Changes in tidal dynamics in response to sea level rise in the Sylt-Rømø Bight (Wadden Sea)

Global sea levels are projected to rise by up to 0.84 m (0.61–1.10 m) by the end of the century under the SSP5-8.5 (Shared Socioeconomic Pathways, scenario 8.5), relative to 1986–2005 (Intergovernmental Panel On Climate Change (IPCC), 2019; Oppenheimer et al. 2019). Despite the SLR projections in the Wadden Sea (0.76 ± 0.23 m by 2100 under SSP5-8.5, Vermeersen 2020) being close to the mean global rates, the impact of the SLR on such low-lying coastal areas can be particularly strong. Stretching along the coasts of the Netherlands, Germany, and Denmark, the Wadden Sea is the world’s largest coherent tidal flat system that entered UNESCO’s list of Natural Heritage List sites in 2009 for its status of "Outstanding Universal Value" (CPSL 2010, 2005, 2001).

The Wadden Sea has historically shown resilience to natural sea level changes due to its large sediment budget and capacity to migrate landward. However, this process is now hindered by human activities, including settlements and coastal protection structures such as dikes and causeways. Furthermore, the historical trend of increasing sediment accumulation in response to sea level changes may not be sufficient to compensate for the accelerating SLR rates in the current days (Becherer et al. 2018; Hofstede 2015; Hofstede et al. 2018). Rising sea levels not only impact the inundation frequency and severity but also substantially alter the hydrodynamic characteristics.

The previous research has largely focused on broader trends in morpho, hydro- and ecological shifts across the Wadden Sea/North Sea. Contrastingly this study aims to provide a localized, in-depth analysis, focusing on the hydrodynamics of the Sylt-Rømø Bight. As a semi-enclosed basin, the Sylt-Rømø Bight is unique due to its limited exchange with adjacent basins, restricted by lateral causeways. This isolation makes it a natural laboratory for exploring the hydrodynamical response to SLR, which has been shown to differ significantly across the Wadden Sea. For instance, Lepper et al. (2024), in their recent study on tidal characteristics shaped by SLR and bathymetry, showed that even neighbouring basins with slight differences in bathymetry and geometry can experience contrasting changes. As a result, while broader-scale studies provide valuable insights, they may not fully capture the localized dynamics of individual basins, highlighting the necessity of complementing regional studies with site-specific investigations. Wachler et al. (2020) also emphasized this variability by examining the differences between ebb and flood tides in the tidal channels of the German Wadden Sea.

This variability is particularly evident in the Sylt-Rømø Bight, where the basin’s isolation has resulted in relatively stable morphodynamic conditions compared to other parts of the German Wadden Sea basins (e.g. Benninghoff and Winter 2019; Hagen et al. 2022; Jordan et al. 2021). While much of the German Wadden Sea has been effectively accumulating sediment at rates that surpass SLR, leading to the expansion of intertidal flats and the deepening and narrowing of subtidal zones over recent decades, the Sylt-Rømø Bight has experienced low sediment accumulation, with its morphodynamics constrained by the causeways. Becherer et al. (2018) project that there will be no significant sediment growth or depletion in the Sylt-Rømø Bight by 2050. By the end of the century, the basin is expected to accumulate a larger amount of sediment, corresponding to an average net bed-level accretion of 6 cm and 1 cm under high and moderate SLR scenarios, respectively (Becherer et al. 2018; Dissanayake et al. 2012; Huismans et al. 2022).

Although the projected sediment accumulation in the Sylt-Rømø Bight is relatively small, changes in hydrodynamics due to SLR can have significant impacts on the local ecosystem. Our preliminary work, collated in a recent rough synthesis on Wadden Sea change (Buschbaum et al. 2024), considers SLR and associated changes as one of the critical threats to the Sylt-Rømø Bight ecosystem. The potential increase in tidal inundation in intertidal flats, which serve as key habitats for many species, may impact biodiversity and disrupt the functioning of ecosystem services. Additionally, shifts in current dynamics—both in intensity and transport patterns—may place additional strain on the resilience of species inhabiting transition zones. As a result, understanding the changing characteristics of tidal dynamics and tidally induced transport patterns becomes crucial for predicting the broader impacts of SLR on this system.

Focusing on the Sylt-Rømø Bight, we use the FESOM-C coastal ocean model setup with a resolution of up to 2 m to investigate basin-scale changes in intertidal dynamics under different SLR scenarios. In addition, we apply Lagrangian tracking experiments to investigate changes in transport patterns of passive tracers. Our study builds on previous work exploring current dynamics and responses to SLR and bathymetric changes in the Wadden Sea region (e.g., Becherer et al. 2018; Benninghoff and Winter 2019; Fofonova et al. 2019; Wachler et al. 2020).

This set of downscaling experiments addresses two unexplored key questions: (1) How will the extent of intertidal flats in the Sylt-Rømø Bight change under future SLR scenarios, which forms the foundation for understanding subsequent changes in tidal dynamics? (2) How will SLR affect tide-driven transport patterns, including changes in current velocities within the basin?

Our results of the Lagrangian tracking experiments reveal a reduction in outflow and limited exchange between different parts of the Sylt-Rømø Bight under SLR scenarios. These findings can be explained by our analysis of tidal asymmetry, which plays a critical role in shaping transport processes and sediment dynamics in the region (e.g., Dronkers 1986). Tidal asymmetry typically arises from complex topography and the strong influence of non-linear processes within the tidal system, including bottom friction, momentum advection, and tidal amplitudes comparable to water depth. In the Sylt-Rømø Bight, where these conditions are prevalent, we observe a general decrease in tidal asymmetry under SLR scenarios. This reduction likely drives the observed decrease in outflow transport and limits the exchange between different parts of the bight. To generalize these findings and assess their robustness, we conducted additional simulations incorporating wind forcing (excluding storm events). Summer wind forcing is not expected to exhibit significant changes over a 50-year time horizon under different IPCC scenarios, in contrast to winter and spring wind conditions (Ortega et al., 2025). While wind influences transport patterns, our results show that the major transport routes in the bights remain unchanged. The findings regarding the outward transport of passive tracers remain valid, confirming that transport dynamics in the bight are predominantly tide-driven under both reference and SLR scenarios.

The following section provides an overview of the study site. The Methods section describes the model setup, and data sources, including the bathymetry and forcing data used in the simulations, as well as the analysis methods, such as tidal asymmetry and thickness-weighted average (TWA) velocities. The Results presents the outcomes of the simulations, focusing on changes in tidal inundation, current velocities, tidal asymmetry, and particle tracer dispersion under various SLR scenarios, all of which are important components contributing to changes in transport processes. The Discussion interprets these findings and explores their implications for coastal management and ecosystem conservation. The Conclusions summarizes the primary insights derived from the study and suggests avenues for future research. Finally, additional supporting information and figures can be found in the Supplementary Materials.

Like most of the Wadden Sea basins, the Sylt-Rømø Bight is a shallow, tidally energetic basin formed behind two barrier islands – the German island of Sylt and the Danish island of Rømø, separated by the tidal inlet Lister Deep (also referred to in the literature as Lister Tief, Lister Dyb; see Fig. 1). However, unlike other back-barrier basins, the Sylt-Rømø Bight is the only semi-enclosed basin in the Wadden Sea, restricted laterally from adjacent basins by artificial causeways on both sides: Hindenburgdamm and Rømødæmningen, built in 1927 and 1948, respectively (Hansen 1956; Wohlenberg 1953). These causeways limit water exchange with adjacent basins, creating a semi-enclosed system that makes the bight an ideal natural laboratory for studying hydrodynamic processes and their responses to external drivers and stressors. Its relative isolation from adjacent basins allows for a controlled analysis of tidal dynamics, water mass transport, and other coastal processes without interference from neighbouring systems.

The tidal range in the bight averages 2 m, based on observations at the List tide gauge (E.U. Copernicus Marine Service Information, https://doi.org/10.48670/moi-00036). Tidal forcing accounts for over 80% of depth-averaged velocity variability under regular wind conditions, in the absence of storms, and over 90% during spring tides (Fofonova et al. 2019). Large tidal amplitudes, shallow depths, and complex topography of the basin contribute to pronounced tidal asymmetry, which plays a significant role in material transport. Tidal asymmetry characterizes the distortion of tidal waves as they propagate over shallow coastal shelves and enter embayments, leading to unequal durations and strengths of the flood and ebb phases.

The bight receives minimal fluvial input, with small rivers such as the Vidå and Brede Å contributing only 4–10 m3/s of freshwater (Purkiani et al. 2015). These inputs are negligible compared to the tidal volumes exchanged through Lister Deep, making the bight primarily tidally dominated. Water exchange with the open North Sea occurs exclusively through Lister Deep, a 2.8 km wide tidal inlet. At the mouth of Lister Deep, a prominent ebb-tidal delta extends seaward, acting as both a sediment trap and a pathway for sediment redistribution within the bight (Dissanayake et al. 2012). However, located at the northern end of the Wadden Sea, the Sylt-Rømø Bight receives only limited sediment input compared to southern basins (e.g. Colina Alonso et al. 2024; Benninghoff and Winter 2019).

The bight spans a total area of approximately 410 km2 and features a highly variable topography, including extensive intertidal flats (> 45%), shallow subtidal zones (~ 35%), and deep tidal channels (~ 10%). The bathymetry of the Sylt-Rømø Bight is characterized by an average water depth of approximately 4 m, with a maximum depth of about 37 m observed in the tidal inlet Lister Deep (Fig. 1).

One of the prominent features in the bay is Königshafen – a small shallow embayment at the northern tip of the island Sylt. The embayment is sheltered from winds and waves, and experiences semidiurnal tides with amplitudes reaching 1.7 m (Kristensen et al. 2000; Reise et al. 1994), which is comparable with the average bathymetry of the Königshafen area as shallow as about 2 m. Schumacher et al. (2014) also reviewed the evolution of its habitats and described changes with respect to the historical records. According to their study, although eutrophication and invasive species have played a significant role, hydrodynamic alterations have been the primary drivers of changes in the nearshore benthic zone. They are expected to intensify as SLR accelerates.

3.1 Framework

This study uses numerical simulations with the coastal ocean model FESOM-C to study the hydrodynamics and transport processes of the Sylt-Rømø Bay under different SLR scenarios using high-resolution bathymetric data. The results are analyzed to identify shifts in the tidal transport dynamics and the causes of these changes. Two key aspects in the Eulerian approach are investigated: tidal asymmetry indicating ebb or flow dominance and depth-weighted flow velocities integrated over the tidal period and divided by the transect area. We extend the analysis to include the Lagrangian paradigm to analyze the transport dynamics more deeply. This approach is necessary due to the large spatial gradients of the velocity components at distances smaller than the tidal displacement of particles, highlighting the strong influence of nonlinear processes. Combining the two approaches in Eulerian and Lagrangian spaces allowed us to identify the main transport pathways in the system, the mechanisms underlying their spatial patterns, and the changes in SLR scenarios. Although this study does not consider sediment dynamics and focuses on the transport of massless passive particles, the findings regarding tidal dominance in the system can be related to the evolution of the system morphology (see Discussion). The methodology is described in more detail in the following sections, including model setup, validation, and scenario-specific adjustments.

3.1.1 Tidal asymmetry definition and analysis

One approach to gaining insight into the possible changes in tidal dynamics and tide-induced transport partly relies on tidal asymmetry analysis. There are multiple ways to describe and quantify this asymmetry, with the choice of method depending on the study’s purpose and sensitivity to specific environmental and non-linear factors of the study site (Wünsche et al. 2024).

In this study, we define tidal asymmetry using two specific metrics: (1) the mean duration asymmetry, which is the relative difference in the average durations of the flood and ebb phases, and (2) the mean discharge asymmetry, which is the relative difference in the average discharges during flood and ebb. The duration asymmetry between flood and ebb phases is commonly used as an indicator of residual transport for fine suspended sediments, whereas velocity asymmetry is more applicable for estimating the transport of coarser suspended materials (Dronkers 1986; Friedrichs and Aubrey 1988; Hagen et al. 2022). These methods allow us to assess spatial variations in tidal asymmetry across the Sylt-Rømø Bight. In the analysis, we distinguish between ebb and flood phases at each spatial point based on elevation behaviour: rising elevation indicates flood, while decreasing elevation indicates ebb. It is also worth mentioning that the asymmetry in discharge and the asymmetry in mean velocity magnitude display visually similar patterns. Therefore, depending on the context, we may interchangeably refer to differences in currents/velocities or discharges during flood/ebb. The tidal asymmetry calculation was based on 114 flood-ebb cycles, representing two full lunar months (2 ×ばつ 29.5 days), with the mean value taken.

Other definitions of tidal asymmetry focus on different aspects of tidal dynamics. For instance, asymmetry is often expressed as the relative phase differences and amplitude ratios of the overharmonics, such as M2 and MS4, and primary harmonics in tidal elevation, such as M4 (Friedrichs and Aubrey 1988; Jordan et al. 2021; Van Maren and Winterwerp 2013; Hagen et al. 2022), which is effective for point-based analysis rather than domain-wide studies. A strong overharmonic signal can contribute to significant tidally induced net transport, which means a substantial net displacement of particles after a tidal cycle. Although a strong overharmonic signal and spatial maps showing the amplitude ratio of overharmonics to primary harmonics offer valuable insights, they do not directly explain transport processes.

As the basin geometry and hypsometry play an important role in shaping hydrodynamic patterns and tidal distortion (de Ruiter et al. 2019), other methods of asymmetry analysis incorporate basin morphology-related metrics, such as the ratio of intertidal storage to subtidal channel volumes or the relationship between tidal amplitude and mean depth (Dronkers 1986; Friedrichs and Aubrey 1988; Hagen et al. 2022). While this approach is effective for describing the overall ebb or flood dominance of a system, it does not capture spatial variability within the basin. While these methods differ, they are not independent, because tidal asymmetry shows various aspects of the role played by nonlinear processes. By focusing on duration and discharge asymmetry, we provide maps of spatially varying flood or ebb dominating areas that offer more comprehensive insights, which can be particularly useful for analyzing transport patterns and further applied by sedimentologists and ecologists. This is partly demonstrated in the Results and Discussion sections.

3.1.2 TWA velocities

While tidal asymmetry provides information about stronger discharges during ebb or flood, it does not give us information about the net local mass discharge at a given location as a result of the flood-ebb cycle. The thickness-weighted average (TWA) velocity over one tidal period is directly related to the net local mass discharge from a current location and helps to identify the direction of this discharge. The TWA represents the residual mass transport velocity from a particular location over one tidal period or the net discharge over one tidal period from a transect of unit length with varying local depth, normalized by the mean transect area over a tidal period:

where \(\:H\) is the total depth, [m], \(\:T\) is the tidal period, [s], and u is the Eulerian velocity vector, [ms^{− 1}]. Note, that mean Eulerian velocity over tidal cycle \(\:\left(\frac{1}{T}{\int\:}_{0}^{T}\varvec{u}\left(t\right)dt\right)\) does not bring directly meaningful characteristics for the Eulerian residual transport in the areas where the elevation is comparable to depth (Zimmerman 1979; Klingbeil et al. 2019), therefore we do not use it in a current study. The TWA velocities were calculated over a period of two synodic months across multiple tidal cycles for each grid cell (in the FESOM-C discretization the vector field is positioned at the centers of the grid cells) in a depth-averaged sense.

A non-zero TWA indicates the presence of flood/ebb discharge asymmetry. However, the presence of asymmetry alone is not enough to result in a large TWA. If we consider two areas with the same degree of discharge asymmetry, but one is deeper and has stronger currents, the TWA will clearly differ, being higher in the area with stronger currents.

3.1.3 Tidally induced net transport and tidal residual circulation

The TWA has the velocity dimension and represents a residual mass transport velocity in the Eulerian framework at the current location. In the Lagrangian framework, the tidally induced net transport is the distance between the final position (after a tidal cycle) and the initial position of a released passive tracer. Note, that the trajectory of a tracer within a cycle can be elliptical or very complex, but we do not consider it here, we are interested only in a net displacement. This net displacement divided by the tidal period will be referred to as tidal residual circulation (note, that the meaning of tidal residual circulation varies across different literature sources). In other words, tidal residual circulation is equal to the Lagrangian residual velocities, when in the system only tidal forcing is present:

where \(\:a\) is considered tracer, \(\:\varvec{x}(a,0)\) represents its initial location in space, \(\:\varvec{x}(a,\:T)\) is final location after a tidal cycle, \(\:{\varvec{u}}^{L}\left(a,t\right)\) is the tracer velocity, [ms^{− 1}], \(\:\langle\:{\varvec{u}}^{L}\left(a,t\right)\rangle\:\) is Lagrangian residual velocity, [ms^{− 1}].

The Lagrangian residual velocity vector from a tracer’s initial position would correspond to the TWA at that position only if the amplitude of the primary tracer displacement within a tidal cycle is small compared to the distance over which the gradient of the TWA stream function varies significantly. In other words, this requires the spatial gradients in the velocity field to remain small over the distance traveled by the passive tracer within a tidal cycle. However, this condition is not met in our domain, characterized by highly irregular bottom topography and the significant influence of bottom friction and momentum advection terms. Transforming from an Eulerian to a Lagrangian framework using a Taylor series requires all, or at least a significant number of, higher-order spatial derivatives of the Eulerian velocity components to be considered (e.g., Zimmerman 1979). In such a shallow and geometrically complex area, Lagrangian residual vectors also depend on the timing of tracer release (e.g., at slack tide, flood, or ebb). While TWA and tidal residual circulation are not identical, analyzing TWA and tidal asymmetry provides useful insights into transport patterns. Increased asymmetry in the whole system, driven by enhanced nonlinear processes, generally indicates greater tidally induced net tracer transport (as in a linear, non-dissipative system, the tracers would return to their initial location after a tidal cycle), whereas decreased asymmetry suggests reduced net tracer transport.

3.2 Experiment description

Three experiments were conducted for this study: a reference scenario (REF), representing current conditions, as well as low (LSLR) and high (HSLR) sea level rise scenarios for 2050 (Table 1). Projections for 2100 were also simulated, though the primary focus of this paper is on 2050 due to relatively lower uncertainties in short-term projections of climate-driven SLR and SLR-related processes, which allows us to make more robust assessments (Oppenheimer et al. 2019). The results of the scenarios for the end of the century are provided in the Supplementary Materials.

We applied a downscaling approach for future scenarios using the projections for future sea level and morphological changes. Each simulation was run for over five months, with the last two lunar months (2 ×ばつ 29.5 days) used for analysis. The first ~ 3 months were treated as a spin-up period and were excluded from the analysis to ensure the model reached dynamic equilibrium.

3.3 Model description and setup

3.3.1 FESOM-C model

The simulations are conducted using the numerical model FESOM-C (Androsov et al. 2019), a specialized coastal branch of the global Finite-volumE Sea Ice–Ocean Model (FESOM-2, Danilov et al. 2017), designed to enhance the resolution and accuracy of coastal oceanographic simulations. It inherits several numerical features from FESOM2, including the finite-volume cell-vertex discretization, but introduces significant modifications tailored for coastal applications. The FESOM-C model functions on various mesh configurations such as triangular, quadrangular, and hybrid. FESOM-C has been validated in numerous idealized and real-world scenarios, demonstrating its robustness and versatility in simulating complex coastal processes (e.g., Fofonova et al. 2021, 2019; Kuznetsov et al. 2024, 2020; Neder et al. 2022; Sprong et al. 2020).

3.3.2 Mesh

The setup utilizes an unstructured hybrid mesh consisting of 208.345 nodes and 211.545 elements. Figure 2a provides an example of element distribution in a zoomed-in view. Due to the semi-enclosed state of the bight, the mesh contains a single open boundary (38 km) at the seaward edge of the domain, connecting the basin with the North Sea. The horizontal spatial resolution varies from up to 2 m in wetting-drying zones to 304 m in the deeper outer part (near the open boundary).

3.3.3 Model parameters

The experiments were carried out running 2D barotropic simulations with the wetting/drying option enabled to capture the periodic submergence and exposure of intertidal areas. The model timestep is set to ~ 0.25 s, with data output every ~ 20 min of simulation time. The bottom friction coefficient is applied as 0.0025, a value identified as optimal in prior studies of the same study area when using TPXO tidal solutions (Fofonova et al. 2019).

3.3.4 Bathymetry and coastline topography

In our reference scenario, we used up-to-date high-resolution bathymetry data. A continuous dataset was produced based on the combination of multibeam echosounder (MBES) survey data recorded with the AWI research vessel "Mya II" (ELAC SeaBeam 1180) for water depths deeper than approximately 5 m and a digital elevation model based on topo-bathymetric LiDAR surveys supplied by LKN.SH. Both data sets were recorded in 2017 and gridded to a resolution of 10 m, after which they were integrated into our model setup (Fig. 2b). While this may not explicitly capture the finest-scale features, e.g. 1–2 m wide micro-channels, it effectively represents the key characteristics of the bight, providing a robust foundation for accurate hydrodynamic modeling.

For future scenarios, geomorphological changes were incorporated into the model by adjusting the bathymetry using data from Becherer et al. (2018). These adjustments account for the expected morphological changes driven by rising sea levels and were applied as differences to the originally surveyed bathymetry to ensure the model accurately reflects the evolving conditions in the basin (Fig. 2c and d). The model was configured with a fixed bed for each scenario, meaning that sediment transport and morphological evolution were not dynamically simulated during the experiments. This approach is particularly suitable for the Sylt-Rømø Bight, where its semi-enclosed nature and limited external sediment supply result in relatively stable bathymetry over the considered timeframe (Benninghoff and Winter 2019; Colina Alonso et al. 2024; Dolch and Hass 2008). Consequently, for the short-term simulation period of two tidal months, the influence of dynamic morphodynamic feedback is expected to be negligible. While this simplifies the analysis, it is assumed that any mismatch between the predefined bathymetry and hydrodynamic forcing has minimal influence on the study’s primary parameters such as tidal dynamics and transport patterns over shorter-term periods.

3.3.5 Tidal forcing

The model was forced tidally at the open boundary to focus specifically on tide-driven flow characteristics. This approach allows for an in-depth investigation of the tidal dynamics in the Sylt-Rømø Bight, particularly the effects of higher harmonics and over-harmonics (e.g., Fofonova et al. 2019; Stanev et al. 2016), which are critical in such a shallow intertidal basin. Therefore, the tidal phases and amplitudes were prescribed for 13 major tidal harmonic constituents (M2, S2, N2, K2, K1, O1, P1, Q1, Mm, Mf, M4, MN4, MS4, 2 N, and S1) and two over-harmonics (M4, MS4) based on TPXO9 tidal atlas (Egbert and Erofeeva 2002), which shows robust and one of the best performance for the North Sea among available tidal solutions (Fofonova et al. 2019). To ensure the validity of boundary conditions, the model results were verified by analyzing time series from tide gauges, with details provided in subsection 3.3.7 Setup Validation.

Although rising sea levels can affect tidal forcing, we opted to keep the tidal forcing unchanged across all scenarios to ensure a consistent comparison of SLR impacts. SLR-induced changes to tidal forcing primarily occur due to modifications in basin resonance and bathymetric gradients, which can amplify or attenuate tidal signals (Idier et al. 2017; Pickering et al. 2012; Müller et al. 2011). However, high-resolution future tidal projections for this region, including all necessary tidal constituents for coastal-scale analysis, remain lacking. Given the current limitations in available literature and datasets, we consider the potential error introduced by maintaining fixed tidal forcing to be smaller than the errors associated with approximating future tidal conditions using lower-resolution data.

3.3.6 Wind forcing

In addition to the tidal-only simulations, we conducted experiments with typical summer wind forcing to evaluate its effect on Lagrangian particle transport in the Sylt-Rømø Bight. The decision to focus on typical summer winds was informed by studies such as Ortega et al. (2025), which indicate that, unlike autumn, winter, and spring winds, summer wind events are not expected to show significant trends under future climate change scenarios like SSP5-8.5. This allowed us to apply the same wind forcing in both the REF and future HSLR scenarios, ensuring consistency across our experiments.

To set up the wind forcing simulations, we utilized hourly wind stress and atmospheric mean sea level pressure data from ERA5 (ECMWF), which provides global atmospheric estimates on a 0.25-degree grid. The data were sourced from the Copernicus Climate Change Service through the Climate Data Store.

Based on the findings from Ortega et al. (2025), we concluded that summer wind forcing does not exhibit significant trends in either magnitude or the number of wind events. A wind event is defined as a period during which the wind speed exceeds a certain threshold and persists from a consistent direction for a sustained duration (Rubinetti et al. 2023). For this study, a wind event is characterized by consistent wind from a particular direction lasting several hours, as observed in typical summer conditions in the Sylt-Rømø Bight. Therefore, as an additional experiment, we conducted a simulation using typical June-July wind forcing, a period of ecological significance as it coincides with the production and dispersion of oyster larvae.

Typical summer wind conditions were determined through an analysis of wind event distributions. The median number of events and their duration for each direction during the June-July period was calculated over the 1959–2021 period (see Table 2). Winds from the northwest (NW) and west (W) accounted for the highest number of hours and frequency of occurrence. We selected the summer of 2009, which closely matched the median distribution of wind events across different directions. Although the total number of wind events exceeded the median, the directional distribution remained similar to the median.

3.3.7 Setup validation

The model was validated using tidal gauge (TG) data from stations List (55.017 N, 8.441 E), Vidå (54.967 N, 8.667 E), and Havneby (55.087 N, 8.565 E), displayed in Fig. 1. These data were obtained from the European Marine Observation and Data Network (EMODnet) database, which aggregates data from the Waterways and Shipping Office in Tönning (for List TG) and the Danish Meteorological Institute and Danish Coastal Authority (for Vidå TG and Havneby TG). For spectral analysis, we used data with a 10-minute resolution covering the period from mid-2014 to the end of 2017 and applied the T-TIDE tool (Pawlowicz et al. 2002). The tidal constituents M2, S2, N2, K1, O1, Q1 and M4 were selected due to their high relevance in the tidal dynamics of the Sylt-Rømø Bight. We tested the selected tidal components’ root mean square deviation (RMSD) against the observations (see Table 3). RMSD is a statistical measure used to quantify the difference between predicted and observed values, which, in our case, provides an estimate of the model’s accuracy by calculating the square root of the average of the squared differences between the predicted and observed amplitudes and phases.

The comparison results show that the maximum amplitude error does not exceed 3% of the total tidal maximum (approximately 120 cm), and the phase shift is about 25 min, occurring only at one of the stations—TG Havneby. This discrepancy can be explained by the insufficient bathymetry data in that zone. It is important to note that the differences in the comparison are significantly smaller at the other two points. Overall, this comparison demonstrates the high robustness of the model in zones with significant tidal nonlinearity.

3.4 Lagrangian module application

To assess the transport patterns within the basin, we used the FESOM-C drift Lagrangian module, which has already been applied in previous studies (e.g., Neder et al. 2022; Sprong et al. 2020). Particle motion is calculated on the same unstructured meshes as the hydrodynamic module (FESOM-C), which is an obvious advantage of using the FESOM-C drift model. In the first stage, all the necessary information on horizontal velocities (average on vertical) is interpolated into points with particle coordinates. Its new coordinate is determined after calculating the velocity vector acting on the particle. After that, the process is repeated at a new time step. The time step of the tracer model in this formulation was 20 min. When going beyond the boundaries of the solid contour to determine the position of the simulated particle, two options are possible: removing it from the calculation or waiting for further movement in the element closest to the boundary. When a particle goes beyond the open boundary, it is lost.

In the first series of experiments, the initial position of the tracers was set for nine separate clouds. The position of these clouds of tracers was chosen depending on the dynamic activity of the entire basin. Each cloud had a radius of 500 m and contained 100 particles. The calculation was performed for three weeks, with the restart of cloud particles being performed every three hours. As a result of the calculations, we had a data set of particle positions consisting of 169 implementations, each three weeks long, to assess the spatial dispersion patterns and probabilities. Such a temporal distribution ensured that we covered all the complex tidal conditions in the basin, making the results statistically reliable.

The second set of experiments was conducted to quantify tracers leaving the domain in REF and HSLR scenarios. In these experiments, tracers were released from all grid elements inside the domain, which do not drain during a tide. We calculate about 90,000 tracers for each realisation with the same time interval and calculation period as in the first experiment. This experiment was conducted to compare the outflow patterns in future scenarios with the reference scenario of the current state. Tracers were considered to have left the domain if they moved to within 1 km of the open boundary.

This section presents a detailed analysis of how an increase in relative sea levels impacts the regular tide-driven transport processes within the basin. We begin by examining changes in tidal inundation, which, although largely a direct consequence of the prescribed sea levels in future scenarios and not requiring complex simulations, sets the stage for subsequent shifts in tidal dynamics. Next, we analyze current velocities by their maximum and thickness-weighted average (TWA, Eq. 1) at both the current stage and in the future, reflecting changes in the current regimes and tidal prism. Following this, we study the change in tidal asymmetry, which can signal potential changes in residual currents in such a well-mixed, tidally energetic basin. Finally, we use Lagrangian representation to demonstrate the overall transport pathways within the basin and assess the impact of the changes under SLR scenarios. Together, these analyses reveal changes in net transport in response to SLR, providing insights into the mechanisms at play and their implications for the future evolution of this coastal system. Additionally, we analyze a scenario with wind forcing to assess its role in the transport patterns.

As the changes in both LSLR and HSLR scenarios by 2050 differ mainly by their intensity, however have similar spatial patterns, the figures represent the REF scenario in the left panels and the difference of HSLR – REF in the right.

4.1 Tidal inundation

In the REF scenario, the intertidal areas of the domain comprise more than 47.5% of the total area (Table 4). In the context of the numerical setup, the wetting/drying in the domain varies from 0 (always wet) to 1 (always dry), and the values in between represent the intertidal areas (Fig. 3a). In future scenarios, the Sylt-Rømø Bight experiences a higher probability of inundation of intertidal zones as SLR accelerates. These results are based on the assumption that the bathymetric adjustments incorporated into the model configuration for future scenarios sufficiently account for the long-term morphological response to SLR. As analysis reveals, the areal extent of intertidal zones is expected to decrease by around 13 km² and 21 km² by 2050 under the LSLR and the HSLR scenarios, which correspond to the − 2.2% and − 3.4% of the total area of the domain or -4.6% and − 7.4% of the current intertidal zones. Not only does the area of current intertidal zones shrink, but the drying probability of the remaining intertidal areas is also decreasing on average by 4% (LSLR) to 6% (HSLR). Although overall change is consistently negative over the domain, it also presents some spatial variability (Fig. 3b), where the most subjected areas being in Sønderstrand, Jordsand bank, the north-eastern coast of the basin, and Königshafen areas.

It should also be noted that these analyses have already accounted for the formation of new intertidal zones as a result of greater "wetting" of previously dry zones (primarily in Sønderstrand), and the still decreasing numbers indicate the coastal squeeze – the inability of the intertidal habitats to migrate further landward due to the settlements and causeways. The consequences further intensify as the sea level keeps rising – by the end of the century, the numbers show a two and four-times increase, namely 29 km² and 84 km² corresponding to 10% and 29.4% of the current intertidal areas, for the respective low and high-emission scenarios.

4.2 Maximum velocities

We calculated the maximum velocity magnitudes within the two synodic months period across all time steps. The maximum velocity in the domain is highest (~ 2.5 m s^{− 1}) in the tidal inlet (Fig. 4a). Overall, the spatial average of peak velocities over the domain in the HSLR increases by ~ 3 cm s^{− 1} compared to the REF scenario, which is attributed to the increase in tidal prism and subsequent increase in the water depth and the associated decrease in the influence of bottom friction in response to SLR. The largest increase in the peak velocity is observed in Sønderstrand and Jordsand Bank due to the flooding of areas that have previously been minimally "wet" or even always "dry". Maximum velocities over the intertidal and shallow subtidal zones also increase considerably. In contrast, the channels experience rather small but consistent opposing patterns of change (-4 to -7 cm s^{− 1}, Fig. 4b).

Such an increase in a spatially averaged peak velocity is expected in light of rising sea levels and with that increasing tidal prism. However, little has been discussed about spatial variations within the basin, where a change in current strength has contrasting trends in the tidal channels as opposed to most subtidal and intertidal zones.

4.3 Tidal asymmetry

4.3.1 Duration asymmetry

In the reference scenario (Fig. 5a), most of the basin experiences an asymmetry with shorter ebb durations compared to the flood. This pattern is particularly pronounced in the northern half of the area (flood-to-ebb duration ratio > 1). This is evident in areas such as the Rømø channel, the tidal inlet, and around the Jordsand Bank. In contrast, in the southern part of the basin, a tidal asymmetry with shorter flood durations is more prevalent (flood-to-ebb duration ratio < 1). This creates a distinct spatial pattern where the northern and central parts of the basin have shorter ebb durations relative to the flood, while the southern parts of the basin show the opposite, with shorter flood phases.

In the future HSLR scenario, the overall flood-to-ebb duration ratio decreases across much of the back-barrier basin, indicating a trend towards longer ebb durations. However, this trend primarily results in a weakening of the initial asymmetry, shifting the system closer to a less asymmetric state. Conversely, in the northeastern part of the basin and certain patches in the south, where the flood durations were originally significantly shorter in the REF scenario, an increase in the flood-to-ebb ratio is observed (Fig. 5b). These contrasting trends suggest that, although tidal asymmetry persists in the HSLR scenario, its intensity is reduced, leading the basin towards a more balanced state between flood and ebb durations.

4.3.2 Velocity/Discharge asymmetry

The map of tidal asymmetry in terms of discharges/currents is presented in Fig. 6a. The ebb-to-flood ratio depicted on these maps indicates the relative strength of the discharges/currents during each tidal phase (ratio > 1 indicates stronger ebb currents; ratio < 1 indicates stronger flood currents). Tidal asymmetry is generally more pronounced in areas with large bathymetry gradients or complex topographic features. For instance, in the REF scenario (Fig. 6a), a significantly stronger ebb is observed along the tidal channels, where there are steep bathymetric gradients, which suggests that in these regions, the outgoing tides have a more significant influence on net water discharge. On the other hand, some areas, such as the southern part of the basin, the Rømø channel, and the region between the tidal inlet and Jordsand Bank, show more substantial flood-related discharges.

Further analysis of the difference between the HSLR and REF scenarios (Fig. 6b) reveals an overall decrease in the ebb-to-flood ratio across much of the basin. Such a decrease in the ebb-to-flood ratio suggests that areas that initially exhibited stronger ebb are now experiencing a less pronounced asymmetry, with the difference between ebb and flood discharges becoming smaller. Meanwhile, in regions where flood currents were already stronger, this decrease in the ebb-to-flood ratio (ratio > 1 means that discharges during flood are more substantial than during ebb) corresponds to an increase in the relative strength of flood currents, further accentuating the asymmetry in favour of flood flows in these areas. Interestingly, this pattern does not apply to the whole domain, and the areas that previously experienced significantly stronger flood currents now show a slight increase in the ebb-to-flood ratio, which suggests a decrease in the strength of the flood currents relative to the ebb currents. This pattern indicates that the asymmetry in these regions is also weakening, contributing to a general trend across the basin towards more homogeneous dynamics. Overall, the basin’s hydrodynamic regime is becoming less asymmetric, with a slight tendency towards relatively stronger flood currents, however, generally preserving the flood/ebb dominance distribution.

4.4 TWA velocities

In the REF scenario (Fig. 7a), larger TWA velocities (Eq. 1) are typically observed in areas where there is a strong discharge asymmetry combined with sufficient depth to sustain the strong currents. For example, the highest TWA velocities in the domain (0.45 m s^{− 1}) occur near the tidal inlet and the edge of the northernmost tip of the island Sylt. The TWA velocities are also notable where a bathymetry gradient is present, particularly around the edges of the tidal channels. In contrast, the central parts of the channels show very low TWA velocities due to low asymmetry there. In intertidal and very shallow subtidal areas, the TWA velocities are near zero for a different reason: despite large discharge asymmetry, the currents are weak, which prevents large TWA.

In the HSLR scenario (Fig. 7b), the distribution of TWA velocities shows a significant shift. There is a general decrease in TWA velocities across many shallow subtidal areas and along the channels. However, the central part of the basin including the tidal inlet and surrounding areas exhibits an increase in TWA velocities. In addition to that, there is also a slight increase in TWA velocities in intertidal areas, particularly near the southern causeway, which is likely associated with the increase in inundation (see Sect. 4.1).

The decrease in TWA in previously shallow areas is attributed to the reduced influence of non-linear effects and the weakening of velocity asymmetry as sea level rises, increasing the depth of the water column (Figs. 6b and 7b). Conversely, newly formed future intertidal zones are characterized by increased TWA both due to appearance of the current and strong asymmetry (Figs. 6b and 7b). A tidal inlet, which is characterized by large velocities and already exhibits a slight pre-existing asymmetry, supports a non-zero TWA (Figs. 6a and 7a). As sea level rises, the asymmetry, though decreasing, does not entirely disappear (Figs. 6b and 7b). The persistence of even minor asymmetry, combined with increasing current magnitude, is sufficient to sustain and enhance TWA velocities in the tidal inlet. The overall increase in TWA in the bottleneck zone signals an increase in the tidal prism - more water enters and exits the area within each tidal cycle. Also, the fact that the whole bottleneck zone is characterized by increasing TWA means that this zone is represented by a combination of ebb- and flood-dominated subareas, which we can also see in Fig. 6a. Combining the maps of velocity asymmetry and TWA velocity magnitude, some information about the direction of the net discharge can be given. For example, if an area is characterized by a stronger velocity magnitude during the flood, it suggests that the TWA at this location will have a positive onshore component (the vector field of the TWA velocity is shown in the Discussion section).

4.5 Tracer dispersion

In this section, transport patterns for both the current and future scenarios are presented, with a focus on examining whether the observed decrease in asymmetry leads to a weakening of tidal residual currents and, consequently, a general reduction in the net tidally induced transport from the bight, despite the increase in tidal prism. In a system characterized by complex topography and spatially variable asymmetry, along with its changes under SLR, the statement about transport dynamics of passive, weightless tracers must be validated using Lagrangian simulations. There is no inherent contradiction between an increasing tidal prism and a potential reduction in net tidal transport. While tracer trajectories may become longer due to stronger currents, the net transport or displacement of tracers could simultaneously decrease. Additionally, larger volumes of water may flow in and out of the domain, increasing the tidal prism, but circulate in a back-and-forth manner without contributing to net transport.

The results are presented as a spread of tracers and dispersion probability (ranging from 0 to 1) of the clouds released in nine different locations in the domain, which are selected to capture areas with different dynamics.

The domain appears to have separated into two systems, one to the north and one south of the Jordsand Bank (Fig. 8). Such a separation can be attributed to duration asymmetry for the reference scenario (Fig. 5), where the southwestern part of the domain represents the area with a longer ebb, while the northeastern part represents the area with a longer flood. Particles released in the northern part hardly travel to the south, and vice versa. It is true even when they spread to and have a high dispersion probability in the tidal inlet regardless of where they were initially released. Second, due to the nature of the shallow subtidal and intertidal environment, in most cases, the particles either end up at the shoreline or circulate primarily through the tidal channels and the tidal inlet. This is expected as the currents have higher magnitudes and therefore a larger transport capacity in deeper channels. However, despite six of the nine clouds indicating a high probability of entering the tidal inlet (1, 3, 6, 7, 8, 9), most of them appear to be trapped circulating within the inlet, following repetitive loop-like trajectories. Among the tracer clouds, the one released at Königshafen (9) demonstrates the relatively high probability to disperse further outside of the basin. Clouds 2, 4 and 5 can generally spread toward the inlet, but they have a higher probability of remaining on the intertidal flats (4, 5) or close to the shore (2).

Additionally, Fig. 8 highlights regions with a higher probability (red colours) of tracer presence, as indicated also by areas of increased tracer concentration. These regions correspond to the primary pathways for passive particle movement throughout the basin, representing dominant transport routes. It is important to note that these routes are relatively narrow and do not directly align with the deepest parts of considered subareas. We observe that areas with relatively large TWA largely correspond to zones with significant particle dispersion, while neighbouring areas with small TWA are proxy for major transport channels, if they are deep enough to sustain relatively strong currents (Figs. 7a and 8). This is not surprising, as a small TWA in areas with large velocities indicates proportionally large water mass transport during both the flood and ebb phases.

At first glance, the dispersion pattern under the HSLR scenario (Fig. 9) appears to be very similar to that of the REF scenario, however, there are important changes to mention. Most notably, the outward transport of particles from the basin is decreasing, as evidenced by the dispersion of clouds 4, 6, 7, and 8. Cloud 4 stands out in particular, as its spread diminishes to the point where it does not even reach the tidal inlet. A similar, though less pronounced, pattern is observed with clouds 2 and 5, which still spread toward the inlet but show minimal dispersion probability. Overall, all the clouds—especially clouds 4, 6, 7, 8, and 9—exhibit a shift in transport directed more toward the intertidal flats. This visual shift supports the implications of changes in velocity patterns and reduced tidal asymmetry with a slight shift toward flood dominance that the system undergoes in response to rising sea levels.

In addition to the particle dispersion maps, we quantified changes in tracer outflow across scenarios through experiments where tracers were distributed throughout the entire domain, excluding areas not consistently submerged during every flood phase. The analysis examined two aspects: the total number of tracers exiting the domain, which represents the cumulative outflow of particles across all release cycles, and the number of unique release locations, which reflects the spatial distribution of source regions contributing tracers to the outflow.

Compared to the REF scenario, the total number of tracers exiting the domain over all release cycles decreases by 9.8%, representing a decline in overall transport efficiency. Similarly, this reduction is accompanied by a 6.2% decrease in the number of unique tracer release sites from which tracers exit the domain (Fig. 10b). This indicates that fewer regions within the bight contribute to the outflow, reflecting a reduction in hydrodynamic connectivity across the domain. By the end of the century, this reduction in total outflow could reach 12.5%. This distinction highlights two key findings: not only is there an overall decrease in the number of tracers leaving the domain, but fewer areas contribute to this outflow (Fig. 10).

To better understand the spatial redistribution of outflow, we divided the domain’s boundary into three equal zones and analyzed the differences between the REF (Fig. 10a) and HSLR (Fig. 10b) scenarios. Tracers released near the tidal inlet predominantly exit through the central and northern zones in both scenarios, with similar distributions. However, notable differences emerge in the inner domain. In the REF scenario, tracers are more densely flushed southward, as indicated by the blue and green markers. Under the HSLR scenario, while the general distribution of outflow zones remains largely unchanged, the overall reduction in outflow reflects a weakening of tidal net transport.

Wind forcing introduces only slight variations in net transport. Under HSLR with wind, total outflow decreases by 8.6%, and the number of tracers from particular release sites drops by 6% compared to REF with wind. Comparing wind-forced scenarios to tide-only experiments shows an 18–18.5% increase in total tracer outflow, while the number of contributing release sites remains unchanged.

5.1 Role of wind forcing

Despite the fact that tidal forcing explains more than 80% of the velocity variability in the bight (e.g., Fofonova et al. 2019), wind forcing can significantly modify tidally induced dynamics. The crucial question we have in mind is whether the major transport channels identified in Figs. 8 and 9 remain relevant when wind forcing is present. Additionally, we need to determine if the result showing a decrease in outflow transport in the bight under SLR pressure still holds when wind forcing is included. To address these questions, we conducted a series of experiments with summer wind forcing. While the composition of wind events in the summer was close to the median, however, the selected year (2009) had more events than a multi-year median, indicating summer wind conditions were stronger than the typical. The period was selected, as mentioned in the Methods, due to its ecological relevance and the absence of future trends (Ortega et al., 2025). The results of these wind-forcing experiments revealed that wind forcing does not significantly alter the transport patterns in the bight. Although it introduces slight variations in particle spread, the primary transport pathways (Fig. 11) and decreasing trend of outward net transport remain consistent with those observed in the tide-only simulations for both the REF and HSLR scenarios.

5.2 Implications

5.2.1 Ecological impact

Our results demonstrate that, despite the projected bathymetry accretion in some areas, tidal inundation is overall increasing in response to SLR. While the basin remains relatively resilient to projected changes by 2050, nearly a third of the intertidal areas—and the corresponding habitats— could be lost under the high-emission SLR scenario (SSP5-8.5) by the end of the century. Historically, the Wadden Sea has shown resilience to sea level changes due to its large sediment budget and the ability of unconfined tidal systems to migrate landward. However, under future SLR scenarios, the intertidal coasts are increasingly susceptible to drowning. This has direct implications for species communities inhabiting transitional zones, which depend on the full tidal range throughout their development. Furthermore, as SLR accelerates, the increased inundation of intertidal flats could have cascading effects on higher trophic levels, such as fish and migratory shorebirds that rely on these zones for feeding and breeding. If the current trajectory of SLR continues, the Wadden Sea’s ability to support its unique biodiversity and ecosystem services could be severely diminished by the end of the century.

In addition to increased inundation, we observe spatially uneven changes and trends in current velocities highlighting the dynamic response of the basin to SLR. The area with a significant increase in maximum velocities, observed along the Lister Ley channel, is currently inhabited by mussel culture plots, whose survival may be significantly challenged by stronger current dynamics.

An interesting insight arises when tidal currents are analyzed in terms of ebb and flood phases. Our results indicate that even within a relatively small area like the Sylt-Rømø Bight, tidal dynamics are patchy, with varying localized patterns of ebb or flood dominance. Overall, we observe a shift from a tidal-dominated system toward a more lagoon-like environment, particularly under the high SLR scenario, with a slight shift toward flood dominance. This observation aligns with long-standing projections that sea-level rise tends to enhance flood dominance due to changes in tidal wave propagation and basin morphology (Dronkers 1986; Pethick 1994; Friedrichs and Aubrey 1988). However, the semi-enclosed nature of the Sylt-Rømø Bight introduces a distinct nuance. While the system does exhibit a slight shift toward flood dominance, the overall strength of tidal asymmetry diminishes. The weakening of tidal asymmetry (both in duration and velocities), which points to a reduction in tidally induced net transport of passive tracers, is a critical finding. This could alter the transport of larvae from the oyster beds in the area, potentially leading to a self-sustaining population.

Further insights into the evolution of transport processes within the basin under future SLR scenarios are provided by Lagrangian particle dispersion experiments. Even in the reference scenario, there is already a separation between the northern and southern parts of the basin: particle tracers released north of the Jordsand Bank rarely migrate to the southern half of the basin, and vice versa. With a reduction in tidal-driven net transport, the basin may become more compartmentalized, reducing the overall hydrodynamic connectivity between different regions of the Sylt-Rømø Bight. This has significant ecological implications, as reduced connectivity could hinder the spread of larvae and other organisms between different parts of the basin, potentially leading to genetic isolation of populations. For example, it would be valuable to investigate the seagrass meadows and other benthic species in the southern part of the basin compared to those in the Danish part. Species like seagrass and benthic fauna that rely on larval dispersal may find it more difficult to recolonize areas that have been disturbed or eroded. This reduced connectivity could also impact the resilience of the ecosystem, making it more vulnerable to localized disturbances such as extreme weather events or pollution. Further research is needed to explore the long-term implications of this reduced connectivity. However, we should note that for this task, we also need to account for wind forcing and its potential changes, as well as modifications not only in the net transport of passive tracers but also in their trajectories inside the bight. In some parts of the bight, these trajectories may become longer, even if the net transport is reduced due to larger velocities. This could partly compensate for the reduction in net transport. Therefore, the connectivity between different parts of the bight requires a separate study.

5.2.2 TWA, tidal asymmetry and morphology

The analysis of maximum and TWA velocities can provide significant insights into local bedforms and potential future morphological changes in the area. The Fig. 12b shows bedform features obtained from multibeam echosounder surveys. The patterns of the TWA velocities in this region (Fig. 12c) accurately reproduce the dynamics favourable for the generation of such a long-term geomorphological feature, e.g. "Fishbones" (Fig. 12b). The direction of the TWA velocities also aligns with the orientation of the underwater dunes in the inlet. The tidal velocity asymmetry (Fig. 12d) and the TWA velocities (Fig. 12c) of the same region, towards the Jordsand Bank, demonstrate the same process – the inflowing flood currents are stronger than outgoing ebb currents. This region, which lies between the tidal inlet and the Jordsand Bank, is characterized by the presence of a flood ramp which might be a result of long-term sediment deposition driven by the aforementioned asymmetry. (Even the Jordsand Bank itself might have been formed by this mechanism). When discussing asymmetry, it is important to consider the co-evolution of bedforms with hydrodynamics, accurate future projection of bedform characteristics in the area will require a coupled sediment module, as well as wind and wave forcing. However, certain features, like ‘Fishbones’ (Fig. 12b), exist due to the particular topography of the area and the way long surface gravity tidal waves enter and propagate within the system. Also, despite the frequent storm events in the area, large transverse dunes have been a stable characteristic of the area for the last 50 years. Projections for the Sylt-Rømø Bight suggest limited changes to bedform growth or depletion by 2050 (Becherer et al. 2018). While our study does not dynamically simulate sediment transport, it is consistent with the projected relative stability of bedforms, as we do not observe significant changes in the distribution of flood-ebb dominance zones, with a few exceptions, but rather a weakening of asymmetry. However, projected higher TWA velocities indicate areas with potentially increased local erosion and bedform peculiarities. The overall slight increase in flood dominance within the system and the reduction in net transport from the bight may potentially indicate a shift toward the accumulation of material in the bight. However, this hypothesis requires further investigation, including an analysis of changes in the inflow transport to the bight and introduction of a proper sediment module with particles of different properties and masses.

5.3 Study limitations

This study is based on some simplifications, which are important to acknowledge and discuss. Firstly, we use the same tidal forcing for all scenarios at the open boundary, which makes the comparison of results between scenarios more straightforward. Predicting future tidal forcing, especially in coastal areas, is challenging due to the multitude of influencing factors operating at different scales and the significant uncertainties associated with these factors. For instance, a single storm event can significantly alter local morphology (e.g., Arns et al. 2015; Bartholomä et al. 2009; Schuerch et al. 2018), and the erosion rates of the islands are largely unknown. Consequently, the prognosis for long-term morphological changes, whose effect subsequently extends to the local tidal characteristics, remains highly uncertain. Due to these uncertainties, only a few future projections of tidal forcing are available for the area, typically coarse in resolution, and rely on many assumptions. It is dubious whether these errors will be insignificant compared to the assumptions that justify using a single database for all experiments.

Secondly, we recognize the potential influence of episodic storm events, which can reshape the basin’s morphology and, in turn, alter the hydrodynamics. However, capturing their effects would require separate experimental setups, and thus, they were not included in this study.

Thirdly, the role of baroclinic effects, such as density-driven circulation and presence of stratification, was not included in this study. Baroclinic processes are known to influence transport patterns in the Sylt-Rømø Bight (e.g., Burchard et al. 2013, 2008; Stanev et al. 2015). However, accounting for these processes is challenging due to the highly varying in time stratification patterns as well as uncertainties in future river discharge projections.

While our study does not claim to comprehensively represent all future transport processes, it provides a robust baseline scenario by limiting uncertainties to a manageable level. Including all uncertain forcings simultaneously would make it harder to isolate and understand the underlying mechanisms driving hydrodynamic changes. Future research could build on this work by incorporating baroclinic effects, evolving storm events, and dynamic morphological feedback, offering a more holistic perspective on the complex interactions shaping hydrodynamics and transport processes under SLR scenarios.

This study investigates how climate-change-driven SLR may affect tidally induced transport dynamics in a semi-enclosed tidal basin. Using the setup with a resolution of up to 2 m in the intertidal areas, we simulated the effects of low and high SLR scenarios by 2050, offering critical insights into the localized impacts on the tidal environment. Specifically, we begin by examining changes in tidal inundation, which establish the baseline for subsequent modifications in the dynamics. This is followed by a detailed analysis of tidal currents from both Eulerian and Lagrangian perspectives, combining these approaches to provide a more comprehensive understanding of the transport pathways of passive tracers.

Our results add to the current knowledge and reveal a few novel insights worth attention. Firstly, under both low (SSP1-2.6) and high (SSP5-8.5) SLR scenarios, we project a substantial reduction in intertidal flats, which are critical habitats for benthic species and migratory birds. By 2050, up to 7.4% of current intertidal zones may be lost under the HSLR scenario, with even more severe reductions (29.5%) expected by 2100. Secondly, in response to higher sea levels, tidal currents undergo a spatial redistribution: maximum velocities increase in intertidal and shallow subtidal zones due to deepening water columns and reduced bottom friction, while deeper tidal channels experience slight reductions in the current strength. Such spatially varying changes in response to SLR scenarios are also observed in TWA velocities and tidal asymmetry. Thirdly, our Lagrangian experiments reveal the major transport pathways in the bight (indicating a high probability of passive tracers occurring) and demonstrate that they remain stable across all considered scenarios, including those incorporating wind forcing. While wind forcing introduces slight variations in particle spread, it does not significantly impact the primary transport pathways or the observed decreasing trend in outward net transport under SLR scenarios These pathways do not necessarily align with the deepest parts of the domain or areas with the highest velocities, but they are approximated by the intersection of zones with small TWA and relatively large discharges. Next, based on this Lagrangian transport analysis, we also found that the northeastern and southwestern parts of the bight remain largely separated in terms of water mass and passive tracer exchange under both present and future SLR scenarios, reflecting the spatial segregation within the bight. And finally, a weakening of tidal asymmetry under SLR in the bight suggests a general reduction in tidally induced net transport of passive tracers. Such a reduction in net transport could potentially lead to more segregated habitat distributions, making them more vulnerable. We observed a transition toward a more lagoon-like system, where outward transport from the basin to the open sea is reduced. Lagrangian particle dispersion experiments provide evidence of this decrease in transport: approximately 9.8% by 2050 and 12.5% by 2100.

The advances in knowledge from this study provide critical insights into the future dynamics of tidal basins under SLR and lays the groundwork for future research. It underscores the importance of such high-resolution, localized modeling. Understanding these patterns is vital for developing effective coastal management and conservation strategies.