How SnotBots, Surveys and NASA Are Saving Our Oceans
One of the biggest problems we face today is the impact we’re having on the natural world.
Covering 70% of our planet we still know so little about our oceans and how we can protect them from further destruction. We can see the devastation around us like deforestation and mass urbanisation etc., but the destruction of the oceans is so easily ignored as it remains largely unseen beneath the waves. It’s so vital to keep our oceans healthy; they are home to millions of species of plant and animal, provide food, financial resources, and even produce half of the oxygen we breathe.
Climate change and coral bleaching, pollution, and overfishing are just a few issues that need urgent attention in order to save our oceans. We desperately need the help of governments, consumers, industries and corporations, and local communities around the world to make the changes necessary before we see the collapse of our ocean habitats.
The growing interdisciplinary field of computational sustainability is playing a part in tackling some of the problems faced. There are many projects using computer vision systems, machine learning and large data sets to hopefully make a difference to our oceans and gain the knowledge to have a real impact on future sustainability.
Warming seas have caused severe damage to our coral reefs, most notably Australia’s Great Barrier Reef, which in 2016 was reported by scientists to have up to 95% irreparable damage. Scientists from the Global Change Institute at the University of Queensland, and Berkeley’s Artificial Intelligence Research Center, have used data from the XL Catlin Seaview Survey to automate the analysis of our coral reefs using deep learning. The process is 900 times faster than manual methods of analysing the photos. Scientists, environmentalists and other organisations involved with saving coral reefs will be able to use this information to learn how pollution, coastal activities, fishing and global warming are affecting them and act upon these findings accordingly. Not just this, but these deep learning algorithms are also trained on identifying different categories of coral and other organisms to better our understanding of what makes up these beautiful ecosystems.
An integral part of our ocean ecosystem are whales, and although they sit at the top of the marine food chain, a staggering six out of the thirteen great whale species are endangered. Artificial Intelligence and drone technologies are being used to monitor the health of these amazing animals, along with their environments by Ocean Alliance using SnotBots. SnotBots are drones that fly over whales and collect the blow from their blowholes when they reach the surface to breathe. This blow (or snot) is home to masses of data on whales like DNA, hormone levels, and bacteria, which is then analysed by powerful algorithms that can identify and inform us on the health of whales, their environments and the impact of these on human health.
From ocean giants to microscopic algae, machine learning is also being used to understand phytoplankton. These tiny plants float on the ocean’s surface and are responsible for producing most of the oxygen we breathe, consuming carbon dioxide and providing a fundamental food source for many species. Scientists at NASA’s Goddard Space Flight Centre are tracking phytoplankton and PACE (Pre-Aerosol Clouds and ocean Ecosystem), a new mission due to launch in 2022 will open up this research by collecting data on different species, and different parameters like temperatures. This information will support efforts in combating climate change and sustaining life in the ocean for many species relying on phytoplankton.
Creating virtually constructed habitats would allow scientists and researchers the chance to experiment in simulated environments on a global scale in order to establish the best way to improve the health of our oceans. EarthCube could be the answer to bringing projects like these together. The National Science Foundation is creating a living model of our entire planet, which will combine a huge collection of data sets from a variety of sources and disciplines. Replicating our ecosystem is a highly complex process. Using sensors to collect data to inform machine learning models would be an approach that could help us better understand how the ecosystem works and provide a base for experimentation. This 3-D model will allow scientists to see how the earth’s systems will respond to changes to the environment and help better predict disastrous events, and also solve many of the problems we face today in preserving our land and ocean habitats.