Extension of the S5P-TROPOMI CCD tropospheric ozone retrieval to mid-latitudes

An advanced CCD (Convective Cloud Differential) algorithm has been developed to retrieve tropospheric ozone columns (TCO) from TROPOMI (Sentinel-5 Precursor–TROPOspheric Monitoring Instrument). The new approach utilizes a local cloud reference sector for TCO retrievals. The standard CCD algorithm relies on cloud data from the Pacific region to determine the stratospheric (above-cloud) column ozone (ACCO), which is later subtracted from the total column in clear-sky scenes in the entire tropics to derive the TCO.

However, the local cloud approach represents a significant advancement by dynamically selecting cloud reference sectors around the retrieval latitudes and longitudes. By extending TCO retrievals to the extra-tropics, the local cloud approach overcomes the constraints of the standard method, which relies on the basic assumption of zonal invariance of stratospheric ozone in the tropics. Additionally, it improves the retrieval accuracy for regions with more pronounced stratospheric ozone variability, such as the extra-tropics.

The advanced CHORA (Cloud Height Ozone Reference Algorithm) local cloud method developed includes three versions: (1) CLC (CHORA Local Cloud) and (2) CLCT (CHORA Local Cloud Theil-Sen) for tropical and subtropical regions and (3) CLCD (CHORA Local Cloud Decision) for mid-latitudes. The development of CLC and CLCT was the first step in extending the local cloud approach to mid-latitudes. Validating the local cloud approach in the tropics against the standard method helps to establish its reliability before its application in mid-latitude regions.

The CLC method uses a longitudinally varying and latitudinally restricted local cloud reference sector for ACCO retrievals, which is well suited for the tropics by effectively upholding the assumption of zonal invariance in stratospheric ozone. The CLCT builds upon the CLC approach and introduces a homogeneity criterion to address variability in stratospheric ozone. This method directly estimates ACCO above the reference altitude of 270 hPa using Theil-Sen regression. Additionally, CLCT can combine another tropical TCO retrieval method, Cloud-Slicing Algorithm (CSA) with CCD to retrieve upper tropospheric ozone. The CSA derives tropospheric ozone volume mixing ratios from the slope of the regression line between collocated ozone columns above clouds and cloud top pressures at varying heights. Meanwhile, the CPC (CHORA Pacific Cloud) algorithm is an improved version of the standard CCD method developed at the University of Bremen, employing the Pacific region as the cloud reference sector to derive the ACCO.

Monthly averaged TCOs for the tropics and subtropics (26◦S–22◦N) were derived from TROPOMI data (2018–2022) using CPC and local cloud algorithms (CLC and CLCT). These results were validated against spatially collocated measurements from NASA/GSFC SHADOZ (Southern Hemisphere Additional Ozonesondes) and the ESA TROPOMI Level-2 tropospheric ozone product. At eight of the nine stations, the CLCT algorithm was in better agreement with ozonesondes than the CPC. In the tropical region (20◦S-20◦N), CLCT achieved a significantly lower overall mean bias and standard deviation of 1±7%, outperforming the CPC (12±10%) and the CCD-based ESA product (22±10%).

We present a new CHORA algorithm, CLCD, for retrieving near-global TCO from TROPOMI. The CLCD algorithm is specifically designed to minimize the influence of stratospheric ozone variability, which is generally more pronounced in the extra-tropics. The method uses a circular local cloud reference sector and is better suited for extra-tropics to estimate the ACCO. In the extra-tropics, variable wind directions and limited ground coverage by geostationary satellites make a circular reference sector more suitable than a longitudinally varying cloud sector (CLC and CLCT), providing better sampling.

In mid-latitudes, where low-level clouds are common, TCOs are calculated from the surface up to 450 hPa (∼middle troposphere). The CLCD method employs two approaches: (1) CLCD-C, which utilizes an ozone climatology, and (2) CLCD-T, which estimates ACCO above 450 hPa using Theil-Sen regression when cloud top heights within the local cloud sector show sufficient variability. The CLCD algorithm dynamically selects between these methods based on cloud properties and applies a homogeneity criterion for total ozone to minimize stratospheric ozone inhomogeneities.

Monthly averaged CLCD-TCOs were derived for the tropics and mid-latitudes (60◦S–60◦N) using TROPOMI data (2018–2022) and validated against spatially collocated ozonesonde measurements from 36 stations within SHADOZ/WOUDC (World Ozone and Ultraviolet Radiation Data Centre)/NDACC (Network for the Detection of Atmospheric Composition Change) HEGIFTOM (Harmonization and Evaluation of Ground-based Instruments for Free Tropo-
spheric Ozone Measurements) networks. Validation indicates robust agreement between CLCD-TCO retrievals and ozonesondes at most stations, with an overall bias of 0.6 DU and a dispersion of 2.5 DU. The maximum observed bias and dispersion across all stations are ∼5 DU and 4 DU, respectively.

The CLCT algorithm is recommended over the CLCD algorithm in the tropics because of its superior performance in retrieving TCO at 450 hPa. Validation against ozonesonde measurements reveals that CLCT shows a lower overall bias and dispersion of 0.2 ± 1.4 DU, outperforming CLCD (0.8 ± 2.8 DU). However, in the subtropics, CLCD outperforms CLCT by reducing the overall bias from -2 DU to -1 DU. In regions like Northeast China and North America, where various emission sources contribute to high tropospheric ozone levels, CLCD effectively detects these enhancements.

This study marks the first successful near-global application of CCD retrievals. Our findings underscore the benefits of using the local cloud reference sector for TCO retrievals in both tropics (CLCT) and extra-tropics (CLCD). The local cloud approach serves as a significant foundation for systematic applications in both current and future satellite missions. The CLCD algorithm is well-suited for geostationary satellites such as GEMS (Geostationary Environment Monitoring Spectrometer, Korea), ESA Sentinel-4, and NASA TEMPO (Tropospheric Emissions: Monitoring of Pollution), which are designed to mostly monitor higher latitudes outside the tropics.