Bayesian Top-Down Pattern-Restricted Estimates of Atmospheric Microplastics Emissions using Gibbs sampler

We present a top-down estimate of atmospheric microplastics (MPs) emissions based on deposition measurements, optimized against an atmospheric transport model (ATM). The central challenge of this work is the severe ill-posedness of the spatial-temporal inverse problem, as emissions cannot be uniquely inferred from the limited number of available measurements. To regularize the inversion, we constrain emissions to follow physically motivated source patterns associated with global road dust, agricultural activities, bare soils, and ocean surface, while estimating their strengths. The relationship between emissions and measurements is established using source–receptor sensitivity (SRS) fields calculated using the ATM Flexpart 11. To estimate source strengths and rigorously quantify uncertainties, we employ a Bayesian inversion framework with a hierarchical prior model, whose parameters are inferred using Gibbs sampler. This approach avoids excessive tuning and enables a realistic representation of uncertainty arising from measurements, transport modeling, and emission assumptions. The inferred atmospheric MPs emissions ranges are broadly consistent with existing literature and measurements from different areas around the world, and the framework provides a transparent and robust quantification of uncertainty in global atmospheric MPs emissions.

 

Acknowledgment:

This research has been supported by the Czech Science Foundation (grant no. GA24-10400S). N.E. were funded by the Norwegian Research Council (NFR) project MAGIC (No.: 334086). FLEXPART model simulations are cross-atmospheric research infrastructure services provided by ATMO-ACCESS (EU grant agreement No 101008004). The computations were performed on resources provided by Sigma2 - the National Infrastructure for High Performance Computing and Data Storage in Norway.