A Queensland University of Technology (QUT) team led by Southern Cross University's Professor Peter Harrison has developed the "LarvalBot" underwater robot that, for the first time, has succeeded in reseeding damaged areas of Australia's Great Barrier Reef with heat-tolerant baby coral polyps to help combat the effects of predators and climate change.

Earlier this year, we reported on an undersea robot developed at QUT under Professor Matthew Dunbabin called RangerBot, which was designed to hunt down and kill the crown-of-thorns starfish responsible for degrading 40 percent of Australia's famous Great Barrier. However, that's only half the story. Now a new generation of RangerBot has been used to try and help repair the damaged reef.

Part of the problem is that repopulating a depleted coral colony isn't like restocking an area with fish or planting a bed of kelp. Coral reproduce in a very unusual and unpredictable ritual called a coral bloom, where colonies individually or together release vast clouds of egg and sperm to unite to form the planktonic larvae that will one day settle down to become the coral-forming polyps.

Exactly when these blooms occur is hard to say. They may happen at night after a full moon during a period when the waters have been warm enough to stimulate the coral to form gametes and may last for days or a week. But exactly which colonies will bloom when is something of a guessing game so collecting the coral spawn is no easy task.

But that's just half the problem. The other half is to find ways to raise the larvae once captured, then disperse them where they'll thrive and do some good in reef rebuilding. It's a new approach, but one that has already been field tested in the Philippines and by Harrison at the Great Barrier Reef in 2016 and 2017. These small-scale tests showed promise, but the next step is to bring in robotics on trials on Vlasoff Reef, near Cairns in north Queensland.

For this, Harrison's team used large floating canopies to capture the coral spawn as it rose from the desired seeding colony. These were then reared in specially designed larval pools, then dispersed over the needed area. For the test, LarvalBot was equipped to handle 100,000 coral larvae per mission, though Harrison plans to scale up to millions of larvae in the near future as his team monitors the results of the first tests next year.

During the test, LarvalBot was tethered as it gently distributed the larvae like an underwater crop duster. However, on larger-scale missions, it will operate autonomously. Weighing 15 kg (33 lb), the 75-cm-long (30-in) LarvalBot is programmed using a tablet app and is equipped with a computer vision system and other sensors for navigation and recording reef pathologies.

According to QUT, the team plans to build larger "mega" spawn-catchers and solar powered floating larval incubation pools for rearing hundred of million of genetically diverse, heat-tolerant coral larvae for seeding on damaged reefs by larval clouds and robots.

The video below discusses the LarvalBot concept.