APPENDIX C

ABSTRACTS OF PRESENTATIONS

[listed in order of presentation at the Workshop]

History and Physical Environment of the Petitcodiac
Denis Haché
Introduction and background session - presentation #1

Note: Elevation measurements are in meters above Geodetic Survey of Canada datum.

In the 1960s three sites were assessed for the causeway location.

1. in front of Place Champlain, a few hundred metres downstream of Halls Creek, five km downstream of the existing causeway,
2. about 1.5 km upstream of Gunningsville bridge (about 500 m downstream of causeway), and
3. about 400 m upstream of existing causeway.

Site no. 1 was not chosen because of the need to provide a navigation channel from the ocean to the existing downtown wharf. The assumption was made that sites 2 or 3 would not result in any significant mud deposition and therefore there would be no impact on navigation. The causeway location was chosen between sites 2 and 3.

Causeway construction started in February 1966 with the following design parameters:

• 5 sliding gates 9.1 m wide by 6.3 m high.
• Gate floor elevations at -1.5 m and ceiling at + 4.8 m.
• Headpond normal operating level at +3.5 m elevation.
• Small high tide range of 4.5 m, average range of 6.2 m., and large range of 7.6 m.

Mud accumulation started immediately upon construction and the extent of mud accumulation was always to the limit of construction. In the fall of 1967, mud accumulation upstream and downstream of the gate was in the order of 6 m deep. The top of the mud was about 1.5 m to 2 m higher than the design headpond normal operating level.

From February 1968 to May 1968, the five causeway gates were left continuously opened to cut through accumulated mud.

At the proposed causeway site no.1, about 5 km downstream of the causeway, the river was narrowed from 680 m in 1961 to 285 m in 1968 to 90 m in 2001. During that time, the bottom of the channel (thalweg) rose from -3.6 m (1961) to +3 m elevation in November 2001.

The causeway gates were designed to leak both ways and that leakage was more than the average monthly flow of the river for most months. The owner decided to operate the headpond from about +6.1 (following freshet) to +2 m elevation (after extended low flow). This resulted in significant salt-water intrusion and mud deposition immediately upstream of causeway.

The causeway gates and the fishway were modified in 1981 to allow for a normal operating level of 6 m. The headpond was dewatered on many occasions for significant periods of time to install infrastructure in the headpond or to help with fish passage. The headpond was operated much lower than the normal operating level on a seasonal basis and on different occasions to protect headpond shoreline from erosion.

In 1988, salt water was allowed in the headpond by keeping anywhere from one to five gates continuously opened for six weeks in the spring and one gate continuously opened for 6 weeks in the fall. In April 1988, five gates were opened during spring freshet and spring tide resulting in a maximum headpond water elevation of about 5.3 m (lower than normal operating level).

Analysis of estuary cross section surveys taken in the fall of 1991 and spring of 1992 show a downstream mud accumulation of about 120 x 106 m3 since construction. Headpond volume below elevation 6.1 m was reduced by about 12 x 106 m3. (from 23 x 106 m3 to 11 x 106 m3).

Monitoring conducted between 1998 to present show that the mud accumulation in the first 2 km downstream of causeway varies from -4 m following spring freshet to +5 m following extended low flow periods usually end of summer.

The survey of November 2001 showed a significant mud accumulation of about 16 x 106 m3 since 1991. Most of sediment deposition was downstream of Halls Creek. The biggest increase was in the bottom 10 km of Petitcodiac estuary located some 30 km downstream of causeway (about 25% increase in river bottom sediment since 1991).

Winter ice conditions (border ice buildups) have resulted in a reduction of the river width during the winter in the first 3 km below causeway from about 120 m down to 20 m wide.

Overview of March 1999 Modelling Feasibility Report
David Willis
Introduction and background session - presentation #2

A hydronumerical model of the Petitcodiac River/estuary would study the physical results of reopening the causeway at Moncton, New Brunswick, constructed in 1968. Such a model is on the edge of 1999 cohesive sediment modelling ability, because:

1. Length (55 km) and narrow width (50 m) of the present low water channel, dictate a fine grid spacing and large number of grid points;
2. Suspended sediment concentrations of the order of 10,000 mg/l result in rapid deposition rates;
3. A tidal bore on the flood implies current velocities of 3 to 6 m/s; and,
4. Erosion processes in the estuary appear to be 'bank erosion' of the steep face of the low water channel banks.

Causeway strategies to be modelled include, but are not limited to:

1. Existing causeway gate operation - the 'status quo' option;
2. Optimize gate operation;
3. Gate removal - fully open;
4. Breach-and-bridge causeway - optimize breach; and,
5. No causeway - pre-1968 conditions.

A questionnaire was sent to 18 numerical modellers, particularly to those in Canada who would be likely to participate in the modelling of the Petitcodiac River/estuary asking them to assess the current state of numerical modelling, with particular reference to the Petitcodiac. From the five completed returns, and comments by others, modelling was determined to be feasible. As a result, specifications and terms of reference were presented for the model study and its field data requirements.

Conclusions

1. Numerically modelling the hydrodynamics, sediment erosion and deposition and floating ice in the Petitcodiac River/estuary is at the 1999 leading edge of feasibility.
2. Modelling will cost less than $350,000 and take less than 20 months to complete.
3. The fieldwork required by the model will cost about $600,000, and should be undertaken between March and October to catch the freshet, ice and sediment flushing, open water, large tides, and autumn storms.
4. Overall project cost will be less than $1,000,000, with a duration of less than two years.

Recommendations

1. Government agencies, in Canada and abroad, may be the best source for modelling, because of the model development required, and the fact that government agencies are willing to absorb part of the development costs.
2. Similarly in Canada, only universities and government agencies currently have experience with the specialized field and laboratory tests required by these models. The fieldwork should be put in the hands of a consortium including: the Acadia Estuarine Research Centre, Geological Survey of Canada - Atlantic, and the National Water Research Institute.
3. As soon as the Petitcodiac mud has thawed in 1999, the Acadia Estuarine Research Centre should carry out a fluid mud susceptibility test on mud and water, regardless of timing of the rest of the project.

Hydrodynamic Modelling of Complex Flow Conditions
David Willis for Norm Crookshank
Hydrodynamics session - presentation #1

(no abstract available)

Vertical Dispersion Mechanisms of Salt and Sediments in Estuary Flows
Dr. Noboru Yonemitsu
Hydrodynamics session - presentation #2

There is a great need in water quality management for reliable numerical models that are capable of predicting the consequences of any changes in estuarine systems. Indeed, models capable of predicting the transport of both dissolved and particulate constituents are the basis of the water quality assessment.

The reliability of model predictions depend on the accuracy with which the physical processes, which have to be simulated, are specified. Unfortunately, estuarine flows are unsteady, have non-uniform turbulent motions, and density differences generally play an important role. These factors contribute significantly to the transport of momentum, heat, and mass, and influence both velocity profiles and the distribution of dissolved and suspended matter. These turbulent processes are highly variable in both space and time, and are highly site-specific. Because of these factors, it is practically impossible to calculate the turbulent transport processes precisely. As such, it is vital to continue to investigate and identify dominant physical processes such as internal waves and hydrodynamic stability. Research of this type will not only allow us to achieve a greater understanding of turbulent transport mechanisms, but will ultimately allow us to make more accurate predictions. This will also allow us to verify current theories with experimental and field measurements.

In the presentation one of these physical processes, hydrodynamic stability, is provided as an example. An important issue is how internal waves (such as Kelvin-Helmholtz and Holmboe waves) influence the production of turbulent energy, and how this turbulent energy affects the vertical dispersion/transport of any constituents. Recently, much attention has been given to this phenomenon in estuaries, where measurements can be made using acoustic devices.

It can be concluded that the fine sediment transport in the Petitcodiac River/estuary may be a very complex process, and will depend on a large number of influencing factors. Insight into the physical processes involved is necessary before sediment transport can be reliably predicted. Unfortunately, quantification of these processes in the Petitcodiac River/estuary has not yet been made. Therefore, it is highly recommended that a basic monitoring program be initiated before any modelling studies are conducted.

High Resolution Finite-Element Modelling of Areas with Drying Inter-tidal Areas
Dr. Dave Greenberg
Hydrodynamics session - presentation #3

Numerical studies centered on the Bay of Fundy are being done with a specially developed three-dimensional finite element computer circulation model. This model divides the geographic domain up into triangles. The variable resolution feature of the finite-element model makes it well suited for covering a wide domain of influence with the required detail in areas of interest needed to resolve local characteristics. The model also has the capability of simulating wetting and drying of inter-tidal areas. Although the generic model code has the capability of including boundary forcing, internal water density and surface winds as current driving forces, the present applications concentrate on the tide, using boundary forcing by the principal diurnal lunar component, M2, which in this area gives a good representation of the mean tide.

Work in the Quoddy region of New Brunswick has centered on tidal properties, residual circulation, and particle tracking. The latter investigations included mixing, dispersion, and differences in particle trajectories depending on depth in the water column. Recent work focuses on diagnostics for specific research interests such as the minimum sustained current for continued oxygen supply and benthic processes. It is hoped that future funding will permit extension of the model to examine more tidal constituents, meteorological forcing and fresh water and salinity influences. Other studies looking at the full Bay of Fundy, have examined the nature of changing tidal constants with rising sea level and the inundation of low lying areas from the combination of high tides and storm surges.

Effluent Dispersion into Tidal Environments
Jochen Schroer
Hydrodynamics session - presentation #4

Opening the Petitcodiac dam/causeway and changing the estuary regime of the Petitcodiac will likely result in changes to the water quality of the estuary. Those changes may require predictive modelling, similar to the modelling of changes in hydraulics and sediment transport. Water quality is affected by the discharge of wastewater from the wastewater treatment plant. The Moncton Sewerage Commission operates a primary wastewater treatment plant at Point Park, Moncton. The effluent from 100,000 houses is passed through the plant and the treated effluent is discharged into the Petitcodiac near Outhouse Point.

Mixing of effluent into estuarine waters is conceptualized by near-field and far-field mixing theories. In the near-field, the energy resulting from momentum flux and buoyancy flux is dissipated. In the process, clean ambient water is entrained into the plume, resulting in dilution. In the far-field, the effluent plume is transported with the ambient current and mixing occurs due to dispersion and diffusion. Typically, a wastewater plume rises until the plume is neutrally buoyant and then spreads uniformly throughout the water column. Suspended solids settle out at some distance from the outfall.

To improve near and far-field mixing, diffusers can be installed. Diffusers divide the effluent flow into several small discharges, each leaving through a single pipe nozzle. To compute effluent dilution in the near and far-field, several mixing models are available. In the example shown the Cornell University mixing model was used for the near-field plume predictions. The RMA 2 and RMA 4 models were used for computing dispersion and dilution in the far-field. To optimize the design of the diffuser, a one-dimensional pipe network model was applied.

The example from the Miramichi River Estuary illustrates the process used for finding the optimum outfall configuration for the outfall at the new Chatham wastewater treatment plant. In order to meet water quality guidelines for the protection of a bathing beach on Middle Island, three outfall locations and several outfall configurations were tried. The resulting concept involves installing a diffuser at ten-metre water depth near the old sewage treatment plant. The diffuser was optimized to provide high discharge velocities and relatively small discharge quantities near the shore.

The principles of effluent dilution are applicable to both the Miramichi and the Petitcodiac estuaries, even though the two estuaries are very different in terms of tidal regime and wastewater discharges. Slide 18 summarizes the major features of the two treatment plants and the two estuaries.

Overview of Candidate Modelling Approaches for the Petitcodiac River/estuary
Andrew Driscoll
Hydrodynamics session - presentation #5

The purpose of this presentation is to discuss in general terms the types of numerical models which may be applicable to studying the complex environmental issues in the Petitcodiac River / estuary. For illustration, several uncalibrated models have been established for the Bay of Fundy / Petitcodiac area.

Model Dimensions

The complexity and computer demands of a model system increase with the number of resolved dimensions. Most basic is the 1D (cross-section network) model. While offering a simplistic description of geometry and physics, 1D models are computationally efficient and thus capable of long-term simulations.

At the opposite end of the spectrum are 3D models, which offer a detailed description of the flow and processes, but are often limited in spatial and temporal coverage due to high CPU requirements.

Types of Model Grids

Numerous types of horizontal grid descriptions are available. The "classical" rectilinear grid can be established with minimal effort, but offers no flexibility in tailoring grid resolution where needed. This can be addressed in large part by dynamically nesting multiple grids.

Curvilinear grids offer a greater ability to focus resolution, and are well suited for modelling sinuous channels and meanders. They also offer an improved description of flow close to land.

An unstructured (finite element) grid offers the greatest flexibility in tailoring resolution.

Implementation of Relevant Mechanisms

Of central interest to the Petitcodiac study are several issues that are unrelated to the dimensions and grid type in the model. Among these are the implementation of mechanisms such as flooding and drying, supercritical flow, and hydraulic structures.

Flooding and drying can be addressed either by altering the characteristics of dried cells (marsh porosity, slot technique) or by adding / removing cells in the model domain as water levels change (point removal).

Supercritical flow can be handled by suppressing convective terms in order to avoid model instabilities at high Froude numbers.

Structures cannot in general be modelled directly, and are normally parameterized as a source-sink point on either side of the structure, where the flow is determined by an empirical or semi-empirical relationship.

Summary

The working Term of Reference for the project as distributed for discussion at the workshop is ambitious and far-reaching. A hybrid application of multiple tools (1D, 2D and 3D) will be required to address all issues.

Cohesive Sediment Transport in Estuaries
Dr. Krish Krishnappan
Sediments session - presentation #1

An overview of modelling approaches for cohesive sediment transport in estuaries is given. It includes the two modelling approaches that are currently available in the literature. The first approach considers the flow region into separate layers such as suspension layer, fluid mud layer, partly consolidated bed layer and fully consolidated bed layer, and solve different governing equations for each layer. The sediment fluxes that transfer sediment between layers are to be specified to these models by making use of empirical relationships that are established for site-specific sediments. The second approach, which is still considered to be in the research stage, considers the flow region as a single layer and solves a two phase flow equation to define the concentration profile in the estuary from the free surface to the stationary bed. This modelling approach requires information on floc properties such as floc size, density, settling velocity, rates of aggregation and disaggregation and effective stress etc. The above information has to be obtained also by direct measurement for site-specific sediments.

Experimental facilities of the National Water Research Institute at Burlington, Ontario, capable of measuring the sediment fluxes between different layers and the floc properties of cohesive sediment are described. These include the rotating circular flume, in-situ particle size analyzer, and the in-situ erosion flume. The rotating circular flume is suited for measuring the erosion and deposition rates for site-specific sediment, rate of consolidation of the fine sediment deposit, floc size, density and settling velocity of the flocculating sediment. The in-situ particle size analyzer is useful for measuring the size distribution of the sediment in suspension and the in-situ erosion flume is suitable for measuring the critical shear stresses for erosion and the erosion rates of the sediment from the bed.

Sediment Behavior in Turbid Environments
Tim Milligan
Sediments session - presentation #2

Extremely rapid deposition of fine-grained sediment has been observed near an artificially created channel between the Edisto and Ashepoo Rivers in South Carolina. Accumulation rates on the order of 0.05 m/month formed a region of fluid mud known as the 'Mud Reach' that was present during a study carried out in May 1998. Image analysis of photographs of the suspended sediment showed that the water column was populated by flocs with a median diameter on the order of 0.3 mm which tended to remain in suspension over the tidal cycle, and flocs on the order of 0.8 mm with calculated settling velocities on the order of 1 cm/s which remained in the near bottom region except at maximum values of shear velocity, u*. Disaggregated inorganic grain size analysis of the suspension and samples from a core collected in the Mud Reach indicated that the accumulation of material on the bottom was the result of floc settling and that in spite of high u* values, little sorting of this material was occurring. These results are consistent with laboratory investigations of fine grained sediment settling which show that maximum deposition of flocculated sediment occurs under high concentration, low turbulence conditions. In the Petitcodiac, extremely rapid settling can be expected to occur as a result of the high sediment load and low turbulence that occurs at slack water. These conditions will promote the creation of fluid mud, further decreasing bottom stress. A feed back mechanism is thus created which can result in accelerated trapping of fine sediment in the Petitcodiac Estuary as shown by the rapid and continuing accretion of the mud flats at Moncton.

Time and Density-Dependent Properties of Near-Bottom Fine Sediment Suspensions
Dr. Richard Faas
Sediments session - presentation #3

Of particular importance to tidal environments, these properties are quite sensitive to tidal asymmetries and are found in suspensions ranging in density from 1.02 to 1.20 mg/m3 (10 to 480 g/l). The properties of yield stress, shear thickening and thinning, thixotropy, and "apparent" viscosity all appear to be related to settling time and density development during periods of slack water. A short slack water period will produce a newly settled muddy deposit of low density, little yield stress, and low thixotropy which will be easily resuspended due to shear thinning flow and be transported on the flooding or ebbing current. Conversely, longer slack water times will allow the same sediment to consolidate to a greater density, acquire a significant yield stress, resist resuspension by shear thinning flow, and remain close to the bottom through a significant portion of the tidal cycle. Modellers should become aware of these peculiarities and the situations within which they occur in order to predict the accumulation and buildup of particles carrying pollutants and/or creating shoals and flats which may necessitate frequent dredging.

The Sediment Budget of Chignecto Bay
Dr. Carl L. Amos
Sediments session - presentation #4

The Chignecto Bay system (inclusive of the Petitcodiac River/estuary) is dominated by silts and clays that move principally in suspension. The origin of this fine material is the floor of Chignecto Bay (7.3 x 106 m3/a) largely due to bed erosion caused by the amplification of the tides over the last 4,000 years (Amos et al., 1991). This massive addition of material to the water column is manifested in the exponentially-decreasing mean concentration (C) in suspension down-estuary and has lead to the formation of extensive salt marshes and mudflats at the head of the Bay: most of these have been reclaimed, thus removing the natural sink of sediment. The distribution of material in suspension is controlled by both the tidal flows and by storm wave activity. The tidal asymmetry (as well as the difference in the rate of change in current speed between high and low water) balances the headward transport of sediment with the seawards dispersion due to eddy diffusion, secondary circulation (Tee and Amos, 1991), and freshwater discharge (Amos and Tee, 1989). Wave activity disrupts this balance causing changes to C which have several effects: firstly, changes in C have a direct effect on the threshold for deposition and on the mass deposition rate, which in turn alters the residual movement of material; secondly, changes in gradient in C have a direct effect on the mass balance down-estuary: if the exponential gradient becomes steeper, sediment is flushed seawards; if the gradient is becomes more shallow, then sediment is moved landwards. Ice formation and break-up has a similar, if not larger, effect on this balance, which has a massive influence on seawards diffusion. In this presentation, I will review the budget of sediments in Chignecto Bay over seasonal, annual, and millennial time-scales and attempt to relate this to the volumes locked within the Petitcodiac mudflat and discuss which factors are important in controlling this budget.

References

Amos, C.L. and Tee, K.T. 1989. Suspended sediment transport processes in Cumberland Basin, Bay of Fundy. Journal of Geophysical Research 94(C10): 14,407-14,417.

Tee, K.T. and Amos, C.L. 1991. Tidal and buoyancy-driven currents in Chignecto Bay, Bay of Fundy. Journal of Geophysical research 96(C8): 15,197-15,216.

Amos, C.L., Tee, K.T. and Zaitlin, B.A. 1991. The post-glacial evolution of Chignecto Bay, Bay of Fundy, and its modern environment of deposition. In D.G. Smith. G.E. Reinson, B.A. Zaitlin and R.A. Rahmani (eds) Clastic Tidal Sedimentolgy. Publ. Canadian Society of Petroleum Geologists 16: 59-90.

European Cohesive Sediment Transport Models
Dr. Erik Toorman
Case Studies Session - European Experiences - presentation #1

Cohesive sediment transport models have been developed in Europe since the 1980s. For engineering purposes they were originally 2-dimensional depth-averaged horizontal (2DH). The evolution of the complexity of the models is directly linked to the progress in hardware capabilities and numerical solvers. Since the mid 1990's extensive efforts have been made to develop 3D hydrostatic models. This allows the description of vertical processes, including turbulent exchange, stratification, and flocculation. Recently there is a trend to remove the restriction of hydrostatics, in particular for stratified flow conditions. A continuous-phase approach is used, which is justifiable as long as concentrations remain smaller than 1% by volume, which usually is the case, except near the bottom, where concentrated suspensions or fluid mud layers can occur.

The major European software packages in use are Delft-3D (Delft Hydraulics, the Netherlands), MIKE3 (Danish Hydraulic Institute, Denmark) and Telemac (Laboratoire Nationale d'Hydraulique et Environnement, France, in collaboration with HR Wallingford, U.K., and SOGREAH, France). They comprise various modules, e.g. for hydrodynamics, waves, sediment transport and water quality. These models are very robust at the cost of accuracy due to the introduction of artificial diffusion to achieve numerically stable calculations.

Efforts are continuously undertaken to improve the model performances. Collaboration between the major commercial, governmental and university research units has been funded in part by the European Commission under their Marine Science and Technology (MAST) programme. The projects G6/8M Coastal Morphodynamics (in part) , 3D-Models and COSINUS have been focusing on Cohesive Sediment Transport (CST) models.

In the COSINUS project several weaknesses in the present models on the parameterisation of various processes have been investigated. Important improvements on the modelling of sediment-turbulence interaction and the occurrence of concentrated benthic suspensions have been achieved. The latter has resulted into a new module for fluid layers in Delft-3D. No significant progress was made regarding the modelling of settling velocity and erosion. Flocculation can be described well by a structural kinetics model, but its calibration is cumbersome. The modelling of erosion remains the weakest part of the model, and probably always will due to the large uncertainty on the corresponding model parameters. Further progress is also to be expected in the modelling of near-bottom effects where high concentrations and laminarisation occur, which are important for improving the estimation of sediment exchange with the bottom.

Within the framework of the G8M and COSINUS project a few model intercomparison exercises have been cared out. The first one, where laboratory flume tests were simulated, illustrated the difficulty of erosion modelling, which could be attributed to a great extend to the lack of knowledge on the bed structure and strength, despite vane strength and density profile data. The COSINUS test cases were academic without experimental reference data. The various models deviate more and more from each other with increasing sediment loading. Differences can be explained, at least partly, by artificial effects generated by the numerical schemes and by the procedure for modelling the near bed layer.

Time and length scales play an important role in the decision-making regarding the modelling approach. The sediment transport models, as discussed above, can be applied at an acceptable computational cost to problems of time scales up to typically a spring-neap cycle. Particularly long-term predictions (i.e. years or decades) require many simplifications and alternative approaches. There is no general consensus on how to do this. Two recent case studies are briefly discussed as an illustration.

In the Netherlands a feasibility study is being conducted on the possible construction of a new international airport on an artificial island to be built in front of the coast. This so-called Flyland project consists of various studies, including the impact of the island on hydrodynamics, morphology, sediment and nutrient transport along the Dutch coast, with special concern for the environmental impact on the Wadden Sea, a large coastal nature reserve of international importance, extending into Denmark. Hydrodynamics is computed with Delft-3D for one spring-neap cycle and subsequently repeated to compute transport processes using a 4-layer 3D grid. This implies that many coupling effects cannot be considered. Corrections for seasonal variation in several parameters and fluxes are accounted for. The uncertainty on the results is large due to many factors.

The second case concerns Mont Saint-Michel in France, a small granite rock island with historic monastery and village, near the coast at the border of Normandy and Brittany, which is a UNESCO world heritage and important tourist attraction. Due to land reclamation in the past, the closed causeway to the island and gate complex on the Couesnon River, the island tends to lose its marine character due to rapid accretion of the neighbouring salt marshes. An investigation has been carried out to modify the hydraulic structures in order to restore and preserve the original character. The time frame here was 50 years. A totally different approach has been adopted. The study has been carried out with a physical scale model, in combination with a numerical model (Telemac). The latter was used to determine boundary conditions for the physical model and to simulate various short-term conditions, such as the flow field and turbidity distribution in the final situation. The physical model has been calibrated with the use of historical data, going back 50 years, and results from the numerical model.

The interpretation of the Cohesive Sediment Transport (CST) model results has to be done with great care, being aware of the limitations of the model. The uncertainties on the results are high due to large physical and numerical model uncertainties, data uncertainties and model parameter uncertainties, which can be attributed to the complex nature of natural sediment transport, for which the large variability in space and time cannot be accounted for in detail. Furthermore, time and budget restrictions put limitations on measurement data and CPU. Nevertheless, CST models have proven to predict the right trends at least qualitatively.

References

Toorman, E.A. (2001). Cohesive sediment transport modelling: European perspective. In: Coastal and Estuarine Fine Sediment Processes (W.H. McAnally & A.J. Mehta, eds.; Proc. INTERCOH'98, Seoul, May 1998), pp.1-18, Elsevier Science, Amsterdam.

Internet Resources
http://www.bwk.kuleuven.ac.be/bwk/cosinus/cosinus.html
http://www.dhisoftware.com/mike3/
http://www.delft3d.com/soft/d3d/index.html
http://www.telemac-system.com/gb/default.html

Examples of Estuary Models and Projects
Dr. Ole Petersen
Case Studies Session - European Experiences - presentation #2

The environment in tidal estuaries is generally the result of a dynamic balance between tides and surges, riverine runoff and the sea salinity, local meteorological conditions, the topography and sediment fluxes and deposits. It is further characteristic, that salt and freshwater meets within the estuary, commonly creating a relatively complex stratified situation. On longer timescales, the topography itself may adjust to the prevailing regime of waves, currents, and sediment fluxes, thus contributing to the delicate balance that establishes the estuary. Further, the interest and usage of estuaries by society are diverse and often conflicting. Quantitative understanding and modelling of such a dynamic environment can therefore be a challenging but often necessary task.

Sediments in estuaries may range from coarse riverine sediments in the upstream part, which are moved by occasional events, to masses of fine marine sediments, which change position with the tidal cycle. Therefore, the processes affecting the estuary occur on a wide range of time- and length scales. This should be recognized in the modelling such that focused models address separate problems, in order to utilize the resources at hand most efficiently.

Today, a number of models exist which are based on assumptions ranging from 1D steady-state models, that can describe the long term development of complete river systems, to detailed 3D models describing accurately the flow and transport around complicated structures. The models usually comprise hydrodynamics, some baroclinic effects, water quality and sediment transport. Some of these models are summarized in Table 1, based on DHI's model system.

Table 1. Examples of Models and Possible Application Types.
The purpose of modelling can be many, for example:

• Provide predictions of future situations, where data does not exist;
• Provide consistent comparisons between scenarios; and/or
• Establish a consistent understanding of a base of observations and assist in defining monitoring programs.

The ability of the model to fulfill this depends, among other things, on the modelling uncertainty, which can be related to the model equations and solution, to uncertainty in the observations used for calibration and to inaccuracies or relevance to the specific situation of the data used for calibration.

Tamar Estuary Study

The Tamar Estuary, UK, is a 40 km long macro-tidal estuary on the south coast of England, which contains a distinct turbidity maximum. As part of the European Union (EU) funded COSINUS research project an extensive study, comprising field observations and modelling, was made to validate 3D models of cohesive sediment transport. One significant development was identification and modelling of high concentrated mud layers, formed at calm periods of the tidal cycle.

Loire Estuary Study

This is a macrotidal estuary, approximately 100 km long with tidal range of up to 5 m. A distinct turbidity maximum exist, containing an estimated 1.5 x 106 m3 of suspended sediment with peak concentrations in the order of 1 g/l. Associated with the suspended sediment are significantly increased oxygen demands.

The purpose of the study has been to establish decision support for the development of restoration strategies for the estuary, comprising among other things, navigational issues, water quality, effects on the natural and urban environments, fish habitats and agricultural uses.

The approach has been the development of a one-dimensional network model, based on MIKE 11, for flow, sediment transport and morphology and water quality. The model was intended, on one hand, to resolve the tidal cycle and the development of a significant tidal asymmetry up through the estuary and on the other hand, to predict the effects after up to 40 years. The 1D model was assisted by several two-dimensional models, addressing local details as bank erosion or effects of alternative dredging schemes.

Mühlenberger Loch, Lower Elbe

The Elbe is a tidal estuary, with a tidal range up to 3 m and some 50 km long from Cuxhaven to the weir at Geesthact. The estuary is the approach to Port of Hamburg and contains considerable amounts of suspended fine sediments.

The purpose of the study was to estimate possible effects following re-opening of a river branch, Alte Süderelbe, entering the environmentally sensitive area of Mühlenberger Loch. The opening was planned to be through two gated structures, thus the study also addressed development of operation strategies for the gates, with respect to flood protection and containment of contaminated sediments.

For the study, a two-dimensional local model of the river branch and the Mühlenberger Loch was developed, describing spring-neap cycles of sediment exchange through the gates, morphology and used for optimization of the gate operations. A 1D model was used to assess long-term evolution of water quality and siltation.

Tidal Asymmetries Mixing Processes and Particle Trapping in Stratified Estuaries
Dr. Bob Chant
Special Tidal Processes Session - presentation #1

Observations in the Navesink River estuary in northern New Jersey demonstrate that buoyancy augments the particle trapping tendencies of flood-dominated systems, because these estuaries heighten tidal period asymmetries in stratification. During the long and slow ebb, which typifies flood-dominated systems, a positive feedback between tidal straining and weak vertical mixing stratifies the estuary. In contrast, during flood, turbulence generated by the stronger tidal currents augments overstraining of the density field and the water column becomes well mixed. The tidal period asymmetries in stratification have profound effects on the vertical structure and transport of suspended matter. During ebb weak vertical mixing allows suspended material to settle downward. In contrast, strong turbulence during flood mixes suspended matter into the water column where it is transported up estuary. Furthermore observations reveal that resuspension events are marked by multiple turbidity spikes, suggestive of multiple limited layers of erodible material. The transport of the turbid waters is consistent with horizontal advection modified by horizontal dispersion. Relatively low levels of turbidity also mark periods of enhanced stratification during the ebb, consistent with more complete settling of suspended material following times of high river discharge.

The interplay between buoyancy and tidal asymmetries are further elucidated with a one-dimensional numerical model featuring a turbulent closure scheme and a passively settling tracer. Model results are generally consistent with the field observations, both emphasizing the robust particle trapping tendencies of a stratified flood-dominated estuary. We speculate that enhanced particle trapping following times of high river discharge may have important biological consequences.

Shipboard and moored data were also presented from the Hudson River Estuary to demonstrate the movement of sediment on both tidal and seasonal time scales. As in the Navesink trapping of sediment is observed over regions of enhanced stratification. However, as the stratification varied with river discharge sediments delivered down stream by the April spring freshet begins moving upstream as river flow reduces and stratification moves upstream. A region of enhanced trapping occurs in the vicinity of a channel contraction where an increase in the horizontal salinity gradient traps suspended material in horizontally converging flows.

Operational Prediction of Storm Surges and Flooding
Dr. Hal Ritchie
Special Tidal Processes Session - presentation #2

For about twenty years the Meteorological Service of Canada (MSC), through the Maritimes Weather Centre (MWC) in its Atlantic (MSC-A) region, has had the responsibility of advising the public when the combination of a high tide and an intense storm produces a risk of higher than normal water levels that may flood some coastal areas. The tides are easily predicted due to their periodic nature, but in the past no numerical model guidance has been available to forecast the additional "storm surge" component of sea level that is caused by the weather systems.

A team of oceanographers at Dalhousie University (Dal) has recently developed a storm surge forecast system for the east coast of Canada. The system is based on a two-dimensional barotropic model that is forced by atmospheric surface pressure and winds. As a collaboration within the Atlantic Environmental Prediction Research Initiative (AEPRI) this system has been transferred from Dal to MWC where it has been implemented as Canada's first operational storm surge prediction system. The operational system is forced by three-hourly atmospheric surface pressure and wind provided by the MSC operational regional weather forecast model and produces storm surge predictions up to 48 hours in advance. The transfer included a preliminary evaluation (as a "hindcast" study) for the period September 1996 to February 1997 and another evaluation for the period October 18 to December 6 1999 when the system first started running at MWC. The system soon demonstrated its potential by producing very accurate forecasts for the January 21, 2000 severe storm surge event that caused considerable coastal flooding and damage in Prince Edward Island (PEI) and eastern New Brunswick. Storm surge prediction specialists in MWC have since produced a storm surge alert system that automatically warns the forecasters when predicted coastal total water levels (surge plus tide) exceed predetermined site-specific thresholds that have been chosen to reflect varying degrees of expected flood damage.

The system was "put to the test" by an event predicted for November 7, 2001 when stage 1 thresholds were forecast to be exceeded in some locations in northern PEI. MWC forecasters were able to successfully identify which areas would be flooded and alert the Emergency Measures Office (EMO) in PEI, who notified the affected areas and monitored the situation as it unfolded. As a result of this inter-agency cooperation, some property damage was avoided.

This storm surge prediction system has also been used to examine the increasing probability of flooding due to storm surges in a changing climate. The presentation will include some results from the study by Thompson et al., 2001.

References

2001 K. Thompson, H. Ritchie, N. Bernier, J. Bobanovic, S. Desjardins, P. Pellerin, W. Blanchard, B. Smith and G. Parkes: "Modelling Storm Surges and Flooding Risk at Charlottetown", Appendix 2, Report on Climate Change Action Fund project CCAF A041 - Coastal Impacts of Climate Change and Sea-Level Rise on Prince Edward Island.

Winter processes in estuaries
Dr. Brian Morse
Special Tidal Processes Session - presentation #3

Very little is known about ice processes in estuaries. To respond to this need, at the suggestion of B. Burrell (New Brunswick Department of the Environment and Local Government) in 1998, the Canadian Committee on River Ice Processes and the Environment (CRIPE) formed a study group under B. Morse. The study revealed that there are three known types of winter estuaries: (1) James Bay (tides in the order of 50 cm); (2) Estuaries emptying into the St. Lawrence River (tidal range 2.5-3.5 m) ; (3) Bay of Fundy estuaries (tidal range 11-14 m). For type 1, the most important consideration is the increase in the extent of the fresh water plume under ice conditions otherwise ice processes are the same as those found in normal rivers. For type 2, such as the Portneuf estuary near Forrestville, ice processes are completely dominated by the tides: there is cracking, stacking, ridging, flooding and jamming - all related to the tides. In turn the ice significantly affects water levels at low tide, dramatically changes the vertical distribution of the current and diminishes the water exchange over a tidal cycle. Ice-water-salinity-temperature dynamics are very complicated.

The Petitcodiac is a type-3 estuary and has been documented in a 1998 report by R. Hudson and V. Partridge at the 2001 CRIPE Ottawa Workshop and which also described some biological processes. However, by far the most authoritative work is presented by Desplanques and Bray (Canadian Journal of Civil Engineering, 1986): They describe 5 ice zones: (a) upstream of the causeway, there is sheet ice; (b) from the causeway to Dover, the channel is dramatically reduced in area (easily 50%) by ice cake depositions forming shear vertical walls; (c) from Dover to Hopewell Cape, there is ice formation and ice cake transport towards zone 2; (d) there is an ice production zone in the shallow delta having salt concentrations less than 27.4 parts per trillion - this ice is transported up the estuary on the flood tide; (e) finally there is the deeper coastal area that is characterized by very little ice formation.

During the winter, ice in the Petitcodiac protects the mud from erosion. It also shelters the biota in the mud from freezing temperatures. During the melt period, ice cakes can locally transport significant pieces of mud (along with its contents). Ice accumulated in the channel substantially reduces the hydrodynamics particularly in zone 2 (just downstream of the causeway), modifies currents and may increase salt content.

A conceptual and/or empirical thermodynamic-hydraulic model of ice formation (particularly in zone 2) is recommended to provide boundary conditions for the hydrodynamic and sediment models during the winter months. During the spring, a thawing/sediment slumping model identifying sources and sinks should be considered to feed into the sediment models (noting that the timing of the melt in relation to the freshet is very important in predicting whether there will be erosion of banks or erosion of the bed). Finally, it may be important to have a simple thermodynamic model to account for the Moncton sewage outfall (roughly estimated at 100 MW), changes in temperature and salinity so that the correct values can be used to estimate flocculation dynamics (required in the sediment transport model).

Of course field measurements are required to support these models: In addition to talking to knowledgeable people (data mining), field techniques should rely on video taping and measurements of conductivity-temperature-depth profiles using probes and currents using Acoustic Doppler Current Profilers (ADCP) and video tracking.

Tides and Water Levels in the Upper Reaches of Fundy
Charlie O'Reilly
Data Collection Session - presentation #1

The upper reaches of the Bay of Fundy experience the highest tides in the world, yet these phenomena are not nearly as well known as popularly believed. Primarily this is because they are not well measured at these locations due to the great difficulties imposed by the large tides. Most tidal knowledge comes from very short observational data sets, typically lasting from several days to perhaps a few months. The only permanent tide gauge in Canadian Fundy waters is operational in Saint John, New Brunswick. Data from this site indicates that the large tides are getting larger still, over and above sea level rise. At the heads of the Bay, there is a great paucity of data to yield any light on long-term behavior. This presentation illustrates what is known, what is not known, and what measures are required to bridge this gap. At present, extreme high water thresholds can be only estimated to within a half-meter at best (a full meter in some locations). Ancillary issues such as the tidal bore phenomena are briefly discussed, as are natural hazards issues such as flooding due to storm surges and accelerated sea level rise. Special emphasis is given to the Petitcodiac estuary, which is an extremely dynamic situation where the large tides greatly affect the river and the river greatly affects the tides.

Hydrographic Surveys of the Estuary and River Channels
Jean-Claude Vautour
Data Collection Session - presentation #2

A series of sections across the Petitcodiac River/estuary have been taken since 1979 mainly by the New Brunswick Department of Transportation (NBDOT). In 2001, the federal department of Public Works & Government Services Canada (PWGSC) carried out for the federal and provincial authorities the same survey scheme. A total of 33 sections span the 35 km of river channel, from the Petitcodiac causeway downstream to the mouth. The data for all 33 sections were only collected in 1991 and 2001. All other surveys only covers a segment of the river and in most cases are in the upper portion of the estuary.

The Petitcodiac estuary has very large tidal fluctuations, strong currents, and very high sediment concentrations in the water column. In the upper section of the estuary, the muddy river bottom almost completely dries out leaving only 2 to 3 hours of survey time per day. Also, crane must be used to launch survey boats, although a boat launch from the beach is possible at high tide at Belliveau Village wharf when working in the lower section of the estuary. All these factors make data collection hazardous and challenging. Due to the high concentration of sediments in the water column, only 33 khz echo sounders can reasonably track the bottom.

All section data has been referenced to the Geodetic Survey of Canada 1st order control network (CGVD28). A dense network of monuments in Moncton, Riverview and along Route #114 to Hillsborough was used for the surveys by the NBDOT and PWGSC. Prior to 2001, tidal elevation was collected by determining directly the elevation of the survey vessel by range-bearing method. In 2001, a real-time kinematic global positioning system (RTK GPS) was utilized. Vertical accuracies of approximately 0.25 m prior to 2001 and 0.20 m in 2001 should be expected for the cross-sections. The sub-bottom penetration of the low frequency echo sounder is a factor that may add to the uncertainty of detecting the true bottom surface.

The horizontal reference datum is based on the Canadian Spatial Reference System North American Datum 83 (CSRS NAD83). The map projection used is the Universal Transverse Mercator (UTM) zone 20. A large number of provincial monuments along both sides of the river were used for the surveys. The horizontal positioning for the work carried out prior to 1997 used range-bearing and bearing-bearing method where the echo sounder paper trace was marked and reduced manually. Horizontal accuracy of about 1 to 2 m is to be expected using these methods. For the survey of 1997 and the following years, all the positioning was done by RTK GPS. The data was logged electronically and HYPACK was used to post-process the data. Horizontal accuracy achieved would be in the order of 0.05 m. The users of the section data should also be aware that the surveyed lines were generally zigzagging within 2 m of the theoretical line and that the boat could have been offline by as much as 5 m in some instances.

The data collection on the Petitcodiac River/estuary is a challenge due to the many environmental constraints where instrumentation and personnel are pushed to their limits. Good planning and experienced staff are essential to succeed on the unforgiving Petitcodiac River.

High Resolution Remote Sensing Mapping of Estuarine Environments
Tim Webster
Data Collection Session - presentation #3

This presentation was intended to introduce remote sensing technology that may be applicable for high-resolution mapping in the estuarine environment. Both satellite and airborne systems were presented involving both active sensors such as synthetic aperture radar (SAR) on board; RADARSAT-1 and laser altimetry from LIDAR, and passive sensors such as Landsat and the Compact Airborne Spectrographic Imager (CASI). High-resolution satellite imagery from Ikonos and QuickBird were also shown for estuarine environments. Some of the areas that remote sensing could aid in the mapping requirements for the Petitcodiac hydrodynamic modelling project were presented including: channel location, landcover (salt marsh, mudflat, etc.), temporal change detection (i.e. channel migration), and detailed topography acquired at low tide. Example datasets were shown from the research group's repository of imagery, many concentrating on the Nova Scotia side of the Bay of Fundy in the Minas and Annapolis Basin estuaries that have some similar characteristics as the Petitcodiac e.g. mudflats, salt marsh, etc. Images acquired at different stages of the tidal cycle were used to demonstrate how the land-water boundary could be extracted to give information on the intertidal geomorphology of the estuary. LIDAR acquired at low tide offers the best potential to produce detailed topographic information of the estuary. A wide range of prices and services exist in the industry for such a survey. The research group, from their recent coastal studies involving such technologies, has learned valuable lessons related to issues such as contract data specifications and suitable environmental conditions for data collection. A list of data providers and a price breakdown of the different technologies was presented at the conclusion of the presentation. Prices vary depending on the volume of data and the availability from an archive.

Acquiring Representative Physical Oceanographic Data in Muddy Rivers/Estuaries
Gary Bugden
Data Collection Session - presentation #4

Some examples of the nearshore studies undertaken by Coastal Ocean Sciences (COS) at the Bedford Institute of Oceanography (BIO) are examination of aquaculture impacts (northern Prince Edward Island), determination of the fate of industrial contaminants (Sydney, Nova Scotia) and general studies in support of Integrated Coastal Zone Management (Bras d'Or Lakes, Nova Scotia). Three types of instrumentation are typically used in these studies to make physical oceanographic measurements such as water currents and the vertical and horizontal distribution of water properties. These are Acoustic Doppler Current Profilers (ADCP's), Conductivity Temperature Depth probes (CTD) and a novel flow-through system that pumps water through a series of sensors onboard a small vessel. Because of shallow depths, high currents, elevated suspended particulate matter loads and mobile bottom sediments the Petitcodiac system presents unique challenges not encountered in other regions. Methods used for instrument deployments in other regions are inapplicable and in some cases it is not clear if the instruments themselves will function properly. Acoustic Doppler Current Profiler (ADCP) technology is particularly appropriate to the questions at hand as, in addition to currents, depth and some indication of the SPM load are also obtained from the instrument. Some suggestions for a physical oceanographic sampling program are discussed and the point is made that it is important to get out and measure something rather than sit around engaging in "unbridled speculation".

Recent Experiences in Modelling Cohesive Sediment Transport
Dr. Doug Scott
Case Studies Session - Other Experiences on Estuary Modelling - presentation #1

This presentation provided a summary of recent project experience in numerically simulating the erosion and deposition of cohesive sediments in both riverine and estuarine environments. Keys issues and observations with relevance to the Petitcodiac estuary were drawn. Two of the examples presented included the numerical modelling of the morphological change and fate of contaminated sediments in the Fox and Sheboygan Rivers in Wisconsin.

In addition, an example of a fast-track design/build project for navigational channel construction within a muddy estuary located in Malaysia was shown. Accurately simulating the fate of estuarine sediments is a critical aspect of the project, and the outline of an integrated modelling and field measurement approach was presented.

The potential variability of defining the cohesive sediment properties in rivers and estuaries was discussed. It is recommended that a range of sediment testing be conducted for the Petitcodiac investigation that involves a variety of laboratories and testing techniques. Accurately defining sediment boundary conditions is also a critical aspect of the modelling procedure, and can lead to much uncertainty in the modelling results.

The importance of developing an integrated team with expertise in numerical modelling, field measurement, and river/estuary geomorphology was stressed, as was the importance of conducting the modelling and field campaigns in repetitive fashion at increasing levels of sophistication. The Petitcodiac modelling should be initiated at a simplistic level and build in complexity as knowledge of the estuary is developed.

It was noted that numerical modelling will not eliminate the uncertainty associated with modifications to the Petitcodiac causeway, and that risks and impacts should be appropriately managed. This may mean that changes to the Petitcodiac estuary may need to be conducted in a controlled fashion with on-going monitoring, modelling and evaluation.

Recent Experience in River Flow and Erosion Modelling - Applicability to the Petitcodiac River
Dr. Michael Davies
Case Studies Session - Other Experiences on Estuary Modelling - presentation #2

(no abstract available)

...table of contents