Ozone I

As shown by recent studies, global atmospheric Ozone (O₃) has begun to recover resulting from the decline in the atmospheric abundance of anthropogenic emissions of halogenated Ozone Depleting Substances (ODS), which is achieved by signing in 1987 Montreal Protocol (and its amendments). Nevertheless, there are other factors that also affect the temporal changes of O₃, such as non-halogen chemical species, solar impact, natural forcings in changing climate, volcanic activities implying significant aerosol loading etc. Therefore, stratospheric O₃ is expected to vary within different timescales as a result of dynamical and chemical forcings.

Contrary to the overall recovery of the global O₃, an unexpected O₃ decline in the tropical mid-stratosphere was observed from the SCanning Imaging Absorption SpectroMeter for Atmospheric CHartographY (SCIAMACHY) measurements during 2004-2012. Similar negative changes of O₃ were also observed from other satellite measurements: HALOE (1991-2005), combined SAGE II-GOMOS (1997-2011), MIPAS (2002-2012), MLS (2004-2013). Our analysis of SCIAMACHY measurements shows that the decrease in O₃ is accompanied by an increase in NO₂.

To reveal the causes of O₃ and NO₂ changes, we performed simulations with the TOMCAT 3D Chemistry Transport Model (CTM) with different chemical and dynamical forcings. TOMCAT reproduces the SCIAMACHY-observed O₃ and NO₂ changes in the tropical mid-stratosphere. The model simulations show that the positive changes in NO₂ (around 7% per decade) are due to similar positive changes in reactive odd nitrogen (NO_y), which are a result of a longer residence time of the source gas N₂O. As the global atmospheric lifetime of N₂O exceeds 100 years, its negative changes of 10% per decade can be described by variations in dynamics in the deep branch of Brewer-Dobson Circulation. However, modelled annual mean age-of-air (AoA) did not show any significant changes in the transport in the tropical mid-stratosphere during 2004-2012. Further analysis of model results demonstrated significant seasonal variations: positive AoA changes, which indicate transport slowdown during autumn months (September-October), cause longer residence time of N₂O with more intense NO_y production which effectively depletes O₃; and negative AoA changes during winter months (January, February), which indicate more intense N₂O transport, with shorter residence time of N₂O and less intense NO_y production via N₂O + O(¹D) to further destroy O₃. Although the changes in AoA cancel out when averaging over the year, the N₂O change remains due to non-linearities in the chemistry-transport interactions.

A continuous monitoring of the stratospheric ozone layer over a global scale is done by means of several techniques. Observations in limb geometry from satellite platforms provide ozone profiles with a good vertical resolution, spatial and temporal coverage.

SCIAMACHY and OMPS-LP are two satellite instruments able to collect shortwave scattered radiance in limb geometry. Ozone profile data sets from SCIAMACHY (2002-2012) and OMPS-LP (2012-2018) were created at the University of Bremen using the same radiative transfer model, spectroscopic databases and a similar retrieval algorithm. The aim of this study is the merging of these data sets, to obtain a consistent time series of ozone global distributions. Since the two missions overlap only for 3 months, a transfer function is needed to overcome issues related to the sensors calibration. To this aim, we chose measurements performed by the MLS instrument as a reference: this sensor has been collecting atmospheric emission in the microwave spectral region in limb geometry since 2004.

Monthly latitude- and longitude-resolved time series of ozone profiles were calculated for the two instruments, exploiting the high spatial resolution of the data sets. Their merging has been then performed minimizing the differences between OMPS-LP and SCIAMACHY ozone number density profiles with respect to MLS values, for each latitude, longitude and altitude independently. The seasonal cycle was not subtracted, because it was found to be consistent enough among the three instruments. Short-term changes in ozone profiles were calculated over 2003-2018 using a multilinear regression (MLR) analysis, including fit proxies as QBO, ENSO and a solar forcing. Positive trends were detected between 35 and 45 km at mid-latitudes, with an increasing ozone concentration up to 2-3% per decade. Negative changes were found in the lower tropical stratosphere but statistically non-significant. A comparison with short-term trends calculated over the SCIAMACHY time period has been done: while a general agreement was found, some discrepancies were seen in the tropical mid-stratosphere.

A merging with SAGE II ozone profiles was also performed: zonal monthly anomalies from the 3 instruments are merged to study ozone trends over the last 35 years. Applying the same MLR analysis, consistent results with previous studies were found: negative trends before 1997 up to -6% per decade at mid-latitudes around 40 km and the expected recovery after to ozone turn-around point at the end of ‘90, related also to the implementation of the Montreal protocol and its amendments.

We have retrieved nadir ozone profiles using SCIAMACHY Level-1 version 9 data with the Ozone ProfilE Retrieval Algorithm (OPERA) developed at KNMI. The v9 data is supposed to be the last version of the SCIAMACHY L1 data. Compared to the nadir ozone profile retrievals performed using v8 and v7 data (Shah et al., 2018), we focus on the degradation correction of the SCIAMACHY measurements. We first applied all calibration options provided in the sciaL1c tool to the L1 v9 spectra. All the analyses are based on the SCIAMACHY L1c child files.

The radiometric degradation of SCIAMACHY Channel 1 and 2 data has been analysed using one day per month of global measurements for the whole mission period (2002-2012). Assuming that the global mean reflectance is relatively stable in the 10-years period, we compared the SCIAMACHY measured reflectances with simulated reflectances using OPERA. It turned out that the degradation of the SCIAMACHY reflectances of wavelengths around 265 nm in Channel 1 amounts to about 30 % in 2012. The degradation of channel 2 is much smaller.

The wavelength calibration and Instrument Spectral Response Function (ISRF) has been analysed for the same SCIAMACHY L1c files using the QDOAS algorithm and our own algorithm. The changes in ISRF and wavelength calibration are very small. These corrections have been taken into accounted in OPERA.

We have processed a selected number of ozone profiles with OPERA taking all these improvements into account. The ozone profiles have been validated with ozonesonde measurements launched at De Bilt. We will show results on the ozone profiles with the latest improvements in the SCIAMACHY data.

The Sentinel-5 Precursor (S5P) mission, launched in October 2017, started the operational atmospheric composition measurements from space as part of the European Copernicus programme. The payload of the S5P mission is the TROPOspheric Monitoring Instrument (TROPOMI) that provides key information on air quality, climate and the ozone layer with high spatial resolution and daily global coverage.

In this presentation we provide an overview of the operational TROPOMI geophysical products including O₃, NO₂, SO₂, HCHO, CO, CH₄, as well as UV, cloud and aerosol properties.

The European teams responsible for the operational products are organized in three groups covering: (i) retrieval algorithms, (ii) data processors being used in the S5P ground-segment for the generation of the operational TROPOMI products, and (iii) routine validation of S5P products using fiducial reference measurements. It is planned to maintain this project organization during the complete mission in order to ensure the timely provision of state-of-science data products that are continuously improved and validated.

Initial versions of the TROPOMI products were already available a few weeks after launch and were presented at the first light event that took place in December 2017. The retrieval algorithms and data processors were optimized during the commissioning phase that lasted until April 2018 and the results of the preliminary validation were presented in June 2018. Finally the first set of operational S5P products was released to the public in July 2018. The release of the remaining S5P products is organized in a staggered approach and will take place during the second part of 2018 and early 2019.

The work on TROPOMI/S5P geophysical products is funded by EU Copernicus, ESA and national contributions from The Netherlands, Germany, Belgium and Finland.

In October 2017, the Sentinel-5P satellite with the TROPOMI instrument was launched into space. Total ozone columns from TROPOMI instrument are retrieved in near real time using the well established DOAS approach, and a subsequent conversion into vertical column densities using an iterative air mass factor calculation. The total ozone algorithm has already been successfully applied to the GOME, SCIAMACHY and GOME-2 missions. To account for the high spatial resolution of the TROPOMI instrument, the treatment of clouds in the total ozone retrieval has been improved.

Besides the total column, a tropospheric ozone column is also provided from TROPOMI. The cloud convective differential method is used to separate the tropospheric and stratospheric ozone column. Deep convective clouds shield the tropospheric ozone from the satellite based observers. Therefore the ozone columns above deep convective clouds are good approximations of the stratospheric ozone column. These data are corrected for the variety in the cloud top altitudes and averaged over a 5 days time period and a clean area. In a last step the stratospheric column are subtracted from the total ozone columns for cloud free observations.

After a short introduction to the retrieval algorithms, first results will be presented for both total and tropospheric ozone columns from TROPOMI. First comparisons to ozone column data from GOME-2 and ground-based observations will be shown as well.

The TROPOspheric Monitoring Instrument (TROPOMI), on board the Sentinel 5 precursor (S5p) satellite, was launched in October 2017. A first set of products, for instance, ozone and cloud properties, became publicly available in July 2018.

Using this operational S5p ozone and cloud dataset, we derive tropical upper tropospheric ozone volume mixing ratios (TTCO). For this, we employ the cloud slicing method [Ziemke, 2001] and compare our results of the cloud slicing algorithm (CSA) with data from the operational CSA version.   

The TROPOMI instrument has a high spatial resolution and a daily coverage of the Earth. We will discuss the choice of the best temporal and spatial resolution that is optimally suited for the intrinsic statistical usage of ozone and cloud measurements for the CSA method. This strongly depends on the instrument characteristics.

The work on TROPOMI/S5P geophysical products is funded by ESA and national contributions from the Netherlands, Germany, Belgium, and Finland.

Drift uncertainties in long-term data records like for many essential climate variables have a large impact on the sensitivity of the data record to detect small trends. A good example is stratospheric ozone which is expected to slowly recovery due to the successful phase-out of ozone-depleting substances (ODS) as regulated by the Montreal Protocol. So far positive ozone trends since 2000 were only statistically significant in the upper stratosphere according to the recent WMO ozone assessment. In addition to the high variability in ozone potential instrumental drifts in satellite data add to trend uncertainty. Stability requirements from 1%/decade (total ozone) to 5%/decade (ozone profiles) have been specified, however, the rationale behind these numbers is not clear. Satellite measurements of ozone are available for nearly four decades. The single lifetime of a typical satellite instrument is on the order of 5-7 years, the longest single instrument record was from SAGE II operating for nearly twenty years. Uncertainties from combining or merging multiple datasets to obtain long-term time series also add to trend uncertainties. The connection between stability requirements and trend detection limits from multiple satellites are quantified by using simple Monte Carlo simulations.