1 Introduction
Nitrogen oxides, and in particular the (NO and ), are important trace gases both in the troposphere and the stratosphere. In the troposphere they are produced mainly by the combustion of fossil and other organic fuels and by the production and use of nitrogen fertilizers for agriculture. They can also have a natural origin, e.g. lightning, biological processes in soils, and biomass burning. The ratio varies with solar illumination primarily, from 0.2–0.5 during the day down to zero at night. are converted to nitric acid and nitrates, which are removed by dry deposition and rain, resulting in a tropospheric lifetime of a few hours to days. Tropospheric are pollutants as well as proxies for other pollutants resulting from the (high-temperature) combustion of organic fuels. They are precursors for tropospheric ozone and aerosols and contribute to acid rain and smog. Because of their adverse health effects, local to national regulations limiting boundary layer concentrations are now in place in a long list of countries across the world. In the stratosphere, are formed by the photolysis of tropospheric nitrous oxide () produced by biogenic and anthropogenic processes and going up through the troposphere and stratosphere. Stratospheric control the abundance of ozone as a catalyst in ozone destruction processes but also by mitigating ozone losses caused by catalytic cycles involving anthropogenic halogens through the lock-up of these halogens in so-called long-lived reservoirs.
The global distribution, cycles, and trends of atmospheric have been measured from space by a large number of instruments on low Earth orbit (LEO) satellites. Since the late 1970s, its stratospheric and sometimes mesospheric abundance have been measured by limb-viewing and solar-occultation instruments working in the UV–visible and infrared spectral ranges: SME, LIMS, SAGE(-II), HALOE, and POAM-2/POAM-3, etc. and, in the last decade, OSIRIS, GOMOS, MIPAS, SCIAMACHY, Scisat ACE, and SAGE-III. Follow-on missions combining limb and occultation measurements are in development, like ALTIUS planned for the coming years. Pioneered in 1995 with ERS-2 GOME , which for the first time brought column measurements into space by Differential Optical Absorption Spectroscopy (DOAS; ), the global monitoring of tropospheric has continued uninterruptedly with a suite of UV–visible DOAS instruments with improving sensitivity and horizontal resolution: Envisat SCIAMACHY , EOS-Aura OMI , and the series of MetOp-A/B/C GOME-2 .
Owing to its cardinal role in air quality, tropospheric chemistry, and stratospheric ozone, and as a precursor of essential climate variables (ECVs), the monitoring of atmospheric on a global scale has been given proper attention in the European Earth Observation programme Copernicus. The Copernicus Space Component (CSC) is developing a constellation of atmospheric composition Sentinel satellites with complementary measurement capabilities, consisting of Sentinel-4 geostationary missions (with hourly monitoring over Europe) and Sentinel-5 LEO missions (with daily monitoring globally), to be launched from 2023 onwards. A measurement channel is also planned for the Copernicus Carbon Dioxide Monitoring mission CO2M for better attribution of the atmospheric emissions. The first element in orbit of this LEO+GEO constellation, the TROPOspheric Monitoring Instrument (TROPOMI), was launched on board of ESA's Sentinel-5 Precursor (S5P) early-afternoon LEO satellite in October 2017. This hyperspectral imaging spectrometer measures the Earth's radiance, at 0.2–0.4 nm resolution in the visible absorption band of , over ground pixels as small as or km (before and after the switch to smaller pixel size on 6 August 2019, respectively) and with an almost daily global coverage thanks to a swath width of 2600 km.
Pre-launch mission requirements for the Copernicus Sentinel data are, for the tropospheric column, a bias lower than 50 % and an uncertainty lower than 0.7 Pmolec cm, and for the stratospheric column, a bias lower than 10 % and an uncertainty lower than 0.5 Pmolec cm . Since the beginning of its nominal operation in April 2018, in-flight compliance of S5P TROPOMI with these mission requirements has been monitored routinely by means of comparisons to ground-based reference measurements in the Validation Data Analysis Facility (VDAF) of the S5P Mission Performance Centre (MPC) and by comparison with similar satellite data from OMI and GOME-2. The Copernicus S5P MPC routine operations validation service is complemented with ground-based validation studies carried out in the framework of ESA's S5P Validation Team (S5PVT) through research projects funded nationally like NIDFORVAL (see details in the Acknowledgements). Ground-based validation of satellite data
Figure 1
Geographical distribution of the UV–visible DOAS spectrometers contributing the ground-based correlative measurements: 26 NDACC ZSL-DOAS instruments in green, 19 MAX-DOAS instruments in blue, and 25 PGN instruments in red.
[Figure omitted. See PDF]
In this paper, we report on the consolidated results of the S5P ground-based validation activities for the first 2 years of nominal operation. The TROPOMI tropospheric, stratospheric, and total column data products under investigation, together with the corresponding ground-based reference data, are described in Sect. . This is followed by a brief assessment of the coherence between the data generated by the near-real-time (NRTI) and offline (OFFL) channels of the operational processors. For clarity, in separate sections we present results for the stratospheric (Sect. ), tropospheric (Sect. ), and total (Sect. ) columns. These three sections include a description of the preparation of the filtered, co-located, and harmonized data pairs to be compared and the comparison results. Robust, harmonized statistical estimators are derived from the comparisons consistently throughout the paper: the median difference as a proxy for the bias and half of the 68 % interpercentile (IP68/2) as a measure of the comparison spread (equivalent to a standard deviation for a normal distribution but much less sensitive to unavoidable outliers). Thereafter, in Sect. , these individual results are assembled and discussed all together, to derive conclusions on their mutual coherence, on the fitness for purpose of the S5P data, and on remaining challenges for the accurate validation of observations from space.
2 Data description2.1 S5P TROPOMI data
The retrieval of (sub)columns from TROPOMI Earth nadir radiance and solar irradiance spectra is a three-step process relying on DOAS and on a chemical transport model (CTM)-based stratosphere–troposphere separation. The TROPOMI algorithm is an adaptation of the QA4ECV community retrieval approach and of the DOMINO/TEMIS algorithm , already applied successfully to heritage and current satellite data records (GOME, SCIAMACHY, OMI, GOME-2). In the first step, the integrated amount of along the optical path, or slant column density (SCD), is derived using the classical DOAS approach . In the second step, the retrieved SCD is assimilated by the TM5-MP CTM to allocate a vertical profile of the concentration, needed for the separation between stratospheric and tropospheric SCDs. This assimilation procedure favours observations over pristine, remote areas where the entire SCD can be attributed to the stratospheric component. Assuming relatively slow changes in the stratospheric field, the model transports information to areas with a more significant tropospheric component. In the third step, the three slant (sub)column densities are converted into vertical (sub)column densities using appropriate air mass factors (AMFs). The CTM can be run either in forecast mode, using 1 d forecast meteorological data from the European Centre for Medium-Range Weather Forecasts (ECMWF), or in a more delayed processing mode, using 0–12 h forecast meteorological data. The former is used for near-real-time (NRTI) processing of the TROPOMI measurements, the latter for the offline (OFFL) production. For full technical details, the reader is referred to the Product Readme File (PRF), Product User Manual (PUM), and Algorithm Theoretical Basis Document (ATBD), all available at
Table 1
Identification of the S5P data versions validated here: near-real-time channel (NRTI), offline channel (OFFL), and interim reprocessing (RPRO). Major updates were those leading to v01.02.00 and to v01.03.00.
Processor | Start | Start | End | End |
---|---|---|---|---|
version | orbit | date | orbit | date |
NRTI | ||||
01.00.01 | 2955 | 9 May 2018 | 3364 | 7 June 2018 |
01.00.02 | 3745 | 4 July 2018 | 3946 | 18 July 2018 |
01.01.00 | 3947 | 18 July 2018 | 5333 | 24 July 2018 |
01.02.00 | 5336 | 24 October 2018 | 5929 | 5 December 2018 |
01.02.02 | 5931 | 5 December 2018 | 7517 | 27 March 2019 |
01.03.00 | 7519 | 27 March 2019 | 7999 | 30 March 2019 |
01.03.01 | 7999 | 30 March 2019 | 9158 | 20 July 2019 |
01.03.02 | 9159 | 20 July 2019 | current version | |
OFFL | ||||
01.02.00 | 5236 | 17 October 2018 | 5832 | 28 November 2018 |
01.02.02 | 5840 | 29 November 2018 | 7424 | 20 March 2019 |
01.03.00 | 7425 | 20 March 2019 | 7906 | 23 April 2019 |
01.03.01 | 7907 | 23 April 2019 | 8814 | 26 June 2019 |
01.03.02 | 8815 | 26 June 2019 | current version | |
RPRO | ||||
01.02.02 | 2836 | 1 May 2018 | 5235 | 17 October 2018 |
Besides very detailed quality flags, the S5P data product includes a combined quality assurance value (qa_value) enabling end users to easily filter data for their own purpose. For tropospheric applications (when not using the averaging kernels), the guideline is to use only data with a qa_value . This removes very cloudy scenes (cloud radiance fraction ), snow- or ice-covered scenes, and problematic retrievals. For stratospheric applications, where clouds are less of an issue, a more relaxed threshold of qa_value is recommended. These data filtering recommendations have been applied here, where the stricter requirement of qa_value has been used for the total column validation as well. Again, further details on this can be found in the PRF, PUM, and ATBD.
2.2 NDACC zenith-sky DOAS dataSince the pioneering ages of column measurements from space with ERS-2 GOME in the mid-1990s, ground-based UV–visible DOAS measurements at twilight have served as a reference for the validation of total column data over unpolluted stations and of stratospheric column data from all nadir UV–visible satellites to date
Table 2
Estimated uncertainties for the different types of ground-based measurements used in this work. Ex ante refers to uncertainties provided with the data, based on a propagation of raw measurement uncertainties and on sensitivity analyses. Ex post refers to uncertainty estimates derived by comparison with other (independent) measurements, which inevitably also contain some representativeness uncertainties. More detail is provided in the dedicated subsections of Sect. .
Instrument | Ex ante | Ex post | Selected |
---|---|---|---|
uncertainty | uncertainty | references | |
ZSL-DOAS | 10 %–14 % | NA | , |
MAX-DOAS | 7 %–17 % | 30 % | , |
PGN | 2.7 | 20 % | , |
NA: not available.
2.3 MAX-DOAS dataSatellite tropospheric column data are compared classically to correlative measurements acquired by Multi-Axis Differential Optical Absorption Spectroscopy (MAX-DOAS) instruments . From sunrise to sunset, MAX-DOAS instruments measure the UV–visible radiance scattered in several directions and elevation angles, from which the tropospheric vertical column density (VCD) and/or the lowest part of the tropospheric profile (usually up to 3 km altitude, and up to 10 km at best) can be retrieved through different techniques
-
Different uncertainty reporting strategy. The reported systematic uncertainty may include only that from the cross sections (approx. 3 %; UNAM, BIRA-IASB, MPIC, AUTH, IUPB), or it may include also a contribution from the VCD retrieval step (up to 14 % in JAMSTEC data and 20 % in KNMI data) and the aerosol retrieval
Chiba U; .- -
Different SCD retrieval. Recommended common DOAS settings are used by all groups in the present study, and when doing so, instrument intercomparison campaigns like CINDI-1 and CINDI–2 (Roscoe et al., 2010; Kreher et al., 2020) revealed relative biases between 3 % and 10 % in the differential slant column density (DSCD).
- -
Different methods to retrieve VCD from DSCD (see also Table ). Using either (1) vertical profile inversion using optimal estimation (BIRA-IASB, UNAM); (2) profile inversion using (an optimal estimation of) parameterized profile shapes (JAMSTEC and Chiba U); (3) direct retrieval via the calculation of a tropospheric AMF (QA4ECV datasets); or (4) direct retrieval using a geometrical approximation can lead to systematic differences in the 5 %–15 % range .
MAX-DOAS data have been used extensively for tropospheric satellite validation, for instance for Aura OMI and MetOp GOME-2
Data are collected either through ESA's Atmospheric Validation Data Centre (EVDC;
The Pandonia Global Network (PGN) delivers direct Sun total column and multi-axis tropospheric column observations of several trace gases, including , from a network of ground-based standardized Pandora Sun photometers in an automated way. In this work, only direct Sun observations are used. These have a random error uncertainty of about 0.27 and a systematic error uncertainty of 2.7 . Studies at US and Korean sites during the DISCOVER-AQ campaign found a good agreement of Pandora instruments with aircraft in situ measurements (within 20 % on average; ), although larger differences are observed for individual sites .
Pandora data have been used before to validate satellite measurements from Aura OMI and TROPOMI .
For the current work, 25 sites have contributed Pandora data, collected either from the ESA Atmospheric Validation Data Centre (EVDC) (
Except at low Sun elevation, the footprint of these direct Sun measurements is much smaller than a TROPOMI pixel. Therefore, as is the case with MAX-DOAS, a significant horizontal smoothing difference error can be expected in the TROPOMI–Pandora comparison, especially in the case of tropospheric gradients and when tropospheric is the largest contributor to the total column.
Three Pandora instruments (Altzomoni, Izaña, Mauna Loa) are located near the summit of a volcanic peak and are therefore not sensitive to the lower lying tropospheric . In this work, their observations are compared to the TROPOMI stratospheric data (see Sect. ).
Table 3
cross section source and temperature for the different data processing used in this work. More detail is provided in Sect. .
Instrument | Reference | Temperature | Comments |
---|---|---|---|
S5P TROPOMI | 220 | With temperature correction in AMF | |
ZSL-DOAS | 220 | ||
ZSL-DOAS | 227 | NIWA instruments | |
MAX-DOAS | 298 | tropospheric retrieval only | |
MAX-DOAS | 298 and 220 | Orthogonalized following | |
PGN | 254.4 | PGN processor v1.7 |
cross section data
A potential source of inconsistencies between the different data products lies in the cross sections that are used. An overview of the different choices made is provided in Table . Most products use the cross sections published by , but there are differences in the choice of temperature at which to take the cross sections. The ZSL-DOAS measurements are processed with cross sections at a fixed 220 or 227 K, i.e. typical stratospheric temperatures. MAX-DOAS data are processed either with cross sections at room temperature (298 K, representing a typical tropospheric temperature) or using an orthogonalized set of cross sections at 298 and 220 when both tropospheric and stratospheric slant columns are retrieved. As the scientific focus of the PGN up until processor version 1.7 (used for this study) was on measuring polluted conditions, i.e. in the presence of moderate to large tropospheric columns, the cross sections used in the processor are scaled to a fixed effective temperature of 254.4 , which corresponds to the situation of approximately equal column amounts in the troposphere and stratosphere. The S5P retrievals use cross sections at 220 but with an explicit correction for the temperature dependence of the cross sections in the AMF: space–time co-located daily ECMWF temperature profile forecasts are used to compute a height-dependent AMF correction factor. The temperature sensitivity parameterized in this correction is approximately 0.32 % K . A posteriori temperature correction of the ground-based data is beyond the scope of this paper, so it must be kept in mind that this may contribute to differences between S5P and ground-based columns. Specifically, we could expect a small seasonal cycle in the stratospheric column comparisons of a few percent due to the seasonal variation in stratospheric temperature not being accounted for in the ZSL-DOAS data processing. PGN columns may either be overestimated by up to 10 % when the column is mostly stratospheric or underestimated by a similar order of magnitude when large tropospheric amounts are present. The MAX-DOAS data may be biased in either direction by a few percent when tropospheric and/or stratospheric temperatures differ strongly from the 298 and 220 default temperatures.
3 Mutual coherence between TROPOMI NRTI and OFFLAs described in Sect. , the main difference between the NRTI and OFFL data processors lies in the use of either 1 d or 0–12 h forecast ECMWF meteorological data as input, which impacts the TM5-MP vertical profiles. The mutual consistency between the NRTI and OFFL data products is monitored routinely using data and tools provided by the S5P MPC Level-2 Quality Control Portal (
Figure 2
(a) Time series of the global means of total column data retrieved with the NRTI (red line) and OFFL (blue line) processors, and their standard deviation, in Pmolec cm, from July 2018 till February 2020. Crosses depict the number of measurements divided by , with the same colour code: red for NRTI, blue for OFFL. Yellow vertical lines indicate the transition dates for processor upgrades and the switch to the smaller ground pixel size. (b) Percent relative difference between NRTI and OFFL global means of total values. The Theil–Sen linear regression line (black) is superimposed.
[Figure omitted. See PDF]
To further assess similarities and differences between the NRTI and OFFL processing channels, values along individual orbits are also compared directly. An illustration is given in Fig. for S5P orbit no. 07407, a randomly selected orbit crossing western Europe on a relatively cloud-free day (19 March 2019). Data were filtered to include only those pixels with a qa_value larger than 0.5 and were gridded to before calculating the differences.
Figure 3
Maps of the difference between the NRTI and OFFL data values for S5P orbit no. 07407 on 19 March 2019. Difference between (a) total column values and (b) stratospheric column values. (c) Close-up of the difference in tropospheric column values over western Europe.
[Figure omitted. See PDF]
The three maps of Fig. show the difference between NRTI and OFFL values for the total, stratospheric, and tropospheric column, respectively, together with the corresponding Pearson correlation coefficient and root-mean-square deviation (RMSD). While the correlation coefficient is high (typically around 0.98), the maps do reveal regions where significant deviations occur, up to Pmolec cm between the NRTI and OFFL stratospheric columns and up to Pmolec cm for both the tropospheric columns and the total columns. North-east of Iceland, NRTI-OFFL differences in stratospheric and in tropospheric columns are of opposite sign, while total column differences are minimal, indicating a different stratosphere–troposphere separation after the slant column retrieval leading. West of Norway, total columns differ significantly between NRTI and OFFL, and these differences are allocated mostly to the tropospheric columns. These features are specific to this particular orbit and not systematic. A more detailed investigation targeted solely at regions and times of significant deviations between NRTI and OFFL would be needed to better reveal the full benefit of the OFFL analysis, but that is beyond the scope of the current paper. What needs to be underlined is that the ground-based validation studies on which the present consolidated results are based upon do not yield significantly different conclusions for the two processing modes. Therefore, all results reported in this paper may be considered as applicable to the two processing channels.
Figure 4
(a) Time series of S5P NRTI stratospheric column data co-located with ground-based SAOZ sunset measurements performed by CNRS/LATMOS at the NDACC mid-latitude station of Observatoire de Haute-Provence (France). The latter were adjusted for the photochemical difference between the S5P and twilight solar local times, while S5P data were averaged over the ground-based twilight air mass. Solid lines represent 2-month running medians. Scatter plot (b) and histogram of the differences (c) with several statistical measures of the agreement between data.
[Figure omitted. See PDF]
4 Stratospheric column validation4.1 Co-location and harmonization
To reduce mismatch errors due to the significant difference in horizontal sensitivity between S5P and ZSL-DOAS measurements, individual TROPOMI stratospheric column data (in ground pixels at high horizontal sampling) are averaged over the much larger footprint of the air mass to which the ground-based zenith-sky measurement is sensitive; see , , and for details. The length of this footprint if of the order of 300–600 in the direction of the Sun, and the width is typically of the order of 50–100 at mid-latitudes, depending on the duration of sunrise and sunset. Note that, as the TROPOMI stratospheric column is a TM5 output, its true resolution is actually much lower than the pixel size. To account for effects of the photochemical diurnal cycle of stratospheric , the ZSL-DOAS measurements at sunset are adjusted to the early-afternoon S5P overpass time using a model-based correction factor. The latter is calculated with the PSCBOX 1D stacked-box photochemical model , initiated by daily fields from the SLIMCAT chemical transport model (CTM). The amplitude of the adjustment factor is sensitive to the effective solar zenith angle (SZA) assigned to the ZSL-DOAS measurements. It is assumed here to be 89.5 or, during polar day and close to polar night, the largest or smallest SZA reached, respectively. This photochemical correction factor is an average based on 10 years of the box-model simulations, and the range of values over these 10 years can be considered an uncertainty estimate. It varies between 1 % and 6 % at the sites considered here, the uncertainty being largest at high latitudes in local winter. This does however not contain any model uncertainty (in the sense of the accuracy of the model in representing the true photochemical variation during the day). Another way to estimate the uncertainty in the adjusted ZSL-DOAS data is by comparing the agreement between sunrise and sunset measurements when both are photochemically adjusted to the S5P overpass time. This does also contain co-location mismatch uncertainty due to transport of air occurring during the period between sunrise and sunset and due to the different air masses that are probed (east or west of the instrument respectively). Moreover, it also contains that part of the measurement uncertainty that is not systematic on a daily (or longer) timescale. We find that sunrise and sunset measurements typically agree within 6 % (standard deviation of the differences). Overall, the 10 %–14 % total uncertainty estimate already presented in Sect. thus seems realistic.
4.2 Comparison results
Figure illustrates the comparison between TROPOMI and ground-based ZSL-DOAS SAOZ data at the NDACC station at Observatoire de Haute-Provence (OHP) in southern France. The time series reveal a small negative median difference for TROPOMI, which is found to be a common feature across the network, but little seasonal structure. The correlation coefficient is excellent, and the histogram of the differences has an almost Gaussian shape.
Figure 5
Difference between the S5P TROPOMI and NDACC ZSL-DOAS stratospheric column data as a function of time, after photochemical adjustment of the ZSL-DOAS sunset data to the S5P SZA. Stations are ordered by increasing latitude (south at the bottom). The dashed vertical line on 6 August 2019 represents the reduction in S5P ground pixel size from to km.
[Figure omitted. See PDF]
Comparison results for the entire ZSL-DOAS network are presented in Fig. . This figure reveals occasionally larger differences in more difficult co-location conditions (e.g. enhanced variability at the border of the polar vortex) but no impact of the TROPOMI pixel size change on 6 August 2019. The latter result must be interpreted with care as, for these comparisons, multiple TROPOMI pixels are averaged over the ZSL-DOAS observation operator before comparison (see Sect. ), and as such any change in the noise statistics of individual pixels will be hidden.
Statistical estimators of the bias (median difference) and scatter per station are presented in box-and-whisker plots in Fig. and in tabular form in Sect. . Across the network, S5P NRTI and OFFL stratospheric column data are generally lower than the ground-based values by approximately 0.2 Pmolec cm, with a station–station scatter of this median difference of similar magnitude (0.3 Pmolec cm). These numbers are within the mission requirement of a maximum bias of 10 % (equivalent to 0.2–0.4 Pmolec cm, depending on latitude and season) and within the combined systemic uncertainty of the reference data and their model-based photochemical adjustment. The IP68/2 dispersion of the difference between TROPOMI stratospheric column and correlative data around their median value rarely exceeds 0.3 Pmolec cm at sites without tropospheric pollution. When combining random errors in the satellite and reference measurements with irreducible co-location mismatch effects, it can be concluded that the random uncertainty on the S5P stratospheric column measurements falls within the mission requirements of max. 0.5 Pmolec cm uncertainty.
Figure 6
Box-and-whisker plots summarizing from pole to pole the bias and spread of the difference between S5P TROPOMI NRTI and NDACC ZSL-DOAS stratospheric columns (SAOZ data in black, other ZSL-DOAS in blue, and PGN in red). The median difference is represented by a vertical solid line inside the box, which marks the 25 % and 75 % quantiles. The whiskers cover the 9 %–91 % range of the differences. The shaded area represents the mission requirement of 0.5 Pmolec cm for the uncertainty. Values between brackets in the labels denote the latitude of the station.
[Figure omitted. See PDF]
The potential dependence of the TROPOMI stratospheric column bias and uncertainty on several influence quantities has been evaluated. Figure shows results for the solar zenith angle (SZA), the fractional cloud cover (CF), and the surface albedo of the TROPOMI measurement. This evaluation does not reveal any variation of the bias much larger than 0.4 Pmolec cm over the range of these influence quantities.
Figure 7
Dependence of the difference between TROPOMI NRTI and ground-based ZSL-DOAS stratospheric column data on the satellite solar zenith angle (SZA), satellite cloud fraction, and satellite surface albedo, including a median and IP68/2 spread per bin (bin widths of 10 in SZA, 0.05 in CF, and 0.1 in surface albedo). Different colours represent different stations, to illustrate the (modest) impact of station–station network inhomogeneity on these analyses.
[Figure omitted. See PDF]
4.3 PGN measurements at high-altitude stationsThree of the PGN direct Sun instruments (see Sect. ) are located near the summit of a volcanic peak: Altzomoni (3985 m a.m.s.l.) in the State of Mexico, Izaña (2360 m a.m.s.l.) on Mount Teide on the island of Tenerife, and Mauna Loa (4169 m a.m.s.l.) on the island of Hawaii. At these high-altitude sites, the total column measured by the ground-based direct Sun instrument misses most of the tropospheric (potentially polluted) part and as such becomes representative of the TROPOMI stratospheric column. These sites have therefore been added to Fig. , illustrating that these comparisons based on direct Sun data yield similar conclusions as those based on zenith-sky data, that is, a minor negative median difference of the order of Pmolec cm. It must be noted that, as discussed in Sect. , the PGN data are processed using cross sections at a temperature of 254.4 , representative of a total column made of equal amounts of in the stratosphere and troposphere. This leads to columns which are about 10 % larger than if they had been processed with cross sections for 220 K. Future processing of the PGN data will address this, and it is expected that this will mostly remove the apparent negative bias for TROPOMI (but lead to a slight inconsistency with the ZSL-DOAS results).
5 Tropospheric column validation
5.1 Co-location and harmonization
TROPOMI data are filtered following the qa_value rule as recommended in the associated PRF (see Sect. ). Then for each day, the pixel over the site is selected. MAX-DOAS data series are temporally interpolated at the TROPOMI overpass time (only if data within 1h exist), and daily comparisons are performed. This short temporal window avoids the need for a photochemical cycle adjustment. Details on the comparison approach are described in for the validation of OMI and GOME-2 column data and in for the validation of the OMI QA4ECV Climate Data Record.
Figure 8
Same as Fig. but now for the S5P OFFL tropospheric column data co-located with ground-based MAX-DOAS measurements performed by BIRA-IASB at the NDACC mid-latitude station of Uccle in Brussels (Belgium).
[Figure omitted. See PDF]
5.2 Comparison resultsAn illustration of the daily comparisons between TROPOMI and ground-based MAX-DOAS measurements between May 2018 and the end of January 2020 is presented in Fig. for the Uccle station (Brussels, B, with moderate pollution levels). The two datasets have a correlation coefficient of 0.75 and a regression slope and intercept of 0.47 and 1.0 Pmolec cm respectively. The (median and mean) difference of about to Pmolec cm corresponds to a median relative difference of about %.
Figure 9
Percent relative difference between the S5P TROPOMI and MAX-DOAS tropospheric column data as a function of time. Stations are ordered by median tropospheric column (lowest median value at the bottom). The dashed vertical line on 6 August 2019 represents the reduction in S5P ground pixel size from to km.
[Figure omitted. See PDF]
Results for the entire MAX-DOAS network are presented in Fig. . This figure reveals mostly (but not only) negative differences, with a fairly significant variability but no clear seasonal features. No impact of the TROPOMI ground pixel size change on 6 August 2019 is observed.
Figure 10
Same as Fig. but now for the difference between S5P TROPOMI OFFL and MAX-DOAS tropospheric columns and ordered as a function of the median ground-based tropospheric column (largest median VCD values on top). The line represents the median difference. Box bounds represent the 25 and 75 percentiles, while whiskers indicate the 9 and 91 percentiles. The shaded area corresponds to the mission requirement of a maximum bias of 50 %.
[Figure omitted. See PDF]
Box-and-whisker plots for the whole network are shown in Fig. , with corresponding numeric values listed in Sect. . Based on measurements from these 19 MAX-DOAS stations, three different regimes can be identified:
- i.
Small tropospheric column values (median values below 2 Pmolec cm), e.g. at the Fukue and Phimai stations, lead to small differences. Typically, these stations show a small median bias ( Pmolec cm), but this can still correspond to up to a % relative bias. The dispersion (IP68/2) of the difference is smaller than 1 Pmolec cm.
- ii.
More polluted sites (median tropospheric columns from 3 to 14 Pmolec cm) experience a clear negative bias. The median difference ranges between and Pmolec cm, i.e. between % (Chiba) and % (Pantnagar). This underestimation is similar to the one identified in the validation of Aura OMI and MetOp GOME-2 tropospheric data by and . The dispersion (IP68/2) of the difference ranges from to Pmolec cm, roughly increasing with increasing tropospheric median VCD.
- iii.
Extremely polluted sites report larger differences. This is the case, for example, at the Mexican UNAM sites (UNAM and Vallejo in/close to Mexico City and Cuautitlan in a more remote part of the State of Mexico), with median tropospheric columns larger than 15 Pmolec cm. These stations experience larger differences ( Pmolec cm, i.e. from % to %). The dispersion (IP68/2) of the difference is also quite large, between 4 and Pmolec cm. Results at these sites need deeper analysis.
The overall bias (median of all station median differences) is Pmolec cm, i.e. %. The median dispersion is 3.5 Pmolec cm, while the site–site dispersion (IP68/2 over all site medians) is 2.8 Pmolec cm. Note that these network-averaged numbers are close to the numbers found for the polluted (Athens to Gucheng) sites. These results are within the mission requirement of a maximum bias of 50 %, but they exceed the uncertainty requirement of at most 0.7 Pmolec cm, which is only satisfied for the clean sites' ensemble. A discussion on the causes of these biases and sometimes large comparisons' spread is provided in Sect. .
Two key influence quantities for observations of tropospheric are aerosol optical depth (AOD) and cloud (radiance) fraction (CRF). The dependence of the differences between MAX-DOAS and TROPOMI tropospheric columns on these two influence quantities is visualized in Fig. . AOD is only retrieved in the processing of a handful of MAX-DOAS instruments, the others using climatological information, hence the limited subset in stations in panel (a) of this figure. No clear dependence of the bias on either property is seen, though in view of the relatively large scatter in these tropospheric column comparisons, this does not preclude more subtle dependencies. The impact of aerosol peak height would also be interesting to assess, but this is impossible to judge within the scope of the current paper as no such information is readily available.
Figure 11
Dependence of the difference between TROPOMI OFFL and ground-based MAX-DOAS tropospheric column data on (a) the MAX-DOAS-retrieved aerosol optical depth (AOD; only available for a subset of the instruments) and (b) the satellite cloud radiance fraction (CRF).
[Figure omitted. See PDF]
6 Total column validation6.1 Filtering, co-location, and harmonization
As was done for the tropospheric column validation in Sect. , only S5P pixels with a qa_value of at least 0.75 are retained. The so-called summed product is used, i.e. the total column computed as the stratospheric plus the tropospheric column values. This summed column differs from the total column product. Only Pandonia measurements with the highest quality label (0 and 10) are used. The average column value within a 1 h time interval, centred on the S5P overpass time, is used. As the ratio varies only slowly around the afternoon solar local time of the TROPOMI overpass, this small temporal window ensures no model-based adjustment is required. A 30 min time interval was tested as well, but this did not change the results significantly. Moreover, only TROPOMI pixels containing the station were considered.
6.2 Comparison results
An example of a time series of co-located TROPOMI and PGN total column measurements, and their difference, is shown in Fig. .
Figure 12
Same as Figs. and but now for the S5P OFFL total column data co-located with ground-based Pandora measurements obtained at the PGN mid-latitude station of Boulder, Colorado.
[Figure omitted. See PDF]
Figure 13
Percent relative difference between the S5P TROPOMI and PGN total column data as a function of time. Stations are ordered by median total column (lowest median value at the bottom). The dashed vertical line on 6 August 2019 represents the reduction in S5P ground pixel size from to km. The three mountaintop sites more suited for the validation of only the stratospheric column are marked with an asterisk.
[Figure omitted. See PDF]
Results for the entire PGN network are presented in Fig. . This figure reveals that the difference, even in relative units, depends strongly on the total column, with low (or slightly positive) biases at low columns and markedly negative biases at high columns. No impact is observed for the TROPOMI ground pixel size switch of 6 August 2019.
Figure 14
Same as Figs. and but now for the difference between S5P TROPOMI (RPROOFFL) and PGN total columns. Stations are ordered by ground-based total median value, like in Fig. . The median difference is represented by a vertical solid line inside the box, which marks the 25 % and 75 % quantiles. The whiskers cover the 9 %–91 % range of the differences. The three mountaintop PGN instruments used for the validation of the stratospheric columns are not included here but in Fig. .
[Figure omitted. See PDF]
Statistical estimators of the comparison results across the network are visualized in Fig. and presented in tabular form in Table . One can distinguish roughly two different regimes.
- i.
The PGN median total column value is between 3 (Alice Springs) and 6 Pmolec cm (New Brunswick). The absolute bias (median difference) is within Pmolec cm in most cases (up to Pmolec cm at Egbert and Helsinki), while the median relative difference is within 5 % in most cases (up to % at Alice Springs, Egbert, Inoe, and Helsinki). Canberra is a deviating case, with larger negative bias ( Pmolec cm; %). The difference dispersion (IP68/2) roughly increases with increasing PGN median VCD, from 0.4–0.6 Pmolec cm at the three cleanest sites to 1–2 Pmolec cm at the other sites.
- ii.
The PGN median total column value is between 8 (Buenos Aires) and 19 Pmolec cm (UNAM, Mexico City). A negative bias is observed, ranging from Pmolec cm ( %) at the Bronx (New York) to Pmolec cm ( %) at Rome Sapienza. The difference dispersion ranges from (Buenos Aires) to 5 Pmolec cm (UNAM).
The median relative difference is mostly within (or bordering) the range for the sites with lower median total column values (Alice Springs to New Brunswick; Canberra is an exception), while it is negative and mostly outside this range, but still within , for the sites with higher median total column value (Buenos Aires to UNAM).
The overall bias over all sites (median over all site medians or site relative medians) is Pmolec cm ( %). The overall dispersion is 1.8 Pmolec cm, while the site–site dispersion (IP68/2 over all site medians) is 2.2 Pmolec cm.
It is however more useful to make the distinction between sites with low (Alice Springs to New Brunswick) and high (Buenos Aires to UNAM). For the low sites, the overall bias is 0.1 Pmolec cm (2 %), the overall dispersion is 1.1 Pmolec cm, and the site–site dispersion is 0.2 Pmolec cm. For the high sites, the overall bias is Pmolec cm ( %), the overall dispersion is 3.3 Pmolec cm, and the site–site dispersion is 1.4 Pmolec cm.
7 Discussion and conclusionsA cross-network summary of the median difference and dispersion for the three S5P (sub)column data is attempted in Table . While the difference between the NRTI and OFFL values can reach up to a few Pmolec cm for individual TROPOMI pixels, the two processing channels do not lead to significantly different validation results, and Table therefore makes no distinction between the two.
Table 4
Cross-network summary of the validation results: bias (median) and dispersion (IP68/2) of the difference with respect to the ground-based correlative measurements (median value over the stations).
Bias | Dispersion | |
---|---|---|
Stratosphere | Pmolec cm; % | 0.3 Pmolec cm |
Troposphere | ||
– low | Pmolec cm; % | 0.7 Pmolec cm |
– high | Pmolec cm; % | 3.4 Pmolec cm |
– extreme | Pmolec cm; % | 7 Pmolec cm |
Total column | ||
– low | 0.1 Pmolec cm; 2 % | 1 Pmolec cm |
– high | Pmolec cm; % | 3 Pmolec cm |
For the stratospheric column, the general picture is a slight negative median difference of TROPOMI with respect to the NDACC ZSL-DOAS network, of the order of -0.2 Pmolec cm on average, with some station–station inhomogeneities and with larger differences in the highly variable conditions of the denoxified polar stratosphere in local winter. This median difference remains within the S5P mission requirements and is similar to the conclusions derived for similar satellite data from other sounders
For the tropospheric and total columns, averaging results over the networks with the hope of obtaining a meaningful global estimate is of limited use as the results depend strongly on the amount of tropospheric . Overall, mission requirements in terms of bias are mostly met, the only exception being the tropospheric columns at extremely polluted sites, which have a bias on the threshold of 50 %. Nevertheless, it is clear that large negative median differences are observed across all sites experiencing significant tropospheric pollution. The dispersion of the difference is well outside of the mission requirements formulated for the tropospheric column data. Nevertheless, these results are consistent with those obtained with completely different validation techniques, such as those explored by over Paris (using ground-based and Eiffel Tower concentrations and a climatology of observed column–surface ratios). Many factors play a role in this apparent disagreement between TROPOMI and the ground-based networks, that can neither be attributed solely to the S5P data, nor to pure area-averaging differences.
First, local horizontal and vertical variations of the field can explain (part of) such discrepancies, as illustrated in , , , and . While the MAX-DOAS picks up small local enhancements, the much larger satellite pixel provides a smoothed perception of the field. In particular for sounders with footprints (much) larger than the emission sources, this generally leads to underestimation in urban conditions while having better agreement in remote locations . showed specific improvements of the S5P comparison results in the case of the Uccle MAX-DOAS when making use of the multiple azimuthal scan mode and when improving the S5P selection criteria to pixels along the MAX-DOAS field-of-view direction and within the effective sensitivity length. Large inhomogeneities around MAX-DOAS sites were also shown by , , , , and . When taking some of these inhomogeneities into account in validation of other sounders, results have been improved . also showed the smoothing of the field when resampling GeoTASO high-resolution airborne measurements to different simulated satellite pixel sizes.
Second, vertical sensitivity (and thus averaging kernels) and a priori vertical profiles are known to be different for MAX-DOAS and nadir UV–visible satellite retrievals , with MAX-DOAS measurements sensitive to layers close to the surface and satellite retrievals sensitive mostly to the free troposphere. The effect of the a priori vertical profile on the comparison was estimated for TROPOMI by for Uccle, showing an increase by about 55 % when recalculating the TROPOMI column with MAX-DOAS daily mean tropospheric profile. Similarly, and show improvement of the agreement between TROPOMI and Pandora total column data for episodes of enhancement, when replacing the coarse a priori profiles with high-resolution profiles from a high-resolution regional air quality forecast model. Somewhat related to the vertical sensitivity is the treatment of aerosol optical depth and its vertical profile. Poor representation of the aerosol opacity has been shown (from simulations) to cause both underestimated in satellite retrievals and overestimated in MAX-DOAS measurements . Satellite-ground discrepancies in previous validation studies have already been attributed to such aerosol issues . Moreover, explicit aerosol corrections in the S5P retrievals have already been shown to improve the agreement .
Third, the treatment of cloud properties can have a significant effect on the retrieval of the TROPOMI tropospheric VCD. discuss the comparison with OMI tropospheric column retrievals and show that on average TROPOMI is lower than OMI by % to % over Europe, North America, and India and up to % over China. This difference is mainly attributed to the different cloud data product used in the retrieval: FRESCO-S derives the cloud top pressure from TROPOMI radiances in the near-infrared band, while for OMI the cloud top pressure is retrieved from the band in the UV–visible. Preliminary validation results (, and Henk Eskes, private communication, 2020) indicate that FRESCO-S is biased high in pressure, especially at altitudes close to the surface. A new version of FRESCO-S with an adapted wavelength window has been implemented and seems to remove most of the 10 %–22 % bias with OMI in polluted regions.
Fourth, although this work, , and all show a generally good coherence of the validation results among the MAX-DOAS instruments across the network and also among MAX-DOAS and Pandora instruments, network homogenization remains an important challenge to focus on to improve the accuracy of future satellite validations (see Sect. for a description of contributors to network inhomogeneity). Intercomparison campaigns, such as the CINDI-1 and CINDI-2 , in-depth intercomparison studies of the retrieval methods , and dedicated projects aiming at the harmonization of the processing and of the associated metadata (such as the FRM4DOAS project of ESA's Fiducial Reference Measurements programme) are an important way to achieve this.
Regarding the mutual consistency of MAX-DOAS- and PGN-based validation results, while it may appear that, at low column values, PGN-based comparisons indicate a smaller bias than the MAX-DOAS comparisons, one must not forget that PGN measures the total column: at stations with a lower total column value, the stratospheric contribution is relatively more important. The better agreement here is therefore consistent with the good agreement found for the TROPOMI stratospheric column vs. ZSL-DOAS and also vs. PGN at pristine mountain sites (Sect. ). For sites characterized by a higher total column, the tropospheric contribution becomes more important, and some of the same effects that make satellite–MAX-DOAS comparisons difficult, such as the smoothing difference error, the lower sensitivity of the satellite close to the surface, and the approximate S5P a priori profile, come into play as well.
In conclusion, the first 2 years of Copernicus S5P TROPOMI column data produced both with the NRTI and OFFL versions 01.0x.xx of the operational processors do meet mission requirements for the bias and, to some extent, with precaution for the uncertainty (dispersion). The different data products available publicly through the Copernicus system are mutually consistent, are in good geophysical and quantitative agreement with ground-based correlative data of documented quality, and can be used for a variety of applications, on the condition that the features and limitations exposed here are taken into proper consideration and that the S5P data are filtered and used according to the recommendations provided in the official Product Readme File (PRF) and associated documentation, also available publicly. Ground-based validation activities relying on the correlative measurements contributed by the NDACC ZSL-DOAS, MAX-DOAS, and PGN global monitoring networks have progressed significantly in recent years and have demonstrated their capacity but also their current limitations in an operational context such as the Copernicus programme. Room does exist for further improvement of both the satellite and ground-based datasets, as well as the intercomparison methodology and its associated error budget. Beyond the methodology advances published here and in aforementioned papers, special effort is needed to understand fully and ever reduce comparison mismatch errors, which so far make the accurate validation of S5P data uncertainty bars difficult. Several updates of the calibration of TROPOMI spectra and of the TROPOMI data retrieval processors are already in development and in implementation. Upcoming data versions should be validated with the same system as used in the current paper, allowing the necessary independent assessment of the S5P data product evolution.
Appendix A Ground networksA1 The NDACC ZSL-DOAS network
Table A1
ZSL-DOAS hosting stations, ordered by latitude, that contribute to the stratospheric column validation. Several measures of the agreement between TROPOMI and the ground-based data are also provided. The bias over all stations (median over all station median differences) is , while the overall dispersion (median over all 1/2IP68) is 0.31 , and the inter-station dispersion (1/2IP68 over all station medians) is 0.30 .
Station | Lat | Long | Altitude | Institute | Processing | Median diff. | Spread | |
---|---|---|---|---|---|---|---|---|
(IP68/2) | ||||||||
() | (m) | (Pmolec cm) | ||||||
a.m.s.l. | ||||||||
Eureka | 80.05 | 610 | U. Toronto | NDACC | 0.04 = 1 % | 0.60 | 0.89 | |
Eureka | 80.05 | 610 | LATMOS-CNRS + U. Toronto | LATMOS_RT | % | 0.20 | 0.97 | |
Ny-Ålesund | 78.92 | 11.93 | 10 | NILU | LATMOS_RT | % | 0.24 | 0.97 |
Scoresbysund | 70.48 | 67 | LATMOS-CNRS + DMI | LATMOS_RT | % | 0.32 | 0.98 | |
Sodankylä | 67.37 | 26.63 | 179 | LATMOS-CNRS + FMI | LATMOS_RT | % | 0.37 | 0.97 |
Harestua | 60.00 | 10.75 | 596 | BIRA-IASB | NDACC | % | 0.36 | 0.95 |
Zvenigorod | 55.69 | 36.77 | 220 | IAP, RAS | NDACC | % | 0.67 | 0.69 |
Bremen | 53.10 | 8.85 | 27 | IUP Bremen | NDACC | % | 0.40 | 0.91 |
Paris | 48.85 | 2.35 | 63 | LATMOS-CNRS | LATMOS_RT | % | 0.56 | 0.59 |
Guyancourt | 48.78 | 2.03 | 160 | LATMOS-CNRS | LATMOS_RT | % | 0.45 | 0.71 |
Haute-Provence (OHP) | 43.94 | 5.71 | 650 | LATMOS-CNRS | LATMOS_RT | % | 0.23 | 0.94 |
Issyk-Kul | 42.62 | 76.99 | 1640 | KNU | NDACC | % | 0.19 | 0.48 |
Athens | 38.05 | 23.86 | 527 | IUP Bremen + NOA | NDACC | % | 0.28 | 0.89 |
Izaña | 28.31 | 2367 | INTA | NDACC | % | 0.14 | 0.95 | |
Saint-Denis | 55.48 | 110 | LATMOS-CNRS + LACy | LATMOS_RT | 0.05 = 2 % | 0.18 | 0.80 | |
Bauru | 640 | LATMOS-CNRS + UNESP | LATMOS_RT | % | 0.19 | 0.80 | ||
Lauder | 169.68 | 370 | NIWA | NDACC | % | 0.28 | 0.92 | |
Kerguelen | 70.26 | 36 | LATMOS-CNRS | LATMOS_RT | % | 0.34 | 0.94 | |
Rio Gallegos | 15 | LATMOS-CNRS | LATMOS_RT | % | 0.28 | 0.95 | ||
Macquarie | 158.94 | 6 | NIWA | NDACC | % | 0.48 | 0.93 | |
Ushuaïa | 7 | INTA | NDACC | 0.09 = 4 % | 0.40 | 0.95 | ||
Marambio | 198 | INTA | NDACC | 0.09 = 3 % | 0.39 | 0.97 | ||
Dumont d'Urville | 140.02 | 45 | LATMOS-CNRS | LATMOS_RT | 0.20 = 5 % | 0.50 | 0.95 | |
Neumayer | 43 | U. Heidelberg | NDACC | % | 0.21 | 0.95 | ||
Dome Concorde | 123.31 | 3250 | LATMOS-CNRS | LATMOS_RT | % | 0.38 | 0.95 | |
Arrival Heights | 166.66 | 184 | NIWA | NDACC | % | 0.25 | 0.90 |
Table A2
MAX-DOAS hosting stations, ordered by increasing median tropospheric column (VCDgb, lowest at the bottom), that contribute to the tropospheric column validation. More details on the QA4ECV datasets can be found at
Station | Lat | Long | Altitude | Institute | Retrieval and | Reference | Med | Med (diff) | Spread | |
---|---|---|---|---|---|---|---|---|---|---|
format type | (VCDgb) | (IP68/2) | ||||||||
() | (m) | (Pmolec cm) | ||||||||
a.m.s.l. | ||||||||||
Vallejo | 19.48 | 2255 | UNAM | OE (MMF), | (a, b) | 29 | ; % | 12 | 0.40 | |
GEOMS | ||||||||||
UNAM | 19.33 | 2280 | UNAM | OE (MMF), | (a, b) | 19 | ; % | 7 | 0.84 | |
GEOMS | ||||||||||
Cuautitlan | 19.72 | 2263 | UNAM | OE (MMF), | (a, b) | 17 | ; % | 4.3 | 0.70 | |
GEOMS | ||||||||||
Gucheng | 39.15 | 115.73 | 13.4 | USTC | GA, ascii | (c, d) | 14 | ; % | 6.5 | 0.86 |
Xianghe | 39.75 | 116.96 | 95 | BIRA-IASB | OE (bePRO), | (e) | 11 | ; % | 5.7 | 0.83 |
GEOMS | ||||||||||
Chiba | 35.60 | 140.10 | 21 | Chiba U | PP, ascii | (f, g, h) | 8.6 | ; % | 6.3 | 0.79 |
Yokosuka | 35.32 | 139.65 | 10 | JAMSTEC | PP, GEOMS | (i) | 8.1 | ; % | 3.7 | 0.85 |
Kasuga | 33.52 | 130.48 | 28 | Chiba U | PP, ascii | (f, g, h) | 7.3 | ; % | 4 | 0.46 |
Mainz | 49.99 | 8.23 | 150 | MPIC | QA4ECV dataset, | 7.3 | ; % | 3.3 | 0.75 | |
GEOMS | ||||||||||
Cabauw | 51.97 | 4.93 | 3 | KNMI | PP, GEOMS | (j) | 6.7 | ; % | 3.5 | 0.40 |
Uccle | 50.80 | 4.36 | 120 | BIRA-IASB | OE (bePRO), | (k) | 5.7 | ; % | 3.3 | 0.75 |
GEOMS | ||||||||||
De Bilt | 52.10 | 5.18 | 20 | KNMI | PP, GEOMS | (j) | 5.4 | ; % | 2.8 | 0.64 |
Bremen | 53.10 | 8.85 | 27 | IUPB | QA4ECV dataset, | 5.2 | ; % | 2.3 | 0.59 | |
GEOMS | ||||||||||
Pantnagar | 29.03 | 79.47 | 237 | Chiba U | PP, ascii | (f, g, h, l) | 4.6 | ; % | 1.6 | 0.33 |
Thessaloniki_lap | 40.63 | 22.96 | 60 | AUTH | QA4ECV dataset, | (m) | 4.6 | ; % | 4.1 | 0.69 |
GEOMS | ||||||||||
Thessaloniki_ciri | 40.56 | 22.99 | 70 | AUTH | QA4ECV dataset, | (m) | 3.6 | ; % | 2 | 0.73 |
GEOMS | ||||||||||
Athens | 38.05 | 23.86 | 527 | IUPB | QA4ECV dataset, | 3.4 | ; % | 3 | 0.66 | |
GEOMS | ||||||||||
Phimai | 15.18 | 102.56 | 212 | Chiba U | PP, ascii | (f, g, h, l) | 2 | ; % | 0.7 | 0.47 |
Fukue | 32.75 | 128.68 | 80 | JAMSTEC | PP, GEOMS | (i) | 0.95 | ; % | 0.6 | 0.01 |
Figure A1
(a) Box-and-whisker plots summarizing the TROPOMI–MAX-DOAS tropospheric VCD difference, per station, ordered as a function of the median ground-based tropospheric column (largest median VCD values on top). Panels (b, c, d) present, respectively, the assumed aerosol optical depth (AOD; either retrieved from the MAX-DOAS measurement or taken from the climatology used in the retrieval), the MAX-DOAS absolute uncertainties, and the relative uncertainties (total median uncertainty in grey bars, random part in black and systematic part in red).
[Figure omitted. See PDF]
A3 The Pandonia Global Network Table A3PGN stations, ordered by median PGN column value, that contribute to the total validation. Mountaintop stations (not sensitive to lower lying tropospheric ) are marked with an asterisk. In the last row, we indicate where the data can be obtained (EVDC or directly from the PGN website). Note that only PGN data from a recent quality upgrade (with file version 004 or 005, where 005 has precedence) was used. The bias over all stations (median over all station medians) is ( %), while the overall dispersion (median over all 1/2IP68) is 1.8 , and the inter-station dispersion (1/2IP68 over all station medians) is 2.2 . Considering the low stations (Alice Springs to New Brunswick) only, the bias is 0.1 (2 %), the overall dispersion is 1.1 , and the inter-station dispersion is 0.2 . For the high stations (Buenos Aires to UNAM), the bias is ( %), the overall dispersion is 3.3 and the inter-station dispersion is 1.4 . Note that the mountaintop stations are not used in the calculation of these overall statistics.
Station code | Full name | Lat | Long | Alt | PGN | med(diff); | 1/2IP68 | archive | |
---|---|---|---|---|---|---|---|---|---|
med(VCD) | med(reldiff) | (diff) | |||||||
() | (m) | () | |||||||
unam | National Autonomous | 19.33 | 2280 | 18.7 | ; % | 4.6 | 0.87 | both | |
University of Mexico | |||||||||
Bayonne | Bayonne | 40.67 | 3 | 15.6 | ; % | 3.2 | 0.88 | EVDC | |
queens_ny | New York Queens | 40.74 | 25 | 14.7 | ; % | 3.6 | 0.84 | EVDC | |
College | |||||||||
sapienza | Rome Sapienza | 41.90 | 12.52 | 75 | 14.2 | ; % | 4.0 | 0.81 | EVDC |
city_college_ny | New York City | 40.82 | 113 | 13.7 | ; % | 3.4 | 0.91 | EVDC | |
College | |||||||||
isacrome | Rome CNR-ISAC | 41.84 | 12.65 | 117 | 10.5 | ; % | 3.2 | 0.85 | both |
bronx_ny | New York – | 40.87 | 31 | 10.3 | ; % | 3.3 | 0.90 | both | |
the Bronx | |||||||||
athens_noath | Athens National | 37.99 | 23.77 | 130 | 10.0 | ; % | 2.8 | 0.70 | PGN |
Observatory | |||||||||
innsbruck | Innsbruck | 47.26 | 11.39 | 616 | 9.8 | ; % | 3.4 | 0.59 | PGN |
buenos_aires | Buenos Aires | 20 | 8.6 | ; % | 2.6 | 0.86 | both | ||
new_brunswick | New Brunswick (NJ) | 40.46 | 19 | 6.4 | ; % | 1.5 | 0.90 | PGN | |
gsfc | Goddard Space | 38.99 | 90 | 5.9 | ; % | 1.3 | 0.80 | both | |
Flight Center | |||||||||
charles_city | Charles City (VA) | 37.33 | 6 | 5.6 | ; % | 2.0 | 0.44 | both | |
boulder | Boulder | 39.99 | 1660 | 5.4 | 0.0; 1 % | 1.6 | 0.87 | both | |
oldfield_ny | New York Old Field | 40.96 | 3 | 5.3 | 0.2; 5 % | 1.1 | 0.93 | both | |
helsinki | Helsinki | 60.20 | 24.96 | 97 | 5.1 | 0.5; 8 % | 1.0 | 0.77 | EVDC |
canberra | Canberra | 149.16 | 600 | 4.8 | ; % | 0.9 | 0.64 | EVDC | |
inoe | Magurele | 44.34 | 26.01 | 93 | 4.7 | 0.3; 8 % | 1.0 | 0.79 | EVDC |
fairbanks | Fairbanks | 64.86 | 227 | 4.7 | 0.1; 3 % | 1.4 | 0.43 | EVDC | |
egbert | Egbert | 44.23 | 251 | 4.3 | 0.5; 12 % | 0.6 | 0.88 | PGN | |
comodoro_rivadavia | Comodoro Rivadavia | 46 | 3.5 | ; % | 0.6 | 0.56 | PGN | ||
izana | Izaña | 28.31 | 2360 | 2.9 | 0.6; 19 % | 0.5 | 0.53 | both | |
mauna_loa | Mauna Loa | 19.48 | 4169 | 2.7 | 0.2; 6 % | 0.5 | 0.43 | both | |
alice_springs | Alice Springs | 133.88 | 567 | 2.7 | 0.2; 8 % | 0.4 | 0.61 | EVDC | |
altzomoni | Altzomoni | 19.12 | 3985 | 2.3 | 0.7; 28 % | 0.6 | 0.64 | both |
Data availability
No data were produced specifically for this research. The S5P data used here are publicly available from the Copernicus Open Access Hub (
Author contributions
TV, SC, and GP carried out the global validation analysis. JCL, KUE, and MVR contributed input and advise at all stages of the analysis. AMF (EVDC), JG (Multi-TASTE), and SN (MPC VDAF-AVS) preprocessed and/or post-processed the ground-based and satellite data. HJE, KFB, PFL, and JPV developed the TROPOMI data processor. AR, MVR, and TW contributed expertise on satellite data retrieval. AC, FH, KK, MT, APa, JPP, and MVR supervise network operation and contributed ground-based scientific expertise. AD, LSdM, and CZ supervise the Copernicus S5P mission, the S5P MPC, and the S5PVT. All other co-authors contributed ground-based data and expertise at ground-based stations. TV, SC, GP, and JCL wrote and edited the paper. All co-authors revised and commented on the paper.
Competing interests
The authors declare that they have no conflict of interest.
Special issue statement
This article is part of the special issue “TROPOMI on Sentinel-5 Precursor: first year in operation (AMT/ACP inter-journal SI)”. It is not associated with a conference.
Acknowledgements
Part of the reported work was carried out in the framework of the Copernicus Sentinel-5 Precursor Mission Performance Centre (S5P MPC), contracted by the European Space Agency and supported by the Belgian Federal Science Policy Office (BELSPO), the Royal Belgian Institute for Space Aeronomy (BIRA-IASB), the Netherlands Space Office (NSO), and the German Aerospace Centre (DLR). Part of this work was carried out also in the framework of the S5P Validation Team (S5PVT) AO projects NIDFORVAL (ID no. 28607, PI Gaia Pinardi, BIRA-IASB) and CESAR (ID no. 28596, PI Arnoud Apituley, KNMI). The authors express special thanks to Ann Mari Fjæraa, José Granville, Sander Niemeijer, and Olivier Rasson for post-processing of the network and satellite data and for their dedication to the S5P operational validation.
The LATMOS real-time processing facility is acknowledged for fast delivery of ZSL-DOAS SAOZ data. Fast delivery of MAX-DOAS data tailored to the S5P validation was organized through the S5PVT AO project NIDFORVAL. The authors are grateful to ESA/ESRIN for supporting the ESA Validation Data Centre (EVDC) established at NILU and for running the Fiducial Reference Measurements (FRM) programme and in particular the FRM4DOAS and Pandonia projects. The PGN is a bilateral project between NASA and ESA, and the NASA funding for the PGN is provided through the NASA Tropospheric Composition Program and Goddard Space Flight Center Pandora project.
The MAX-DOAS, ZSL-DOAS, and PGN instrument PIs and staff at the stations are thanked warmly for their sustained effort on maintaining high-quality measurements and for valuable scientific discussions. Aleksandr Elokhov and Aleksandr Gruzdev acknowledge national funding from RFBR through the project 20-95-00274. IUP Bremen acknowledges DLR Bonn for funding received through project 50EE1709A. The SAOZ network acknowledges funding from the French Institut National des Sciences de l'Univers (INSU) of the Centre National de la Recherche Scientifique (CNRS), Centre National d'Etudes Spatiales (CNES), and Institut polaire français Paul Emile Victor (IPEV). Work done by Hitoshi Irie was supported by the Environment Research and Technology Development Fund (2-1901) of the Environmental Restoration and Conservation Agency of Japan, JSPS KAKENHI (grant nos. JP19H04235 and JP17K00529), the JAXA 2nd Research Announcement on the Earth Observations (grant no. 19RT000351), and JST CREST (grant no. JPMJCR15K4). The University of Toronto ZSL-DOAS measurements at Eureka were made at the Polar Environment Atmospheric Research Laboratory (PEARL) by the Canadian Network for the Detection of Atmospheric Change (CANDAC), with support from the Canadian Space Agency (AVATARS project), the Natural Sciences and Engineering Research Council (PAHA project), and Environment and Climate Change Canada.
Financial support
This research has been supported by the ESA/ESRIN (grant no. 4000117151/16/I-LG) and the BELSPO/ESA ProDEx (TROVA-E2 (PEA grant no. 4000116692)).
Review statement
This paper was edited by Jhoon Kim and reviewed by two anonymous referees.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
© 2021. This work is published under https://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Abstract
This paper reports on consolidated ground-based validation results of the atmospheric
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
Details





















1 Royal Belgian Institute for Space Aeronomy (BIRA-IASB), Ringlaan 3, 1180 Uccle, Belgium
2 Royal Netherlands Meteorological Institute (KNMI), Utrechtseweg 297, 3730 AE De Bilt, the Netherlands
3 Institute of Environmental Physics (IUP), University of Bremen, Otto-Hahn-Allee 1, 28359 Bremen, Germany
4 Norsk Institutt for Luftforskning (NILU), Instituttveien 18, 2007 Kjeller, Norway
5 Science & Technology Corporation (S&T), Delft, the Netherlands
6 Goddard Space Flight Center (NASA/GSFC), Greenbelt, MD, USA; LuftBlick, Kreith, Austria; Institute of Meteorology and Geophysics, University of Innsbruck, Innsbruck, Austria
7 LuftBlick, Kreith, Austria; Institute of Meteorology and Geophysics, University of Innsbruck, Innsbruck, Austria
8 Laboratoire Atmosphères, Milieux, Observations Spatiales (LATMOS), UVSQ Université Paris-Saclay/Sorbonne Université/CNRS, Guyancourt, France
9 Laboratory of Atmospheric Physics, Aristotle University of Thessaloniki (AUTH), Thessaloniki, Greece
10 Royal Netherlands Meteorological Institute (KNMI), Utrechtseweg 297, 3730 AE De Bilt, the Netherlands; Meteorology and Air Quality group, Wageningen University, 6700 AA Wageningen, the Netherlands
11 Department of Physics, University of Toronto, 60 St. George Street, Toronto, Ontario, M5S 1A7, Canada
12 European Space Agency/Centre for Earth Observation (ESA/ESRIN), Frascati, Italy
13 Max-Planck-Institut für Chemie (MPI-C), Hahn-Meitner-Weg 1, 55128 Mainz, Germany
14 A.M. Obukhov Institute of Atmospheric Physics (IAP), Russian Academy of Sciences, Moscow, Russian Federation
15 Centro de Ciencias de la Atmósfera, Universidad Nacional Autónoma de México (UNAM), Mexico City, Mexico
16 National Observatory of Athens, Lofos Nymphon – Thissio, P.O. Box 20048 – 11810, Athens, Greece
17 Norsk Institutt for Luftforskning (NILU), P.O. Box 6606 Langnes, 9296 Tromsø, Norway
18 Center for Environmental Remote Sensing, Chiba University (Chiba U), Chiba, Japan
19 Danish Meteorological Institute (DMI), Lyngbyvej 100, 2100 Copenhagen, Denmark
20 Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Yokohama, Japan
21 Space and Earth Observation Centre, Finnish Meteorological Institute, Tähteläntie 62, 99600 Sodankylä, Finland
22 BK Scientific GmbH, Astheimerweg 42, 55130 Mainz, Germany
23 Royal Netherlands Meteorological Institute (KNMI), Utrechtseweg 297, 3730 AE De Bilt, the Netherlands; University of Technology Delft, Mekelweg 5, 2628 CD Delft, the Netherlands
24 Department of Precision Machinery and Precision Instrumentation, University of Science and Technology of China, Hefei, 230026, China
25 Atmospheric Research and Instrumentation, National Institute for Aerospace Technology (INTA), Madrid, 28850, Spain
26 Laboratoire de l'Atmosphère et des Cyclones (LACy), Université de La Réunion, Saint-Denis, France
27 National Institute of Water and Atmospheric Research (NIWA), Private Bag 50061, Omakau, Central Otago, New Zealand
28 University of Manchester, Oxford Rd, Manchester, M13 9PL, United Kingdom
29 Kyrgyz National University of Jusup Balasagyn (KNU), 547 Frunze Str., Bishkek, Kyrgyz Republic
30 Meteorology and Air Quality group, Wageningen University, 6700 AA Wageningen, the Netherlands