Coral Bleaching

What is Coral Bleaching?

Coral bleaching occurs when corals become stressed and expel zooxanthellae—the microscopic algae living in their tissues. Zooxanthellae provide corals with up to 90% of their food energy through photosynthesis and give reefs their vibrant colors. Without these algae, corals become white (bleached), weakened, and vulnerable to starvation and disease.

The primary trigger for bleaching is elevated water temperature. When sea surface temperatures rise even 1-2°C above the coral's normal range for several weeks, it stresses the coral-algae partnership. Other stressors include freshwater runoff, excessive sunlight, ocean acidification, and pollution. In the 1998 El Niño event, approximately 16% of the world's coral reefs died in a single bleaching event. More recently, the 2016 global bleaching event affected two-thirds of the Great Barrier Reef, demonstrating both the severity of the threat and the value of systematic monitoring.

Bleaching is not immediately fatal. If temperatures cool within days to weeks, corals can recover by reabsorbing zooxanthellae. However, if stress persists for months, most bleached corals die. Marine scientists track bleaching through satellite monitoring, boat surveys, and underwater imagery to understand when and where stress occurs and which reefs are most vulnerable.

How AI Classifies Bleaching Events

Artificial intelligence, specifically deep learning neural networks, can automatically detect and classify coral bleaching from satellite imagery with accuracy comparable to expert marine scientists. These AI systems are trained on large datasets of labeled satellite images—some showing bleached reefs, others showing healthy reefs—to learn the visual patterns associated with each condition.

The process begins with satellite image preprocessing: scientists collect raw multispectral data and align it with geographic coordinates. They manually label a subset of images as "bleached," "partially bleached," or "healthy" based on field surveys and expert analysis. An AI model then learns the pixel patterns and color combinations that predict each classification. Once trained and validated, the model can automatically classify new satellite images as they arrive, flagging regions likely experiencing bleaching within hours of image acquisition.

This automated approach enables monitoring at scales impossible with manual surveys. Scientists can track thousands of reef sites simultaneously, identify bleaching events in remote regions within days, and provide early warnings to conservation organizations. The AI doesn't replace human expertise—marine scientists still validate predictions and conduct ground-truth surveys—but it dramatically extends the reach of monitoring programs.

The Data: Temperature, Light, and Stress

Effective coral bleaching analysis combines multiple types of environmental data. Temperature data comes from satellite infrared sensors and in-situ ocean buoys. Light data measures photosynthetically active radiation (PAR) penetrating the water column—excessive light during heat stress amplifies bleaching. Salinity data from conductivity sensors indicates freshwater runoff that stresses corals. Chlorophyll concentration reflects plankton abundance and water clarity.

Students working through this module encounter real datasets: monthly sea surface temperature anomalies, weekly Degree Heating Weeks calculations, satellite imagery in visible and infrared wavelengths, and ocean color measurements. They learn to recognize what healthy and stressed conditions look like in raw data, interpret scientific units and calculations, and understand the limitations of each data source.

For example, satellite temperature readings are estimates based on physics calculations from infrared emissions—they have uncertainty ranges of ±0.5°C. Students learn to ask: How does this uncertainty affect our ability to detect 1°C temperature anomalies? What happens if our threshold for bleaching prediction is wrong by half a degree? This builds scientific reasoning alongside data literacy.

Teacher Resources & Downloads

The Teacher's Guide (linked below) includes detailed lesson plans for each level, discussion prompts, classroom activities, assessment rubrics, and answers to all embedded questions. All datasets used in the module are available for download in CSV and GeoTIFF formats, allowing teachers to create their own extensions and activities.

We provide three types of satellite imagery: natural-color photos (what a camera would see), false-color composites highlighting water conditions, and processed thermal data. Teachers can display these side-by-side to show students what different data representations reveal about the same location.

The module is designed for middle school through advanced high school, with differentiation strategies for younger and older students included in the teacher guide. All activities align with Next Generation Science Standards (NGSS) for life science, earth science, and engineering practices.

Ready to dive in? Start with Level 1 for foundational reef ecology, or jump to the level matching your students' prior experience.