Mass coral bleaching threatens biodiversity across western Indian Ocean

Kenya’s coral reefs have been hit by the most severe bleaching event ever recorded in the country, with more than three-quarters of reef-building corals showing signs of thermal stress during March–April 2024. The findings reveal catastrophic declines in healthy coral cover along 600 km of the nation’s Swahili coastline.

The study, titled “A Citizen Science Approach for Documenting Mass Coral Bleaching in the Western Indian Ocean”, deployed a combined citizen science and artificial intelligence (AI) methodology to measure the impact of an unprecedented marine heatwave. By training the CoralNet AI platform to analyze thousands of underwater images collected across 22 sites, the research offers a rapid-response assessment of ecological damage in a region where conventional surveys often lag behind unfolding crises.

How the 2024 heatwave redefined coral bleaching thresholds

The investigation began in early March 2024 after local divers and scientists reported bleaching despite only minimal degree-heating weeks (DHWs) according to NOAA’s Coral Reef Watch estimates. These anomalies prompted Mayfield to conduct on-the-ground surveys, capturing 2,300 reef images between Mombasa and Lamu. The images were processed by AI trained on hundreds of manually annotated samples, achieving over 80% classification accuracy.

The results revealed a consistent underestimation of bleaching severity by NOAA’s models. While satellite-based measurements indicated up to 13.5 DHWs by mid-April, in situ dive computer readings suggested temperatures 1.3°C higher, equating to nearly 22 DHWs—levels unmatched since NOAA began tracking the region in 1985. This gap in thermal stress estimates proved critical: bleaching thresholds were crossed earlier and more intensely than forecasts predicted, leaving corals with little time to acclimate.

Temperature stress was not evenly distributed. Mombasa’s reefs, surveyed earlier in the month with fewer DHWs, showed significantly less bleaching than Lamu’s reefs, which were surveyed later after prolonged heat exposure. Nevertheless, all regions experienced widespread bleaching, and depth patterns emerged as a secondary factor. Shallow reefs under intense light exposure, particularly in Watamu and Lamu, suffered the most damage, compounding the effects of extreme heat.

AI-backed citizen science offers new monitoring model

This approach blended grassroots participation with automation to overcome the logistical limitations of traditional monitoring programs. By using AI-assisted image analysis, the study was able to quantify coral health and bleaching prevalence across an estimated 11–12 hectares of reef, far more spatial coverage than is typically possible in formal transect-based surveys.

Tourist and long-term monitoring sites showed similar results, with mean bleaching prevalence of around 77% and healthy coral cover averaging just 7–8% of the benthos. This figure marks a steep decline from more than 30% healthy coral cover in the late 1990s. Even when adjusting for the AI’s false-positive bias, where partially pale corals were sometimes classified as bleached, the healthiest scenario still left only 13–14% healthy coral cover, representing a 50% drop from the previous year and a two-thirds decline since the 1990s.

Importantly, the study’s meta-analysis of historical data contextualized the 2024 crisis against past bleaching events. The 1998 event wiped out over 90% of shallow corals in some areas, while the 2016 global bleaching event caused significant but less widespread losses. By comparison, 2024’s bleaching levels were higher than 2016’s in nearly every surveyed region. In Lamu, where the 2016 event had already reduced coral cover by more than half, the 2024 event bleached nearly 90% of coral tissues.

This rapid-assessment model, relying on accessible equipment such as consumer-grade underwater cameras and free AI platforms, could be replicated in other reef-dependent nations. By training divers, both professional and recreational, in key techniques like white-balancing and targeted image capture, large-scale ecological monitoring can be conducted even in areas lacking formally trained marine scientists.

Long-term consequences and urgency for action

The study warns that the full impact of the 2024 bleaching will only be measurable in subsequent mortality surveys. If all bleached corals perish, Kenya’s reefs will face an average 75% reduction in healthy coral cover compared to the late 1990s. Even if half recover, the losses would still be in the range of 35–50%.

This scale of decline has major ecological and socio-economic implications. Coral reefs in the Western Indian Ocean support critical biodiversity, fisheries, and tourism, yet they are increasingly vulnerable to climate-driven marine heatwaves. The 2024 event demonstrated how quickly conditions can deteriorate, outpacing both predictive models and the capacity of traditional monitoring efforts.

While some sites showed localized improvements due to active restoration programs, such as those run by REEFoLution, these gains were outweighed by regional losses. Historical data compiled by the study indicate that healthy coral cover across Kenya has trended downward for three decades despite intermittent recovery periods.

The research underscores the importance of integrating citizen science into national reef monitoring frameworks, particularly in the context of rapid-onset environmental events. Empowering thousands of divers with basic training and AI-assisted analysis tools could provide near-real-time data for conservation agencies, policymakers, and climate resilience planners.

Looking ahead, the reduced heat stress recorded in early 2025 may give surviving corals a reprieve, but the study cautions that future marine heatwaves are inevitable. Without swift adaptation measures, ranging from improved predictive modeling to expanded restoration and monitoring, the next bleaching event could further erode the resilience of Kenya’s coral reef ecosystems.