FRESCO-B: a fast cloud retrieval algorithm using oxygen B-band measurements from GOME-2

Abstract. Nitrogen dioxide (NO2) is one of the main data products measured by the Tropospheric Monitoring Instrument (TROPOMI) on the Sentinel-5 Precursor (S5P) satellite, which combines a high signal-to-noise ratio with daily global coverage and high spatial resolution. TROPOMI provides a valuable source of information to monitor emissions from local sources such as power plants, industry, cities, traffic and ships, and variability of these sources in time. Validation exercises of NO2 v1.2–v1.3 data, however, have revealed that TROPOMI's tropospheric vertical column densities (VCDs) are too low by up to 50 % over highly polluted areas. These findings are mainly attributed to biases in the cloud pressure retrieval, the surface albedo climatology and the low resolution of the a priori profiles derived from global simulations of the TM5-MP chemistry model. This study describes improvements in the TROPOMI NO2 retrieval leading to version v2.2, operational since 1 July 2021. Compared to v1.x, the main changes are the following. (1) The NO2-v2.2 data are based on version-2 level-1b (ir)radiance spectra with improved calibration, which results in a small and fairly homogeneous increase in the NO2 slant columns of 3 % to 4 %, most of which ends up as a small increase in the stratospheric columns. (2) The cloud pressures are derived with a new version of the FRESCO cloud retrieval already introduced in NO2-v1.4, which led to a lowering of the cloud pressure, resulting in larger tropospheric NO2 columns over polluted scenes with a small but non-zero cloud coverage. (3) For cloud-free scenes a surface albedo correction is introduced based on the observed reflectance, which also leads to a general increase in the tropospheric NO2 columns over polluted scenes of order 15 %. (4) An outlier removal was implemented in the spectral fit, which increases the number of good-quality retrievals over the South Atlantic Anomaly region and over bright clouds where saturation may occur. (5) Snow/ice information is now obtained from ECMWF weather data, increasing the number of valid retrievals at high latitudes. On average the NO2-v2.2 data have tropospheric VCDs that are between 10 % and 40 % larger than the v1.x data, depending on the level of pollution and season; the largest impact is found at mid and high latitudes in wintertime. This has brought these tropospheric NO2 closer to Ozone Monitoring Instrument (OMI) observations. Ground-based validation shows on average an improvement of the negative bias of the stratospheric (from −6 % to −3 %), tropospheric (from −32 % to −23 %) and total (from −12 % to −5 %) columns. For individual measurement stations, however, the picture is more complex, in particular for the tropospheric and total columns.

Nitrogen dioxide (NO2) is one of the main data products measured by the Tropospheric Monitoring Instrument (TROPOMI) on the Sentinel-5 Precursor (S5P) satellite, which combines a high signal-to-noise ratio with daily global coverage and high spatial resolution. TROPOMI provides a valuable source of information to monitor emissions from local sources such as power plants, industry, cities, traffic and ships, and variability of these sources in time. Validation exercises of NO2 version v1.2-v1.3 data, however, have revealed that TROPOMI's tropospheric vertical columns (VCDs) are too low by up to 50 % over highly polluted areas. These findings are mainly attributed to biases in the cloud pressure retrieval, the surface albedo climatology and the low resolution of the a-priori profiles derived from global simulations of the TM5-MP chemistry model. This study describes improvements in the TROPOMI NO2 retrieval leading to version v2.2, operational since 1 July 2021. Compared to v1.x, the main changes are: (1) The NO2-v2.2 data is based on version 2 level-1B (ir)radiance spectra with improved calibration, which results in a small and fairly homogeneous increase of the NO2 slant columns of 3 to 4 %, most of which ends up as a small increase of the stratospheric columns; (2) The cloud pressures are derived with a new version of the FRESCO cloud retrieval already introduced in NO2-v1.4, which lead to a lowering of the cloud pressure, resulting in larger tropospheric NO2 columns over polluted scenes with a small but non-zero cloud coverage; (3) For cloud-free scenes a surface albedo correction is introduced based on the observed reflectance, which also leads to a general increase of the tropospheric NO2 columns over polluted scenes of order 15 %; (4) An outlier removal was implemented in the spectral fit, which increases the number of good quality retrievals over the South-Atlantic Anomaly region and over bright clouds where saturation may occur; (5) Snow-Ice information is now obtained from ECMWF weather data, increasing the number of valid retrievals at high latitudes. On average the NO2-v2.2 data have tropospheric VCDs that are between 10 and 40 % larger than the v1.x data, depending on the level of pollution and season; the largest impact is found at mid- and high-latitudes in wintertime. This has brought these tropospheric NO2 closer to OMI observations. Ground-based validation shows on average an improvement of the negative bias of the stratospheric (from −6 % to −3 %), tropospheric (from −32 % to −23 %) and total (from −12 % to −5 %) columns. For individual measurement stations, however, the picture is more complicated, in particular for the tropospheric and total columns.

Abstract. Stratospheric aerosols play a key role in atmospheric chemistry and climate. Their particle size is a crucial factor controlling the microphysical, radiative, and chemical aerosol processes in the stratosphere. Despite its importance, available observations on aerosol particle size are rather sparse. This limits our understanding and knowledge about the mechanisms and importance of chemical and climate aerosol feedbacks. The retrieval described by Malinina et al. (2018) provides the stratospheric particle size distribution (PSD) from SCIAMACHY (SCanning Imaging Absorption spectroMeter for Atmospheric CHartographY) limb observations in the tropics. This algorithm has now been improved and extended to work on the entire globe. Two PSD parameters of a unimodal lognormal PSD, the median radius and the geometric standard deviation, are retrieved between 18 and 35 km altitude from SCIAMACHY limb observations by a multiwavelength nonlinear regularized inversion. The approach assumes an aerosol particle number density profile that does not change during the retrieval. The effective Lambertian surface albedo pre-retrieved from coinciding SCIAMACHY nadir observations is integrated into the retrieval algorithm to mitigate the influence of the surface albedo on the retrieval results. The extinction coefficient and the effective radius are calculated from the PSD parameters. The aerosol characteristics from SCIAMACHY are compared with in situ balloon-borne measurements from Laramie, Wyoming, and retrievals from the satellite instruments of the Stratospheric Aerosol and Gas Experiment series (SAGE II and SAGE III) and Optical Spectrograph and InfraRed Imager System (OSIRIS). In the Northern Hemisphere, the median radius differs by less than 27 % and the geometric standard deviation by less than 11 % from both balloon-borne and SAGE III data. Differences are mainly attributed to errors in the assumed a priori number density profile. Globally, the SCIAMACHY extinction coefficient at 750 nm deviates by less than 35 % from SAGE II, SAGE III, and OSIRIS data. The effective radii from SCIAMACHY, balloon-borne measurements, and SAGE III agree within about 18 %, while the effective radius based on SAGE II measurements is systematically larger. The novel data set containing the PSD parameters, the effective radius, and the aerosol extinction coefficients at 525, 750, and 1020 nm from SCIAMACHY observations is publicly available.

TROPOspheric Monitoring Instrument observations of total column water vapour: Algorithm and validation

Abstract. The sensitivity of satellites to air pollution close to the sea surface is decreased by the scattering of light in the atmosphere and low sea surface albedo. To reliably retrieve tropospheric nitrogen dioxide (NO2) columns using the TROPOspheric Monitoring Instrument (TROPOMI), it is therefore necessary to have good a priori knowledge of the vertical distribution of NO2. In this study, we use an aircraft of the Royal Belgian Institute of Natural Sciences equipped with a sniffer sensor system to measure NOx (= NO + NO2), CO2 and SO2. This instrumentation enabled us to evaluate vertical profile shapes from several chemical transport models and to validate TROPOMI tropospheric NO2 columns over the polluted North Sea in the summer of 2021. The aircraft sensor observes multiple clear signatures of ship plumes from seconds after emission to multiple kilometers downwind. Besides that, our results show that the chemical transport model Transport Model 5, Massively Parallel version (TM5-MP), which is used in the retrieval of the operational TROPOMI NO2 data, tends to underestimate surface level pollution – especially under conditions without land outflow – while overestimating NO2 at higher levels over the study region. The higher horizontal resolution in the regional CAMS (Copernicus Atmosphere Monitoring Service) ensemble mean and the LOTOS-EUROS (Long Term Ozone Simulation European Operational Smog) model improves the surface level pollution estimates. However, the models still systematically overestimate NO2 levels at higher altitudes, indicating exaggerated vertical mixing and overall too much NO2 in the models over the North Sea. When replacing the TM5 a priori NO2 profiles with the aircraft-measured NO2 profiles in the air mass factor (AMF) calculation, we find smaller recalculated AMFs. Subsequently, the retrieved NO2 columns increase by 20 %, indicating a significant negative bias in the operational TROPOMI NO2 data product (up to v2.3.1) over the North Sea. This negative bias has important implications for estimating emissions over the sea. While TROPOMI NO2 negative biases caused by the TM5 a priori profiles have also been reported over land, the reduced vertical mixing and smaller surface albedo over sea make this issue especially relevant over sea and coastal regions.

The Scanning Imaging Absorption spectroMeter for Atmospheric ChartographY (SCIAMACHY) instrument, launched on the Envisat satellite in March 2002, measures the earthshine radiance, simultaneously from the ultraviolet (UV) to the near infrared (NIR), in the three viewing geometries: nadir, limb, and occultation. These measurements are used to retrieve both the total amount and vertical profiles of a large number of atmospheric constituents. In this paper, stratospheric ozone profiles between 15 and 40 km altitude are retrieved on 3 km grids from SCIAMACHY limb scattered radiance in the Chappuis-Wulf band. The study employs a new multiplicative algebraic reconstruction technique (MART) coupled with the radiative transfer model SCIATRAN. This technique is outstanding in that more than one measurement vector element can be used to retrieve the ozone density at any altitude. Furthermore, it is straightforward to understand, easy to implement and likely to produce stable results. Radiance normalization and wavelength pairing is applied to radiance as an intermediate step, using the wavelengths 525 nm, 600 nm and 675 nm. The sensitivity of ozone retrieval by this method to tangent altitude pointing, surface albedo, aerosol and cloud parameters is studied, and the results show that the retrieval impact due to tangent altitude pointing bias is the biggest up to 75% with 1 km shift, and the impact of albedo is limited within 5%. The effect of boundary visibility and cloud parameters can be ignored since their impact is too small. The effectiveness of the retrieval is demonstrated using a set of coincident SCIAMACHY products at Hefei that shows a mean bias of less than 12% between 15 and 40 km, and with a better accuracy of 5% from 16 to 36 km.

Abstract. We present a new tropospheric ozone dataset based on TROPOspheric Monitoring Instrument (TROPOMI)/Sentinel-5 Precursor (S5P) total ozone measurements combined with stratospheric ozone data from the Belgian Assimilation System for Chemical ObsErvations (BASCOE) constrained by assimilating ozone observations from the Microwave Limb Sounder (MLS). The BASCOE stratospheric data are interpolated to the S5P observations and subtracted from the TROPOMI total ozone data. The difference is equal to the tropospheric ozone residual column from the surface up to the tropopause. The tropospheric ozone columns are retrieved at the full spatial resolution of the TROPOMI sensor (5.5×3.5 km2) with daily global coverage. Compared to the Ozone Mapping and Profiler Suite Modern-Era Retrospective analysis for Research and Applications 2 (OMPS-MERRA-2) data, a global mean positive bias of 3.3 DU is found for the analysed period April 2018 to June 2020. A small negative bias of about −0.91 DU is observed in the tropics relative to the operational TROPOMI tropical tropospheric data based on the convective cloud differential (CCD) algorithm throughout the same period. The new tropospheric ozone data (S5P-BASCOE) are compared to a set of globally distributed ozonesonde data integrated up to the tropopause level. We found 2254 comparisons with cloud-free TROPOMI observations within 25 km of the stations. In the global mean, S5P-BASCOE deviates by 2.6 DU from the integrated ozonesondes. Depending on the latitude the S5P-BASCOE deviate from the sondes and between −4.8 and 7.9 DU, indicating a good agreement. However, some exceptional larger positive deviations up to 12 DU are found, especially in the northern polar regions (north of 70∘). The monthly mean tropospheric column and time series for selected areas showed the expected spatial and temporal pattern, such as the wave one structure in the tropics or the seasonal cycle, including a summer maximum, in the mid-latitudes.

Abstract. On 13 October 2017, the Tropospheric Monitoring Instrument (TROPOMI) was launched on the Copernicus Sentinel-5 Precursor satellite in a sun-synchronous orbit. One of the mission's operational data products is the total column concentration of carbon monoxide (CO), which was released to the public in July 2018. The current TROPOMI CO processing uses the HITRAN 2008 spectroscopic data with updated water vapor spectroscopy and produces a CO data product compliant with the mission requirement of 10 % precision and 15 % accuracy for single soundings. Comparison with ground-based CO observations of the Total Carbon Column Observing Network (TCCON) show systematic differences of about 6.2 ppb and single-orbit observations are superimposed by a significant striping pattern along the flight path exceeding 5 ppb. In this study, we discuss possible improvements of the CO data product. We found that the molecular spectroscopic data used in the retrieval plays a key role for the data quality where the use of the Scientific Exploitation of Operational Missions – Improved Atmospheric Spectroscopy Databases (SEOM-IAS) and the HITRAN 2012 and 2016 releases reduce the bias between TROPOMI and TCCON due to improved CH4 spectroscopy. SEOM-IAS achieves the best spectral fit quality (root-mean-square, rms, differences between the simulated and measured spectrum) of 1.5×10-10 mol s−1 m−2 nm−1 sr−1 and reduces the bias between TROPOMI and TCCON to 3.4 ppb, while HITRAN 2012 and HITRAN 2016 decrease the bias even further below 1 ppb. HITRAN 2012 shows the worst fit quality (rms = 2.5×10-10 mol s−1 m−2 nm−1 sr−1) of the tested cross sections and furthermore introduces an artificial bias of about -1.5×1017 molec cm−2 between TROPOMI CO and the CAMS-IFS model in the Tropics caused by the H2O spectroscopic data. Moreover, analyzing 1 year of TROPOMI CO observations, we identified increased striping patterns by about 16 % percent from November 2017 to November 2018. For that, we defined a measure γ, quantifying the relative pixel-to-pixel variation in CO in the cross-track and along-track directions. To mitigate this effect, we discuss two destriping methods applied to the CO data a posteriori. A destriping mask calculated per orbit by median filtering of the data in the cross-track direction significantly reduced the stripe pattern from γ=2.1 to γ=1.6. However, the destriping can be further improved, achieving γ=1.2 by deploying a Fourier analysis and filtering of the data, which not only corrects for stripe patterns in the cross-track direction but also accounts for the variability of stripes along the flight path.

Current and future satellite sensors provide measurements in and around the oxygen A-band on a global basis. These data are commonly used for the determination of cloud and aerosol properties. In this paper, we assess the information content in the oxygen A-band for the retrieval of macrophysical cloud parameters using precise radiative transfer simulations covering a wide range of geophysical conditions in conjunction with advance inversion techniques. The information content of the signal with respect to the retrieved parameters is analyzed in a stochastic framework using two common criteria: the degrees of freedom for a signal and the Shannon information content. It is found that oxygen A-band measurements with moderate spectral resolution (0.2 nm) provide two pieces of independent information that allow the accurate retrieval of cloud-top height together with either cloud optical thickness or cloud fraction. Additionally, our results confirm previous studies indicating that the retrieval of cloud geometrical thickness (CGT) from single-angle measurements is not reliable in this spectral region. Finally, a sensitivity study shows that the retrieval of macrophysical cloud parameters is slightly sensitive to the uncertainty in the CGT and very sensitive to the uncertainty in the surface albedo.

As part of the UV-Indien project, a station for measuring ultraviolet radiation and the cloud fraction was installed in December 2019 in Moroni, the capital of the Comoros, situated on the west coast of the island of Ngazidja. A ground measurement campaign was also carried out on 12 January 2020 during the ascent of Mount Karthala, located in the center of the island of Ngazidja. In addition, satellite estimates (Ozone Monitoring Instrument and TROPOspheric Monitoring Instrument) and model outputs (Copernicus Atmospheric Monitoring Service and Tropospheric Ultraviolet Model) were combined for this same region. On the one hand, these different measurements and estimates make it possible to quantify, evaluate, and monitor the health risk linked to exposure to ultraviolet radiation in this region, and, on the other, they help to understand how cloud cover influences the variability of UV-radiation on the ground. The measurements of the Ozone Monitoring Instrument onboard the EOS-AURA satellite, being the longest timeseries of ultraviolet measurements available in this region, make it possible to quantify the meteorological conditions in Moroni and to show that more than 80% of the ultraviolet indices are classified as high and that 60% of these are classified as extreme. The cloud cover measured in Moroni by an All Sky Camera was used to distinguish between the cases of UV index measurements taken under clear or cloudy sky conditions. The ground-based measurements thus made it possible to describe the variability of the diurnal cycle of the UV index and the influence of cloud cover on this parameter. They also permitted the satellite measurements and the results of the simulations to be validated. In clear sky conditions, a relative difference of between 6 and 11% was obtained between satellite or model estimates and ground measurements. The ultraviolet index measurement campaign on Mount Karthala showed maximum one-minute standard erythemal doses at 0.3 SED and very high daily cumulative erythemal doses at more than 80 SED. These very high levels are also observed throughout the year and all skin phototypes can exceed the daily erythemal dose threshold at more than 20 SED.

The retrieval of atmospheric composition from space-based measurements, by e.g., Sentinel-5p TROPOMI, is strongly affected by radiative interferences with clouds. Dedicated cloud data products, typically retrieved from measurements by the same sounder, are therefore essential. Cloud information is used to filter data and as input to the modelling of atmospheric radiative transfer and the conversion of slant column densities into vertical column densities. The three main TROPOMI cloud retrieval algorithms are: (i) L2_CLOUD OCRA/ROCINN CAL (Optical Cloud Recognition Algorithm/Retrieval of Cloud Information using Neural Networks; Clouds-As-Layers), (ii) L2_CLOUD OCRA/ROCINN CRB (Clouds-as Reflecting Boundaries), and (iii) the S5P support product FRESCO-S (Fast Retrieval Scheme for Clouds from Oxygen absorption bands for Sentinel). The cloud variables provided by these products (radiometric cloud fraction, cloud (top) height, and cloud albedo/cloud optical thickness) are subsequently used in the retrieval of the TROPOMI trace gas products. The quality of cloud products and trace gas products is routinely assessed by the ESA/Copernicus Atmospheric Mission Performance Cluster (ATM-MPC) validation service, with ad hoc support from Sentinel-5p Validation Team (S5PVT) AO projects. Version upgrades have had a significant impact on the characteristics of S5P cloud data. The change of the wavelength window in the FRESCO product since version 1.4 (‘FRESCO-wide’) leads to a clear increase in the height of low clouds with a large impact on the tropospheric NO2 retrieval (van Geffen, 2022), and improving the validation results regarding the tropospheric and total NO2 column. The first upgrades of the ROCINN products (from v1 to v2.1-v2.3) led to an increase in correlation with CLOUDNET cloud height, but to a more negative bias for the low clouds, with ROCINN CRB cloud height even dropping below the CLOUDNET cloud base height on average. However, this effect seems alleviated with the latest upgrade to v2.4. The impact on the HCHO validation results is investigated but is less clear compared to the NO2 case. To resolve the discontinuities due to the processor version jumps, a full mission reprocessing is currently ongoing and largely carried out for the L2_CLOUD and FRESCO-S products. The reprocessed ROCINN data have a lower dispersion and higher correlation with respect to the CLOUDNET cloud heights. The bias of the L2_CLOUD OCRA/ROCINN CAL CTH becomes more negative, but that of L2_CLOUD OCRA/ROCINN CRB CH bias improves. Finally, we also discuss the impact of the FRESCO-S reprocessing on the validation results.

Operational retrievals of tropospheric trace gases from space-borne spectrometers are based on one-dimensional radiative transfer models. To minimize cloud effects, trace gas retrievals generally implement Lambertian cloud models based on radiometric cloud fraction estimates and photon path length corrections. The latter relies on measurements of the oxygen collision pair (O2-O2) absorption at 477 nm or on the oxygen A-band around 760 nm. In reality however, the impact of clouds is much more complex, involving unresolved sub-pixel clouds, scattering of clouds in neighboring pixels and cloud shadow effects, such that unresolved three-dimensional effects due to clouds may introduce significant biases in trace gas retrievals. In order to quantify this impact, we study NO2 as a trace gas example, and apply standard retrieval methods including approximate cloud corrections to synthetic data generated by the state-of-the-art three-dimensional Monte Carlo radiative transfer model MYSTIC. A sensitivity study is performed for simulations including a box-cloud, and the dependency on various parameters is investigated. The most significant bias is found for cloud shadow effects under polluted conditions. Biases depend strongly on cloud shadow fraction, NO2 profile, cloud optical thickness, solar zenith angle, and surface albedo. Several approaches to correct NO2 retrievals under cloud shadow conditions are explored. We find that air mass factors calculated using fitted surface albedo or corrected using the O2-O2 slant column density can partly mitigate cloud shadow effects. However, these approaches are limited to cloud-free pixels affected by surrounding clouds. A parameterization approach is presented based on relationships derived from the sensitivity study. This allows identifying measurements for which the standard NO2 retrieval produces a significant bias, and therefore provides a way to improve the current data flagging approach.

Operational retrievals of tropospheric trace gases from space-borne spectrometers are based on one-dimensional radiative transfer models. To minimize cloud effects, trace gas retrievals generally implement Lambertian cloud models based on radiometric cloud fraction estimates and photon path length corrections. The latter relies on measurements of the oxygen collision pair (O2-O2) absorption at 477 nm or on the oxygen A-band around 760 nm. In reality however, the impact of clouds is much more complex, involving unresolved sub-pixel clouds, scattering of clouds in neighboring pixels and cloud shadow effects, such that unresolved three-dimensional effects due to clouds may introduce significant biases in trace gas retrievals. In order to quantify this impact, we study NO2 as a trace gas example, and apply standard retrieval methods including approximate cloud corrections to synthetic data generated by the state-of-the-art three-dimensional Monte Carlo radiative transfer model MYSTIC. A sensitivity study is performed for simulations including a box-cloud, and the dependency on various parameters is investigated. The most significant bias is found for cloud shadow effects under polluted conditions. Biases depend strongly on cloud shadow fraction, NO2 profile, cloud optical thickness, solar zenith angle, and surface albedo. Several approaches to correct NO2 retrievals under cloud shadow conditions are explored. We find that air mass factors calculated using fitted surface albedo or corrected using the O2-O2 slant column density can partly mitigate cloud shadow effects. However, these approaches are limited to cloud-free pixels affected by surrounding clouds. A parameterization approach is presented based on relationships derived from the sensitivity study. This allows identifying measurements for which the standard NO2 retrieval produces a significant bias, and therefore provides a way to improve the current data flagging approach.

A regional forecast model of the ultraviolet (UV) index has been developed by using a radiative transfer model and a multiple linear regression model to forecast total ozone over East Asia. This is a difficult and challenging task because of frequent cloud cover and atmospheric aerosols. The new, improved total ozone forecast model was constructed over each grid point in East Asia based on extensive investigation of the correlation between the total ozone and predictors related to the variation in total ozone. Root mean square errors (RMSEs) in the UV index between the forecast and satellite observations from the Scanning Imaging Absorption Spectrometer for Atmospheric Chartography (SCIAMACHY) for clear sky conditions range from 0.27 to 1.50 with an average of 0.79 in monthly statistics. Although the patterns of the corrected UV index when applying the cloud modification factor (CMF) and the aerosol modification factor (AMF) compare reasonably well with those of the UV index measured with the Ozone Monitoring Instrument (OMI), the performance depends on the accuracy of forecast for cloud and aerosol optical depth (AOD). Additional consideration of surface albedo and cloud optical depth are required to further refine and improve the accuracy of the prediction.

Algorithm for retrieving surface pressure from hyper-spectral measurements in oxygen A-band