British Columbia’s coast holds some of the most extensive kelp forests on the Pacific side of North America, but it is also one of the cloudiest places to try to watch them from space. A new method published this month by a University of Victoria remote-sensing team, led by Marta Konik and senior researcher Maycira Costa at the school’s SPECTRAL Remote Sensing Laboratory, works around that problem by trading cameras for radar. The study, appearing in the journal Ecological Informatics with co-authors Eduardo Loos, Alejandra Mora-Soto, Sarah B. Schroeder, Lianna Gendall, Lauren Man, Luba Reshitnyk, Vanessa Caron, Michael Henley, and Kim Pearce, shows that satellite-based synthetic aperture radar, or SAR, can map floating kelp canopy along the BC coast about as well as the optical imagery scientists have relied on for years, without needing a clear sky to do it.
That distinction matters more on this coastline than almost anywhere else kelp grows. The tools researchers have used to track kelp canopy extent over time are mostly optical sensors, like those aboard the Sentinel-2 and Landsat satellites, and they need an unobstructed view of the sea surface. BC’s coast is frequently blanketed in cloud for weeks at a stretch, and every cloudy pass is a gap in the record. This kind of tracking is only useful if it happens consistently, so missed passes make it harder to catch the early signs of a kelp bed thinning out or a new one taking hold.
Radar sidesteps the weather problem because it does not depend on reflected sunlight. A SAR satellite bounces its own radar pulses off the sea surface and reads what bounces back. A mat of floating kelp blades sitting just centimeters above the water changes the roughness of that surface in a way the radar can detect, cloud cover or not. Konik’s team paired SAR readings from the European Space Agency’s Sentinel-1 satellites with matching optical Sentinel-2 scenes of the BC coast to build and test a classification model that identifies where kelp canopy is present. They then published the underlying training data and classification model through the University of Victoria’s public research data collection, so other groups can use or build on it.
The point isn’t to replace optical monitoring so much as to patch its blind spots. The authors frame SAR and optical imagery as complementary tools, each catching what the other misses, with the combination pointing toward a more continuous, weather-independent record of kelp canopy along the coast. That kind of steady record is what an early-warning system for kelp decline would actually need, and it’s the piece that’s been missing for a coastline where the clouds themselves have been part of the monitoring problem. The team’s next step is refining the approach and seeing how well it extends to other persistently overcast coastlines nearby, where the same cloud-cover gap has likely been hiding just as much.