GENERAL COMMENTS

The manuscript and accompanying data set submitted to ESSD by M.A. van der Lugt et al. concern a low-energy beach on the lee side of a barrier island of the Dutch coast. The paper presents an extensive data set based on hydrodynamic measurements carried out across and along the beach, repetitive beach profile measurements and surface sediment sampling carried out during a six-week field experiment, these data being complemented by meteorological data.

Detailed hydrodynamic studies conducted on sheltered beaches are limited compared to the ones carried out on open coasts, which makes this data set novel and certainly useful for future research in order to increase our understanding of beach morphodynamics in low-energy coastal environments. In addition, this data set will surely be useful to validate sediment transport models and morphodynamic models that were, in most cases, calibrated and validated with field data from higher-energy open coasts.

The hydrodynamic measurements were carried out with state-of-the-art instrumentation and data processing techniques. The methods are generally described in sufficient details, but additional information could be useful concerning the different sampling frequencies that were used with the hydrodynamic measurements depending on the instruments (see “Specific comments” below). Some explanations may also be useful regarding grain-size analyses of biogenic and siliciclastic sediments (see “Specific comments” below).

In addition to the hydrodynamic measurements, the series of high-frequency (bi-daily) beach profile measurements allows for a detailed analysis of the beach face response to hydrodynamic forcing. The beach profile data set also include beach profile data collected before the experiment (7 days between 16/10/2020 and 18/08/2021) and after the experiment (11 days from 17/12/2021 to 11/03/2023) that are probably not essential in the data set.

The sediment data include cumulative percentile values of the grain-size distibution, but also the mean grain-size, sorting, skewness and kurtosis, based on the geometric graphical measures of Folk and Ward (1957), which are the most commonly used grain-size parameters in the scientific community. Since surface sediment samples were collected on several dates during the field experiment, the data can be used for analyzing sediment sorting processes by combining sediment grain-size statistics and hydrodynamic measurements

The data are easily accessible with the given identifier and appears to be complete. High attention was given to the verification of the quality of the hydrodynamic data, which resulted in the rejection of some suspicious or clearly erroneous data. Error estimates are given and possible sources of errors are discussed in the paper. Specifically, turbidity measurements could not be adequately calibrated to be converted into SSC and the authors preferred to include only the unconverted turbidity signal in the published data set.

Overall, the data set is of high quality and the data are provided in usable formats. The metadata are appropriate, but an indication of the location of the individual beach profiles could be added in the paper (see “Specific comments” below). The paper is well structured and clearly written and the figures are of very good quality, although some elements of context could probably be added regarding the need for implementing a coastal defense at the study site (Prins Hendrik Zanddijk) in 2019 (see “Specific comments” below).

SPECIFIC COMMENTS

- line 13: The authors consider that “sheltered coastlines are traditionally defended by hard coastal structures made of concrete, asphalt or stones”. This statement is a little surprising since sheltered coastlines are by definition protected from the action of high-energy waves that are responsible for coastal erosion. I would rather think that hard coastal structures would commonly be implemented along open coasts as a protection against high-energy waves.

- lines 23-24: It is stated that “beaches in estuaries and on the lee-side of islands are not truly fetch-limited, as refracting ocean swell waves may still reach the shore…”. Although it is true that some beaches in estuaries are not truly fetch-limited, other beaches located in an estuary may be fetch-limited where the mouth of the estuary is protected from ocean swells (by an island for example).

- lines 74-75: it is written that the data were collected “shortly after construction”. After construction of what? Presumably, this refers to the sand nourishment that took place in 2019, but this is mentioned later in the paper.

- lines 89-90: The authors indicate that “the campaign was undertaken at the Prins Hendrik Zanddijk (PHZD), a sandy coastal defense along the Wadden Sea coast of the island Texel constructed in 2019”. Maybe some elements of context could be useful here. What motivated the realization of a beach nourishment on a sheltered beach? Was the coast eroding even if it is partially protected from offshore waves from the North Sea? Was the dyke damaged which required further coastal protection?

- lines 93-97: Some information is given here about waves and tides at the study site, stating that most waves reaching the site are locally generated short waves, but that in certain conditions swell waves can propagate into the basin and can refract towards the coastline, but with diminishing amplitude. Given the configuration of the site, one would expect that wave heights would be strongly lower when reaching the beach and would not have enough energy for causing significant erosion. Is there any information on this?

It is also stated that tidal velocities can reach 1 m/s along the PHZD (or even more under the influence of local winds). Could tidal currents be a major cause of beach erosion at this site instead of waves?

- lines 125-131 (and table 1): It is indicated that the different hydrodynamic instruments recorded data at different sampling frequencies, ranging from 4 to 16 Hz. Any reason why the instruments were programmed to make measurements at different sampling frequencies?

- lines 179-180: the authors indicate that “shells or their fragments were not removed…, so the analysed samples are composed of both biogenic and non-biogenic minerals”. What is the justification for not removing the shells or shell fragments? Although, some scientists prefer to carry out grain-size analysis on the total sediment samples without removing biogenic sediments (because the biogenic material is also a part of the sediment transported by waves and currents), some researchers prefer to remove the biogenic fraction from samples when measuring sediment grain-size. The authors mention that the non-biogenic sand fractions were much more abundant than the biogenic sediments, but it would have been better to carry out grain-size analyses on samples including both biogenic and non-biogenic sediments and on non-biogenic sand and provide the results of both analyses.

- lines 289-290: It is stated that “alongshore (tidal) velocities first decrease with water depth, but then the wave-driven alongshore component increases the total alongshore velocity again in even shallower water (Figure 8f)”. Although Figure 8f actually shows an increase in alongshore current velocity with decreasing water depths on the beach face (from approximately 200 to 210 m), which is presumably related to a wave-driven alongshore component, the figure does not show a decrease in current velocity with water depth over the rest of the beach, but rather similar current velocities.

- Figure 2: It could be useful to add the number of the beach profile on each transect shown in the figure. This would be helpful for quickly locating the position of the profile along the beach when using the beach profile data.

- Figure 8: The graph 8f shows mean current speeds in the cross shore and alongshore directions with positive and negative values. It could be useful to indicate if positive (negative) values of cross shore currents are onshore-directed (or offshore-directed); this could be mentioned in the figure caption or directly in the graph. The same for alongshore currents: do positive or negative values indicate eastward- or westward-flowing currents?

General comments:

The manuscript by van der Lugt et al.. presents a comprehensive datasets on the hydro- and morphodynamics of a sheltered beach located on the lee side of the barrier island Texel in the Netherlands. This dataset mostly originates from a nearly 6 week-long field campaign organised during Fall of 2021, where intensive efforts were made to collect hydrodynamic measurements (bottom pressure and currents), repetitive beach profiles, turbidity and surface sediment sampling. The originality of this dataset lies in the scarcity of studies on the morphodynamic of sheltered beaches in comparison to their open coast counterparts.

Though a few things can be clarified in the paper and the .nc files (see more detailed comments), I found that the dataset was well organised and all interested users will find their way easily into it. There is no doubt that the dataset will be useful in future studies to study the morphological evolution of sheltered beaches in response to various forcings (wind, current, waves and their combinations). This is especially relevant for temporal scales corresponding to meteorological events, but also potentially up to interannual timescales, with the aid of validated numerical models.

Overall, the manuscript is relatively well written and organised, but I found that the introduction could be improved. I feel that the authors fail to find the right balance between the specificity of the site, and how the dynamics at this location is representative and similar to other low-energy sites around the globe: in other words, is this dataset going to help find generic processes, which can in turn be useful elsewhere? Or instead, is this study more relevant to the North Sea context? There is a sort of review on "sheltered" beaches, whose definition is not completely clear (is this the role of this paper to define such morphological features anyway, I wonder?). Part of my feeling also comes from phrases jumping from the Dutch context to generalities, which are sometimes extrapolated. For instance, the first phrase: “Sheltered coastlines are traditionally defended by hard coastal structures made of concrete, asphalt or stones.”. I am not a specialist but I do not think this actually the case everywhere.

Besides the introduction, I have two other major comments that I think should be addressed before a future resubmission :

1 - The authors often refer to “sediment transport” as measured or, at least, that can be studied with the present dataset. But is this really the case? Turbidity measurements could not be transformed into SSC, so how can the dataset actually serve to validate sediment transport (not morphological changes)? Because of this, I would tone down the sediment transport part, and rather focus on a dataset capturing the morphological evolution of the site as well as describing the spatio-temporal evolution of sediment characteristics. It does not mean that the dataset is less useful to model the site morphodynamics, there just is less means to validate sediment transport directly.

2 – I have concerns over the processing of the bottom pressure, and the bulk wave parameters provided in the paper. My attention got caught especially in Figure 8, with the much larger Hm0 offshore compared to other sensors, without an obvious explanation. Reconstructing the free surface elevation from bottom pressure is a problem that I know well, and I am fully aware of the challenges in the present dataset for reconstructing a signal with such short waves, and sometimes in the presence of relatively strong non-linearities. In short, there is no ideal solution, but at least the problem should be acknowledged, and the related uncertainties into bulk parameters quantified. First, there is no mention of the cutoff frequency for the linear correction, except at line 209 where 1 Hz is noted but this is surely not the correct one. Here, I analysed several bursts of data at the OSSI L2C10 (similar behaviour is observed at C9, I have not checked at other sensors) as follows. 40-min of data were taken during 4 situations of “low” and “high” energy conditions (2 each), including the one chosen. Note that the behaviour I describe next could be found in most bursts I extracted and tested. Over this 4 bursts of data, I computed power spectra of the hydrostatic (ζ_hyd  in figures)and linear (ζ_L) reconstructions of the free surface elevation from the detrented pressure signal using 60 Hann-tappered blocks of 64 s overlapping by 50% (effective d.o.f. are ~112). From these, the significant wave height H_m0 and mean wave period T_m02 are computed using a range of different cutoff frequencies. Lastly, the bicoherence b, a measure of the bound energy at a given frequency, was computed from ζ_hyd following Hagihira et al. (2001). From the provided plots, several important remarks can be made:

• Bulk wave parameters are extremely sensitive to the cutoff frequency, especially H_m0. Let us consider the energetic example of Figure 8 (19 September 2021 around 08:15): integrating up to 0.5 Hz gives Hm0 ~ 0.5 m, while integrating up to 0.8 Hz gives more than the double! In terms of energy, that is a factor four. Here, it should be noted that by looking at the hydrostatic reconstruction, choosing 0.8 Hz as a cutoff does not look incorrect, since it still corresponds to energy levels 2 orders of magnitude above noise level.

• The cutoff frequency chosen by the authors seems to vary in time, and sometimes even between parameters. In order to retrieve the values stored in the 'tailored' data, a cutoff frequency at 0.6 Hz should generally be applied, but it actually seems to vary between 0.4 Hz (13 October 00:15) to 0.9 Hz (1 October 02:45), and potentially less or more, respectively, since I only checked a limited amount of data bursts.

• Non-linearities seem relatively strong, and interactions around the peak frequency or between distinct bands of frequencies seem both intense and pretty common in the present dataset. Those explain large fractions of bound energy at high frequencies, leading to a potentially large overestimation of the pressure-to-surface transfer function (e.g., Martins et al., 2021), which affects the computation of bulk parameters. As shown in some of the examples provided, the blow-up of the linear reconstruction occurs rather “fast” in some cases, e.g. well below frequencies where noise start to dominate the pressure signal.

As mentioned above, I acknowledge the fact that there currently does not exist a method adapted to such wave conditions (non-linear reconstructions being currently limited to weakly or moderately dispersive regimes). Furthermore, the raw pressure data is provided, so that informed users will be able to decide themselves how to compute bulk parameters. However, a large fraction if not most users will not be fully aware of the issues, and will be mostly interested in bulk products as computed by the authors. In this context, and in my opinion, the issues related to the reconstruction of such short waves often undergoing non-linear processes should be acknowledged, and the uncertainty on bulk quantities shoud be estimated and discussed. Here I chose bursts of data at high tide, where we expect minor influence from the current, but considering the wave periods measured here, this is definitely another aspect that need to be discussed as it unsure whether the authors account for it for computing wavenumbers.

The figures are attached in the supplementary file.

Several other comments or suggestions are provided below

More specific comments:

#1 – L19: along the lines of my comment above on the introduction, what then makes them different than other sandy beaches? What about winds?

#2 – The cross-shore and longshore referencing used is really hard to follow in the paper, as it changes between Figures. I found no information on the coordinate system used (only vertical referencing), and in this regard, maybe the mean coastline orientation could be provided so that users could convert easily between real and local coordinate systems (includent long- and cross-shore directions).

#3 – The unit of raw pressure was not intuitive (Pa around a mean Pa value relative to NAP). At this point, they could be provided directly in m (hydrostatic) relative to NAP or please clarify somewhere.

#4 – L214: why only resticting to such a narrow frequency band? First, it becomes really sensitive to the definition of the peak frequency, which is clearly hard to define in such wave conditions. Second, the bottom pressure “contains” what reaches the bottom in terms of non-linearities so that including frequencies up to the noise level will be much more representative of wave non-linearities.

#5 – Sections 8.2 and 8.3 feel light. There is no mention of Figure 9 in section 8.2.

#6 – All data files have different time origins. Is this the best choice for processing multiples files?

#7 – The table indicates 30 min bursts for the ADV L2C10 while the data actually suggests it is 10 min. Please verify that the figures provided in the table are correct.

#8 – In relation to #7: the Table indicates that the sampling frequency of L2C10SOLO is 8 Hz, while the data file indicates 10 Hz. By comparing with the pressure from the ADV, it clearly is 8 Hz. Please verify that all netcdf files include the correct information. As reviewers, we cannot verify every individual files…

#9 – I felt that more information could be provided within the netcdf files, and fields could be more descriptive in general.

References:

Hagihira, S., M. Takashina, T. Mori, T. Mashimo and I. Yoshiya, 2001. Practical issues in bispectral analysis of electroencephalographic signals. Anesthesia and analgesia 93, 966–970. doi: 10.1097/00000539-200110000-00032.

Martins, K., P. Bonneton, D. Lannes and H. Michallet, 2021. Relation between orbital velocities, pressure, and surface elevation in nonlinear nearshore water waves. Journal of Physical Oceanography 51(11), 3539–3556. doi: 10.1175/JPO-D-21-0061.1