Trial of crown-of-thorns starfish (COTS) and coral monitoring design collected by Reef Joint Field Management Program

This dataset contains Crown-of-thorns starfish (COTS) and coral monitoring data collected from 18 Great Barrier Reef locations during two pilot surveillance deployments in January-February 2025 as part of the Crown-of-thorns starfish Control Innovation Program (CCIP). The deployments tested integrated monitoring approaches across the Swain Reefs (9 reefs, January 2025) and Mackay region (9 reefs, February 2025), combining traditional visual surveys with emerging technologies. The dataset includes: (1) manta tow survey data comprising 613 individual tows covering approximately 117 km of reef margin, with observations of COTS abundance, feeding scars, and categorical hard coral cover estimates; (2) ReefScan underwater camera data from both transom-mounted (272,874 images, 875 GB) and deep-towed (361,970 images, 838 GB) systems, processed using machine learning for COTS detection and benthic composition analysis; (3) environmental DNA (eDNA) summary results from 432 water samples across 18 reef locations, providing molecular detection metrics for COTS presence; and (4) comparative coral cover estimates derived from each monitoring method. \nLineage: The pilot deployments were conducted aboard the Queensland Parks and Wildlife Service (QPWS) vessel Reef Resilience. Reef locations were selected to approximate a spatially balanced cluster sampling design, with three clusters of three reefs in each region, though operational constraints required alignment with pre-scheduled Reef Joint Field Monitoring Program (RJFMP) activities. The January deployment (9-13 January 2025) surveyed nine reefs in the northern Swain Reefs complex (cross-latitude 21°S), while the February deployment (7-12 February 2025) targeted nine reefs in the Mackay region (north of 20°S latitude) under challenging monsoonal weather conditions.\nManta tow surveys followed standard RJFMP protocols, with observers recording COTS presence, feeding scar categories, and categorical hard coral cover estimates using established methods (Miller and Müller 1999). Data were aggregated to reef level by calculating total COTS observed, proportions of tows with feeding scars or COTS presence, and average hard coral cover derived from category midpoint values. ReefScan camera systems were deployed concurrently: the Transom unit collected surface imagery during manta tows, while the Deep unit conducted independent 1-2 nautical mile transects where operationally feasible. ReefScan imagery was processed using automated machine learning detection systems (Bainbridge et al. 2025), with high-confidence COTS sequences (maximum confidence ≥0.75, median confidence ≥0.75, length ≥5 images) validated manually to eliminate false positives. As each ReefScan unit employs dual cameras operating independently, only detections from a single camera should be retained in any analysis to prevent duplicate counts. Hard coral cover from ReefScan data was derived from benthic analysis using the Reef Cloud platform, averaged across cameras within transects and then across all transects per reef. Environmental DNA samples were collected following protocols developed during the COTS Control Innovation Program, with 12 replicate water samples filtered at each reef location. Laboratory processing and PCR analysis were conducted by the Australian Institute of Marins Science (AIMS), yielding metrics including proportion of positive samples and average PCR concentration values. All spatial data are referenced to WGS84 datum in decimal degrees.