Conceptual Approach for Positioning of Fish Guidance Structures Using CFD and Expert Knowledge

Round  1

Reviewer 1 Report

GENERAL COMMENTS

The submission presents a procedure aiming at providing a practical design solution for positioning fish guidance structures (FGS) such as bar racks in hydro power plants. The procedure essentially relies on full-scale 3D CFD of the flow across the unmodified site, whose results are confronted to the fish biological specificities in order to validate, unvalidate or improve pre-determined positions of the future fish-guidance structure.

The proposed procedure is conveniently described, adequately considering fish fauna as the first entry parameter which will condition the subsequent decision-making process primarily based on the CFD results and on secondary considerations.

However, the structure of the submission is not adequate. I reckon that the case study should be separated from the Methods Section, and that the Results Section should better be written as an application of the method to the case study, reflecting the proposed workflow.

While the submission title and content mention "optimal positioning", the procedure is far from adopting a definite optimization process. It is also not proven that the validated FGS positioning will actually comply with the biological constraints. The hydraulic effect of the bar rack to be installed is only qualitatively adressed, relying on rather unrelated laboratory results, so neither its influence on the upstream nor on the downstream flow are evaluated.

On the numerics side, the specifics of the 3D simulation are not adequately adressed. More information is needed on the case setup, and the results should be presented in detail, and commented, before discussing the bar rack positioning. If available, a confrontation with existing field results, even from a very coarse distibution, should be provided.

On the whole, the procedure itself lacks robustness. First, in the test case, three tentative positions for the FGS are given, according "configuration assessments" refering to two German language documents that are not accessible (refs. 60 and 61). For the procedure to be reliable, a minimum number of criteria should be provided that support these preliminary choices. Alternatively, qualitative explanations are needed, at least. Then, the procedure ends by invalidating 2 of these positions and only slightly modifying the remaining one to provide a more adequate choice. As a final step, it is said that in any case, not taking into account the FGS in the CFD is eligible because the bar rack will necessarily improve fish guidance. As a matter of fact, one could add that not taking into account the bypass (which is what is done here) and its discharge acts in the same fashion. However, if a precise result is sought these two elements should be more adequately considered as their influence may favour other FGS positioning that would be invalidated otherwise.

The deliberate omission of the FGS in the CFD raises another problem: although the objective is to determine a suitable positioning of the FGS, the authors very briefly discuss existing knowledge on the effect of bar racks at the end of Section 3 only. Since the FGS is the central element in the procedure, this discussion shall be more extensively developped prior to the Results section, and shall contain a better argumented review of how FGS do modify a flow. In my opinion this part shall appear in Section 2 (Methods) as predicting the effect of the FGS is fully part of the methodology.

An additional critic can be raised: given an existing hydropower plant with no existing FGS, the CFD can be carried out prior to discussing any possible positioning of the FGS, and provide an overall view of the flow across the head race canal. Then, velocity maps can be analysed to identify which zones may be considered for installing FGS, and which may be not. Such an approach would be a lot more comprehensive than restricting the choice to pre-selected locations, as it would be done at no additional computational cost since the same CFD is already needed by the procedure. The only additional effort would then be to identify the range of positioning options that are to be considered.

I also consider that the proposed approach can only be seen as the initialization stage of a proper optimization strategy, or as a one-way design scheme. In my opinion, if optimality is sought, then it shall be based on an iterative scheme starting from the undisturbed flow and taking into acount, even in a restricted way, the presence of the FGS and, at least, the bypass.

I should add that nowadays, any (serious) consulting company that would be subcontracted to design part or all of an hydropower plant uses combined CAD, structural and CFD studies. On that respect there is no real novelty in the proposed procedure.

However, the authors effort to provide a definite FGS positioning procedure must be acknowledged as this topic needs to be robustly addressed in the hydropower community. I therefore strongly advise the authors to enrich their approach by making it more comprehensive, less arbitrary, and with fewer approximations.

There are too many publicly unavailable, and German language, references in the submission. When such references are cited, the authors should provide brief descriptions of their topics and results, especially when they serve as design or decision tools. Such short reviews would be valuable to the community if they were included in the Introduction.

From my point of view the submission should be deeply revised before it is accepted in a scientific journal. Please consider my detailed comments below.

DETAILED COMMENTS

Abstract

A brief description of the procedure should be given in the abstract. In the submission it is only stated that it relies on fauna, structural and hydraulic conditions, but there is no indication of the type of approach that is used. Please add a few sentences explaining how the approach does use biologic, kinematic and structural constraints to determine which zones of the unperturbed flow may be suitable for installing an FGS.

It is also misleading as it is said that the FGS is not considered in detail in the numerical model. In fact the FGS is not considered at all in the CFD, which makes a significant difference. The authors should rephrase this sentence as it does not adquately characterize their approach.

1-Introduction

p1 last line : There have been more than these 4 references dealing with downstream migration devices recently, don't restrict the bibliographic references on that subject to the last 2 years (unless the authors only consider biological tests here, in this case it must be said).

p2 l17: Why using the terme 'louvres' instead of 'louvers' as in most of the english language litterature? It is not an issue, but is it useful?

p2 l18: Louvers are generally oriented perpendicular to the main flow direction, but not systematically. The description should be less restrictive here.

p2 l19: The authors should clearly define what they mean by rack axis, e.g. by refering to corresponding axis systems that can be added to Fig 1a. Alternatively this type of rack can be described as having vertical bars oriented normal to the rack plane.

Fig.1 : b,c and d subfigures do not clearly show both the flow direction and bar orientation. On the 3D views, the up- and downstream sides of the water block look like solid frontiers that are attached to the racks. I suggest either enlarging the water block to longer up- and downstream distances to the rack and changing the elevation angle of the view, either switching to a more conventional 2D top view (especially for c & d).

p2 l23: It is stated p2 l14 that bar spacing is generally in the range 10-20mm, but large bar spacings are mentionend here. What do the authors consider as small and large in this context? and which fish species does it apply to?

p3l1 : Please be more specific about the results provided by German language references 12 and 13. Especially, for which conditions is it observed that fish can cross the rack?

p3l8: The hydraulic and guiding performances of FGS with vertical bars must be more extensively discussed here.

p3l10: As reference 11 is in German language and reference 18 is currently unavailable, the authors should briefly describe what improvements or changes are provided by the MBR and MCR arrangements. Since MCR are seen as a possible solution in Section 3, their hydraulic and guiding performances have to be discussed and compared to those of conventional bar racks.

Additionaly, other bar racks types with non-normal orientation have been studied in the past, they shall be included in this paragraph too. E.g.: inclined bar racks (inclination angle w.r.t. the river bed plane), vertical bar racks with bars oriented in the main flow direction, horizontal bar racks (as in, apparently, reference [63], which has to be presented since it is seen as  possible solution in section 3).

The author do not mention inclined bar racks at all, whereas it is a very common solution. Is there a specific reason for even not considering this option? In any case, this option should be part of the global approach and considered in the "Draft design" block of Fig2.

2-Methods

Fig. 2: The proposed scheme is relatively clear, though it could be better presented as a conventional workflow chart. My main suggestion here would be to modify the connections and ordering beteween the "Structural" and "Hydraulic" conditions blocks.
First, it seems to me that the "Structural conditions" condition the hydraulics across the hydropower plan, so there should be a first connector from "Structural Cond." to "Hydraulic Cond.". Then, I think that the dashed connector from "Hydraulic Cond." to "Draft design" comes to early in the workflow and that it is the "CFD" and the "Structural Cond." that should lead to "Draft design", or at least some elementary data processing that is more reliable than a "visual assessment".
Another suggestion, following the comment made p12 l11 on the suitability of FGS types w.r.t the target species: there should be connectors between the "Inflow criteria" and "Draft design" and between "Body proportion" and "Geometric properties" (why is it dashed here?)

p5 (2.1.2) : The different notions evoked here are rather vague. I can understand that a vast range of FGS can be considered, but the authors should at least make a provisory list of objective parameters. For example, inclined FGS are can be too costly to be installed in very deep head race channels, and angled one become very costly in large channels, there must be room for a thrash rack cleaner, etc. Some simple geometric ratios can help designers to choose between different technologies, and they must be part of the procedure.

p5 (2.1.3) : The criterions given here are also too vague to serve as decision-making parameters. What is contained in "visual assessment" that can favour one or antoher solution.? The authors should be more specific here.

Similarly, when the authors assert that experience syas that best results are achieved when visual assessment and measurements or simulations are carried out, they should include sor bibliographis material to support this.

Finally, as the presented case study extensively relies on CFD, I strongly think that it is a mistake to reduce the preliminary choice of possible FGS locations to the structural and visual assessments. At most, one might admit as a first guess that the geometric angles of V1 and V2 must theoretically satisfy the vt/vn and vn criterions. Moreover, the CFD will be computed in the following step anyway, so it shouldn't be used as a simple validation tool. Instead it should serve as a real design tool, with a thorough analysis leading to an assessment of the pros and cons of different FGS types in different zones of the head race canal, and not only in 3, rather disputable, locations.

p5 (2.1.4) : In my opinion this step of the procedure is the minimal work that should be done to start discussing possible FGS locations: it has to condition the "draft design" block of Fig.2. Then, once a tentative configuration is found, an optimization process can be started that includes, at least, the bypass, and, to a certain extent, a simplified model of the rack. In the present state, this part of the procedure can only serve as a pre-dimensioning tool as it may lead, for example, to over angled FGS (in reference to the authors comment p11 l6-7).

p6 (2.1.5/6/7) For the procedure to be as robust as possible, these three steps have to be carried out for each choice of a FGS location. For example, if one FGS type is prematurely rejected in 2.1.2 or 2.1.3 and another type is selected, that turns out needing a huge investment cost, or leading to an unacceptable head losses, etc. , then there is no way to comparatively assess whether the rejected type of FGS could have been a wiser final choice, only because its performances have never been evaluated.

This goes along with my general comment that the procedure should be more inclusive and contain some optimization loop or iterative branches. Also, a comprehensive dimensioning and validation table could be  derived to summarize the compatibility of selected FGS with the key points of the procedure.

Numbering the scheme presented in Fig. 2 and linking it with the descriptions of Section 2 is advisable.

I suggest inserting a separate section for the case study instead of including it in Section 2. Subsections could then contain the description of fish fauna and structural parameters. Then a results subsection could contain the necessary description of the CFD case and results. The following section (instead of "3-Results") could be "3-test case" and focus on the CFD, with more emphasis that what is done in the submission.

Then a 4th Section would present the application of the proposed approach to the case study and its results. This Section should be structed as indicated in Section 2 and Figure 2.

3-Results

p7 l7: From this point the authors start to describe the CFD case, but they have not given sufficient informations on the hydro power plant itself, appart from its discharge and some elements of its structural arrangement. The only vague dimensions that are given can be found in figs 2a,b,c. Even fig. 5 has no dimensional information

There is absolutely no dimensional information about the head race channel and intakes, which is strongly lacking. All the key dimensions should be given before the CFD is presented, especially for the third simulation (water depth, height, width, length, bed profile, etc.).

Figs. 4a-c could be shown with depth contour level.

The authors are discussing a complex, large scale, high discharge, presumably highly turbulent, fluid mechanics problem whithout providing any dimensionless number. This is no routine work, and they can, at best, expect to obtain realistic indicative results in such a case. At least, indicative (but well defined) Froude and Reynolds numbers should be given.

p7 l19: In this paragraph, all the CFD specifics must be given. I suppose that Fig 5a corresponds to the first simulation, Fig 5b to the second and Fig 5c,d to the third. As the first and second simulation results serve as boundary conditions for the second and third one, respectively, the authors should indicate what kind of boundary conditions are used and what kind of data are transfered. This includes water fraction (field alpha in OpenFoam), velocity and turbulence. Is there any inlet turbulence level?  Is synthetic turbulence used at inlet? or is just a mean inlet velocity profile with no turbulence used?

Concerning the third simulation, I suppose that the utility snappyHexMesh is used to include the HPP in a parallelepipedic base mesh. Then, many design elements are lacking.
- where do the river bed shape and HPP geometry come from?
- what are the size and resolution of the intial mesh?
- Were there boundary layers defined around the obstacles, how many cells, with which growing parameter?
The revised version should list this.

Other informations that are lacking: boundary conditions for water fraction alpha, pressure and velocity, initial conditions on water fraction and velocity.
The information on how the water fraction is handled is essential, as VOF method are very diffusive and sensitive to boundary conditions as well as to mesh resolution at the air-water interface. OpenFOAM offers various ways to handle the air/water fractions, and the authors should provide this information for the submission to be useful for the community.

Also, why isn't the flow exiting through the fishpass represented? This would have a significant influence on V1 and V1*, at no significant additional computational cost.

Another critical information concerns the convergence of the simulation. interFoam is a transient solver that needs a large number of iterations to converge to a steady state (which is not always attained), so the authors must show an indicator of this convergence in time and/or iteration number.
With respect to this, are the presented results averages over some time steps or do they correspond to a single snapshot taken once the simulation has converged?

References [20-31], and especially [27] in which OpenFOAM is used to simulate a full-scale case, provide more suitable descriptions (though not always totally sufficient) and should be used as guidelines by the authors to list the items that are needed for a suitable case description.

The main results of the third simulation should be given at this stage, whithout reference to the FGS first. This includes:
-Global view of the flow across the HPP (e.g. top view of the velocity field - either depth-averaged or at a chosen depth - across the HPP) :  this needs to be presented to give an overview of the flow orientation so that possible FGS locations can be envisioned.
-Joint vertical profiles of the water fraction and velocity at chosen loations (e.g. at some stations at the begining of the straight portion of the canal and ahead of the canal inlet): this is essential as the validity of water velocity values quickly decreases near the air-water interface when alpha is less than 0.95. So the transition zones from water (alpha=1) to air (alpha=0) must be of very short extent (a few cells only) to be reliable.
-How is the air-water interface considered (alpha=0.5?)? Show a perspective view of the predicted water level at the HPP.
-The outcome of the simulation must be somehow validated. Are the flow results and water level consistent with the operating parameters of the HPP? Are there any experimental data available?

See e.g. https://skemman.is/bitstream/1946/13932/1/Agust_Gudmundsson_Model_Investigations_SFO.pdf in which CFD parameters are results are more edequately presented

Once the main results have been discussed, the application of the procedure to the test case can be presented.

p8 l4 :
- The criteria for choosing these 3 positions should be given. Referencing german language refs 60 and 61 is not sufficient. The foreseen bypass positions shall be given too.
- I don't subscribe to this methodology in which FGS positions are choosen without reference to experimental data or to CFD results when they are available. Choices should be made with the maximum available information. The cost of defining possible positions using the full CFD is minimal as soon as the CFD is already computed. I advise the authors to modifiy this part of their procedure to take into account the CFD as a pre-design tool to derive first FGS choices (e.g. using averaged values), and then refine the approach using complete velocity maps

Fig 6:
-For such figures, the authors should state what they consider as the air-water interface and add it to their plots so that the water fraction is clearly identified. Air velocity maps are irrelevant here so they must be either removed, or separated frome the water fraction.
-The coloring bands are not adequate, the criterion can clearly appear as a green-to-red gap, but more bands are needed to illustrate the evolution of vt/vn and vn along the cut-planes.
-The modeled turbulent kinetic energy (k field in OpenFOAM) should be shown in similar maps, as it gives an indication of the variability of vt and vn along the future FGS position.

p8 l24: what do the author mean by "a" ratio of vt to vn here? There is indeed a ratio because it is defined as such in previous experiments (and p2 below Fig.1). Even if the ratio is 0 or infinity, there is a ratio. There also are ratios of the same kind in other types of FGS, intakes, etc.

p8 l28: Isotachs do not show sensitivity nor inaccuracies, they only indicate arbitrary limits that somehow help interpreting the distribution of vn values. Sensitivity may be illustrated by slightly varying the discharge and/or water level of the whole case. Inaccuracies may be evidenced by studying the sensitivity to mesh resolution or by using some recent advanced post-processing technique.

p9 l1: If global CFD results were shown and discussed in a previous Section, the reader could actually verify by himself that separation does occur at the inlet pier. In the present state there is only some trace of a vn>0.77m/s somewhere in the vicinity of the pier. This is not adequately represented. Actually, if CFD had been considered prior to pre-selecting V1,V2,V3 arrangements then V3 would never have been in the shortlist.

p9 l3: If a different configuration is investigated, then the corresponding results must be shown.

p9 l3 : Discussing the effect of FGS using a comparison with lab experiment in straight channels at lower flows regimes should be strongly justified. This can be done if it is verified that the flow impinging on the FGS are comparable in both case. This is a possibility here since the head race canal has a significant straight portion, but this has to be backed up by the CFD results.

From what is said here, the MCR at 30° isn't in any case representative of V1,V2, nor V3, but the horizontal FGS may be representative of V1 and V2. Why discussing this MCR here? I suggest to remove it or to use an angle of 42° or 45°.

Additionaly there is no information given about the lab FGS structure (curvature and angular variation (MCR), tpe of bar, bar thickness and spacing, number of bars, number, type and size of bar supports): these have a strong influence on the velocity distribution upstream of the FGS.

Another point is lacking here: the effect of the bar racks on the lab-scale simulations is only qualitatively adressed, whereas their quantitative influence on vt/vn and vn are already known in these cases. So, their deviation from their initial values (i.e. with no FGS) cand be used to quantify what effect can be anticipated along the HPP FGS.

However, as neglecting the FGS is meant to save computational effort, which one can admit, adding a reduced series of bars, or no-slip baffles, to represent the rack at a coarser resolution, but respecting its global solidity and orientation, would still be feasable. Or, it would be possible to perform a similar simulation of the full HPP at lab scale (i.e. at a reduced Reynolds number) with a realistic lab-scale rack, which would not demand much additional computational effort than the existing lab-scale simulations. This way, results could be compared and discussed.

p11 l6 : This last sentence is disputable as the cited references also show that at some locations, FGS can increase vn and reduce vt/vn compared to configurations with no-FGS (see fig. 8 for example). So it cannot be objectively said that omitting the FGS is eligible whithout further validity assessments.

4-Discussion

p11 l14 : Omitting the FGS is the main advantage of the procedure in terms of computational cost, but also its main limitation. Indeed the authors admit that further investigations are needed to assess the impact of FGS, and that, consequently, head losses must be evaluated from empirical formulae.

p11 l21 : This part of the discussion on head losses is not coherent with Fig.2 and subsection 2.1.7. Either the authors complete their test case by choosing an FGS type and evaluate its head losses, otherwise the procedure cannot be completed, either they remove this step from the methodology and leave the headlosses as an external constraint (which I wouldn't recommend).

p11 l23 : If the CFD result had been studied prior to pre-select configuration V1, the main orientation of the flow on each side of the bridge pier had been known, and more convenient angles would have been chosen. Therefore a configuration such as V1* could have been selected earlier.
In the present methodology, it is not possible to state if a modified V2 configuration, with an angle of the order of 35°, would satisfy the biological criteria: this would probably be cheaper than manufacturing a two-part V1* grid. Or, anoher V1 configuration with a higher angle fully downstream of the bridge pier maybe another choice. Such various possibilities cannot be considered in the given procedure, simply because (apparent) arbitrary choices have been made earlier. This may be the main limitation of the proposed approach, unless the pre-selected choice (from subsection 2.1.2) are objectively argumented.

p12 l11 : These considerations of the choice of an FGS type should be formally included in the procedure, not just mentioned at the end.

It is also surprising that none of the pre-selected V1,V2,V3 is compatible with the fish pass intake. As the author have based their preliminary choices on visual assessments (Fig. 2) and on unreported considerations (p8 l3), a practical and cost-saving criterion would have been to place the downstream end of V1 downstram of the fish pass at first instance, and ensuring sufficient space for a bypass to be installed. But none of this seems taken into account. Why?

p12 l18: It can be admitted that the full design of the bypass may be done once the FGS has been chosen. However, the bypass intake location and discharge shall be included in the FGS choice, otherwise, all the CFD work must be reconsidered by including these new features afterwards.

5-Conclusion

p12 l22: Qualifying the submitted procedure as a "new" conceptual approach is a bit overrated here: other publications have focused on real scale trashracks and/or bypasses assessements using comparable procedures, CFD and empirical laws (e.g.http://www.tmt.unze.ba/zbornik/TMT2013/115-TMT13-019.pdf or https://skemman.is/bitstream/1946/13932/1/Agust_Gudmundsson_Model_Investigations_SFO.pdf for the easily available ones). Appart from the use of CFD (which is however included in many projects nowadays), the procedure is very conventional and just conforms to the general requirements for the insertion of FGS.

Also, optimality is not ensured: the procedure is closer to a trial-and-error test ending up by identifying a suitable configuration on the basis of a pre-selected FGS, which does not prove to carry any optimality.

p22 l20 : The rack is not considered at all in the numerical model, please remove "in detail" unless the procedure is modified by adding some representation of the rack (and bypass) in the simulations.

p23 l11 : typo: RCK instead of CRK

Author Response

Dear Reviewer 

Please find our answer in the attached pfd file. 

Best regards, 

Linus Feigenwinter
David Vetsch
Stephan Kammerer
Carl Robert Kriewitz
Robert Boes

Author Response File: Author Response.pdf

Reviewer 2 Report

General Comments

The paper brings an important issue relate to migratory fish problems mitigations in dams. It brings a conceptual approach for determining the best position of fish guidance considering multiple criteria (fish swimming capability, hydraulic conditions, head losses). However, it is not clear in the methods the description of this criteria or condition (expect the swimming capability of target fish) even how to consider these criteria in this approach. For example, it was not clear what means structural conditions: is it related to support of FGS concerning the efforts? I suppose it should be linked to stability of structure but it is not clear. The same occurs items “hydraulic conditions” and flow conditions in cross section.

The authors applied a powerful tool, CFD model in a open source software, in a case study but some parts of methods became clear only in Results part. In this part, some lab results are introduced but no methods were described in the last item. Three CFD scenarios were described in the methods but the results of just one is presented. I suggest not present the other two in the methods.

Finally, the results of study case are not discussed nether the proposed approach. Just limitations of research were presented and possible further adjust. This a paragraph of results in this part also.

This paper has potential to be published in the “Sustainability” but major corrections should be done.

Some minor comments:

Abstract

I couldn’t understand what mean structural conditions.

No results or discussion at the end of abstract

Introduction

2002 and 2010 – Not so actual to refer as current research”

To refer the maximum sustained swimming velocity as vopt is weird....I believe it make confusion with the optimal velocity proposed by Castro-Santos 2005

For Civil Engineering, structural aspect is related to capability of some structures keep up. Maybe layout configuration conditions” suit better as definition for this criteria.

Methods

What is unfavorable flow-related FGS?

Best results concerning what?

Results

The first paragraph in results is methodology

This part in results is methodology

In addition, an operating condition with combined turbine and weir flow was investigated.

There are other comments in the pdf

Comments for author File: Comments.pdf

Author Response

Dear Reviewer 

Please find our answer in the attached pfd file. 

Best regards, 

Linus Feigenwinter
David Vetsch
Stephan Kammerer
Carl Robert Kriewitz 
Robert Boes

Author Response File: Author Response.pdf

Reviewer 3 Report

In this manuscript, the authors describe a methodological framework to evaluate the potential design selection and site of vertical and horizontal bar screens for guiding downstream migrating fish at hydropower plants.  The authors used a hydropower facility in Switzerland as a case study of the application of their new methodology.  Overall, the manuscript outlines a basic list of design considerations for bar screens and initial assessment of hydraulic variables generated from CFD models.  While the authors provide a thorough review of scientific literature and extol on the benefits of CFD modelling in design, to which I agree, the approach ultimately fails to provide any novel research or approach.  The basic design steps are already recommended by regulators in the US (USBR, Fish Protection at Water Diversion) and UK (Environment Agency, Screening for Intake and Outfalls: a best practice guide).

In the discussion, the authors argue the “main advantage of the proposed procedure is the possibility to assess locations of FGS without considering them in the numerical model.” This is an intriguing argument, that if true could yield significant time reductions in modelling time; however, there is little to no data to support this statement in the manuscript.  I would have expected to see an analysis between CFD models with and without a FGS.  Then direct comparisons between computational resources, amount and quality of design data, and adaptability of the two modelling approaches could be done.

While the proposed methodology may be useful for preliminary engineering evaluations for bar screens (i.e., white paper), the manuscript does not report on any new experiments or provide substantially new information in its current form.  I therefore, cannot recommend the manuscript be accepted for publication based on the journals requirements.

Additional detailed comments are provided in the attached document.

Comments for author File: Comments.pdf

Author Response

Dear Reviewer 

Please find our answer in the attached pfd file. 

Best regards, 

Linus Feigenwinter
David Vetsch
Stephan Kammerer
Carl Robert Kriewitz 
Robert Boes

Author Response File: Author Response.pdf

Round  2

Reviewer 1 Report

The paper is almost ready for publication, though still of moderate interest on the technical point of view.

The CFD part is now better addressed, as I required.The decision process is better described, but additional validation computations would still be needed.

The results also are now more adequately presented. 

Author Response

Dear Reviewer

Please find our answer in the attached pdf file. 

Kind regards, 

Linus Feigenwinter

David Vetsch

Stephan Kammerer

Carl Robert Kriewitz

Robert Boes

Author Response File: Author Response.pdf

Reviewer 2 Report

Please, correct the call of Table 1 in page 10.

Comments for author File: Comments.pdf

Author Response

Dear Reviewer

Please find our answer in the attached pfd file.

Kind regards,

Linus Feigenwinter

David Vetsch

Stephan Kammerer

Carl Robert Kriewitz 

Robert Boes

Author Response File: Author Response.pdf

Reviewer 3 Report

Overall, the authors did an excellent job re-framing the structure of the manuscript to better suite their material.  I think the objective of developing a conceptual model for the design and potential installation of FGS is far clearer and well achieved.  I recommend the manuscript be accepted for publication after a few minor edits.  Please see specific comments in the attached document.

Comments for author File: Comments.pdf

Author Response

Dear Reviewer

Please find our answer in the attached pfd file.

Best regards,

Linus Feigenwinter

David Vetsch

Stephan Kammerer

Carl Robert Kriewitz 

Robert Boes

Author Response File: Author Response.pdf