Global Plastic Watch (GPW), a tool developed by Minderoo Foundation, uses satellite data and machine learning to show a map of the Earth’s plastic pollution.
The satellite data comes from the European Space Agency and the machine learning tool was made in collaboration with Earthrise Media.
The tool focuses on land based plastic waste that threatens to leak into the marine environment and can detect waste sites that are as small as 5m2. The AI scans satellite imagery looking for signs of plastic dumping sites like grey-brown areas showing mounds of waste and entry roads for heavy vehicles.
It is the largest freely available open-source dataset of plastic waste and is currently identifying plastic waste sites across 25 countries.
“The destination for every piece of environmental plastic is nano-plastic, which has both poisonous and cutting attributes able to mutilate cells and possibly even penetrate the human blood-brain barrier. Preventing illegal and legal plastic waste stockpiles entering the oceanic environment is critical to limit this harm. Once in the ocean, through both mixing, absorption and ingestion by animals, this plastic will officially enter the human environment,” said Dr Andrew Forrest the co-founder of Minderoo Foundation.
One in five of the sites mapped so far are within 250 metres of waterways which will take the plastic to the ocean. “We know that land-based leakage contributes up to 91 per cent of the plastic waste that enters the ocean,” said Minderoo Foundation’s Head of Technology and Innovation Fabien Laurier.
The tool has mapped out all south-east Asian countries as well as Australia and the countries identified by research published Science Advances as countries responsible for high rates of plastic pollution into the ocean.
The Indonesian government has been working with Minderoo Foundation to help manage its waste problem as the fifth worst global plastics polluter, but Fabien Laurier said, “It’s not about naming and shaming,” but “empowering governments”.
