Abstract
Abstract. This paper evaluates and discusses the quality of the stratospheric ozone analyses delivered in near real time by the MACC (Monitoring Atmospheric Composition and Climate) project during the 3-year period between September 2009 and September 2012. Ozone analyses produced by four different chemical data assimilation (CDA) systems are examined and compared: the Integrated Forecast System coupled to the Model for OZone And Related chemical Tracers (IFS-MOZART); the Belgian Assimilation System for Chemical ObsErvations (BASCOE); the Synoptic Analysis of Chemical Constituents by Advanced Data Assimilation (SACADA); and the Data Assimilation Model based on Transport Model version 3 (TM3DAM). The assimilated satellite ozone retrievals differed for each system; SACADA and TM3DAM assimilated only total ozone observations, BASCOE assimilated profiles for ozone and some related species, while IFS-MOZART assimilated both types of ozone observations. All analyses deliver total column values that agree well with ground-based observations (biases < 5%) and have a realistic seasonal cycle, except for BASCOE analyses, which underestimate total ozone in the tropics all year long by 7 to 10%, and SACADA analyses, which overestimate total ozone in polar night regions by up to 30%. The validation of the vertical distribution is based on independent observations from ozonesondes and the ACE-FTS (Atmospheric Chemistry Experiment – Fourier Transform Spectrometer) satellite instrument. It cannot be performed with TM3DAM, which is designed only to deliver analyses of total ozone columns. Vertically alternating positive and negative biases are found in the IFS-MOZART analyses as well as an overestimation of 30 to 60% in the polar lower stratosphere during polar ozone depletion events. SACADA underestimates lower stratospheric ozone by up to 50% during these events above the South Pole and overestimates it by approximately the same amount in the tropics. The three-dimensional (3-D) analyses delivered by BASCOE are found to have the best quality among the three systems resolving the vertical dimension, with biases not exceeding 10% all year long, at all stratospheric levels and in all latitude bands, except in the tropical lowermost stratosphere. The northern spring 2011 period is studied in more detail to evaluate the ability of the analyses to represent the exceptional ozone depletion event, which happened above the Arctic in March 2011. Offline sensitivity tests are performed during this month and indicate that the differences between the forward models or the assimilation algorithms are much less important than the characteristics of the assimilated data sets. They also show that IFS-MOZART is able to deliver realistic analyses of ozone both in the troposphere and in the stratosphere, but this requires the assimilation of observations from nadir-looking instruments as well as the assimilation of profiles, which are well resolved vertically and extend into the lowermost stratosphere.
Highlights
The presence of a high-altitude ozone layer in the atmosphere, which protects the Earth system against the harmful ultraviolet (UV) light from the Sun, was first determined in the 1920s from observations of the solar UV spectrum
Ozone analyses produced by four different chemical data assimilation (CDA) systems are examined and compared: the Integrated Forecast System coupled to the Model for OZone And Related chemical Tracers (IFS-MOZART); the Belgian Assimilation System for Chemical ObsErvations (BASCOE); the Synoptic Analysis of Chemical Constituents by Advanced Data Assimilation (SACADA); and the Data Assimilation Model based on Transport Model version 3 (TM3DAM)
The assimilated satellite ozone retrievals differed for each system; SACADA and TM3DAM assimilated only total ozone observations, BASCOE assimilated profiles for ozone and some related species, while IFS-MOZART assimilated both types of ozone observations
Summary
The presence of a high-altitude ozone layer in the atmosphere, which protects the Earth system against the harmful ultraviolet (UV) light from the Sun, was first determined in the 1920s from observations of the solar UV spectrum. Two coupled systems were created in MACC: IFS-TM5 and IFS-MOZART (Flemming et al, 2009; Stein et al, 2013) These coupled dynamics-chemistry DAS are run at the European Centre for Medium-Range Weather Forecasts (ECMWF) in near-real-time (NRT) for monitoring present and near-future atmospheric conditions up to 5 days ahead, through analyses and forecasts of carbon monoxide (CO), formaldehyde (HCHO), nitrogen oxides (NOx, i.e. NO+NO2), sulfur dioxide (SO2) and ozone (O3). Our study is similar to the intercomparison of ozone analyses realised in the Assimilation of Envisat Data (ASSET) project (Geer et al, 2006; Lahoz et al, 2007), with some major differences: here the DAS were configured primarily to satisfy operational constraints and deliver NRT products (and in the case of IFS-MOZART to deliver several tropospheric products in addition to stratospheric ozone); we assimilated a large variety of data sets while ASSET used only observations from Envisat (Environmental Satellite); and the investigated period is much longer (3 years instead of 5 months). This section includes a description of the data sets used in the validation of the four analyses
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
