Global land carbon sink halves in 2024, AI model reveals shocking findings
A groundbreaking study led by researchers at Peking University's Institute for Carbon Neutrality (ICN) has uncovered a startling revelation about our planet's carbon cycle. Using advanced AI models, the team, comprising Wang Heyuan and Wang Kai, has determined that the global land carbon sink has drastically diminished due to a sudden and extreme global temperature rise. Their research, titled 'AI-tracked halving of global land carbon sink in 2024', was published in the prestigious Science Bulletin.
The significance of this discovery cannot be overstated. Terrestrial ecosystems, which absorb nearly one-third of human-induced carbon emissions annually, are crucial in regulating the global carbon cycle. As climate extremes become more frequent, the need for accurate and timely tracking of these ecosystems' changes is paramount.
The AI models employed by the team offer a remarkable solution. By overcoming the typical one-year lag associated with traditional assessment methods, these models enable near-real-time detection and diagnosis of carbon cycle responses. This breakthrough enhances our understanding of the Earth's coupled carbon-climate system, paving the way for more informed and responsive policy-making.
Key findings from the study are eye-opening. In 2024, the land carbon sink plummeted to less than half of its average level over the previous decade, with the tropics experiencing particularly severe reductions. Within tropical ecosystems, grasslands and savannas suffered proportionally greater losses compared to tropical rainforests, indicating that semi-arid regions under prolonged drought are not as resilient as previously thought. Further analysis suggests that heat and drought-induced vegetation productivity declines are the primary drivers behind the tropical carbon sink reduction.
These findings have profound implications. They highlight the vulnerability of tropical land systems, especially semi-arid grasslands and savannas, which may be more susceptible to extreme events than previously assumed. This vulnerability could have far-reaching consequences for atmospheric CO₂ levels, potentially exacerbating global warming. As these AI-driven insights are integrated with atmospheric inversions and ground observations, they can inform adaptive land-management strategies, stress-test various climate scenarios, and facilitate more agile policy interventions.
The study's DOI is 10.1016/j.scib.2025.10.015, and it can be accessed at [https://dx.doi.org/10.1016/j.scib.2025.10.015]. The research underscores the urgent need for innovative approaches to monitoring and managing our planet's carbon sinks in the face of a rapidly changing climate.
