the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 Oct 2025
Cross-calibration of GOME and SCIAMACHY Spectrometers Enhanced by Polarization Monitoring Devices Data
Abstract. Spectrometer instruments have significantly contributed to monitoring atmospheric composition and climate change for decades. Among them, the Global Ozone Monitoring Experiment (GOME, 1995–2011) and the Scanning Imaging Absorption Spectrometer for Atmospheric Chartography (SCIAMACHY, 2002–2012) were two well-known sensors whose missions overlapped by nearly a decade. Both instruments provided valuable data for atmospheric applications. However, to ensure data consistency and extend long-term time series, cross-calibration between the two instruments was important. The Fundamental Data Record for Atmospheric Composition (FDR4ATMOS) project, initiated by the European Space Agency (ESA), aims at harmonizing GOME and SCIMACHY Level 1 data, i.e., irradiance and reflectance measurements.
This paper presents, for the first time, the cross-calibration methodology for spectrometers used in the FDR4ATMOS project. Several challenges, such as differing spatial resolutions, lack of exact spatiotemporal overlap, and the need to preserve spectral structure, were addressed using targeted strategies. This process involved selecting scenes with minimal acquisition time differences over Pseudo-Invariant Calibration Sites (PICS) characterized by stable meteorological and atmospheric conditions. A key step involves spatially weighted averaging of SCIAMACHY pixels within each GOME footprint and computing spectral channel-wise ratios over Bands 2B, 3, and 4, which represent ultraviolet, visible, and near-infrared (UV/VIS/NIR) wavelengths. Additionally, the paper presented an analysis approach based on Polarization Monitoring Devices (PMDs) data to investigate the spatial homogeneity of pixels used in the cross-calibration and its influence on the performance of the cross-calibration.
Observations under near-clear-sky conditions from 2003 were collocated over PICS and used to derive transfer functions (TFs). Polynomial TFs were fitted for Bands 2B and 3, while a constant TF was used for Band 4. The TFs showed dependence on viewing zenith angle (VZA), degradation, and wavelength. The uncertainty of TFs increased with wavelength, corresponding to reduced homogeneity in PMD measurements. Using PMD measurements from cross-calibrated pixels as an indicator to filter out non-homogeneous pixels of the main spectral channels resulted in an uncertainty reduction up to 70 % in the TFs.
Overall, the presented cross-calibration approach and PMD-based analysis provide a pathway toward generating consistent and long-term spectrometer records. This work highlights the potential for expanding future TFs derivation beyond ideally suited scenes, increasing robustness across the varied surface and atmospheric conditions.
- Preprint
(18662 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 28 Dec 2025)
- RC1: 'Comment on egusphere-2025-4942', Anonymous Referee #1, 09 Dec 2025 reply
-
RC2: 'Comment on egusphere-2025-4942', Anonymous Referee #2, 19 Dec 2025
reply
Owda et.al. describe a method to harmonize the reflectances of the spectrometers GOME and SCIAMACHY. Goal is a harmonized time-series of both instruments for reflectances. The authors uses measurements over PICS sites with co-located reflectance measurements from both instruments. The method includes utilizing the spatially higher resolved measuerments of the PMDs of both instrument to ensure homogenous scenes used for the harmonization.
General commentsIn general, the paper is well written. The ideas behind the method are well described.
Long term data records are an important topic in atmospheric science. Harmonizing the data from different instrument is especially an issue for satellite based measurements. Harmonizing reflectances is a new approach to harmonize the measurements on spectrometer level.This paper describes a method to harmonize the reflectances of the satellite instruments GOME
and SCIAMACHY. A method to harmonize the irradiance is not covered. This is not clearly stated in the abstract (according the the abstract, the FDR4ATMOS project aims to provide also irradiance). Please clarify this in the abstract, for example in line 7:`This paper presents, for the first time, the cross-calibration methodology for the reflectance of the spectrometers used in the FDR4ATMOS project. You might also consider adding the term reflectance to the title of the paper.`
In the introduction, I miss the motivation for the selected spectral windows. This is driven by the common spectral windows, which contain spectral absorbtions of important trace gases. I suggest to add the relavant trace gases for each band also in Table 2.In Figure 7, transfer functions (TFs) for the whole band 4 (including O2A) is shown (dashed line). In Figure 10, only the two separate TFs for outside the absorbtion are shown. The TF for band 4 needs some clarifications:
- How is the TF for the whole channel build?
- Which TF will be used for the harmonization? This is finally stated in the conclusion: `constant value for the whole band`. This needs to be clarified already here.
- Therefore, the whole channel 4 TFs need to be added to Figure 10, because that is the used one.
Related to this: In Table 1, for the NIR the wavelength intervalls are given as 756–757 & 773–774 nm. These are the intervall used for the TF calculation, but the FDR4ATMOS product will contain the intervall 756-774nm including the O2A Band. Here you need to distinguish between the spectral window in the harmonized product and the windows used in your calculations. You also need to descripe the calculation of the final TF for band 4.
Specific and technical comments
- p5, l120:
The SCIAMACHY Level1 product does not contain cloud information. Probably, your cloud information comes from the corresponding Level 2 products. Please clarify. - p5, 122:
The cluster concept refers to a subdivision of a channel containing a specific wavelength region and detector exposure time, aiming to identify certain important spectral windows in the data.
The goal is not to identify important windows. The goal is to optimize the data rate towards the important sprectral windows:
..., aiming to optimize the data rate towards important spectral windows in the data. - p5, l127:
with the scial1c tool developed by DLR (the link to the tool is under Section 8).
I suggest to change this to a normal reference and add the URL to the references, something like:
with the scial1c tool developed by DLR (ESA, 2025c). - p7, Fig. 2:
Add to the text in the building blocks the variable names used in text/formulas, just as in the first block with (Rg) and (Rs)
- SCIA2GOME Pseudo reflectance -> SCIA2GOME Pseudo reflectance (RSCIA2GOME)
- Ratio of refectance (SCIAMACHY/GOME) -> Ratio of refectances (Ratio)
- Transfer functions ... -> Transfer functions (TF) ...
I suggest to add a block for "Outlier removal", that step is missing in the sketch. - p7, eq(1):
Please use a proper multiplication sign in equation (1), this seem to be a simple dot (.).
Should be \cdot in Latex math mode. Same for all further equations. - p7, Figure 3:
There is a gap between subsequent SCIAMACHY scanlines. This is not expected, there is (almost) no gap between the scanlines (similar to the two GOME ground pixels in the plot). Please check your figure. - p9, Table2:
This table is a duplication of Table 1, only provide and refer to Table 1. - p 24, l397:
the spectral channels with strong absorption features: Only the O2A absorbtion band is excluded, so I suggest to clarify here, that the O2A window is the one excluded. - p 24, l401:
It found that... -> It has been found that... - 25, Section 8:
Instead of a list of links (with wrong indentations, use the `itemize` environment for lists), make a short text and move the links to the references. Something like:
SCIAMACHY and GOME Level-1 data are available from ESA (ESA, 2025c; ESA, 2025d), etc.
Citation: https://doi.org/10.5194/egusphere-2025-4942-RC2 -
RC3: 'Comment on egusphere-2025-4942', Anonymous Referee #3, 19 Dec 2025
reply
Major comments
This paper presents the post-calibration method of GOME reflectance through cross-calibration with SCIAMARCHY data. However, I believe that additional analyses are required to enhance the scientific significance of the study.
- This paper cites a reference for the long-term degradation of GOME in contrast to the radiometric stability of SCIAMACHY. The authors should perform and present their own evaluation results. This would also help establish a more robust calibration strategy.
- The paper emphasizes the necessity of correcting GOME degradation but presents transfer function (TF) results only for the year 2003. The authors should include year-by-year results at least and clearly specify whether and how this methodology is applied on an annual or monthly basis.
- This paper also addresses how to deal with the post-calibration for GOME 1995-2022.
- They should provide a validation results of the L2 product with and without TF-based correction.
- The manuscript requires careful proofreading. I have indicated only a few of the total typos in the technical correction section.
Specific comments
Section 1.3
#Line 82 the necessity to preserve the spectral structure of the observations
->Please clarify. This say that “the spectral resolution of the two datasets needs to be matched.” ?
#Line 83-84: These challenges distinguish spectrometer cross-calibration from that of imaging satellite instruments, which typically do not face such constraints.
-> The meaning of this statement is unclear to me.
#Imaging sensors generally have much higher spatial resolution and do not require strict preservation of spectral structure
-> The comparison implied in this statement is unclear. Limb? It is hard to agree with “high spatial resolution measurements do not require strict preservation of spectral structure.”
Section 2.1
# The PMDsare broadband detectors specifically designed to capture polarization information with high temporal resolution
-> High temporal resolution or spatial resolution? Line 140 highlights the spatial resolution of PMD.
5.2 Transfer functions
#The ratios were higher for the overlapping pixels with eastward VZA compared to those with nadir and westward VZA
-> Interpreting the results solely based on the magnitude of the ratio is not appropriate. A ratio close to unity indicates consistency between the two satellite reflectances, whereas deviations above or below 1 should be interpreted as discrepancies rather than simply as “higher” or “lower” values.
#. On average, the uncertainties in Bands 3 and 4 were about 4 times greater than the uncertainties in the Band 2B nadir TFs
-> Please provide deeper insight. Why Band3/4 have more SDs than band 2B. GOME/SCIAMACY spectral dependent SNR information could be sourced.
# Section 6.1: “Third, the PMD-based analysis approach was applied as an indicator of pixel homogeneity in the cross-calibration. This approach helped exclude non-homogeneous pixels and ensured more uniform pixel textures, thereby reducing the uncertainty of TF. This was possible thanks to the higher sampling frequency of PMD measurements compared to the main spectral channels”
-> This does not appear to be consistent with the analysis results presented in Section 5.5. Section 5.5 shows no significant difference in the transfer functions (TFs) with and without PMD-based filtering. The impact of the filtering is instead reflected in the standard deviations (SDs). However, the correction factor used for cross-calibration depends only on the TFs. Therefore, I believe that the PMD-based filtering should be removed in order to enhance the sampling. Moreover, this study already selects analysis regions where spatial homogeneity is ensured.
#Section 6.3 GOME PMDs had roughly double the spectral bandwidth compared to SCIAMACHY, and the optics are different, (ii) the GOME measurements at the equatorial crossing time of 10:30 against SCIAMACHY at 10:00, hence, the GOME PMDs would receive larger signals, (iii) PMD data on Level 1b were not corrected with the cos (phi) factor (as the reflectance was), (iv) the larger values in PMDs
-> It should be described in data description first, not in discussion section.
Technical comments
Line 54 (Cosnefroy et al., 1996) è Cosnefroy et al. (1996)
Line 55 km2 à km2
Line 66 remove ”two spectrometers”
Line 67 (Owda and Lichtenberg, 2025) -> Owda and Lichtenberg, (2025)
Line 66 (Coldewey-Egbers et al., 2018) -> Coldewey-Egbers et al. (2018).
Line 77 have recently been è have been or was recently
Line 78 please add a reference for FDR4ATMOS and introduce any related outcomes from companion papers.
Line 79 the: ultra-violet è theultra-violet
Line 80 The cross-calibration between the satellite spectrometers GOME and SCIAMACHY è This cross-calibration
Line 98 scattered in the Earth’s atmosphere è scattered by the Earth’s atmosphere
131 remove “(“, “)”
Section 3 is better to be placed in Section 4.
Line 155 cross-calibration on Rs and Rg è cross-calibration between Rs and Rg,
Line 160-164. The information is correct, but need to be revised. For example, from “Learth is the calibration radiance as measured by the instrument.”, “as measured by the instrument” is unnecessary.
Line 166. The information is correct, but need to be revised.
Line 167: Figure 4 or Figure 3 ?
Line 179: difference in è difference between
Line 183: need to be revised for FDR4ATMOS manual (2024)
Why both table 1 and table 2 are provided ?
5.1 is better to be placed in section 4.
need to be revised for For Band 2B (Fig. (8a),
the correlation between PMD measurements from GOME and SCIAMACHY è the correlation of PMD measurements between GOME and SCIAMACHY
Citation: https://doi.org/10.5194/egusphere-2025-4942-RC3
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 185 | 52 | 23 | 260 | 19 | 18 |
- HTML: 185
- PDF: 52
- XML: 23
- Total: 260
- BibTeX: 19
- EndNote: 18
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
See attachement