the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 29 Jan 2024
Fusion of Lagrangian drifter data and numerical model outputs for improved assessment of turbulent dispersion
Abstract. Transport and dispersion processes in the ocean are crucial, as they determine the lifetime and fate of biological and chemical quantities drifting with ocean currents. Due to the complexity of the coastal ocean environment, numerical circulation models have difficulties to accurately simulate highly turbulent flows and dispersion processes, especially in highly energetic tidal basins such as the eastern English Channel. A method of improving the results of coastal circulation modeling and tracer dispersion in the Dover Strait is proposed. Surface current velocities derived from Lagrangian drifter measurements in November 2020 and May 2021 were optimally interpolated in time and space to constrain a high-resolution coastal circulation MARS model, with careful attention given to selecting ensemble members composing the model covariance matrix. The space-time velocity covariances derived from model simulations were utilized by the Optimal Interpolation algorithm to determine the most likely evolution of the velocity field under constraints provided by Lagrangian observations and their error statistics. The accuracy of the velocity field reconstruction was evaluated at each time step. The results of the fusion of model outputs with surface drifter velocity measurements show a significant improvement (by ~50 %) of the model capability to simulate the drift of passive tracers in the Dover Strait. Optimized velocity fields were used to quantify the absolute dispersion in the study area. The implications of these results are important, as they can be used to improve existing decision-making support tool or design new tools for monitoring the transport and dispersion in coastal ocean environment.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(2897 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2897 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-176', Anonymous Referee #1, 26 Feb 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-176/egusphere-2024-176-RC1-supplement.pdf
-
AC1: 'Reply on RC1', Sloane Bertin, 15 Mar 2024
Dear reviewer,
Thank you very much for dedicating time to review our manuscript, and for your clear and pertinent remarks. Please find our point-by-point response concerning the major comments. Concerning the minor comments, all of them have been considered and clarified in the revised version of the manuscript (given in red).
Major comments
- Line 125 in page 5, authors just present the tidal conditions for model’s boundary. What are the other boundary conditions for simulation? Such as gradient, Clamped, Flather or other boundary conditions?
The numerical model utilizes nested configurations with progressive resolutions: (i) 2 km covering the Northeast Atlantic (level 0), (ii) 700 m at the regional scale, encompassing the English Channel (level 1), and (iii) 250 m for the Eastern English Channel (level 2). This nesting technique enables the accurate capture of interactions between large-scale and small-scale processes. This enables the transfer of all resolved fields from lower resolution levels to the open boundaries of higher resolution levels.
The model accounts for kinematic free-surface and bottom boundary conditions, contingent upon friction terms (Lazure and Dumas, 2008). The turbulence closure employed in the model follows the approach described in Gaspar et al. (1990).
All these details will be added to the part 2.3 ‘Current velocity from numerical model’.
- Two one-year long model runs are implemented in this research. Can authors provide more detailed modeling settings to make model stable in such long simulation.
Comprehensive information regarding model equations, the coupling of barotropic and baroclinic modes, model discretization, solving methods, computational stability according to CFL criterion (table 1, Lazure and Dumas, 2008), and costs are meticulously outlined in Lazure and Dumas (2008). To maintain CFL stability, the modeling timestep was set to 30 seconds for the level 2 model.
All these details will be added to the part 2.3 ‘Current velocity from numerical model’.
- In Figure 5, authors just present the model absolute errors at S2-4. May authors present the absolute errors using “Box-Whisker” plot over all drifters of S2?
Thank you for this advice. I modified Figure 5 and used box-whiskers to present the model absolute error for all the drifters during S1 (Fig. 5a) and S2 (Fig. 5b). Consequently, I modified the descriptive paragraph L266-280 and legend.
- Line 280-281, authors describe that the drifter S2-4 is well reproduced. Can authors plot all drifters of S1and S2 in Figure 6 to make reader directly understand the simulation results. There is similar problem in Figure 7. In Figure 7, authors use drifters S1-2 and S2-1, why is it different from Figure 6? Can authors also present all drifters of S1 and S2 in Figure 7? Or using “Box-Whisker” plot to present separation distance in Figure 7? Furthermore, how about the wind-corrected trajectories of other drifters in S1 and S2? Can present other drifters in Figure 8? If other OI- or corrected trajectories are similar to S1-1 and S2-1, please describe the related statement of other trajectories in S1 and S2.
I modified Figures 6 and 8 by presenting all the drifters’ trajectories (2 for S1 and 4 for S2) in order to make the reader directly understand the simulation results. I also modified Figure 7 by using box-whiskers to present the separation distance results for all the drifters simultaneously. Consequently, the paragraphs and legends concerning these three figures have been modified.
Thank you,
Best regards.
Sloane Bertin
Citation: https://doi.org/10.5194/egusphere-2024-176-AC1
-
AC1: 'Reply on RC1', Sloane Bertin, 15 Mar 2024
-
RC2: 'Comment on egusphere-2024-176', Anonymous Referee #2, 24 May 2024
General comments:
This study used surface velocity measurements from Lagrangian drifters to constrain an ocean model, improving model accuracy and predictive capability, which is crucial for studying material transport and dispersion in coastal environments. The results indicate that through optimal interpolation and additional correction for wind-induced velocity component, the discrepancy between observations and model simulations was significantly reduced. However, this still has some shortcomings. For example, the description of the drifters is inadequate, and the methods and data selection for model assessment are not rigorous enough. Therefore, minor revisions are required for this manuscript.
Major comments:
- Line 106 in page 5, the author mentions that all drifters were equipped with an anchor to allow them to drift with surface currents. However, it is not specified whether there were sensors to monitor the presence of the anchor or if there was an assessment of the anchor's stability in the marine environment.
- In the process of optimizing model evaluation, this study extensively utilized fused data sources to assess fusion outcomes. However, such an evaluation process may not objectively reflect the effectiveness of the fusion method and the characteristics of the real ocean current field. Given the scarcity of high-resolution observational data in the study area, buoy data can be partitioned into training and validation sets. The "cross validation" method mentioned at line 285 is an effective approach for dataset partitioning, which could be considered as a core method to extend across various stages of model evaluation, illustrated through figures and charts.
- Line 285 in page 12, the author mentioned the "cross validation experiment," where one drifter was used for model optimization and the others for validation. However, the cross-validation method imposes high requirements on the randomness and independence between the training and validation sets, often employing random sampling. Can simply selecting one drifter as the training set meet these requirements? For example, considering that drifters released during the same period exhibit highly similar and repetitive trajectories due to minimal differences in release times and geographic distances, would the cross-validation method remain effective in such scenarios?
Minor comments:
- Line 103 in page 5, the construction of laboratory-made drifters was described. It would be beneficial to also introduce the construction of Nomad drifters and provide a comparison between these two types of drifters.
- Line 106 in page 5, the author mentioned that all drifters were equipped with an anchor of 0.5 m long positioned in the water layer between 0.8 and 1.3 m depth. Would it be feasible to calculate the overall center of buoyancy depth, including the anchor?
- Line 108 in page 5, the author mentions that observed surface current velocities were estimated from the drifter trajectories. Please describe the specific method used.
- In Figure 1,would it be better to align the display area of Figure 1b with the measurement area outlined in Figure 1a?
- Issues with image consistency. For example, the image sizes and font sizes of axis labels in Figures 1a and 1b are inconsistent. Additionally, the positioning of subplot identifiers in Figures 1 and 9 is inconsistent.
Citation: https://doi.org/10.5194/egusphere-2024-176-RC2 - AC2: 'Reply on RC2', Sloane Bertin, 06 Jun 2024
-
RC3: 'Comment on egusphere-2024-176', Anonymous Referee #3, 31 May 2024
The paper proposes a method to enhance coastal circulation modelling by integrating Lagrangian drifter data with a high-resolution numerical model using Optimal Interpolation (OI). OI is appealing for its simplicity although more sophisticated DA schemes exist. This fusion significantly improves the accuracy of simulating surface current velocities and tracer dispersion in the Dover Strait, reducing model errors by approximately 50%. Authors propose that the improved modelling technique can be applied to develop better tools for monitoring and decision-making in coastal ocean environments. Interestingly a wind-forcing correction scheme provides a significant uplift to model performance.
I believe the paper is of interest to the community but minor revisions are required to improve the clarity of the paper.
- Authors state that “The wind significantly affects the local circulation” in the region. However, results indicate that the model reconstruction is not sensitive to the number of ensembles used (with ensembles generated from perturbing wind forcings). Does this indicate that the assimilation scheme is not sensitive to the number of ensembles used or that perturbing winds do not significantly impact model trajectory? The domain is described as a “tide dominated basins” in line 254 but later results suggest that modifying the wind-forcing scheme has a significant impact on model outputs.
- There is a concern that the data being assimilated into the model is also being used to evaluate the performance of the model. Does this adversely impact the evaluation and bias towards assimilation schemes that weigh more heavily towards observation data? A cross-validation scheme is implemented where only one drifter is assimilated but it isn’t obvious that these are independent. Are there other independent datasets such as in-situ sensors that could be used to confirm the robustness of the findings?
- Does the wind correction scheme in (3) require information on drifter & model data for the entirety of the period? i.e. are corrections being made at time t using information available at time t+1?
Citation: https://doi.org/10.5194/egusphere-2024-176-RC3 - AC3: 'Reply on RC3', Sloane Bertin, 06 Jun 2024
-
RC4: 'Comment on egusphere-2024-176', Anonymous Referee #4, 15 Jun 2024
Review of the Manuscript "Fusion of Lagrangian drifter data and numerical model outputs for improved assessment of turbulent dispersion" by Sloane Bertin, Alexei Sentchev, and Elena Alekseenko
This paper presents a method to enhance coastal circulation modeling and tracer dispersion simulations in the Dover Strait. It combines Lagrangian drifter data with a high-resolution coastal circulation MARS model, using Optimal Interpolation (OI) to integrate drifter observations into model outputs, significantly improving the accuracy of simulated velocity fields. The results show a significant improvement (by ~50%) with the fusion of model outputs, which is quite promising. The manuscript is well-written and concise. I recommend it for acceptance with minor revisions.
Concerns:
-
Discussion of 3D Processes: The study is based on a 2D model, which ignores all baroclinic processes. It is necessary to include a brief discussion in section 2.1 about 3D processes, with references, and clearly state that 3D processes can be ignored compared to tidal currents.
-
Stokes Drift: You mention that the Stokes drift impacts the drift and is ignored. You need to provide estimated values of surface Stokes drift based on wind speed or available modeled results (e.g., WWIII results from Ardhuin Fabrice’s group).
-
Drifter Parameters: You listed some parameters of the drifter used in this study. However, it is unclear whether the drifter measured surface trajectories or averaged depth trajectories. Please clarify this.
-
Relevance of Paragraph L30-L40: This paragraph does not seem closely related to the study. Your study is mostly based on a 2D model focusing on tidal dynamics in coastal areas. The paragraph discusses mesoscale and submesoscale processes in the ocean, which are mainly associated with 3D baroclinic processes. Consider revising or removing this paragraph for better alignment with the study's focus.
Citation: https://doi.org/10.5194/egusphere-2024-176-RC4 - AC4: 'Reply on RC4', Sloane Bertin, 17 Jun 2024
-
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-176', Anonymous Referee #1, 26 Feb 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-176/egusphere-2024-176-RC1-supplement.pdf
-
AC1: 'Reply on RC1', Sloane Bertin, 15 Mar 2024
Dear reviewer,
Thank you very much for dedicating time to review our manuscript, and for your clear and pertinent remarks. Please find our point-by-point response concerning the major comments. Concerning the minor comments, all of them have been considered and clarified in the revised version of the manuscript (given in red).
Major comments
- Line 125 in page 5, authors just present the tidal conditions for model’s boundary. What are the other boundary conditions for simulation? Such as gradient, Clamped, Flather or other boundary conditions?
The numerical model utilizes nested configurations with progressive resolutions: (i) 2 km covering the Northeast Atlantic (level 0), (ii) 700 m at the regional scale, encompassing the English Channel (level 1), and (iii) 250 m for the Eastern English Channel (level 2). This nesting technique enables the accurate capture of interactions between large-scale and small-scale processes. This enables the transfer of all resolved fields from lower resolution levels to the open boundaries of higher resolution levels.
The model accounts for kinematic free-surface and bottom boundary conditions, contingent upon friction terms (Lazure and Dumas, 2008). The turbulence closure employed in the model follows the approach described in Gaspar et al. (1990).
All these details will be added to the part 2.3 ‘Current velocity from numerical model’.
- Two one-year long model runs are implemented in this research. Can authors provide more detailed modeling settings to make model stable in such long simulation.
Comprehensive information regarding model equations, the coupling of barotropic and baroclinic modes, model discretization, solving methods, computational stability according to CFL criterion (table 1, Lazure and Dumas, 2008), and costs are meticulously outlined in Lazure and Dumas (2008). To maintain CFL stability, the modeling timestep was set to 30 seconds for the level 2 model.
All these details will be added to the part 2.3 ‘Current velocity from numerical model’.
- In Figure 5, authors just present the model absolute errors at S2-4. May authors present the absolute errors using “Box-Whisker” plot over all drifters of S2?
Thank you for this advice. I modified Figure 5 and used box-whiskers to present the model absolute error for all the drifters during S1 (Fig. 5a) and S2 (Fig. 5b). Consequently, I modified the descriptive paragraph L266-280 and legend.
- Line 280-281, authors describe that the drifter S2-4 is well reproduced. Can authors plot all drifters of S1and S2 in Figure 6 to make reader directly understand the simulation results. There is similar problem in Figure 7. In Figure 7, authors use drifters S1-2 and S2-1, why is it different from Figure 6? Can authors also present all drifters of S1 and S2 in Figure 7? Or using “Box-Whisker” plot to present separation distance in Figure 7? Furthermore, how about the wind-corrected trajectories of other drifters in S1 and S2? Can present other drifters in Figure 8? If other OI- or corrected trajectories are similar to S1-1 and S2-1, please describe the related statement of other trajectories in S1 and S2.
I modified Figures 6 and 8 by presenting all the drifters’ trajectories (2 for S1 and 4 for S2) in order to make the reader directly understand the simulation results. I also modified Figure 7 by using box-whiskers to present the separation distance results for all the drifters simultaneously. Consequently, the paragraphs and legends concerning these three figures have been modified.
Thank you,
Best regards.
Sloane Bertin
Citation: https://doi.org/10.5194/egusphere-2024-176-AC1
-
AC1: 'Reply on RC1', Sloane Bertin, 15 Mar 2024
-
RC2: 'Comment on egusphere-2024-176', Anonymous Referee #2, 24 May 2024
General comments:
This study used surface velocity measurements from Lagrangian drifters to constrain an ocean model, improving model accuracy and predictive capability, which is crucial for studying material transport and dispersion in coastal environments. The results indicate that through optimal interpolation and additional correction for wind-induced velocity component, the discrepancy between observations and model simulations was significantly reduced. However, this still has some shortcomings. For example, the description of the drifters is inadequate, and the methods and data selection for model assessment are not rigorous enough. Therefore, minor revisions are required for this manuscript.
Major comments:
- Line 106 in page 5, the author mentions that all drifters were equipped with an anchor to allow them to drift with surface currents. However, it is not specified whether there were sensors to monitor the presence of the anchor or if there was an assessment of the anchor's stability in the marine environment.
- In the process of optimizing model evaluation, this study extensively utilized fused data sources to assess fusion outcomes. However, such an evaluation process may not objectively reflect the effectiveness of the fusion method and the characteristics of the real ocean current field. Given the scarcity of high-resolution observational data in the study area, buoy data can be partitioned into training and validation sets. The "cross validation" method mentioned at line 285 is an effective approach for dataset partitioning, which could be considered as a core method to extend across various stages of model evaluation, illustrated through figures and charts.
- Line 285 in page 12, the author mentioned the "cross validation experiment," where one drifter was used for model optimization and the others for validation. However, the cross-validation method imposes high requirements on the randomness and independence between the training and validation sets, often employing random sampling. Can simply selecting one drifter as the training set meet these requirements? For example, considering that drifters released during the same period exhibit highly similar and repetitive trajectories due to minimal differences in release times and geographic distances, would the cross-validation method remain effective in such scenarios?
Minor comments:
- Line 103 in page 5, the construction of laboratory-made drifters was described. It would be beneficial to also introduce the construction of Nomad drifters and provide a comparison between these two types of drifters.
- Line 106 in page 5, the author mentioned that all drifters were equipped with an anchor of 0.5 m long positioned in the water layer between 0.8 and 1.3 m depth. Would it be feasible to calculate the overall center of buoyancy depth, including the anchor?
- Line 108 in page 5, the author mentions that observed surface current velocities were estimated from the drifter trajectories. Please describe the specific method used.
- In Figure 1,would it be better to align the display area of Figure 1b with the measurement area outlined in Figure 1a?
- Issues with image consistency. For example, the image sizes and font sizes of axis labels in Figures 1a and 1b are inconsistent. Additionally, the positioning of subplot identifiers in Figures 1 and 9 is inconsistent.
Citation: https://doi.org/10.5194/egusphere-2024-176-RC2 - AC2: 'Reply on RC2', Sloane Bertin, 06 Jun 2024
-
RC3: 'Comment on egusphere-2024-176', Anonymous Referee #3, 31 May 2024
The paper proposes a method to enhance coastal circulation modelling by integrating Lagrangian drifter data with a high-resolution numerical model using Optimal Interpolation (OI). OI is appealing for its simplicity although more sophisticated DA schemes exist. This fusion significantly improves the accuracy of simulating surface current velocities and tracer dispersion in the Dover Strait, reducing model errors by approximately 50%. Authors propose that the improved modelling technique can be applied to develop better tools for monitoring and decision-making in coastal ocean environments. Interestingly a wind-forcing correction scheme provides a significant uplift to model performance.
I believe the paper is of interest to the community but minor revisions are required to improve the clarity of the paper.
- Authors state that “The wind significantly affects the local circulation” in the region. However, results indicate that the model reconstruction is not sensitive to the number of ensembles used (with ensembles generated from perturbing wind forcings). Does this indicate that the assimilation scheme is not sensitive to the number of ensembles used or that perturbing winds do not significantly impact model trajectory? The domain is described as a “tide dominated basins” in line 254 but later results suggest that modifying the wind-forcing scheme has a significant impact on model outputs.
- There is a concern that the data being assimilated into the model is also being used to evaluate the performance of the model. Does this adversely impact the evaluation and bias towards assimilation schemes that weigh more heavily towards observation data? A cross-validation scheme is implemented where only one drifter is assimilated but it isn’t obvious that these are independent. Are there other independent datasets such as in-situ sensors that could be used to confirm the robustness of the findings?
- Does the wind correction scheme in (3) require information on drifter & model data for the entirety of the period? i.e. are corrections being made at time t using information available at time t+1?
Citation: https://doi.org/10.5194/egusphere-2024-176-RC3 - AC3: 'Reply on RC3', Sloane Bertin, 06 Jun 2024
-
RC4: 'Comment on egusphere-2024-176', Anonymous Referee #4, 15 Jun 2024
Review of the Manuscript "Fusion of Lagrangian drifter data and numerical model outputs for improved assessment of turbulent dispersion" by Sloane Bertin, Alexei Sentchev, and Elena Alekseenko
This paper presents a method to enhance coastal circulation modeling and tracer dispersion simulations in the Dover Strait. It combines Lagrangian drifter data with a high-resolution coastal circulation MARS model, using Optimal Interpolation (OI) to integrate drifter observations into model outputs, significantly improving the accuracy of simulated velocity fields. The results show a significant improvement (by ~50%) with the fusion of model outputs, which is quite promising. The manuscript is well-written and concise. I recommend it for acceptance with minor revisions.
Concerns:
-
Discussion of 3D Processes: The study is based on a 2D model, which ignores all baroclinic processes. It is necessary to include a brief discussion in section 2.1 about 3D processes, with references, and clearly state that 3D processes can be ignored compared to tidal currents.
-
Stokes Drift: You mention that the Stokes drift impacts the drift and is ignored. You need to provide estimated values of surface Stokes drift based on wind speed or available modeled results (e.g., WWIII results from Ardhuin Fabrice’s group).
-
Drifter Parameters: You listed some parameters of the drifter used in this study. However, it is unclear whether the drifter measured surface trajectories or averaged depth trajectories. Please clarify this.
-
Relevance of Paragraph L30-L40: This paragraph does not seem closely related to the study. Your study is mostly based on a 2D model focusing on tidal dynamics in coastal areas. The paragraph discusses mesoscale and submesoscale processes in the ocean, which are mainly associated with 3D baroclinic processes. Consider revising or removing this paragraph for better alignment with the study's focus.
Citation: https://doi.org/10.5194/egusphere-2024-176-RC4 - AC4: 'Reply on RC4', Sloane Bertin, 17 Jun 2024
-
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 347 | 155 | 34 | 536 | 13 | 16 |
- HTML: 347
- PDF: 155
- XML: 34
- Total: 536
- BibTeX: 13
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Cited
1 citations as recorded by crossref.
Sloane Bertin
Alexei Sentchev
Elena Alekseenko
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2897 KB) - Metadata XML