A trained-Steward field protocol for DuskWatch sessions at the Kakadu Beach Constructed Roost. The same fields appear in the in-app form. Sessions can be recorded on paper and entered later, or directly in the WebApp.
Before you go
Date selection
High tide within 125 minutes of civil twilight, ideally 30–60 minutes apart. Priority for dates within 2–3 days of full moon, which allows arrival counting to continue past civil twilight by moonlight.
What to bring
Binoculars (8×42 standard issue). Spotting scope optional. Red-filter head-torch for note-taking after dark — white light disturbs roosting birds. Recording form (paper or in-app). Watch synced to Australian Eastern Standard Time. Tide and twilight times from BoM (Caloundra Bar tide reference, Sandstone Point sunrise–sunset).
Two-person rule
Sessions running past civil twilight require a two-volunteer minimum. The after-dark return-to-vehicle pathway is agreed at training.
At the observation point
Arrival
Arrive 60 minutes before sunset. Set up at a fixed position from which the open central area of KBR and at least one of the mangrove end-zones is visible. Record start time, conditions, and position on the session header.
Session window
Run from session start (sunset −60 min) through to civil twilight +30 minutes, or until two consecutive 5-minute periods register no arrivals.
What you record
The form has four parts.
A. Session header
Filled at session start.
| Date | YYYY-MM-DD |
|---|---|
| Observer | Your steward pseudonym (auto-generated in the app) |
| Position | Open-ground viewpoint / mangrove-end viewpoint / camera segment ID |
| Sunset / civil twilight | Local time, BoM Sandstone Point |
| High tide | Time and height in metres, BoM Caloundra Bar |
| Tide–twilight gap | Minutes between high tide and civil twilight (signed: negative if tide is before twilight) |
| Moon phase | % illuminated, rise time, cloud cover at session start |
| Wind / weather | Speed (Calm / Light breeze / Windy), direction (N…NW), cloud (eighths), rain |
| Pressure (now) | Barometric pressure in hPa at session start, from BoM Caloundra Aero or Brisbane Aero |
| Pressure (6 h ago) | Pressure 6 hours before session start, in hPa, same BoM source |
| Pressure (24 h ago) | Pressure 24 hours before session start, in hPa, same BoM source |
| Session window | Start time (sunset −60 min), planned end (civil twilight +30 min) |
B. Time-block counts
Every 30 minutes, in line with the published Sanders and Handmaker resolution. Within each block, log every arrival or departure event with these fields:
- Time of event (HH:MM)
- Direction of first detection (N / NE / E / SE / S / SW / W / NW)
- Altitude of approach (low over water / mid / high)
- Flight pattern (direct / circling / ringing)
- Species (or unidentified)
- Flock size (count or bracket: 1–5, 6–20, 21–50, 51–100, 100+)
- Landing zone (open ground / mangrove edge / mangrove canopy / not-settled)
At the end of each 30-minute block, sum net arrivals by species. The in-app form does this automatically once events are entered.
C. Behaviour scans
At three fixed time-points — sunset, civil twilight, civil twilight +30 minutes (where light permits) — scan the settled flock and record, for the focal species (Terek), an estimated percentage in each of the postures below.
| Posture | What it looks like | What it indicates |
|---|---|---|
| Head-up | Head and neck raised, eyes open, scanning | Active vigilance. Common during arrival window or after disturbance. |
| Head-forward | Head held forward at shoulder height, body relaxed | Resting but alert. The default state at a settled roost. |
| Bill-tucked (back-rest) | Head turned back, bill placed on the back and covered by plumage | Deep rest. Often paired with one-leg standing. |
| One-leg standing | One leg raised into plumage, the other supporting | Heat conservation in cooler conditions. Often paired with bill-tucked. |
Plus three additional behaviour categories: preening, walking or re-positioning, and out-of-view (in mangrove canopy). Percentages should sum to 100 within each scan.
Optional fourth scan at full dark, by ear only: calling activity present, absent, or continuous.
D. Disturbance log
A new row for every disturbance event during the session.
| Time | Source | Distance (m) | Response | Resettle (min) |
|---|---|---|---|---|
| HH:MM | Dog / walker / boat / jet-ski / drone / raptor / corvid / human swimmer / other | Estimate to nearest 25 m against landmarks | None / Head-up only / Walk-away / Partial flight / Full alarm flight | Time from flight to settled, with destination if relocated |
E. Flag observations
If the scope shows individually marked birds in the settled flock during the behaviour-scan window, log them here. Standard QWSG / AWSG fields apply.
| Field | What to record |
|---|---|
| Time | HH:MM of sighting |
| Species | Terek Sandpiper or other (species list as for arrival events) |
| Leg position | Left tibia / right tibia / left tarsus / right tarsus / both / unclear |
| Flag colours and arrangement | e.g. white plain, yellow over orange, green engraved |
| Engraved code | e.g. AAJ, H21, T8 — only if readable |
| Notes | Plumage state, photo file reference, anything that helps verification |
Flag-marked birds are read with the spotting scope while the flock is settled and visible — usually in the early scans before light drops. Engraved codes need 30× or 40× and good light.
Submit resightings beyond DuskWatch. The flyway-wide colour-marked wader portal is BirdMark, which holds banding and flagging data from VWSG, AWSG, Global Flyway Network, NTWSG and QWSG. Queensland-specific reports go to the QWSG flag-sighting coordinator at QWSG leg-flag observation. Records entered into the DuskWatch form remain in the project's own dataset for site-specific analysis; submission to BirdMark/QWSG is the second step that closes the flyway loop.
After the session
If you used the WebApp, the record queues to your device and syncs when the device next has connectivity. If you used paper, enter the record into the WebApp at the next opportunity. Paper sheets are retained as a backup.
The 30-minute net-arrival rows are the input format the Sanders/Handmaker R analysis expects. The behaviour-scan rows feed the settling-curve. The disturbance rows feed the alarm-flight-rate calculation, with the Lilleyman et al. (2016) coefficient applied to give a per-session and per-season energy-cost estimate. The pressure-trend rows allow a test of whether arrival at KBR correlates with falling barometric pressure ahead of foul weather.
Why this protocol
The methodology starts from Sanders, Handmaker, Johnson and Senner (2021), Wader Study 128(2):117–124 — a published, peer-reviewed Whimbrel nocturnal-roost protocol with an open-source R analysis toolkit at github.com/mainahandmaker/nocturnal-roost-counts. Six features of their core design are kept: multiple counting positions, civil twilight as the dusk reference, tide–twilight gap of 125 minutes or less for date selection, moon-phase sensitivity, 30-minute time-block resolution, and disturbance-event subtraction from running totals.
Three layers of behavioural observation are added: per-disturbance variables from Lilleyman, Franklin, Szabo and Lawes (2016), Emu 116(2) — same flyway, very similar context to Moreton Bay; posture-state scoring from Ryeland, Weston and Symonds (2017, 2019), Functional Ecology; and direction-and-flight-pattern fields that allow partial reconstruction of the catchment area.
Anticipatory or reactive: why pressure is recorded
The pre-storm KBR observations from November 2021 and December 2022 are consistent with a Terek roost shift in advance of foul weather. Two mechanisms are possible. Anticipatory: birds detect a falling barometer and shift to the constructed roost ahead of the front arriving. Reactive: birds shift only after wind and wave action have made the mangrove fringe unusable. Pressure-trend data alongside the wind/cloud/rain conditions allows a test of which mechanism is at work. The conservation implication differs: an anticipatory shift implies the birds have learned KBR as a strategic option; a reactive shift implies KBR is a fall-back when the preferred habitat fails.
Peters and Otis 2007 (Journal of Applied Ecology) is the closest published precedent: at South Carolina coastal roosts, daily-scale roost selection by migratory shorebirds was driven primarily by wind speed and the wind-shelter provided by individual roosts. Birds shift between roosts on a daily basis based on weather. DuskWatch tests whether Terek at Pumicestone Passage do the same, and whether the trigger is reactive (wind already up) or anticipatory (pressure falling).
Flag observations and Terek migration context
Branson, Shigeta, Chiang and Minton 2010 (Wader Study Group Bulletin 117: 27–34) is the foundational reference on Terek movements in the East Asian-Australasian Flyway, drawing on flag sightings and band recoveries from Australia, Japan, Taiwan, Hong Kong, China, Indonesia and Russia. Their key finding for Moreton Bay's purposes: Terek wintering in eastern Australia have a different migration strategy from those wintering in north-west Australia. Eastern birds occur widely along the SE Asia mainland and Japan on both legs; NW birds initially stage in Taiwan. So Moreton Bay Tereks belong to a behaviourally distinct cohort from the better-studied Roebuck Bay (Broome) population. Site-level data from Pumicestone Passage is not substitutable from Broome data.
Flag observations recorded during DuskWatch sessions feed both the project's local dataset and the flyway-wide BirdMark database. Any individual flag-marked Terek detected at KBR adds to a Terek migration evidence base that, as of 2025, remains thinner than it is for godwits, knots, or even tattlers.
Camera role
A roost camera operating at KBR is available as a supporting evidence stream, subject to the operator's permission. Camera-derived counts are entered into the same data structure as observer counts, with a source-flag column distinguishing the two.
References
Branson, N.J.B.A., Y. Shigeta, C.Y. Chiang and C.D.T. Minton. 2010. Movements of Grey-tailed Tattlers and Terek Sandpipers in the East Asian-Australasian Flyway. Wader Study Group Bulletin 117: 27–34.
Coleman, J.T., D.A. Milton and H. Akutsu. 2018. The migratory movements of Grey-tailed Tattler Tringa brevipes from Moreton Bay, south-east Queensland. Stilt 72: 2–8.
Fuller, R.A., R.S. Clemens, B.K. Woodworth, D. Moffitt, R. Steven and B.A. Simmons. 2021. Managing Threats to Migratory Shorebirds in Moreton Bay. Final report to Healthy Land and Water. University of Queensland, Brisbane.
Higgins, P.J. and S.J.J.F. Davies (eds). 1996. Handbook of Australian, New Zealand and Antarctic Birds. Volume 3: Snipe to Pigeons. Oxford University Press, Melbourne.
Lawler, W. 1995. Guidelines for the design of artificial high-tide roosts for shorebirds. Queensland Wader Study Group.
Lilleyman, A., D.C. Franklin, J.K. Szabo and M.J. Lawes. 2016. Behavioural responses of migratory shorebirds to disturbance at a high-tide roost. Emu — Austral Ornithology 116(2): 111–118.
Peters, K.A. and D.L. Otis. 2007. Shorebird roost-site selection at two temporal scales: is human disturbance a factor? Journal of Applied Ecology 44(1): 196–209.
Ryeland, J., M.A. Weston and M.R.E. Symonds. 2017. Bill size mediates behavioural thermoregulation in birds. Functional Ecology 31(4): 885–893.
Ryeland, J., M.A. Weston and M.R.E. Symonds. 2019. The importance of wind and water temperature for shorebird thermoregulation. Functional Ecology 33: 477–486.
Sanders, F.J., M.C. Handmaker, A.S. Johnson and N.R. Senner. 2021. Nocturnal roost on South Carolina coast supports nearly half of Atlantic coast population of Hudsonian Whimbrel. Wader Study 128(2): 117–124.