Estimating Submicron Aerosol Mixing State at the Global Scale with Machine Learning and Earth System Modeling

This study integrates machine learning and particle-resolved aerosol simulations to develop emulators that predict sub-micron aerosol mixing state indices from the Earth System Model (ESM) simulations. The emulators predict aerosol mixing state using only ESM bulk aerosol species concentrations, which do not by themselves carry mixing state information. Here we used PartMC as the particle-resolved model and NCARs CESM as the ESM. We trained emulators for three different mixing state indices for sub-micron aerosol in terms of chemical species abundance (χa), the mixing of optically absorbing and non-absorbing species (χo), and the mixing of hygroscopic and non-hygroscopic species (χh). Our global mixing state maps show that there is considerable spatial and seasonal variability in mixing state indices, ranging between 23% and 96% for χa, between 49% and 95% for χo, and between 19% and 90% for χh, with averages of 76%, 75%, and 63%, respectively. High values in one index can be consistent with low values in another index depending on the grouping of species and their relative abundance, meaning that each mixing state index captures different aspects of the population mixing state. Although a direct validation with observational data has not been possible yet, our results are consistent with mixing state index values derived from ambient observations. This work is a prototypical example of using machine learning emulators to add information to ESM simulations.