Machine learning uncovers aerosol size information 1 from chemistry and meteorology to quantify 2 potential cloud-forming particles 3

5 Cloud condensation nuclei (CCN) are mediators of aerosol–cloud interactions, which 6 contribute to the largest uncertainty in climate change prediction. Here, we present 7 a machine learning/artificial intelligence model that quantifies CCN from variables of 8 aerosol composition, atmospheric trace gases, and meteorology. Comprehensive multi9 campaign airborne measurements, covering varied physicochemical regimes in the tro10 posphere, confirm the validity of and help probe the inner workings of this machine 11 learning model: revealing for the first time that different ranges of atmospheric aerosol 12 composition and mass correspond to distinct aerosol number size distributions. Ma13 chine learning extracts this information, important for accurate quantification of CCN, 14 additionally from both chemistry and meteorology. This can provide a physicochemi15 cally explainable, computationally efficient, robust machine learning pathway in global 16 climate models that only resolve aerosol composition; potentially mitigating the un17 certainty of effective radiative forcing due to aerosol–cloud interactions (ERFaci) and 18 improving confidence in assessment of anthropogenic contributions and climate change 19 projections. 20

[1]  Andrew R. Whitehill,et al.  Investigation of factors controlling PM2.5 variability across the South Korean Peninsula during KORUS-AQ , 2020, Elementa.

[2]  J. Dramsch,et al.  70 years of machine learning in geoscience in review , 2020, Advances in Geophysics.

[3]  F. Yu,et al.  Using machine learning to derive cloud condensation nuclei number concentrations from commonly available measurements , 2020, Atmospheric Chemistry and Physics.

[4]  E. Ray,et al.  A large source of cloud condensation nuclei from new particle formation in the tropics , 2019, Nature.

[5]  J. Crounse,et al.  ATom: L2 In Situ Data from Caltech Chemical Ionization Mass Spectrometer (CIT-CIMS) , 2019 .

[6]  P. Pilewskie,et al.  Neural network for aerosol retrieval from hyperspectral imagery , 2019, Atmospheric Measurement Techniques.

[7]  A. Segers,et al.  Machine learning for observation bias correction with application to dust storm data assimilation , 2019, Atmospheric Chemistry and Physics.

[8]  D. Cziczo,et al.  A machine learning approach to aerosol classification for single-particle mass spectrometry , 2018, Atmospheric Measurement Techniques.

[9]  M. Facchini,et al.  Identification of new particle formation events with deep learning , 2018, Atmospheric Chemistry and Physics.

[10]  Yongguo Yang,et al.  Comprehensive Evaluation of Machine Learning Techniques for Estimating the Responses of Carbon Fluxes to Climatic Forces in Different Terrestrial Ecosystems , 2018 .

[11]  Matthew West,et al.  Machine Learning to Predict the Global Distribution of Aerosol Mixing State Metrics , 2018 .

[12]  M. A. Zaidan,et al.  Predicting atmospheric particle formation days by Bayesian classification of the time series features , 2018 .

[13]  Rintaro Okamura,et al.  Feasibility study of multi-pixel retrieval of optical thickness and droplet effective radius of inhomogeneous clouds using deep learning , 2017 .

[14]  K. Jucks,et al.  Planning, implementation, and scientific goals of the Studies of Emissions and Atmospheric Composition, Clouds and Climate Coupling by Regional Surveys (SEAC4RS) field mission , 2016 .

[15]  J. Peischl,et al.  The Deep Convective Clouds and Chemistry (DC3) Field Campaign , 2015 .

[16]  A. Sorooshian,et al.  On the competition among aerosol number, size and composition in predicting CCN variability: a multi-annual field study in an urbanized desert. , 2015, Atmospheric chemistry and physics.

[17]  Douglas M. Bates,et al.  Fast and Elegant Numerical Linear Algebra Using the RcppEigen Package , 2013 .

[18]  Glenn E. Shaw,et al.  The Arctic Research of the Composition of the Troposphere from Aircraft and Satellites (ARCTAS) mission: design, execution, and first results , 2010 .

[19]  A. Nenes,et al.  Scanning Flow CCN Analysis—A Method for Fast Measurements of CCN Spectra , 2009 .

[20]  J. Hudson.  Variability of the relationship between particle size and cloud‐nucleating ability , 2007 .

[21]  M. Andreae,et al.  Size Matters More Than Chemistry for Cloud-Nucleating Ability of Aerosol Particles , 2006, Science.

[22]  S. Menon,et al.  The radiative influence of aerosol effects on liquid-phase cumulus and stratiform clouds based on sensitivity studies with two climate models , 2006 .

[23]  A. Nenes,et al.  A Continuous-Flow Streamwise Thermal-Gradient CCN Chamber for Atmospheric Measurements , 2005 .

[24]  George Tselioudis,et al.  GCM Simulations of the Aerosol Indirect Effect: Sensitivity to Cloud Parameterization and Aerosol Burden , 2002 .

[25]  V. Ramanathan,et al.  Aerosols, Climate, and the Hydrological Cycle , 2001, Science.

[26]  Ronald,et al.  Drizzle Suppression in Ship Tracks. , 2000 .

[27]  V. Ramanathan,et al.  Reduction of tropical cloudiness by soot , 2000, Science.

[28]  Rosenfeld,et al.  Suppression of rain and snow by urban and industrial air pollution , 2000, Science.

[29]  J. Hansen,et al.  Radiative forcing and climate response , 1997 .

[30]  Olivier Boucher,et al.  The sulfate‐CCN‐cloud albedo effect , 1995 .

[31]  R. Pincus,et al.  Effect of precipitation on the albedo susceptibility of clouds in the marine boundary layer , 1994, Nature.

[32]  D. W. Johnson,et al.  The Measurement and Parameterization of Effective Radius of Droplets in Warm Stratocumulus Clouds , 1994 .

[33]  B. Albrecht.  Aerosols, Cloud Microphysics, and Fractional Cloudiness , 1989, Science.

[34]  T. L. Wolfe,et al.  An assessment of the impact of pollution on global cloud albedo , 1984 .

[35]  S. P. Lloyd,et al.  Least squares quantization in PCM , 1982, IEEE Trans. Inf. Theory.

[36]  J. A. Hartigan,et al.  A k-means clustering algorithm , 1979 .

[37]  S. Twomey.  The Influence of Pollution on the Shortwave Albedo of Clouds , 1977 .

[38]  S. Twomey.  Pollution and the Planetary Albedo , 1974 .

[39]  J. W. Fitzgerald.  Dependence of the Supersaturation Spectrum of CCN on Aerosol Size Distribution and Composition , 1973 .

[40]  C. Junge,et al.  Relationship of Cloud Nuclei Spectra to Aerosol Size Distribution and Composition , 1971 .

[41]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[42]  Masson-Delmotte,et al.  The Physical Science Basis , 2007 .