Type “Aurogorgia tasmaniensis” – a new species of soft coral found off Tasmania – into a search engine and you are unlikely to yield many results.
But for CSIRO’s researchers from our Oceans and Atmosphere team in Hobart, discovering this species embodies the possibilities of science to illuminate life in the deep-sea.
The Biodiversity, Imagery, Annotation and Automation (BIAA) team research the benthic zone, the region at the bottom of the sea and the life it supports. While much of their work relies on imagery from cameras that can be “flown” above the sea floor by pilots onboard a ship and withstand crushing amounts of water pressure, it is grounded in taxonomy, the oldest discipline in biology.
The importance of deep sea imagery
CSIRO surveys launched from Hobart, many with researchers from museums, government and other institutions onboard, have been finding new species for decades.
“Imagery makes it possible to capture species seldom collected because they are rare or prone to damage by physical sampling,” Candice Untiedt from CSIRO’s BIAA team, said.
“For instance, after Aurogorgia tasmaniensis coral was observed in images, and compared with samples collected in the region, the images and specimens were used by CSIRO taxonomist Dr Phil Alderslade and colleagues in the USA and Brazil to describe the species.”
The coral is endemic to the region around Tasmania’s seamounts and the prized type specimens, on which the species is based, are lodged at the Tasmanian Museum and Art Gallery in Hobart.
“Our research has significantly increased our understanding of octocorals, which are commonly called soft corals,” Untiedt said.
“Through our foundational taxonomic work over the last 30 years, we have increased the known number of octocorals around Australia from 160 to over 700 species, most of which are new to science and found from shallow shelf to deep abyssal depths.
How are new species discovered?
In late 2018, far from the tropics most Australians envision when they think of coral, CSIRO researchers aboard the RV Investigator took part in a survey of submerged mountains off Tasmania.
The expedition with Parks Australia and the Australian Government's National Environmental Science Program Marine Biodiversity Hub traversed Tasmania’s seamounts, which include extinct volcanoes. Using a sophisticated underwater camera system developed by CSIRO, researchers collected video and images of more than 250km of seabed. More than 70,000 images captured information from the seabed, revealing rich benthic communities of marine invertebrates. It was a fresh and refined look at the region’s habitats and diversity, 10 years after the first extensive survey.
“Analysis of the imagery found many cold-water coral reefs, forming large structures commonly 950m to 1,350m below the surface,” Franzis Althaus, a benthic ecologist, explained.
“The reefs were made up of single species, Solenosmilia variabilis, which support a wide variety of soft corals, sponges, echinoderms and other invertebrates. They are defined as Vulnerable Marine Ecosystems.”
The seamounts have been impacted by deep-sea bottom trawl fishery. Some of the seamounts are still open to fishing but most have been protected for 10 years and are enclosed in offshore marine parks.
Importantly, Australian Marine Parks off southern Tasmania, especially the Huon Marine Park, protect a large fraction of the total regional pool of seamounts, including many with light or no signs of impact.
“Using Image data gives us insights into these areas. Collecting visual samples from the field enables us to quantify and map habitats, invertebrate and fish communities,” Althaus said.
“By collecting, annotating and archiving images, we have the opportunity to build up a record over time that can be pulled up and compared to imagery in the future.
“This is invaluable for monitoring changes in the environment and changes due to, for example, climate change. It will also assist with ongoing management.”
The role of taxonomy
As scientists rely more on imagery-based underwater surveys, species can be identified more accurately by targeting and co-locating the collection of specimens for taxonomic analysis.
Taxonomic information is crucial for ecologists to understand and manage ecosystems. Science has been classifying living species for more than 250 years, but an estimated 91 per cent of marine species remain undescribed.
“Specimens are sent to national museums where expert taxonomists identify existing species or describe new ones,” Kylie Maguire from the BIAA team said.
“Identifying species hinges on taxonomists, museums and their rigour as a bedrock of knowledge of deep-sea biodiversity and the distribution of species around Australia.
“It demonstrates the importance of taxonomy as a foundation for monitoring changes in response to climate change and physical disturbance from human-induced impacts.”
In other words, for Australians to understand and conserve the deep-sea riches they only see in pictures, they first need to know what is there.
Mighty machine learning
The BIAA team is increasingly tackling the mammoth task of data processing beyond the ability of any human using machine learning. This makes data analysis quicker and more efficient. The key is to train the machine learning tools correctly. So far, the results are encouraging.
“We recently repurposed a dataset for a machine learning exercise to detect coral reef substrate, a proxy for Vulnerable Marine Ecosystem habitat,” Maguire said.
“We are very happy with the results and are now looking into object detection for selected deep-sea species.”
As machines shoulder more of the load of detecting Australia’s underwater species, CSIRO’s BIAA researchers are determined to keep expanding what is known of the deep-sea void.