Constraints on Local-­to-­Regional Anthropogenic CO2 from Satellite Retrievals of Combustion-­related Trace Gases: Initial Assessment Using Observing System Simulation Experiments (OSSEs)

Abstract:

Quantifying anthropogenic sources of CO2 is imperative yet challenging. Here, we present data assimilation experiments to assess the information gain in using current satellite observations of atmospheric constituents that are co-­emitted during a combustion process. In particular, OSSEs will be conducted to investigate synergistic information from GOSAT (and OCO-2) CO2, MOPITT (and IASI) CO and OMI NO2 retrievals in constraining sources of anthropogenic combustion at city, state, and regional spatiotemporal scales. These experiments will be carried out using ensemble-­‐based data assimilation (DA) system comprising of a regional air quality/weather model, WRF-­Chem, global climate/chemistry model, CAM-Chem, and a data assimilation software package, DART. The ensemble-based DA system, which mimics a numerical weather prediction with chemistry, provides a means to statistically estimate local sensitivities across modeled meteorological and chemical states (CO2, CO, NO2, and related species) and parameters (including surface fluxes) using the ensemble statistics derived from dynamical, physical, and chemical perturbations in the model. We take advantage of these sensitivities in fully exploiting the synergistic information provided by the enhancement ratios that are indicative of combustion characteristic for a given location sampled by these space-­based observations. Here, we introduce a two-­step approach in optimizing anthropogenic CO2. This includes: 1) analysis of atmospheric CO2 distribution using multi-­‐species DA, and 2) Bayesian synthesis time-independent inversion using the CO2 analysis as observational constraints. We evaluate the performance of this approach within an OSSE framework, where one realization of simulated atmosphere is assumed to be the ‘truth’. Synthetic observations are then derived from this atmosphere using sampling and error characteristics of the retrievals. The assimilation of these synthetic observations will be evaluated based on how well the analysis reproduce the ‘true’ anthropogenic CO2. Initial results will be presented comparing the relative contributions of different satellite retrievals in constraining anthropogenic CO2. We will also discuss appropriate sampling characteristics to this effect.