TL;DR
This paper quantifies the ocean and land carbon dioxide (CO2) sinks over the 1980s and 1990s, revealing that while atmospheric CO2 growth was similar, the combined sinks absorbed significantly more CO2 in the 1990s. The research highlights a more robust and steadily increasing oceanic sink compared to a more variable land sink, offering crucial insights for understanding and predicting future climate change.
What Problem Does This Paper Address?
The global carbon cycle is a complex system where atmospheric CO2 concentrations are influenced by emissions from human activities (like burning fossil fuels) and uptake by natural sinks (oceans and land). Accurately quantifying these sinks is vital for predicting future atmospheric CO2 levels and, consequently, the pace of climate change. While fossil fuel emissions increased significantly from the 1980s to the 1990s, the rate of atmospheric CO2 increase remained surprisingly similar. This implies that the combined ocean and land sinks must have absorbed more CO2 in the 1990s. However, precisely how much each sink contributed to this increased uptake and how their variability differs remained unclear. This paper aims to resolve this by comparing multiple state-of-the-art methods to quantify the ocean and land CO2 sinks for these two decades.
Methods
The authors employ a multi-faceted approach, comparing four distinct methods to estimate the CO2 sinks:
- Atmospheric Inversions: These methods use atmospheric CO2 measurements and sophisticated transport models to infer the geographical distribution and temporal variability of CO2 fluxes (sinks and sources) on land and at sea.
- CO2-only inversion (CO2-INV-IPSL): Relies solely on atmospheric CO2 concentration data.
- Multi-tracer inversions (MULTI-INV-CSIRO and MULTI-INV-SCRIPPS): Incorporate additional isotopic data (like 813C) and/or oxygen measurements (O2/N2) to better distinguish between land and ocean fluxes and improve regional estimates. These are updates of previous inversion techniques.
- Ocean Models: These use physical and biogeochemical models of the ocean to simulate the exchange of CO2 between the atmosphere and the ocean.
- OGCM-interannual: A three-dimensional ocean general circulation model (OGCM) forced with re-analyzed atmospheric and oceanic data to capture interannual variability.
- OGCM-clim: The same OGCM but forced with climatological data to represent the long-term trend in the oceanic CO2 sink due to increasing atmospheric CO2, without natural variability.
- 1-D models: Simpler, one-dimensional models used to isolate the effect of factors like sea surface temperature on CO2 uptake.
The paper also discusses adjustments made to account for the transport of CO2 from land to oceans via rivers, which is not directly captured by atmospheric measurements.
Datasets
The study utilizes several key datasets:
- Atmospheric CO2 concentrations: Provided by GLOBALVIEW, including interpolated data.
- Atmospheric 813C measurements: Used in multi-tracer inversions, with data from stations like Cape Grim.
- Atmospheric O2/N2 measurements: Also used in multi-tracer inversions.
- Oceanic data: Including temperature, salinity, and nutrient observations (e.g., Levitus et al. datasets).
- Meteorological data: Wind stresses and fluxes from ERS satellite and NCEP/ECMWF re-analyzed data.
- Historical CO2 concentrations: From ice cores for periods before direct measurements.
The paper does not introduce new benchmark datasets but rather synthesizes and compares results from existing observational and modeling efforts.
Key Insights and Ideas
- The 1990s were a "super sink" decade: Despite a significant increase in fossil fuel emissions, the combined ocean and land CO2 sinks absorbed substantially more carbon in the 1990s than in the 1980s, leading to a similar atmospheric CO2 growth rate.
- Oceanic sink is more stable and increasing: When short-term variability is smoothed out, the oceanic CO2 sink shows a more consistent and steadily increasing trend over the two decades.
- Land sink is more variable: The land sink exhibits larger year-to-year fluctuations, with notable anomalies like a reduced sink during the 1990-1994 period.
- Oceanic variability is less pronounced but poorly constrained: While the ocean's CO2 sink is less variable than the land's, the exact amplitude and phase of this variability remain uncertain across different estimation methods.
- Convergence on the trend, divergence on the mean: While there's good agreement on the increase in the oceanic sink from the 1980s to the 1990s, the absolute mean sink value for each decade still carries significant uncertainty.
Key Results
- Decadal difference: The ocean sink was estimated to be 0.3 Pg C yr⁻¹ larger in the 1990s than the 1980s, while the land sink was 0.7 Pg C yr⁻¹ larger.
- Global sink increase: The total ocean and land sink was approximately 1 Pg C yr⁻¹ larger in the 1990s compared to the 1980s.
- Oceanic sink trend: Ocean models consistently show an increasing oceanic CO2 sink of about 0.3 Pg C yr⁻¹ per decade, driven by rising atmospheric CO2.
- Variability ratio: All methods indicate that the land sink has significantly higher interannual variability (1/3 to 1/2 more) than the oceanic sink.
- Mean oceanic sink estimate: Supported by O2 observations, a mean oceanic sink of 1.9 Pg C yr⁻¹ for the 1990s is suggested, though with an uncertainty of ±0.7 Pg C yr⁻¹.
Limitations and Future Work
- Uncertainty in atmospheric transport models: This remains a significant challenge for atmospheric inversion methods, impacting the accuracy of regional flux estimates.
- Data quality and station network: The accuracy of multi-tracer inversions is sensitive to the quality and distribution of isotopic and CO2 measurements.
- Ocean model limitations: Ocean models can underestimate variability in ocean circulation, especially at high latitudes, and are subject to uncertainties in parameterizations and climate feedbacks.
- Regional flux determination: While global trends are becoming clearer, accurately quantifying regional CO2 fluxes and their variability remains a challenge, particularly for distinguishing between ocean and land contributions.
- Future work: The authors suggest that a more accurate quantification of oceanic CO2 flux and variability at the regional level is needed to reduce the uncertainty in mean ocean and land CO2 sinks. Further comparisons of atmospheric inversion and ocean model results are also necessary to resolve discrepancies.
Summary updated