How AI can help save hidden ‘biological hotspots’ the planet needs

"There are folks like the Tulalip Tribes using this tool to broadly ask how water is held across the landscape, and where wetlands can be restored to store water longer through the season or cool down streams that are salmon-bearing," says Maureen Ryan, multidisciplinary development lead for TealWaters.

Other organizations and agencies are using the tool to study how wetlands can mitigate floods, increase green spaces, nourish wildlife and support climate goals. For small cities without a wetland ecologist, the tool has become especially helpful for providing much-needed expertise.

"One of the interesting uses coming up is can we speed up the understanding of what wetlands do and the buffer distances required for urban planning and permitting, which can be an expensive process," Halabisky says.

In addition to using machine learning, TealWaters is exploring Microsoft AI tools for computer vision and convolutional neural networks for finding patterns in geospatial data more accurately and efficiently. It's also looking into Microsoft's Planetary Computer, a platform for accessing global environmental data. And it plans to incorporate data on water quality, floodwater absorption, habitats and other information.

The team is initially focused on creating a high-resolution map of wetlands in Washington and the state's first-ever map of high-carbon wetlands, which tend to be peatlands and forested, hard-to-find wetlands. Halabisky says the diversity of local ecoregions - rainforests, mountains and grasslands, to name a few - makes Washington a "perfect test area" for developing a tool for national and global use.