Spatially Resolving Methane Emissions In California: Constraints From The Calnex Aircraft Campaign And From Present (gosat, Tes) And Future (tropomi, Geostationary) Satellite Observations

Abstract

We apply a continental-scale inverse modeling system for North America based on the GEOS-Chem model to optimize California methane emissions at 1/2°x2/3° horizontal resolution using atmospheric observations from the CalNex aircraft campaign (May-June 2010) and from satellites. Inversion of the CalNex data yields a best estimate for total California methane emissions of 2.86 ± 0.21 Tg a-1, compared with 1.92 Tg a-1 in the EDGAR v4.2 emission inventory used as a priori and 1.51 Tg a-1 in the California Air Resources Board (CARB) inventory used for state regulations of greenhouse gas emissions. These results are consistent with a previous Lagrangian inversion of the CalNex data. Our inversion provides 12 independent pieces of information to constrain the geographical distribution of emissions within California. Attribution to individual source types indicates dominant contributions to emissions from landfills/wastewater (1.1 Tg a-1), livestock (0.87 Tg a-1), and gas/oil (0.64 Tg a-1). EDGAR v4.2 underestimates emissions from livestock while CARB underestimates emissions from landfills/wastewater and gas/oil.

Generated Summary

This study presents a comprehensive analysis of methane emissions in California, focusing on constraints from the CalNex aircraft campaign and satellite observations. The research employs an inverse modeling system based on the GEOS-Chem Eulerian chemical transport model (CTM) to optimize methane emissions at a high spatial resolution of 1/2°x2/3°. The primary goal is to assess the potential of current (GOSAT, TES) and future (TROPOMI, geostationary) satellite observations to constrain the spatial distribution of methane emissions at the state level. The study also aims to compare the constraints achievable with the CalNex aircraft observations to those achievable from satellite data and to evaluate the impact of different inversion methodologies. The core of the methodology involves utilizing the GEOS-Chem model, driven by GEOS-5 meteorological data, to simulate methane concentrations and employing a Bayesian optimization approach to minimize the cost function, which weighs the constraints on emissions from satellite observations against a priori estimates from EDGAR v4.2. The study also used the CalNex aircraft campaign (May-June 2010) and satellite observations to validate the SCIAMACHY observations for North America in an inversion of methane sources. Further, the research performs observation system simulation experiments (OSSEs) to evaluate the capabilities of the TROPOMI and geostationary satellite instruments in constraining California methane emissions. The findings provide insights into the spatial distribution of methane emissions and the effectiveness of different monitoring approaches.

Key Findings & Statistics

• The study found that the best estimate for total California methane emissions is 2.86 ± 0.21 Tg a¯¹, compared to 1.92 Tg a¯¹ in the EDGAR v4.2 emission inventory and 1.51 Tg a¯¹ in the California Air Resources Board (CARB) inventory.
• The inversion of the CalNex data provides 12 independent pieces of information to constrain the geographical distribution of emissions within California.
• Attribution to individual source types indicates that the dominant contributions to emissions come from landfills/wastewater (1.1 Tg a¯¹), livestock (0.87 Tg a¯¹), and gas/oil (0.64 Tg a¯¹).
• EDGAR v4.2 underestimates emissions from livestock, while CARB underestimates emissions from landfills/wastewater and gas/oil.
• The study showed that livestock emissions increase statewide by 92% relative to EDGAR, landfill/wastewater by 28%, and gas/oil by 26%.
• GOSAT observations were used for the CalNex period, 1 May to 22 June 2010, with 257 GOSAT and 133 TES observations on the GEOS-Chem grid.
• The study found that the inversion had 1.3 DOFS (Degrees of Freedom for Signal) for the GOSAT observations, compared to 12.2 DOFS for the CalNex observations.
• In the Los Angeles Basin, the study found an emission estimate of 0.42 ± 0.08 Tg a¯¹.
• The study notes that the best-fit of the US anthropogenic emissions is from the 1000 cluster inversion, with US anthropogenic emissions at 32.0 ± 1.3 Tg a¯¹.
• The model-observation root mean square difference (RMSD) decreased from 11.6 to 9.7 ppb, while R increased from 0.65 to 0.76, demonstrating improvement limited by the random noise in the SCIAMACHY measurements.
• The model-observation RMSD for individual observations decreased from 33.5 to 28.5 ppb, while R increased from 0.73 to 0.74, with the improvement limited by small-scale model and representation error for individual observations.
• The resulting model-observation RMSD weighted by the number of INTEX-A observations in each 8°x10° grid cell decreased from 23.2 to 12.3 ppb when using the optimized versus the a priori emissions.
• The study found that the optimized emission estimate for the Canadian wetlands is lower than the a priori, but consistent with other recent studies.
• The optimized US anthropogenic emissions were 32.0 ± 1.3 Tg a¯¹, with the best estimate from the 1000-cluster inversion and its error standard deviation from the ensemble of inversions with different numbers of clusters in Figure 3.6.
• The study showed that the state total emissions is larger than their 2.37 ± 0.27 Tg a¯¹ but this appears to reflect their use of a lower a priori inventory.
• The study reveals that the patterns of the correction factors from the inversion are highly spatially correlated through the use of homogeneous perturbations in large blocks.

Other Important Findings

• The study’s findings indicate that the state total livestock emission of 0.87 Tg a¯¹ is in close agreement with CARB, but much higher than EDGAR, and lower than the 1.29 Tg a¯¹ estimate of Santoni et al. (2013).
• The research found that the patterns of correction factors cannot be explained simply by the EDGAR v4.2 source types.
• The study observed that the TROPOMI may perform as well as a dedicated aircraft campaign (CalNex), and is thus superbly positioned to constrain emissions at the state level.
• The GEO-CAPE inversion has 26.5 DOFS, much higher than CalNex and TROPOMI, reflecting the greater density of observations.
• The study also finds that the patterns of correction factors from the inversion in Figure 3.5 reveal structure that cannot be simply explained by the EDGAR v4.2 source types.
• The study highlights the large discrepancies between CARB and EDGAR for different source types.
• The study concludes that current satellite observations of methane from GOSAT and TES are too sparse to quantitatively constrain California emissions.
• The study indicates that the TROPOMI instrument will provide global daily coverage with 7×7 km² nadir resolution, and it’s expected to improve the ability to monitor methane emissions from space.

Limitations Noted in the Document

• The study acknowledges that the accuracy of the inversion is limited by the accuracy of the a priori inventory, since the method relies on the EDGAR v4.2 emission inventory.
• The study mentions that the high spatial correlation of a priori errors might limit the effectiveness of the inversion in correcting emission estimates.
• The research recognizes that the reliance on a single averaging kernel from GOSAT to generate synthetic observations for both instruments could introduce uncertainties.
• The study indicates that the chosen error covariance matrix for the observations and a priori emissions could influence the final emission estimates, as it relies on a uniform relative error standard deviation.
• The findings are subject to limitations associated with the spatial resolution of the GEOS-Chem model, which might not fully capture small-scale variability in emissions.

Conclusion

The study’s findings underscore the importance of accurately characterizing methane emissions, particularly from livestock, and highlight the potential of satellite observations for improving methane monitoring. The research provides a robust estimate of California’s methane emissions, showing a value of 2.86 ± 0.21 Tg a¯¹, which is higher than previous estimates from CARB. The analysis reveals the critical need for improving emission inventories and the benefit of using different datasets and methods for verification and validation, with the optimized emissions, indicating a significant impact on the contribution from livestock, landfills/wastewater, and gas/oil sectors. The study’s insights into the spatial distribution of California methane emissions, supported by high-resolution observations, provide valuable insights. The TROPOMI instrument, as well as geostationary satellite observations, are highlighted as promising advancements. The study’s emphasis on the limitations of the current monitoring techniques and the importance of understanding the spatial variability of emissions provides valuable insights for future monitoring efforts and supports the use of independent measurements in inversion methodologies, highlighting the need for more detailed and high-resolution emission data for effective climate mitigation strategies. The researchers emphasize that the study supports the use of independent measurements, such as satellite data, in inversion methodologies, thereby enabling a greater spatial detail in assessing methane emissions on a continental scale and supporting more accurate emission estimations.