High-Resolution Mapping of Fossil Fuel CO ₂ Using Plant Radiocarbon and Bayesian Inversion: Toward a City-Scale Emission Audit

Accurate quantification of fossil fuel carbon dioxide (CO_2ff) emissions is essential for evaluating mitigation progress and informing climate policy. However, urban-scale inventories often face significant uncertainties, highlighting the urgent need for robust, independent validation. Here we present a high-resolution mapping of CO_2ff emissions in Shenzhen using radiocarbon (¹⁴C) observations integrated with Bayesian inversion. We collected 70 herbaceous plant samples across the city in 2022 on a 5 × 5 km² grid to capture spatial gradients in CO_2ff. The plant ¹⁴C-derived CO_2ff signals were assimilated in a Bayesian framework, and multiple sensitivity tests were performed to assess robustness. The inferred emissions show clear dominance of industrial sources, supported by strong spatial correlations with industrial facility density and coemitted pollutants (PM_2.5, PM₁₀, and NO₂). The inversion yields a citywide total of 59.2 ± 4.0 Mt CO₂/year, bracketed by the lower MEIC inventory and the substantially higher ODIAC estimate. These results demonstrate that herbaceous plant ¹⁴C measurements provide a cost-effective, scalable constraint on urban CO_2ff at fine spatial resolution. When coupled with Bayesian inversion, this observation-driven approach enables robust, independent validation of bottom-up inventories and supports city-scale carbon auditing and science-based governance, particularly in data-limited urban regions.