Retrieval of Aerosol Properties Using Invertible Neural Networks

Retrieval of aerosol properties from in situ, multi-angle light scattering measurements using invertible neural networks with Romana Boiger and Rob Modini from the Paul Scherrer Institute

Abstract:
Atmospheric aerosols have a major influence on the earth’s climate and public health. Hence, studying their properties and recovering them from light scattering measurements is of great importance. State of the art retrieval methods such as pre-computed look-up tables and iterative, physics-based algorithms can suffer from either accuracy or speed limitations. These limitations are becoming increasingly restrictive as instrumentation technology advances and measurement complexity increases. Machine learning algorithms offer new opportunities to overcome these problems, by being quick and precise. In this work we present a method, using invertible neural networks to retrieve aerosol properties from in situ light scattering measurements. In addition, the algorithm is capable of simulating the forward direction, from aerosol properties to measurement data. The applicability and performance of the algorithm are demonstrated with simulated measurement data, mimicking in situ laboratory and field measurements. With a retrieval time in the millisecond range and a weighted mean absolute percentage error of less than 1.5%, the algorithm turned out to be fast and accurate. By introducing Gaussian noise to the data, we further demonstrate that the method is robust with respect to measurement errors. In addition, realistic case studies are performed to demonstrate that the algorithm performs well even with missing measurement data.

Romana Boiger is a tenure-track scientist at Paul Scherrer Institute (PSI) in Villigen, Switzerland. She studied mathematics in Austria and obtained a PhD in the field of inverse problems from the University of Klagenfurt, Austria, in 2016. Following her doctoral studies, she gained experience working on several industrial research projects as mathematician and data scientist. In 2020, Romana joined PSI as part of the Laboratory for Simulation and Modeling, where she conducted research related to machine learning for particle accelerators and aerosol physics. This year, in February, she obtained a tenure-track position at the Laboratory of Waste Management at PSI. Her current research interests lie in the fields of machine learning, surrogate modeling, process optimization, and computer vision, with a specific focus on their applications in waste management.

Rob Modini is a senior scientist in the Aerosol Physics Group in the Laboratory of Atmospheric Chemistry at the Paul Scherrer Institute in Villigen, Switzerland. He obtained his PhD in Australia (Queensland University of Technology) and completed his postdoctoral studies in San Diego, USA (Scripps Institution of Oceanography) and Lausanne, Switzerland (Ecole Polytechnique Federale de Lausanne). Rob’s research focuses on many different aspects of atmospheric aerosols including their physical and optical properties, their interactions with clouds, and their health effects. He specializes in aerosol measurements and instrumentation, but also works actively at the interface between aerosol measurements and models.