To manage the global carbon budget, we must know not only how much carbon dioxide we emit but also where the "sink"—the portion absorbed by the land and oceans—actually goes. While the physics of atmospheric accumulation is straightforward, estimating the strength of these natural sinks involves a complex tug-of-war between different scientific perspectives. Recent research reveals that the largest uncertainties in our climate projections often stem from fundamental disagreements between how models, inventories, and satellites "see" the Earth.
The Land Sink Paradox
The most striking discrepancy in carbon accounting is the massive gap between global climate models and national greenhouse gas inventories. As Grassi et al. demonstrate, these two approaches differ by approximately 6.7 GtCO2 per year—a figure large enough to flip the land sector from a perceived carbon source to a net sink.
This is not necessarily a failure of measurement, but a difference in definition. Global models typically attribute carbon uptake on "managed" land to natural environmental responses (like CO2 fertilization), whereas national inventories often attribute this same uptake to human management. Obermeier et al. argue that these conflicting definitions create a paradox where the same patch of forest can be classified as a source or a sink depending on the methodology used. Harmonizing these views is essential for policymakers to track progress toward international climate goals accurately.
Top-Down vs. Bottom-Up Divergence
Beyond definitional issues, researchers face a persistent mismatch between "top-down" and "bottom-up" estimation techniques.
- Bottom-up methods, such as those used by Lauerwald et al., rely on process models and ground-based inventories to build an estimate from the soil up.
- Top-down methods, like atmospheric inversions, work backward from observed atmospheric CO2 concentrations to infer where the gas must have been absorbed.
In Europe, Lauerwald et al. and the authors of the consolidated European synthesis have found that bottom-up estimates of the land sink significantly exceed top-down estimates. This suggests that either our ground-based models are overestimating the land's capacity to absorb carbon, or our atmospheric models are missing regional nuances in how CO2 circulates. As noted by Friedlingstein et al. in the 2021 Global Carbon Budget, this lack of agreement is particularly acute in the northern extra-tropics, complicating our understanding of regional carbon sinks.
The Ocean Sink and Interannual Variability
The ocean is generally considered a more stable carbon sink than the land, but it is not immune to methodological conflict. Le Quéré et al. observed that while the oceanic sink is steadily increasing, different estimation methods yield varying amplitudes and phases for its year-to-year variability.
This uncertainty persists in more recent assessments. Friedlingstein et al. highlight that discrepancies remain regarding the mean strength of the ocean sink over the past decade. Furthermore, Le Quéré et al. point out a gap between observational estimates and model outputs concerning interannual variability, suggesting that our current biogeochemistry models may not yet fully capture the ocean's dynamic response to climate change.
The Limits of Remote Sensing
Satellite data offers a potential bridge between these methods, but it introduces its own set of discrepancies. For instance, Shim et al. found that the Tropospheric Emissions Spectrometer (TES) could track regional pollution plumes but struggled with precise day-to-day variability compared to in-situ aircraft measurements. This "smoothing" effect in satellite data can obscure the sharp gradients needed to accurately map carbon sources and sinks at a local level.
Toward a Unified Budget
The recurring theme across these studies—from the 2016 Global Carbon Budget to the 2024 update—is that while our global estimates are becoming more robust, our regional and methodological precision is still lagging. The path forward, as suggested by Grassi et al. and Obermeier et al., lies in "translation" frameworks that adjust model outputs to match the conceptual definitions used in national reporting, thereby closing the gap between scientific theory and policy practice.
Key Foundational Work
Two decades of ocean CO2 sink and variability (2003) — contributed 2 concept elements to this direction.
Satellite observations of Mexico City pollution outflow from the Tropospheric Emissions Spectrometer (TES) (2008) — contributed 1 concept elements to this direction.