Patrick Selmer, programmer/analyst at Science Systems and Applications, Inc, was the featured speaker of the Climate / Atmospheric Science & Engineering (CASE) Colloquium series on Dec. 5, 2022. The presentation was titled "Backscatter Lidar Retrieval Improvement with Machine Learning."

The CASE Colloquium is presented by the University of Iowa's Center for Global and Regional Environmental Research and the Iowa Technology Institute. Find past CASE presentations here.

Abstract:

Lidar remote sensing of the atmosphere provides vertically resolved profiles of clouds and aerosols. Heights and optical properties can be retrieved, contributing to climate and modeling research. Advances in Machine Learning (ML) since the early 2000s, particularly as related to image processing, can improve retrievals and overcome limitations of traditional lidar data processing techniques. Certain Convolutional Neural Networks (CNNs) can improve feature detection, classification, and perform denoising. Applications of other ML methods such as clustering to feature classification are discussed. Results show improvement in feature detection, typing, retrievals, and processing speed.