Off the coast of Tampa Bay and along the shores of Southern California, microscopic algae blooms can turn the water toxic in a matter of days—poisoning dolphins and sea lions, sickening swimmers, and costing coastal economies tens of millions of dollars every year. Now, NASA scientists have built an AI tool that fuses data from five satellites to spot these dangerous blooms before they spread, giving coastal communities the early warning they desperately need.
The problem has plagued these regions for decades. In Florida's Gulf waters, a species called Karenia brevis thrives in conditions that kill wildlife and foul beaches with sickening toxins. On the West Coast, blooms of Pseudo-nitzschia have poisoned hundreds of marine animals in recent years, and their airborne toxins can even trigger respiratory illness in humans. The stakes are high, which is why health agencies spend countless hours testing waters by hand—collecting samples from boats, sending them to labs for analysis, a process that takes a day or more and leaves huge gaps in monitoring coverage.
The challenge that NASA researchers Michelle Gierach, Kelly Luis, and Nick LaHaye set out to solve was elegantly simple in concept but fiendishly complex in execution: how could artificial intelligence learn to recognize an algal bloom across multiple satellite datasets at once? Today's Earth-orbiting satellites already detect these blooms through different lenses. NASA's PACE satellite uses a hyperspectral sensor that identifies algal communities by their size, shape, and pigment. The TROPOMI instrument picks up on the faint red glow emitted by K. brevis as it photosynthesizes. But bringing all this information together into one coherent warning system required teaching a machine to learn from diverse data streams without human guidance.
The team developed a self-supervised machine learning system trained on satellite data collected in 2018 and 2019, then validated against real-world field and lab measurements. The approach was groundbreaking because it enabled the AI to recognize patterns and relationships between different data sources without needing any advance labeling—a crucial advantage when dealing with the raw volume of satellite information pouring in continuously. The results, published recently in the Earth and Space Science journal, showed that the tool could correctly identify and map harmful blooms, even in murky coastal waters swirling with sediment, plants, and runoff.
"At the very least, a tool like this can help us know where and when to collect water samples as an algal bloom is starting," said Gierach, a scientist at NASA's Jet Propulsion Laboratory in Southern California. More ambitiously, she noted it could drive collaboration between specialists and foster new ways to deliver decision-support products to communities on the front lines.
The implications extend beyond Florida and California. NASA scientists are now expanding the tool with more data from additional coastlines and testing its capabilities on lakes and other water bodies. The goal is to make it accessible to the decision-makers—water quality managers, public health officials, and emergency responders—who need to act fast when a bloom appears on the horizon. For coastal communities that have wrestled with toxic algae for decades, this represents a real chance to move from reactive crisis management to proactive protection.