Marine Carbon Sequestration: Modelling the Global Burial Hotspots

Continental margins may capture a staggering 367 million tonnes of carbon annually. This is the central claim of a new assessment integrating global datasets with predictive algorithms. Historically, quantifying marine carbon sequestration has been a logistical nightmare, plagued by the sheer scale of the ocean and the difficulty of obtaining physical samples from the seabed. Previous estimates were often fragmented, leaving researchers to guess at the total capacity of these underwater storage vaults.

Defining Marine Carbon Sequestration Limits

The study moves beyond simple data collection. By employing spatial machine learning and three-dimensional reaction–transport modelling, the researchers attempt to fill the gaps left by sparse observations. The model suggests that continental-margin sediments act as one of the planet's most persistent sinks, exceeding the storage capacity of lakes, reservoirs, and vegetated coastal ecosystems combined. However, the efficiency of this trap appears to decay over time. The estimated flux drops from 367 Tg C yr⁻¹ on centennial scales to 246 Tg C yr⁻¹ on multi-millennial scales, indicating that while sediment captures carbon quickly, it does not hold all of it indefinitely.

A distinct methodological shift separates this work from earlier attempts. Traditional assessments relied heavily on direct point-source observations—taking a core sample here and there and extrapolating the results across thousands of kilometres. This 'old method' is prone to massive error bars due to environmental heterogeneity. The new method contrasts this by using machine learning to synthesise those points into a continuous global map. Instead of relying on raw, scattered data (akin to the biological specificity of gene markers in a different field), the model uses reaction–transport equations to simulate how carbon behaves chemically and physically (similar to analysing broad GC content trends, but applied here to sediment flux). This computational layer aims to smooth out the noise of 'extraordinary environmental variability', offering a cohesive, albeit simulated, view of global burial.

Scepticism remains necessary regarding the geographical distribution. The model allocates approximately 70% of carbon burial to the Northern Hemisphere and 50% to equatorial regions. While 'pronounced hotspots' are identified in tropical margins, one must ask if this Northern skew reflects geological reality or a bias in the training data. If the majority of historical observations come from the Northern Hemisphere, the machine learning algorithm will naturally weigh these regions more heavily. The study provides a spatial framework, but until these hotspots are physically verified, they remain mathematical probabilities rather than confirmed reservoirs.

These findings refine the coastal carbon budget. They imply that ocean-atmosphere CO₂ exchange is more heavily influenced by sediment burial than previously thought. For policymakers, this suggests that protecting continental margins is not just a matter of biodiversity, but a requirement for maintaining the global climate thermostat.