Wave-induced sediment resuspension potential in the Finnish Archipelago, Baltic Sea: integrating field measurements with large-scale numerical model simulations

Björkqvist, Jan-Victor; Savela, Mari; Pettersson, Heidi; Alari, Victor; Norkko, Alf

doi:10.5194/os-22-2123-2026

Articles | Volume 22, issue 4

https://doi.org/10.5194/os-22-2123-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/os-22-2123-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 22, issue 4

Research article

|

06 Jul 2026

Research article |

| 06 Jul 2026

Wave-induced sediment resuspension potential in the Finnish Archipelago, Baltic Sea: integrating field measurements with large-scale numerical model simulations

Jan-Victor Björkqvist, Mari Savela, Heidi Pettersson, Victor Alari, and Alf Norkko

Download

Final revised paper (published on 06 Jul 2026)
Preprint (discussion started on 03 Jul 2025)

Interactive discussion

Status: closed

RC1: 'Comment on egusphere-2025-2936', Anonymous Referee #1, 07 Aug 2025

Review of: “Wave-induced sediment resuspension in the Finnish Archipelago, Baltic Sea: Combining small-scale in situ measurements and large-scale numerical model simulations”

Dear Editor,

The authors present an interesting case study on wave resuspension of sediment in the Finnish Archipelago. The authors have used a spectral wave model to calculate bottom orbital velocities and bed shear stresses, and use this to estimate the threshold of motion for the sediments in the region. Interestingly, the authors have also measured the threshold of motion from sediment samples using a laboratory which has been published (Joensuu et al. 2020). As I have little experience of spectral wave modelling, I will keep my reviews focused on the sediments. However I would ask if wind forcing at a temporal resolution of 3 hours is enough to adequately resolve the storms?

The authors find a large discrepancy between the predicted and measured critical shear stresses, and attribute this to biological and/or biogeochemical properties of the sediments. Concerningly, there is no reporting of the biogeochemical properties of the sediments, so it is difficult assess how accurate their claim is.

The authors seem to think this is the only way to explain the differences in measured and predicted threshold of motion for the sediments – but this is unfounded. Bulk density of sediment has been shown to have the largest effect on the threshold of motion (Thompson et al. 2019). Mixed grain sizes (for example, a bimodal distribution) can increase the threshold of motion (Staudt et al. 2017; McCarron et al. 2019). In short, the manuscript is, at present, not representing the state of the science. The authors need to address their simplification of the problem.

Upon reading (Joensuu et al. 2020), I see that all the relevant information has been collected, but not used in the present work. I encourage the authors to utilise this information, particularly to follow the work of Thompson et al. (2019), who were able to adjust the threshold of motion for their sediment based upon an array of sedimentary and biological variables. As such, I recommend major revisions as a reanalysis of the data is necessary for the manuscript to be up to date with the state of the science.

Comments to the authors.

Dear authors, I read your manuscript with interest. This is tricky subject, and I think you need to include some of the sedimentary information from (Joensuu et al. 2020) into your estimate of the threshold of motion of the sediments and use the work of (Thompson et al. 2019) to recalculate your thresholds of motion - it seems like an approach similar to their “model 1” would work well. You seem to have limited your explanation of the difference between modelled and measured threshold of motion to only biological processes, yet there is little reason given in the manuscript to justify this. There are numerous reasons why such a difference could occur, much of this is covered in Thompson et al. 2019, and other papers.

Some specific comments:

Equation 8 – the Soulsby Whitehouse equation is incorrectly written, it should be:

(see attached image)

I also highlight to the authors that this is the equation for initiation of bedload, not suspended load. There is a Soulsby-Whitehouse like equation for this, fitted by Van Rijn (unpublished, the source is his website):

(see attached image)

I have calculated the suspension values for their grain sizes (Their table 1, my numbers in red). It would appear they have used the correct version of the Soulsby Whitehouse equation, and it was just written wrong in the manuscript.

Sediment class	Grain size (mm)	theta crit (N m⁻²)	theta sus(N m⁻²)
Mud to muddy sand	0.09	0.14	0.16
Sand	0.34	0.21	0.37
Coarse sediments	2.00	1.20	3.06
Mixed sediments	0.15	0.16	? mixed ? (0.21)
Boulders	200.00	178.10	323

The difference between theta crit and theta sus could account for most of the discrepancy the authors have found, especially for the larger grain sizes. It is unclear how they have arrived at their bed shear stress estimate for “mixed” sediments, it appears to be just using the median particle size, which is unjustified. This is unwise as larger clasts (such as gravels and boulders) can “armor” the bed and reduce mobility of all sediments (Wiberg et al. 1994; Vericat et al. 2006). Likewise fine sediments (< sand size) can add cohesion to the bed and reduce hyporheic flow, impacting sediment mobility (Blois et al. 2014; Fox et al. 2014; Parsons et al. 2016; Perret et al. 2018; 2023).

In particular, i recommend to the authors that they read the paper by Thompson et al., (2019), as this paper works through many of the issues with defining a threshold of motion in complicated sediments, including those with biological controls (for instance, Thompson et al., include the concentration chlorophyll-A in their “Model 1”). Moreover, that work found that the bulk density and porosity of the sediment was the overriding control on benthic resuspension, I suggest you try and include this in your work based on the data available in (Joensuu et al. 2020).

References.

Blois, Gianluca, James L. Best, Gregory H. Sambrook Smith, and Richard J. Hardy. 2014. ‘Effect of Bed Permeability and Hyporheic Flow on Turbulent Flow over Bed Forms’. Geophysical Research Letters 41 (18): 6435–42. https://doi.org/10.1002/2014GL060906.

Fox, Aryeh, Fulvio Boano, and Shai Arnon. 2014. ‘Impact of Losing and Gaining Streamflow Conditions on Hyporheic Exchange Fluxes Induced by Dune‐shaped Bed Forms’. Water Resources Research 50 (3): 1895–907. https://doi.org/10.1002/2013WR014668.

Joensuu, Mari, Conrad A. Pilditch, and Alf Norkko. 2020. ‘Temporal Variation in Resuspension Potential and Associated Nutrient Dynamics in Shallow Coastal Environments’. Estuaries and Coasts 43 (6): 1361–76. https://doi.org/10.1007/s12237-020-00726-z.

McCarron, Connor J., Katrien J.J. Van Landeghem, Jaco H. Baas, Laurent O. Amoudry, and Jonathan Malarkey. 2019. ‘The Hiding-Exposure Effect Revisited: A Method to Calculate the Mobility of Bimodal Sediment Mixtures’. Marine Geology 410 (April): 22–31. https://doi.org/10.1016/j.margeo.2018.12.001.

Parsons, Daniel R., Robert J. Schindler, Julie A. Hope, et al. 2016. ‘The Role of Biophysical Cohesion on Subaqueous Bed Form Size’. Geophysical Research Letters 43 (4): 1566–73. https://doi.org/10.1002/2016GL067667.

Perret, Emeline, Céline Berni, Benoît Camenen, Albert Herrero, and Kamal El Kadi Abderrezzak. 2018. ‘Transport of Moderately Sorted Gravel at Low Bed Shear Stresses: The Role of Fine Sediment Infiltration’. Earth Surface Processes and Landforms 43 (7): 1416–30. https://doi.org/10.1002/esp.4322.

Perret, Emeline, Benoit Camenen, Céline Berni, Kamal El kadi Abderrezzak, and Benjamin Renard. 2023. ‘Uncertainties in Models Predicting Critical Bed Shear Stress of Cohesionless Particles’. Journal of Hydraulic Engineering 149 (4): 04023002. https://doi.org/10.1061/JHEND8.HYENG-13101.

Staudt, Franziska, Julia C. Mullarney, Conrad A. Pilditch, and Katrin Huhn. 2017. ‘The Role of Grain-Size Ratio in the Mobility of Mixed Granular Beds’. Geomorphology 278 (February): 314–28. https://doi.org/10.1016/j.geomorph.2016.11.015.

Thompson, C.E.L., M.E. Williams, L. Amoudry, et al. 2019. ‘Benthic Controls of Resuspension in UK Shelf Seas: Implications for Resuspension Frequency’. Continental Shelf Research 185 (September): 3–15. https://doi.org/10.1016/j.csr.2017.12.005.

Vericat, Damia, Ramon J. Batalla, and Celso Garcia. 2006. ‘Breakup and Reestablishment of the Armour Layer in a Large Gravel-Bed River below Dams: The Lower Ebro’. Geomorphology 76 (1–2): 122–36. https://doi.org/10.1016/j.geomorph.2005.10.005.

Wiberg, Patricia L., David E. Drake, and David A. Cacchione. 1994. ‘Sediment Resuspension and Bed Armoring during High Bottom Stress Events on the Northern California Inner Continental Shelf: Measurements and Predictions’. Continental Shelf Research 14 (10–11): 1191–219. https://doi.org/10.1016/0278-4343(94)90034-5.

Citation: https://doi.org/10.5194/egusphere-2025-2936-RC1

AC1: 'Reply on RC1', Jan-Victor Björkqvist, 06 Jan 2026

Response to review comments. Our response in bold.

Review of: “Wave-induced sediment resuspension in the Finnish Archipelago, Baltic Sea: Combining small-scale in situ measurements and large-scale numerical model simulations”

Dear Editor,

The authors present an interesting case study on wave resuspension of sediment in the Finnish Archipelago. The authors have used a spectral wave model to calculate bottom orbital velocities and bed shear stresses, and use this to estimate the threshold of motion for the sediments in the region. Interestingly, the authors have also measured the threshold of motion from sediment samples using a laboratory which has been published (Joensuu et al. 2020). As I have little experience of spectral wave modelling, I will keep my reviews focused on the sediments. However I would ask if wind forcing at a temporal resolution of 3 hours is enough to adequately resolve the storms?

Our response: If a 3 hour wind forcing resolution is high enough to resolve the storms is a valid question. Björkqvist et al. (2018) made Baltic Sea wide simulations using a 6 h temporal resolution in the wind, and that model validated well when compared to extensive measurements. However, in this study the 3 hour resolution was only used for 2017 (based on available data at the time). The model was run for 2017 mainly to compare with available wave measurements, and this validation was good. For the times actually overlapping with the sediment samples, we used an 1 h resolution. This means that the examples that are seen in Figures 7 and 8 are generated using a 1 h temporal resolution in the forcing wind field.

The authors find a large discrepancy between the predicted and measured critical shear stresses, and attribute this to biological and/or biogeochemical properties of the sediments. Concerningly, there is no reporting of the biogeochemical properties of the sediments, so it is difficult assess how accurate their claim is.

The authors seem to think this is the only way to explain the differences in measured and predicted threshold of motion for the sediments – but this is unfounded. Bulk density of sediment has been shown to have the largest effect on the threshold of motion (Thompson et al. 2019). Mixed grain sizes (for example, a bimodal distribution) can increase the threshold of motion (Staudt et al. 2017; McCarron et al. 2019). In short, the manuscript is, at present, not representing the state of the science. The authors need to address their simplification of the problem.

Upon reading (Joensuu et al. 2020), I see that all the relevant information has been collected, but not used in the present work. I encourage the authors to utilise this information, particularly to follow the work of Thompson et al. (2019), who were able to adjust the threshold of motion for their sediment based upon an array of sedimentary and biological variables. As such, I recommend major revisions as a reanalysis of the data is necessary for the manuscript to be up to date with the state of the science.

Our response: Thank you for your constructive review. Based on your suggestions we have made significant changes to how we conduct our study. In summary, we tested (and rejected) the model from Thompson et al. (2019), but instead determined our own similar model using both physical and biological properties directly from the in situ samples. This serves to directly quantify the role of biology based on the samples, and also allows for the biological component to vary in time when the model is applied. It also opens up for a discussion on why the model from Thompson et al. (2019) doesn’t necessarily translate to our data. For more details, please see our comments below.

Comments to the authors.

Dear authors, I read your manuscript with interest. This is tricky subject, and I think you need to include some of the sedimentary information from (Joensuu et al. 2020) into your estimate of the threshold of motion of the sediments and use the work of (Thompson et al. 2019) to recalculate your thresholds of motion - it seems like an approach similar to their “model 1” would work well. You seem to have limited your explanation of the difference between modelled and measured threshold of motion to only biological processes, yet there is little reason given in the manuscript to justify this. There are numerous reasons why such a difference could occur, much of this is covered in Thompson et al. 2019, and other papers.

Our response: Thank you for taking the time to review our manuscript. Following your suggestions we have gone back to the raw data of Joensuu et al. (2018, 2020) and used this to both test the model of Thompson et al. (2019) to the degree it was possible, but, more importantly, constructed our own similar model for the critical shear stress based on the data.

Some specific comments:

Equation 8 – the Soulsby Whitehouse equation is incorrectly written, it should be:

(see attached image)

I also highlight to the authors that this is the equation for initiation of bedload, not suspended load. There is a Soulsby-Whitehouse like equation for this, fitted by Van Rijn (unpublished, the source is his website):

(see attached image)

I have calculated the suspension values for their grain sizes (Their table 1, my numbers in red). It would appear they have used the correct version of the Soulsby Whitehouse equation, and it was just written wrong in the manuscript.

Our response: We appreciate the attention to detail. This indeed looks like a notational issue and will be corrected in the manuscript.

Sediment class	Grain size (mm)	theta crit (N m−2)	theta sus(N m−2)
Mud to muddy sand	0.09	0.14	0.16
Sand	0.34	0.21	0.37
Coarse sediments	2.00	1.20	3.06
Mixed sediments	0.15	0.16	? mixed ? (0.21)
Boulders	200.00	178.10	323

The difference between theta crit and theta sus could account for most of the discrepancy the authors have found, especially for the larger grain sizes. It is unclear how they have arrived at their bed shear stress estimate for “mixed” sediments, it appears to be just using the median particle size, which is unjustified. This is unwise as larger clasts (such as gravels and boulders) can “armor” the bed and reduce mobility of all sediments (Wiberg et al. 1994; Vericat et al. 2006). Likewise fine sediments (< sand size) can add cohesion to the bed and reduce hyporheic flow, impacting sediment mobility (Blois et al. 2014; Fox et al. 2014; Parsons et al. 2016; Perret et al. 2018; 2023).

Our response: We have modified our approach based on the extensive comments we got in the review stage. All grain sizes are now based on the in situ measurements only, and the modelled data from EMODNET is only used as a classification. Since no in situ data is available for coarse sediments and gravel, no results are given for those areas. Since we don’t have the grain size distribution, we had no choice but to rely on the median grain size also for the mixed sediments class. However, since our model now also includes the bulk density and chlorophyll a, the results are not only reliant on that particular measure for the grain size.

In particular, i recommend to the authors that they read the paper by Thompson et al., (2019), as this paper works through many of the issues with defining a threshold of motion in complicated sediments, including those with biological controls (for instance, Thompson et al., include the concentration chlorophyll-A in their “Model 1”). Moreover, that work found that the bulk density and porosity of the sediment was the overriding control on benthic resuspension, I suggest you try and include this in your work based on the data available in (Joensuu et al. 2020).

Our response: We have tested Model 1 of Thompson et al. (2019) to the degree it was possible. For the parameters not available to us we used similar realistic values to those found in Thompson et al. (2019). This model did not work well for our data. Especially, the authors found that bulk density explained a lot of the variation in critical stress, and show a model (their Fig. 3) with a positive quadratic relationship. Nonetheless, in their Model 1, the coefficient for the bulk density is negative. Although their model fit their data, we got unphysical negative values when applying it on our data, because our bulk density was higher. The conclusion that biological factors are not that important also doesn’t necessarily translate between their and our dataset, since their data was collected at significantly larger depths.

We therefore fit our own model to our dataset using three variables: grain size, bulk density and Chlorophyll a. We made a fit for each available bottom type (mud, sand, mixed sediment) separately (Figure 1, all except top left), and a fit to all the data (Figure 1, top left). Here, the grain size and bulk density represents the physical conditions, while Chlorophyll A is meant to capture the biological activity. We can see from Fig.1 (top left) that the model can especially explain the variability between the groups.

(See attached file)

Figure 1. Our three variable model (grain size, dry bulk density and ChlorophyllA) to model the critical shear stress. The model can especially explain the variations between the different classes (top left).

After we fit the model we then determined representative values for grain size and bulk density for each class based on the in situ measurements. While we assumed that the physical properties are constant, we allowed for the Chl A to vary over time. These values were determined as monthly values based on the in situ data (made possible by the temporal dataset of Joensuu et al. 2020).

This spatio–temporal model can now be used for predicting the critical threshold for each class, and these thresholds can be compared to the shear stresses from the numerical model. This significant modification accomplishes several things: a) it puts our work in respect to other similar work, especially that of Thompson et al. (2019), b) it makes a stronger connection between using a combination of measurements and models, where measurements are directly used to determine a critical threshold, while models are used to generalize these using the EMODNET seabed model, and compared to the stresses from the numerical simulations by SWAN, c) it highlights the difference between the physical and biological properties, by allowing the latter to vary in time.

Citation: https://doi.org/10.5194/egusphere-2025-2936-AC1

RC2:
'Comment on egusphere-2025-2936', Anonymous Referee #2, 06 Oct 2025

General comments
Mixed sediments are a complex problem and not well integrated into numerical models of sediment dynamics and transport. In situ data are limited and often site specific due to the variability of biogeochemical properties of mobile sediment. Improving understanding of the differences between modelled and observed sediment erosion and resuspension should lead to the development of better models of sediment processes. This work uses a novel high spatial resolution model to highlight the amount of uncertainty in coastal models of hydrodynamics and sediment processes, when compared to in situ samples, and variability on small spatial scales.
The figures could be better thought out, to make it easier for the reader to compare and interpret them. Many statements are given without references to support them and the reader cannot follow them up. The work of Joensuu et al. is relied on for the in-situ data, but their contribution and the authors contribution is not always clear. Key information on the source of the bathymetry and its accuracy and quality is missing.
Specific comments
Line 26: Exceeding the critical shear stress for motion doesn’t always lead to resuspension, it might just be bedload. It implies here that once the critical limit is exceeded suspension will happen. Please clarify.
Section 2.1.2 mentions ‘variation in sediment surface characteristics (e.g. bedforms, biofilms) across sites’, but this data is not shown or used. Grain size class % and biological factors are key factors in sediment resuspension, which were measured by Joensuu et al. (2018; 2020), but not fully integrated into this work.
Section 4 – Discussion: What was the bathymetry source used? And were there any problems with it? Also, the section could be better referenced, many statements are made that aren’t supported by the results. Joensuu et al. give % clay fraction but there is no mention of how cohesive sediment could affect these results. Biological factors are highly variable, and effects difficult to account for, but as the clay has been measured could it be used to reduce uncertainty in the model results? The EMODNET data set is not explicitly mentioned in the discussion and, given that this represents the greatest uncertainty, I wanted to know more about where the data came from and how it was put together. Although it is out of scope of this work, examining or re-analysing the raw data used by EMODNET could result in a better understanding of the problem of sediment variability. What survey work would be needed to get a data set that could significantly reduce the uncertainty of a resuspension model?

Access to the in-situ sediment sample data set was not made available for the reviewers to check, though the Zenodo record exists. The details of where the wave buoy data is archived were not given.

Technical corrections
Lines 17-20: citations?
Line 80: Citation for DWR-G4 Directional Waverider? Is it the same for the Waverider Mk-III?
Figure 2: What is the data source of the bathymetry contours and coastline? The area of this figure doesn’t correspond to the bottom classification and model figures, which is a bit confusing when trying to link the bathymetry to the bed type and model results. Site 13 isn’t listed in Table 2.
Line 93: citations here for those unfamiliar with EROMES coring.
Lines 109-110: What is the data source of the bathymetric data and land-sea mask?
Figure 4: where are the latitude and longitude tick marks, and land-sea mask? It is difficult to relate this to figure 2. Marking the sites of the cores would be useful to understand how good the EMODNET data is?
Line 141: Citation to back this up. Finer grains are generally the first to be suspended, but coarser grains will be mobilized first as bedload. Please clarify that you are writing about wave suspension. Is this just for waves and non-cohesive sediment? This statement seems a bit over simplified without a citation to back it up.
Line 149: citation for this equation.
Line 165: citation please.
Line 168: Soulsby and Whitehouse improved Shield’s curve with extra data, not the parameter.
Line 193: Tvärminne research station is not marked on the model data figures. 23° 06’ -23° 12’is easier to find. Please mark the areas you are writing about on the figures so that readers can understand better. Bathymetry contour lines would help with the interpretation of shallow and deep areas, though this could clutter up the figure, contours every 10- 20 m depth could make the interpretation easier than switching between figures 1 & 4 and model figures.
Figure 6: The figure has a title of Tvärminne but it is not shown in the figure. Place labels, e.g. for Tvärminne and Täktbukten could be used to highlight interesting areas where the exceedance is high (as mentioned in line 200) and discuss these individual sites.
Lines 252-4: Were these measurements carried out by Joensuu et al.? Cite them if so.
Line 259: The sample sites aren’t marked on figures 7 or 8, so it is hard to see this.
Table 2: the grain size measurements are from Joensuu et al. (2018; 2020) and should be cited here.
Line 299: remove ID from ‘stIDress’.
Lines 340-367: Appendix A - Citations please.
Line 442: ‘https://doi.org/’ is duplicated.
Line 445: Full reference so that it may be found. An English translation can be found here: http://authors.library.caltech.edu/25992/1/Sheilds.pdf

Citation: https://doi.org/10.5194/egusphere-2025-2936-RC2
- AC2: 'Reply on RC2', Jan-Victor Björkqvist, 06 Jan 2026
  
  Response to review comments. Our response in bold.
  General comments
  Mixed sediments are a complex problem and not well integrated into numerical models of sediment dynamics and transport. In situ data are limited and often site specific due to the variability of biogeochemical properties of mobile sediment. Improving understanding of the differences between modelled and observed sediment erosion and resuspension should lead to the development of better models of sediment processes. This work uses a novel high spatial resolution model to highlight the amount of uncertainty in coastal models of hydrodynamics and sediment processes, when compared to in situ samples, and variability on small spatial scales.
  The figures could be better thought out, to make it easier for the reader to compare and interpret them. Many statements are given without references to support them and the reader cannot follow them up. The work of Joensuu et al. is relied on for the in-situ data, but their contribution and the authors contribution is not always clear. Key information on the source of the bathymetry and its accuracy and quality is missing.
  Our response: Thank you for taking the time to review our manuscript. Based on the comments you and the second reviewer have given, we have heavily edited the manuscript. The main addition is that we have now used the data from Joensuu et al. (2018, 2020) to make our own model for the critical shear stress. This makes it evidently clear what is the new contribution of this manuscript, lifts the scientific novelty and better connects the measurements with the numerical results.
  Specific comments
  Line 26: Exceeding the critical shear stress for motion doesn’t always lead to resuspension, it might just be bedload. It implies here that once the critical limit is exceeded suspension will happen. Please clarify.
  Our response: We have edited the text to reflect that exceeding the threshold creates a potential for resuspension.
  Section 2.1.2 mentions ‘variation in sediment surface characteristics (e.g. bedforms, biofilms) across sites’, but this data is not shown or used. Grain size class % and biological factors are key factors in sediment resuspension, which were measured by Joensuu et al. (2018; 2020), but not fully integrated into this work.
  Our response: You are right that there was a lack of connection between the in situ data, and how the critical threshold were modelled in our study. Based on the review comments we have made significant changes to our study. In summary, we tested (and rejected) the model from Thompson et al. (2019), but instead determined our own similar model using both physical and biological properties directly from the in situ samples. This serves to directly quantify the role of biology based on the samples, and also allows for the biological component to vary in time when the model is applied. It also opens up for a discussion on why the model from Thompson et al. (2019) doesn’t necessarily translate to our data. For more details, please see our comments below.
  Section 4 – Discussion: What was the bathymetry source used? And were there any problems with it? Also, the section could be better referenced, many statements are made that aren’t supported by the results. Joensuu et al. give % clay fraction but there is no mention of how cohesive sediment could affect these results. Biological factors are highly variable, and effects difficult to account for, but as the clay has been measured could it be used to reduce uncertainty in the model results? The EMODNET data set is not explicitly mentioned in the discussion and, given that this represents the greatest uncertainty, I wanted to know more about where the data came from and how it was put together. Although it is out of scope of this work, examining or re-analysing the raw data used by EMODNET could result in a better understanding of the problem of sediment variability. What survey work would be needed to get a data set that could significantly reduce the uncertainty of a resuspension model?
  
  Our response: The bathymetric data has been compiled by FMI using nautical charts and the Velmu depth model (which is a part of the Baltic Sea Bathymatric Database). In additions, the depth measurements made at the sites (Joensuu et al. 2018; 2002) were incorporated into the wave model grid. Because of defence purposes, the accuracy of the depth information in the area cannot be guaranteed. While there is also an uncertainty in the EMODENET data, our updated methodology only relies on the EMODNET data for classification, while the actual properties for the classes are based directly on the in situ measurements. As mentioned, a more in depth review of the quality of the EMODNET data falls without the scope of our work. Connecting the in situ measurements and the numerical data is a challenging work in itself, so we have to rely on taking the EMODNET data “as is”. However, we do find som shortcoming in the classification based on the samples we used, and raise this in the discussion.
  Access to the in-situ sediment sample data set was not made available for the reviewers to check, though the Zenodo record exists. The details of where the wave buoy data is archived were not given.
  Our response: These data sets will be opened up upon the acceptance of the manuscript.
  Technical corrections
  Lines 17-20: citations?
  Our response: Thank you for pointing this out. We have added appropriate citations to the lines 17-20.
  Line 80: Citation for DWR-G4 Directional Waverider? Is it the same for the Waverider Mk-III?
  Our response: Thank you for pointing this out. The two buoys do have different manuals. The reference to Mk-III has been updated and the reference to G4 has been added.
  Figure 2: What is the data source of the bathymetry contours and coastline? The area of this figure doesn’t correspond to the bottom classification and model figures, which is a bit confusing when trying to link the bathymetry to the bed type and model results. Site 13 isn’t listed in Table 2.
  Our response: The source for the bathymetry is nautical charts and the Velmu depth model (https://ckan.ymparisto.fi/dataset/velmu-syvyysmalli). The coastline data has been rasterized from the polygon data in the dataset (https://ckan.ymparisto.fi/dataset/ranta10-rantaviiva-1-10-000). We have now added tick marks to the bottom classification figure which makes it easier to compare to the bathymetry. We will remove site 13 from the map.
  Line 93: citations here for those unfamiliar with EROMES coring.
  Our response: Thank you for the suggestion. We have added a citation that describes the EROMES system to provide context for readers unfamiliar with it.
  Lines 109-110: What is the data source of the bathymetric data and land-sea mask?
  Our response:The source for the bathymetry is nautical charts and the Velmu depth model (https://ckan.ymparisto.fi/dataset/velmu-syvyysmalli). The coastline data has been rasterized from the polygon data in the dataset (https://ckan.ymparisto.fi/dataset/ranta10-rantaviiva-1-10-000)
  Figure 4: where are the latitude and longitude tick marks, and land-sea mask? It is difficult to relate this to figure 2. Marking the sites of the cores would be useful to understand how good the EMODNET data is?
  Our response: We have redrawn the figure of the classes to have tick marks. We also marked the region used in the other figures. We did not add the stations to all figures, but we added them to figure 7. This shows some limitations in the accuracy of the EMODNET classification, and this is raised in the edited discussion.
  Line 141: Citation to back this up. Finer grains are generally the first to be suspended, but coarser grains will be mobilized first as bedload. Please clarify that you are writing about wave suspension. Is this just for waves and non-cohesive sediment? This statement seems a bit over simplified without a citation to back it up.
  Our response: Thank you for your valuable feedback. We agree that particle motion is influenced by grain size, however, our study focuses on particles that can be resuspended rather than coarser sediments or boulders, which were excluded from our dataset as they fall outside the scope of this work. The critical shear stress values obtained using the EROMES system are based on the suspended motion of the particles.
  While the primary focus of this study is on wave-induced sediment resuspension, we also aim to cover a range of sediment types. Natural sediments typically consist of a mixture of particle sizes, and the erosion process can vary depending on sediment properties (e.g. floc erosion versus suspension). In addition to grain size, other factors such as cohesion and organic content influence erosion behavior.
  Despite these complexities, finer particles are generally the first to be resuspended under wave action, which is the phenomenon we sought to capture in natural submerged sediments. We have clarified this point in the manuscript and added appropriate citation.
  Line 149: citation for this equation.
  Our response: We assume this refers to Eq (1). Citation was added.
  Line 165: citation please.
  Our response: We assume this refers to Eq (7). Citation was added.
  Line 168: Soulsby and Whitehouse improved Shield’s curve with extra data, not the parameter.
  Our response: This was changed to “The parameterization of the Shields parameter, as improved by Soulsby and Whitehouse (1997),
  reads:”
  Line 193: Tvärminne research station is not marked on the model data figures. 23° 06’ -23° 12’is easier to find. Please mark the areas you are writing about on the figures so that readers can understand better. Bathymetry contour lines would help with the interpretation of shallow and deep areas, though this could clutter up the figure, contours every 10- 20 m depth could make the interpretation easier than switching between figures 1 & 4 and model figures.
  Our response: We appreciate that the figures are not that easy to interpret. We have made an effort to harmonize the presentation and marking the main area in the figure presenting the seabed classes. We however decided not to include contour lines, since it adds too much information to the figures. In the new figures it should be clearer which areas are e.g. of type sand, which makes it easier to follow the results.
  Figure 6: The figure has a title of Tvärminne but it is not shown in the figure. Place labels, e.g. for Tvärminne and Täktbukten could be used to highlight interesting areas where the exceedance is high (as mentioned in line 200) and discuss these individual sites.
  Our response: In the edited manuscript we have moved away from discussing individual sites. However, we have marked the Tvärminne research station on the figure, since it is mentioned in the text.
  Lines 252-4: Were these measurements carried out by Joensuu et al.? Cite them if so.
  Our response: Thank you for pointing this out. We confirm that these measurements were carried out by Joensuu et al., and we have added the appropriate citations to acknowledge their work.
  Line 259: The sample sites aren’t marked on figures 7 or 8, so it is hard to see this.
  Our response: This sentence has been rewritten in the revised manuscript.
  Table 2: the grain size measurements are from Joensuu et al. (2018; 2020) and should be cited here.
  Our response: Thank you for pointing this out. We have added the appropriate citations.
  Line 299: remove ID from ‘stIDress’.
  Our response: This typo has been corrected
  Lines 340-367: Appendix A - Citations please.
  Our response: We have added citations to the Appendix
  Line 442: ‘https://doi.org/’ is duplicated.
  Our response: Thank you for catching that duplication. We’ve reviewed and cleaned the references to remove the extra https://doi.org/ entries.
  Line 445: Full reference so that it may be found. An English translation can be found here: http://authors.library.caltech.edu/25992/1/Sheilds.pdf
  Our response: Thank you for the suggestion. We have added a citation to the original Shields (1936) work and included a reference to the English translation for accessibility. We have noted the translation link in the reference for readers who wish to access it.
  
  Citation: https://doi.org/10.5194/egusphere-2025-2936-AC2

Peer review completion

AR – Author's response | RR – Referee report | ED – Editor decision | EF – Editorial file upload

AR by Jan-Victor Björkqvist on behalf of the Authors (14 Jan 2026) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (20 Jan 2026) by John M. Huthnance

RR by Anonymous Referee #2 (23 Jan 2026)

Suggestions for revision or reasons for rejection

General comments

Major changes and improvements to the model used for estimating the critical threshold have been carried out since the initial submission, and the manuscript re-written. Monthly variation in threshold due to biological cohesion has been included in a new model for critical stress. In addition, the manuscript is much better referenced.

I found it interesting that wave stress is focused by channels between islands leading to peaks in stress, which once passed allow re-deposition of fine sediment resulting in mud beds. While near the shore higher mean/background stress in shallower water results in coarse sediment at the bed, as would be generally expected. The re-suspension events and re-deposition seem to be controlled by the coastal bathymetry and morphology. For me, to ‘demonstrate that sediment mobilization is highly episodic, concentrated in exposed shallow areas and constricted channels, with implications for turbidity, nutrient fluxes, and benthic habitats’ is a key result made possible by the high spatial resolution of their model. The integration of the high-resolution model with the EMODnet data and in-situ measurements, factoring in the sediment classes, is also a highlight of the work.

Specific comments

Section 4.3
The non-stationary physical–biological model could be better described, with more statistical details about the multiple linear regression please. The number of measurements used to create the model, RMS error/difference and the overall p-value. These would help the reader assess the uncertainty and significance of the model. The uncertainty isn’t mentioned and if there is a wide range of possible critical stress values within the physical–biological model, then the spatial model maps could look different at the extremes of the range. In particular, Chl a data can have a lot of scatter, due to patchiness over the sediment bed, and this could propagate to a large uncertainty in the critical stress values.

Correlation between variables (collinearity) can be tested for (see Zuur et al. 2010, step 5, https://besjournals.onlinelibrary.wiley.com/doi/10.1111/j.2041-210X.2009.00001.x), but understanding the variables and keeping the number to a minimum few key variables reduces its likelihood. ‘The choice of which covariates to drop can be based on the VIFs, or perhaps better, on … knowledge.’ (Zuur et al, 2010).

Technical corrections

Figure 5 caption: remove ‘mathrm’ from ‘mathrmChlA’

L223-4: ‘While the areas with sand bottom are highlighted also in the maximum shear stress values, the highest values (exceeding 2 N m−2) can be found at mud bottoms’ where is this shown? Is this comparing figures 4a and 6c? Please clarify.

Conclusions: ‘This study underlines the importance of incorporating both physical and biological factors – and their temporal dynamics– into sediment transport models to achieve reliable predictions of erosion thresholds and resuspension potential.’ could be added to the end of the abstract as a key conclusion.

Hide

RR by Anonymous Referee #1 (09 Feb 2026)

Suggestions for revision or reasons for rejection

Dear editor,
The manuscript “Wave-induced sediment resuspension in the Finnish Archipelago, Baltic Sea: Integrating field measurements with large-scale numerical model simulations” by Björkqvist et al is much improved from the initial submission, but there are still some refinements to the paper and clarity needed around modelling validation for it to be accepted.
The spurious bed shear stresses and suspension of boulders have been removed from the results, primarily as the authors do not have field measurements of their suspension. This focusing of the model results onto what relevant field data they have has greatly improved the scope of the manuscript. The authors have ingested the paper of (Thompson et al., 2019) well, and have created their own model for the initiation of motion for sediments with differing biological properties and compaction which now explains a larger proportion of the sediments’ mobility. In doing this they have also been able to define the difference between sediment mobility with and without the biological properties, which is a good new result for this work.
There are still questions as to other processes which are controlling the threshold of motion for their sediments – questions which the wider literature also has. As such the paper is not much more up to date with the state of the science.
I currently still have reservations about two aspects of the manuscript
1) The lack of any distinction between bedload and suspended load in the calculations. The authors are only using the threshold of motion, and not suspension, yet always talk about resuspension. I advise them to either adjust their method or change the focus to initiation of motion. They may well compare both methods – which would also be fine.
2) The poor validation of wave period in their model compared to observation (Figure 3). Whilst bottom orbital velocity is well matched with observation, wave period is not well compared to the point where wave period is often> 50% lower in the model compared to the field measurement. Considering the orbital velocities are good, this suggests wave heights are not well modelled - presumably during period of low wave height conditions. Figure 3 needs to be expanded to include wave height observations and model prediction. An estimate of wave energy as well would let readers have a clear reason why the authors are confident in ignoring this discrepancy in their model – as presumably the wave energy during the periods of low correlation is also low.
There is also a concern that the period of model validation does not appear to be used in the rest of the paper. The 2017 validation period is unused in section 5. The 2017 model coincides with available wave measurement data – whilst the sediment collection data (and model runs for these) cover periods in 2014 and 2015. Figures 6 -7 (modelling period undefined) and Figures 8-9 (a storm event on 28 November 2015) are presumably all from the same 2015 model run. All things being equal, this shouldn’t be an issue. A validation period of a different year should also work, my concern however is that the range of bottom orbital velocities shown in the November 2015 storms is far higher than any during the validation period; 0.8 m/s (Figure 8) in 2015 vs a maximum of 0.1 m/s in 2017 (Figure 3). Given the absence of observations for this period, additional justification is needed to demonstrate that the model is operating in a physically realistic way under these higher-energy conditions. A comparison of offshore wave climate between years and a discussion of whether the validated conditions adequately represent the extremes simulated would be useful.
Given these concerns i still think the manuscript requires a major revision.

Thompson, C.E.L. et al. (2019) ‘Benthic controls of resuspension in UK shelf seas: Implications for resuspension frequency’, Continental Shelf Research, 185, pp. 3–15.

Referee Report: PDF

Hide

ED: Reconsider after major revisions (10 Feb 2026) by John M. Huthnance

AR by Mari Savela on behalf of the Authors (23 Mar 2026) Author's response Author's tracked changes Manuscript

ED: Reconsider after major revisions (31 Mar 2026) by John M. Huthnance

AR by Mari Savela on behalf of the Authors (11 May 2026) Author's response Author's tracked changes Manuscript

ED: Publish subject to technical corrections (15 May 2026) by John M. Huthnance

Dear Authors
Thank-you for your further revisions. I am now asking for “Technical Corrections” as in the “Detailed comments” below. Then please upload your final version to the Copernicus / OS editorial system for publication. There should be proof reading but there will be no copy-editing by Copernicus.
Thank-you for publishing in Ocean Science.
Yours sincerely
John Huthnance (editor)

Detailed comments.
Line 35 “Consequently, critical shear stress and resuspension potential vary across space and time . .” You use the expression “resuspension potential” several times; this first (?) use suggests that it may be something different from “critical shear stress” but you don’t define it. In view of Referee 1’s concern I think you should provide clarification.
Line 70. Could omit first “and”
Line 72. Better “higher” –> “finer”?
Line 74. “up to” –> “as fine as”.
Figure 3 caption 2nd line. “high” –> “long”
Line 114. “higher” –> “larger”.
Line 146. “is” –> “was”? “have been” –> “were”?
Title Line “Table 2. . .”, line 196 etc. “The value for chlorophyll a is . .” and “a” should be upper or lower case independent of whether “chlorophyll” is abbreviated/
Line 183. “. .greater bulk density . .” I do not understand “higher fines”. Do you mean “. . greater proportions of fines and organic carbon.”?
Lines 247-248. Better “. . moderately larger shear stress values, the largest instantaneous maxima . .” [Or “stronger”, “strongest”}
Line 250. Better “low” –> “small” or “weak”
Line 261. “. . include . .”
Line 277. “beyond” –> “exceeding”
Line 279. Better “highest” –> “strongest”
Line 303. Right justify
Line 318. “are” –> “is”
Line 326. “. . enough to” or delete “be enough to”
Line 329. “emerged” –> “emerge”
Line 348. “meter” –> “metre”
Line 351. Add Holthuijsen (2007) reference as you say in your response.
Line 378. “in” –> “and”
In your response to the .pdf referee comments “Figure 3 is concerning. . .”, you state “Figure C1 in Appendix C show that when the significant wave height was over 0.30 m the agreement between the modelled and measured peak period was good.” However, the reader does not know which points in Figure C1 have significant wave height over 0.30 m. What is the distinction between the red and black points (please explain in the caption)?

Hide

AR by Mari Savela on behalf of the Authors (22 May 2026) Author's response Manuscript

Short summary

Strong motions caused by surface waves can set the material at the bottom in motion. How strong the wave motions need to be depends on the bottom type, for example mud or sand. We estimated how often waves can lift particles from the bottom by comparing wave model results to sea floor samples in the laboratory. We included the effect of seasonal biological activity, which was found to be significant.