Evaluating In Situ and OCO-3 CO2 Observations in the Sichuan Basin Using WRF-Chem Simulations and Footprint Analysis

Abstract

Revealing the spatiotemporal patterns and controlling factors of atmospheric CO2 in the Sichuan Basin (SCB) is a prerequisite for the scientific regulation of regional carbon sources and sinks. This study investigates surface and column-averaged CO2 concentrations in SCB and surrounding regions during August to December 2024, using in situ measurements, Orbiting Carbon Observatory-3 (OCO-3) column-averaged CO2 retrievals, and a Weather Research and Forecasting model coupled with Chemistry (WRF-Chem) simulations and footprint analysis. At the national atmospheric background site JinFoShan (JFS), observed CO2 concentrations rose from ~404 ppm in mid-August to around 440 ppm by late December. Comparisons between observations at JFS and WRF-Chem show that simulations driven by Open-source Data Inventory for Anthropogenic CO2 (ODIAC) fossil fuel emissions and Vegetation-Global-Atmosphere-Soil (VEGAS) biosphere fluxes yield the highest Pearson correlation coefficient (r = 0.77). The diurnal and seasonal variabilities at JFS are primarily controlled by biospheric fluxes, with fossil fuel contributions weak in August to September and slightly enhanced in October to December, partly due to regional transport from downtown Chongqing revealed by footprint analysis. By contrast, Yongchuan station, closer to urban Chongqing, shows mean CO2 levels ~10 ppm higher than JFS. For satellite observations, OCO-3 Snapshot Area Map (SAM) measurements over the SCB suffer from substantial missing samples and high spatial noise. We find no spatial correlation exists between SAM retrievals and simulations over Chongqing, and only weak positive correlations appear over Chengdu using ODIAC and Gridded Fossil Emissions Datasets (GridFEDs), whereas results based on the Multi-resolution Emission Inventory for China (MEIC) show no correlation, likely related to biased suburban emission spatial distributions. Overall, the spatial correlations between SAM retrievals and model simulations are weak, ranging from −0.22 to 0.31. To our knowledge, this is the first study to systematically compare in situ observations and OCO-3 SAM retrievals with WRF-Chem simulations in the Sichuan Basin.