the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 May 2026
Measurement report: Global Total Ozone Records – part 1: ground-based monitoring networks performance assessment and status review
Abstract. Total column ozone (TCO) has been observed since the 1920s, with global monitoring established during the International Geophysical Year (1957–1958). We compile and assess TCO records from six major ground-based instrument types: Dobson and Brewer spectrophotometers, Filter ozonometers, zenith-sky DOAS (UVVIS), Fourier-transform infrared (FTIR) spectroscopy, and Pandora spectrometers. Data are drawn from the World Ozone and Ultraviolet Radiation Data Centre, Network for the Detection of Atmospheric Composition Change, Pandonia Global Network, and European Brewer Network. Using harmonized statistical criteria and daily comparisons with multiple satellite products and four reanalysis datasets, we evaluate site-level performance in five-year intervals from 1940 to 2024. Metrics include mean bias, variability of daily and monthly differences, seasonal amplitude, and the range of annual means, with percentile-based thresholds used to classify data quality.
Ground-based annual means generally agree with satellite and reanalysis benchmarks within ±2 %, with typical variability near 2 %. Larger discrepancies occur in the pre-satellite era, where reanalyses show biases of up to −5 % relative to Dobson observations. Network-wide distributions of daily mean differences indicate comparable internal consistency for Brewer and Pandora (standard deviations generally <2 %), while Filter, FTIR and UVVIS exhibit slightly broader spreads (<3 %), especially at high latitudes.
Network capacity has evolved substantially since the 2000s, with a decline in Dobson sites and expansion of Brewer and Pandora observations. By providing station-level flags and thresholds, this assessment helps users identify robust records, prioritize calibration and reprocessing, which ultimately strengthens their confidence in long-term ozone trend detection and satellite validation.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Atmospheric Chemistry and Physics.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(5669 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 27 Jun 2026)
- RC1: 'Comment on egusphere-2026-2009', Anonymous Referee #1, 12 Jun 2026 reply
Data sets
NASA/GFSC total ozone satellite datasets NASA https://avdc.gsfc.nasa.gov/pub/data/satellite/
Brewer, Dobson, and Filter total ozone records WOUDC/ECCC https://woudc.org/
FTIR and UVVIS records NDACC https://www-air.larc.nasa.gov/missions/ndacc/data.html
Brewer total ozone records EUBrewnet https://eubrewnet.aemet.es/eubrewnet
Pandora total ozone records PGN https://www.pandonia-global-network.org/home/documents/pgn-data/
TROPOMI overpass total ozone files ECCC https://hpfx.collab.science.gc.ca/~deg001/tropomi_ovp/
ERA5 reanalysis total ozone ECMWF https://cds.climate.copernicus.eu/datasets/reanalysis-era5-single-levels?tab=overview
MERRA-2 reanalysis total ozone Global Modeling and Assimilation Office (GMAO) https://disc.gsfc.nasa.gov/datasets?project=MERRA-2
JRA3Q reanalysis total ozone JMA https://gdex.ucar.edu/datasets/d640000/dataaccess/#
MSR2 reanalysis total ozone KNMI https://www.temis.nl/protocols/o3field/o3field_msr2.php
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 323 | 175 | 13 | 511 | 20 | 15 |
- HTML: 323
- PDF: 175
- XML: 13
- Total: 511
- BibTeX: 20
- EndNote: 15
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The paper is of high relevance for the scope of ACP. It assesses the quality of the largest ground-based total column ozone observation networks in a time-dependent manner and allows to identify problematic stations and times. It also provides intercomparisons between the ground-based records and the major satellite and re-analysis products.
The paper is based on a sound methodology and well written. I recommend it for publication after minor revisions.
specific comments:
- l. 37-39: differences to what, re-analysis? what is the internal consistency for Dobsons?
- l. 65/191: you might also want to put this into context with the upcoming activities with the BTS array spectrometers with WMO / SAG ozone mandating some evaluation of its capacity as potential 3rd generation ozone reference instrument as an information to the reader. But as there is no network associated to it (at least for now) I leave it to you if you to decide if you think such a statement would be sensible.
- l. 80-81: what do you exactly mean by "assimilation-driven shifts in the reanalysis datasets"? Some clarification and also the mention which datasets are (not) assimilated in which re-analyses will help to better understand not only this sentence but the whole paper.
- l. 91: similar to the above comment. Can you explain what you mean by time-dependent shifts in data assimilation
- l. 113: "from their entire 85-year records" -> as you mention in the introduction some sites started measuring in the 1920ies already, there are active stations with > 85 years of data.
- Fig 1. Please specify in the caption what the size of the circle means. Without this information the figure is not clear to me. Additionally a small cosmetic comment: the figure you provide seems to have lost its aspect ratio in the sense that the "circles" are actually ellipses (higher than wide)
- l.125-126: "it employs a double monochromator and optical wedge system to provide high measurement precision" --> you might also want mention the long optical path within the instrument which is an advantage in terms of stray light suppression
- l. 125: "e.g. zenith sky" --> also mention Umkehr
- l. 129: I would argue that "IGY" would be a more adequate and widely used abbreviation for the International Geophysical Year.
- l. 129: "Dobson network" --> "The Dobson network"
- l. 131: Mention that D083 is itself calibrated using Langley plots under the excellent atmospheric conditions at Mauna Loa
- l. 125-133: please specify which ozone cross-sections and ozone effective temperatures were used in the retrieval of the data in the actual paper. This is especially relevant for Dobson spectrophotometers where the often-used Bass and Paur cross-sections with constant temperature are known to introduce a artificial seasonal cycle (and an offset at polar sites) which may interfere with your seasonality analysis. Please also make a statement in your analysis how this change would affect it.
- l. 136-137: why do you capitalize "World" and "Regional" but not "calibration centres". This looks strange to me.
- l. 167-168: I do not understand why using radiative transfer for the airmass is a "cost" in itself. In my eyes, it would rather be the input data to the model that can introduce uncertainty, but I would argue that you have these assumptions almost everywhere. Dobsons and Brewers calculate an airmass based on geometrical considerations and atmospheric profiles.
- l. 206: please make clear from the beginning that (almost) no ground-based data is assimilated in any of the re-analyses, so that these can be considered as independent validation data for the ground-based observations, except for MSR2 before 1979 (and to some degree maybe also afterwards due to the correction you mention)
- l. 266-269, Fig. 2: A comment on the almost zero pre-satellite bias of MSR2 with Dobson assimilated might be insightful (to guide the reader in first place). If additionally you can think of a way to assess whether this is only due to quite local tearing of the model towards the assimilated observations which are also used for verification or by how much this is a real general improvement of the model quality (i.e. if there are non-assimilated ground-based observations for verification, e.g. filter ozonometer), this would be a very interesting finding.
- Fig. 3: please make more clear that in panel b) you are showing the std over all stations' annual means (at least that is what I understood/guessed from your plot).
- l. 304-306. This is almost the same sentence as at the beginning of section 2.4. Please rephrase and/or consider avoiding duplication.
- l. 330-331: this is a bit like comparing peers to apples and might advantage the dataset with the shortest records, because satellite and re-analysis quality could be expected to be best in the more recent years. Could you please indicate the same values for a coincident time period of all instruments as well?
- l. 337: From l. 228-235 I had understood that MSR2 just assimilated Dobson in the pre-satellite era, so I am confused with the statement "This is because MSR2 assimilated the Brewer and Dobson observations from some sites" made here. Please clarify this and the reason for the better agreement.
- Fig C1: The fact of showing lines in this graph is confusing as it suggests that the different datasets are somehow related in a series. Why not showing this as stacked boxplots or something the like?
l. 369-381: you tend to use "will" here, but I assume you are rather speculating on what "would" happen if...
- l.389: I don't understand your condition 2). Can you describe it a little more, so that the reader is not obliged to check Fioletov et al. just to understand the main working idea of your method. Finally it is your main classification method and would deserve a little more description (although it is similar to Fioletov et al.)
- l. 399-401: the parameters given in the parentheses seem more important to me than to be given in form of such a sidenote. Either move to some methodology section or a table.
- Fig. 6/7: why is Davos not mentioned as a continuation of the Arosa time series? The operators and PI's of this site are presenting this time series as one with a generic Arosa-Davos station and there are ongoing homogenisation efforts for an improuve merge of Arosa and Davos. Stübi et al. 2017 and 2021 analysed the re-location and concluded the effect is < 0.5%.
- l. 501-502: reformulate the part related to FTIR in the sense that it becomes clear that you have primarily used the reprocessed time series in the present work. Are the newer data (after 2023) processed by the individual institutes again?
-l. 523-524: "..., which their PI might lack..." sounds strange and not really causal to me. Please re-formulate
-l. 549: "The sites host ..." please reformulate this sentence/the tense of it.