Black carbon (BC) is a harmful air pollutant released from burning fossil fuels and biomass. It contributes to climate warming and negatively impacts human health. However, estimating BC emissions accurately is difficult because of their large variability over space and time and limited monitoring data—especially for tracing pollution back to specific sources. While many monitoring networks report elemental carbon (EC) as a proxy for BC, these data often lack the detail needed for source attribution.

To address this, we modified a chemical transport model called CMAQ by adding capabilities to track EC emissions from individual sources, such as power plants, vehicles, and residential wood burning. We then applied a logarithmic Bayesian inversion method, which combines model outputs with measured EC and BC data, across California and a special focus on the San Francisco Bay Area. This approach improves the stability of the results and helps refine emission estimates at high spatial resolution.

Our preliminary results suggest current emission inventories may underestimate sources like power generation and diesel vehicles, while overestimating emissions from residential wood burning. This study demonstrates how combining models with observations through advanced statistical methods can improve emission estimates and support more effective air quality and climate policies.