The TROPOspheric Monitoring Instrument (TROPOMI) spectrometer aboard the polar orbiting Copernicus Sentinel-5P (S5P) satellite, launched on 13 Oct. 2017, measures several trace gases (notably O3, NO2, SO2, HCHO, CH4, CO) as well as some aerosol and cloud properties over the entire globe. Measurements are performed in a 2600 km wide swath, with a resolution of 7 x 3.5 km2 at nadir, achieving global coverage each day.

Nitrogen dioxide (NO2) tropospheric and total column data derived from satellite-based observations are of great value for air quality and climate studies and essential for emission estimates. TROPOMI NO2 tropospheric columns are retrieved using a DOAS algorithm combined with an integrated modelling / retrieval-assimilation approach based on the TM5-MP chemistry-transport model (operating at 1ºx1º resolution and driven by ECMWF meteo data) to estimate stratospheric NO2 columns and to derive air-mass factors. Developments from the EU QA4ECV project (www.qa4ecv.eu) have been included in the retrieval to ensure long-term data record consistency.

This contribution describes the TROPOMI NO2 SCD retrieval results, focussing on their long-term stability, and it investigates NO2 SCD values, the uncertainties following from the DOAS fit, as well as the uncertainties using an independent statistical approach {Zara et al, 2018], in comparison with QA4ECV results for the Ozone Monitoring Instrument (OMI), aboard the EOS-Aura satellite.

The uncertainty on the TROPOMI NO2 SCDs are much smaller than OMI’s (see the graph), because of the much higher signal-to-noise ratio: the TROPOMI SNR is about 1500, the OMI SNR was about 900-1000 back in 2005, despite the much small ground pixel size. TROPOMI shows very small across-track variation, much smaller than the “striping” in the OMI NO2 data. There appears to be some small smooth across-track variation, which gives the Level 0-to-1b team valuable information for improving the (ir)radiance calibration.

Off-line NO2 data of the operational (E2) phase, which started 30 April 2018, and near-real time (NRT) data as of 3 July 2018 are available from the Copernicus Open Access S5P data hub (s5phub.copernicus.eu).

Figure caption: Along track averages over the nadir latitude range [-20º:+20º] of the NO2 slant column error estimate, following from the DOAS fit, as function of the across-track viewing zenith angle, for the Pacific Ocean orbit of TROPOMI on 4 July 2018 (red solid line) and the corresponding OMI orbit (blue solid line, where rows affected by the so-called ’row anomaly’ have been omitted). For comparison, results for the Pacific Ocean orbit of OMI on 1 July 2005 are also shown (dark-gray dashed line).The OMI data is from QA4ECV v1.1.

Area D - Board 361

The TROPOMI L01b processor was developed well prior to the start of the Sentinel-5 Precursor mission on the 13th of October 2017. Rather than using prototype code, the L01b processor was already employed during the on-ground calibration campaign which started in December 2014.
The processor was then used to derive self consistently all necessary calibration key data. The key data serves as an input for the processor to convert the raw input data into calibrated spectral radiances and irradiances.
The geolocation, so the calibrated and validated longitude and latitude for each measured groundpixel, is also allocated to the radiance data in the L01b processor.
As described in the Algorithm Theoretical Basis Document (http://www.tropomi.eu/document/tropomi-l01b-atbd), the processor represents an implementation of the reverse instrument model, including all newly observed features.
The output of the L01b processor is radiance, irradiance and special calibration data. The radiance and irradiance data is used in the L2 processors to derive global information on concentrations of trace gases and aerosols important for air quality, climate forcing, and the ozone layer. The irradiance data by itself is also used for research on the solar output.
The radiance and irradiance output is part of both the near-real time and the offline products.
The architecture of the L01b processor includes the use of multi-threading: calculations are performed in parallel to allow for fast computing. An important part of the L01b processor is to allow for constant monitoring of calibration parameters. To that end special algorithms are implemented and data averaging is performed. The output calibration data is ingested into a database which supports the web based monitoring portal (see talk "In-flight performance of TROPOMI" by A. Ludewig et al.).
In-flight monitoring and re-calibration showed that the in-flight instrument calibration can be improved with respect to the on-ground calibration key data. In-flight degradation and new instrument features need to be represented by the L01b processor's algorithms and calibration key data.
The new version (v2.0.0) of the L01b processor addresses these issues. In the talk we present instrument features and how the processor update addresses these with new algorithms and time dependent calibration key data.

Area D - Board 362

The ESA Atmospheric Validation Data Centre (EVDC) serves as a central, long-term repository for archiving and exchange of correlative data for validation of atmospheric composition products from satellite platforms. The EVDC builds on the previous ENVISAT Cal/Val database system in operation at NILU since the early 2000s and provides tools for extraction, conversion and archival. The objective of the current ESA funded project lead by Skytek [1] with the partnership of NILU [2] and ICHEC [3], is to provide an online information system that supports users in managing and exploiting campaign datasets for Earth Observation missions and applications.
Through the web portal https://evdc.esa.int users can access to a large variety of data from campaigns, in-situ ground-based measurements, aircraft, balloons and, in general, from a wide range of stations and measurements principles for validation of the satellite atmospheric composition products. An important aspect for the Cal/Val data is the standardization of the format to enhance the usability of correlative data and ensure an extensive quality control. With this aim the portal offers numerous conversion tools and specific support for conversion to the GEOMS format [4].
EVDC provides, moreover, access to satellite Level-2 products for specific missions in particular Sentinel-5P and AEOLUS. Some new features and tools are available online via the EVDC portal e.g: the Orbit Predictor Overpass Tool (OPOT) and the Sub-Setting tool. The OPOT is based on the TLEs (Two-Line Element set) [5] and uses the Simplified General Perturbation Model (SGP4) [6] to predict and store their future orbits. Given a location or a region of interest (defined as a polygon) the OPOT produces a list of overpasses for that region and satellite for a future time range. The Sub-Setting facility uses the HARP toolkit [7] as its backbone. Users searching for Sentinel-5P data, can choose to perform a sub-setting operation on the file instead of downloading the entire file, moreover the data extraction can be scheduled specifying location(s), temporal window, distance from spatial reference through a systematic service which make the data package available once the service job is completed.

REFERENCES
[1] Skytek http://www.skytek.com/; info@skytek.com
[2] NILU Norwegian Institute for Air Research. http://www.nilu.no/; nadirteam@nilu.no.
[3] ICHEC Irish Centre for High-End Computing. https://www.ichec.ie/
[4] GEOMS documentation: http://evdc.esa.int/documentation/geoms/
[5] Two Line Element set format: https://www.celestrak.com/NORAD/documentation/tle-fmt.php
[6] Models for Propagation of NORAD Element Sets. F. R. Hoots R. L. Roehrich 1980. https://www.celestrak.com/NORAD/documentation/spacetrk.pdf
[7] HARP documentation: https://cdn.rawgit.com/stcorp/harp/master/doc/html/index.html

Area D - Board 363

Launched on October 13, 2017, ESA’s Sentinel-5 Precursor (S5P) is the first atmospheric composition satellite of European Union's Copernicus Earth Observation programme. On board of S5P, the UV/VIS/NIR/SWIR spectrometers of TROPOspheric Monitoring Instrument (TROPOMI) measure on the global scale, on a daily basis, and at unprecedented horizontal resolution, the atmospheric abundance of species related to air quality, climate forcing, stratospheric ozone, UV radiation and volcanic hazards: tropospheric and stratospheric O₃, NO₂, HCHO, SO₂, CO, CH₄, clouds, aerosols... As part of the Copernicus Space Component, S5P contributes data to the production of advanced information services on our environment and security - e.g., through data assimilation in the Copernicus Atmosphere Monitoring Service (CAMS) run at ECMWF. Therefore, data quality is essential and compliance with mission requirements must be verified appropriately over the mission lifetime.

Procured by an international consortium contracted by ESA and supported by The Netherlands, Germany and Belgium, the S5P Mission Performance Centre (MPC) provides an operational service-based response to the mission requirements for quality control, calibration, validation and end-to-end system performance monitoring during the Routine Operations Phase of the S5P mission. The Routine Operations Validation Service provided by the S5P MPC relies in large parts on the Validation Data Analysis Facility (VDAF) and its Automated Validation Server. Building upon a two-decade heritage of geophysical validation research with precursor instruments (GOME, SCIAMACHY, OMI, GOME-2A/B) and on recent advances in Cal/Val practices and validation systems (e.g., EC FP7 QA4ECV and H2020 GAIA-CLIM), VDAF has been tailored to the TROPOMI and Copernicus needs, with special emphasis on operational and service-oriented aspects. VDAF ingests Fiducial Reference Measurements (FRM) archived at ESA’s Validation Data Centre (EVDC), and other quality-assured validation data collected from ground-based monitoring networks (GO3OS, NDACC, SHADOZ, TCCON, TOLNET…), satellites, dedicated field campaigns and numerical modelling. Its Automated Validation Server compares those data to TROPOMI data following community-endorsed protocols in an automated environment, to assess the overall quality, the compliance with mission requirements and the validity of uncertainty estimates of the TROPOMI data products. This service monitors validation results on a cyclic basis and produces every three months the Routine Operations Consolidated Validation Report (ROCVR), available publicly. The service also contributes quality assessment support to the evolution of S5P L2 algorithms and data processors. Validation of the operational Level-1 and Level-2 data products is documented on the portal of the TROPOMI Validation Data Analysis Facility at http://mpc-vdaf.tropomi.eu .

In this paper we report on one year of achievements of the MPC validation service for the S5P TROPOMI operational data products delivered publicly. After a short description of the heritage systems and of the scientific and technical advances implemented in the S5P VDAF, latest validation results are summarized for all S5P operational data products to date. This paper concludes with remaining challenges and with perspectives on further operationalisation of satellite validation facilities for the Copernicus Sentinels.

Area D - Board 364

Continuous monitoring of the health and performance of the TROPOMI instrument is a key ingredient to ensure continuous production of high quality of data products over the various modules.
After launch of TROPOMI on board of ESA's Sentinel-5 Precursor satellite, the first six months in-orbit were reserved to perform inflight calibration of both the UVN and SWIR module. This phase was an essential step in acquiring the calibration necessary to deliver high quality products and complemented the calibration done on-ground. In this presentation, we will present the results of all measurements relevant to the SWIR module carried out during the inflight calibration.

Measurements in this phase also provided initial conditions required to monitor the health of the SWIR module during the full lifetime of TROPOMI.
Monitoring is done using a range of measurements during the eclipse side of an orbit. This includes responses of the various on-board calibration sources.

Measurements to derive the dark flux, offset and noise were planned in the eclipse side of each orbit. However, the sensitivity of the SWIR detector revealed contaminats in the form of Earthshine and/or point sources to be affect the derivation. As such, it was chosen to perform these measurements with the folding mirror closed at less regular intervals. These measurements provide a quantification of the SWIR detector performance on a per-pixel basis.

Dark measurements are affected by the impacts of cosmic rays, which potentially harm the SWIR detector on the long term. We will present the rate at which pixels are lost during a full year, providing a clear outlook on the future of the SWIR module.

Using daily measurements of the solar irradiance in combination with the responses of onboard calibration sources, the stability of the radiometry is monitored. Radiometry includes the transmission of incoming signal to the detector as well as the amount and distribution of any stray light.

Last but not least, the stability of the Instrument Spectral Response Function, which is used in the SWIR L2 retrievals, such as CO and CH4, is monitored. Straylight and ISRF are monitored using dedicated on-board diode lasers.

We will show that the TROPOMI-SWIR module is in excellent health. It also shows that it will likely continue to produce very high quality L1b data suitable for L2 retrievals over the full lifetime of TROPOMI

Area D - Board 365

The Copernicus Sentinel-5 Precursor (S-5P) is the first of a series of atmospheric chemistry missions within the European Commission’s Copernicus Programme, launched successfully in October 2017. S-5P entered its operational Phase at the end of April 2018 and provides continuity in the availability of global atmospheric data products between its predecessor missions SCIAMACHY (Envisat) and OMI (AURA) and the future Sentinel-4 and -5 series.

S-5P delivers unique data regarding the sources and sinks of trace gases with a focus on the lower Troposphere including the planet boundary layer due to its enhanced spatial, temporal and spectral sampling capabilities as compared to its predecessors.
The S-5P satellite carries a single payload, namely TROPOMI (TROPOspheric Monitoring Instrument) that was jointly developed by The Netherlands and ESA. Covering spectral channels in the UV, visible, near- and short-wave infrared, it measures various key species including tropospheric/stratospheric ozone, NO2, SO2, CO, CH4, CH2O as well as cloud and aerosol parameters.

The geophysical validation and characterization of the TROPOMI Level 1 and Level 2 data products during the phase E2 is conducted by ESA at different levels. The so-called Mission Performance Center carries out the routine validation throughout the mission life-time and rely on the availability of independent data sets for example from ground based measurements or the so-called Fiducial Reference Measurement (FRM) data sets, which are a suite of independent, fully characterized and traceable ground measurements. FRMs will progressively become available on the ESA Atmospheric Validation Data Centre (EVDC). In addition, within the framework of its Copernicus missions, ESA opened a call for the Sentinel-5 Precursor Calibration Validation Team (S5PVT), that aims to engage leading expertise for the Calibration and Validation of the S5P in the mission validation team, providing independent experimental data, analysis and recommendations to critically assess the end-to-end performance of the instrument and its products.

The overall ESA S5P Validation approach during the operational phase E2 will be presented in this paper.

Area D - Board 366

The TROPOspheric Monitoring Instrument (TROPOMI) has been launched on October 13th, 2017, aboard the polar orbiting platform Sentinel-5 Precursor (S5p). TROPOMI measures the Earth’s radiance in the ultraviolet, visible, near and short-wave infrared spectral ranges with an unprecedented spatial resolution of 7x3.5km², providing important information on natural and anthropogenic emissions of trace gases and aerosols. Although currently not part of the suite of operational products, glyoxal tropospheric columns can be retrieved from TROPOMI measurements in the visible spectral range. The low optical depth of glyoxal makes its retrieval prone to instrumental noise and sensitive to spectral interferences. Nevertheless, this challenging product offers the potential to provide additional quantitative information on VOC emissions.

The BIRA-IASB glyoxal algorithm, successfully applied in the past to GOME-2A/B and OMI, has been transferred to TROPOMI and we present here results for more than one year of data. Based on comparisons with OMI, we illustrate the benefit of the excellent TROPOMI spatial resolution and signal-to-noise ratio to better identify and characterize the sources of this tropospheric trace gas. We also compare the retrieved TROPOMI glyoxal data with columns simulated using the state-of-the-art CTM MAGRITTE in different regions of the world. The glyoxal/formaldehyde ratio is often presented as a metric enabling to discriminate different types of VOC emissions. Using the operational TROPOMI HCHO products also relying on a DOAS algorithm developed at BIRA-IASB, those ratios are computed at the global scale for different seasons and we investigate whether systematic patterns can be identified as a function of the emission type.

Area D - Board 367

We present a comprehensive assessment of carbon monoxide (CO) total column retrievals from Sentinel-5P (S5P) Short-Wave Infrared (SWIR) observations. The Level1-2 processing is performed using the Beer Infrared Retrieval (BIRRA) which performs a least squares fit of Earth’s radiance (essentially transmission) to retrieve the molecular column densities (essentially density scaling factors) along with some auxiliary parameters (reflectivity, etc.) [Gimeno Garcia et al., AMT 2011]. It has been developed at DLR since about 2005 and its computational core modules have been integrated in the operational SCIAMACHY (Scanning Imaging Absorption Spectrometer for Atmospheric Chartography) Level1b-2 processor to retrieve CO from channel 8 (2.3 μm) and methane from channel 6 (1.6 μm) nadir observations. In the framework of DLR's SCIAMACHY activities BIRRA CO and CH4 columns were thoroughly examined in several intercomparisons [e.g. Hochstaffl et al., Remote Sens. 2018].

The latest version of the BIRRA prototype is featuring new enhancements in the GARLIC (Generic Atmospheric Radiation Line-by-line Infrared Code) [Schreier et al., JQSRT 2014] forward model. It utilizes, among other things, additional line shape parameters given in the latest spectroscopic databases such as the SEOM-IAS (Scientific Exploitation of Operational Missions-Improved Atmospheric Spectroscopy), a project that intended the improvement of spectroscopic information in preparation for the S5P mission. An updated framework is providing appropriate auxiliary information (e.g. a priori atmospheric profiles of pressure, temperature, and concentrations of interfering species, cloud information, topography, the latest molecular spectroscopy data, etc.).

Finally, the results are assessed by a thorough comparison with other data products. More specifically, an estimation of the quality of our CO product in view of the accuracy requirements defined in the mission preparation phase (i.e. 15% for CO) is given using NDACC (Network for the Detection of Atmospheric Composition Change) or TCCON (Total Carbon Column Observing Network) ground-based observations.

Area D - Board 368

The Tropospheric Monitoring Instrument (TROPOMI) was lunched on the 13 October 2017 on ESA’s Sentinel 5 Precursor satellite. One of the primary targets of the mission is to measure the atmospheric total column concentration of carbon monoxide (CO). The operational CO retrieval algorithm of TROPOMI was developed by SRON the Netherlands institute for Space Research using TROPOMI’s shortwave infrared (2.3 μm) measurements. In this contribution, we report on the status of the TROPOMI CO dataset and discuss recent improvements e.g. the use of the IAS molecular spectroscopic dataset. We assess the impact of the spectroscopy on the CO data product analyzing both spectral residuals of the fit and the validation with measurements of the TCCON network. Moreover, we focus on the novelty of the TROPOMI dataset, which provides CO with daily global coverage and a high spatial resolution of 7x7 km2. Because of the high signal-to-noise ratio of individual measurements, pollution hot-spots can be detected from measurements of single orbit overpasses, which opens up new opportunities for emission monitoring of pollution point sources on daily scales. For a set of cases, we demonstrate this application in more detail discussing CO emission estimates from wild fires and industrial activities using TROPOMI observations.

Area D - Board 369

Within the ESA project TCCON4S5P (ID 28603) accurate ground reference for CO and CH₄ measurements of the Tropospheric Monitoring Instrument (TROPOMI) on-board the ESA satellite Sentinel-5 Precursor (S5P) is provided using ground-based FTIR soundings from the Total Carbon Column Observing Network (TCCON).

When comparing retrievals of two independent remote sensing instruments, large biases can occur, depending on the dynamic situation affecting the vertical concentration profile of the gas species observed. This intercomparison error arises because the a priori information used in the retrievals of the two instruments is both different (instrument 1 vs. instrument 2) and, to some extent, unrealistic (instrument 1/2 vs. reality). Problems typically occur where the real vertical profiles exhibit an extreme variability in shape: E.g. Ostler et al. (2014) found unacceptably large biases in column-averaged CH₄ (up to 30 ppb), comparing mid- and near-infrared FTIR (NDACC vs. TCCON) measurements over the Arctic observation site Ny Ålesund during stratospheric subsidence events. Such effects were found to diminish towards lower latitudes, i.e. the intercomparison error produced a significant latitudinal bias. Also seasonal biases turned out to critically depend upon the detailed choice of priors. Obviously, such biases are also expected to occur in the comparison of TROPOMI vs. TCCON, effectively rendering TCCON validation data from high-latitude sites less suitable for validation purposes.

To overcome this drawback, the goal of this work is to reduce the intercomparison error via the use of independent profile information. This effort is led by KIT/University of Augsburg, and related commitments are fulfilled within the German DLR-funded project S5P-FTIR-UniA. We try to find an optimum reduction of the intercomparison error by correcting the satellite and ground retrievals to a common prior which resembles the true concentration profile at the moment of observation as closely as possible. Past work like Sussmann et al. (2013) and Ostler et al. (2014) demonstrated the potential of this method to reduce the intercomparison error by applying modelled vertical profiles from a chemistry transport model as common prior for NDACC and TCCON measurements of CH₄. Similarly, in our current work we investigate both the use of the Transport Model, version 5 (TM5; Krol et al., 2005) providing profiles for CH₄ and CO, and/or quasi-simultaneous NDACC (mid-infared) profile retrievals for the case of CO. After elaborating the best correction method for CH₄ and CO, respectively, we plan to produce corrected data series for both TROPOMI and TCCON. Our poster will show the current status of our work.

References

Krol, M., Houweling, S., Bregman, B., van den Broek, M., Segers, A., van Velthoven, P., Peters, W., Dentener, F., and Bergamaschi, P.: The two-way nested global chemistry-transport zoom model TM5: algorithm and applications, Atmos. Chem. Phys., 5, 417–432, doi:10.5194/acp-5-417-2005, 2005.

Ostler, A., Sussmann, R., Rettinger, M., Deutscher, N. M., Dohe, S., Hase, F., Jones, N., Palm, M., and Sinnhuber, B.-M.: Multistation intercomparison of column-averaged methane from NDACC and TCCON: impact of dynamical variability, Atmos. Meas. Tech., 7, 4081–4101, doi:10.5194/amt-7-4081-2014, 2014.

Sussmann, R., Ostler, A., Forster, F., Rettinger, M., Deutscher, N. M., Griffith, D.W. T., Hannigan, J.W., Jones, N., and Patra, P. K.: First intercalibration of column-averaged methane from the Total Carbon Column Observing Network and the Network for the Detection of Atmospheric Composition Change, Atmos. Meas. Tech., 6, 397–418, doi:10.5194/amt-6-397-2013, 2013.

Area D - Board 370

We present the first results of the comparison between satellite-based TROPOMI (TROPOspheric Monitoring Instrument) NO₂ products and ground-based observations in Helsinki. TROPOMI NO₂ total (summed) columns are compared with the measurements performed by the Pandora spectrometer in Helsinki (Finland) during April-September 2018. We find a general agreement between satellite- and ground-based data, but also that TROPOMI total columns underestimate ground-based observations for relatively large Pandora NO₂ total columns, corresponding to episodes of localized and relatively elevated pollution. This is expected because of the different representativeness of a TROPOMI ground pixel compared to the relatively small field-of-view of the Pandora instrument.

We also analyse the NO₂ surface concentrations measured from the Helsinki-Kumpula air quality station (located a few metres from the Pandora spectrometer). We find temporal correlation between TROPOMI, Pandora and NO₂ surface concentrations, with NO₂ enhancements observed during the same days. The daily variability is mostly well captured by the TROPOMI NO₂ retrievals. Furthermore, we compare the weekly cycles from TROPOMI, Pandora and NO₂ surface concentrations. Both satellite- and ground-based data show a similar weekly cycle, with lower NO₂ levels during the weekend compared to the weekdays. This is expected for an urban site such as Kumpula (Helsinki), since emissions from traffic and industrial activities are reduced during the weekend.

Finally, we demonstrate the added value of TROPOMI NO₂ observations in detecting point sources in Finland. This is achieved by rotating individual pixels about a supposed point source to achieve a unified wind direction, and time-averaging the results to a predefined high-resolution grid.

This is one of the first works in which TROPOMI NO₂ retrievals are validated against ground-based observations, and the results provide an early evaluation of their applicability for monitoring polluting emissions. TROPOMI retrievals have proved to be able to describe air quality features and variability similar to surface observations, even for relatively small pollution sources. The work has been carried on within the Sentinel-5P validation team activities (HIGHval project).

Area D - Board 371

The TROPOspheric Monitoring Instrument (TROPOMI) was launched on 13 October 2017. The first TROPOMI measurement was obtained on 7 November 2017. The objective of TROPOMI is to provide accurate and timely observations of key atmospheric species, for services on air quality, climate, and the ozone layer. TROPOMI will also contribute to services on volcanic ash for aviation safety.

TROPOMI validation is carried out using a routine validation approach and with campaign-based efforts. The validation using ship measurements is important for evaluating TROPOMI performance over ocean and covering a large geophysical area. Three measurement cruises on board the German research vessels R.V. Sonne and R.V. Maria Merian for cross Atlantic and Pacific were planned. The first ship measurement for the validation of TROPOMI has been carried out on the R.V. Sonne from Emden (Germany) to Buenos Aires (Argentina) (17 December 2017 – 9 January 2018) . The other two cruises were planned in December 2018 from Cape Verde to Panama and in February 2019 from Fiji to Mexico.

We explored atmospheric properties of trace gases and aerosols using MAX-DOAS instruments (from KNMI and MPI-C) and 2 Calitoo sunphotometers. Besides our own instruments, there were visible and infrared cloud cameras from MPI-M and a ceilometer from MPI-C onboard to measure clouds. There were NASA sunphotometers (Microtops) to measure aerosols and water vapor. We can also get data from the ship, such as the pitch, roll, latitude, longitude, and standard meteorological observations. During the first cruise, TROPOMI was still in the commissioning phase. However, we managed to get some collocated TROPOMI NO₂ and UVAI products to compare with the MAX-DOAS and Calitoo measurements. More MAX-DOAS and Calitoo data are to be collected for the TROPOMI validation in the 2018 and 2019.

In the presentation, we will describe the MAX-DOAS and Calitoo measurements on the ships and show some preliminary results of the validation of TROPOMI NO₂ using MAX-DOAS measurements.

Area D - Board 372

TROPOsheric Montoring Instrument (TROPOMI) onboard the Copernicus Sentinel-5 Precursor is expected to provide information on the global atmospheric compositions. Furthermore, it could play as a traveling standard for upcoming sensors onboard geostationary satellite missions such as GEMS, TEMPO, and Sentienl-4. To assess its potential as an inter-comparison reference, it is important to characterize and validate its observation data. Here, we use a series of high-resolution solar reference spectra (SAO, KNMI, QAS, and JPL) to evaluate the TROPOMI solar observations. These references spectra are selected taking into account of the spectral coverage and resolution of TROPOMI. As the sun can be seen in TROPOMI’s solar irradiance port for about two minutes around orbit phases 0.75 in all orbits, the TROPOMI measures the solar irradiance approximately once a day, every 15 orbits. All observed irradiances from the launch to the present are compared. For the evaluation, the reference solar spectra are convolved with the instrument spectral response function (ISRF) of TROPOMI. Since the ISRF is a key parameter characterizing the spectral features, on-orbit ISFR is derived and compared with the ground measured ISRF. During the on-orbit ISRF derivation, variation of the ISRF characteristics due to the choice of high-resolution solar reference spectra are also investigated. The sensitivity of retrieval performance due to the high-resolution solar reference spectra used are found to be insignificant. The cross-track variations of width parameter are smaller than the variations of the shape parameter. The comparison between measured solar irradiance and the reference solar irradiance shows a good agreement. Correlation coefficients range from 0.97 to 0.99 for all wavelengths. The radiometric discrepancies along the wavelength are less than 2 % except band 1 (270 to 300 nm). Though it is in a good shape, a few issues regarding the TROPOMI solar irradiances need to be identified. The noticeable inconsistencies are observed in overlapping wavelengths between some spectral bands (band 2 to 3, and band 5 to 6). The presentation will cover detailed validation results and characteristics of on-orbit ISRF.

Area D - Board 373

NOAA Suomi National Polar Partnership (SNPP) Visible Infrared Imaging Radiometer Suite (VIIRS) aerosol detection product (ADP) identifies the presence of smoke and dust in the atmosphere. The ADP algorithm uses measurements at 412 nm, 440 nm, and 2.25 µm to derive absorbing aerosol index (AAI) and Dust Smoke Discrimination Index (DSDI) to identify the presence of dust and smoke in the atmosphere. The ADP has been operational at NOAA since July 2017 and is shown to have 80% probability of correct detection (POCD) by comparing to Cloud Aerosol Lidar with Orthogonal Polarization (CALIOP) vertical feature mask. Aerosol Robotic Network (AERONET) Angstorm Exponent data are also used to classify the observations into smoke or dust based on certain criteria for coarse and fine mode particles. The VIIRS ADP has a lot of false dust detections when the instrument views the Earth in a nadir view, especially over bright desert surfaces. To minimize the false detections, the Brightness Temperature Difference (BTD) tests between various infrared bands (3.9 µm, 11 µm, 12 µm) are used to identify the presence of dust when satellite is in nadir view. The highly absorbing smog over Asia is also falsely identified as smoke in the VIIRS ADP product and we are attempting to find ways to discriminate smoke (from biomass burning) from urban pollution (smog).
The NOAA proposed activity to European Space Agency (ESA) for Sentinel 5 Precursor Tropospheric Ozone Monitoring Instrument (S5P TROPOMI) calibration/validation activities include exploring the possibility of a synergistic retrieval using measurements from VIIRS and TROPOMI to improve the accuracy of VIIRS ADP. We have begun comparing AAI between TROPOMI and VIIRS for several biomass burning events in July and August 2018 in the United States and Canada. Early analysis indicates that TROPOMI AAI values have a low bias and thus missed detecting the full extent of smoke plume correctly. The full statistical analysis of the two months of VIIRS ADP and TROPOMI AAI with results stratified for dust, smoke, and haze will be presented.

Area D - Board 374

Several new portable spectrometers have become available during the recent years. However the performances of these instruments have not been fully characterized. The ESA project “Fiducial Reference Measurements for Ground-Based Infrared Greenhouse Gas Observations (FRM4GHG)” aims at characterizing several portable spectrometers for measuring the greenhouse gases CO2, CH4 and CO, and at using their data for validation of the ESA Sentinel-5 Precursor (S5P) instrument (TROPOMI). The spectrometers have been collocated at the Sodankylä site of the Total Carbon Column Observing Network (TCCON; Wunch et al., 2011) and they have been performing greenhouse gas measurements during the years 2017 and 2018, next to the standard TCCON instrument. The latter has been operated in both normal high-resolution mode and also in the special low-resolution mode, similar to the portable spectrometers. The remote sensing total column measurements from the spectrometers are accompanied by balloon borne AirCore (Karion et al., 2010) in-situ measurements of the vertical profiles of the greenhouse gases up to about 25 km altitude. These AirCores have been launched on a regular basis from the Sodankylä site. The acquired profiles are calibrated to WMO standards, and are used as reference profiles for the remote sensing measurements.

In this presentation, the almost final results of the project will be shown, including results of the validation of S5P using the low resolution spectrometer data. We will also discuss the non-linearity effects that have been discovered in the TCCON spectra, and the methodologies that have been developed to correct for these effects.

The presentation will include future perspectives regarding the ground-based networks for remote sensing measurements of greenhouse gases, and their exploitation for satellite and model validation purposes and for carbon cycle studies.

References

Karion, A., Sweeney, C., Tans, P., Newberger, T., AirCore: An Innovative Atmospheric Sampling System, J. Atmos. Ocean. Technol., 27, doi:10.1175/2010JTECHA1448.1, 2010.

Wunch, D., Toon, G.C., Blavier, J.F.L., Washenfelder, R.A., Notholt, J., Connor, B.J., Griffith, D.W.T., Sherlock, V., Wennberg, P.O., The Total Carbon Column Observing Network, Phil. Trans. R. Soc. A, 369, doi:10.1098/rsta.2010.0240, 2011.

Area D - Board 375

Ground-based remote sensing multi-axis MAXDOAS and direct-sun PANDORA instruments measuring in the UV-Vis region are nowadays widely used to monitor atmospheric NO₂ columns. Owing to the multiple geometries used, the technique can be used to differentiate between total, tropospheric and stratospheric NO₂ content. Under favorable conditions they can also provide coarse information on the vertical distribution of this molecule in the low troposphere. The technique provides an excellent source of correlative data for validation of tropospheric and total NO₂ columns retrieved from nadir sensors, such as GOME-2, OMI and TROPOMI, that strongly rely on a-priori knowledge of the NO₂ vertical distribution, surface albedo and information on cloud cover and height.

In this study we combine ground-based remote sensing correlative measurements available from over 50 sites distributed worldwide from different networks (institute-based network, QA4ECV network, S5PVT NIDFORVAL uv-vis network) to address the validation of (1) the GOME-2 GDP 4.8 NO₂ products generated within the EUMETSAT AC SAF project, (2) the OMI TEMIS product and (3) the operational TROPOMI product. A negative bias is generally found between the different satellites and the ground-based datasets, which is different for the suburban and remote locations case and in the urban locations case. In order to investigate how much these biases are due to the station locations and the satellite pixel sizes (GOME-2 40x80km², OMI 13x24km² and TROPOMI 7x3.5km²), an OMI NO₂ QA4ECV high-resolution grid data is created and used to quantify the dilution happening around each station. A dilution correction method is proposed and applied to the different validation results, showing global improvement for GOME-2 and OMI, especially in the urban locations. The impact of the pixels size and location is investigated and the extension of this method to TROPOMI comparisons is investigated. Tropospheric NO₂ columns are also calculated from Direct-sun total NO₂ data by removing the satellite-estimated stratospheric contribution, and the coherence to MAXDOAS tropospheric comparisons is then discussed.

Area D - Board 376

Within the framework of the Network for the Detection of Atmospheric Composition Change (NDACC), we use ground-based infrared Fourier transform spectrometers (FTIR), to measure the concentrations of greenhouse gases in the atmosphere, such as carbon dioxide (CO2), methane (CH4), nitrous oxide (N2O) as well as stratospheric trace gases that are important for ozone chemistry, such as ozone (O3), hydrogen chloride (HCl), hydrogen fluoride (HF), carbon monoxide (CO), nitric acid (HNO3), formaldehyde (HCHO) and chlorine nitrate (ClONO2). The FTIR sites are Kiruna (Sweden), Izaña on Tenerife and Karlsruhe on the Campus Nord of the Karlsruhe Institute of Technology (KIT).

ESAs Sentinel-5 precursor satellite (S5P) was launched in October 2017. It is an atmospheric composition satellite of the Copernicus programme, which carries the Tropospheric Monitoring Instrument (TROPOMI). Since then the spectrometer collects, among other compounds, the trace gases ozone, carbon monoxide, methane and formaldehyde. Satellite remote sensing techniques are needed to obtain information on the distribution of atmospheric trace gases on a global and regional scale. The validation of the measurements is of great importance for the quality assurance of satellite-based remote sensing systems. As part of this work, space-based column observations are compared with robust ground-based FTIR column observations. Within the scope of the project, the three FTIR sites provide atmospheric column contents of CO, CH4 and HCHO. For methane and carbon monoxide, the ground-based measurements also provide information on the height distribution of the gases. We will use these datasets for the validation of ESA's Sentinel-5 precursor satellite data.

Therefore, we want to present our first comparisons of the used evaluation setups of the S5P target gases as well as our first results of the time series of these trace gases, in particular of CO (total column) derived from the ground-based observations. Accurate monitoring of this species is necessary as CO is the key to a better understanding of the chemistry and transport of air pollutants. Here the long-term trend as well as short-term changes and the comparison with the model data are of interest.

Area D - Board 377

We present a global comparison of CO (carbon monoxide) measurements acquired by the TROPOMI and MOPITT instruments since November 2017 (close to the onset of the TROPOMI campaign).

CO is mostly produced by incomplete fuel combustion, biomass burning, and oxidation of methane and other hydrocarbons. CO oxidation results in greenhouse gases (carbon dioxide, ozone, and -indirectly- methane); hence, CO has a positive indirect radiative forcing. Because of its approximately 2-month lifetime, CO is not well mixed in the troposphere, which makes it a good tracer to monitor pollution plumes. Understanding CO is, thus, important in climate studies as well as in air quality monitoring. Satellite measurements are key to analyzing CO’s global distribution, transport, sources and sinks. MOPITT and TROPOMI are the only two satellite instruments currently retrieving CO from reflected infrared measurements (i.e., from CO’s 2.3 μm overtone band).

TROPOMI (the TROPOspheric Monitoring Instrument), onboard ESA’s Sentinel-5 Precursor platform, is a spectrometer measuring ultraviolet, visible, and reflected-infrared radiances. From the latter, total (partial) CO columns can be retrieved over land under clear (cloudy) conditions; additionally, partial CO columns can be retrieved over water if clouds are present. TROPOMI achieves close-to-global coverage daily with a spatial resolution at nadir near 7x7 km².

MOPITT (the Measurements of Pollution In The Troposphere instrument), onboard NASA's Terra satellite, has been measuring CO on the global scale since 2000. MOPITT’s dataset constitutes the longest global CO record to date. Global coverage is achieved every three days, approximately, with 22x22 km² spatial resolution at nadir. MOPITT derives tropospheric CO concentrations (total column and profiles) from cloud-free radiance measurements acquired in the reflected- and thermal-infrared regions. MOPITT's unique capability for multispectral observations provides enhanced sensitivity near the surface.

MOPITT retrievals have been systematically validated with respect to in situ measurements from aircraft at fixed stations (mainly over North America) and in numerous field campaigns over the last 18 years. Comparing MOPITT and TROPOMI is necessary to evaluate the continuity of the global CO time series and to investigate effects related to the two retrieval algorithms. Our results indicate that, on average, there is good agreement between the two datasets, with differences near the ~10% accuracy requirement.

Area D - Board 378

An Elastic/Raman LIDAR is installed in the Atmospheric Physics Laboratory of University of Rome “La Sapienza” (APL-Sapienza), at the last floor of the Department of Physics Fermi Building. The building is inside the University Campus, located downtown Rome and surrounded by civil, commercial and administration buildings. On the roof of the same building other instruments for atmospheric observation are operative, most of them included in the Boundary-layer Air Quality Using Network of INstruments (BAQUNIN) supersite equipment.

The LIDAR is equipped with a NdYAG laser source, emitting pulses at 30 Hz of repetition rate, at three wavelengths (1064, 532 and 355 nm), with a total power of about 50W. The receiver is based on 4 telescopes, in order to collect signals with intensity spanning over many orders of magnitude. The receivers of the signal at 532 nm are designed to study the depolarization produced by a-spherical scatterers.
This LIDAR can perform measurements of elastic backscattering from aerosols and clouds at the three laser wavelengths, and extinction at 386 nm, as well as observations of vertical profiles of water vapour mixing ratio from N2 and H2O Raman signals; it can operate in day and night conditions. The efficiency of the system in terms of vertical and time resolutions, 7.5 m and 10 sec respectively, permits to study the principal atmospheric processes connected to the dynamics of the Urban Boundary Layer, included the time evolution of the nocturnal and diurnal boundary layer thickness, the formation of convective plumes, the interaction between sea breeze and the UBL. The large vertical range of selected signals (up to 30 km of altitude for the 355 and 532 nm elastic signals) allows the observation of low and high level clouds, the study of desert aerosol transport and of high-troposphere/low-stratosphere interaction. The precision in the retrieval of altitude, boundaries and optical thickness of clouds and aerosol layers is crucial for satellites Cal/Val activities. Therefore, a full error propagation scheme has been developed and implemented in the retrieval chain.

System layout, performances, error estimates and retrieval algorithms are described in this work. Examples of results from observations obtained in coincidence of satellites and airplane overpasses are reported. Finally the actions planned for improving the overall instrumental performance and accuracy are discussed.

Area D - Board 379

The lidar system is widely used in atmospheric aerosol and boundary layer (BL) studies, and for the detection of cloud boundaries. However automatic and accurate identification of cloud top and bottom heights and BL height presents some difficulties, especially for low signal to noise ratio values, and when aerosol layers are observedon the top of BL. In addition, the disentanglement of cloud and aerosol contribution to Lidar signal is not trivial.
In this work, a signal threshold approach is presented, starting from the range corrected signal (RCS) and using its spatial and temporal variations. The approach has been tested using the BAQUNIN (Boundary-layer Air Quality using Network of INstruments) Lidar measurements.

A RCS threshold is selected to retrieve the presence of aerosol or cloud layer. This threshold depends on the characteristics of the specific Lidar characteristics, in particular the signal to noise ratio. Considering the RCS as a matrix, with altitude and time as dimensions, the elements of this matrix which have a value greater than the threshold are “labeled” as possible aerosol or cloud elements. In a second step, the spatial and temporal variations of the RCS are considered: if the neighbor elements of a selected “labeled” element are also “labeled”, then the selected element is confirmed to be a cloud/aerosol signal. This procedure excludes “single bin” cloud or aerosol elements in the RCS, possibly due to instrumental noise.
The number of elements to be considered for the 2-D analysis depends on the spatial and temporal resolution of the Lidar. In our case, these are 7.5 m and 10 s, therefore we selected a grid of 5x5 elements centered on the selected elements. This algorithm is applied to the complete set of Lidar measurement session, producing a matrix of “labeled” elements. The bottom and top heights of the layers are detected as the height of first and last labeled elements of each layer and profile. To distinguish between aerosol and cloud layers, we use the temporal variance of the RCS, calculated considering profiles inside the interval of 5 minutes centered on each profile. The possible values of the variance are known to be larger in the presence of clouds with respect to aerosol layers.Thus, when the estimated variance is larger than a selected value (also depending on Lidar signal to noise ratio), than a cloud is detected and its boundaries are retrieved.
The uncertainty of retrieved cloud parameters is obtained from a rigorous error propagation scheme, in order to provide detailed cloud information to the satellite Cal/Val communities.
This method was teste for different atmospheric conditions and results are discussed in this work.

Area D - Board 380