Detecting Features in Spaceborne Backscatter Lidar Data: Spatial Averaging vs. Machine Learning Based Denoising

^(a) NASA Godard Space Flight Center, Greenbelt, MD, USA
^(b) Science, Systems, and Applications, Inc., Lanham, MD, USA
^(c) Department of Atmospheric and Oceanic Sciences, Univ. of Maryland, College Park, MD, USA
^(d) Earth System Science Interdisciplinary Center, Univ. of Maryland, College Park, MD, USA Lead Author e-mail address: This email address is being protected from spambots. You need JavaScript enabled to view it.

Published online: 9 April 2026

Abstract

Space-based lidar systems provide critical information about the vertical distributions of clouds and aerosols that greatly improve our understanding of the climate system. However, daytime spaceborne lidar signals are degraded by solar background. To overcome this issue, data is averaged during science processing at the expense of spatial resolution. New machine learning tools for denoising daytime spaceborne lidar data enable improvements in signal-to-noise ratio and data products at finer resolutions. Here we use airborne data and spaceborne simulations of backscatter lidar systems to quantify the performance of cloud detection frequencies and cloud top heights using spatial averaging and DDUNet autoencoder denoising.