Coastal Sediment Transport Research Articles

MODELING SEDIMENT TRANSPORT IN THE YELLOW RIVER MOUTH

Yellow River is famous for its high sediment concentration, which carries a huge amount of sediment into Bohai Sea during the flood season. In the mouth of Yellow River, the suspended sediment concentration (here after: SSC) are relatively high and the average depth is generally shallow. Recent studies indicated that in the shallow and micro-tide area, wave usually is a primary mechanism for bottom sediment resuspension. Most numerical studies of sediment transport in the Yellow River mouth didn't include the wave effects. To analyze the sediment transport in the Yellow River mouth with wave effect, a Coastal Sediment Transport Modeling System (CSTMS) was applied in this study.

Prediction and assimilation of surf-zone processes using a Bayesian network: Part I: Forward models

Prediction of coastal processes, including waves, currents, and sediment transport, can be obtained from a variety of detailed geophysical-process models with many simulations showing significant skill. This capability supports a wide range of research and applied efforts that can benefit from accurate numerical predictions. However, the predictions are only as accurate as the data used to drive the models and, given the large temporal and spatial variability of the surf zone, inaccuracies in data are unavoidable such that useful predictions require corresponding estimates of uncertainty. We demonstrate how a Bayesian-network model can be used to provide accurate predictions of wave-height evolution in the surf zone given very sparse and/or inaccurate boundary-condition data. The approach is based on a formal treatment of a data-assimilation problem that takes advantage of significant reduction of the dimensionality of the model system. We demonstrate that predictions of a detailed geophysical model of the wave evolution are reproduced accurately using a Bayesian approach. In this surf-zone application, forward prediction skill was 83%, and uncertainties in the model inputs were accurately transferred to uncertainty in output variables. We also demonstrate that if modeling uncertainties were not conveyed to the Bayesian network (i.e., perfect data or model were assumed), then overly optimistic prediction uncertainties were computed. More consistent predictions and uncertainties were obtained by including model-parameter errors as a source of input uncertainty. Improved predictions (skill of 90%) were achieved because the Bayesian network simultaneously estimated optimal parameters while predicting wave heights.

Development of a Coupled Ocean–Atmosphere–Wave–Sediment Transport (COAWST) Modeling System

Understanding the processes responsible for coastal change is important for managing our coastal resources, both natural and economic. The current scientific understanding of coastal sediment transport and geology suggests that examining coastal processes at regional scales can lead to significant insight into how the coastal zone evolves. To better identify the significant processes affecting our coastlines and how those processes create coastal change we developed a Coupled Ocean–Atmosphere–Wave–Sediment Transport (COAWST) Modeling System, which is comprised of the Model Coupling Toolkit to exchange data fields between the ocean model ROMS, the atmosphere model WRF, the wave model SWAN, and the sediment capabilities of the Community Sediment Transport Model. This formulation builds upon previous developments by coupling the atmospheric model to the ocean and wave models, providing one-way grid refinement in the ocean model, one-way grid refinement in the wave model, and coupling on refined levels. Herein we describe the modeling components and the data fields exchanged. The modeling system is used to identify model sensitivity by exchanging prognostic variable fields between different model components during an application to simulate Hurricane Isabel during September 2003. Results identify that hurricane intensity is extremely sensitive to sea surface temperature. Intensity is reduced when coupled to the ocean model although the coupling provides a more realistic simulation of the sea surface temperature. Coupling of the ocean to the atmosphere also results in decreased boundary layer stress and coupling of the waves to the atmosphere results in increased bottom stress. Wave results are sensitive to both ocean and atmospheric coupling due to wave–current interactions with the ocean and wave growth from the atmosphere wind stress. Sediment resuspension at regional scale during the hurricane is controlled by shelf width and wave propagation during hurricane approach.

Modelling of wave–current interactions at the mouths of the Danube

The target of the present study is the entrance to the Danube Delta in the Black Sea. The wave conditions in this coastal sector are usually significant from an energetic point of view and the relatively strong currents induced there by the outflow from the Danube lead to interactions between waves and currents. This process modifies considerably both the magnitude and direction waves, affecting also coastal navigation and sediment transport patterns. In order to assess the effects of the wave–current interactions, the simulating waves nearshore (SWAN) model was considered for developing a multilevel wave prediction system. Validations against measured data were carried out for each computational level. Five case studies corresponding to the most relevant patterns of the environmental matrix were analyzed. Finally, in order to assess the current effect for a longer timescale, an analysis concerning the variation of the main wave parameters was performed for a 3-month period considering some reference points. The results show that the currents produce considerable changes in the wave field, especially as regards the significant wave heights, mean wave directions and wavelengths. The Benjamin–Feir index was also estimated. The analysis of the variation induced by the current over this spectral shape parameter indicates that, in certain conditions, in the target area the wave heights cannot be considered Rayleigh distributed and freak waves may also occur.

Effects of Low Water Temperature on Longshore Sediment Transport on a Subarctic Beach, Hudson Bay, Canada

Measurements of longshore sand transport, wave breaker height, longshore current velocity, and water temperature and salinity were carred out on a subarctic beach under varying water temperature conditions during summer and fall. Water temperature during summer varied from 9°C to 17°C, while temperature ranged from −1.75°C to 1.25°C during the fall experiments. During periods with low water temperatures, considerably higher rates in both bedload and suspended sediment transport were observed compared with the results obtained during summer with similar wave heights and current velocities. The increase in suspended sediment transport associated with low water temperatures is likely due to lower sediment fall velocities resulting from significant increases in fluid kinematic viscosity, which ranged from 1.1 to 1.37 × 10−6 m2 s−1 in summer and from 1.74 to 1.92 × 10−6 m2 s−1 in the fall. Less but still significant increase in bedload sediment transport in cold water is interpreted to be the result of temperature-induced variations in fluid viscosity that affect turbulence and boundary layer dynamics as well as bed configuration. Frazil ice occurred on some occasions during the fall experiments, but with the small number of samples collected when frazil ice crystals were present in the water column, it was not possible to establish a relationship between sediment transport rates and the presence or absence of ice. Our results nevertheless indicate that in cold climate regions extremely low water temperatures may cause a substantial increase of coastal sediment transport.

Stretched Eulerian coordinate model of coastal sediment transport

A one-dimensional vertical model of sediment and dissolved material transport in a coupled, vertically resolved benthic-pelagic system has been developed. An apparently unique feature of the model is that the governing equations are first formulated in a time-varying Eulerian frame and then solved numerically, which provides a consistent approach for incorporating resuspension/deposition, swelling/consolidation, and biodiffusion processes within a single modelling framework. Numerical experiments were carried out with varying grids to test diffusive properties of the model. A stretched numerical grid that maintains high vertical resolution of the top sediments and constrains numerical diffusion in deep layers is an integral part of the model. The developed model provides a test bed for testing scientific hypothesis pertaining to coastal sediment transport. The model can also be used to support biogeochemical studies in a coupled benthic-pelagic system.

Numerical study of circulation on the inner Amazon Shelf

We studied the circulation on the coastal domain of the Amazon Shelf by applying the hydrodynamic module of the estuarine and coastal ocean model and sediment transport. The first barotropic experiment aimed to explain the major bathymetric effects on tides and those generated by anisotropy in sediment distribution. We analyzed the continental shelf response of barotropic tides under realistic bottom stress parametrization (C d ), considering sediment granulometry obtained from a faciologic map, where river mud deposits and reworked sediments areas are well distinguished, among others classes of sediments. Very low C d values were set in the fluid mud regions off the Amapa coast (1.0 10 − 4), in contrast to values around 3.5 10 − 3 for coarser sediment regions off the Para coast. Three-dimensional experiments represented the Amazon River discharge and trade winds, combined to barotropic tide influences and induced vertical mixing. The quasiresonant response of the Amazon Shelf to the M2 tide acts on the local hydrodynamics by increasing tidal admittance, along with tidal forcing at the shelf break and extensive fluid mud regions. Harmonic analysis of modeled currents agreed well with the analysis of the AMASSEDS observational data set. Tidal-induced vertical shear provided strong homogenization of threshold waters, which are subject to a kind of hydraulic control due to the topographic steepness. Ahead of the hydraulic jump, the low-salinity plume is disconnected from the bottom and acquires negative vorticity, turning southeastward. Tides act as a generator mechanism and topography, via hydraulic control, acts as a maintainer mechanism for the low-salinity frontal zone positioning. Tidally induced southeastward plume fate is overwhelmed by northwestward trade winds so that they, along with background circulation, probably play the most important role on the plume fate and variability over the Amazon Shelf.

P. S. Balson & M. B. Collins(eds) 2007. Coastal and Shelf Sediment Transport. Geological Society Special Publication no. 274. v + 162 pp. London, Bath: Geological Society of London. Price £65.00, US $130.00; GSL members' price £32.50, US $65.00; AAPG/SEPM/GSA/RAS/EFG/PESGB members' price £39.00, US $78.00 (hard covers). ISBN 9781 86239 217 5.

The Geological Society of London, as the oldest of such societies, is looked towards (not least by me) for leadership and excellence in Earth Sciences. Coastal and Shelf Sediment Transport is the 274th Special Publication of that learned society. The book continues the tradition of the Society through a compilation of selected recent studies on varying approaches used to tackle a complex and fundamentally important subject within a rapidly evolving sector of Earth Science. The book results from a stimulating conference held …

Time-averaged near-bed suspended sediment concentrations under waves and currents: Comparison of measured and model estimates

Profiles of suspended sediment concentration and velocity were measured over a 15-day period at a near-shore site off Santa Cruz, CA in Monterey Bay. The concentration and velocity data were collected with an Acoustic Backscattering System (ABS) and Acoustic Current Profiler (ACP) that were mounted on a bottom tripod. High-resolution bottom scanning sonar was also attached to the tripod to provide images of bed features during the experiment. Hourly time-averaged near-bed concentrations of suspended sediment were calculated from three models and compared with the measurements. Surface waves and currents that were generated by a storm of moderate intensity caused bed stresses that exceeded threshold stress for D 50=0.02 cm, the median size of the moderately well-sorted bottom sediment, over a period of about 7 days. Estimates of the concentration at 1 cm above the bottom, C a1, were obtained using the ABS measurements. These observations have been compared with predictions for the concentration at 1 cm above the bottom, C 1. Nielsen's models for reference concentration C o [Nielsen, P., 1986. Suspended sediment concentrations under waves. Coastal Engineering 10, 32–31; Nielsen, P., 1992. Coastal Bottom Boundary Layers and Sediment Transport, Advanced Series on Ocean Engineering. World Scientific, Hackensack, NJ.] are purely wave-based and do not include effects of bottom currents on bed stress and bedform scales. C 1 calculated from this model compared well with measured C a1 when currents were weak and small oscillatory ripples were observed in the sonar images. However, during the 3-day period of highest bottom stresses modeled C 1 did not compare well to C a1. The other two models for C 1, Glenn and Grant [Glenn, S.M., Grant, W.D., 1987. A suspended sediment stratification correction for combined wave and current flows. Journal of Geophysical Research 92(C8), 8244–8264.] and van Rijn and Walstra [Van Rijn, L.C., Walstra, D.J.R., 2004. Description of TRANSPOR2004 and implementation in Delft3D-ONLINE. Interim Report prepared for DG Rijkswaterstaat, Rijksinstituut voor Kust en Zee. Delft Hydraulics Institute, The Netherlands.], accounted for combined wave-current stresses and included different formulations for predicting bedform scales. C 1 calculated from these models compared well with measurements throughout the early and middle period of the storm, but underpredicted measured values during the latter part of the experiment. An alternative method based on dimensional analysis provided a straightforward way to relate C 1 to bed stress and sediment parameters. This new relationship based on dimensional analysis explained about 77% of the variability in C a1 derived from the ABS data.

Field Signatures of the SE-Asian Mega-Tsunami along the West Coast of Thailand Compared to Holocene Paleo-Tsunami from the Atlantic Region

The Andaman-Sumatra Tsunami of Dec. 26, 2004, was by far the largest tsunami catastrophe in human history. An earthquake of 9 to 9.3 on the Richter scale, the extension of waves over more than 5000 km of ocean and run-ups up to 35 m are its key features. These characteristics suggest significant changes in coastal morphology and high sediment transport rates. A field survey along the west coast of Thailand (Phuket Island, Khao Lak region including some Similan Islands, Nang Pha mangrove areas and Phi Phi Don Islands) seven to nine weeks after the tsunami, however, discovered only small changes in coastal morphology and a limited amount of dislocated sediments, restricted to the lower meters of the tsunami waves. This is in striking contrast to many paleo-tsunami’s events of the Atlantic region. Explanations for this discrepancy are sought in: a. Mechanics of the earthquake. A rather slow shock impulse on the water masses over the very long earthquake zone, b. Shallow water in the earthquake zone, and c. Bathymetry of the foreshore zone at the impacted sites. Shallow water west of Thailand has diminished wave energy significantly. The differences in geomorphological and sedimentological signatures of this tsunami compared with many paleo-tsunami worldwide makes it unsuitable to be used as a model for old and future tsunami imprints by an event of this extreme energy and extension.

The surf zone distortion of beach profiles in small–scale coastal models

In recent costal modeling practice physical models are constructed in undistorted scales preserving the Froude similarity. Correction for scale effect is obtained a posteriori by means of a distortion coefficient, which multiplies the final quasi–equilibrium profile abscissa and provides a correction of the predicted prototype beach slope. The choice of this distortion coefficient is therefore crucial as it highly affects the reliability of the prototype reproduction. Furthermore, it has been observed that, because of the complexity of coastal sediment transport phenomena and beach profiling in the surf zone, a single coefficient does not succeed in correcting the whole beach profile. The scale effects affecting the surf zone of Froude models has been examined in this paper.A method for correcting scale effects in the surf zone is proposed. It is based on a distortion parameter, αsurf, to be applied only to the beach surf zone, leaving the offshore and foreshore parts undistorted, above and below the surf zone. An empirically based formula is found for calculating αsurf as a function of the ratio between a parameter representing bed load transport and the characteristics of the model sand, expressed by the mean diameter D 50 and the Kurtosis parameter of the particle–size curve. It is commonly accepted that scale effects can be minimized if, besides Froude similarity, the Dean parameter is the same both in the prototype and in the model. However, this often implies very large model scales, not compatible with laboratory facilities. It is shown that the introduction of the αsurf parameter allows to employ a more practical small model scale by properly correcting scale effects, even though Dean similarity is not respected.

An evaluation of combined geophysical and geotechnical methods to characterize beach thickness

Beaches provide sediment stores and have an important role in the development of the coastline in response to climate change. Quantification of beach thickness and volume is required to assess coastal sediment transport budgets. Therefore, portable, rapid, non-invasive techniques are required to evaluate thickness where environmental sensitivities exclude invasive methods. Site methods and data are described for a toolbox of electrical, electromagnetic, seismic and mechanical based techniques that were evaluated at a coastal site at Easington, Yorkshire. Geophysical and geotechnical properties are shown to be dependent upon moisture content, porosity and lithology of the beach and the morphology of the beach–platform interface. Thickness interpretation, using an inexpensive geographic information system to integrate data, allowed these controls and relationships to be understood. Guidelines for efficient site practices, based upon this case history including procedures and techniques, are presented using a systematic approach. Field results indicated that a mixed sand and gravel beach is highly variable and cannot be represented in models as a homogeneous layer of variable thickness overlying a bedrock half-space.

“Coastal dynamics and human impact: south-eastern North Sea”, an overview

In these times of increasing human usage, coastal zones in particular have come under ever increasing socioeconomic pressure. The exploitation of the potential of the land–sea interface in terms of settlement, traffic, constructions, harbour development, tourism, fisheries, offshore structures, etc. goes along with severe impacts on natural environments. Past experiences have shown the vulnerability of the marine environment, e.g. in the form of largescale response to small-scale coastal engineering (cf. Pilkey and Dixon 1996). Thus, management must aim at a sustainable development of coastal zones based on detailed knowledge of the natural system, the underlying processes and their response to external forcing at a variety of temporal and spatial scales. In order to deepen our understanding of coastal dynamics, the Department of Marine Science of Senckenberg Institute in Wilhelmshaven and the Research Center Ocean Margins (RCOM) at Bremen University, Germany, in cooperation with the Institute of Geography at the University of Copenhagen, Denmark, are carrying out a research programme on hydrodynamic, sediment dynamic and morphodynamic processes on different spatial scales in the south-eastern North Sea, and how these may be affected by human impact. In a variety of marine environments in this region, detailed studies are underway using state-ofthe-art technical instrumentation and numerical models to describe, analyse and interpret the complex interrelations of coastal currents, suspended and near-bed sediment transport, and bed evolution at various scales. The aim of this thematic issue of Geo-Marine Letters, no. 26(3), which is guest-edited by Winter and Bartholoma (2006), is to provide an overview of the research aims of the programme and to introduce the articles and their authors. The research was variously funded by the Deutsche Forschungsgemeinschaft as part of the DFG Research Center Ocean Margins (RCOM) at the University of Bremen (Germany), the Senckenberg Natural History Society in Frankfurt (Germany), the GKSS Research Centre, in Geesthacht (Germany), and the Science Research Council of Denmark.

Quasi-hydrodynamic lubrication effect of clay particles on sand grain erosion

Minor changes in the mass physical properties of submerged sand beds can have significant consequences relative to bed stability against erosion. To examine the effect of small amounts of clay-sized particles in bed pore water on the critical shear stress τ c for the erosion of sand grains, flume experiments were carried out on the erosion of quartz sand beds impregnated with clay particles. Starting with no clay, as the clay mass fraction ψ was increased, τ c was found to decrease below the value for pure sand τ co at ψ = ψ m, then reverted to τ co at ψ = ψ r and continued to increase above τ co as ψ was increased further. Post-experimental analysis suggests that ψ r is the pore space-filling fine sediment fraction above which sand erosion is significantly influenced by clay. In the range of ψ ≥ ψ m, slider-bearing type lubrication due to the viscosity of the clay-laden interstitial fluid appears to govern the dependence of τ c on ψ, mimicking Petroff's law of thick-film lubrication. When ψ < ψ m, as ψ decreases lubrication is increasingly curtailed by grain asperities, and τ c reverts ultimately to τ co at ψ = 0. An equation relating τ c to ψ is proposed in analogy with the quasi-hydrodynamic Stribeck function for lubrication. The observed effect of clay particles appears to be significant enough to require its consideration in coastal and estuarine sediment transport modeling. It may also be a factor in the estimation of bed stability when biological activity in the benthic boundary layer introduces fine particles in clean sand beds.

Fall velocities of natural sediment particles: a simple mathematical presentation of the fall velocity law

The fall velocities of large natural sediment particles (shingle) and of commonly used model sediments (anthracites) have been measured using a video imaging technique. The study covers a Reynolds number ranging from 35 to 23,000, a significant extension to the existing database for natural sand particles (up to 1500). Making use of the extended database, a simple empirical equation has been derived giving the relationship between the particle fall velocity and sieve diameter. Additional analysis of the drag coefficient has led to a revised formula based on that of Fredsoe and Deigaard (Mechanics of Coastal Sediment Transport. Advanced Series on Ocean Engineering, World Scientific, Singapore), providing a better fit for the data obtained at high Reynolds numbers.

A 3-D Wave-Current Driven Coastal Sediment Transport Model

Most of the existing sediment transport models are not synchronously driven by both the wave field and the flow field. This paper describes a 3D sediment transport model with waves and currents directly coupled within the model to continuously account for different-scale activities especially those that have significant contribution to local sediment transport processes such as formation of sediment plumes and turbidity maxima. A practical issue in modeling coastal sediment transport, besides the concern of model accuracy, is the efficiency of the model. In the present model, the wave action equation, instead of the computational demanding elliptic mild-slope equation, is used to calculate the wave parameters. The wave action equations take into account wave refraction and diffraction as well as the tidal hydrodynamic modification. The calculation of the wave and current forcing is coupled during the time marching process so that the effects due to short-term activities can be considered. The model has been verified against laboratory measurements and has also been applied to simulate actual sediment transport situations in the Pearl River Estuary (PRE), China. It has been quantitatively shown that the suspended sediment concentration in the PRE increases significantly when waves are present. Sediment deposition occurs at the upstream region of the PRE while erosion takes place mostly at the down-estuary region due to exposure to wave actions.

Density gradation in cross-shore sediment transport

One of the fundamental properties of a coastline is its sediment composition. Coastal sediments are rarely composed of one type of sediment. Due to these differences, the sediments are sorted on the beach and foreshore. The effect of density variations of the sediment on coastal sediment transport has been studied in a wave flume experiment. Two sands with an equal grain size distribution but with different densities (heavy minerals and quartz) have been used. Detailed measurements of profile evolution and of sediment composition have been used to assess the sediment transport rates for each sediment fraction. The experiments show that the presence of heavy minerals in the sediment results in reduced erosion on the beach face; the breaker bar is smaller and its crest is more pronounced. The transport rates of different sediment fractions are not only determined by the availability at the bed. In the inner surf zone close to the beach, armouring occurs; in the inner surf zone close to the breaker bar, sediments are entrained more easily than expected. For effective armouring, the heavy mineral fraction in the bed should exceed a certain level. This threshold concentration is reached first close to the beach and extends later in seaward direction.

Eddy viscosity and Eulerian drift over rippled beds in waves

This paper reports the results of an experimental study of the Eulerian drift induced by weakly asymmetrical progressive waves in the bottom boundary layer over rippled beds in the transitional flow regime. Fourteen tests were carried out in a wave flume and one test in a steady flow flume for the estimation of the Nikuradse roughness length k s of the artificial rippled bed which was used. The fluid velocities were measured with a two-component laser-Doppler anemometer. The edge of the bottom wave boundary layer is estimated to be located at one Nikuradse roughness length above the level midway between crest and trough of the ripples. Momentum transfer in this boundary layer is dominated by organized vortices. An analytical model based on a time-varying eddy viscosity was developed by Davies and Villaret [J. Geophys. Res. 104 (C1) (1999) 1465] for the estimation of the wave-induced Eulerian drift above rippled and very rough beds in the turbulent flow regime. Present data show that the applicability of this model can be extended to the “lower” part of the transitional flow regime in the parameter ranges 1000≤ R≤6500 and 1≤ a/ k s≤3, where R is the flow Reynolds number and a is the orbital amplitude of fluid, if adjusted values of the model coefficients which represent the variation amplitude of the symmetrical and asymmetrical time-varying components of the eddy viscosity are used. The drift at the edge of the boundary layer is oriented in the direction of wave propagation in the transitional flow regime, while this drift is oriented in the opposite direction in the turbulent flow regime. The vertical profiles of Eulerian drift, horizontally averaged over a ripple length, are also compared with Longuet-Higgins' [Philos. Trans. R. Soc. Lond., A 245(903) (1953) 535] solution for a laminar flow above a smooth bed and with Nielsen's [Coastal bottom boundary layers and sediment transport. Advanced Series on Ocean Engineering, Vol. 4, World Scientific, Singapore, Chap. 1, pp. 40–60, 1992] solution involving a time-invariant eddy viscosity for turbulent flows over fairly rough beds.

On fine sediment transport by long waves in the swash zone of a plane beach

We calculate suspended sediment transport, erosion and deposition under reflected long waves on a plane beach. Our method employs the shallow-water equations, and is based on the calculation of the concentration field in Lagrangian coordinates: this allows results to be obtained readily throughout the whole domain, including the often troublesome shoreline region. These results suggest that the tendency of wave-dominated muddy shores to export sediment and erode over long time scales is not due to wave forcing at infragravity frequencies, since the sediment transport under such waves is localized close to the shoreline, and is principally directed shorewards. They also provide easily reproduced test cases against which to validate the numerical methods used in more detailed studies of coastal sediment transport and morphodynamics.

Development of a regional neural network for coastal water level predictions

This paper presents the development of a Regional Neural Network for Water Level (RNN_WL) predictions, with an application to the coastal inlets along the South Shore of Long Island, New York. Long-term water level data at coastal inlets are important for studying coastal hydrodynamics sediment transport. However, it is quite common that long-term water level observations may be not available, due to the high cost of field data monitoring. Fortunately, the US National Oceanographic and Atmospheric Administration (NOAA) has a national network of water level monitoring stations distributed in regional scale that has been operating for several decades. Therefore, it is valuable and cost effective for a coastal engineering study to establish the relationship between water levels at a local station and a NOAA station in the region. Due to the changes of phase and amplitude of water levels over the regional coastal line, it is often difficult to obtain good linear regression relationship between water levels from two different stations. Using neural network offers an effective approach to correlate the non-linear input and output of water levels by recognizing the historic patterns between them. In this study, the RNN_WL model was developed to enable coastal engineers to predict long-term water levels in a coastal inlet, based on the input of data in a remote NOAA station in the region. The RNN_WL model was developed using a feed-forwards, back-propagation neural network structure with an optimized training algorithm. The RNN_WL model can be trained and verified using two independent data sets of hourly water levels. The RNN_WL model was tested in an application to Long Island South Shore. Located about 60–100 km away from the inlets there are two permanent long-term water level stations, which have been operated by NOAA since the1940s. The neural network model was trained using hourly data over a one-month period and validated for another one-month period. The model was then tested over year-long periods. Results indicate that, despite significant changes in the amplitudes and phases of the water levels over the regional study area, the RNN_WL model provides very good long-term predictions of both tidal and non-tidal water levels at the regional coastal inlets. In order to examine the effects of distance on the RNN_WL model performance, the model was also tested using water levels from other remote NOAA stations located at longer distances, which range from 234 km to 591 km away from the local station at the inlets. The satisfactory results indicate that the RNN_WL model is able to supplement long-term historical water level data at the coastal inlets based on the available data at remote NOAA stations in the coastal region.