Abstract. Atmospheric CO2 concentrations in urban areas reflect a combination of fossil fuel emissions and biogenic fluxes, offering a potential approach to assess city climate policies. However, atmospheric models used to simulate urban CO2 plumes face significant uncertainties, particularly in complex urban environments with dense populations and vegetation. This study addresses these challenges by analyzing CO2 dynamics in the Metropolitan Area of São Paulo (MASP) using the Weather Research and Forecasting model with Chemistry (WRF-Chem). Simulations were evaluated against ground-based observations from the METROCLIMA network, the first greenhouse gas monitoring network in South America, and column concentrations (XCO2) from the OCO-2 satellite spanning February to August 2019. To improve biogenic fluxes, we optimized parameters in the Vegetation Photosynthesis and Respiration Model (VPRM) using eddy covariance flux measurements for key vegetation types, including the Atlantic Forest, Cerrado, and sugarcane. Results show that at the urban site (IAG), the model consistently underestimated CO2 concentrations, with a negative mean bias of −9 ppm throughout the simulation period, likely due to the complexity of vehicular emissions and urban dynamics. In contrast, at the vegetated site (PDJ), simulations showed a consistent positive mean bias of 5 ppm and closely matched observations. Seasonal analyses revealed higher CO2 concentrations in winter, driven by greater atmospheric stability and reduced vegetation uptake estimated by VPRM, while summer exhibited lower levels due to increased mixing and higher agricultural productivity. A comparison of biogenic and anthropogenic scenarios highlights the need for integrated emission modeling and improved representation of biogenic fluxes, anthropogenic emissions, and boundary conditions for high-resolution modeling in tropical regions.
Alves et al. (2025) studied this question.