the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 14 Oct 2025
MAESTRO instrument operation and performance over two decades in orbit
Abstract. MAESTRO (Measurement of Aerosol Extinction in the Stratosphere and Troposphere Retrieved by Occultation), a dual UV (ultraviolet) and Visible-NIR (visible-near-infrared) spectrometer, has been operating aboard the Canadian satellite SCISAT for over 21 years. Recently, MAESTRO version 4.5 data products were released, consisting of volume mixing ratio profile data for NO2 and O3 retrieved from the UV channel, as well as a separate O3 product from the Visible-NIR channel. An aerosol extinction product is currently under development. Motivated by the instrument’s longevity, this paper will review the MAESTRO operations and performance over its lifetime, examining the key issues that impact its retrievals of atmospheric constituents. These include: a) the design of the MAESTRO spectrometer measurement schemes for sunset and sunrise occultations, including the role of long-term changes in the measured spectral intensities from the UV and Visible-NIR spectrometers, b) the determination of the position of the MAESTRO field of view (FOV) on the Sun, c) changes in the MAESTRO FOV position during occultations, including the impacts of this on the Level 1 transmittance calculation(s), d) understanding the relative position between the MAESTRO FOV and the ACE-FTS (Atmospheric Chemistry Experiment – Fourier Transform Spectrometer) FOV which are crucial for incorporating the input ACE-FTS atmosphere data in the Level 2 retrieval, e) verifying on-orbit the wavelength assignment at the detector arrays affected by a sudden persistent change in the instrument’s thermal environment, and f) the approaches taken to determine the MAESTRO measurement tangent heights, which are the pivotal steps in the MAESTRO retrieval of atmospheric constituents.
- Preprint
(10607 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-4636', Robert Damadeo, 01 Dec 2025
-
AC1: 'Reply on RC1', Jiansheng Zou, 02 Feb 2026
Response to Referee 1 – “MAESTRO instrument operation and performance over two decades in orbit” by Jiansheng Zou et al.
We thank Dr. Damadeo for his comments, which are given below with our responses provided in bold and italic font text.
This paper discusses the MAESTRO instrument onboard the SCISAT satellite (alongside the ACE-FTS instrument). Given the over 20 year long operational duration of the mission, there are a number of already published papers that detail various aspects of the instrument's operation, technical challenges, and corrections over the years. This paper summarizes these works and discusses updates as part of the most recently released v4.5 of the data product. The focus herein is not the impact on the data products, but rather a discussion of the nuances of some of the technical challenges faced over the years regarding the instrument and how those have been characterized and addressed. While I do have some questions regarding why some methods were chosen and perhaps some recommendations, this paper simply details what was done for an already released version of data. As such, those questions do not need to be addressed here and are better suited for offline conversations. Overall, the paper is well written and I do not see any need for modification.
Some general questions or points for the authors to consider (again, not necessary to add to the paper):
1) I would like to remind the authors that the blanket statement that solar occultation is "self-calibrating" is not necessarily true. Yes, the overall methodology is less sensitive to changes in instrument performance over orbit-to-orbit, month-to-month, and year-to-year, but not during the course of an occultation itself. It appears that there are several potential transients in the observation, particularly from thermal impacts and pointing knowledge/stability.
Agreed. The term “self-calibration” is used here in a general sense without detailed explanation you’ve brought up; however, in practice, changes in the thermal environment during an occultation are still accounted for. Each spectrum collected during an occultation is individually calibrated for its wavelength assignment, which varies with temperature. This is done by fitting the spectrum to a high-order polynomial of the solar spectrum derived from the high sun measurement taken during the same occultation.
Given the current measurement geometry, MAESTRO still faces challenges in correcting shifts in the slit positions projected onto the Sun, particularly at low altitudes. Although Sections 3.2 and 3.3 discuss these issues, no correction method has yet been established.
2) It appears that the horizontal extent of the UV and Visible-NIR slits are different. Why was the instrument manufactured this way?
This misalignment originated during ground testing. During testing the UV detector was found to have an electronic malfunction, necessitating its replacement, after which the pointing adjustment was partially redone. The cross-slit direction was properly controlled, and the lateral direction was assumed to remain aligned; however, the entire slit evidently shifted by about a quarter of a degree in the lateral direction—a displacement that was not detected when the re-positioned detector was realigned.
3) I would imagine the fact that the UV slit does not fully envelope the Sun to be particularly problematic regarding how the retrieval algorithm works and the impact on the accuracy and precision of the data products in the presence of a highly refracted Sun and/or pointing jitter.
The UV slit does not fully cover the solar disk, which introduces errors that are difficult to quantify. However, MAESTRO provides both UV and Visible–NIR O₃ measurements for the same atmospheric conditions, allowing the two products to be cross-validated. Comparisons with independent satellite measurements show that the Visible–NIR O₃ data exhibit better overall quality than the UV O₃ data (Jeffery et al., 2024).
4) Looking at figures 15 and 16, I'm not convinced that the UV measurements are "recovering". It's more likely you're seeing something spurious happen.
Regarding the degradation of the UV spectrometer, there is some speculation about the underlying cause. Because the spectral intensities at overlapping UV and Visible wavelengths (e.g., around 530 nm) exhibit similar attenuation, the issue is likely located outside the MAESTRO instrument itself. One possible scenario is that cutting oil used during the manufacture of the satellite base plate may have deposited on the main pointing mirror. The FTS instrument also shows spurious absorption features in the infrared, which supports this interpretation. Over time, the supply of this contaminant would diminish, and the material may either evaporate under solar radiation or undergo chemical transformation.
To assess whether the recent increase in UV sensitivity represents a true recovery or a spurious effect, we examined the UV NO₂ and O₃ data quality. We found that the data products generated after 2024 are reasonably robust, and these have therefore been released for public access.
5) I really don't understand the approach of trying to figure out the time shift and FOV offset between MAESTRO and ACE-FTS simultaneously. While it makes sense that the MAESTRO FOVs may have moved during launch, it's highly unlikely there was any sufficient mechanical impulse to shift them again post-launch. You should be able to nail down that overall move by analyzing exoatmospheric measurements over the lifetime of the mission. If they do continue to move, then you have an even bigger problem on your hands. After that, it is really strange that a timing offset between the two instruments would have a time-of-day dependence. The most reasonable way to assess it if it were static would be to look at altitude registration offsets (between the two instruments) in successive sunrises and sunsets. Again, if it's changing that often and that rapidly between those measurements, you have a bigger problem on your hands and I would have to question the validity of the entire data set.
Determinations of the pre-launch and post-launch slit positions were carried out independently, using different methods. The two approaches did not yield identical estimates, but the results were close. We examined the long-term exo-atmospheric spectral data and found that they are dominated by natural variations—daily, beta-angle, and annual cycles—as well as gradual long-term changes due to system degradation. No sudden changes or anomalies were detected. The exo-atmospheric spectra therefore appear stable, and so do the corresponding slit positions projected on the Sun.
The paper shows that the differences between the ACE-FTS and MAESTRO fields of view can be reduced to time shifts between the two instruments. Analysis of the data products (e.g., time-shift histograms) indicates that these time shifts are approximately constant, with variations likely resulting from effects of the differing vertical resolutions of the two instruments. This is notable because the fitting scheme—matching MAESTRO O₃ slant columns to those retrieved by the ACE-FTS for each occultation—was initially motivated by a time-of-day dependence observed in earlier versions; however, this dependency could have been influenced by the smaller dataset available earlier in the mission. With continued operation, and hence a larger dataset, a more statistically robust set of time shifts could be determined, as done in the most recent version of the MAESTRO dataset (v4.5). In the paper, we did not discuss these details, and more thorough work on this issue may be needed to definitively close the book on this issue.
Initially, there was concern that the ability to make an independent measurement of these species could be compromised because of the correlation introduced between the two instruments. However, there are more than 50 ozone measurements and only one degree of freedom associated with the applied time shift used to generate the needed correlation for producing a reliable profile. For the UV retrievals, the time shift derived from O₃ slant-column fitting also yields good NO₂ retrievals, further demonstrating the feasibility of the method.
Citation: https://doi.org/10.5194/egusphere-2025-4636-AC1
-
AC1: 'Reply on RC1', Jiansheng Zou, 02 Feb 2026
-
RC2: 'Comment on egusphere-2025-4636', Emmanuel Dekemper, 06 Jan 2026
General review
This paper explains the difficulties encountered since the beginning of operations of the ACE-MAESTRO solar occultation UV-VIS-NIR spectrometer. As highlighted in the manuscript, MAESTRO was the only satellite instrument of its kind between 2006 and 2017. Therefore, any effort to improve the mission’s geophysical products should be supported. High-quality O3 and aerosol profiles retrieved from MAESTRO’s solar occultation measurements would help fill the significant gap in the SAGE time series. A merged, multi-decadal solar occultation dataset would be a valuable tool for the atmospheric science community.
In this context, this work will be particularly useful to researchers outside the core MAESTRO team, as it describes several challenges encountered by the instrument during its lifetime and the in-depth analyses conducted to characterize key acquisition parameters, such as the spectrometer’s actual pointing on the solar disk. Any effort to improve understanding of the L1 product (the spectral atmospheric transmittance profiles) will enhance the mission’s scientific return. I recommend publishing the manuscript after addressing the minor comments below.
Specific comments:
At the end of the first paragraph of the introduction, it is stated that a paper by Jeffery et al. 2025 concluded to a "good agreement" between "measurement products" from "other instruments"... Even if the cited publication can definitely be consulted by the interested readers, I would prefer that a highlight of the conclusions by Jeffery et al. is given. This would allow to immediately capture the stakes of the work. Two sentences might suffice to point out where problems remain, and from there, explain how the submitted report contributes to improving the situation.
Section 3.1 discusses the "action tables," which, as I understand, are sequences of acquisition settings designed to maintain high-quality radiometric measurements from the bright high Sun to the dim low Sun. However, this discussion occupies too much space in the manuscript. The situation is clear: the "A" set of action tables was poorly calibrated, and the "B" set corrected this. Combining Figures 2 and 3 (panels (a) and (b) only) would allow readers to quickly assess the improvements introduced by the "B" set. I recommend using a logarithmic scale on the left y-axis to better visualize the altitude range of the acquisitions.
What the manuscript does not clearly explain is the "tic" and "pixel grouping" concepts. I could not fully grasp whether the detector pixels are exposed to sunlight for only 0.2 µs, with the exposure repeated up to 1,024 times, or if other exposure durations can be commanded. I also recommend converting exposure times to "seconds" or "milliseconds" instead of "tics," as currently done in lines 157–158. Finally, in the caption of Figure 2, could you clarify whether the geodetic coordinates refer to the tangent point or the spacecraft’s location?Section 3.2 and 3.3 tackle one of the most difficult aspects of the mission: the determination of the pointing of the spectrometer, and how it compares with ACE-FTS. The topic is complex but the authors have succeeded in showing the large efforts put into getting the pointing right (or at least understood). In particular, this is a pretty neat example of the synergistic use of the three payloads on the Scisat satellite. My main remarks concern the figures. First, I'm questioning the interest for the readers to have the x-axes of Fig 4 expressing scanning time. I would have expected the axis to show the commanded pitch or yaw angles offset. That would allow to reconnect with the physical angular dimensions of the solar disk and the slit acceptance angles. Second, the discussion on the differences of co-alignement of the fields of view of the MAESTRO channels and ACE-FTS is difficult to follow. While the supporting figure 6 is quite clear, the statement that it is "indicating that as the satellite moves in elevation during an atmospheric scan, errors will arise in the normalized spectra" is unclear. Could you explain what is meant by the satellite elevation move ?
Section 3.5.1 analyzes the impact of the star tracker failure on the satellite’s attitude. The evidence presented in Figure 13 convincingly shows that the failure did not significantly affect the spacecraft’s attitude control during solar occultations. However, the final sentences of the last paragraph in this section are confusing: if the primary effect on the L1 data is the change in the orientation of the slits during the solar occultation, why isn’t this addressed as a standalone point in the paper? Additionally, the use of "high" and "low altitudes" in line 439 is unclear. Could you clarify this?
Citation: https://doi.org/10.5194/egusphere-2025-4636-RC2 - AC2: 'Reply on RC2', Jiansheng Zou, 02 Feb 2026
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 221 | 137 | 30 | 388 | 25 | 24 |
- HTML: 221
- PDF: 137
- XML: 30
- Total: 388
- BibTeX: 25
- EndNote: 24
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This paper discusses the MAESTRO instrument onboard the SCISAT satellite (alongside the ACE-FTS instrument). Given the over 20 year long operational duration of the mission, there are a number of already published papers that detail various aspects of the instrument's operation, technical challenges, and corrections over the years. This paper summarizes these works and discusses updates as part of the most recently released v4.5 of the data product. The focus herein is not the impact on the data products, but rather a discussion of the nuances of some of the technical challenges faced over the years regarding the instrument and how those have been characterized and addressed. While I do have some questions regarding why some methods were chosen and perhaps some recommendations, this paper simply details what was done for an already released version of data. As such, those questions do not need to be addressed here and are better suited for offline conversations. Overall, the paper is well written and I do not see any need for modification.
Some general questions or points for the authors to consider (again, not necessary to add to the paper):
1) I would like to remind the authors that the blanket statement that solar occultation is "self-calibrating" is not necessarily true. Yes, the overall methodology is less sensitive to changes in instrument performance over orbit-to-orbit, month-to-month, and year-to-year, but not during the course of an occultation itself. It appears that there are several potential transients in the observation, particularly from thermal impacts and pointing knowledge/stability.
2) It appears that the horizontal extent of the UV and Visible-NIR slits are different. Why was the instrument manufactured this way?
3) I would imagine the fact that the UV slit does not fully envelope the Sun to be particularly problematic regarding how the retrieval algorithm works and the impact on the accuracy and precision of the data products in the presence of a highly refracted Sun and/or pointing jitter.
4) Looking at figures 15 and 16, I'm not convinced that the UV measurements are "recovering". It's more likely you're seeing something spurious happen.
5) I really don't understand the approach of trying to figure out the time shift and FOV offset between MAESTRO and ACE-FTS simultaneously. While it makes sense that the MAESTRO FOVs may have moved during launch, it's highly unlikely there was any sufficient mechanical impulse to shift them again post-launch. You should be able to nail down that overall move by analyzing exoatmospheric measurements over the lifetime of the mission. If they do continue to move, then you have an even bigger problem on your hands. After that, it is really strange that a timing offset between the two instruments would have a time-of-day dependence. The most reasonable way to assess it if it were static would be to look at altitude registration offsets (between the two instruments) in successive sunrises and sunsets. Again, if it's changing that often and that rapidly between those measurements, you have a bigger problem on your hands and I would have to question the validity of the entire data set.