Simon Fraser University researchers have received nearly $1 million in special funding from the Digital Research Alliance of Canada to develop an artificial intelligence-powered system that forecasts whale movements in busy shipping corridors.
Download high-res images and video
The Humans and Algorithms Listening for Orcas (HALLO) project aims to help the Port of Vancouver and vessel pilots make more informed decisions about when and where to slow down for the Salish Sea's endangered Southern Resident killer whales.
The system integrates real-time acoustic and visual data, vessel tracking, and citizen-scientist whale spotting reports to track not only where the Southern Resident J, K, and L pods currently are, but forecast where they'll be over the next few hours.
"We're using AI to support faster detection, more reliable classification of Southern Resident killer whales in and adjacent to shipping lanes," says Ruth Joy, principal investigator on the HALLO team.
"We're looking at the feasibility of a dynamic slowdown trial to test if real-time detection of these killer whales coupled with a movement forecast system could trigger timely ship slowdowns when and where the whales are present, rather than at fixed times when whales may not be present."
Right now, the Port of Vancouver coordinates a voluntary static slowdown from June to November, encouraging B.C. coast pilots to reduce speed through key whale habitats like Haro Strait and Boundary Pass.
"But slower transit can cause total transit times to exceed a pilot's regular eight-hour shift, requiring a second pilot. Ships are being asked to do a lot even at times we know whales are not present," Joy says.
Noise a 'controllable threat'
Historical data shows the critically endangered Southern Resident killer whales use clear "commuting" corridors in the Salish Sea, which includes the Strait of Georgia, the Strait of Juan de Fuca, Boundary Pass, and Haro Strait.
They share these heavily trafficked waterways with freighters and tanker ships, cruise ships, fishing boats, passenger ferries, tourism vessels, and recreational boats.
Vessel noise overlaps with the sound frequencies the whales use for communication and echolocation - essential for finding the fast-moving salmon that make up the bulk of their diet, explains Joy, assistant professor of environmental science.
"Noise is a controllable threat that compounds the impact of Chinook salmon scarcity. More fish doesn't help if the whales can't 'hear' them," she says.
Wider role for real-time whale detection and forecasting
HALLO researcher Fabio Frazao is building an AI-based classifier to distinguish transient killer whales from the Southern Residents, while Teng-Wei Lim is using AI to forecast a pod's direction and speed to predict when the whales will be moving into the shipping lanes.
But the HALLO project's potential impact extends well beyond reducing vessel speeds in shipping lanes, Joy says.
The real-time whale detection and movement forecast system can also be used to inform port expansions, pile-driving projects and other noise-generating development along B.C.'s coastline.
Beyond the Salish Sea, the forecasting system could also be used on the east coast, supporting protection measures already in place to combat whale entanglement and vessel strikes, the two biggest threats facing the critically endangered North Atlantic right whale.
Using AI to bridge conservation and industry
The funding comes through the Special Funding Opportunity for Research Software AI Enhancement, awarded by the Digital Research Alliance of Canada to four national research projects.
The program supports research platforms and portals in adopting AI techniques and delivering AI-enabled services to their user communities.
HALLO researchers say the funding injection comes at a critical moment, aligning with the port's interest in piloting dynamic slowdowns.
"Right now, our tracking and forecasting system is available through a web app, but the team is still discussing the most useful format for vessels and the Port of Vancouver," says Bruno Padovese, researcher on the HALLO team.
"Our role isn't to transmit alerts directly to vessels, but to help authorities access real-time forecasting data and decide how best to inform vessel pilots."
SFU experts available
RUTH JOY, assistant professor, School of Environmental Science, associate member, statistics
BRUNO PADOVESE, researcher, School of Environmental Science
(out-of-province, available by Zoom)
