Bayesian inversion of urban sectoral scale anthropogenic CO ₂ emissions coupled with weather research and forecasting (WRF) and stochastic time-inverted lagrangian transport (STILT) models in Chengdu-Chongqing economic circle

Greenhouse gas emissions from anthropogenic activities, especially CO2, are the primary cause of global warming. Accurate estimation of urban-scale anthropogenic carbon emissions is critical for developing emission reduction policies and achieving carbon neutrality. This study focuses on the top-down inversion of anthropogenic CO2 emissions in the Chengdu-Chongqing Economic Circle (CCEC) from December 2019 to May 2020, using a coupled Weather Research and Forecasting (WRF) model and Stochastic Time-Inverted Lagrangian Transport (STILT) model. The model integrates EDGAR and GRACED prior carbon emission inventories with CO2 concentration observations and applies a multi-ratio factor Bayesian optimization algorithm to invert sectoral carbon emissions. Results show clear temporal and spatial variations in footprint weights, and the WRF-STILT model effectively simulates CO2 concentrations at hourly and daily scales. Simulations based on GRACED are closer to observed values than those from EDGAR, with enhancements ranging from 8 to 47 ppm. The industrial sector contributes most to CO2 increases, followed by the power sector. CO2 concentrations from GRACED show a correlation of over 0.94 with observations, indicating strong tolerance to concentration errors. The WRF-STILT model enables accurate sectoral emission inversion with a small constraint (±2 ppm) on atmospheric CO2 concentrations.