Mapping Underwater Meadows From the Skies
Hakai scientists use drones and computer savvy to track changes in seagrass ecosystems.
Scientists collect clues to piece together what’s going on in the natural world. To separate fact from fiction requires data, and the right people to make sense of it. Since May, Hakai researcher Thea Mai pored over mosaics made up of thousands of aerial photos of one of the coast’s most productive ecosystems—seagrass meadows—to discover if the meadows are growing, shrinking, or staying the same. It may seem simple, but getting to the answers is not so straightforward.
“Mapping anything in coastal areas is challenging,” says Mai. “The variability in how habitats look is so high. You have to problem-solve quite a bit.”
Seagrass meadows can be particularly difficult to map. Hakai geospatial scientists fly a drone in a grid pattern over a meadow, taking photos that are then stitched together into one image. Even if conditions are perfect—low tide, minimal sun glare, not too many waves—finding the exact edge of the seagrass meadow can be murky. Some seagrass is exposed in the intertidal at low tides, but it can grow down to a depth of five meters on the Central Coast of British Columbia, and the meadow’s edge can overlap and be obscured by the presence of kelps and other algae.
Once the images are stitched together, Mai inputs them into software that analyzes each photo mosaic based on color, texture, and the similarity of neighboring pixels. As best as it can, the software draws a shape around what it thinks is seagrass, much the same way as someone would draw by hand. But the computer algorithms can’t do everything.
“I have to supervise the software,” says Mai. “All of the objects are twisty. There aren’t many clean edges between habitats. Sometimes it took weeks to work up one scene.”
Through this detail-oriented process, Mai found that how seagrass meadows are changing in size varies across the Central Coast of British Columbia. In some places, such as Pruth Bay on Calvert Island, the meadows have grown since 2012. In other places, meadows have shrunk. And in some locales, the overall extent has stayed the same, but the density of seagrass has declined. Where density is lower the image looks lighter green in color and the edges of the meadow are less defined, rather than the dark green and crisp edges of a meadow with densely growing plants.
A drone’s-eye view of the seagrass meadow at Superstition Point on Hunter Island in 2019 and 2020. While the extent of the meadow did not change much, the density of the seagrass declined. Aerial image by Luba Reshitnyk. Inset image by Margot Hessing-Lewis
The Hakai Institute is partnering on a project with the US National Oceanic and Atmospheric Administration to look at how seagrass meadows change over time and help inform management practices, as well as a coastwide project funded by the US National Science Foundation to examine the possible drivers of observed changes in seagrass beds—from both warming waters and an emerging wasting disease.
It didn’t take long for Mai to relish a chance to visualize the seagrass story. But when family members asked what she was working on, Mai realized that she had to expand her explanation beyond just her mapping pursuits.
“People don’t necessarily appreciate how important [seagrass meadows] are ecologically,” says Mai. “There’s just so much life down there.”
Mai’s time mapping seagrass with the Hakai Institute is ending shortly. She’s excited to use her honed skills for a new project with an entirely different focus. She’s begun a master’s program in anthropology at the University of Victoria to develop a model to find old shorelines and look for habitation sites in Haida Gwaii dating as far back as 10,000 years. And that will undoubtedly bring countless more data points to pore over.
Having all of that data is great. But it takes people like her to make sense of it.