TL;DR
This paper investigates the drivers behind the recent increase in atmospheric methane concentrations between 2006 and 2008. By combining two atmospheric inversion models and a wetland ecosystem model, the study finds that tropical natural wetlands were the primary source of this increase, particularly in 2007, with significant contributions from high latitudes.
What Problem Does This Paper Address?
For a decade prior to 2007, atmospheric methane concentrations had been relatively stable. However, starting in 2007, a noticeable increase in the growth rate of atmospheric methane was observed. This surge prompted scientific inquiry into its causes, as methane is a potent greenhouse gas. The challenge lies in pinpointing the specific sources responsible for this interannual variability. While anthropogenic emissions tend to change gradually, natural sources like wetlands and biomass burning are more susceptible to year-to-year climate fluctuations. Distinguishing between these sources and quantifying their contributions to the observed methane increase is crucial for understanding and predicting future climate change.
Methods
The study employs a multi-pronged approach:
- Two Atmospheric Inversion Models (INV1 and INV2): These models use atmospheric observations of methane and a chemical transport model to infer the distribution of methane sources and sinks.
- INV1: A computationally efficient, large-region inversion model that separates emissions by source type (wetlands, other natural, biomass burning, anthropogenic). It uses monthly mean observations from a network of stations.
- INV2: A variational inversion model that uses a finer grid and assimilates daily mean observations, providing a more detailed spatial resolution of emissions. It infers total methane flux rather than separating source types.
- Wetland Ecosystem Model (ORCHIDEE): This model simulates methane emissions from natural wetlands based on physical processes like precipitation, temperature, and soil moisture, incorporating factors like wetland extent and flux density.
The study compares the results from these models to attribute the observed methane anomalies during 2006-2008 to specific sources and regions.
Datasets
- Atmospheric Methane Observations: Data from surface monitoring networks (NOAA/ESRL, CSIRO, IPSL/LSCE) were used for the inversion models.
- SCIAMACHY Satellite Data: Used for independent validation of atmospheric methane column-averaged mixing ratios, particularly over tropical regions.
- Climate Data: CRUNCEP dataset (combining CRU and NCEP data) was used to drive the ORCHIDEE wetland model.
- Inventories: Prior emissions were taken from existing inventories (e.g., EDGAR4.1, Matthews and Fung, van der Werf).
- Satellite-derived Inundation Maps: Used to constrain wetland extent in the ORCHIDEE model.
Key Insights and Ideas
- Wetlands as the Primary Driver: The study strongly suggests that natural wetlands, particularly in tropical regions, were the main cause of the increased atmospheric methane growth rate observed from 2007 onwards. This highlights the significant sensitivity of wetland methane emissions to climate variability.
- Tropical Dominance: Tropical emissions, largely driven by wetlands, were the dominant contributor to the global methane anomaly in 2007, accounting for 60-80% of the increase.
- Latitudinal Partitioning Matters: The study emphasizes the importance of atmospheric transport in interpreting flux anomalies. While tropical emissions are diluted by atmospheric mixing, high-latitude emissions have a more pronounced impact on surface concentrations, even if their absolute magnitude is smaller.
- Model Intercomparison is Crucial: Discrepancies between the ecosystem model and inversions, especially in 2006 and 2007, underscore the need for continued refinement of both approaches and better integration of observational constraints.
- OH Sink Stability: Changes in the hydroxyl radical (OH), the primary methane sink, were found to be minimal (<1%) during the study period, indicating they were not a significant factor in the observed methane increase.
Key Results
- Global Anomalies:
- 2006: Negative anomaly (-10±4 Tg for INV1, -3 Tg for INV2).
- 2007: Positive anomaly (21±4 Tg for INV1, 20 Tg for INV2).
- 2008: Positive anomaly (17±4 Tg for INV1, 19 Tg for INV2).
- Overall average for 2006-2008 compared to 1999-2006: -7 Tg (2006), +21 Tg (2007), +18 Tg (2008).
- Tropical Contribution: In 2007, tropical emissions accounted for 60-80% of the global anomaly, with natural wetlands contributing ~2/3 of this.
- High Latitude Contribution: High latitudes contributed approximately 25% to the 2007 global anomaly.
- Model Agreement: Agreement between the wetland ecosystem model (ORCHIDEE) and inversions was good in 2008 but less satisfactory in 2006 and 2007.
- Key Regions: Tropical South America and Boreal Eurasia were consistently identified as major contributors to methane emission variations across models.
Limitations and Future Work
- Latitudinal Partitioning Discrepancies: While global trends were consistent, the agreement on the latitudinal distribution of flux anomalies was only fair in 2006 and not good in 2008, highlighting challenges in regional source attribution.
- Model Uncertainties: The ORCHIDEE model's representation of wetland extent and its sensitivity to precipitation, particularly in the tropics, may be an oversimplification. The paper suggests that tropical floodplains might not be explicitly represented, and the model might overstate the direct impact of precipitation on methane emissions.
- Data Sparsity: Limited surface observations in key tropical regions (South America, Africa) make inversion estimates for these areas more uncertain.
- Future Work: The authors suggest that further investigation is needed into the relative weighting of different observation types in inversion models. They also point out that while natural wetlands can modulate methane growth rates over a few years, the continued increase in anthropogenic emissions means methane is likely to continue rising on decadal timescales, even without considering potential releases from marine hydrates or permafrost. The study also highlights the need for better modeling of atmospheric transport to accurately attribute fluxes to their source regions.
Summary updated