the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 20 May 2026
Comparing AERONET integrated water vapor with GNSS over the Arctic in the context of former comparisons with radiosondes and reanalyses
Abstract. The Integrated Water Vapor (IWV) determination in the Arctic is very important for atmospheric-climate studies. Different measurements and techniques are used for this purpose, having in general low spatial and temporal resolutions, except the Global Navigation Satellite Systems (GNSS), especially in the Arctic due to extreme weather conditions. In this work, the comparison of IWV measured by sun photometers (SP) from AERONET (IWVSP) and IWV measured by GNSS (IWVGNSS) at Arctic sites, is carried out. Taking coincident hourly averaged data IWVSP from 13 AERONET and 35 IWVGNSS sites were selected, covering the period 1997 to 2023. The comparison IWVGNSS versus IWVSP presents very good statistic indicators with the range of absolute and relative Mean Bias Error MBE (rMBE) of 0.007(0.8%)–0.091(15.4%) cm and standard deviation STD (rSTD) of 0.047(6.5%)–0.123(14.4%) cm. The positive bias at all sites detects a dry bias of SP respect to GNSS. Linear regression slopes are greater than 1 at nine sites, and lower at four. Those slopes lower than 1 appears related to higher frequency of IWV lower values (very dry conditions). All sites present Pearson correlation coefficients higher than 0.96, showing a low data dispersion. Results show that for IWV low values, less than 1.5 cm, SP predominates over GNSS, and the opposite when IWV values increase. This work complements two previous ones, where IWV from radiosondes and reanalysis were compared with IWVSP. Although the different data-bases, for coincident hourly, the highest values of IWV are given by GNSS, radiosonde, SP and reanalysis in this order.
- Preprint
(2537 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 02 Oct 2026)
- RC1: 'Comment on egusphere-2026-2374', Anonymous Referee #1, 15 Jul 2026 reply
-
RC2: 'Comment on egusphere-2026-2374', Anonymous Referee #2, 04 Sep 2026
reply
This study presents a comparison of atmospheric integrated water vapour measurements obtained using different observing systems: GNSS, AERONET sun photometers, radiosondes, and ECMWF ERA5 reanalysis. The measurement sites are distributed over a large area covering different parts of the Arctic, including northern Europe, Siberia, northern Canada, and Greenland. The measurements cover the period 1997–2023.
This study is important because water vapour contributes significantly to the radiative budget in a key region experiencing amplified climate warming. It is therefore crucial to obtain accurate and well-documented long-term water vapour measurements series in the Arctic. Studies such as this one contribute to the characterization of the different datasets, their uncertainties, and the biases between them.
The main results that I take from this study are that there is good agreement between GNSS and sun photometer measurements, the sun photometers exhibit a dry bias relative to GNSS, (IWV < 0.5 cm: large dispersion and less reliable behaviour of GNSS relative to the sun photometer; 0.5–1.7 cm: the best agreement, and importantly, this range contains the majority of the observations; > 1.7 cm: the dispersion increases again, with GNSS generally yielding higher IWV values than SP), radiosonde measurements are closer to the sun photometer measurements than GNSS measurements, while ERA5 generally provides higher IWV values. Finally, a clear seasonal cycle is observed, with higher IWV values in summer and lower values in winter. If the authors agree with highlighting these main messages, the abstract could be rewritten and clarified accordingly.
I support publication of this paper, provided that the authors address the following comments and adequately answer the questions below:
1) The measurement uncertainty associated with each observing technique should be documented more thoroughly. The authors give typical uncertainties in Section 1 (SP and GNSS uncertainties are of the order of 10%). However, if both instruments have uncertainties of approximately 10%, is a 7.5% bias necessarily statistically significant at the individual measurement level? It would be useful to clarify this point.
2) Please specify the variables and units reported in Table 4.
3) A maximum distance of 100 km is used for the co-location analysis. However, water vapour is known to exhibit substantial spatial variability on relatively small scales. Do the authors observe closer agreement for the closest sites and larger IWV differences for sites that are farther apart? A sensitivity analysis as a function of the distance between the SP and GNSS sites could be useful.
4) Are the numbers of measurements relatively homogeneous among the different sites? If the sites have very different numbers of observations, is this taken into account when calculating the overall bias for the Arctic?
5) The authors state that the sun photometer operates only when the Sun is sufficiently high and there is sufficient direct solar radiation, and that observations are further filtered for clouds. This is not the case for GNSS measurements. Consequently, SP and GNSS do not sample exactly the same meteorological conditions. This sampling difference could potentially contribute to the observed bias. It would therefore be interesting to perform the comparison after selecting measurements corresponding to similar meteorological conditions (e.g. cloudiness, weather regime, and precipitation).
6) The bibliography is incomplete. The authors could compare and discuss their results in relation to the following studies:
Zhang, C., Wang, S., Zhao, Y., Xu, Y., Zhang, J., Mo, Y., & Yu, H. (2025). Evaluation of water vapor from CARRA reanalysis based on GNSS and radiosonde observation in the Arctic. Journal of Atmospheric and Solar-Terrestrial Physics, 268, 106431. https://doi.org/10.1016/j.jastp.2025.106431
Negusini, M., Petkov, B. H., Tornatore, V., Barindelli, S., Martelli, L., Sarti, P., & Tomasi, C. (2021). Water Vapour Assessment Using GNSS and Radiosondes over Polar Regions and Estimation of Climatological Trends from Long-Term Time Series Analysis. Remote Sensing, 13, 4871. https://doi.org/10.3390/rs13234871
Citation: https://doi.org/10.5194/egusphere-2026-2374-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 159 | 75 | 25 | 259 | 24 | 25 |
- HTML: 159
- PDF: 75
- XML: 25
- Total: 259
- BibTeX: 24
- EndNote: 25
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General comments
The manuscript addresses an interesting and relevant topic. Comparisons of this type are still relatively scarce in the Arctic, and the use of an original photometer dataset represents a valuable contribution. However, I have several major concerns that should be addressed before the manuscript can be considered for publication.
First, the English language requires substantial revision, and the overall organization of the manuscript should be improved. Several sections are difficult to follow, and the presentation of the methodology and results is not always sufficiently clear. Second, a number of statements and conclusions are not adequately supported by the existing literature.
In addition, important methodological details regarding the datasets are missing. In particular, the manuscript should provide:
Maybe these methodological aspects should preferably be included in the Data or Instrumentation section of the manuscript.
I therefore recommend acceptance after major revision.
Specific comments
Line 22: For consistency with the rest of the manuscript, please use subscripts for IWVGNSS and IWVSP.
Line 27: Please reformulate the sentence "the positive bias at all sites detects a dry bias".
Line 29: The expression "a low data dispersion" is unclear. It may be more appropriate to refer to the consistency or agreement between the time series.
Lines 46-57: Please revise the wording and improve the flow of ideas in this paragraph, particularly the transitions between sentences.
Line 47: The expression "More recently" is not appropriate for a technique introduced in 1992 (Bevis et al., 1992).
Section "Instruments": Consider renaming this section to "Instruments and Data". Why not also introduce the radiosonde and reanalysis datasets in this section?
Subsection "Sun Photometer": Please provide more information about the dataset. What is the temporal resolution? What are the operating conditions and measurement schedule? What are the main uncertainties associated with the technique?
Lines 81-82: Please briefly summarize the main findings of the cited studies.
Subsection "GNSS": Please note that these data may be affected by limitations related to the relatively low spatial and temporal resolution of the VMF1 grids used by NGL. The authors may wish to discuss this issue and refer to: https://doi.org/10.5194/essd-15-723-2023.
Lines 110-111: Why is the uncertainty associated with IWV_SP discussed here rather than in the previous subsection?
Lines 113-114: It would be useful to briefly recall the method used for the vertical extrapolation of IWV.
Table 1: Why not include the radiosonde information in this table as well?
Lines 132-140: Please reformulate this paragraph for improved clarity.
Line 156: Please provide statistics describing the agreement among the GNSS datasets used to generate the average product (e.g., mean differences, standard deviation, minimum and maximum values).
Figure 2: I suggest moving this figure later in the manuscript, closer to the discussion in Section 4.2 where it is first described and interpreted.
Subsection 3.3: This section does not seem to warrant a dedicated subsection and could be moved to an appendix.
Lines 197-200: This sentence is difficult to understand and should be reformulated.
Table 3: Please include the number of comparison points used in each analysis to allow readers to assess the statistical significance of the results. In addition, it may be useful to separate the reanalysis comparisons by region in order to provide more insight into the spatial variability of the results.
Section 4.3: Consider merging Sections 4.3 and 4.4 into a single Discussion section.
Lines 239-240: Please reformulate this sentence.
Line 270: Please add an appropriate reference.
Lines 288-292: This paragraph is not clear and should be rewritten.
Line 293: I suggest replacing "demonstrated" with "confirmed".
Subsection 4.5: I recommend moving this subsection before the current Sections 4.3 and 4.4, especially if those sections are merged into a Discussion section.
Line 312: Please add an appropriate reference.