San Francisco Bay has become an unexpected refuge for migrating gray whales, but the region's heavy shipping traffic creates a deadly hazard. A collaboration between government agencies and researchers launched an automated whale detection system in May that uses thermal cameras and artificial intelligence to alert nearby vessels of marine mammals in the water, potentially preventing fatal collisions during peak migration seasons.

According to IEEE Spectrum AI, the system was developed by WhaleSpotter, a Somerville-based startup, and monitors the bay from a Coast Guard tower on Angel Island. The technology detects thermal signatures from whale breath plumes and sends verified alerts to the U.S. Coast Guard's Vessel Traffic Service, which then informs ships to reduce speed or change course. Human experts review all AI detections before alerts are issued, eliminating false alarms.

A Climate Crisis Reshaping Migration Patterns

The fundamental challenge stems from shifting environmental conditions. Gray whales undertake one of nature's longest migrations, traveling 15,000 to 20,000 kilometers annually between Arctic feeding grounds and Mexican breeding lagoons. Historically, these animals ate extensively before departure and fasted during transit, requiring no intermediate stops.

Since 2018, however, over 100 gray whales have deviated into San Francisco Bay during northbound journeys. Researchers believe record-low Arctic sea ice, driven by rapid climate warming, has decimated the microscopic algae supporting the food chain these whales depend on. Unable to accumulate sufficient reserves in traditional feeding zones, the animals now pause in the bay for weeks to forage before continuing north.

The consequences have been severe. Ship strikes accounted for 40 percent of gray whale deaths in the region during 2025, when 21 whales died in and around the bay. An April study estimated an 18 percent mortality rate for whales entering the bay. Douglas McCauley, director of the Benioff Ocean Science Laboratory at UC Santa Barbara, characterized last year as "truly a crisis for gray whales."

How Thermal Detection Outperforms Human Observation

The technology originated at Woods Hole Oceanographic Institution as a method for recognizing whale spouts. When cetaceans surface to breathe, they expel air warmer than surrounding water. Thermal imaging captures these heat signatures even when they occupy just one or two pixels in footage. The system monitors a water cone extending approximately 7 kilometers from its vantage point, operating continuously through night, fog, and inclement weather.

WhaleSpotter reports its systems reduce ship-strike risk by 90 percent. The company currently operates on eight container vessels operated by Honolulu-based Matson Shipping, but the San Francisco installation represents the first integration of land-based and vessel-based monitoring networks in a single, continuously surveilled region during peak migration season.

Expanding the Detection Network

  • Primary monitoring station located on Angel Island's Blunt Point
  • Secondary camera planned for a ferry operating between San Francisco and Vallejo
  • Real-time data shared with marine researchers and available on public online maps
  • UC Santa Barbara and Marine Mammal Center utilizing detection data for population studies
"Gray whales are trying to be a brave new whale in a weird world. This is a front-row view to climate change." , Douglas McCauley, UC Santa Barbara

The initiative reflects a broader adaptation to climate-driven ecological changes. As traditional migration corridors become less viable, human infrastructure must evolve to accommodate species displacement. The San Francisco system demonstrates that combining thermal sensing, machine learning, and human expertise can meaningfully reduce anthropogenic threats to vulnerable populations navigating an altered world.