Item

Abstract

Tropospheric ozone (O3) profiles, especially within the boundary layer, are essential for studying the vertical, temporal, spatial variations, as well as the formation sensitivity and regional transport of O3. However, it is rare to find continuous tropospheric O3 profiles with high temporal and spatial resolutions without blind areas using current remote sensing technologies, with issues such as low near-surface sensitivity or systematic blind areas from satellite and LiDAR observations, respectively, being encountered. In this study, multi-source data including stratospheric O3 profiles from external datasets and local monthly dependent a priori profiles were fused in the retrieval algorithm, then vertical O3 profiles from the near-surface to the free troposphere were retrieved from multi-axis differential optical absorption spectroscopy (MAX-DOAS) measurements. With the aim of deriving a set of best practice recommendations for applying a profile inversion algorithm to long-term observations, we comprehensively investigated the influence of various settings on profile retrieval, with emphasis on the selection of a Fraunhofer reference spectrum and appropriate a priori profiles in the upper troposphere. These recommendations are essential for applying the algorithm to real long-term measurements. In this study, tropospheric O3 profiles were retrieved for operational MAX-DOAS observations in Beijing, and the results, especially for the boundary layer, were evaluated in detail with respect to well-established independent O3 datasets, including one-year ozonesonde profiles and tower-based in-situ measurements at different altitudes. A good level of agreement was found for both near-surface and elevated-altitude results, and the MAX-DOAS O3 profiles were able to reproduce the vertical distributions measured by ozonesonde. However, MAX-DOAS measurements are less sensitive to the upper troposphere and converge to the a priori profile, and the smoothing effect induced by the optimal estimation method makes it difficult to reproduce steep vertical gradients in real atmosphere.