On the assessment of channel deepening impacts in micro-meso tidal estuaries: A systematic analysis

Contents lists available at ScienceDirect Ocean Modelling journal homepage: www.elsevier.com/locate/ocemod On the assessment of channel deepening impacts in micro-meso tidal estuaries: A systematic analysis Guillermo Martín-Llanes ∗, Alejandro López-Ruiz Departamento de Ingeniería Aeroespacial y Mecánica de Fluidos, Universidad de Sevilla, Camino de los Descubrimientos s/n, 41092, Seville, Spain ARTICLE INFO Dataset link:Delft3D - FLOW and bathymetry f iles (Original data) Keywords: Dredging Tidal propagation Salt intrusion Basin management Estuary ABSTRACT The need for efficient maritime transportation in estuaries has led to the development of diverse dredging strategies to accommodate vessels with deep drafts. Most recent studies assessing the environmental impacts of channel deepening use advanced, tailored models to simulate the long-term response to historical bathymetric changes in estuaries worldwide. However, these models are often time-consuming and highly specific to local conditions, limiting the broader applicability of their results. In addition, a common limitation is the significant time gap between the bathymetric data used, often exceeding 100 years. This makes it challenging to quantify the effects of isolated deepening operations, which is essential for understanding the influence of human intervention on estuarine dynamics. To overcome this limitation while ensuring efficient and adaptable modelling, this paper presents a three-dimensional idealised model (Delft3D) to quantify the short-term, e.g., weeks, hydrodynamic and salinity response to dredging operations in micro-meso tidal, well-mixed estuaries. Implications on channel operativity are also discussed. The numerical experiments examine variations in both channel depth and dredging length. Key findings suggest that dredging length is critical in the estuarine response. Specifically, dredging length has a greater influence on tidal amplification than channel depth. Changes in the flow structure are primarily driven by changes in the barotropic pressure gradient and bed shear forces, which vary spatially along the estuary, defining three distinct regions of behaviour. In addition, salt intrusion increases linearly with channel depth and becomes particularly sensitive to dredging length in shorter operations. Regarding basin management, results reveal that landward operativity is compromised by dredging in the lower river. 1. Introduction Estuaries constitute fertile and densely populated areas (Syvitski and Saito,2007) that provide multiple social, economic and environmental benefits (Orton et al.,2015). In recent decades, estuaries have been increasingly affected by multiple pressures from climate change and human activities, such as river regulation and changes in river morphology (Mitchell et al.,2015;Siemes et al.,2024). In particular, the expansion of maritime transportation has accelerated the development of channel deepening projects to improve navigational efficiency. These modifications have negatively altered salt and sediment transport (Ralston and Geyer,2019;Reid et al.,2022), impacting water quality and ecosystem composition. In response to these challenges, extensive research has focused on three key areas: (1) analysing the current consequences of historical bathymetric changes (Ralston et al., 2019;Van Maren et al.,2015), (2) assessing the vulnerability and environmental risks associated with future dredging operations (Bai et al.,2003;Alba et al.,2014;Gómez et al.,2014;Paarlberg et al., ∗Corresponding author. E-mail addresses: [email protected] (G. Martín-Llanes), [email protected] (A. López-Ruiz). 2015;Álvarez et al.,2017;Zarzuelo et al.,2019;He et al.,2024) and (3) developing mitigation and adaptation strategies (Orton et al.,2015; Li et al.,2016;Hoagland et al.,2020;Hendrickx et al.,2024). Regarding the first question, changes in estuarine dynamics due to channel deepening arise from the response of the tide and the river flow to changes in the channel depth. Channel deepening alters the structure of tides, typically resulting in tidal amplification (DiLorenzo et al., 1993;Ralston et al.,2019), increased current amplitudes (Sirviente et al.,2023), and higher wave celerity (Zhang et al.,2021), while also influencing tidal asymmetry (Winterwerp and Wang,2013;Winterwerp et al.,2013). Additionally, dredging can modify the estuarine response to river floods and storm surges, reducing peak water levels during river floods but facilitating the further propagation of surge waves (Ralston and Geyer,2019;Bao et al.,2022). Further research in Tampa Bay (Zhu et al.,2015;Meyers et al.,2017), the Seine Estuary (Grasso and Le Hir,2019), and the Santos Estuary System (Reid et al.,2022) suggest that changes in the river-tide interactions lead to an increase in https://doi.org/10.1016/j.ocemod.2025.102552 Received 31 December 2024; Received in revised form 14 March 2025; Accepted 12 April 2025 Ocean Modelling 196 (2025) 102552 Available online 25 April 2025 1463-5003/© 2025 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/bync-nd/4.0/ ). G. Martín-Llanes and A. López-Ruiz estuarine circulation. Additional insights are provided by Chant et al. (2018), whose research on Newark Bay highlights that the impact of channel deepening on exchange flow is determined by the sensitivity of horizontal salinity gradients to channel depth 𝐻, which, in turn, depends on the dredging length. In particular, when the dredging length is short relative to the tidal wave length, the horizontal salinity gradient is insensitive to the change in channel depth, and the exchange flow is proportional to 𝐻3, consistent with the theory of Hansen and Rattray Jr. (1966). An increase in exchange flow enhances the landward transport of salinity and sediment, leading to greater salt intrusion (Chen et al.,2019;Ralston and Geyer,2019;Reid et al.,2022; Zhao et al.,2022), elevated suspended sediment concentrations, and increased turbidity (Talke et al.,2009;Ralston et al.,2012;De Jonge et al.,2014;Van Maren et al.,2015;Eidam et al.,2021,2022). Hence, precisely quantifying the effects of channel deepening on estuarine hydrodynamics is essential for assessing its impact on salinity and sediment concentrations, both of which play a critical role in determining ecological health and water quality. However, much of the existing literature relies on analytical or semi-analytical models (Talke et al.,2009;Cai et al.,2012a,b;Di Risio et al.,2017) and observational data studies (Chant et al.,2018;Bao et al.,2022), which often lack precision (Zhang et al.,2011) and are unable to accurately represent complex estuarine systems (Winterwerp and Wang,2013). In contrast, numerical modelling emerges as a powerful alternative for assessing the effects of channel deepening (Bai et al.,2003;Alba et al., 2014;Gómez et al.,2014;Paarlberg et al.,2015;Van Maren et al., 2015;Chen et al.,2019;Zhao et al.,2022;Yi et al.,2024). Among their advantages, numerical models enable the evaluation of spatially variable risks associated with human interventions (Familkhalili and Talke,2016;Meyers et al.,2017;Familkhalili et al.,2020) and to overcome linearisation assumptions that are not acceptable for the analysis of strong non-linear processes, such as tidal asymmetry (Guo et al.,2014). However, these studies often face three key limitations: (1) the significant computational time required for high-detail models, (2) their strong dependence on local conditions, which restricts the generalisation of results to other estuarine systems, and (3) the extensive time gap in the bathymetric data used to simulate the estuarine response to channel deepening (Meyers et al.,2017;Ralston et al., 2019;Familkhalili et al.,2020;Eidam et al.,2021,2022). This time gap, often exceeding 100–150 years, makes it challenging to isolate the specific effects of channel deepening from other environmental and anthropogenic changes that may have occurred over the same period (Reid et al.,2022). For instance, Eidam et al. (2021) suggested that several morphological modifications, such as land reclamation and the disposal of dredged material within the estuary, took place during the analysed time frame, further complicating the attribution of observed changes to dredging alone. Additionally, while many studies have focused on long-term trends, the short-term response to individual dredging operations have received relatively little attention (Vellinga et al.,2014). In this context, this paper develops a three-dimensional idealised model (Delft3D) to isolate the short-term hydrodynamic and salinity response to dredging operations. Numerical experiments assess the influence of changes in channel depth and dredging length, a factor that has received limited attention (Chant et al.,2018). Specifically, 15 dredging scenarios are simulated in a micro-meso tidal, well-mixed estuary (Martín-Llanes and López-Ruiz,2024) to quantify changes in tidal structure and residual flow, which are further analysed through momentum balance variations. The effects of these hydrodynamic alterations on salt intrusion and channel operativity (i.e., the time fraction during which water level is higher than a certain vessel draft) are also discussed. By using an idealised modelling approach, the proposed methodology eliminates bathymetric irregularities and local effects, thereby improving the applicability of findings while reducing computational time. This makes it particularly relevant for estuarine port systems subjected to microand meso-tidal estuaries, typically found in European regions (Garel and D’Alimonte,2017;Díez-Minguito et al.,2012), providing valuable insights for port management and the potential environmental risks associated with dredging activities. The content of the paper is organised as follows. Section 2describes the physical domain and the numerical framework. The numerical experiments representing the baseline and dredging scenarios are defined in Section 3. Section 4analyses the impact of dredging in estuarine hydrodynamics and salt intrusion. The implications of the obtained results on the management of navigational channels are discussed in Section 5. Finally, the main conclusions are presented in Section 6. 2. Materials and methods 2.1. Physical domain The physical domain in which the numerical model is implemented follows the idealised geometry formulation (Savenije,2012), which is represented by a river width (𝐵) that varies exponentially along the longitudinal, linear axis as: 𝐵(𝑥) =𝐵0exp (−𝑥 𝑏)(1) where 𝐵0is the mouth width, 𝑥represents the axial (along-channel) distance (positive landward) from the mouth to the upstream section of the estuary at 𝑥=𝐿, and 𝑏is the convergence length. According to Eq. (1), the geometry of the estuary is obtained using 𝐵0= 1000 m,𝑏= 157 k mand 𝐿= 80 k m(Fig. 1a). This geometry represents an estuary with moderate convergence according to MartínLlanes and López-Ruiz (2024), where more details on the geometric characteristics can be found. The estuary thalweg 𝐻𝑚𝑎𝑥(𝑥)and the coastal shelf are defined with constant slopes of 𝑆= 2.5 × 10−5 and 𝑆𝑛= 2.5 × 10−3, respectively. Hence, considering the average water depth at the mouth 𝐻0= 10 m, average depths of 𝐻𝑟= 8 mand 𝐻𝑛= 25 mare obtained in the upstream and offshore boundaries, respectively. The cross-channel geometry is defined following a Gaussian shape: 𝐻(𝑥, 𝑦) =𝐻𝑚𝑎𝑥(𝑥) exp (−𝑦2 2𝑐2)(2) where 𝑦is the cross-channel coordinate, which varies from −𝐵(𝑥)∕2 to 𝐵(𝑥)∕2. The shape of the Gaussian is defined using a standard deviation of 𝑐= 152 m, which results in a cross-channel average slope of 0.04. 2.2. Model description and setup 2.2.1. Model description Hydrodynamics and salt transport in the estuary are obtained using the Delft3D model (Lesser et al.,2004). This modelling framework has been widely applied in recent studies on morphodynamics, hydrodynamics, and salt transport (Martyr-Koller et al.,2017;Ruiz-Reina and López-Ruiz,2021;Zarzuelo et al.,2021), particularly in response to climate change and human interventions (Mulligan et al.,2019;Yin et al.,2019;Wu et al.,2021;Arevalo et al.,2022), including channel deepening (Alba et al.,2014;Guo et al.,2014;Paarlberg et al.,2015; Van Maren et al.,2015;Zarzuelo et al.,2015,2019;Reid et al.,2022; Zhao et al.,2022). The hydrodynamic FLOW-module of the Delft3D model is used to solve the unsteady shallow water equations in 3D and the advection–diffusion equation for salt transport coupled to a turbulence closure model. Further details of these equations are found in Martín-Llanes and López-Ruiz (2024). A sigma-layer scheme with a constant number of layers is used for the vertical discretisation. Ocean Modelling 196 (2025) 102552 2 G. Martín-Llanes and A. López-Ruiz Fig. 1. Model setup and numerical experiments. (a1) Estuary bathymetry. (a2) Detail of the red area highlighted in (a1). (b) Cross-channel geometry for the different depth increases. (c) Thalweg profiles after dredging operations extending 15 km (𝛥𝐻 = 15,30,40%). (d) Thalweg profiles after dredging operations with 𝛥𝐻 = 40% (d =5, 10, 15, 20, 30 km). Black dashed line in (c, d) indicates mean water level. 2.2.2. Model setup The model setup consists of a regular grid covering both the estuary and the shelf. Grid resolution varies from 854 ×328 m2near the offshore and cross-shore boundaries to 109 ×24 m2near the mouth and at the upstream sections of the river. To ensure adequate vertical resolution, the initial 𝜎-layer scheme (Martín-Llanes and López-Ruiz, 2024), consisting of 10 vertical layers with thickness varying according to local depth, was assessed against a 20-layer scheme. In particular, harmonic analysis of water levels and velocities was conducted along the channel. Additionally, axial salinity profiles at neap and spring tides and the temporal evolution of salt intrusion were analysed for both vertical resolutions. The 10-layer scheme demonstrated sufficient accuracy in all sensitivity tests and was thus adopted for this study. Further details on the layer sensitivity analysis are provided in Section 1 of the Supplementary Material. The time step, which ensures stability and accuracy, given the spatial grid resolution and depths, is 6 s. Boundary conditions are imposed in four open boundaries. The offshore boundary prescribes an astronomical water level forcing with two semi-diurnal components: (1) 𝑀2(𝐴= 1.00 m;𝜙= 180◦) and (2) 𝑆2(𝐴= 0.25 m;𝜙= 90◦). Hence, the estuary is considered micromeso tidal. At the cross-shore boundaries, a Neumann-type condition is imposed with a zero longshore water level gradient (Roelvink et al., 2004). Finally, a total discharge of 𝑄𝑟= 150 m3s−1 with uniform vertical distribution of velocities is considered in the upstream section of the estuary. This value is established according to the mean tidal prism (6.47 × 107m3) and the period of the main tidal harmonic to reproduce well-mixed conditions, i.e., with a Canter-Cremers Number below 0.1, according to Dyer (1973). The definition of the physical parameters used in the model can be found in Section 2 of Supplementary Material. The model is set at the equator and then the Coriolis effects are neglected. 3. Numerical experiments The numerical experiments comprise 15 simulations where the estuary described in Section 2and Fig. 1a (Scenario 0) is modified by varying the channel depth (see Table 1) while preserving the crosssectional shape. Specifically, three different depth increases of 𝛥𝐻 = 15,30,40% (Fig. 1b,c) are considered, relative to the original mouth depth. These values conform a representative sample of common depth increases observed in different estuaries affected by dredging operations, which usually range between 10 − 40% (Chant et al.,2018;Ralston et al.,2019;Ralston and Geyer,2019;Amorim et al.,2023). The increased depth extends over five different lengths: 𝑑= 5,10,15,20,30 k m (Fig. 1d), which correspond to a fraction of the maximum salt intrusion in Scenario 0 (0.35, 0.7, 1, 1.35 and 2, respectively). From this point, a constant transition slope of 1∕200 is applied to connect the dredged plane with the original channel bathymetry. Since this slope remains fixed across all 15 operations, the extent of the transition zone varies depending on each dredging scenario. Simulations are set considering the boundary conditions described in Section 2.2.2. Initial conditions for each scenario were obtained from prior spin-up simulations with identical external forcing, ensuring steady-state hydrodynamic and transport conditions. The time frame for the numerical experiments is 15 days, so that differences in the flow structure due to neap-spring cycles are captured. Ocean Modelling 196 (2025) 102552 3 G. Martín-Llanes and A. López-Ruiz Table 1 Summary of numerical experiments. Depth increase 𝛥𝐻 [%] Dredging length 𝑑[km] Scenario 0 0 0 Scenario 1 15 5 Scenario 2 15 10 Scenario 3 15 15 Scenario 4 15 20 Scenario 5 15 30 Scenario 6 30 5 Scenario 7 30 10 Scenario 8 30 15 Scenario 9 30 20 Scenario 10 30 30 Scenario 11 40 5 Scenario 12 40 10 Scenario 13 40 15 Scenario 14 40 20 Scenario 15 40 30 4. Results: impact of channel deepening on hydrodynamics and salinity distribution 4.1. Hydrodynamics and salinity distribution in the unaltered estuary This section examines hydrodynamics and salinity distribution at the unaltered scenario (Scenario 0). The tide is first characterised through the harmonic analysis of water level and velocity at 10 points along the channel, using the t_t ide MATLAB package (Pawlowicz et al., 2002). Results are shown in Fig. 2. Regarding water level amplitudes (Fig. 2a), the damping of the semi-diurnal component 𝑀2is significant up to half of the channel (𝑥= 40 k m); after this section, the higher influence of the river discharge compensates the damping of the tidal wave and leads to a synchronous behaviour with water level amplitudes approaching 0.4 m. Reduced external overtide 𝑀4amplitudes are obtained along the channel, with a maximum of 0.1 m in the upstream section. Tidal currents at the mouth of the estuary are shown in Fig. 2b. Maximum values during the flood (positive landward) reach 1.1 and 0.6 ms−1 during spring and neap tide cycles, respectively. On the other hand, maximum currents during the ebb (negative seaward) reach 1.6 and 1.2 ms−1 in spring and neap tide cycles, respectively. These values decrease monotonically along the channel (Fig. 2c); in the upstream section, the amplitude of the semidiurnal component 𝑀2is less than one fifth of the value at the mouth. Tidal interactions with bathymetry and river discharge lead to tidal asymmetry (LeBlond,1991;Parker,1991), which is reflected in an imbalance between flood and ebb durations and maximum flow velocities (Guo et al.,2014). The strength of the asymmetry is measured by the amplitude ratio between the overtide 𝑀4and the semidiurnal component 𝑀2while the direction of the resulting net transport is a function of the phase lag between these two components. The results in Fig. 2d,f show an increasing ebb tidal asymmetry along the channel (2𝜙𝑀2−𝜙𝑀4≈ 350◦), enhanced by the landward depth reduction, which is particularly notorious during neap tides (Fig. 2b), when maximum ebb velocities are twice those achieved at the flood. This ebb dominant behaviour is common in strongly dissipative tide dominated estuaries affected by river discharge (Lanzoni and Seminara,1998). Finally, the phase lag between water level and velocities (Fig. 2e) increases along the channel from 0.93 h at the mouth to 2.79 h at the upstream section. Consequences for the transport of conservative substances (e.g. salt transport) derived from this result will be significant when the role of the river-tide relation is relevant, i.e. when river discharge changes with a time scale close to the tidal period are analysed. Horizontal salinity distribution in the unaltered estuary is quantified in terms of salt intrusion, i.e., the landward spatial boundary where the salinity of the bottom layer is reduced to 1 psu. This magnitude changes periodically in relation to the tide, reaching a maximum value of 15 km, approximately, at spring tide. For the tidal and river discharge conditions defined in Section 2.2.2, the water column is well-mixed. The results indicate that the estuary is tide dominated, exhibiting pronounced ebb tidal asymmetry. In consequence, the hydrodynamic response and the transport changes due to dredging will be mainly determined by changes in the tidal properties. The tide can be modified by changes in river discharge, convergence of river margins, channel depth and frictional effects. Dredging will directly modify the channel depth and frictional effects, but will also induce changes in the rivertide relationship and the role of convergence, all of which determine estuarine dynamics in the new configuration. 4.2. Tidal propagation and asymmetry The impact of channel deepening on hydrodynamics is first assessed by analysing the ratio of the 𝑀2amplitude (𝜂) and phase (𝜖) between the dredged scenarios and the unaltered estuary along the channel (Fig. 3). The reduction of friction leads to tidal amplification (Fig. 3a,c,e) and a general decrease in the tidal phase (Fig. 3b,d,f). Changes are proportional to the dredging length and depth increase, with the latter having a greater influence on tidal amplification. As shown in Fig. 3e, the maximum amplitude is reached for a 40% depth increase and a dredging length of 30 km, resulting in a tidal amplification of 60%. In terms of tidal phase, the reduction in bottom friction due to dredging increases the tidal wave celerity. The minimum phase ratio is thus obtained for scenario 15 and results in 93.5% (i.e. a reduction in travel time of 0.178 h after dredging), as shown in Fig. 3f. In contrast, when analysing tidal propagation across different dredging lengths for a given depth increase (e.g. for 𝛥𝐻 = 30% in Fig. 3c,d), two different results are observed. First, tidal amplification and phase reduction increase along the dredged section of the channel, reaching their peak at 𝑥=𝑑. However, the dredging length that produces the highest tidal amplification at a given location varies depending on that location position. For example, at 𝑥= 17 k mthe greatest amplification occurs for 𝑑= 20 k m, whereas at 𝑥= 7 k m, the maximum is observed for 𝑑= 10 k m. On the other hand, in the unaltered section of the channel, i.e. for 𝑥 > 𝑑, tidal amplification and phase reduction decrease, approaching an asymptotic value that differs from the corresponding value in the unaltered estuary. Asymptotic tidal amplitude increases with dredging length and depth increase, while the opposite is observed for the tidal phase. Consequently, even when dredging is limited to a small fraction of the channel, it induces significant alterations in the tidal structure throughout the entire estuary: with the maximum depth increase and the longest dredging length, tidal amplification reaches nearly 40% in the upstream section of the channel. This behaviour was observed by Ralston and Geyer (2019) in the Hudson River estuary, where a local change in channel depth at the mouth affected 100 km upstream. Changes in tidal asymmetry after dredging along the estuary are quantified in Fig. 4. This parameter is represented as the ratio between the amplitude of the overtide harmonic 𝑀4and that of the principal tidal constituent 𝑀2. In terms of water level, channel deepening results in a reduction of the 𝑀4amplitude (while increasing the 𝑀2amplitude, as seen in Fig. 3), leading to an overall decrease in tidal asymmetry along the channel compared to the unaltered estuary (Fig. 4a,d,g). This effect becomes more pronounced with both the increase in depth and the dredging length, achieving a maximum reduction of 45% at 𝑥= 33 k mfor 𝛥𝐻 = 40% and 𝑑= 30 k m(scenario 15, Fig. 4g). However, dredging does not alter the overall trend in the 𝑀4∕𝑀2 ratio along the channel, which increases from the estuary mouth up to 𝑥= 33 k m, decreases until 𝑥= 49.5 k mand increases again in the remaining section of the estuary. The increasing asymmetry up to 𝑥= 33 k magrees the hyposynchronous behaviour (decreasing tidal amplitude) of the estuary seen in Fig. 2a. In this region, two distinct linear slopes are observed. The first slope, which is smaller, corresponds Ocean Modelling 196 (2025) 102552 4 G. Martín-Llanes and A. López-Ruiz Fig. 2. Hydrodynamics in the unaltered estuary. (a) 𝑀2and 𝑀4water level amplitudes along the channel. (b) Surface currents at the estuary mouth at neap (NT) and spring (ST) tide. (c) 𝑀2and 𝑀4current amplitudes along the channel. (d) Tidal asymmetry (strength). (e) Phase lag between current and elevation. (f) Tidal asymmetry (direction). Fig. 3. Amplitude and phase ratios along the estuary. (a),(c),(e) Tidal amplitude ratio. (b),(d),(f) Tidal phase ratio. Each row corresponds to a different depth increase. Each curve corresponds to a different dredging length. Subindex 0 indicates magnitudes at the unaltered estuary. to the dredged stretch. Since the 𝑀2amplitude is higher near the mouth and the overtide is small, tidal asymmetry increases slightly in the dredged area. In addition, more significant differences between the unaltered and dredged scenarios are observed in the hyposynchronous frame compared with the remainder of the channel. Since both tidal amplitudes and tidal amplification are smaller in the upper estuary, changes in tidal asymmetry are mainly attributed to the harmonic 𝑀2, i.e. dredging leads to more symmetric cycles (compared to the unaltered scenario) when tidal amplitudes are higher and less symmetric cycles when tidal range is smaller. Tidal current asymmetry is quantified by the ratio 𝑈𝑀4∕𝑈𝑀2. As shown in Fig. 4b,e,h, the magnitude of tidal asymmetry is reduced throughout the estuary proportionally to both the increase in depth and the dredging length. In addition, considering the phase lag between both harmonics, Fig. 4c,f,i shows ebb dominance in all scenarios (2𝜙𝑀2−𝜙𝑀4≈ 350◦). Both results imply that dredging leads to a weaker ebb asymmetry, compared to the unaltered estuary. This is consistent with the results from Winterwerp and Wang (2013), who observed higher flood dominance after dredging. As the ebb dominance is reduced, the tidally averaged seaward transport of water Ocean Modelling 196 (2025) 102552 5 G. Martín-Llanes and A. López-Ruiz Fig. 4. Tidal asymmetry along the channel. (a),(d),(g) Strength of water level asymmetry. (b),(e),(h) Strength of currents asymmetry. (c),(f),(i) Direction of currents asymmetry. Each row corresponds to a different depth increase. Each curve corresponds to a different dredging length, as indicated in Fig. 3. Light blue dashed line indicates magnitudes in the unaltered estuary. will be weaker in the new configurations, consequently enhancing the landward transport of salt (Chen et al.,2019;Zhao et al.,2022) and suspended sediment (Van Maren et al.,2015). This behaviour is further analysed in the following sections. 4.3. Along-channel residual flow structure A further analysis to describe the water transport involves the evaluation of the residual (tide averaged) unit width water flux (RUWF) (Chen et al.,2019): RUWF =1 𝑇∫𝑇 0  𝑄 𝑑 𝑡(3) where  𝑄is the instantaneous rate of water transport per unit width through the water column, 𝑇is the 𝑀2tidal period, and 𝑡is time. The difference in the RUWF between the dredged scenarios with 𝛥𝐻 = 40% and the unaltered estuary is shown in Fig. 5. Dredging behind the salt intrusion limit increases the exchange flow, defined as the difference between the net bottom inflow and net surface outflow, in the dredged area, and the net outflow in the unaltered part of the estuary (Fig. 5a,b,c), leading to an increase in salt intrusion. In contrast, when the dredging length is greater than the salt intrusion length (Fig. 5d,e), a different behaviour is obtained between the residual salt intrusion limit and the transition slope, where negligible residual flux difference is observed. Since residual currents are different from zero in both the unaltered and dredged scenarios, results indicate equivalent tidally averaged hydrodynamic conditions in both cases. Increasing the dredging length results in greater salt intrusion, a slight reduction in the exchange flow variance within the dredged region, and an enhanced outflow in the unaltered section of the channel. On the other hand, the magnitude of the exchange flow and the salt length increase proportionally with the depth increment. In particular, for 𝑑= 5 k mthe net inflow is increased by 0.08 m3s−1, 0.15 m3s−1 and 0.2 m3s−1 for 𝛥𝐻 = 15,30 and 40%, respectively. Further details on the impact of salt intrusion are presented in Section 4.5. To better understand the underlying mechanisms driving changes in the residual flow structure, a detailed analysis of tidal currents is conducted. Specifically, a representative cross-section from each of the three regions identified is selected, and the difference in the vertical distribution of the axial current (thalweg) between Scenario 14 and the unaltered estuary is examined over two tidal cycles (Fig. 6). Representative sections are located at 𝑥= 7.5 k m,𝑥= 18 k m,𝑥= 22.5 k m, respectively. In the estuarine circulation region (Fig. 6a), which corresponds to the dredged area below the salt intrusion limit, the current difference shifts sign between the flood and ebb tides. On the one hand, a negative difference is observed at flood tide. Since these current values are positive, this result indicates a reduction in the flood current throughout the entire water column after dredging. The maximum negative difference (−0.25 ms−1) is observed between the maximum flood and the high water slack (HWS) at the surface, and it decreases gradually towards the bottom. On the other hand, a positive difference is observed during the ebb tide, indicating a decrease in the ebb current throughout the water column, except at the surface during the maximum ebb. Maximum positive values (+0.20 ms−1) are obtained between maximum ebb and low water slack (LWS) at the bottom, decreasing towards the surface. As a result, time averaging leads to a net outflow at the surface and a net inflow at the bottom, as shown in Fig. 5. Considering the dredged area between the salt intrusion limit and the transition slope (Fig. 6b), the same behaviour is observed, with negative values mainly during the flood, and positive values mainly during the ebb. However, the vertical distribution of the axial current difference is uniform, and the maximum positive and negative values are equivalent. This means that flood and ebb currents decrease by the same amount throughout the water column, with the magnitude of the change varying depending on the tidal phase. The maximum negative difference occurs near high water slack (HWS), while the maximum positive difference is observed near low water slack (LWS). Symmetric changes in tidal currents lead to equivalent residual current configurations in both the dredged and unaltered estuaries, resulting in a negligible difference in the RUWF across the entire water column in both scenarios (Fig. 5). Finally, in the unaltered stretch of the estuary (Fig. 6c), the sign shift is reversed, resulting in larger flood and ebb currents. This result is consistent with the amplification of the 𝑀2current within this region, leading to tidal cycles with higher amplitudes of both flood and ebb Ocean Modelling 196 (2025) 102552 6 G. Martín-Llanes and A. López-Ruiz Fig. 5. Difference in the residual unit width water flux between the dredged scenarios with a depth increase of 𝛥𝐻 = 40% and the unaltered estuary. Each panel corresponds to a different dredging length. Solid black line represents the dredged bathymetry; dashed black line represents the unaltered estuary. Solid red line indicates residual salt intrusion after dredging; dashed red line indicates residual salt intrusion in the unaltered estuary. Positive values indicate inflow. currents. This amplification is uniform throughout the water column. In addition, since ebb currents increase more than flood currents and persist for a greater fraction of the tidal cycle, a net outflow (negative RUWF difference) is observed beyond the transition slope in Fig. 5. This result aligns with the ebb dominance shown in Fig. 4c,f,i. 4.4. Momentum balance This section examines variations in the momentum balance to identify the primary factors driving hydrodynamic impacts in dredged estuaries. The Delft3D momentum equation in the horizontal axial direction, neglecting the acceleration due to Coriolis force, reads (Deltares, 2016): 𝜕 𝑢 𝜕 𝑡+𝑢𝜕 𝑢 𝜕 𝑥+𝑣𝜕 𝑢 𝜕 𝑦+𝜔 𝐻 𝜕 𝑢 𝜕 𝑧+𝑔𝜕 𝜂 𝜕 𝑥+𝑔𝐻 𝜌0 𝜕 𝜌 𝜕 𝑥−𝐹𝑥−1 𝐻2 𝜕 𝜕 𝑧(𝜈𝜕 𝑢 𝜕 𝑧)−𝑀𝑥= 0(4) where 𝑢, 𝑣, 𝜔are the along-channel, cross-channel and vertical currents, 𝜂is the water level, 𝑔is the gravitational acceleration constant, 𝐻is the water depth measured along the z-coordinate, 𝜌0is the seawater density, 𝜌is the water density, and 𝜈is the eddy viscosity. These terms in Eq. (4) represent the local acceleration, the tidal stress, the acceleration due to lateral transport, the vertical advection of momentum, the barotropic pressure gradient, the baroclinic gradient, the acceleration due to viscosity (𝐹𝑥), the vertical diffusion of momentum and the bed shear force (𝑀𝑥), respectively. Momentum terms were first evaluated during two tidal cycles at the channel thalweg of the three representative sections defined in Section 4.3. The obtained results, which are shown in Section 3 of the Supplementary Material, suggest that the barotropic pressure gradient and bed shear forces are the dominant factors, so that impacts of channel deepening on hydrodynamics can be assessed in terms of changes in these two key terms. Specifically, alterations in the flow structure at the surface are closely linked to changes in the barotropic pressure gradient, while modifications in bottom currents are primarily influenced by changes in bed shear forces. This is shown in Fig. 7, where the barotropic gradient and bed shear forces in both the unaltered and dredged (Scenario 14) situations are represented along two tidal cycles. First, as shown in Fig. 7a, the pressure gradient is reduced in the dredged area below the salt intrusion limit, and exhibits a 2hour phase lag compared to the results in the unaltered estuary. A clear correlation with the axial current is observed, as the sign of the difference in the barotropic term between the dredged and unaltered configurations is negative during flood tide and positive during ebb Ocean Modelling 196 (2025) 102552 7 G. Martín-Llanes and A. López-Ruiz Fig. 6. Difference in the vertical distribution of the axial current between the dredged (Scenario 14) and unaltered scenarios along two tidal cycles. (a) 𝑥= 7.5 k m. (b) 𝑥= 18 k m. (c) 𝑥= 22.5 k m. Positive values indicate inflow. Black contours indicate zero velocity. The right axis represents the surface current at the thalweg. tide, which is consistent with the vertical distribution of currents at the surface (Fig. 6a). Moreover, the time evolution of the barotropic term suggests a clear ebb-dominated asymmetry. Since this term is proportional to the surface current, this result implies a negative tidally averaged current, i.e., a net surface outflow, which is consistent with observations in Fig. 5. Regarding the frictional term in this region (Fig. 7b), since bed shear opposes the bottom current, maximum positive values are observed at ebb tide and vice versa. Dredging leads to a general decrease in the frictional term, which is particularly relevant at ebb tide. Asymmetry changes from ebb-dominated (positive bed shear acceleration) to flood-dominated (negative bed shear acceleration) after dredging, leading to a positive net inflow at the bottom, as shown in Fig. 5. In Fig. 6a, stronger positive current differences were also observed at the bottom. The overall result, considering changes in both surface and bottom currents, is an increase in estuarine circulation, which is proportional to the depth increase. In the dredged part of the channel beyond the salt intrusion limit, the barotropic gradient shows a slight phase lag (∼ 1hour) compared to the unaltered estuary (Fig. 7c), with no significant changes in amplitude. Increasing channel depth reduces the water surface slope. However, as friction decreases due to dredging, tidal velocities and effective drag increase, offsetting the effect of the increased depth (Ralston et al., 2019). Consequently, the pressure gradient undergoes a time shift in this region, without significant changes in its amplitude. Regarding Fig. 7d, dredging results in a lower and more symmetric time distribution of bed shear. Values at the surface and bottom suggest equivalence between the tidally averaged hydrodynamic conditions at the dredged and unaltered estuaries, resulting in a negligible difference in the RUWF (Fig. 5). Finally, in the unaltered part of the channel, the pressure gradient increases during the ebb tide and shows no significant changes during the flood tide. This indicates that the increased depth compensates for the flood tide increase, but it does not balance the increased ebb (Fig. 7e). Bed shear acceleration also increases at the ebb (Fig. 7f) in response to the larger increase in bottom ebb current. Both results highlight the dominance of the ebb current throughout the entire water column, leading to a net outflow, as observed in Fig. 5. 4.5. Impacts on salt intrusion In this section, the effects of channel deepening on salt intrusion are addressed. Fig. 8shows the ratio of the maximum salt intrusion in the different dredged scenarios to the value obtained for the unaltered scenario. The overall results indicate that salt intrusion increases with both channel depth and dredging length. The maximum ratio is observed in Scenario 15, where salt intrusion increases in 15.5%. Additionally, a linear relationship between salt intrusion and depth increase is observed, with the slope varying between 0.19 and 0.37 depending on the dredging length. Attending to Fig. 8b, a local minimum is observed for a dredging length equal to the maximum salt length in the unaltered estuary (𝑑=𝑋𝑚𝑎𝑥,0= 15 k m). This behaviour, occurring for any depth increase, suggests a distinct response in salt intrusion depending on whether deepening is considered short or long, relative to the maximum salt length in the unaltered estuary. Specifically, salt intrusion exhibits higher sensitivity to the dredging length for short operations, with this effect becoming more pronounced as 𝛥𝐻 increases. In contrast, this effect weakens for long-distance operations. Salt intrusion as a function of the dredging length can be expressed by Eq. (5): 𝛥𝑋[%] =⎧ ⎪ ⎨ ⎪ ⎩ 𝑏𝑑 𝑋𝑚𝑎𝑥,0 +𝑏0𝑑 < 𝑋𝑚𝑎𝑥,0 𝑐𝑑 𝑋𝑚𝑎𝑥,0 +𝑐0𝑑 > 𝑋𝑚𝑎𝑥,0 (5) where 𝑏= [4.392,8.864,11.4] and 𝑐= [3.95,6.281,7.45] for 𝛥𝐻 = [15,30,40]%. Notably, the difference in slope between short (b) and long (c) operations for 𝛥𝐻 = 40% is ten times greater than that obtained for 𝛥𝐻 = 15%. This result highlights that the extent of dredging plays a more significant role in modulating salt intrusion than the depth increase alone. Specifically, the increase in salt intrusion between the shortest and longest dredging lengths for 𝛥𝐻 = 40% surpasses the increase observed between any of the depth increments. 5. Discussion: Implications on the management of navigational channels Many ports worldwide are located in estuaries and may consequently be affected by numerous coastal processes, including sediment transport and deposition. Sediment infilling reduces the available water depth, thus shortening the time for which vessels can navigate through the estuary and limiting port operativity (Álvarez et al.,2017), i.e., the fraction of time during which water depth is higher than the required draft for a certain port operation (Zarzuelo et al.,2019). Regarding this issue, dredging operations are periodically planned (Sepehri et al., 2024). This section explores variations in the channel operativity after dredging, which is determined by changes in the water depth. Water depth after dredging results from the ensemble consideration of: (1) the increased elevation of bed level below the mean sea level (𝛥𝐻) and (2) the resulting changes in the water level due to the altered tide. Although the effect of the first term is limited to the dredged portion of the channel (𝑥≤𝑑), changes in the tidal level affect the entire system. As seen in Section 4.2 (Fig. 3), dredging leads to tidal amplification, resulting in increased high and low water levels and different durations of the tidal cycle. Regarding the latter, increased low levels lasting for a longer period of time play an unfavourable role in respect to channel operativity. Although this effect is expected to be negligible compared to the increased bed level elevation in the dredged portion of the channel, channel operativity for low target drafts in certain regions of the estuary where 𝑥 > 𝑑could be compromised by the duration of the low tide. The change in channel operativity (𝛥𝑂 𝑝𝑡𝑣) for a given target draft (𝑦𝑜𝑏𝑗 ) is expressed as the absolute difference between operativity in the dredged estuary (𝑂 𝑝𝑡𝑣) and operativity in the unaltered estuary (𝑂 𝑝𝑡𝑣0). Both quantities represent the time fraction, over a 15-day Ocean Modelling 196 (2025) 102552 8 G. Martín-Llanes and A. López-Ruiz Fig. 7. Momentum balance variation: Barotropic pressure gradient and bed shear forces along two tidal cycles. (a), (b) 𝑥= 7.5 k m. (c), (d) 𝑥= 18 k m. (e), (f) 𝑥= 22.5 k m. Blue solid and dashed lines represent the dredged (Scenario 14) and unaltered estuaries, respectively. Solid black line represents surface current (right axis). Fig. 8. Difference in the maximum salt intrusion between the unaltered (𝑋𝑚𝑎𝑥,0) and dredged (𝑋𝑚𝑎𝑥) scenarios. Curves in (a) correspond to each dredging length; curves in (b) correspond to the different depth increases. period, during which water depth exceeds the target draft. Hence, 𝛥𝑂 𝑝𝑡𝑣(𝑦𝑜𝑏𝑗 )≤100%; positive values indicate an increased operativity after dredging; and negative values indicate the opposite. In addition, 𝛥𝑂 𝑝𝑡𝑣(𝑦𝑜𝑏𝑗 ) = 0implies that the operativity is the same in both the unaltered and dredged scenarios. This situation will occur, for instance, for low target drafts for which there will always be a greater water depth in both the unaltered and dredged bathymetries. In addition, 𝛥𝑂 𝑝𝑡𝑣(𝑦𝑜𝑏𝑗 ) = 0will also occur for high target drafts for which water depth is always lower in both the unaltered and dredged bathymetries. Fig. 9represents 𝛥𝑂 𝑝𝑡𝑣 across the estuary in Scenario 5 considering a range of target drafts scaled with the dredged channel depth (𝐻). As illustrated, the magnitude and the sign of 𝛥𝑂 𝑝𝑡𝑣 change abruptly from the dredged (𝑥≤30 k m) to the unaltered (𝑥≥30 k m) part of the channel. In addition, these two variables are directly related to the value of the target draft. First, considering the region affected by channel deepening, the function 𝛥𝑂 𝑝𝑡𝑣(𝑦𝑜𝑏𝑗 )is positive and adopts a parabolic shape at each point of the estuary. This result implies that operativity in this region increases or maintains its value before dredging, depending on the value of the target draft. As aforementioned, changes in channel operativity are determined by changes in the water depth. In this part of the channel, the contribution of the increased bed level elevation causes the post-dredge water depth time series (𝑊𝑑(𝑡)) to be higher than the water depth time series at the unaltered estuary (𝑊𝑑 ,0(𝑡)) for each time (t) (Fig. 10a). Hence, the function 𝛥𝑂 𝑝𝑡𝑣(𝑦𝑜𝑏𝑗 )takes a different value depending on the value of 𝑦𝑜𝑏𝑗 with respect to 𝑊𝑑(𝑡)and 𝑊𝑑 ,0(𝑡), as illustrated in Fig. 10b. Attending to this figure, five different regions within the 𝑊𝑑−𝑡space are distinguished: Region (1), where 𝑦𝑜𝑏𝑗 < 𝑊𝑑 ,0(𝑡); Region (2), where 𝑦𝑜𝑏𝑗 is within the interval defined by 𝑊𝑑 ,0(𝑡); Region (3), where 𝑦𝑜𝑏𝑗 > 𝑊𝑑 ,0(𝑡)and 𝑦𝑜𝑏𝑗 < 𝑊𝑑(𝑡); Region (4), where 𝑦𝑜𝑏𝑗 is within the interval defined by 𝑊𝑑(𝑡)and Region (5), where 𝑦𝑜𝑏𝑗 > 𝑊𝑑(𝑡). The extreme regions (1, 5) are characterised by a null change in the channel operativity; in Region (1) target drafts are sufficiently low to have an operativity of 100% in both the undredged and dredged channels, whereas in Region (5) high drafts lead to a operativity of zero in both cases. Both situations lead to 𝛥𝑂 𝑝𝑡𝑣(𝑦𝑜𝑏𝑗 ) = 0, which means that no improvement in operativity has been achieved for these drafts. On the other hand, operativity after dredging shifts from Ocean Modelling 196 (2025) 102552 9