PhD student enlists citizen scientists and Stanford AI researchers to study fish in Georgian Bay

Climate change created a problem for Biology Professor Pat Chow-Fraser’s Wetlands Ecosystem Research Lab.

In 2019, after Chow-Fraser and her team of student researchers had used nets for years to keep tabs on fish communities in Georgian Bay’s wetlands, the  nets came up empty.

Climate change was causing unprecedented and extreme swings in water levels that had led to a new zone of flooded dead trees and shrubs that blocked fish from moving into their wetland nursery habitat. The fish were staying in deeper water that nets couldn’t reach.

Chow-Fraser set aside money from a Great Lakes Protection Initiative grant from Environment and Climate Change Canada and asked biology PhD student Danielle Montocchio if she wanted to help solve the problem.

Growing up, Montocchio had gone on family camping trips around Georgian Bay. She had studied Group of Seven landscapes in art classes. She had realized in high school that she wanted to be an environmental scientist. And once she learned about wetlands, it was clear what she wanted to study and protect.

“Wetlands are unique biodiversity hotspots. It’s where land meets water and everything from fish to birds coexist in such a relatively small space.”

Montocchio was certainly the right choice, Chow-Fraser says.  “This project’s been a labour of love with years worth of dogged determination by Danielle. I’m not sure anyone else could have pulled this off.”

Montocchio got right to work. “Each time I suggested a solution, Pat gave me the resources to run with it. The solution we finally ended up with was probably Plan D on our journey, so Pat’s patience and support to get there was invaluable.”

She came up with three breakthrough ideas.

What if the team caught fish on camera instead of in nets? They could record fish in deeper water, with the added bonus of doing it in a less labour-intensive way. And there’d be no risk of fish dying in nets.

She also knew who to ask for help. “Ever since I was little, I’d go fishing with my dad. And now he could help me figure out a way to view fish in their own underwater world.”

Together, they MacGyvered underwater camera stands out of PVC pipes and cement. Each stand was tricked out with two cameras. One camera pointed out into the bay and the other one pointed toward shore.

Montocchio dropped the stands into 15 wetland locations along the southeastern coast of Georgian Bay. By the end of 2021, she had collected around 1,000 hours of underwater video.

Montocchio is dogged and driven, but even she couldn’t watch that much footage on own. This led to her second big idea.

Chow-Fraser’s lab has a long tradition of getting undergrads and community members involved in its research. This fish survey project would need a bigger boatload of volunteers.

So Montocchio pitched Zooniverse on a Where’s Walleye? project. Zooniverse is billed as the world’s largest and most popular platform for people-powered research. It has more than a million volunteers around the world who help researchers like Montocchio.

Zooniverse approved Where’s Walleye? and the project is scheduled to go live this week. Volunteers will watch 360,000 10-second video clips, marking which clips feature fish and identifying which species are in the footage.

When they’re finished, they’ll have a massive database.

Which leads to Montocchio’s third big idea: Could that dataset be used to train AI to watch and categorize future fish footage? Montocchio couldn’t find any open-access AI platforms that biologists could easily use. So she reached out to machine learning experts at Stanford University, where she hopes researchers can come up with a solution.

“We’re lucky to be living in a time where the advances in technology and knowledge sharing mean pretty much any problem can be solved with a bit of creativity and a willingness to ask for help,” says Montocchio. “Reaching out to researchers, coders and engineers has shown me the power of collaboration.”

Chow-Fraser predicts that what Montocchio’s working on will be a game changer for more than just researchers. Footage from her videos has already been used as evidence during a hearing of the Ontario Land Tribunal. That evidence helped protect a fragile coastal wetland from development.

“I truly think that what Danielle is working on will become the future for wetland monitoring and protection.”