The two field campaigns were designed to understand and quantify sediment exchange at two different spatial scales: (1) between different basins, and (2) between channels and shoals within a basin. These objectives steer (i) the spatial design, (ii) measurement period, (iii) measurement duration, and (iv) instrumentation. These design aspects and the installation of the measurement frames are explained in greater detail below, followed by more detailed descriptions of the content of the resulting Winter 2023–2024 and Spring 2025 campaigns in the next sections.

All measurements were carried out south of a barrier island (Ameland) centrally located in the Dutch Wadden Sea (Fig. ). The scientific motivations to select this part of the Wadden Sea were (1) the presence of channel-shoal transitions in close proximity to a tidal divide; (2) the large spatial variation of sandy and muddy sediment beds, with the northern part of the basin being generally sandier and the southern part muddier ; and (3) the influence of anthropogenic activities, including frequent dredging of the navigation channel and associated sediment disposal in the basin, such that sediment dynamics are controlled by both natural and anthropogenic processes. Practical considerations include the accessibility of the site via the Holwerd ferry pier and the Ameland marina, and the availability of complementary data from other monitoring campaigns and programs (discussed in Sect. ). The Winter 2023–2024 campaign (A stations in Fig. ) was primarily done to better understand the spatial variability in flow and sediment transport between two basins, hence we selected the measurement locations around the tidal divide and located in both the west- and eastward basin. The Spring 2025 campaign (B stations in Fig. ) was done to further improve our understanding of the exchange between basins, but also to better understand the (fine-)sediment exchange processes between channel and shoals in the south of the tidal basin. All measurement locations were focused on the back of the basin near the tidal divide, thereby prioritizing a detailed investigation of this area at the expense of capturing the full spatial variability of intertidal areas across the basin.

The key factor controlling the selection of the measurement period was the seasonal variability in wind climate. Wind induces (residual) flow across tidal divides , and wave stirring mobilizes sediment on tidal flats . To sample both storm-driven dynamics (when transport over the tidal divides and erosion of the tidal flats are expected) and calm-weather conditions (typically characterized by accreting tidal flats), both measurement campaigns were scheduled at transitional periods: at the beginning of winter (the Winter 2023–2024 campaign; Sect. ) and spring (the Spring 2025 campaign; Sect. ). The summer season is not captured in the two measurement campaigns, such that effects of biota on sediment dynamics that peak in summer e.g., are not included in the data.

The duration of measurement campaigns reflect a trade-off between deployment duration and sampling frequency, constrained by instrument battery capacity. Extending deployments through servicing was not feasible for practical and financial reasons. Deployment and retrieval were further limited to periods with sufficient emergence of the intertidal measurement locations during daylight hours. For both campaigns, the duration of measurements amounted to approximately 8 weeks.

Measurement locations were equipped with a selection of instruments to measure flow velocities, waves, suspended sediment concentrations, and the local bed level, although not all parameters could be measured at every location due to the limited availability of instruments. Flow velocity profiles were measured with Acoustic Doppler Current Profilers (ADCPs). Acoustic Doppler Velocimeters (ADVs) were deployed for high-frequent measurements of flow velocities (capturing the oscillatory near-bed flows generated by waves) and water level variations. The ADVs were mounted vertically, such that they also measure the distance to the bed, hence the variation in bed elevation, at the start and end of each burst. The acoustic backscatter of the ADVs may be used as a measure for the suspended sediment concentration (SSC) , but all measurement locations were also equipped with at least one optical backscatter sensor (OBS) to determine the turbidity, used to estimate the SSC (see Sect. ). Local water levels are derived from the pressure measurements of ADCP and ADV instruments. Some of the measurement stations in the second (Spring 2025) campaign were equipped with an A-SED for acoustic measurements of the distance to the bed, similar to the ADV bed level measurements but carried out at an acoustic frequency of 400 kHz instead of 6 MHz. Hourly atmospheric pressure measurements at the permanent meteorological measurement station at Hoorn, Terschelling (Fig. ) are available to correct pressure recordings for variations in atmospheric pressure. During the Spring 2025 campaign, however, we installed a pressure sensor in Holwerd (i.e., closer to the measurement site).

Most measurements stations were installed on intertidal areas, where frames and instruments were mounted in situ during low tide, when the seabed was exposed. Locations on a fringing mudflat were accessed on foot from the Holwerd ferry pier. The other intertidal locations (Fig. ) were reached by grounding a ship (prior to low tide) followed by transport on foot, with materials transported on (floating) sledges. All frames were marked with surface buoys. The single subtidal station (Spring 2025 campaign) was installed by lowering an instrumented lander from a vessel during high tide. The lander was secured in position using a weighted anchor and marked with multiple surface buoys. A series of photographs in Fig. illustrates the installation of the measurement stations.

The measurement stations in the Winter 2023–2024 campaign were installed on 21 November 2023, and retrieved on 16 January 2024. The observed wind conditions over this period are illustrated in Figs. and a and the water level variation at a nearby (permanent) measurement station (Nes) is illustrated in Fig. b. Winds were relatively strong during the measurement period (mean wind speed 7.7 m s⁻¹ compared to an average wind speed of 6.2 m s⁻¹) and frequently directed from the west (Fig. ). Winds from the northern and southern quadrants were hardly observed. The passing of four storms (Elin on 9 December, Pia on 21 December, Gerrit on 27 December, and Henk on 2–3 January) caused multiple periods with strong winds during the observation period (with the maximum observed hourly mean wind speed exceeding 20 m s⁻¹ on 3 January) and elevated water levels. Storm Pia on 21 December induced an exceptionally high water level set-up along the Dutch coast, such that all storm surge barriers in the Netherlands were closed for the first time ever. The resulting water levels in the Dutch Wadden Sea (3.19 m NAP at Nes; Fig. ) were the highest since 2006.

Measurement stations A2 and A4 were located on the shoals along two different channel distributaries in the basin west of the tidal divide of Ameland (Fig. ). Stations A1 and A3 were located on the tidal divide, in between the channel ends of both tidal basins. Station A5 was located just east of the tidal divide. Each measurement station was equipped with an upward looking ADCP (Nortek Aquadopp Profiler) and a sideward looking OBS (Campbell OBS-3+ Turbidity Sensor). The ADCPs were placed within the seabed, with only the head of the instrument emerging above the bed (Fig. ), allowing flow velocity measurements in relatively shallow water. The OBSs were mounted such that their measurement height corresponds to the lowest bin of the ADCPs. In addition, stations A2, A3 and A4 were equipped with a downward looking ADV (Nortek Vector). Unfortunately, the five ADCP + OBS sets stopped measuring after approximately one month because of power failures. Further details of the deployment and the configuration of the different instruments are listed in Table .

The measurement stations in the Spring 2025 campaign were installed on 27 and 28 January (locations B6 & B7), on 6 and 7 February (B1–B5), and on 13 February (B8), 2025, and retrieved on 26–31 March 2025. The observed wind conditions over this period are illustrated in Figs. and a and the water level variation at a nearby (permanent) measurement station (Nes) is illustrated in Fig. b. During most of the measurement period, winds were relatively weak (<10 m s⁻¹) and from varying directions. The strongest winds of approximately 14 m s⁻¹ were from the east, which is exceptional when comparing to the average wind climate (i.e., strong winds from the east are uncommon), causing a water level set-down exceeding 1 m at the measurement site.

Measurement stations B5, B6 and B7 were located at the tidal-divide (Fig. ), with the bed elevation (at deployment) ranging from -0.77 m NAP to -0.56 m NAP (Table ). Station B1 was located at the end of a tidal channel. Stations B2–B4 and B8 were located on the fringing mudflat adjacent to this channel, with stations B3, B2, and B8 in a transect from the channel edge towards the salt marsh. Each station was equipped with an upward-looking ADCP (Nortek Signature, Aquadopp Profiler, or HR Profiler), either mounted on a frame or placed within the seabed, and one (B7), two (B1, B3–B6 and B8) or three (B2) turbidity sensors (Campbell OBS-3+, Seapoint STM, or WiMo turbidimeter). Frame-mounted ADVs (Nortek Vector) were installed at all stations except B7 (Fig. ), with A-SEDs at stations B1, B2, B5, B6 and B8. The two WiMo Multiparameter Sondes (MPP) at B1 were not only logging turbidity, but also conductivity and water temperature. Configurations of frames and instruments varied slightly among the stations due to equipment availability. Full details are provided in Table .

Bed sediment samples were collected during both campaigns at all measurement locations except from A1, A5, and B1, to determine the (dry) sediment density and grain size distribution (GSD), and for the Spring 2025 campaign also the organic content and shell material fraction. The samples were taken by filling 50 mL jars with sediment from the top ± 2 cm of the seabed. Wet and dry densities (Table ) are determined based on the loss of mass after (overnight) drying in the oven at a temperature of 105 °C, assuming a water density of ρw=1025 kg m⁻³ and a density of solids of ρsol=2600 kg m⁻³.

The grain size distribution (GSD) of the dried samples of the Winter 2023–2024 campaign was determined using a Malvern Mastersizer 3000 after pre-treating the samples using a combination of deflocculant (Na4O7P2⋅10H2O) and ultrasone. Two subsamples of the Spring 2025 samples were analysed for GSD using a Malvern Mastersizer 2000, one without and one with treatment with HCl (1 mol L⁻¹) to remove shell fragments. A third subsample was analysed with a TGA701 Thermogravimetric Analyzer for the organic content and the (mass) fraction of carbonate minerals. The fraction organic matter (%OM) and carbonate minerals (%CaCO₃) are computed from the loss of sample mass during heating as follows: 1%OM=mdry-m550°Cmdry⋅1002%CaCO3=m550°C-m800°Cmdry⋅MCaCO3MCO2⋅100,MCaCO3MCO2=100.0944.01≈2.273 with mdry the mass of the sampling after drying at 105 °C, m550 °C and m800 °C the mass of the sample after heating the sample until 550 and 800 °C, respectively, and MCaCO3 and MCO2 the molar mass of CaCO₃ and CO₂ in [g mol⁻¹]. Several characteristics of the sediment samples are listed in Table , indicating the muddy (clayey) measurement locations in the south and sandy measurement locations towards the north of the tidal basin.

The OBS turbidity sensors were calibrated in a tank using suspensions with known sediment concentrations to convert the analogue sensor output to suspended sediment concentration (SSC). Calibration suspensions, ranging from clear water to 8000 mg L⁻¹, were prepared from local bed samples from locations B3 and B4 (with psand = 9 %–23 % and pmud = 91 %–77 %; Table ), because collecting water samples during deployment and retrieval was not feasible at intertidal measurement locations. Instrument-specific calibration curves are derived by fitting a linear relationship between the OBS signal (in counts or mV) and SSC (in mg L⁻¹) at low SSC, and a quadratic relationship at higher SSC. Figure illustrates the calibration curves for the Seapoint STM and Campbell OBS-3+ deployed at location B3. The OBS-3+ has a linear response up to approximately 1000 mg L⁻¹, whereas the STM remains linear up to about 1500 mg L⁻¹. At SSC > 4000 mg L⁻¹, the STM backscatter intensity decreases for increasing SSC, indicating that concentrations exceed the instrument's reliable operating limit.

Burst-averaged pressure measurements from ADVs and ADCPs are used to determine the local water level, and the high frequent (16 Hz) ADV pressure measurements are used to determine wave spectral statistics (as discussed in Sect. ). The pressure is converted into a water level in two steps. First, the offset of the different pressure sensors follows from the difference between the recorded pressure and the atmospheric pressure at a moment at which the instruments are not immersed (tref, typically the first low-water period after deployment), and is subtracted from the data. Next, the pressure recordings are corrected for atmospheric pressure variations, by subtracting the recorded atmospheric pressure relative to tref. The water depth above the pressure sensors is determined by dividing the air-pressure corrected recordings with the mean water density (ρ=1025 kg m⁻³) and the gravitational acceleration (g=9.81 m s⁻²), and yields the local water level after adding the measured elevation of the pressure sensors in the field.

During emerged conditions, some of the pressure recordings from our ADVs and ADCPs were obviously erroneous (> several kPa). The pressure sensors in these instruments are temperature-compensated, and we expect the errors were caused by rapid, uneven heating of the instruments during sunny days, or cooling during cold nights (i.e., when the air temperature is significantly different from the water temperature; Fig. ). Hence, the pressure data could not reliably indicate the periods of inundation of the instruments. Low-water periods are therefore removed from the data based on the amplitude of the acoustic signal of the velocity measurements, which is (too) low when the instruments were not immersed. Low-water periods are removed from the ADCP data by applying a minimum amplitude threshold in the lowest velocity bin (see Sect. ).

Processing of the ADV and ADCP velocity measurements consists of three steps: (1) filtering outliers; (2) excluding data from periods when the instruments, or, for ADCPs, individual velocity bins, were not immersed; and (3) converting velocities from probe coordinates (XYZ) to field coordinates (ENU).

ADV recordings are flagged as outliers if the instantaneous inter-beam correlation does not exceed 62 %, if the instantaneous velocity exceeds the instrument's velocity range, or if the instantaneous measurements are more than 3 standard deviations from the mean over a burst. The correlation threshold of 62 % follows from the criterion proposed by : 3cmin⁡=0.3+0.4⋅sfsf0⋅100% with sf the sampling frequency of ADV measurements of 16 Hz and sf0 the reference frequency of this criterion of 25 Hz. The horizontal and vertical velocity range of 3.5 and 1 m s⁻¹, respectively, follow from the nominal velocity range of ±2.0 m s⁻¹ that is set while programming the ADVs. The ADCP measurements are processed with a minimum inter-beam correlation threshold of 50 %, which satisfies the criterion for acceptable measurement accuracy by for the sampling frequency of 1 Hz. Only the burst-averaged ADCP measurements are included in the filtered and tailored datasets. The almost continuous Aquadopp HR Profiler measurements at location B6 are averaged over 10 min intervals.

Low-water periods, during which instruments were not submerged, are excluded from the dataset based on the acoustic signal amplitude rather than the local water level, as pressure recordings during emerged conditions proved to be unreliable (see Sect. ). For the ADVs, low-water periods were defined as intervals where the burst-averaged signal-to-noise ratio (SNR) did not exceed 10 dB (Fig. ). For the Aquadopp (HR) Profiler ADCPs, a beam amplitude of 110 counts in the lowest bin is used as threshold for inundation of the instrument, and a threshold of 75 dB is used for the Signature ADCP. Inundation of velocity bins beyond the first is subsequently verified using the local water level, as submergence of the instrument, hence reliable pressure recordings, is already ensured by the first-bin check.

The ADV velocity measurements are converted from XYZ to ENU coordinates by rotating the reference frame, using instrument orientations measured in the field with DGPS. The raw ADCP data are already in ENU coordinates, owing to the instruments' internal compasses, but a correction for the Earth's magnetic declination is applied because true north slightly differs from magnetic north: 2.6° for the Winter 2023–2024 campaign and 2.8° for the Spring 2025 campaign.

Wave spectral statistics are included in the tailored dataset and were derived per burst from ADV measurements using linear wave theory. Sea-surface elevation was reconstructed from near-bed pressure recordings following the approach of , which accounts for short-crested wind sea waves. The variance density spectrum of the sea surface was determined according to: 4PSS=SW2Phydrostatic5SW=cosh⁡(kd)cosh⁡(kh) with PSS the variance density spectrum of the sea surface, Phydrostatic the variance density of the hydrostatic surface elevation, k the wave number, d the local water depth, and h the instrument height above the bed. Sea-surface reconstruction was performed over the frequency band 0.125–1 Hz, which encompasses the short-crested wind sea waves observed at the measurement site, with peak frequencies up to 0.5 Hz. Energy at frequencies below 0.125 Hz is negligibly small at our measurement site and therefore omitted. To limit amplification of high-frequency noise within the integration range in the power spectrum, the linear transfer function SW was capped at a maximum value of 5 and tapers to zero between 1 and 1.5 Hz. Further details on the sea-surface reconstruction method are provided by .

The spectral analysis yields the significant wave height Hm0=4m0, three mean wave periods Tm-1,0=m-1m0, Tm0,1=m0m1, and Tm0,2=m0m2, the peak wave period Tp, and the smoothed peak wave period Tps. The mean wave direction θmean (and directional spreading σθ) are determined following the method of maximum entropy . All wave statistics are included in the tailored dataset.

Instrument-specific calibration curves (see Sect. ) are used to convert the analogue signals of turbidity sensors into estimates of suspended sediment concentrations (SSC). To ensure values remain within each sensor's reliable operating range, SSC derived from Seapoint STMs is capped at 4000 mg L⁻¹, whereas SSC from the other turbidity sensors is capped at 8000 mg L⁻¹, corresponding to the upper limit of the calibration data. Periods during which turbidity instruments were exposed above the surface are excluded, consistent with the filtering applied to velocity measurements at the same height above the bed. No further filtering or despiking is applied. Suspended sediment transport rates are calculated by multiplying SSC with the corresponding flow velocities (i.e., [m s⁻¹] ⋅ [kg m⁻³] = [kg m⁻² s⁻¹]).

ADV and A-SED acoustic measurements of the instrument-to-bed distance are combined with instrument elevations obtained from DGPS measurements to produce time series of the local bed level, referenced either to NAP or to the bed level at deployment.

The dataset enables quantitative analysis of flow and sediment transport under varying forcing conditions by tides, wind, and waves. This section illustrates the dataset content and its potential applications using data samples of measured current velocities, and bed elevation dynamics and SSC. Only a subset of the measurement locations is highlighted here; similar data are available for the other measurement locations.

Limitations of the measurement data originate from instrument accuracy, turbidity sensor calibration, and data processing. The accuracy of high-frequent acoustic flow velocity measurements (bias <1 % of the measured velocity; ) is considered negligible for typical applications in coastal and estuarine systems. Turbidity sensors have an upper operating limit due to signal saturation , but this limit was not reached during the field measurements. The main source of uncertainty in SSC timeseries follows from changes in the suspended sediment composition over time. The relationship between turbidity and SSC (Sect. ) is sensitive to the sediment characteristics, as particles of different sizes, shapes, and mineral compositions scatter light differently . Optical turbidity sensors were calibrated with one type of sediment mixture. Therefore, the calibration does not reflect temporal changes and spatial variations in suspended sediment composition. Resulting SSC values (in mg L⁻¹) may therefore deviate from the actual concentrations. Changes in sediment composition may be estimated from acoustic and optic instruments deployed in parallel , but such an analysis has not been performed yet.

Wave spectral statistics are derived from near-bed pressure recordings, using a wave attenuation correction factor (SW in Sect. ) to reconstruct the short-crested wind sea waves. A cap on the correction factor prevents blow-up of noise, but also underestimates the variance density of high-frequency waves. Wave spectral statistics are therefore unreliable when high-frequency waves (>1 Hz) are dominant. Such conditions occurred frequently during the calm weather periods in the Spring 2025 campaign.