Interrogation of Heterogeneous Atmospheric Structure Through Statistical Signal Processing

[09-Jan-2023] Abstract  Photon counting is commonly employed as an optical detection technique when the instantaneous optical power is very low. This technique is highly sensitive to low light levels. For lidar, photon counting enables the capture of backscatter signals while operating low pulse energy lasers, allowing for the quantitative capture of atmospheric properties with reliable, low-cost, and eye-safe transmitters. It is also an essential technique for capturing targets at large distances and has and will likely continue to be employed in future space-based lidar missions such as AOS/ACCP. While photon-counting enables the capture of very low optical signals, it encounters limitations in regions of high backscatter, where non-ideal detector behavior becomes significant. In single photon counting modules (SPCMs), detector deadtime and afterpulsing can contribute to errors in the captured signal. Due to these effects, averaging over non-uniform scenes can bias the captured signals, which can, in turn, bias derived products. The effects of these errors are poorly understood, unquantified and unaccounted for in lidar error analysis. This issue is of particular concern for moving platforms (e.g. airborne and spaceborne sensors), where the distance between laser shots can result in dramatically different sample volumes. With the eventual goal of developing improved detection and data acquisition solutions that account for non-ideal detector behavior and subgrid-scale atmospheric variability, we have developed an approach to acquire and process very high-resolution photon counting data from the MicroPulse DIAL (MPD). MPD data is captured from SPCMs at very high resolution (0.04 s x 0.75 m) using an NCAR developed time tag acquisition system. The data is processed using Poisson Total Variation, a regularized maximum likelihood estimator, while employing a detector noise model that accounts for deadtime. We will present atmospheric observations enabled by this effort and discuss implications for future lidar development.