Machine Learning Uncovers Aerosol Size Information From Chemistry and Meteorology to Quantify Potential Cloud‐Forming Particles. Issue 21 (9th November 2021)

• Record Type: Journal Article
• Authors: Nair, Arshad Arjunan; Yu, Fangqun; Campuzano‐Jost, Pedro; DeMott, Paul J.; Levin, Ezra J. T.; Jimenez, Jose L.; Peischl, Jeff; Pollack, Ilana B.; Fredrickson, Carley D.; Beyersdorf, Andreas J.; Nault, Benjamin A.; Park, Minsu; Yum, Seong Soo; Palm, Brett B.; Xu, Lu; Bourgeois, Ilann; Anderson, Bruce E.; Nenes, Athanasios; Ziemba, Luke D.; Moore, Richard H.; Lee, Taehyoung; Park, Taehyun; Thompson, Chelsea R.; Flocke, Frank; Huey, Lewis Gregory; Kim, Michelle J.; Peng, Qiaoyun
• Is Part Of: Geophysical research letters. Volume 48:Issue 21(2021)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-11-09
• Subjects: Cloud condensation nuclei (CCN) -- particle size distribution (PNSD) -- aerosols -- aircraft campaign observations -- machine learning -- explainable artificial intelligence (xAI); Geophysics -- Periodicals; Planets -- Periodicals; Lunar geology -- Periodicals; 550
• DOI: 10.1029/2021GL094133
• Languages: English
• ISSNs: 0094-8276

Abstract

Abstract: Cloud condensation nuclei (CCN) are mediators of aerosol‐cloud interactions, which contribute to the largest uncertainty in climate change prediction. Here, we present a machine learning (ML)/artificial intelligence (AI) model that quantifies CCN from model‐simulated aerosol composition, atmospheric trace gas, and meteorological variables. Comprehensive multi‐campaign airborne measurements, covering varied physicochemical regimes in the troposphere, confirm the validity of and help probe the inner workings of this ML model: revealing for the first time that different ranges of atmospheric aerosol composition and mass correspond to distinct aerosol number size distributions. ML extracts this information, important for accurate quantification of CCN, additionally from both chemistry and meteorology. This can provide a physicochemically explainable, computationally efficient, robust ML pathway in global climate models that only resolve aerosol composition; potentially mitigating the uncertainty of effective radiative forcing due to aerosol‐cloud interactions (ERFaci ) and improving confidence in assessment of anthropogenic contributions and climate change projections. Plain Language Summary: The largest uncertainties in climate change modeling are linked with cloud condensation nuclei (CCN). These tiny atmospheric particles modulate cloud formation and thus affect the Earth's energy budget. A machine learning/artificial intelligence model that accurately quantifies CCN can potentially reduce these uncertainties. Comprehensive multi‐campaign aircraft measurements over varied atmospheric environments validate this model. Importantly, the inner workings of this model are teased out to reveal that its decisions are rooted in physical and chemical principles. Key Points: Machine learning (ML) derived cloud condensation nuclei numbers in strong agreement with comprehensive multi‐campaign aircraft observations First demonstration that aerosol size information is contained in aerosol mass speciation, chemistry and meteorology, and is extractable by ML A physicochemically explainable (xAI) and robust ML avenue to mitigate aerosol‐cloud interaction uncertainties in climate models is realized