GEOS-Chem Atmospheric Transport Model Research Articles

Ocean-driven interannual variability in atmospheric CO2 quantified using OCO-2 observations and atmospheric transport simulations

Interannual variability (IAV) in the atmospheric CO2 growth rate is caused by variation in the balance between uptake by land and ocean and accumulation of anthropogenic emissions in the atmosphere. While variations in terrestrial fluxes are thought to drive most of the observed atmospheric CO2 IAV, the ability to characterize ocean impacts has been limited by the fact that most sites in the surface CO2 monitoring network are located on coasts or islands or within the continental interior. NASA’s Orbiting Carbon-Observatory 2 (OCO-2) mission has observed the atmospheric total column carbon dioxide mole fraction (XCO2) from space since September 2014. With a near-global coverage, this dataset provides a first opportunity to directly observe IAV in atmospheric CO2 over remote ocean regions. We assess the impact of ocean flux IAV on the OCO-2 record using atmospheric transport simulations with underlying gridded air-sea CO2 fluxes from observation-based products. We use three observation-based products to bracket the likely range of ocean air-sea flux contributions to XCO2 variability (over both land and ocean) within the GEOS-Chem atmospheric transport model. We find that the magnitude of XCO2 IAV generated by the whole ocean is between 0.08-0.12 ppm throughout the world. Depending on location and flux product, between 20-80% of the IAV in the simulations is caused by IAV in air-sea CO2 fluxes, with the remainder due to IAV in atmospheric winds, which modulate the atmospheric gradients that arise from climatological ocean fluxes. The Southern Hemisphere mid-latitudes and low-latitudes are the dominant ocean regions in generating the XCO2 IAV globally. The simulation results based on all three flux products show that even within the Northern Hemisphere atmosphere, Southern Hemisphere ocean fluxes are the dominant source of variability in XCO2. Nevertheless, the small magnitude of the air-sea flux impacts on XCO2 presents a substantial challenge for detection of ocean-driven IAV from OCO-2. Although the IAV amplitude arising from ocean fluxes and transport is 20 to 50% of the total observed XCO2 IAV amplitude of 0.4 to 1.6 ppm in the Southern Hemisphere and the tropics, ocean-driven IAV represents only 10% of the observed amplitude in the Northern Hemisphere. We find that for all three products, the simulated ocean-driven XCO2 IAV is weakly anti-correlated with OCO-2 observations, although these correlations are not statistically significant (p>0.05), suggesting that even over ocean basins, terrestrial IAV obscures the ocean signal.

WOMBAT v1.0: a fully Bayesian global flux-inversion framework

Abstract. WOMBAT (the WOllongong Methodology for Bayesian Assimilation of Trace-gases) is a fully Bayesian hierarchical statistical framework for flux inversion of trace gases from flask, in situ, and remotely sensed data. WOMBAT extends the conventional Bayesian synthesis framework through the consideration of a correlated error term, the capacity for online bias correction, and the provision of uncertainty quantification on all unknowns that appear in the Bayesian statistical model. We show, in an observing system simulation experiment (OSSE), that these extensions are crucial when the data are indeed biased and have errors that are spatio-temporally correlated. Using the GEOS-Chem atmospheric transport model, we show that WOMBAT is able to obtain posterior means and variances on non-fossil-fuel CO2 fluxes from Orbiting Carbon Observatory-2 (OCO-2) data that are comparable to those from the Model Intercomparison Project (MIP) reported in Crowell et al. (2019). We also find that WOMBAT's predictions of out-of-sample retrievals obtained from the Total Column Carbon Observing Network (TCCON) are, for the most part, more accurate than those made by the MIP participants.

Seasonal variability of tropospheric CO2 over India based on model simulation, satellite retrieval and in-situ observation

In this study, investigation of the seasonal cycle of the tropospheric CO2 concentration over India was carried out using the GEOS-Chem atmospheric transport model, Greenhouse gas Observation SATellite (GOSAT) retrievals, and in-situ measurements. The model simulation is highly coherent with the satellite and in-situ datasets, and it shows a distinct seasonal cycle of the tropospheric CO2 tendency over India with a negative phase (decreasing concentration) during April–August and a positive phase (increasing concentration) during September–March. The model diagnostics were analyzed to estimate budgets of the surface layer CO2, up to 650 hPa pressure level, for the two-phases of the seasonal cycle. A mean tendency, equivalent to −0.70 ppmv month−1, observed during April–August, which results from the loss of CO2 content in the surface layer through horizontal advection (−2.25 ppmv month−1) and vertical diffusion (−0.20 ppmv month−1), that dominates the gain from vertical advection (1.53 ppmv month−1). The negative contribution of horizontal advection in this period comes from the transport of CO2 depleted air-parcels over the oceanic region to India by the southwest monsoon winds and the positive contributions of vertical advection comes from upwelling of CO2 enriched air-parcels. The mean tendency, equivalent to 1.01 ppmv month−1, during September–March results from the gain through vertical advection (0.78 ppmv month−1) and horizontal advection (0.37 ppmv month−1) and a small contribution of vertical diffusion (−0.15 ppmv month−1). In this period, positive contribution of horizontal advection is due to the transport of CO2 enriched air-parcels from the southeast Asian region to India by north-east monsoon winds. At the annual scale, CO2 content of the surface layer over India has a net gain of 0.75 GtC that comes from 14.31 GtC through vertical advection that exceeds the loss due to horizontal advection (−11.10 GtC) and vertical diffusion processes (−2.46 GtC). This net gain is almost 85% higher than the input of 0.4 GtC through surface fluxes, which composed of 0.61 GtC anthropogenic emission and −0.21 GtC net terrestrial ecosystem exchanges. Additional sensitivity experiment was carried out to elucidate the semi-annual features of the seasonal cycle of CO2 for north India, in contrast to the annual characteristics of the seasonal cycle for south India in relation to the GOSAT observation.