Underwater ‘doorbell’ helps scientists catch coral-eating fish in Florida
Marine scientists in Florida working to help reverse a calamitous decades-long decline in coral reefs caught fishy “porch pirates” in the act with an innovative underwater doorbell-style surveillance camera.
The footage showed that three corallivorous species – redband parrotfish, foureye butterflyfish and stoplight parrotfish – were responsible for eating more than 97% of coral laid as bait by the researchers at an offshore reef near Miami.
The findings, they say, can help inform coral reef repopulation efforts following a 90% decline in Florida’s coral cover since the 1970s following unprecedented bleaching events caused by the climate emergency, particularly record ocean heat over the last two summers.
“Intense fish predation on newly outplanted corals has emerged as a major restoration bottleneck. The main goal was to address our lack of knowledge of the fish species that target corals after outplanting,” said Diego Lirman, a global leader in coral restoration research. He is a project leader and associate professor at the University of Miami’s (UM) Rosenstiel school of marine, atmospheric and earth science.
“Identifying the fish species responsible for coral predation would allow practitioners to avoid reef sites or areas within sites with high abundances of those species and, similarly, select the right coral species for the right outplanting site,” Lirman said.
“The coral-baited underwater cameras provide insight into corallivore behavior and preferences and allow documentation of predation at various sites rapidly and without incurring the cost of outplanting.”
Lirman’s team, funded by a grant from the Fish & Wildlife Foundation of Florida, designed and built a number of recording devices using GoPro cameras in waterproof housing attached to a PVC frame, and with lead weights for stability. After shaky preliminary results, divers secured later models to the seabed at Paradise Reef, close to Key Biscayne in Miami-Dade county, using masonry nails and cable ties.
The so-called C-Bruvs (coral-baited remote underwater video system) were set to record time-lapse video, and footage was collected at 24- and 48-hour periods after deployment and weekly thereafter for the duration of the six-week project.
Lirman said the team overcame initial setbacks including overheating external batteries and leaks causing flooding in the camera housing, while the research was also not immune to petty porch pirate-style pilfery familiar to homeowners above ground who have recorded thefts on their doorstep.
“One of the C-Bruvs with a surface buoy attached was stolen from the reef,” he said.
Overall, however, the researchers considered their experiment a success. After analysis of the imagery, they determined redband parrotfish, which are prolific in the Caribbean and waters off Florida, were the most voracious predators, accounting for 56.3% of bites on the nine coral species put out in fragments as bait.
Next came foureye butterflyfish with 36.9%, and stoplight parrotfish with 4%. The three species, Lirman said: “showed clear preferences” for two or three particular types of coral, which received more than 65% of all bites recorded.
UM marine scientist and research master’s graduate Erin Weisman presented the findings to the Reef Florida symposium of conservation leaders at Miami’s Phillip and Patricia Frost Museum of Science in November.
“By identifying, for the first time, the main fish predators as well as their preferred diet, reef restoration practitioners can select sites and species that would minimize predation impacts and maximize restoration success before large-scale, costly outplanting is implemented,” Lirman said.
Future similar research, he added, could introduce elements of artificial intelligence.
“Analyses of the videos were extremely time-consuming, requiring a constant rewinding and stopping of the footage to record and annotate coral/fish interactions,” he said.
“It will be beneficial to explore AI software that can be trained to identify fish and their behaviors to automate the analysis process.”
About The Author
