and a separate grey reef shark". (7) **Giant clam (P3150250), spotted moray (P3150279), emperor angelfish (P3150274), two-spot red snapper (P3150308), coral hind (P3150307)** — the macrofauna list the old description started, now completed and frame-referenced. (8) **Black corals (antipatharians) in ≥5 frames** — as at BIEA18's gorgonians, octocorals and antipatharians are structurally present but invisible to a hard-coral point scheme. (9) **P3150310 is the best herbivory frame** — a surgeonfish school grazing directly on the substrate. (10) No per-frame depth data.

**Benthic data:** [Download BIEA20 benthic survey data (CSV)](../benthic_data/BIEA20_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).