Cloud condensation nuclei (CCN) connect aerosol populations to cloud formation. My work asks how observations and physical-model knowledge can help estimate CCN when detailed particle-size information is unavailable.
Published foundation
In first-author work published in Atmospheric Chemistry and Physics in 2020, I developed machine-learning estimates of CCN number concentrations from commonly available measurements. A 2021 first-author study in Geophysical Research Letters extended the analysis to aircraft observations across contrasting environments. It showed that aerosol chemistry and meteorology contain information about particle size that can help quantify cloud-forming particles.
These studies combine atmospheric physics with data-driven inference: the scientific goal is to understand which measured properties carry useful information about the aerosol population. In collaborative work published in 2022, we also examined machine learning in a climate model to reduce uncertainty in predicted particle number and aerosol indirect radiative forcing.
Next questions
My next priority is to develop and evaluate efficient representations of aerosol optical properties and particle-size evolution, including coagulation. I aim to integrate these into chemical transport modeling and determine where they can be trusted as emissions and meteorology change. I plan to test transfer by withholding entire campaigns, seasons, or atmospheric regimes, and to distinguish agreement with a reference simulation from agreement with independent measurements.
Those tests will guide the use of physical constraints, predictor selection, and explicit uncertainty estimates. Evaluation will consider CCN error and bias, the reliability of predicted uncertainty, physical consistency, and computational cost against a stated baseline. A useful outcome is a model with clearly established limits as well as demonstrated skill.
This direction connects my experience in aerosol microphysics and observational evaluation to a focused aim: making efficient atmospheric models easier to assess and use responsibly.
Selected papers
- CCN from commonly available measurements — Atmospheric Chemistry and Physics, 2020.
- Aerosol-size information from chemistry and meteorology — Geophysical Research Letters, 2021.
- Machine learning in climate-model aerosol predictions — Geophysical Research Letters, 2022.