Dive 3 — Re-occupation and the Noise Floor

Safety and gas bind before the survey does. Adapt to the conditions or stop — a partial ring is always a valid outcome.

Dive 3 is the dive that turns a survey into a measurement. The team returns to the hub surveyed in Dive 2, re-establishes the same north, and runs the same six-station ring a second time — about twenty-four hours later, on Day 2. On a reef that cannot have changed meaningfully overnight, the difference between the two readings is not biology; it is the survey's own noise floor — how much the number moves when nothing real has. That spread is the error bar for everything the method will ever report: a future survey counts as real change only when it moves by more than this floor. Without this dive the course would teach a survey but never establish what a survey result means.

Why about twenty-four hours. Re-running the ring an hour after Dive 2 would capture only short-term counting scatter. Re-running it the next day lets the noise floor span a full tidal cycle and the day-to-day movement of the resident fish, which is a more honest estimate of the repeatability that a seasonal or annual comparison will actually face. The reef's residents move and mill; the structure does not. A 24-hour repeat on unchanged structure therefore isolates measurement-and-sampling noise from real ecological change.

What re-occupation requires. Three things must be re-established, in order:

1. The same hub. The hub is the fixed point the whole ring is built around, so it must be relocated, not chosen afresh. It is re-found from the surface GPS fix taken at descent and from a recognizable fixed natural feature noted on Dive 2 — never from a marker left on live substrate, which the course forbids. Where the site is mapped under Reef Cartographer, the structure map locates the hub directly.

2. The same north. Every bearing is read from magnetic north, which is stable day to day, so the same compass headings re-create the same ring orientation. Set and confirm north exactly as on Dive 2.

3. The day's calibration. Visibility is re-measured for Day 2's conditions — today's kick count K and today's radius r — because a kick cycle is not a fixed distance and the water may differ. The ring is then R = 2r for today's r.

On comparable conditions. The cleanest noise floor comes from re-surveying under conditions like Dive 2's — similar visibility, similar time of day. If today's visibility is close to yesterday's, today's ring (R = 2r) is nearly the same size, so the six stations fall on nearly the same spots and the comparison is between like and like. If visibility differs substantially, the density comparison is still valid — density is per-area and the r/K normalization removes the kick-length difference (Appendix A.7) — but the ring then covers a different-sized patch, which adds to the apparent spread. Record the conditions for both dives; a large visibility difference is itself noted as a caveat on the noise floor.

Recognizing the stations. The resident, site-attached organisms the method counts are also the check that the team is in the right place: a station that held a recognizable resident on Dive 2 should hold it again on Dive 3. This biological site-fidelity — the same premise that makes re-occupation meaningful at all — doubles as confirmation of station identity.

Performance requirements. By the end of this dive the student will have:

1. Re-located the Dive 2 hub from the surface GPS fix and a recognized fixed feature, with no marker placed on live substrate.

2. Re-established the same north and re-run the complete six-station ring (out 2r on 000°, then the +60° edge headings), a panorama at each center.

3. Re-measured the day's calibration (K, r) and applied R = 2r for today.

4. Produced a reconcilable two-observer count at each station, counting only reef-associated organisms.

5. Computed (topside) the six densities, the site index D̄, and the spread between the Dive 2 and Dive 3 readings — the noise floor.

Equipment for this dive. As Dive 2: a nitrox-set computer, the 360° camera, a compass, and the slate carrying the Dive 2 hub reference, north, K, r, and the heading table; plus the marked line for the Day-2 visibility calibration. No reel for the survey; no marker placed on live substrate.

Pre-dive briefing. Review how the hub will be re-found (the GPS fix, the fixed feature noted on Dive 2, the Reef Cartographer reference if available), and confirm that north will be set identically and that visibility is re-measured today (R = 2r re-derived). Confirm the gas plan and the no-contact, no-marker discipline. Brief the comparison explicitly: the two readings will be differenced to estimate the noise floor, so both dives must be run the same way.

The dive.

1. Take the surface GPS fix; descend to the re-located hub; establish neutral buoyancy and trim; set and confirm north.

2. Re-calibrate. Deploy the marked line; obtain today's K (kick count over the line) and today's r (marker fade); derive R = 2r.

3. Run the ring. Out 2r on 000° to center 1, panorama; then add 60° (120°), swim 2r, panorama; repeat through 180°, 240°, 300°, 000° — six panoramas; the closing 060° leg returns the team to center 1.

4. The recorder keeps the in-water tally; recognizable residents at each station confirm its identity.

5. Ascend with a safety stop.

Post-dive. Offload and count the six panoramas, reconciling against the in-water tally as on Dive 2. Compute the six densities, the Dive 3 site index, and compare it with the Dive 2 site index. The difference between them — read together with the within-dive station spread — is the noise floor: the magnitude of variation the method produces on unchanged structure. Record it explicitly, because it becomes the decision rule for the whole survey program: a later reading is treated as real change only when it exceeds the noise floor. Debrief the re-occupation (was the hub re-found cleanly? did residents confirm the stations?), the count reconciliation, and the size of the noise floor relative to the readings.

Assessment. The dive is satisfactory when the student has re-located the hub and re-established north, re-run the complete ring under today's calibration, produced reconcilable counts, computed both site indices and their difference, and can state in plain terms what the resulting noise floor means — that a change smaller than this spread cannot be distinguished from measurement noise.

Depth range. 6–13.5 m / 20–45 ft (Day 2), consistent with the gas plan (gas and no-decompression time both hard limits, either reached first).