A Regional Operational Model for the North East Atlantic: Model Configuration and Validation
Round 1
Reviewer 1 Report
This paper presents the efforts to achieve a relatively high resolution operational model to hindcart and forecast circulation and physical oceanography parameters including water temperature and salinity in the northeast Atlantic that including Ireland’s territorial waters. The Regional Ocean modeling System (ROMS) with a spatial resolution of 1.1 km was forced by the global ECMWF wind and outputs from a lower resolution NEMO model along the open boundaries. The model was verified versus satellite SST and SSH as well as Argo, CTD, tide gage data and maps of EKE and MKE. The verification in general showed an appropriate model performance. The paper is well-written with a reasonable and standard approach toward the main idea and I would like to recommend this paper for publication in the Journal of Marine Science and Engineering after applying “minor modifications” as follows:
General Comments:
- Regarding the fact that the study area and generally the north Atlantic is affected by large surface waves throughout the year, especially during winter time (please see Allahdadi et al., 2019: Predicting ocean waves along the US east coast during energetic winter storms: sensitivity to whitecapping parameterizations), including the effect of waves on ocean mixing could be important in simulating the salinity and temperature across the water column. Although the authors showed that the model skill in simulating SST and the average Argo and CTD temperature profiles is relatively high, still remains the question that how the simulated mixed layer compares with reality? In this regard, wave effect could be very important (Please see for example Aijaz et al., 2017: Nonbreaking wave-induced mixing in upper ocean during tropical cyclones using coupled hurricane-ocean-wave modeling). The reviewer understand that at this stage, including wave mixing was not in the scope of the study, but still the authors need to add a short paragraph to mention the probable effect of waves on the mixing in the study area with appropriate references especially when mentioning differences between model and observations.
Use of K-profile parametrization for including the vertical eddy viscosity and diffusivity is a popular approach in ROMS. However, the vertical mixing and thereby water temperature/salinity across the water column and also ocean currents inside the upper mixed layer can be significantly affected by choice of the vertical turbulent closure and the associated parameters(Please see Allahdadi and Li., 2017, a and b, see the complete citation below). In several studies this parameter has been used to tune the vertical profile of currents and temperature/salinity (for example Keen and Glenn, 1998 ad 1999, in addition to Allahdadi and Li.2017 a,b). So, it would be relevant if the authors pay more attention to this very important modeling issue in different parts of the paper including where they introduce KPP model. Also, they need to mention this issue as a potential source of deficiencies between modeling results and observations especially temperature and salinity. Please support your dissuasions with appropriate reference(s).
Allahdadi, M. N; Li, C, 2017a. Numerical Simulation of Louisiana Shelf Circulation under Hurricane Katrina Journal of Coastal Research, https://doi.org/10.2112/JCOASTRES-D-16-00129.1
Allahdadi, M.N; Li, C, 2017b. Effect of Stratification on Current Hydrodynamics over the Louisiana shelf during Hurricane Katrina. Water Science and Engineering, 10(2), 154-165
Detail comments
(Ln stands for Line)
Ln 26: “0.17 for salinity” : a unit for salinity is required
Ln 37, 46, and 47: Please show “Nordic Sea” , “Celtic Sea”, and “Irish Sea” on the map of figure1. Please also show any other geographical names/regions related to the study area on this map if they were mentioned in the text but not shown on the map.
Ln 72-75: These lines are not consistent with the title of the paper. Based on the title, the main objective of the paper is preparing an operational ocean model for the study area. Studying the ICC is a secondary objective that should be mentioned after pointing out the major objective!
Ln 167: “with 3 hour frequency” should be “with 3 hour temporal step”
Ln 189: correct
Ln 192: Did you also introduce river water temperature/salinity ( very small or zero) to the model? If yes, include them in the table. If not explain why?
Ln 236: Is SST data the daily averaged or measured at a specific time each day? Please make it clear
Ln 259-260: Argo data are measured at specific times, not as a daily average. Why model data were daily-averaged for comparing with Argo data? Why the outputs at the exact time as Argo data were not extracted?
Ln 283: define all parameters in the Coriolis factor formulation
Ln 336-337 (Figure4) and 320-321: comparing the pattern of SST overestimation with Figure 1 that shows model bathymetry shows that the regions of SST overestimation are consistent with the very locally deep regions, especially the deep trench between coasts of France and Spain. So, the SST deficiencies could be a model resolution problem, either the model spatial step or bathymetry data. Please discuss this effect in relation to these type of resolutions.
Ln 366: Although a reference was presented for surge calculation, some main principals of deducing the surge from the total water level should be mentioned in the text,
Ln 391-394: What quantity was compared from Argo to the model? Was it the depth-averaged temperature/salinity or values at a specific depth? The same questions for CTD data
415-416: Why model failed to properly simulate the mixing here? Is it because of the vertical eddy diffusivity approach that is used? Boundary condition? Or something else. Please discuss
Figure 7. Left panel: what is the colorbar title and unit? Temperature? Salinity? right panel: a higher quality location map is required
Ln 453-456: is there any reference/evidence showing the underestimation of river discharges used in the present modeling?
Ln 470-471, 481-482: By looking at different panels in figure9, it is not easy to conclude that NEA_ROMS results are consistent with observations than the CMEMS. Please present a quantitative metric showing this merit.
Ln 497-499: more details on calculating ICC maps are required.
Ln 507-508: Why currents are more intense in 2019? Is it due to stronger atmospheric forces? Stronger tidal forces?
Figure 10: Current vector are not clear. Please re-plot the figure with more clear current vectors.
Author Response
Dear Editor/Reviewer
Authors wish to thank you very much for considering our paper for publication pending suitable minor revision. We wish also to thank the anonymous reviewers for their constructive comments, which helped us to improve the manuscript. Below, we address all comments point-by-point, discussing the subsequent modifications. All suggested changes have incorporated.
Thank you
On behalf of all co-authors,
Hazem Nagy
Reviewer (1)
General Comments:
1- Regarding the fact that the study area and generally the north Atlantic is affected by large surface waves throughout the year, especially during winter time (please see Allahdadi et al., 2019: Predicting ocean waves along the US east coast during energetic winter storms: sensitivity to white capping parameterizations), including the effect of waves on ocean mixing could be important in simulating the salinity and temperature across the water column. Although the authors showed that the model skill in simulating SST and the average Argo and CTD temperature profiles is relatively high, still remains the question that how the simulated mixed layer compares with reality? In this regard, wave effect could be very important (Please see for example Aijaz et al., 2017: Nonbreaking wave-induced mixing in upper ocean during tropical cyclones using coupled hurricane-ocean-wave modeling). The reviewer understand that at this stage, including wave mixing was not in the scope of the study, but still the authors need to add a short paragraph to mention the probable effect of waves on the mixing in the study area with appropriate references especially when mentioning differences between model and observations.
A:. Authors would like to thank you very much for your important notes about model errors came from large surface waves. So Authors have added and highlighted in the MS the following sentences and paragraphs with the appropriate mentioned references:
In section 3.1 simulating SST. The SST deficiencies in the Bay of Biscay could be due to the model spatial resolution or inaccuracies in the bathymetry data. This region features very deep waters and steep topography. Also, the north Atlantic is affected by large surface waves throughout the year, especially during winter time as mentioned by [74]. The surface wave is not included in our model and nonbreaking wave‐induced mixing effect could be important in simulating SST as described in [75]. This may lead to significant cooling of the simulated SST [75].
In section 3.3 ‘Also, the absence of the wave effect in our model could be an important reason for poor representation of the salinity and temperature across the water column as described in [74].
Use of K-profile parametrization for including the vertical eddy viscosity and diffusivity is a popular approach in ROMS. However, the vertical mixing and thereby water temperature/salinity across the water column and also ocean currents inside the upper mixed layer can be significantly affected by choice of the vertical turbulent closure and the associated parameters(Please see Allahdadi and Li., 2017, a and b, see the complete citation below). In several studies this parameter has been used to tune the vertical profile of currents and temperature/salinity (for example Keen and Glenn, 1998 ad 1999, in addition to Allahdadi and Li.2017 a,b). So, it would be relevant if the authors pay more attention to this very important modeling issue in different parts of the paper including where they introduce KPP model. Also, they need to mention this issue as a potential source of deficiencies between modeling results and observations especially temperature and salinity. Please support your dissuasions with appropriate reference(s).
Authors have included your important note in section 2 as you recommended with the appropriate references in introducing KPP model as follow
‘’ Vertical diffusion is calculated using the K-profile parameterization (KPP) [42] and a modified Jacobian method [43] is used to calculate horizontal pressure-gradient forces. The associated parameters with KPP vertical turbulent closure scheme has been enhanced to tune the vertical profile of currents, temperature and salinity by [44-47]’’.
In section 3.3 ‘’ This may be as indirect effect of model excess vertical mixing due to use of the associated parameters with KPP vertical turbulent closure scheme as described in [46,47]. This parameters has been used to tune the vertical profiles of currents, temperature and salinity [46,47]
In section 3.4 ‘’ Other possibility, as previously mentioned, may be due to the use of KPP vertical turbulent closure scheme inside ROMS model which can affect the vertical mixing as described in [46,47]. This issue is a potential source of deficiencies between model output and observations.
In conclusions ‘’ The use of KPP vertical turbulent closure scheme in ROMS model can affect the vertical mixing of the salinity and temperature profiles as described in [46,47]. This issue was a potential source of deficiencies between model predictions and observations.
Allahdadi, M. N; Li, C, 2017a. Numerical Simulation of Louisiana Shelf Circulation under Hurricane Katrina Journal of Coastal Research, https://doi.org/10.2112/JCOASTRES-D-16-00129.1
Allahdadi, M.N; Li, C, 2017b. Effect of Stratification on Current Hydrodynamics over the Louisiana shelf during Hurricane Katrina. Water Science and Engineering, 10(2), 154-165
Detail comments
(Ln stands for Line)
2- Ln 26: “0.17 for salinity” : a unit for salinity is required
A:. Authors have include salinity unit and highlighted change in ln 26 to be ‘’while it is 0.17 PSU for salinity’’.
3- Ln 37, 46, and 47: Please show “Nordic Sea” , “Celtic Sea”, and “Irish Sea” on the map of figure1. Please also show any other geographical names/regions related to the study area on this map if they were mentioned in the text but not shown on the map.
- Authors have changed Fig.1 with more names and regions such as Nordic Seas, Celtic Sea, Irish Sea, North Sea, Bay of Biscay, English Channel, Norway and Iceland. Also we have refined, changed the map resolution and colormap for better contrast. The authors also changed the place of Figure 1 to be directly after the first paragraph. So the readers can be familiar with places and locations we mentioned in MS. Thanks for your important note. We hope it’s much better now.
4- Ln 72-75: These lines are not consistent with the title of the paper. Based on the title, the main objective of the paper is preparing an operational ocean model for the study area. Studying the ICC is a secondary objective that should be mentioned after pointing out the major objective!.
Authors have changed and highlighted Ln 72-75 to be ‘This study presents the set-up and results from the validation of a high-resolution numerical operational model (hereafter called NEA_ROMS) developed at Irish Marine Institute Validation of NEA_ROMS was carried out for the time period from January 2016 until December 2019 using observational data from various sources. The model has been compared with satellite SST, tide gauges, Argo, and CTD temperature and salinity profiles. In addition, the Geostrophic and Ekman Current Observatory GEKCO surface data was used to validate model velocity fields. Also, this study will focus on the Irish coastal waters and the authors will examine the representation of the Irish coastal current (ICC) in the model. Section 2 describes the model implementation, and nesting procedures, Section 3 presents the validation against observational data, Section 4 describes the model results related to the ICC, and Section 5 provides the discussion and conclusions.
5- Ln 167: “with 3 hour frequency” should be “with 3 hour temporal step”
A:. Authors have corrected and highlighted 3 hour frequency with 3-hourly temporal step
6- Ln 189: correct
A:. Authors have corrected and highlighted ln 189 .
7- Ln 192: Did you also introduce river water temperature/salinity ( very small or zero) to the model? If yes, include them in the table. If not explain why?
A:. Authors have include river salinity to be zero. But we didn’t introduce river temperature for two main reasons data availability and model stability. In ROMS there is an option to not introduce the river temperature. Authors have included the river salinity in the table and wrote it in the text . ‘The NEA_ROMS rivers salinity was set to zero and the rivers temperature was not prescribed.
8- Ln 236: Is SST data the daily averaged or measured at a specific time each day? Please make it clear .
A:. As mentioned in the ODYSSEA product document https://resources.marine.copernicus.eu/documents/QUID/CMEMS-SST-QUID-010-025.pdf
‘’ SST data valid for a particular day (from previous day 12:00 to current day 12:00) so should be daily averaged. Authors have corrected and highlighted Ln 236 to be
‘’The product consists of daily averaged SST values. The data valid for a particular day (from previous day 12:00 to current day 12:00) at a horizontal resolution of 0.02o x 0.02o for a domain covering the European North West Shelf and Iberia-Biscay-Irish Seas’’.
10-Ln 259-260: Argo data are measured at specific times, not as a daily average. Why model data were daily-averaged for comparing with Argo data? Why the outputs at the exact time as Argo data were not extracted?
A:. Yes you are absolutely right it’s a mistake in the text we are using the model hourly-snapshots files for ARGO comparison. The Authors would like to thank your very much about this note and comment, We have corrected and highlighted in text to be hourly snapshots model output file corresponding to the date the ARGO profile was acquired and the grid point closest in location to the ARGO profile
11-Ln 283: define all parameters in the Coriolis factor formulation
A:. Authors have defined and highlighted Ln 283 to be ‘Where is the absolute dynamic topography anomaly in meters, g is the gravity acceleration in m/s2 and f=2Ω sinɸ, where Ω = 2π/T is the earth angular velocity in sec-1 , T is the earth periodic time = 86400 sec (1 day in sec) and ɸ is the latitude in degrees.
12- Ln 336-337 (Figure4) and 320-321: comparing the pattern of SST overestimation with Figure 1 that shows model bathymetry shows that the regions of SST overestimation are consistent with the very locally deep regions, especially the deep trench between coasts of France and Spain. So, the SST deficiencies could be a model resolution problem, either the model spatial step or bathymetry data. Please discuss this effect in relation to these type of resolutions.
Authors would like to thank reviewer for his/her note. A paragraph has been added as you suggested
The SST deficiencies in the Bay of Biscay could be due to the model spatial resolution or inaccuracies in the bathymetry data. This region features very deep waters and steep topography. Also, the north Atlantic is affected by large surface waves throughout the year, especially during winter time as mentioned by [74]. The surface wave is not included in our model and nonbreaking wave‐induced mixing effect could be important in simulating SST as described in [75]. This may lead to significant cooling of the simulated SST [75].
13- Ln 366: Although a reference was presented for surge calculation, some main principals of deducing the surge from the total water level should be mentioned in the text,
A:. We hope we well understood your point, I mentioned and highlighted in the text ‘. Surge is caused by atmospheric pressure and wind. In order to obtain the surge signal we are removing tides. The tidal signal in the Sea Surface Height (SSH) data was dominated by three semi-diurnal constituents (M2, S2, N2,) and three diurnal constituents (K1 O1 and Q1)
14-Ln 391-394: What quantity was compared from Argo to the model? Was it the depth-averaged temperature/salinity or values at a specific depth? The same questions for CTD data
A:. We are comparing ARGO/CTD in the same way. Using hourly averaged model output file corresponding to the date the ARGO profile. Acquired the grid point closest in location to the ARGO profile. We are comparing values at a specific depth find bias and RMSE.
15-415-416: Why model failed to properly simulate the mixing here? Is it because of the vertical eddy diffusivity approach that is used? Boundary condition? Or something else. Please discuss
A:. Authors have added and highlighted explanation (in ln 415-416) as you suggested in your comments ‘This may be as indirect effect of model excess vertical mixing due to the associated parameters with KPP vertical turbulent closure scheme as described in [46,47]. This parameters has been used to tune the vertical profiles of currents, temperature and salinity [46,47]
16- Figure 7. Left panel: what is the colorbar title and unit? Temperature? Salinity? right panel: a higher quality location map is required
A:. Authors have put PSU on the colorscale to be clear for the readers.
A:. Authors have changed Figure 8 a-d with different map and color scale for Bias and RMSE also we increased the resolution to avoid any misleading. Hope it looks better now. Same happened for Figure 6 a-d to be consistent.
17-Ln 453-456: is there any reference/evidence showing the underestimation of river discharges used in the present modeling?
A:. Thanks a lot for your good question. Unfortunately we don’t have any reference or evidence showing the underestimation. It’s just our prospection because we are using climatological river values not real time. But certainly we are going to investigate this important issue. Any way , we have added to the text ’ Other possibility, as previously mentioned, may be due to the use of KPP vertical turbulent closure scheme inside ROMS model which can affect the vertical mixing as described in [46,47]. This issue is a potential source of deficiencies between model output and observations.
18- Ln 470-471, 481-482: By looking at different panels in figure9, it is not easy to conclude that NEA_ROMS results are consistent with observations than the CMEMS. Please present a quantitative metric showing this merit.
A;. For Ln 470-471 the authors think you are right, it’s hard to conclude that NEA_ROMS reproduces better mean circulation than CMEMS so we have corrected and highlighted to be ‘’Both NEA_ROMS and CMEMS Global models produce similar mean circulation to the GEKCO currents’’
A:. For Ln 481-82 the authors have corrected highlighted to be Figure 9e-g represents the mean EKE fields for GEKCO data, NEA_ROMS, and CMEMS global respectively. The basin averaged EKE for GEKCO data, NEA_ROMS and CMEMS were 0.0063 m2 s-2 0.0153 m2 s-2 and 0.0147 m2 s-2 respectively. The EKE comparison suggests the NEA_ROMS has the highest EKE field.
19- Ln 497-499: more details on calculating ICC maps are required.
A:. Authors have added and highlighted more details on calculating ICC maps ‘’ We are using the model output hourly snapshots velocity field. To produce annual maps we get velocities away from ICC region out of the picture.
20- Ln 507-508: Why currents are more intense in 2019? Is it due to stronger atmospheric forces? Stronger tidal forces?
A:. Most probably due to strong wind. Unfortunately we have no evidence for this hypothesis. we just know the ICC is driven by the front .That’s why We didn’t add in the text ‘’This may be due to strong wind in 2019’’.
21- Figure 10: Current vector are not clear. Please re-plot the figure with more clear current vectors.
A:. Authors have re-plot Fig 10 with more clear vectors. Also we have changed the colormap for better contrast. Authors would like to thank you very much because this modification will enrich the quality of our MS.
Author Response File: 
Author Response.docx
Reviewer 2 Report
First of all I would like to apologize for the long delay of my review.
The submitted manuscript presents a new operational implementation of ROMS over the North East Atlantic. The paper is very well written and clear in its porposes and methodologies.
The introduction is clear and reports the most significant literature on the area and processes associated. My only observation is on the paragraph starting on line 33: in this paragraph there are a lot of references to currents and front. It could be helpful for readers not familiar with the area to show the locations and path of the main ones on figure 1 or to provide a second map with the oceanographycal features.
The model implementation is described clearly and in details and validation is outstandingly done with a lot of observations from different sources of data. The model do show some significant BIAS especially for salinity but it is well inside an expected and acceptable range.
It could be interesting to add in the conclusion section a paragraph on the future application of this implementation.
I can just list some very minor observations on the plots:
- figure 1: there is no need to show every longitude and latitude degree. The grid on the map could be more refined. Some of the labels are not easily read (i.e. Ireland, Shannon) try to use a different color for text.
- figure 6: figure 6 a-d show bias and RMSE for T and S. I would suggest to use different color scale for BIAS and RMSE. In the present form since the ranges between figures are quite different having a single color scale could be misleading.
- Figure 7: even if it is clear that the colors cale is salinity it should be specified on plot.
I want to say that it is rare in my experience to review such an high quality manuscript on the first round of review. I was especially pleased to read a detail description of the model set up and characteristics and to see such an extensive validation!
Author Response
Dear Editor/Reviewer
Authors wish to thank you very much for considering our paper for publication pending suitable minor revision. We wish also to thank the anonymous reviewers for their constructive comments, which helped us to improve the manuscript. Below, we address all comments point-by-point, discussing the subsequent modifications. All suggested changes have incorporated.
Thank you
On behalf of all co-authors,
Hazem
Reviewer (2)
First of all I would like to apologize for the long delay of my review.
The submitted manuscript presents a new operational implementation of ROMS over the North East Atlantic. The paper is very well written and clear in its porposes and methodologies.
The introduction is clear and reports the most significant literature on the area and processes associated.
A:. Authors would like to thank you very much for your kind words .
My only observation is on the paragraph starting on
- 1- line 33: in this paragraph there are a lot of references to currents and front. It could be helpful for readers not familiar with the area to show the locations and path of the main ones on figure 1 or to provide a second map with the oceanographycal features.
- Authors have changed Fig.1 with more names and regions such as Nordic Seas, Celtic Sea, Irish Sea, North Sea, Bay of Biscay, English Channel, Norway and Iceland. Also we have refined, changed the map resolution and colormap for better contrast. The authors also changed the place of Figure 1 to be directly after the first paragraph. So the readers can be familiar with places and locations we mentioned in MS. Thanks for your important note. We hope it’s much better now.
- 2- The model implementation is described clearly and in details and validation is outstandingly done with a lot of observations from different sources of data. The model do show some significant BIAS especially for salinity but it is well inside an expected and acceptable range. It could be interesting to add in the conclusion section a paragraph on the future application of this implementation.
A:. Authors have added and highlighted paragraph on the future application of this implementation’’
Future work requires the use of high frequency near real time river discharges. The inclusion of data assimilation in the NEA_ROMS operational system is considered to be the next logical step in its ongoing development. Recent applications also include the modelling studies of the biogeochemical cycling. Also, indices supporting the implementation of the Marine Strategy Framework Directive (MSFD) are usually obtained from in-situ data. This is a major difficulty in oceanic areas where data are scarce. Validated numerical models can fill this gap. Tools will be developed to obtain indices of interest directly from model results, e.g. areas of upwelling, fronts, eddy index, primary production, trophic status, or even conditions for propagation of noise.
I can just list some very minor observations on the plots:
4- - figure 1: there is no need to show every longitude and latitude degree. The grid on the map could be more refined. Some of the labels are not easily read (i.e. Ireland, Shannon) try to use a different color for text.
- Authors have changed Fig.1 with more names and regions such as Nordic Seas, Celtic Sea, Irish Sea, North Sea, Bay of Biscay, Norway and Iceland. Also we have refined, changed the map resolution and colormap for better contrast. Hope it’s much better now.
figure 6 a-d show bias and RMSE for T and S. I would suggest to use different color scale for BIAS and RMSE. In the present form since the ranges between figures are quite different having a single color scale could be misleading.
Authors Have changed Figure 6 a-d with different color scale for Bias and RMSE also we increased the resolution to avoid any misleading. Authors have changed Figure 8 a-d with different clear color scale for Bias and RMSE also we increased the resolution to avoid any misleading. Hope it’s better looking now.
6- - Figure 7: even if it is clear that the colors cale is salinity it should be specified on plot.
A:. Authors have put PSU on the colorscale to be clear for readers.
7- I want to say that it is rare in my experience to review such an high quality manuscript on the first round of review. I was especially pleased to read a detail description of the model set up and characteristics and to see such an extensive validation!
A:. Thank your very much for your nice words and your constructive/encouragement comments.
Author Response File: Author Response.docx
