SCIENCE

Predicting coral bleaching events has been key to reef conservation management efforts. Current satellite-based bleaching prediction tools offer effective regional-scale alerts of bleaching risk, but lack reliability at the reef-scale. Bleaching models focus on predicted heat exposure during summer, omitting critical factors that influence heat stress responses and the subsequent coral reef community bleaching severity. The IPCC framework however assesses the susceptibility of a system to be harmed by climate change based on exposure, sensitivity, and adaptive capacity. In this perspective, we propose integrating the IPCC vulnerability framework to develop a holistic coral bleaching prediction model that accounts for reef-scale exposure to heat stress, species-specific sensitivity, and adaptive capacity. We specifically recommend: 1) incorporating historical temperature metrics to account for acclimatisation responses, 2) including community composition metrics to better reflect variations in sensitivity at the reef scale, and 3) addressing environmental conditions to identify potential refugia and refine predictions. We discuss these factors and the feasibility to inform metrics for use in prediction tools. Historical temperature is identified as a primary target, with community composition and environmental drivers recommended for further exploration as data availability improves. Future assessments of these sensitivity metrics should be integrated into an experimental framework to further refine and improve prediction tools. This perspective underscores the urgency of refining coral bleaching prediction models and directly supports reef conservation efforts in the face of climate change.

There is an urgent need for improved monitoring approaches to rapidly and accurately assess sea cucumber populations at ecologically relevant scales. Timely surveys are critical for informing effective fisheries management and decision-making. Traditional surveys, undertaken via snorkelling, manta tows, or SCUBA, are limited to shallow and accessible areas; however, sea cucumbers inhabit a broad range of depths, including areas beyond safe diving limits and exposed shallow waters inaccessible by boat. To overcome these limitations and increase the rapidity of field collection, we propose the use of remote sensing technologies to survey sea cucumber populations across a range of depths. Here, we evaluated the effectiveness of aerial drones and in-water remote operated vehicles (ROVs) for assessing sea cucumber species and abundances across various depth ranges (< 50 m) on the Great Barrier Reef, Australia. Aerial drone orthomosaics and ROV video footage were compared to more traditional snorkel and SCUBA-based assessments conducted at similar depths. The vast majority of pairwise comparisons between in-water ROV video counts and snorkel or SCUBA assessments found no significant differences in sea cucumber assemblages. Counts from aerial drone-derived orthomosaics, however, were significantly lower, counting approximately half as many sea cucumbers as snorkel counts. This was largely attributed to poor weather during the drone surveys. Remote methods were significantly faster in the field for surveying a given area than traditional methods. Given that towed ROVs can efficiently cover a broader depth range and aerial drones are effective for survey shallow areas under suitable weather conditions, we recommend using a combination of aerial drones and towed ROVs to survey sea cucumbers, with tool selection guided by prevailing weather conditions. This approach offers the advantages of collecting multiple types of data from a single data source, vastly increasing survey efficiency, and providing a historical record for future assessments. The methods have the potential to be used to survey other benthic–associated species.

Predicting coral bleaching events has been key to reef conservation management efforts. Current satellite-based bleaching prediction tools offer effective regional-scale alerts of bleaching risk, but lack reliability at the reef-scale. Bleaching models focus on predicted heat exposure during summer, omitting critical factors that influence heat stress responses and the subsequent coral reef community bleaching severity. The IPCC framework however assesses the susceptibility of a system to be harmed by climate change based on exposure, sensitivity, and adaptive capacity. In this perspective, we propose integrating the IPCC vulnerability framework to develop a holistic coral bleaching prediction model that accounts for reef-scale exposure to heat stress, species-specific sensitivity, and adaptive capacity. We specifically recommend: 1) incorporating historical temperature metrics to account for acclimatisation responses, 2) including community composition metrics to better reflect variations in sensitivity at the reef scale, and 3) addressing environmental conditions to identify potential refugia and refine predictions. We discuss these factors and the feasibility to inform metrics for use in prediction tools. Historical temperature is identified as a primary target, with community composition and environmental drivers recommended for further exploration as data availability improves. Future assessments of these sensitivity metrics should be integrated into an experimental framewor to further refine and improve prediction tools. This perspective underscores the urgency of refining coral bleaching prediction models and directly supports reef conservation efforts in the face of climate change.

Abstract:

Coral reefs, as biologically diverse ecosystems, hold significant ecological and economic value. With increased threats imposed on them, it is increasingly important to monitor reef health by developing accessible methods to quantify coral cover. Discriminating between substrate types has previously been achieved with in situ spectroscopy but has not been tested using drones. In this study, we test the ability of using point-based drone spectroscopy to determine substrate cover through spectral unmixing on a portion of Heron Reef, Australia. A spectral mixture analysis was conducted to separate the components contributing to spectral signatures obtained across the reef. The pure spectra used to unmix measured data include live coral, algae, sand, and rock, obtained from a public spectral library. These were able to account for over 82% of the spectral mixing captured in each spectroscopy measurement, highlighting the benefits of using a public database. The unmixing results were then compared to a categorical classification on an overlapping mosaicked drone image but yielded inconclusive results due to challenges in co-registration. This study uniquely showcases the potential of using commercial-grade drones and point spectroscopy in mapping complex environments. This can pave the way for future research, by increasing access to repeatable, effective, and affordable technology.