the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 26 Feb 2024
Impact of two high resolution altimetry mission concepts for ocean forecasting
Abstract. Observing System Simulation Experiments (OSSEs) with the Mercator Ocean/Copernicus Marine global 1/12° data assimilation system has been carried out to compare and quantify the expected performance of two high resolution altimetry mission concepts envisioned for the long-term evolution (post-2032) of the Copernicus Sentinel-3 topography mission. The two mission concepts are a constellation of two wide-swath altimeters and a constellation of 12 nadir altimeters. These two configurations greatly improve ocean forecasting and monitoring capabilities. Compared to a constellation of three nadir altimeters (the present configuration), analysis and forecast errors are reduced by a factor of 2. Our results also show that a constellation of two wide-swath altimeters has better performance than a constellation of 12 nadirs. Compared to a constellation of 12 nadirs, the error of the Sea Surface Height (SSH) forecast of a two wide swath constellation will be reduced by 14 % overall. Improvements are also observed when analysing surface currents and Lagrangian diagnostics. A constellation of two wide-swath altimeters thus seems to be a very promising concept for the long-term evolution of the Sentinel-3 topography mission.
- Preprint
(1726 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2024-420', Anonymous Referee #1, 24 Mar 2024
The use of Observing System Simulation Experiments (OSSEs) with the Mercator Ocean/Copernicus Marine global 1/12° data assimilation system offers groundbreaking insights into the future trajectory of satellite oceanography. The authors' innovative approach and the clarity with which the findings are presented are commendable. I would recommend the manuscript be accepted for publication subject to minor revisions.
I have the following suggestions:
1. The results articulate a clear advantage of the wide-swath altimeters in reducing SSH error variance and improving forecast skill. However, a more detailed statistical analysis would enrich the findings. Specifically, addressing the statistical significance of the improvements and the potential for overfitting in high-variability areas like the Gulf Stream and Kuroshio regions could provide deeper insights into the robustness of the wide-swath altimeters' superiority.
2. The methodological approach to scale separation for assessing the impact of altimeter data on different spatial scales is innovative. Nonetheless, the criteria for choosing the specific cutoffs (<200 km and <500 km) warrant further elaboration. An exploration of how these scale choices influence the overall analysis and whether alternative scale separations might yield different insights would contribute to a more nuanced understanding of the wide-swath altimeters' performance.
3. The reported improvements in model dynamics, such as surface and depth velocities and system mass, highlight the potential of wide-swath altimeter data assimilation. However, a discussion on the mechanistic reasons behind these improvements, particularly how they translate from better SSH data assimilation to enhanced model dynamics, would be valuable. This could help in understanding whether the improvements are due to better spatial coverage, reduced error variance, or other factors.
4. The Lagrangian diagnostics provide compelling evidence of the wide-swath altimeters' ability to capture mesoscale structures more effectively. While the improvement in particle drift analysis is notable, further discussion on the implications of these findings for practical oceanographic applications, such as marine ecosystem management or oil spill tracking, would underscore the real-world impact of the proposed satellite configurations.
5. The study's focus on two specific high-energy regions and its implications for global ocean forecasting raise questions about the generalizability of the results. Insights into how the wide-swath altimeter configuration might perform in lower-energy regions or under different oceanographic conditions would be beneficial. Moreover, considerations of the operationalization of these satellite configurations, including cost, technological feasibility, and integration into existing satellite constellations, would provide a more comprehensive picture of their potential for real-world implementation.
Citation: https://doi.org/10.5194/egusphere-2024-420-RC1 - AC2: 'Reply on RC1', Mounir Benkiran, 15 May 2024
-
RC2: 'Comment on egusphere-2024-420', Anonymous Referee #2, 27 Mar 2024
- AC1: 'Reply on RC2', Mounir Benkiran, 15 May 2024
-
RC3: 'Comment on egusphere-2024-420', Anonymous Referee #3, 29 Mar 2024
This article assesses the impact of the sea surface height (SSH) observation for the global ocean data assimilation system. The observation system simulation estimations (OSSEs) are performed and compared the impact of constellation of two wide-swath altimeters and twelve nadir altimeters to the current operating three nadir altimeters. Nature run for the OSSEs is obtained from the results without assimilation based on the latest operational model configurations, while the OSSEs uses the previous version of the model settings. The OSSEs show that the error reduction of SSH in the 2Swaths experiment is larger than other experiments. The impacts are also evaluated through the error reduction of the temperature and salinity profiles and surface velocity. The satellite constellation proposed in this study will be a next generation for the altimetry mission. Thus, the OSSEs have significant role on the decision of the future mission. The article almost satisfies the quality for publication, but there are several issues and concerns which should be improved. The detail is described below.
Major comments
Please specify the background SSH variance from Nature Run in the Fig. 6. It can be a lead time of the SSH forecast. For example, if the background SSH variance is about 20 cm2, the lead time is about 6 days (5days) in the 2Swaths (12Nadirs) experiment. Firstly, I expected that the large error reduction in the western boundary regions led to the more improvement of the separation distance (Figure. 16 B and C). However, the results show that the improvement is greater in the low latitudes not in the western boundary regions. Is it related to lead time for the OSSEs?
In Figure. 15, the surface velocities in the low latitudes have been improved. However, the additional SSH impact in the low latitudes is less effective compared to the middle and high latitude regions as shown in Figure 4 and 5. Why the surface velocities in the low latitudes are improved? There are two possible reasons. One is that the impact of SSH assimilation just on the low latitudes directly improves surface velocities. The other is that the SSH improvement in subtropical regions have adjusted the dynamical balance in the equator. Which is the more plausible reason?
With respect to the surface velocity, there is another concern. Tchonang et a., 2021 shows that the OSSE with single swath data had negative impact on the zonal velocity error in the equator in their Figure 12. However, Figure 15 in this study shows the improvement of the surface velocities in the low latitude. Does it mean that the more observation data obtained from the altimeter constellation can explain the improvement?
Anyway, there seems to be no explanation and discussion about Figure 15 in the manuscript. Please also check this point.
How many SSH data are used for assimilation? Could you describe the number of SSH observation in the manuscript with respect to Figure 1? Which has more observations in Fig.1 C or D? As described in the manuscript, there is more improvement in 2Swaths run than 12Nadirs run. It will be related with the difference in the number of SSH data and/or the spatial area available for observation (swath v.s. nadir). Which is more efficient for reducing the SSH error?
Figure 7 shows the interesting result, indicating the SSH variations in the Gulf Stream area is difficult to control by SSH assimilation comparing to the Kuroshio extension region shown in Figure 9. The SSH RMS error of the constellation experiments is almost similar to the 3Nadir experiment. What caused the comparable SSH error in early April for 2Swaths and in early July for 12Nadirs? Is it related with the position of the Gulf Stream axis?
Minor comments
P2 L44: Benkiran et al., (2012) is (2021) or (2022)?
P3 L3: A random noise is set to 2 cm in the manuscript, but 3cm in the previous studies (e.g., Benkiran et al., 2021, 2011). Is this correct?
P3 L19: The article focuses on the impact of the SSH assimilation. Therefore, it is better to describe the method of the assimilation scheme, especially for SSH variable. In Benkiran et al., (2021), SEEK filter, which is one of the sequential assimilation schemes, for the short-term variations is used for assimilation scheme. It will be easier for the readers to find out the forecast and analysis cycles. In addition, how was the mean surface height for the ocean model obtained? The mean surface height has a critical role of the SSH assimilation.
P5 L34: If the authors deal with the black box area in Figure. 7A, it is better to use “Kuroshio extension” to specify the area.
P6 L36 “Figures 15 B and C” is “Figures 16 B and C”
P8-9: Please check the order of the references. Sometimes the order should be reverse such as Vergara et al., 2019 and Ubelmann et al., 2015.
P12 Table2: The SSH variance error for wavelengths smaller than 500km in the 2Swatths experiment is 84 in Table 2. The value seems to be wrong.
P14 & P15 Fig. 7 D and Fig. 9 D: Why does the figure start from February?
P15 Fig. 10: The title of the left figure is (A)?
P18 Fig. 16 and 17: Please add (A), (B), (C), etc in the figure title.
Citation: https://doi.org/10.5194/egusphere-2024-420-RC3 -
AC3: 'Reply on RC3', Mounir Benkiran, 15 May 2024
We've taken your comments into account when editing the paper. For the separation of scales, we tested different separation cuts to highlight the impact, we have this set (200km and 500km) because this is where the impact is clearest. We used a Lanczos low-pass filter (python) to separate these scales. The improvement in surface velocities is due to better spatial coverage (2D fields) along the swaths, unlike the 12nadirs where we have to interpolate between the traces. with the 2swaths, the shape of the structures is better defined. This improvement in velocities will lead to better estimation of vertical velocities, which will have a significant impact on biogeochemical models (not shown here), as well as on applications such as slick drift (object tracking). We have set this up with particle drift.
Citation: https://doi.org/10.5194/egusphere-2024-420-AC3 -
AC4: 'Reply on RC3', Mounir Benkiran, 15 May 2024
Thank you for your corrections and comments. We have taken your comments into account in correcting the paper. We have incorporated the omissions and corrections into the article.
Citation: https://doi.org/10.5194/egusphere-2024-420-AC4
-
AC3: 'Reply on RC3', Mounir Benkiran, 15 May 2024
-
EC1: 'Comment on egusphere-2024-420', Bernadette Sloyan, 19 Apr 2024
The three reviewers have provided substantial comments and suggestion for the authors. I strongly encourage the authors to consider these comments and consider how they will address the issues identified by the reviewers
Citation: https://doi.org/10.5194/egusphere-2024-420-EC1
Status: closed
-
RC1: 'Comment on egusphere-2024-420', Anonymous Referee #1, 24 Mar 2024
The use of Observing System Simulation Experiments (OSSEs) with the Mercator Ocean/Copernicus Marine global 1/12° data assimilation system offers groundbreaking insights into the future trajectory of satellite oceanography. The authors' innovative approach and the clarity with which the findings are presented are commendable. I would recommend the manuscript be accepted for publication subject to minor revisions.
I have the following suggestions:
1. The results articulate a clear advantage of the wide-swath altimeters in reducing SSH error variance and improving forecast skill. However, a more detailed statistical analysis would enrich the findings. Specifically, addressing the statistical significance of the improvements and the potential for overfitting in high-variability areas like the Gulf Stream and Kuroshio regions could provide deeper insights into the robustness of the wide-swath altimeters' superiority.
2. The methodological approach to scale separation for assessing the impact of altimeter data on different spatial scales is innovative. Nonetheless, the criteria for choosing the specific cutoffs (<200 km and <500 km) warrant further elaboration. An exploration of how these scale choices influence the overall analysis and whether alternative scale separations might yield different insights would contribute to a more nuanced understanding of the wide-swath altimeters' performance.
3. The reported improvements in model dynamics, such as surface and depth velocities and system mass, highlight the potential of wide-swath altimeter data assimilation. However, a discussion on the mechanistic reasons behind these improvements, particularly how they translate from better SSH data assimilation to enhanced model dynamics, would be valuable. This could help in understanding whether the improvements are due to better spatial coverage, reduced error variance, or other factors.
4. The Lagrangian diagnostics provide compelling evidence of the wide-swath altimeters' ability to capture mesoscale structures more effectively. While the improvement in particle drift analysis is notable, further discussion on the implications of these findings for practical oceanographic applications, such as marine ecosystem management or oil spill tracking, would underscore the real-world impact of the proposed satellite configurations.
5. The study's focus on two specific high-energy regions and its implications for global ocean forecasting raise questions about the generalizability of the results. Insights into how the wide-swath altimeter configuration might perform in lower-energy regions or under different oceanographic conditions would be beneficial. Moreover, considerations of the operationalization of these satellite configurations, including cost, technological feasibility, and integration into existing satellite constellations, would provide a more comprehensive picture of their potential for real-world implementation.
Citation: https://doi.org/10.5194/egusphere-2024-420-RC1 - AC2: 'Reply on RC1', Mounir Benkiran, 15 May 2024
-
RC2: 'Comment on egusphere-2024-420', Anonymous Referee #2, 27 Mar 2024
- AC1: 'Reply on RC2', Mounir Benkiran, 15 May 2024
-
RC3: 'Comment on egusphere-2024-420', Anonymous Referee #3, 29 Mar 2024
This article assesses the impact of the sea surface height (SSH) observation for the global ocean data assimilation system. The observation system simulation estimations (OSSEs) are performed and compared the impact of constellation of two wide-swath altimeters and twelve nadir altimeters to the current operating three nadir altimeters. Nature run for the OSSEs is obtained from the results without assimilation based on the latest operational model configurations, while the OSSEs uses the previous version of the model settings. The OSSEs show that the error reduction of SSH in the 2Swaths experiment is larger than other experiments. The impacts are also evaluated through the error reduction of the temperature and salinity profiles and surface velocity. The satellite constellation proposed in this study will be a next generation for the altimetry mission. Thus, the OSSEs have significant role on the decision of the future mission. The article almost satisfies the quality for publication, but there are several issues and concerns which should be improved. The detail is described below.
Major comments
Please specify the background SSH variance from Nature Run in the Fig. 6. It can be a lead time of the SSH forecast. For example, if the background SSH variance is about 20 cm2, the lead time is about 6 days (5days) in the 2Swaths (12Nadirs) experiment. Firstly, I expected that the large error reduction in the western boundary regions led to the more improvement of the separation distance (Figure. 16 B and C). However, the results show that the improvement is greater in the low latitudes not in the western boundary regions. Is it related to lead time for the OSSEs?
In Figure. 15, the surface velocities in the low latitudes have been improved. However, the additional SSH impact in the low latitudes is less effective compared to the middle and high latitude regions as shown in Figure 4 and 5. Why the surface velocities in the low latitudes are improved? There are two possible reasons. One is that the impact of SSH assimilation just on the low latitudes directly improves surface velocities. The other is that the SSH improvement in subtropical regions have adjusted the dynamical balance in the equator. Which is the more plausible reason?
With respect to the surface velocity, there is another concern. Tchonang et a., 2021 shows that the OSSE with single swath data had negative impact on the zonal velocity error in the equator in their Figure 12. However, Figure 15 in this study shows the improvement of the surface velocities in the low latitude. Does it mean that the more observation data obtained from the altimeter constellation can explain the improvement?
Anyway, there seems to be no explanation and discussion about Figure 15 in the manuscript. Please also check this point.
How many SSH data are used for assimilation? Could you describe the number of SSH observation in the manuscript with respect to Figure 1? Which has more observations in Fig.1 C or D? As described in the manuscript, there is more improvement in 2Swaths run than 12Nadirs run. It will be related with the difference in the number of SSH data and/or the spatial area available for observation (swath v.s. nadir). Which is more efficient for reducing the SSH error?
Figure 7 shows the interesting result, indicating the SSH variations in the Gulf Stream area is difficult to control by SSH assimilation comparing to the Kuroshio extension region shown in Figure 9. The SSH RMS error of the constellation experiments is almost similar to the 3Nadir experiment. What caused the comparable SSH error in early April for 2Swaths and in early July for 12Nadirs? Is it related with the position of the Gulf Stream axis?
Minor comments
P2 L44: Benkiran et al., (2012) is (2021) or (2022)?
P3 L3: A random noise is set to 2 cm in the manuscript, but 3cm in the previous studies (e.g., Benkiran et al., 2021, 2011). Is this correct?
P3 L19: The article focuses on the impact of the SSH assimilation. Therefore, it is better to describe the method of the assimilation scheme, especially for SSH variable. In Benkiran et al., (2021), SEEK filter, which is one of the sequential assimilation schemes, for the short-term variations is used for assimilation scheme. It will be easier for the readers to find out the forecast and analysis cycles. In addition, how was the mean surface height for the ocean model obtained? The mean surface height has a critical role of the SSH assimilation.
P5 L34: If the authors deal with the black box area in Figure. 7A, it is better to use “Kuroshio extension” to specify the area.
P6 L36 “Figures 15 B and C” is “Figures 16 B and C”
P8-9: Please check the order of the references. Sometimes the order should be reverse such as Vergara et al., 2019 and Ubelmann et al., 2015.
P12 Table2: The SSH variance error for wavelengths smaller than 500km in the 2Swatths experiment is 84 in Table 2. The value seems to be wrong.
P14 & P15 Fig. 7 D and Fig. 9 D: Why does the figure start from February?
P15 Fig. 10: The title of the left figure is (A)?
P18 Fig. 16 and 17: Please add (A), (B), (C), etc in the figure title.
Citation: https://doi.org/10.5194/egusphere-2024-420-RC3 -
AC3: 'Reply on RC3', Mounir Benkiran, 15 May 2024
We've taken your comments into account when editing the paper. For the separation of scales, we tested different separation cuts to highlight the impact, we have this set (200km and 500km) because this is where the impact is clearest. We used a Lanczos low-pass filter (python) to separate these scales. The improvement in surface velocities is due to better spatial coverage (2D fields) along the swaths, unlike the 12nadirs where we have to interpolate between the traces. with the 2swaths, the shape of the structures is better defined. This improvement in velocities will lead to better estimation of vertical velocities, which will have a significant impact on biogeochemical models (not shown here), as well as on applications such as slick drift (object tracking). We have set this up with particle drift.
Citation: https://doi.org/10.5194/egusphere-2024-420-AC3 -
AC4: 'Reply on RC3', Mounir Benkiran, 15 May 2024
Thank you for your corrections and comments. We have taken your comments into account in correcting the paper. We have incorporated the omissions and corrections into the article.
Citation: https://doi.org/10.5194/egusphere-2024-420-AC4
-
AC3: 'Reply on RC3', Mounir Benkiran, 15 May 2024
-
EC1: 'Comment on egusphere-2024-420', Bernadette Sloyan, 19 Apr 2024
The three reviewers have provided substantial comments and suggestion for the authors. I strongly encourage the authors to consider these comments and consider how they will address the issues identified by the reviewers
Citation: https://doi.org/10.5194/egusphere-2024-420-EC1
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 433 | 134 | 53 | 620 | 24 | 26 |
- HTML: 433
- PDF: 134
- XML: 53
- Total: 620
- BibTeX: 24
- EndNote: 26
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 248 | 40 |
| France | 2 | 57 | 9 |
| China | 3 | 55 | 8 |
| Germany | 4 | 41 | 6 |
| United Kingdom | 5 | 30 | 4 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 248
