A vertically-integrated approach to climate science : from measurements and machine learning to models and policy

Abstract

The role anthropogenic aerosol particles play in the formation and persistence of ice clouds remains one of the most uncertain aspects of understanding past, present, and future climate. Studying how these particles influence ice cloud formation requires careful measurement of their ice nucleating ability as well as robust uncertainty quantification of experimental results. These measurements and their corresponding uncertainties form the basis for parameterizations used in climate models to probe how anthropogenic particle emissions affect climate through ice cloud formation. This type of investigation can help to inform policy decisions about controls on anthropogenic particle emissions. This study aims to clarify the human role in the climate system by 1) developing instrumentation to perform ice nucleation measurements, 2) quantifying the uncertainty associated with these measurements using machine learning algorithms, 3) incorporating measurements and uncertainty quantification in climate model simulations, and 4) using the modeled climate response to help inform policy decisions for anthropogenic particle emissions.