the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 Apr 2025
EarthCARE’s Cloud Profiling Radar Antenna Pointing Correction using Surface Doppler Measurements
Abstract. The Earth Cloud Aerosol and Radiation Explorer (EarthCARE) mission, a joint effort between the European Space Agency (ESA) and the Japan Aerospace Exploration Agency (JAXA), aims to advance our understanding of aerosols, clouds, precipitation, and radiation using a comprehensive active and passive sensors payload. A key component of the payload is the 94-GHz Cloud Profiling Radar (CPR), which provides the first-ever Doppler velocity measurements collected from space. Accurate knowledge of the CPR antenna pointing is essential for ensuring high quality CPR Doppler velocity measurements. This study focuses on the geolocation assessment and antenna mispointing corrections during EarthCARE's commissioning phase and beyond, using Earth’s surface Doppler velocity measurements collected over the first nine months of the mission. While the instrument footprint is proven to be properly geolocated within about 100 meters, surface Doppler velocity observations reveal mispointing trends influenced by solar illumination cycles and thermoelastic distortions on the antenna. Correcting these effects significantly reduces biases, ensuring better Doppler velocity measurements, essential for understanding cloud microphysics and dynamics. The results, validated through the analysis of Doppler velocities in ice clouds, underline the critical role of pointing corrections for the success of the EarthCARE mission.
- Preprint
(2723 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
CC1: 'Comment on egusphere-2025-1680 - Commonalities to Doppler wind lidar on Aeolus', Oliver Reitebuch, 27 Apr 2025
I consider this as a very important and innovative contribution to the emerging field of active remote sensing from space of atmospheric winds with Doppler radar and lidar. Actually, a similar behavior of the wind bias along the orbit and with seasonal changes was found with the first Doppler wind lidar in space on Aeolus launched in August 2018. Here the thermo-elastic deformations of the primary mirror itself caused some complex wind bias variations along the orbit for the horizontal line-of-sight (HLOS) winds of several m/s as illustrated in Rennie et al. (2021), e.g. Fig. 4 or Weiler et al. (2021), e.g. Fig. 5. Although the vertical wind bias for CPR and the HLOS wind bias of Aeolus are caused by the antenna resp. primary mirror, there are some significant differences. The bias of the CPR is caused according to the authors by thermo-elastic deformations of the CPR antenna resulting in a mis-pointing of the LOS direction. This is mainly driven by the direct illumination or shading of the CPR antenna along the orbit by the Sun, which shows a seasonal dependency. The situation for Aeolus is more complex: Here the wind-bias is caused by a thermo-elastic deformation of the primary mirror shape, resulting in a change in the illumination (angle of incidence and divergence) of the Fabry-Perot and Fizeau interferometers used for detecting Doppler shifts from Rayleigh and Mie scattering. This thermo-elastic deformation of the Aeolus primary mirror is caused by changes of the infrared albedo of the Earth (outgoing long-wave radiation) - and thus depending on the atmosphere properties and sun illumination of the atmosphere. Despite this complex influence, it was possible to correlate the Aeolus wind bias with the temperature sensors mounted on the backside of the primary mirror (Fig. 2 in Weiler et al. 2021), which was a major breakthrough for correcting the wind biases (Rennie et al. 2021).
I would propose to include this commonality (and possibly also differences) in the introduction and discussion of the paper with the corresponding references.
It would be also nice to show a correlation of this bias with temperature sensors on the satellite structure and antenna, which could reflect this behavior along the orbit. Potentially also a correction of the LOS mis-pointing with these temperature sensors could be performed in the future.
A correction using the ground-return velocity from Aeolus is also discussed in the paper by Weiler et al. (2021), e.g. Fig. 13. As high-quality ground-returns from Aeolus are mainly limited to the polar regions with high albedo from ice and snow in the ultraviolet spectral region, and sea-surface returns can not be used due to their low SNR and non-negligible movement, a correction purely based on ground-returns is challenging for Aeolus. But such a correction scheme is still under investigation for the operational follow-on mission Aeolus-2, where higher SNR and higher vertical resolution for ground-return sampling is expected.
References:
Rennie, M., L. Isaksen, F. Weiler, J. de Kloe, Th. Kanitz, O. Reitebuch (2021): The impact of Aeolus wind retrievals in ECMWF global weather forecasts. Q. J. Roy. Meteorol. Soc., Vol 147, Issue 740, 3555-3586, https://doi.org/10.1002/qj.4142
Weiler, F., M. Rennie, Th. Kanitz, L. Isaksen, E. Checa, J. de Kloe, Ngozi Okunde, and O. Reitebuch (2021): Correction of wind bias for the lidar on-board Aeolus using telescope temperatures. Atm. Meas. Tech., 14, 7167–7185, https://doi.org/10.5194/amt-14-7167-2021
Citation: https://doi.org/10.5194/egusphere-2025-1680-CC1 -
AC1: 'Reply on CC1', Bernat Puigdomènech Treserras, 16 May 2025
Thank you for this very thoughtful and insightful comment.
We fully agree that the comparison with Aeolus is highly relevant and appreciate the detailed explanation of the underlying mechanisms and corrective approach. We will include a discussion of these commonalities and differences in the introduction and discussion sections, along with the references to Rennie et al. (2021) and Weiler et al. (2021). This context not only enriches the interpretation of the observed biases but also underscores the relevance of thermal deformation effects in spaceborne active remote sensing systems.
We also appreciate the suggestion regarding the use of onboard temperature sensors to trace and potentially correct LOS mispointing biases. The methodology presented here is independent of onboard temperature telemetry and demonstrates how natural targets can serve as valuable references for geolocation and antenna pointing correction. This approach is currently applied to the ESA L2a CPR data products, and the idea of incorporating temperature data in earlier processing stages, such as L1b, is indeed compelling from both engineering and scientific perspectives. It aligns with ongoing plans and discussions within the mission teams.
Citation: https://doi.org/10.5194/egusphere-2025-1680-AC1
-
AC1: 'Reply on CC1', Bernat Puigdomènech Treserras, 16 May 2025
-
RC1: 'Comment on egusphere-2025-1680', Anonymous Referee #1, 09 May 2025
This study calibrates the antenna pointing of EarthCARE’s 94 GHz Cloud Profiling Radar (CPR), the first spaceborne Doppler weather radar, to ensure accurate Doppler velocity measurements. Even small misalignments can bias cloud and precipitation velocities, so the authors focus on two key tasks during the commissioning phase: (1) verifying CPR geolocation and (2) identifying and correcting off-nadir pointing errors using Doppler signals from stationary ground targets. Analyzing surface returns from coastlines, mountains, and snow/ice over the first months of observations (Aug 2024–Feb 2025), they find the CPR is well geolocated (~100 m accuracy) but exhibits subtle, orbit- and season-dependent mispointing. These biases correlate with thermal cycles and are corrected using a climatological mispointing model, reducing velocity errors to within 5–7 cm/s (90% <10 cm/s). Validation using ice cloud data confirms the correction removes spurious Doppler biases. The study concludes that EarthCARE’s CPR is now accurately calibrated for high-precision cloud dynamics research.
This work is scientifically rigorous and addresses a vital calibration problem for EarthCARE. The authors thoroughly ground their study in prior literature on sources of Doppler error (spectral broadening, non-uniform beam filling, and pointing uncertainty) and build on pre-launch plans for EarthCARE’s calibration (citing Kollias et al. 2023 for broadening/NUBF corrections and earlier studies like Tanelli et al. 2005 for pointing issues). The methods used are appropriate and appear very robust. The strategy of using Earth’s surface as a calibration target is sound: a stationary ground return should have zero Doppler shift (aside from known platform motion components), so any systematic offset directly indicates a pointing error. A potential weakness in the methodology is that some choices and corrections are referenced to other documents and could be explained in more detail for completeness. For instance, the geolocation technique could be explained a little bit more. In general the article is well written, it requires mostly minor corrections, and I find just one major issue:
- The conclusion that thermoelastic deformation from solar heating causes the mispointing is based on circumstantial evidence (correlation with day/night cycle and seasonal repetition). The authors have made a strong case for it, but direct evidence (e.g. temperature measurements on the radar structure) was not presented. EarthCARE likely has temperature sensors on the CPR or nearby structure. A correlation between the measured antenna/baseplate temperatures and the inferred pointing bias could conclusively link cause and effect
Minor points:
1. L67 – Methodology for Mispointing Detection in Areas with Large Elevation Gradients:
Please clarify the methodology used to detect mispointing in regions with complex topography. Specifically:- What are the "artificial mispointing errors" referred to here?
- Are these based on simulations of surface returns assuming perfect satellite geolocation, where only antenna azimuth and elevation are varied and then compared to the actual radar signal?
- If so, please describe this process more explicitly, including assumptions and limitations.
For the coastline analysis, explicitly state that the land and ocean have distinct radar backscatter signatures (σ₀), which allows the land–ocean transition to be used for detecting pointing biases.
2. Figure 1 – Description and Interpretation
This figure needs a more detailed explanation:- Does panel brepresent the optimal mispointing correction for the entire domain shown in panel a, or is it specific to a selected location along the track?
- Clarify how both panels relate to the region over the Greek Islands.
- Additionally, please provide an equivalent of panel busing the coastline detection method, for direct comparison between methods.
3. L80+ – Time-Varying Pointing Correction
You show that antenna mispointing varies over time. This temporal evolution challenges the coastline-based detection method, which requires several months of data to achieve sufficient spatial sampling. Please discuss this limitation more clearly and consider quantifying the error introduced when using long-averaged coastline data under varying pointing conditions.4. L116: Spell out “100s” as "hundreds" for clarity
5. L120: Mention explicitly that the surface Doppler velocity analysis is performed globally, without separating land and ocean scenes.
6. L131: Clarify what is meant by "surface Doppler velocity."
- Is it the Doppler velocity at the radar signal peak, or a mean over a defined range around the peak?
- Given that the CPR's point target response is broad and flat, explain how the surface location is selected in the Doppler spectrum and how consistent this is across scenes.
7. Provide Parametrization in Usable Form:
- Please provide the Fourier expansion of the normalized temporal trend (ranging from -1 to 1) for a reference day (e.g., January 1).
- Include the same expansion for the minimum, maximum values and temporal shift.
- Express the correction directly in terms of Doppler velocity, not in antenna angle, so users can directly apply it to Level 1 data without relying on Level 2 products.
8. Section 4 – Comparison with High-Gradient Land Surface Method
- Include a marker (e.g., star/square/dot) on the correction plots to indicate results from the mispointing estimates derived over topographically complex land areas (as described earlier).
- Assess how these compare with the Doppler-based correction estimates.
- Use consistent marker colours for the same observation week across methods to visually indicate agreement or differences.
9. Section 5: In Figure 8, please add a line showing the global V–Z (Doppler velocity–reflectivity) relationship derived from all valid cirrus cloud observations (i.e., those not affected by the demodulation bias), not just January data. Alternatively, provide a supplementary figure showing the V–Z relationship for cirrus clouds, including the standard deviation envelope for context. Use all the data together over all frames and provide a polynomial fit to the formula.
Citation: https://doi.org/10.5194/egusphere-2025-1680-RC1 - AC3: 'Reply on RC1', Bernat Puigdomènech Treserras, 14 Jun 2025
-
RC2: 'Comment on egusphere-2025-1680', Anonymous Referee #2, 14 May 2025
General Assessment
This paper makes an important contribution to EarthCARE and the achievement of quality Doppler measurements from space. The methodology for addressing satellite pointing accuracy is generally sound, however I have identified several concerns regarding the fundamental assumptions about surface Doppler velocity measurements that should be addressed before publication.
Major Concerns
My primary concern stems from the statement on Page 5 Lines 118 that "Therefore any departure from the expected 0 m/s velocity indicates a potential mispointing." I guess this is somewhat true, but the quantitative level of departure will depend not only on the pointing angle of the antenna but also the NRCS of the surface at different angles. The surface backscatter will vary statistically with angle, and this will have the effect of causing non-zero-mean NUBF.
There is significant discussion of Figure 3, but very little discussion of Figure 3-a. Figure 3-a appears to show that while the mean surface Doppler velocity of the oceans are consistent with latitude, the Doppler velocity of the land surface varies significantly. This does not appear to be a noise issue, as the standard deviation in Figure 3-b does not show the same features as Figure 3-a. This matches my expectation above.
Having the surface Doppler technique potentially not work as well over land does not surprise me, but I'm concerned that if the remainder of this work includes the land surface Doppler velocity, it will cause added uncertainty on the order of 0.5 m/s (a visual estimate of per-latitude mean Doppler changes between land and ocean at the same latitude).
The current text discusses variability (around Line 160) and states that flat surfaces are expected to introduce no vertical motion at nadir. I do not agree with this statement due to how NRCS changes with angle. Further, the data in Figure 3-a do not appear to show a lack of flat-terrain-induced apparent vertical motion. Certain areas of flat land (such as the Great Plains in North America) show significant mean difference from the ocean data, while the Rocky Mountains to their west match the mean ocean velocity.
The rest of the paper is quite good, but because all the remaining data are effectively zonally averaged, this question of land-vs-ocean remains as a constant source of uncertainty, particularly in the Northern hemisphere.
On page 11 Line 255 the paper discusses a technique to ingest 250 km along-track averaged surface Doppler velocity observations. This seems like a good approach (over ocean) but there are no data shown about how this works. The statement around Line 265 that the 90th percentile of residuals remain below 0.00077 degrees is very promising, but is there some data that shows this? I don't see how it can work with the average surface velocity over land varying by ~0.5 m/s as compared to ocean.
Recommendations
Please address how variations in mean surface Doppler velocity, particularly over land, impact the overall analyses. I recommend performing these same analyses with an ocean-surface mask to determine how land-ocean discrepancies affect the results.
Please provide a more detailed physical explanation for the observed land-surface Doppler velocity variations.
Minor Comments
Figure 7 - The interpretation of this plot is unclear. It appears to show residual mispointing after removing seasonal effects, but the units are not specified. Please clarify what is being represented.
Line 80 and Figure 1 - Please explain how cross-track geolocation is accomplished with a single overpass using terrain. If this is one overpass it could be informative to show the range-to-surface vs track distance plot combined with the terrain.
Figure 2 - This figure is challenging to interpret. I recommend:
- separating the plot into two (ascending and descending) to clarify the different clustering patterns,
- making the stars indicating the mean values more prominent, and
- replacing the yellow text with a color that provides better contrast.Conclusion
While the paper represents an important contribution to the field, addressing these concerns - particularly regarding surface Doppler velocity assumptions and land-ocean discrepancies - would significantly strengthen the work.
Citation: https://doi.org/10.5194/egusphere-2025-1680-RC2 - AC2: 'Reply on RC2', Bernat Puigdomènech Treserras, 16 May 2025
Status: closed
-
CC1: 'Comment on egusphere-2025-1680 - Commonalities to Doppler wind lidar on Aeolus', Oliver Reitebuch, 27 Apr 2025
I consider this as a very important and innovative contribution to the emerging field of active remote sensing from space of atmospheric winds with Doppler radar and lidar. Actually, a similar behavior of the wind bias along the orbit and with seasonal changes was found with the first Doppler wind lidar in space on Aeolus launched in August 2018. Here the thermo-elastic deformations of the primary mirror itself caused some complex wind bias variations along the orbit for the horizontal line-of-sight (HLOS) winds of several m/s as illustrated in Rennie et al. (2021), e.g. Fig. 4 or Weiler et al. (2021), e.g. Fig. 5. Although the vertical wind bias for CPR and the HLOS wind bias of Aeolus are caused by the antenna resp. primary mirror, there are some significant differences. The bias of the CPR is caused according to the authors by thermo-elastic deformations of the CPR antenna resulting in a mis-pointing of the LOS direction. This is mainly driven by the direct illumination or shading of the CPR antenna along the orbit by the Sun, which shows a seasonal dependency. The situation for Aeolus is more complex: Here the wind-bias is caused by a thermo-elastic deformation of the primary mirror shape, resulting in a change in the illumination (angle of incidence and divergence) of the Fabry-Perot and Fizeau interferometers used for detecting Doppler shifts from Rayleigh and Mie scattering. This thermo-elastic deformation of the Aeolus primary mirror is caused by changes of the infrared albedo of the Earth (outgoing long-wave radiation) - and thus depending on the atmosphere properties and sun illumination of the atmosphere. Despite this complex influence, it was possible to correlate the Aeolus wind bias with the temperature sensors mounted on the backside of the primary mirror (Fig. 2 in Weiler et al. 2021), which was a major breakthrough for correcting the wind biases (Rennie et al. 2021).
I would propose to include this commonality (and possibly also differences) in the introduction and discussion of the paper with the corresponding references.
It would be also nice to show a correlation of this bias with temperature sensors on the satellite structure and antenna, which could reflect this behavior along the orbit. Potentially also a correction of the LOS mis-pointing with these temperature sensors could be performed in the future.
A correction using the ground-return velocity from Aeolus is also discussed in the paper by Weiler et al. (2021), e.g. Fig. 13. As high-quality ground-returns from Aeolus are mainly limited to the polar regions with high albedo from ice and snow in the ultraviolet spectral region, and sea-surface returns can not be used due to their low SNR and non-negligible movement, a correction purely based on ground-returns is challenging for Aeolus. But such a correction scheme is still under investigation for the operational follow-on mission Aeolus-2, where higher SNR and higher vertical resolution for ground-return sampling is expected.
References:
Rennie, M., L. Isaksen, F. Weiler, J. de Kloe, Th. Kanitz, O. Reitebuch (2021): The impact of Aeolus wind retrievals in ECMWF global weather forecasts. Q. J. Roy. Meteorol. Soc., Vol 147, Issue 740, 3555-3586, https://doi.org/10.1002/qj.4142
Weiler, F., M. Rennie, Th. Kanitz, L. Isaksen, E. Checa, J. de Kloe, Ngozi Okunde, and O. Reitebuch (2021): Correction of wind bias for the lidar on-board Aeolus using telescope temperatures. Atm. Meas. Tech., 14, 7167–7185, https://doi.org/10.5194/amt-14-7167-2021
Citation: https://doi.org/10.5194/egusphere-2025-1680-CC1 -
AC1: 'Reply on CC1', Bernat Puigdomènech Treserras, 16 May 2025
Thank you for this very thoughtful and insightful comment.
We fully agree that the comparison with Aeolus is highly relevant and appreciate the detailed explanation of the underlying mechanisms and corrective approach. We will include a discussion of these commonalities and differences in the introduction and discussion sections, along with the references to Rennie et al. (2021) and Weiler et al. (2021). This context not only enriches the interpretation of the observed biases but also underscores the relevance of thermal deformation effects in spaceborne active remote sensing systems.
We also appreciate the suggestion regarding the use of onboard temperature sensors to trace and potentially correct LOS mispointing biases. The methodology presented here is independent of onboard temperature telemetry and demonstrates how natural targets can serve as valuable references for geolocation and antenna pointing correction. This approach is currently applied to the ESA L2a CPR data products, and the idea of incorporating temperature data in earlier processing stages, such as L1b, is indeed compelling from both engineering and scientific perspectives. It aligns with ongoing plans and discussions within the mission teams.
Citation: https://doi.org/10.5194/egusphere-2025-1680-AC1
-
AC1: 'Reply on CC1', Bernat Puigdomènech Treserras, 16 May 2025
-
RC1: 'Comment on egusphere-2025-1680', Anonymous Referee #1, 09 May 2025
This study calibrates the antenna pointing of EarthCARE’s 94 GHz Cloud Profiling Radar (CPR), the first spaceborne Doppler weather radar, to ensure accurate Doppler velocity measurements. Even small misalignments can bias cloud and precipitation velocities, so the authors focus on two key tasks during the commissioning phase: (1) verifying CPR geolocation and (2) identifying and correcting off-nadir pointing errors using Doppler signals from stationary ground targets. Analyzing surface returns from coastlines, mountains, and snow/ice over the first months of observations (Aug 2024–Feb 2025), they find the CPR is well geolocated (~100 m accuracy) but exhibits subtle, orbit- and season-dependent mispointing. These biases correlate with thermal cycles and are corrected using a climatological mispointing model, reducing velocity errors to within 5–7 cm/s (90% <10 cm/s). Validation using ice cloud data confirms the correction removes spurious Doppler biases. The study concludes that EarthCARE’s CPR is now accurately calibrated for high-precision cloud dynamics research.
This work is scientifically rigorous and addresses a vital calibration problem for EarthCARE. The authors thoroughly ground their study in prior literature on sources of Doppler error (spectral broadening, non-uniform beam filling, and pointing uncertainty) and build on pre-launch plans for EarthCARE’s calibration (citing Kollias et al. 2023 for broadening/NUBF corrections and earlier studies like Tanelli et al. 2005 for pointing issues). The methods used are appropriate and appear very robust. The strategy of using Earth’s surface as a calibration target is sound: a stationary ground return should have zero Doppler shift (aside from known platform motion components), so any systematic offset directly indicates a pointing error. A potential weakness in the methodology is that some choices and corrections are referenced to other documents and could be explained in more detail for completeness. For instance, the geolocation technique could be explained a little bit more. In general the article is well written, it requires mostly minor corrections, and I find just one major issue:
- The conclusion that thermoelastic deformation from solar heating causes the mispointing is based on circumstantial evidence (correlation with day/night cycle and seasonal repetition). The authors have made a strong case for it, but direct evidence (e.g. temperature measurements on the radar structure) was not presented. EarthCARE likely has temperature sensors on the CPR or nearby structure. A correlation between the measured antenna/baseplate temperatures and the inferred pointing bias could conclusively link cause and effect
Minor points:
1. L67 – Methodology for Mispointing Detection in Areas with Large Elevation Gradients:
Please clarify the methodology used to detect mispointing in regions with complex topography. Specifically:- What are the "artificial mispointing errors" referred to here?
- Are these based on simulations of surface returns assuming perfect satellite geolocation, where only antenna azimuth and elevation are varied and then compared to the actual radar signal?
- If so, please describe this process more explicitly, including assumptions and limitations.
For the coastline analysis, explicitly state that the land and ocean have distinct radar backscatter signatures (σ₀), which allows the land–ocean transition to be used for detecting pointing biases.
2. Figure 1 – Description and Interpretation
This figure needs a more detailed explanation:- Does panel brepresent the optimal mispointing correction for the entire domain shown in panel a, or is it specific to a selected location along the track?
- Clarify how both panels relate to the region over the Greek Islands.
- Additionally, please provide an equivalent of panel busing the coastline detection method, for direct comparison between methods.
3. L80+ – Time-Varying Pointing Correction
You show that antenna mispointing varies over time. This temporal evolution challenges the coastline-based detection method, which requires several months of data to achieve sufficient spatial sampling. Please discuss this limitation more clearly and consider quantifying the error introduced when using long-averaged coastline data under varying pointing conditions.4. L116: Spell out “100s” as "hundreds" for clarity
5. L120: Mention explicitly that the surface Doppler velocity analysis is performed globally, without separating land and ocean scenes.
6. L131: Clarify what is meant by "surface Doppler velocity."
- Is it the Doppler velocity at the radar signal peak, or a mean over a defined range around the peak?
- Given that the CPR's point target response is broad and flat, explain how the surface location is selected in the Doppler spectrum and how consistent this is across scenes.
7. Provide Parametrization in Usable Form:
- Please provide the Fourier expansion of the normalized temporal trend (ranging from -1 to 1) for a reference day (e.g., January 1).
- Include the same expansion for the minimum, maximum values and temporal shift.
- Express the correction directly in terms of Doppler velocity, not in antenna angle, so users can directly apply it to Level 1 data without relying on Level 2 products.
8. Section 4 – Comparison with High-Gradient Land Surface Method
- Include a marker (e.g., star/square/dot) on the correction plots to indicate results from the mispointing estimates derived over topographically complex land areas (as described earlier).
- Assess how these compare with the Doppler-based correction estimates.
- Use consistent marker colours for the same observation week across methods to visually indicate agreement or differences.
9. Section 5: In Figure 8, please add a line showing the global V–Z (Doppler velocity–reflectivity) relationship derived from all valid cirrus cloud observations (i.e., those not affected by the demodulation bias), not just January data. Alternatively, provide a supplementary figure showing the V–Z relationship for cirrus clouds, including the standard deviation envelope for context. Use all the data together over all frames and provide a polynomial fit to the formula.
Citation: https://doi.org/10.5194/egusphere-2025-1680-RC1 - AC3: 'Reply on RC1', Bernat Puigdomènech Treserras, 14 Jun 2025
-
RC2: 'Comment on egusphere-2025-1680', Anonymous Referee #2, 14 May 2025
General Assessment
This paper makes an important contribution to EarthCARE and the achievement of quality Doppler measurements from space. The methodology for addressing satellite pointing accuracy is generally sound, however I have identified several concerns regarding the fundamental assumptions about surface Doppler velocity measurements that should be addressed before publication.
Major Concerns
My primary concern stems from the statement on Page 5 Lines 118 that "Therefore any departure from the expected 0 m/s velocity indicates a potential mispointing." I guess this is somewhat true, but the quantitative level of departure will depend not only on the pointing angle of the antenna but also the NRCS of the surface at different angles. The surface backscatter will vary statistically with angle, and this will have the effect of causing non-zero-mean NUBF.
There is significant discussion of Figure 3, but very little discussion of Figure 3-a. Figure 3-a appears to show that while the mean surface Doppler velocity of the oceans are consistent with latitude, the Doppler velocity of the land surface varies significantly. This does not appear to be a noise issue, as the standard deviation in Figure 3-b does not show the same features as Figure 3-a. This matches my expectation above.
Having the surface Doppler technique potentially not work as well over land does not surprise me, but I'm concerned that if the remainder of this work includes the land surface Doppler velocity, it will cause added uncertainty on the order of 0.5 m/s (a visual estimate of per-latitude mean Doppler changes between land and ocean at the same latitude).
The current text discusses variability (around Line 160) and states that flat surfaces are expected to introduce no vertical motion at nadir. I do not agree with this statement due to how NRCS changes with angle. Further, the data in Figure 3-a do not appear to show a lack of flat-terrain-induced apparent vertical motion. Certain areas of flat land (such as the Great Plains in North America) show significant mean difference from the ocean data, while the Rocky Mountains to their west match the mean ocean velocity.
The rest of the paper is quite good, but because all the remaining data are effectively zonally averaged, this question of land-vs-ocean remains as a constant source of uncertainty, particularly in the Northern hemisphere.
On page 11 Line 255 the paper discusses a technique to ingest 250 km along-track averaged surface Doppler velocity observations. This seems like a good approach (over ocean) but there are no data shown about how this works. The statement around Line 265 that the 90th percentile of residuals remain below 0.00077 degrees is very promising, but is there some data that shows this? I don't see how it can work with the average surface velocity over land varying by ~0.5 m/s as compared to ocean.
Recommendations
Please address how variations in mean surface Doppler velocity, particularly over land, impact the overall analyses. I recommend performing these same analyses with an ocean-surface mask to determine how land-ocean discrepancies affect the results.
Please provide a more detailed physical explanation for the observed land-surface Doppler velocity variations.
Minor Comments
Figure 7 - The interpretation of this plot is unclear. It appears to show residual mispointing after removing seasonal effects, but the units are not specified. Please clarify what is being represented.
Line 80 and Figure 1 - Please explain how cross-track geolocation is accomplished with a single overpass using terrain. If this is one overpass it could be informative to show the range-to-surface vs track distance plot combined with the terrain.
Figure 2 - This figure is challenging to interpret. I recommend:
- separating the plot into two (ascending and descending) to clarify the different clustering patterns,
- making the stars indicating the mean values more prominent, and
- replacing the yellow text with a color that provides better contrast.Conclusion
While the paper represents an important contribution to the field, addressing these concerns - particularly regarding surface Doppler velocity assumptions and land-ocean discrepancies - would significantly strengthen the work.
Citation: https://doi.org/10.5194/egusphere-2025-1680-RC2 - AC2: 'Reply on RC2', Bernat Puigdomènech Treserras, 16 May 2025
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 457 | 50 | 23 | 530 | 10 | 33 |
- HTML: 457
- PDF: 50
- XML: 23
- Total: 530
- BibTeX: 10
- EndNote: 33
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1