Optimization of the GSFC TROPOZ DIAL retrieval using synthetic lidar returns and ozonesondes – Part 1: Algorithm validation

• Published in: Atmospheric measurement techniques Vol. 8; no. 10; pp. 4133 - 4143
• Main Authors: Sullivan, J. T., McGee, T. J., Leblanc, T., Sumnicht, G. K., Twigg, L. W.
• Format: Journal Article
• Language: English
• Published: Katlenburg-Lindau Copernicus GmbH 09.10.2015; Copernicus Publications
• Subjects: Air quality; Algorithms; Analysis; Computer simulation; Constants; Detectors; Dials; Differential absorption lidar; Emission measurements; Error correction; Lidar; Noise; Optical radar; Optimization; Ozone; Ozonesondes; Photochemical ozone; Photochemicals; Photochemistry; Pollution monitoring; Remote sensing; Resolution; Retrieval; Saturation; Sensors; Space flight; Troposphere; Tropospheric ozone; Uncertainty; Uncertainty analysis; Vertical distribution
• ISSN: 1867-8548; 1867-1381; 1867-8548; 1867-1381
• DOI: 10.5194/amt-8-4133-2015

The main purpose of the NASA Goddard Space Flight Center TROPospheric OZone DIfferential Absorption Lidar (GSFC TROPOZ DIAL) is to measure the vertical distribution of tropospheric ozone for science investigations. Because of the important health and climate impacts of tropospheric ozone, it is imperative to quantify background photochemical ozone concentrations and ozone layers aloft, especially during air quality episodes. For these reasons, this paper addresses the necessary procedures to validate the TROPOZ retrieval algorithm and confirm that it is properly representing ozone concentrations. This paper is focused on ensuring the TROPOZ algorithm is properly quantifying ozone concentrations, and a following paper will focus on a systematic uncertainty analysis. This methodology begins by simulating synthetic lidar returns from actual TROPOZ lidar return signals in combination with a known ozone profile. From these synthetic signals, it is possible to explicitly determine retrieval algorithm biases from the known profile. This was then systematically performed to identify any areas that need refinement for a new operational version of the TROPOZ retrieval algorithm. One immediate outcome of this exercise was that a bin registration error in the correction for detector saturation within the original retrieval was discovered and was subsequently corrected for. Another noticeable outcome was that the vertical smoothing in the retrieval algorithm was upgraded from a constant vertical resolution to a variable vertical resolution to yield a statistical uncertainty of <10 %. This new and optimized vertical-resolution scheme retains the ability to resolve fluctuations in the known ozone profile, but it now allows near-field signals to be more appropriately smoothed. With these revisions to the previous TROPOZ retrieval, the optimized TROPOZ retrieval algorithm (TROPOZopt) has been effective in retrieving nearly 200 m lower to the surface. Also, as compared to the previous version of the retrieval, the TROPOZopt had an overall mean improvement of 3.5 %, and large improvements (upwards of 10–15 % as compared to the previous algorithm) were apparent between 4.5 and 9 km. Finally, to ensure the TROPOZopt retrieval algorithm is robust enough to handle actual lidar return signals, a comparison is shown between four nearby ozonesonde measurements. The ozonesondes are mostly within the TROPOZopt retrieval uncertainty bars, which implies that this exercise was quite successful.