SAER-5347

I I

EXON PRODUCTlON RESEARCH COMPANY

I I I 1 I Abu APi kkmd Causeway Extension Study

.. . . .. . . , , .~ : . ~

R. B. Gordon Y. K. Vyas

H. M. Shabib

I I I I lr I 1 I AER# 5347

Copyright©Saudi Aramco 2009. All rights reserved.

POSTOFFICEBOX 2189 HOUSTON, TEXAS77252-2189 B J.F. WOLFE MANAGER

April 25, 1988

Mr. A. A. Dafas Supervisor, Environmental Unit Process and Control Systems Department Aramco Box 53 DHAHRAN, 31311 Saudi Arabia

Dear Mr. Dafas:

The enclosed report, "Abu Ali Island Causeway Extension Study," completes the project authorized by Aramco in April 1987. The report provides our assessment of the impact of the proposed Abu Ali causeway extension on the current, salinity, and sedimentation patterns in the vicinity of Abu Ali Bay.

We have enjoyed working with Hassan Shabib on this project. Please let us know if we may be of any further assistance.

Very truly yours,

J. F. Wolfe RBG/rhm Enclosure File: 25292


DISTRIBUTION:

, ARAMCO

A. A. Dafas (20)

EXXON PRODUCTION RESEARCH COMPANY

Foundation and Environmental Analysis Section Library (2) EPR Library (2)


fi Offshore Division EPR. 14PS.88 April 1988

li a

Exxon Production Research Company

Abu Ali Island Causeway Extension Study

R. B. Gordon Y. K. Vyas

H. M. Shabib

FOR ARAMCO USE ONLY This report containa Aramco information; copies s h d d not begivsn to c0nlPsni.r other th.nAnmco without their prior wfitten consent.


TABLE OF CONTENTS

EXECUTIVE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

CONCLUSIONS AND RECOMMENDATIONS ................................. 2

1 . INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 . HYDRODYNAMIC MODEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 . 1 Description of the hydrodynamic model ...................... 4

2 . 2 Grid coverage .............................................. 4

2 . 3 Bathymetry ................................................. 5

2 . 4 Boundary tidal data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 . 5 Model calibration .......................................... 5

3 . SALINITY MODEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

3.1 Description of the transport-dispersion model . . . . . . . . . . . . . . 7

3 . 2 Grid coverage .............................................. 9

3 . 3 Model calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

4 . SEDIMENT TRANSPORT ANALYSIS PROCEDURE ........................... 10

4.1 Wave modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

4 . 2 Sediment transport analysis ................................ 1 2

5 . EFFECTS OF PROPOSED CAUSEWAY EXTENSION ON CURRENTS, SALINITY AND SEDIMENT TRANSPORT ................................. 14

5.1 Tidal circulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

5 . 2 Salinity distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 5

5 . 3 Sediment transport analysis results ........................ 16

5 . 3 . 1 Wave modeling results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

5 . 3 . 2 Sediment budget analysis results . . . . . . . . . . . . . . . . . 17

ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19


1 4 8. J a c 8 , 3 ' 1' I

4 8 J 1' d

APPENDICES

A. Representation of causeway culverts

B.

C.

D. Sediment transport computations

Inferred bathymetry from Landsat imagery

Methodology and data sources for calculating evaporation rates


LIST OF FIGURES

1.1

1 . 2

2 . 1 Study area map.

2 . 2

2 . 3

2 . 4 Bathymetry map on 100 m wave model grid; (a) existing causeway, (b)

Map showing existing causeway and proposed extension.

Flow chart of the modeling procedure.

Bathymetry map on 900 m current model grid.

Bathymetry map on 300 m current model grid.

existing causeway with proposed extension.

2 . 5 Location of Aug. 1987 tide and current meter stations.

2 . 6 Comparison of modeled and predicted tide height at Abu Ali Pier and Berri Platform.

2.7 Comparison of modeled and measured current speed at (a) current meter station A, and (b) current meter station B. Data from Aramco's August 1987 survey.

3 . 1 Schematic representation of the Everett twelve-point finite difference scheme.

3 . 2 Location of August 1987 salinity measurement stations.

3 . 3 Computed evaporation rate at Abu Safah GOSP for the year prior to Aramco's August 1987 survey. Dotted line is a cubic spline fit.

3 . 4 Predicted vs. measured salinities. Data from Aramco's August 1987 survey. Numbers on middle panel indicate stations (see Fig. 3 . 2 ) .

5.1 Contours of the one-year maximum tidal current speed; (a) existing causeway only, and (b) with causeway extension.

5 . 2 Positions for which current speed time history plots were prepared in relation to (a) existing causeway, and (b) existing causeway with proposed extension.

5 . 3 Time history plots of current speed at select locations for the one-year maximum spring/neap cycle; (a) existing causeway only, (b) existing causeway with proposed extension.

5 . 4 Salinity contours for the "worst case" scenario. Results from the 900 m grid; (a) existing causeway only, and (b) with causeway extension.

- iv-


5.5 . .

5.6

5.7

5.8

5.9

Same as Fig. 5.4 but for results from the 300 m grid; (a) existing causeway only, and (b) with causeway extension.

Comparison of the predicted change in bay salinity due to the causeway extension with changes due to the natural variability as determined from Aramco’s “bioaccumulation” data set.

Vector plot of wave height, amplitude and direction from refraction analysis for the one-year storm: (a) existing causeway only; (b) existing causeway plus extension.

Vector plot of wave height amplitude and direction from refraction analysis for the ten-year storm: (a) existing causeway only; (b) existing causeway plus extension.

Erosion (a) and deposition (b) under the one-year storm (existing causeway only).

5.10 Same as Fig. 5.9 except for existing causeway plus extension.

5.11 Net changes in bottom topography due to the causeway extension for the one-year storm.

5.12 Net changes in bottom topography due to the causeway extension for the ten-year storm.

C.l Ras Tanura evaporation rate (cm/month), 1960-1977; Aramco formulation.

C.2 Ras Tanura evaporation rate (cm/month), 1960-1977; Price formulation - air/sea temperature difference set to zero.

C.3 Ras Tanura evaporation rate (cm/month), 1960-1977; Price formulation - air/sea temperature difference based on climatological values for Ras Tanura.

-V-


LIST OF TABLES

2.1 Harmonic constants at points along the 900 m hydrodynamic model boundary. H is amplitude in centimeters, g is phase relative to the lunar crossing at Greenwich in degrees, and LAT is lowest astronomical tide in centimeters.

5.1 Summary of sensitivity runs for salinity.

-vi-


LIST OF PLATES

1. Sate l l i t e image of the study area (Sept. 4 , 1 9 7 2 ) .

2 . Sa te l l i t e image of the study area (Jan. 2 6 , 1973).

3 . Sa te l l i t e image of the study area (Oct. 6 , 1 9 7 8 ) .

4 . Sa te l l i t e image of the study area (June 6 , 1 9 7 9 ) .

5 . Sa te l l i t e image of the study area (May 11, 1984).

6 . Sa te l l i t e image of the study area (Apr. 1 2 , 1 9 8 5 ) .

7 . Sa te l l i t e image of the study area (May 4 , 1 9 8 7 ) .

- v i i -


EXECUTIVE SUMMARY

'i'

/

.. 4

I

The Abu Ali Causeway Study was conducted in order to determine the likely impact of the proposed Abu Ali causeway extension on the current, salinity, and sedimentation patterns in the vicinity of Abu Ali Bay.

Current and salinity transport computer models (SYSTEM 21) were licensed from the Danish Hydraulic Institute and installed on the EPR mainframe computer. The wave refraction modeling program, REFRACT, was used to develop wave fields for sedimentation analysis, and Dr. Madsen of the Massachusetts Institute of Technology was retained as a consultant for the sedimentation work. The current and salinity models were calibrated with actual field data acquired by the Aramco Environmental Unit during August 1987. Historical aerial and satellite photographs were used to qualitatively assess the effect of the existing causeway on sedimentation patterns.

The calibrated current and salinity models were used to determine the long-term changes in circulation and salinity patterns associated with the primary causeway extension option. The results indicate that maximum spring tidal currents will be increased from 1.0 to 1.5 meters per second in the near-vicinity of the tip of the proposed causeway extension. The increase in peak current speed is quite localized. The resulting long-term changes in salinity due to the causeway extension are small, and are within the statistical and measurement uncertainties associated with historical data sets acquired by Aramco.

For the sedimentation analysis, examination of the prevailing circulation currents and wave patterns suggests that the long-term effects of the causeway extension on sedimentation patterns would be small, due to the shelter provided by Abu Ali Island to the prevailing northwesterly wind and circulation. Therefore, the work examined what effects the causeway extension would have on sedimentation (erosion and deposition) during rare storm events from the southeast (exposed) direction. The analyses developed one- and ten-year return period current and wave fields, determined if sediment grains could be picked up in these storms, and computed the net sediment transport for each storm case. The results indicate that less than 30 cm of sediment erosion or deposition would occur during either storm case, with most of the activity localized near the tip of the causeway extension.

The study concludes that implementation of the primary causeway extension option will have negligible effect on the salinity and sedimentation patterns of the area, with an increase in tilaximum tidal curr\nts localized near the tip of the extension.

-1-


CONCLUSIONS AND RECOHHENDATIONS

The primary causeway extension option results in some small, localized change in the current patterns in Abu Ali Bay.

The primary causeway extension option does not result in an appre- ciable change in bay salinity.

The primary causeway extension option will have little effect on the sedimentation pattern of the area.

The DHI System 21 hydrodynamic model is an appropriate tool for nearshore circulation studies. However, some shortcomings in the System 21 transport-dispersion model need to be rectified by DHI before we can recommend its use by Aramco.

e EPR recommends that a "minimum" level environmental monitoring program (salinity and perhaps other variables) be undertaken before, during and after causeway construction. A well designed program would be relatively inexpensive and would allow Aramco management to make informed decisions on future constructions and/or remedial actions.

- 2 -


..

1. INTRODUCTION

In April 1987, Aramco authorized Exxon Production Research Company to determine the effects that a proposed extension to the Abu Ali Island causeway would have on salinity, circulation patterns and sedimentation in and around Abu Ali Bay. The existing causeway and the proposed extension are shown in Fig. 1.1.

Because of the complexity of the regional circulation and salt and sediment balances, the only possible way of addressing what effects the extension would have was through the use of numerical models. Hydrody- namic and mass transport models (System 21) were licensed from the Danish Hydraulic Institute and installed at EPR. The wave refraction modeling program, REFRACT, was used to develop wave fields which, along with modeled currents and sea levels, were used in the sedimentation analysis. An overview of the entire modeling procedure is presented as a flow chart in Fig. 1.2.

The main body of this report is organized as follows. Chapters 2 and 3 present descriptions of DHI's hydrodynamic and transport-dispersion models, their setup for the Abu Ali application and the results of their calibration.

Chapter 4 discusses the analysis procedures used to assess the impact of the proposed Abu Ali Island causeway extension on the sediment erosion/deposition in the area.

In Chapter 5, we present the details of how we assessed the impacts that the causeway extension will have on currents, salinity and sediment transport.

For completeness, we have included a number of appendices which document, in detail, those methods or data that were developed or obtained especially for this project, or are not readily available elsewhere.

- 3 -


2. HYDRODYNAMIC MODEL

2.1

The hydrodynamic model is a finite difference approximation to the two-dimensional (vertically averaged) equations of mass and momentum conservation for a homogeneous, incompressible fluid. The numerical model (DHI's S21HD/MK83) solves for water level and the two components of depth-averaged horizontal velocity. Input data to the model includes bathymetry and relevant ancillary information (bottom drag coefficient, horizontal eddy viscosity, wind, open boundary conditions, etc.).

The model has a built-in interactive refinement of scale facility which allows for increased grid resolution in areas of special interest. Interactive nesting assures consistency (in a numerical sense) between solutions obtained on the different grid meshes. The interactive nesting facility was not used in this study because of computer storage con- straints. Rather, solutions on fine grids were obtained by imposing coarse grid solutions as boundary conditions. The adequacy of this approach can be judged by reviewing the model calibration results presented below.

The model also allows for the flooding and drying of shallow areas. This facility was used in order to represent the extensive region of tidal flats within Abu Ali Bay.

A special modification to the model was made by DHI specifically for this project. The modification allows for the representation of causeway culverts whose widths are smaller than the model's grid spacing. A description of this modification is given in Appendix A. Since our modeling results indicated that the solid fill causeway extension had minimal impact on the region's salinity and sediment transport, analysis of cases including causeway culverts was not performed.

Additional details on the numerical model are given in DHI's System 21 MK83 Users' Guide (DHI, 1984).

2.2 Grid Coverage

Three grids were used in the hydrodynamic modeling: (1) a coarse grid, with 900 m spacing, encompassing the entire study area (Figs. 2.1 and 2.2); (2) an intermediate grid (300 m spacing) covering all of Abu Ali Bay (Fig. 2.3); and (3) a fine grid (100 m spacing) which includes the area immediately surrounding the existing causeway and proposed extension (Fig. 2 . 4 ) . The coarse grid coverage was designed to make the best use (in an economic fashion) of Aramco's tidal height data (Fig. 2.1) in specifying model open boundary conditions.

Description of the Hydrodynamic Model

-4-


2.3 Bathymetry

The bathymetry for all three model grids was prepared from a common bathymetric data base. This data base included measurements from the following sources: (1) Aramco's high resolution survey charts; (2) inferred bathymetry from processed IANDSAT imagery; and (3) DHI gridded bathymetry from the Arabian Gulf Hindcast Study. These data were combined into a single master file which was interpolated onto a 100 m grid using EPR's GPMAP program. The gridded data was then subsampled at 300 and 900 m intervals in order to obtain bathymetric matrices for the 300 and 900 m current model grids, respectively.

The use of LANDSAT imagery for obtaining bathymetry is relatively novel. A description of the technique used at EPR is presented in Appendix B.

2.4 Boundary Tidal Data

Boundary tidal data were interpolated from the sea level tidal constants available at the stations shown in Fig. 2 . 1 . The interpolation was facilitated by making use of the cotide charts for the M2, S2, N 2 , K1, and 01 constituents prepared for Aramco by Dr. John Boone (Virginia Institute of Marine Science). The procedure was based on linear interpolation of the data points and the cotide charts. The procedure was somewhat subjective due to the coarse spatial coverage of the data. Acceptability of the specified boundary conditions was judged by comparing modeled tides at Abu Ali Pier and Berri 46/51 with the predicted tides at these locations (see discussion below).

The eight most significant constituents were specified at the model boundary. The amplitude and phase for each constituent at various points along the 900 m grid open boundary are presented in Table 2.1. The procedure for specifying boundary conditions was to predict the total tide at these points for the desired period (using the harmonic constants) and then to linearly interpolate the predicted tide between these points in order to establish a specified tide height for every boundary grid cell. Errors due to neglecting additional tidal constituents are relatively small.

2.5 Model Calibration

The tidal model was first calibrated by adjustment of boundary data and bottom drag coefficient until a best match was obtained against predicted tide height at Abu Ali Pier and Berri 46/51 (see Fig. 2.5 for station locations). A comparison of modeled and predicted tide height at these two locations is presented in Fig. 2 . 6 .

Unfortunately, the modeled tide height at both of these locations was relatively insensitive to large changes in bottom drag coefficient, although the currents within the bay were found to be quite sensitive. Consequently, the current meter data obtained by Aramco during their August 1987 survey were also used in the model calibration. The location of the current meter stations is shown in Fig. 2.5.

. .

- 5 -


A thirteen hour period was selected from the month-long current meter records for use in the calibration. The period chosen is one of spring tides during which the winds were light so that winds could be neglected in the simulations. The calibration was performed on the 300 m grid in order to adequately resolve the bathymetry surrounding the current meter mooring locations.

Numerous runs were performed in which different bottom drag coefficient formulations (Chezy or Manning) and different values for the coefficients were used. The best match obtained is shown in Fig. 2.7. These were obtained with a Manning's n of 32 m%/s. Current speed at Station A is slightly overestimated but in phase with the measurements while those at Station B are somewhat underestimated and are roughly an hour out of phase with the measurements (model leads the data). Current directions are in excellent agreement with the measurements indicating that the local bathymetry is correctly represented in the model. The sources of error in modeled results are due, in part, to large spatial gradients in the flow field in the vicinity of the measurements (entrance to the bay). The 300 m grid resolution apparently does not capture this variability perfectly. Calibration runs using the 100 m grid were not performed since they would have been prohibitively expensive (roughly 30 times the CPU time required on the 300 m grid). The error in phasing at Station B may be due to errors in the representation of the bathymetry within the bay (which were derived indirectly from LANDSAT imagery). The effects of these errors on subsequent calculations of the bay's salinity are small (Chapters 3 and 5).

Aramco's August 1987 survey included a water level recorder located inside the bay (Fig. 2.5). Modeled water levels at this location were in poor agreement with the measurements during low tide. This is because the region surrounding the gauge is extremely shallow and becomes exposed at low water. Bathymetric data inside the bay was only available to Lowest Astronomical Tide (LAT). Bathymetry above LAT was extracted as best as possible from available charts.

-6-


3 . SALINITY MODEL

3.1

The transport-dispersion model is a finite difference approximation to the two-dimensional (vertically averaged) mass transport equation. The model equation includes a source term representing the increase in salinity due to evaporation. This term enters the model equation (rather than being applied as a boundary condition) due to the vertical integra- tion.

The numerical model (DHI's S21TD) solves for concentration of a conservative (e.g.; salinity) o r non-conservative (e.g.; dissolved oxygen) constituent. Input data to the model includes velocities, water depths, horizontal eddy dispersion coefficients, evaporation rates, etc. A complete description of the numerical model is given in DHI's S21TD manual (DHI, 1984).

The finite difference approximations used to represent the model equation are as follows: (1) advective terms are approximated by a 12-point "Everett" interpolation scheme; ( 2 ) diffusion terms are approximated by standard "second order" central differences; and (3) time stepping is accomplished by standard differencing techniques. Although the Everett scheme yields fairly high order accuracy (3rd order), it is cumbersome (and, in fact, inaccurate) adjacent to land boundaries. (The order of accuracy of a numerical scheme is defined by the number of terms retained by the numerical approximation of the Taylor series expansion of a given differential operator.) This problem arises because the scheme is Lagrangian in nature, and has difficulty in representing the no-flux boundary condition required at land boundaries. This can be seen in Fig. 3 . 1 in which the shaded grid cells represent a thin land barrier such as a causeway. The scheme advects a water parcel from point A to point B in one time step. The salinity at point A is determined by interpolation using values at the twelve points marked by an 'x'. Some of these points fall on land cells so that an "appropriate" salinity for them must be set. In practice, the code is not capable of setting these salinities to exactly satisfy the no-flux boundary condition. Also, note that two of the grid points used in the interpolation are actually located on the opposite side of the land barrier. This difficulty is exacerbated when the land boundaries are represented on a "coarse" grid. In the case of both the 300 and the 900 m grids used in the Abu Ali simulations (described below), the Everett scheme resulted in numerical instabilities so that results from the model were unusable.

This difficulty was discussed with DHI. They confirmed that such a problem could occur on a "coarse" grid, and stated that a degenerate advection scheme was available at DHI (i.e.; one that used fewer points in space for the interpolation, and consequently has a lower order accuracy in the finite difference sense) that might circumvent this problem.

Description of the Transport-Dispersion Model

7 -


Since the schedule was very tight at the time that this problem was identified and discussed with DHI, EPR decided that it was necessary to replace the Everett scheme with a far simpler and widely used finite difference approximation to the advective terms: upwind differencing (Roache, 1982). Although this scheme gives rise to numerical dispersion, the amount of the artificial dispersion is predictable by theory. In fact, there is more than a passing similarity between the functional relationship between the numerical dispersion coefficient and modern descriptions of the physical dispersion coefficient, which represents turbulent mixing and shear effects (see, for example, Bowden, 1983).

DHI's "degenerate" scheme may be of use to Aramco in future investiga- tions; but our recommendation is that it should be carefully tested before any license is undertaken. It may very well be that the simple upwind differencing scheme used here is the most appropriate one to use in any future coastal mixing studies.

Boundary conditions for all model simulations were handled identically; boundary salinities on inflow into the model domain were specified as 42 parts per thousand (ppt) based on the salinity data made available to us by Aramco. S21TD extrapolates salinities linearly from inside the model domain on outflow. In practice, this procedure worked well since the model boundaries were set "far enough" away from the regions of large salinity increases due to evaporation (shallow portions of the bay).

S21TD allows for a constant in time and space evaporation rate to be specified. The S21TD code was modified by EPR so that a time varying evaporation could be specified. This was necessary in order to conduct both the salinity model calibration and "worst-case" scenario simulations. Evaporation rate is a computed quantity which is a function of a large number of meteorological and oceanographic parameters. Appendix C gives a full description of the methodology and data sources used in this project to calculate evaporation rates.

There is virtually no well-accepted formulation for the specification of horizontal dispersion coefficients. In a 2-D model, these coefficients represent both the mixing action of sub-grid scale turbulent fluctuations and the "shear effect" (G.I. Taylor, 1954) which arises due to the unrepresented shear in the vertical profile of horizontal velocities due to the vertically integrated form of the equations used in the model. Existing guidance is highly empirical, is a strong function of a multi- tude of local conditions, and compares poorly with available field data. A relatively recent review of the state-of-the-art in estuary mixing processes is given in the textbook by Fischer, et al. (1979). Bowden (1983) presents a particularly simple form: D - UH, where U is the depth-mean current velocity, H the water depth and D the dispersion coefficient.

In the Abu Ali study, the dispersion coefficient was taken as a free parameter and was adjusted until it gave an acceptable match to Aramco's August '87 salinity measurements within the Bay.

-8-


3.2 Grid Coverage

Two grids were used in the salinity modeling: (1) a coarse grid with 900 m spacing, encompassing the entire study area (but not encompassing all of the coarse grid used in tlie S21HD simulations); and ( 2 ) a fine grid (300 m spacing) which includes the area immediately surrounding the existing causeway and proposed extension (again, this grid is a subset of the 300 m grid used in the hydrodynamic modeling). The grid coverage was chosen so as to give the finest resolution that could be considered while keeping the open boundaries far enough away from the bay so that errors in their specification would not overly affect results in the bay.

An extensive set of numerical experiments were conducted (mainly using the 900 m grid) in order to gain an understanding of the relative impor- tance of the modeling assumptions. Use of the 900 m grid for this purpose facilitated the calculations since it is 27 times less costly to compute than the 300 m grid. The extent of each of the grids can be seen in Figs. 5.4 and 5.5.

3.3 Model Calibration

The primary source of data used to calibrate the dispersion coefficient in the transport-dispersion model was the salinity measurements made by Aramco during August 1987. The positions of the sampling stations are shown in Fig. 3.2. Since there was an insufficient amount of data to adequately specify an initial salinity condition throughout the model domain, we chose to specify the initial salinities at a uniform value of 42 ppt (equal to the time-independent boundary value). This means that it was necessary to run the simulation long enough so that transients in the salinity field (resulting from the poorly specified initial condition) are removed from the system. Based on numerical experimentation with the model (see Section 5.2), one year was deemed sufficient to "wash out" these transients.

Figure 3.3 shows the year-long time history of evaporation rate used to drive the calibration simulation. As is to be expected, there is a strong annual cycle corresponding to the annual cycles of air temperature, humidity and other less important parameters. Results of the calibration simulations are shown in Fig. 3.4 where modeled salinities are plotted against measured salinities for three different values of the dispersion coefficient (D): 2, 3, and 4 mz/s. As can be seen from the figure, the level of agreement is extremely sensitive to the value of the dispersion coefficient, with D=3 mz/s giving the best fit. This indicates that, especially in the inner portions of the bay where tidal stirring is minimal, dispersion is a dominant mixing mechanism. However, the modeling problem is not one of dispersion alone, as advection plays an important role as well.

-9-


4. SEDIHENT TRANSPORT ANALYSIS PROCEDURE

The primary goal was to quantify the effect of the causeway extension on the existing equilibrium in sediment transport in the area (i.e.; the potential for large changes in bottom topography).

For the purpose of this study, the severity of sediment transport in the region was quantified based on an analytical approach only. Due to large uncertainties stemming from scale-effects and the high cost of physical models, an experimental approach was not considered practical. Professor J. W. Kamphuis of Queen's University, Canada, in his 1985 state-of-the- art review paper (Kamphuis, 1985) on physical modeling of sediment transport still maintains his earlier conclusion (Kamphuis, 1975) :

"Owing to the variety and magnitude of scale effects, which can only be fully understood by an experimenter with experience, modeling coastal areas will continue to appear an art."

The analytical approach involves quantification of combined effects of waves and currents on the initiation and transport of bottom sediments. While the analytical approach itself is based on several simplifying assumptions, its reasonableness is judged by the results on the severity of sediment transport obtained for the existing causeway. The results are in general agreement with the overall existing equilibrium in sediment transport in the region as implied by the Landsat images shown in Plates 1 to 7 covering the study region for a time span between September 1972 and May 1987. The same topographic features one can see in the vicinity of the causeway in Plate 1 (Sept. 1972) are also visible in Plate 7 (May 1987) .

The rest of Chapter 4 discusses the procedure used in numerical modeling of waves in the region. This is followed by the discussion of the procedure used to compute the sediment budget under the combined effects of waves and currents. The currents used are the total currents composed of tidal and wind-generated currents as established using the procedure discussed in Chapter 2. The wave-induced current due to breaking waves in the surf zone was not explicitly accounted for due to its very small magnitude compared to the tidal current.

The assessment of the severity and extent of topographical changes induced by the causeway extension requires quantification of not only the effects of the extension on the wave and current patterns but also the interaction of the resulting fluid motion with the bottom sediments. In the latter, the key questions to be asked are whether or not the sediment is going to move due to the combined action of waves and currents and if so, at what rate? Section 4.1 discusses the procedure used to investi- gate the effect of the causeway extension on the waves in the region, while Section 4 . 2 discusses the procedure used to compute the sediment transport. Both the wave and current modeling and the associated sediment transport calculations were performed for one-year and ten-year

-10-


I 8 1 c I.

storm conditions. Since the primary interest is to examine the long-term (-15-20 years) impact of the causeway extension, it seemed reasonable to model severe conditions that are expected to occur several times during the time span of interest. For the purpose of sediment transport analysis, a long term cumulative effect is of greater interest and can be studied by the use of one-year storm conditions. The effects of more severe storm conditions (10-year) are also assessed. The combined effect of waves and currents on the bottom sediments for a return period of less than one year are not expected to be significant due to the shelter provided by Abu Ali Island to the prevailing northwesterly wind and circulation. Therefore return periods less than one year were not examined.

4.1 Wave Modeling

The input conditions used in the one-year and ten-year wave models were representative of southeasterly storms. This was judged to be the direction of the dominant wind-generated waves influencing sediment transport outside the bay. Actually, the dominant wind direction is northwesterly, but the waves from that direction are fetch limited due to land blockage and are not expected to have any significant impact on sediment transport outside the bay.

The effect of the causeway extension on waves representing one-year storm conditions was examined using the computer program REFRACT (Martin et al., 1987), developed by Professor R. A. Dalrymple. REFRACT is based on wave action equations which include the effects of wave refraction, shoaling, partial breaking, and bottom friction. Professor Dalrymple has also developed a computer program called REF/DIF 1 (Martin et al., 1987) which can account for the combined effects of refraction and diffraction: however, the model is expensive in terms of computer cost due to the small grid sizes required. At least five computational grid points per wave length are required in the REF/DIF 1 analysis. The water depths in the vicinity of the causeway being less than 1 meter required grid spacing of less than 2 m. The REF/DIF 1 analysis for such a fine grid was not only very expensive, but due to uncertainties in water depths interpolated to so fine a grid from the original 100 m grid, the combined refracted/diffracted wave height results were questionable.

The REFRACT model is a much simpler computer program and does not have the limitation of requiring at least five computational grid points per wave length. However, the energy dissipation due to bottom friction and partial breaking in REFRACT are not predicted as well as by REF/DIF 1. In general, REFRACT is expected to yield a more conservative estimate of wave heights than REF/DIF 1. For this particular study, due to the questionable nature of the REF/DIF 1 results, the sediment budget analysis relied upon the REFRACT results.

REFRACT was implemented on a 100 m x 100 m grid with the existing causeway (Fig, 2.4a) and a similar grid with the causeway extension also included (Fig. 2.4b). For both grids, the following input conditions, derived from the Danish Hydraulic Institute hindcast results (DHI, 1985)

-11-


were used in REFRACT runs for one and ten-year conditions. The statistics in the table represent the southeasterly storm. REFRACT assumes a sinusoidal wave having the wave height as indicated in the table.

Input to Program REFRACT 1-Year Model 10-Year Model

Significant Wave Height 1.55 m 3.40 m Wave Period (Zero-Crossing) 4.0 sec 6.0 sec Maximum Surge 2.34 m above U T 2.50 m above LAT Minimum Surge 0.80 m above LAT 0.90 m above IAT

4.2 Sediment Transport Analysis

The procedure developed by Professor 0. S . Madsen and Dr. W. D. Grant (Madsen and Grant, 1976; Grant and Madsen, 1986) to predict the net sediment transport under the combined action of waves and currents was used in this study. The above references contain detailed derivations of the equations and therefore only a brief outline of the procedure is presented below. Further details are presented in Appendix D.

Firstly, it should be noted that the procedure is applicable to only cohesionless sediments and to "non-breaking" wave conditions. The former of these limitations does not seem to be severe in view of the fact that the bottom sediments in the study region are fine sand. The latter limitation excludes the direct application of the results in the immedi- ate vicinity of the structure (one-grid spacing) and in the surf zone where the conditions are complicated by breaking waves. The procedure can be modified to account for wave-induced current due to breaking waves when the surf zone is extensive. Using the procedure discussed in Appendix D, the magnitude of the wave-induced current was found to be very small (-5 cm/sec) and therefore was not explicitly considered in the analysis.

The question of whether or not a sediment particle would be moved under the combined action of waves and currents was resolved by computing the Shields Parameter (see E q . (D.12) in Appendix D) and comparing its value with the critical value for initiation of sediment movement obtained from the Shields diagram (Madsen and Grant, 1976). For this study, the critical value of 0.05 was used in all the analysis.

When the Shields parameter exceeded - 0.05 at - a grid point, the average net volume of sediment transport qxt and q Y i along the local x and y directions (in-line and normal to wave direction), respectively, were computed as outlined in Eqs. (D.5) and (D.6) in Appendix D. Next, using the sediment continuity equation given below for each grid cell, an estimate of deposition/erosion within a grid cell was obtained:

- (1 - E ) 5 9 + -x+ a< % - 0

a t ax ay

-12 .


# P 1 'I I t I i Y i il 1 Y

1 1 I 1 s

i

where t is the porosity of the bottom sediment. X and Y are the grid axes. The rate of erosion/deposition within a grid cell corresponding to a given year return period wave and current environment was then computed by evaluating ?.?l at each one-hour time interval and then sum- ming the hourly contributions over the entire tidal cycle ( 2 4 hours). The one year currents also included wind driven currents.

The procedure outlined above was carried out for the existing and the extended causeway conditions. Comparison of the severity of erosion/deposition for the existing and extended causeway was used to assess the extent of topographical changes induced by the causeway extension under both one-year and ten-year storm conditions.

at

-13-


5 . EFFECTS OF PROPOSED CAUSEWAY EXTENSION ON CURRENTS, SALINITY AND SEDIMENT TRANSPORT

5.1 Tidal Circulation

The predominant forcing mechanism for flow within the bay and its sur- roundings is the astronomical tide. Because of this, we chose to define a "worst case" scenario to define the effects of the causeway extension on bay flow conditions as that associated with the maximum spring/neap tidal cycle over the course of one year. 1987 was taken as a representative year, and Aramco's tide tables (Aramco, 1986) were reviewed for the Abu Ali Pier and Berri Platform locations in order to find the spring/neap cycle having the greatest range in vertical cide. The period chosen was January 1-15, 1987. Note that the spring/neap cycle is approximately 14 days; the first day was used to "spin up" the hydrodynamic model from a state of rest.

Boundary tide heights for this period were generated from interpolated harmonic constants (see Section 2 . 4 ) . A full run was first made on the 900 m grid with results stored five times per hour, thereby resolving all important tidal frequencies. Boundary conditions for the 300 m grid were then extracted from the 900 m grid results, and the hydrodynamic model was rerun on the 300 m grid for the full period.

Maximum current speeds over the 14 day spring/neap cycle (discounting the initial first day of the run which was used for model spin-up) were determined by scanning the time histories of current speed at each grid cell in the 300 m grid. The resulting maximum current speeds are contoured in Figs. 5.la (existing causeway) and 5.lb (existing causeway plus extension). Only speeds greater than 0.6 m/s are contoured, and the contour interval is 0 . 2 m/s. Without the extension (Fig. 5.la). speeds in excess of 0.6 m/s are seen to occur at the eastern tip of Abu Ali Island (lower right hand corner of the figure) and within the entrance to the bay. Note that the high speeds to the east of Abu Ali are not an artifact of the 300 m grid since these results are consistent with those obtained on the much larger 900 m grid.

In the entrance to the bay, the speeds reach a maximum of 1.0 m/s right at the tip of the existing causeway. The effect of the causeway extension is to increase the current speeds in the bay entrance since the cross sectional area across the entrance is substantially reduced. Quantitatively, the spatial extent of the 0.6 m/s contour is somewhat reduced because of the constraint to flow imposed by the extension. However, the current speeds in the entrance region increase substantially, reaching a maximum of 1.5 m/s at the tip of the extension. Time history plots of current speed at select locations over the first ten days of the simulation are presented in Fig. 5.3. Figure 5 . 2 shows the location of these stations.

-14-


t 1)

S i 1 t I i Y 1 S i I c 11 I I 1 i

5 . 2 Salinity Distribution

A simple scaling calculation and preliminary model runs both indicared that roughly one year would be required for the bay's salinity field to reach an equilibrium state with the annual evaporation cycle. The relatively long "spin-up" time was needed since an initial salinity distribution was not available from measurements so that an artificial one had to be adopted. [This artificial distribution was chosen to be a uniform one with the salinity set equal to the open boundary value of 42 ppt. The 42 ppt value is a good representative boundary value based on the salinity data Aramco made available.]

Knowing that integrations of the transport-dispersion equation of a year or more would be required, we next addressed how to specify a representative flow field for the calculations. Since the tide dominates the flow within the area, and since the wind-induced currents are intermit- tent and are small in comparison to the tidal currents, we chose to represent the flow field using a "mean" tidal cycle of 15 hours in duration. This mean tidal cycle was selected from the Aramco 1987 tide tables with the criterion for selection being that the range of the tide at Abu Ali Pier and Berri Platform be about the average of all diurnal cycles over the entire year. The period chosen was January 5 - 6 , 1987.

The procedure taken to carry out the hydrodynamic simulations was iden- tical to that described above with the exception that the boundary tidal heights were generated for the chosen period. These 15-hour long boundary data were repeated for 5 cycles (75 hours) so that the hydrodynamic model could be spun up. The results (water depths and currents) for the last cycle were archived at one hour intervals for subsequent repeated use in the transport-dispersion calculations.

In order to assess the long-term effect of the causeway extension on the bay's salinity, which may take years to fully manifest itself, it was necessary to specify some meaningful representation of the variation of evaporation rate with time. We chose the following specification:

E = c[(a + b) + b cos (ut - * ) I 5.1

where E is the evaporation rate and a, b, and c are constants. The period of rhe sine wave was taken as one year. The constants were chosen to represent the mean evaporation rate at Ras Tanura plus one standard deviation, as computed from the seventeen years of monthly averaged meteorological data reported in the Aramco "metocean report" (Williams, 1979). This results in the following values for the constants: a = 7 cm/month, b - 6 cm/month, c = 1. Evaporation rates were calculated using the "bulk" equation described in Appendix C. The dispersion coefficent was specified as 4 m2/s based on the results of the salinity model calibration (Section 3 . 3 ) .

There is considerable uncertainty in specifying both the evaporation rate and the dispersion coefficient. Also, we had questions about how sensitive the model results would be to: (1) grid extent (i.i., how far the

-15-


model open boundaries are from the bay), (2) duration of the simulation, and (3) grid resolution. Consequently, we performed a number of simulations to establish how sensitive our results were tv uncertainties in the modeling procedure. An inventory of the simulations is given in Table 5.1. Although salinities varied significantly in some of the sensitivity runs, the change in salinity resulting from the presence of the causeway extension did not vary to any significant degree from run to run. This indicates that the final assessment of the causeway extension impact on bay salinities is robust, since it is not sensitive to key modeling assumptions.

Salinity contour maps for the 900 m grid are shown in Fig. 5.4a (without extension) and 5.4b (with extension). The contours are plotted for August conditions when salinities reach their annual maximum in the bay. Comparing the two figures show that the salinities change only a small amount due to the presence of the causeway extension. Essentially the same salinity distributions appear in the 300 m grid results (Figs. 5.5a and 5.5b) illustrating the fact that the 900 m grid resolution is ade- quate for these simulations. The extension causes salinities to rise slightly (<1 ppt) on the bay side of the extension and to fall by roughly 2 ppt at the head of the bay. We do not, as yet, have an explanation for the slight reduction in upper bay salinities when the extension is in place. However, the model results have been carefully checked to make certain that a proper balance to the mass transport equation is being maintained. This check gives a good level of assurance that the finite difference representation and coding is correct. [The very good agreement with August 1987 salinity measurements obtained as part of the model calibration (Section 3.3) substantiates this.]

In any event, Fig. 5.6 illustrates that the predicted change in bay salinity due to the causeway extension (shown on the right of the figure) is small when compared with the natural variability in bay salinity, as seen from the bioaccumulation measurements (Jan. 1982-Nov. 1984) and the August 1987 measurements (center). We therefore conclude that changes in bay salinity due to the presence of the causeway extension are negligibly small.

5.3 Sediment Transport Analysis Results

5.3.1 Wave Modeling Results

Figs. 5.7a and 5.7b show vector plots of wave directions for a one-year storm where the length of the vector represents the wave amplitude, for before and after causeway extension, respectively. Figs. 5.8a and 5.8b show similar results but for ten-year storm conditions. The vector plots essentially show refraction of waves due to bottom bathymetry with increase in wave height in shallower water. In the one-year storm, the causeway extension has an imperceptible effect on waves. In the ten-year storm, the causeway extension causes some wave height increases near the tip of the causeway inside the bay.

16.


5.3.2 Sediment Budget Analysis Results

Figs. 5.9a and 5.9b present the contour plots of erosion and deposition rates, respectively, for the existing causeway in the one-year storm. Figs. 5.10a and 5.10b are similar plots but for the case with the causeway extension. The erosion/deposition patterns are very similar. Fig. 5.11 is a 3-D perspective view showing the net changes in the bottom topography caused as a result of the causeway extension, i.e., the difference in bottom changes (erosion/deposition) before and after the causeway extension.

The above figures show that on average, erosion/deposition rates are on the order of 1 m/day for one-year storm conditions, and that differences in erosion/deposition due to the causeway extension are greatest at the tip of the extended causeway where they are about 0.25 m/day:

Figure 5.12 is similar to Figure 5.11 but for 10-year storm conditions showing that on average, the 10 year storm conditions can give rise to maximum erosion/deposition differences of 0.30 cm/day near the tip of the extended causeway. These differences in erosion/depositon rates with and without the causeway extension are small (-25%) relative to the absolute erosion/deposition rates without the causeway extension.

-17-


ACKNOWLEDGEKENTS

There are a lot of people who have contributed substantially to this project. We take pleasure in acknowledging their contributions.

Ben Merenbeck and Bob Brovey of EPR’s Exploration Concepts Division created much of the bathymetric data used in the study. Ben wrote Appendix B which documents his methodology for inferring bathymetry from Landsat images. Without Ben and Bob’s timely and creative response to our request, we would not have completed this project as quickly or as well as we did.

Kent Allen (Aramco) was extremely helpful in almost all aspects of this project; from its definition, through the data collection stage and in consulting on interpretation of model results. His quick response in. supplying meteorological data and information on the meteorological instrumentation was very helpful to us. Bruce North (Aramco) also supplied assistance during the course of the project.

Our Offshore Division colleagues: Robert Haring, John Heideman, and John Vermersch, carefully reviewed our progress during the course of the project and gave us excellent guidance and ideas. Without the assistance of Bob Haring and John Vermersch, we could not have adequately presented our results at the Dhahran review meeting last November. We thank all three for their very significant contribution to this project.

The current and salinity models used in our analyses were licensed from the Danish Hydraulic Institute. We acknowledge the assistance of Asker Kej (DHI) and Keith Bell (EPR) in the preparation of the license agreement. Lars Behrendt (DHI) very capably served as project manager for the installation and testing of the software at EPR and was the author of the coding for representation of sub-grid scale culverts in the S21HD model.

Ole Madsen (MIT) was our consultant for the sediment transport analysis procedures.

We also thank Ruth McHugh, Monte Dunton, Joe Klingshorn, Glen Turner, John Wilson, Ivy Kaminsky, and Mike Chicca (all of EPR) for their assistance.

-18-


REFERENCES

Aramco (1977), "Evaporation Rates for Aramco Operating Areas," AER-4054, Environmental Unit, Mechanical Systems Division.

Berry, S. A. and N. R. Beers (1945), Handbook of Meteorology, McGraw- Hill, N.Y.

Bowden, K.F. (1983), Physical Oceanography of Coastal Waters, John Wiley & Sons, New York.

Danish Hydraulic Institute (1984), "System 21 MK83 Users Guide -FORTRAN."

Danish Hydraulic Institute (1984), "System 21 Transport-Dispersion Model Users Guide."

Danish Hydraulic Institute (1985), "Arabian Gulf Hindcast Study, TS1 57-174 Offshore Design Criteria," Vol. 1, Report Prepared by Danish Hydraulic Institute for Aramco Overseas Company.

Fischer, H.B., E.J. List, R.C.Y. Koh, J. Imberger, N.H. Brooks (1979). Mixing in Inland and Coastal Waters, Academic Press, New York.

Grant, W. D., and Madsen, 0. S. (1986). "The Continental-Shelf Bottom Boundary Layer," Ann. Rev. Fluid Mech., pp. 265-305.

Hargreaves, C. H. (1970), "Consumptive Use Derived from Evaporation Pan Data," Transactions of the American Society of Civil Engineers, Vol. 135.

Kamphuis, J. W. (1985), "On Understanding Scale Effect in Coastal Mobile Bed Models," Physical Modelling in Coastal Engineering, Edited by R. A. Dalrymple.

Kamphuis, J . W. (1975), "The Coastal Mobile Bed Model - Does It Work?" Proc. Modelling '75, San Francisco, pp. 993-1009.

Madsen, 0. S., and Grant, W. D. (1976). "Sediment Transport in the Coastal Environment," MIT, Civil Engineering Dept. Report No. 209.

Madsen, 0. S . , Ostendorf, D. W., and Reyman, A. S. (1978), "A Long Shore Current Model," Coastal Zone '78, Proc. of the Symposium on Techni- cal, Environmental, Socioeconomic and Regulatory Aspects of Coastal Zone Management (ASCE), San Francisco.

Martin, C. J., Dalrymple, R. A,, and Miller, M. C. (1987), "A Verifica- tion Procedure for Wave Propagation Models," Modelling the Offshore Environment, Society of Underwater Technology, London.

Miller, A. and J. C. Thompson (1975). Elements of Meteorology, 2nd Edition, Charles E. Merrill Publishing, Columbus, Ohio.

-19-


Price, J. F. (1982). "Program ONEDX Manual," report prepared for EPRCo.

Roache, P. J. (1982), Computational Fluid Dynamics, Hermosa Publishers, Albuquerque, N.M.

Rouse, H. (1946), Elementary Mechanics of Fluids, Dover Publications, New York, 376 pp.

Taylor, G.I. (1954), "The Dispersion of Matter in Turbulent Flow through a Pipe, Proc. R. SOC. London Ser. A, 223, 446-468.

Williams, R. 0. (1979), "Aramco Meteorological and Oceanographic Data Book for the Eastern Province Region of Saudi Arabia," Aramco Environmental Unit, Mechanical Systems Division.

-20-


APPENDIX A

REPRESENTATION OF CAUSEWAY CULVERTS

In the initial scoping of this study, Aramco was considering the analysis of a causeway extension containing culverts. Because of this, EPR had the Danish Hydraulic Institute (DHI) make the modifications to their S21HD model necessary to represent culverts which are less than one grid spacing in width (i.e., "sub-grid scale").

The culvert flow representation is based on the Manning formula for open channel flow (e.g., Rouse, 1946). Coding of the Manning formula within S21HD was performed by Dr. Lars Behrendt of DHI and was fully tested during installation at EPR.

The following pages are the notes left by Dr. Behrendt to document the finite difference implementation and its use within S21HD.

A- 1


page 1 of 3 . 03.07.87/LAB

Sub-Grid Scale Internal Boundary Conditions.

This facility is based on the so-called Manning Formula for flow in a channel of rectangular cross section:

p/h - M R**(2/3) SQRT(ds/dx) (1)

where:

p is the depth-averaged flux, h is the water depth, M is the Manning Number (m**(l/3)/s), R is the hydraulic resistance radius - the hydraulic radius for a rectangular channel

= the water depth for a wide rectangular channel, s is the surface elevation, x is a coordinate in the flow direction.

Defining :

K = M R**(2/3)

and rearranging we get:

p - K h SQRT(ds/dx)

This equation can be discretized as follows:

pn+l = -K ( s n + k - sn+%) SQRT(~.O/(~X*ABS(S:+~ - s n + h ) ) ) j + l j ~ + l j

( 4 ) *0.5 (hj+l + sj+l n + hj + sj) n

giving us the coefficients to replace the original momentum-equation coefficients:

AM0 = -0 .5 K SQRT(l.O/(dx*ABS(Sj+, n+k - Sj n+k )))(hj+l + Sj+l n + hj + Sj) " BMO - 1.0 CMO = -AM0 DMO = 0.0

A- 2


page 2 of 3 . 03.07.87/LAB

Use of the Facility.

In the NBMP-namelist given in the BMP-file, three new logicals can be given. These logicals, which all liave the default value .FALSE. are:

BMANPO, set to .TRUE. if there are any Manning Points in domain 0. BMANPl, set to .TRUE. if there are any Manning Points in domain 1. BMANP2, set to .TRUE. if there are any Manning Points in domain 2 .

Then, for each domain containing Manning Points, these must be specified in the BMP-file just before the specification of the topography. The Manning Point specification consists of:

A head line of text A namelist MANPTS giving the two parameters NXMP and NYMP. These

are the number of x-direction and y-direction Manning Points, respectively. These must both be given, since they have no default values.

A namelist MANVAL for each Manning Point giving j, k, and the constant K (see e.g., eq. 2 ) . X-direction Manning points first, thereafter y-direction Manning Points.

Example:

Test2a: Project 5766, EPRCO, Houston, Licensing of S21 Package. &NBMP JEXTRO-8, KESTRO-17, DXO-200, DYO-200, NEXTR-010, DT-90., BMANPO=T, BMANP1-F, BMANP2-F,

b

b

&END X-M. P. in (j,k)-(4,9) and (2,2), Y-M. P. in (j,k)-(6,6) and (3,l). &MANPTS NXMP-2, NYMP-2 &END W A L JMP-4, KMP-9, KVMP-00.00 &END W A L JMP-2, KMP-2, KVMP-07.00 &END W A L JMP-6, KMP-6, KVMP-11.00 &END &MANVAL JMP-3, KMP-1, KVMP-02.00 &END COARSE GRID Bathymetry: &N2D I2D-7, &END .

b

b

A- 3


page 3 of 3 03.07.87/LAB

General WarninEs and Recommendations.

With this facility, the following approach is recommended:

Estimate the flow expected through the sub-grid structures using simple engineering calculations. Several culverts must here be equivalences with a rectangular channel one grid spacing width.

Perform simple S21 tests with a channel to determine the constant K.

After the S21-run check the flow through the sub-grid structures.

Warnings.

The Manning points in the x-direction must be located within a y-direction one-point wide section. The Manning points in the y-direction must be located within a x-direction one-point wide section.

This means that neighboring water points cannot be Manning Points for the same flow direction.

Manning Points should not be located close to the outer boundaries of the model.

Manning Points should not be located close to inner boundaries in the model where fine-grid and coarse-grid sections meet.

If the difference in water level across the Manning Point at time n is less than 0.001 m, the flux at n+l will for stability reasons be set to zero.

A - 4


page 1 of 4 . 17.07.87/LAB

Culvert Point Formulation.

The Manning Formula for flow in a wide rectangular channel:

where p is the depth-averaged flux h is the water depth M is the Manning number R is the Resistance radius

ds/dx is the gradient (- depth for a wide channel)

A - 5


page 2 of 4 . 18.07.87/LAB

In S21 the Manning formula can be used to calculate the flux in specific points. These points must be water points. In these points the momentum equation is replaced by the Manning formula. The Manning formula can be represented in S21 in two ways:

Case 1 BMANH- .FAJSE.

Defining: K - M R + Eq. (1) can then be discretized as:

%[h + sn + h. + sn] j+l j+l J j

n p”+l

Case 2 BMANH- .TRUE.

Defining: K - M Eq. (1) can then be discretized as:

where we have set:

R = total depth in the flux-point. (OK only for wide channels)

A- 6

( 3 )

( 4 )


1 1

E 1 1

page 3 of 4 . 17.07.87/L.4B

Remarks :

In ( 3 ) and (5) the S-values under the square root should really be at n+%. This, however, is not possible since we need to maintain the implicit term linearly formulated.

Furthermore, the centering of flux p is pushed ahead to the new time step, n-1, where, in the original momentum equation, it is at n+%. This is done trying to make the formulation as stable as possible by putting more weight on the implicit terms.

A- 7


page 4 of 4 . 17.07.87

To use the Manning-formulation for the Domain (subdomain zero) set:

BMANPO-T , in the NBMP namelist.

Then just before the topography in the BMP file insert:

"Header - line of description" &MANPTS NXMP- . . . , NYMP- . . . , BMANH- . . . , &END &MANVAL JMP- . . . , KMP- . . . , KVMP- . . . , &END

With the same definitions as on the sheets dated 3-7-87. Restrictions and recommendations from sheets dated 3-7-87 still apply!

A- 8


APPENDIX B

INFERRED BATHYMETRY FROM LANDSAT IMAGERY By B. F. Merembeck

The image processing problem in this project was whether bathymetry could be accurately predicted in areas where it was not measured. To address this question, two data sets were investigated: high resolution gridded bathymetry, supplied by ARAMCO, and Thematic Mapper Landsat data. The Thematic Mapper (TM) instruments have flown on Landsats 4 and 5. They have 30 meter ground spatial resolution and seven bands of spectral resolution. Band specifications are:

Band Bandpas s

1 0 . 4 5 - 0 .52 micrometers 2 0 . 5 2 - 0 . 6 0 micrometers 3 0 .63 - 0 . 6 9 micrometers 4 0 . 7 6 - 0 . 9 0 micrometers 5 1 . 5 5 - 1 . 7 5 micrometers 6 1 0 . 4 - 1 2 . 5 micrometers 7 2 .08 - 2 .35 micrometers

Bands 6 and 7 are reversed due to band 7 being added in the final design stage after the first 6 bands had been selected.

Of the 7 TM bands, only four were considered useful for this project; bands 1, 2, and 3 for predicting bathymetry, and band 4 for masking land areas from processing. Bands 1, 2 , and 3 cover the visible spectrum from green to red, and are the only TM bands that have any significant pene- tration through water. Band 4 , near infrared, is almost totally absorbed by water and is well suited to delineate land-water boundaries.

ARAMCO bathymetric observations were gridded in GPMAP and a Grid to Data file (GTOD) written to tape for processing. This tape was unpacked and reformatted for our VAX/IDIMS image processing system. Inspection of the data indicated that it was an unusually high resolution grid and had been carefully prepared. To integrate bathymetry and TM data, the coordinate systems of both data sets had to be registered. As the bathymetry was already in the project coordinate system, Thematic Mapper (TM) data were registered to the high resolution bathymetric grid. This was done using base maps of the study area to match GPMAP Grid to Data File X, Y coordinates with phenomena observed on TM; causeways, unique shoreline characteristics, etc. Final registration was an eight point first order fit of the form:

x - a,, + alxx'y' i aZxx' Y - aoy + aIyx'y' + a 2yY'

B-1


where x, y are in base map coordinates, and x' y' are in TM coordinates. Bathymetry in ungridded areas was estimated using least squares multiple regression. The matrix form of the estimate is

- Y - X b

where is the estimated bathymetry, X is the design matrix of the form - 1

1

TM,, TM,, . . . .

TM,. TM,, . . . . Vector b contains the estimated coefficients used to relate TM observations with bathymetry and is of the form

Estimating coefficient vector b is a least squares procedure having the standard form

b = (X'X)-lX'Y

where Y is the vector of gridded bathymetric observations. Since a very large model design matrix, X, was required for this much data, use of a standard statistical package such as SAS was not feasible. Such packages require that all data be kept in memory during computation. A least squares image processing program was written to solve this problem. This program makes use of the fact that X'X and X'Y can be computed simulta- neously for each observation, requiring only two rows of X and a single observation in Y. In taking this approach, any model can be incorporated without requiring large internal program storage by properly constructing and ordering the input images.

Initial processing results revealed that TM bands 1, 2 and 3 accounted for virtually all bathymetric variability. A s mentioned, this was not unexpected because water absorbs the longer wavelengths sensed by the other TM bands. Some nonlinearity in predicted bathymetry was also encountered. This was solved by including natural logarithms from TM bands 1 and 2 into the model.

The final model presented to the regression program was a six band image. The dependent variable was gridded bathymetric data with znils in ungridded pixels. The program ignores znils when computing regression coefficients. Independent variables were TM1, TM2, TM3, Ln TM1, and Ln

B-2


TM2. The regression relationship defined for the pixel in image row i and column j is then

- Yij - a. + al TMlij + a2 TM2ij + a3 TM3,j +

a4 (In TMlij) + a5 (ln TM2ij)

Program output was an image of predicted bathymetry.

Results were analyzed by subtracting predicted from gridded bathymetry and displaying that residual image for inspection. Residual image inspection showed very good bathymetric prediction down to below 30 meters. The TM data would not support estimation much below 30 meters and deeper estimates should be viewed with suspicion.

Due to the time-critical nature of the project, more analytical measures such as the coefficient of determination (R squared) and mean squared error were not incorporated into the regression program. These tools, and others, can be added if this particular image processing technique is used in the future. Final processing output was a GTOD format tape containing predicted bathymetry nested in the same x y coordinate system supplied with the original grid. All land and structures, such as causeways, are set to mil.

B- 3


.

APPENDIX c METHODOLOGY AND DATA SOURCES FOR CALCUIATING EVAPORATION RATES

This appendix describes the formulas and data used in calculating evaporation rates for this study. Our ability to describe the transfer of liquid water into vapor is highly empirical; especially so over the coastal ocean, where the air-sea interface resides within a turbulent boundary layer. The best available engineering approach - the "bulk formula" approach is presented below. This approach best conforms to modern, high quality measurements over the sea given relatively easily available "bulk" measurements. For purposes of comparison, we also present the formula in comon use by Aramco. The Aramco formula is based on pan evaporation data, and is therefore of questionable use for de- scribing conditions over the coastal ocean.

The bulk formula used is taken from Price (1982):

pa

pw E - - Cq(S,o S s a t ) W qw

where E - evaporation rate (m/s) pa - air density taken as 1.23 Kg/m3 pv - water density taken as 1024 Kg/m3 C, - eddy exchange coefficient for water vapor and heat

taken as 1.3 x - specific humidity at 10 m elevation

S a a t - saturation specific humidity W - wind speed at 10 m elevation (m/s)

H.,.+. - heat of vaporization of water equal to 2.47 joules/Kg.

The specific humidities (Slo and S , , , ) are calculated using empirical relationships taken from the Handbook of Meteorology (Berry and Beers, 1945). They are functions of air temperature, surface water temperature and barometric pressure. A program for calculating evaporation rates is included at the end of this appendix.

The Aramco formula is taken from Hargraves (1970), and is given by:

E = 0.97d(1.0 - H)(T - 3 2 ) ( C . 2 )

where E is the evaporation rate in centimeters/month d is monthly daytime coefficient H is mean monthly, relative humidity at noon T is the average monthly temperature degrees F

A comparison between the two formula may be made by viewing Figs. C.l-C.3. Each of these figures shows Ras Tanura evaporation rate statistics by month based on 17 years of monthly mean meteorological data furnished in the Aramco metocean report (Williams, 1979). Fig. C.l shows

c- 1


evaporation rate statistics computed using Eq. (C.2) while Figs. C.2 and C.3 show evaporation rate statistics computed using Eq. (C.l) with an air/sea temperature difference equal to zero (Fig. C.2) or set to its climatological value for Ras Tanura (Fig. C.3). Qualitatively, the two formula yield roughly the same values when the air/sea temperature difference is set to zero in Eq. (C.1). However, including the air/sea temperature difference substantially reduces the evaporation rate (see Fig. C.3). In the salinity analyses, we made use of Eq. (C.l) with the Ras Tanura climatological value of air/sea temperature difference.

c-2


C ws C C BP C

WIND SPEED4 WS

SURFACE EARMETRIC PRESSURE MB .

C A t r u e v a l v e for one of t h e f o l l o w i n g two v a r i a b l e s must be g i v e n : C C TDP DEW POINT TEMPERATURE* DEG. C C (SET TO -99. IF NOT AVAILABLE) C C RH RELATIVE HWIIDITY* PERCENT C (SET TO -99. I F NOT AVAILABLE C C INDICATES VALUES AT 10 M E L W T I O N C c-------------------------------------------------------------------------

L


2/9/88

c I. C DEW POINT TEMP.(DP) AS FNC. OF A I R TEMP.(DB) & WET BULB DEPRESSION (DWB)

C RELATIVE HUMIDITY (RHI AS FNC. OF A I R TEMP.(DB) & C WET BULB DEPRESSION (DWB) I

I I 4 I 0 I I I I I I I 1 1 I I I I I Copyright©Saudi Aramco 2009. All rights reserved.

m 1 1 I I I I 1 I I I 1 I 1 I I I 1 I

c

L

C WRITE THINGS OUT


C

C C

270

271

500 C

C

C

C C

OPEN(UNIT=16,FILE=‘PLOT MONTH’) OPEN(LNIT=17,FILE=’PLOT EWPI’) OPEN(LMIT=18,FILE=’PLOT EWP2’)

DO 500 I=l,ICNT WRITE(16,270) WTH(1) FORMAT(15) WRITE(l7,271) EUAPl(1) FORMAT ( F 1 0 . 1 1 WRITE(18,271) EWP2CI) CONTINUE

CLOSE(16) CLOSE(17) CLOSE(l8)

STOP END

n 1 I t J I I I I I I I 1 I I 1 1 I I Copyright©Saudi Aramco 2009. All rights reserved.

II 1 li 1 1 I I I I I I I I I I I I I I

c


EWIP Page6 2/9/88

C SEARCH FOR WB BIN USING EITHER TDB OR RH, WHICHNER IS AVAILABLE C C INTERPOIATE WET BULB DEPRESSION (TldBD) TO CORRECT TDB OR RH UALUE

C

40 50

C

C

60 70

IF (RH .EQ, -99.) THEN DO 40 J=1,14

JJ = J IF (TDP .LE. DPCCJ) .WD. TDP .GT. DPC(J*l)) GO TO SO

CONTINUE CONTINUE

FACTOR = (TDP - DPC(JJ))/(DPC(JJ+l) - DPCCJJ)) RH (RHC(JJ*l) - RHC(JJ))*FACTOR + RHCCJJ)

END IF

IF (TDP .EQ. -99.) THEN DO 60 J=1,14

JJ = J

C M I N U E C M I N U E FACTOR = (RH - RHC(JJ))/(RHC(JJ+I) - RHC(JJ)) TDP = (DPC(JJ+l) - DPC(JJ))*FACTOR + DPCCJJ)

IF (RH .LE. RHC(J) . M D . RH .GT. RHC(Jt1)) GO TO 70

I L

TTWB = TWB10.55555556 - 17.777777778 RETURN END

I 1 b I 1 I 1 I I i I I 1 I I I I I I Copyright©Saudi Aramco 2009. All rights reserved.

I 1 I I 1 I I 1 I I I 1 I I I I 1 1 I

DEW P D I N l TABLE 2/9/88

1 2 3 4 5 6 7 8 9 1 0 2 0 1 6 1 2 8 2 - 7 - 2 1 0 0 0 0 0 25 22 19 15 10 5 -3 -15 -51 0 0 0 30 27 25 21 18 1 4 8 2 -7 -25 0 0 35 33 30 28 25 21 17 13 7 0 - I f 0 40 38 35 33 30 28 25 21 18 13 7 0 45 43 41 38 36 34 31 28 25 22 18 0 50 48 46 4 4 42 40 37 34 32 29 25 -2 55 53 51 50 48 45 43 41 38 36 33 14 60 58 57 55 53 51 49 47 45 43 40 25 65 63 62 60 59 57 55 53 51 49 47 34 70 69 67 65 64 62 61 59 57 55 53 42 75 74 72 71 69 68 66 64 63 61 59 49 80 79 77 76 74 73 72 70 68 67 65 56 85 84 82 81 80 78 77 75 74 72 71 62 90 89 87 86 85 83 82 81 79 78 76 69 95 94 93 91 90 89 87 86 85 83 82 74

100 99 98 96 95 94 93 91 90 89 87 80 105 104 103 101 100 99 98 96 95 94 93 86 110 109 108 106 105 104 103 102 100 99 98 91 115 114 113 112 110 109 108 107 106 104 103 97

15 20 25 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

-11 0 0 14 0 0 26 -14 0 36 15 0 44 28 -7 52 39 19 59 48 32 66 56 43 72 63 52 78 70 61 84 77 68 90 83 75

Pagel

0 0 0 0 0 0 0 0 0 0 0 0 0 0 1

24 37 40 57 66

30 35


RELATIVE HLMlDlTY TABLE 2/9/88

1 2 3 4 5 6 7 8 20 85 70 55 40 26 12 0 0 0 25 87 74 62 49 37 25 13 1 0 30 89 78 67 56 46 36 26 16 6 35 91 81 72 63 54 45 36 29 19 40 92 83 75 68 60 52 45 37 29 45 93 86 78 71 64 57 51 44 38 50 93 87 80 74 67 61 55 49 43 55 94 88 82 76 70 65 59 54 49 60 94 89 84 78 73 68 63 58 53 65 95 90 85 80 75 70 66 61 56 70 95 90 86 81 77 72 68 64 59 75 96 91 86 82 78 74 70 66 62 80 96 91 87 83 79 75 72 68 64 85 96 92 88 84 80 76 73 69 66 90 96 92 89 85 81 78 74 71 68 95 96 93 89 85 82 79 75 72 69

100 96 93 89 86 83 80 77 73 70 105 97 93 90 87 83 80 77 74 71 110 97 93 90 87 84 81 78 75 73 115 97 94 91 88 85 82 79 76 74

9 10 0 0 0 0 0 0

10 0 22 0 31 0 38 10 43 19 48 26 52 31 55 36 58 40 61 44 62 46 65 49 66 51 68 54 69 55 70 57 71 58

15 0 0 0 0 0 0 0 0 5

12 19 24 29 32 36 38 41 43 46 47

20 25 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 9 0

15 3 20 8 24 13 27 17 30 21 33 23 36 26 37 28

Pagr 1

30 35 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 7

12 15 18 21

4 i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I Copyright©Saudi Aramco 2009. All rights reserved.

APPENDIX D

SEDIMENT TBANSPORT COMPUTATIONS

Detailed derivations of the equations given below may be found in Madsen and Grant (1976) and Grant and Madsen (1986).

From the wave and current models, one would have the following information at selected grid points:

Waves : R.M.S. Wave Height - H Mean Zero-Crossing Wave Period - T,

Wave Direction - 8 ,

Current : Depth Averaged Current - U,, at z-h/e, where e is 2.7182 Current Direction - 8 , Total Water Depth - h

Sediment: Sediment grain size - d,, (50% of sediment by weight is finer than d50)

For each time-step (1-hour interval) in the tidal cycle, the skin- friction wave and current shear stress r W m and r C , respectively, are obtained as:

T,, - P f,, ub’ . . . . (D.1)

T c = P f,, u: . . . . (D.2)

where p is the density of sea water, f,, is the wave friction factor and f,, is the current friction factor, and U, is the maximum bottom wave orbital velocity which can be obtained using linear wave theory as provided below:

where L is the wave length, and T is the wave period.

To compute rwm and r C in Eq. (D.2), one also needs f,, and f,, which are obtained from the following boundary layer equations [ 6 ] :

D-1


w m and U*

where k, - bottom roughness - 302, - d,,

6,, = KU'm , K = 0 . 4 w

. . . . (D.5)

. . . . (D.6)

. . . . (D.7)

. . . . (D.8)

(D.lO)

where Z, i s the depth a t which U, i s specified

To compute rYm and r C , one only needs t o know f,, and p . Eqs. ( A . 4 ) through ( A . l O ) a r e solved i t e r a t ive ly u n t i l the f i n a l value of p i s within 1% of i t s previous value. p-0 i s used as the s t a r t i n g value.

Maximum combined shear s t r e s s is then given by:

rm - T,, 11 + 2P cos bC + PZj,,

D-2

(D.11)

I I I I 1 I I I I I I I 1 I 1 I 1 I I Copyright©Saudi Aramco 2009. All rights reserved.

I 1 I I I I 1 B I I I I I I I I I 1 I

Shields Parameter

. . . . (D.12)

where ps is the density of sediment, p is density of sea water, g is acceleration due to gravity and d is sediment grain size taken as d5,.

If 3, > the critical value of the Shields Parameter, the sediment moves. The critical value of the Shield's parameter can be obtained from Fig. 5 in Madsen and Grant (1976). A critical value of = 0.05 was used in this study. If it does, the average net transport in {, and q, (in the x' and y') directions are: -

- q x * = 40 32, [3/2 p + p S ] wf d COS dWc . . . . (D.13)

- qyr = 40 3&, [ k p + p 3 ] w,d sin 4wc . . . . (D.14)

Note that the x' and y' are the local direction at a grid point; x' is the wave direction and y' is direction normal to the wave direction.

where 3,, = 3,/[l + 2p cos dc + p 2 I k . . . . (D.15)

- angle between wave and current vectors = ( 8 , - 8,)

wf - fall velocity of sediment obtained from DHI (1985) The direction of the transport vector +s is:

.(D.16)

- Having obtained local { x ~ and q y t at each grid point, the net transport along the grid axes x and y can be obtained which, when used with the finite-difference scheme and the following continuity equation, will yield erosion/deposition rates for a grid cell in terms of the elevation change :

- - ( l - E ) 2 2 + a q x + a q y = 0

a t a, a,

D-3

. (D.17)


The above equation is solved at every hour during the tidal cycle and a q / a t is then accumulated over the tidal cycle. The actual results are provided in terms of erosion/deposition in units of cm/day.

The above procedure can be modified when breaking waves occur causing an extensive surf-zone, as occurs during the 10-year storm. Here, one also needs to account for wave-induced current in addition to wind and tide generated currents. The wave-induced current may be accounted for approximately by using the following formula (Madsen et al., 1978):

V,, - 1 2 . 5 x x sin ( 8 , ) x tan (m) x JgHb . . . . (D.18)

where V, = long shore current which flows in the direction parallel to the causeway

8 , = angle between wave vector and the breaker line (identified from where wave breaking begins using output from the REFRACT program)

m = bottom slope

H, - breaking wave height at the breaker line. The magnitude of the computed wave-induced current for the ten-year storm was found to be very small (-5 cm/s) and was therefore not included in the computation of the bottom shear stress.

D-4

I I 1 I I I 1 # I I I t I I 1 I I I I Copyright©Saudi Aramco 2009. All rights reserved.

Table 2.1. Harmonic constants at points along the 900 n hydrodynamic model boundary. H is amplitude in centimeters, g is phase relative to the lunar crossing at Greenwich in degrees, and IAT is lowest astronomical ti& in centimeters.

STATION IATITUDE IDNGITLTDE IAT la s2 N2 K1 01 Sa P1 K2 ID DEG. HIN. DEG. HIN. (cn) H g H g H g H g H g H g H g H g

1 27

2 27

3 27

4 27

5 27

6 27

7 27

8 26

9 27

10 27

11 27

18.3 -49 14.3

30.0 -49 27.1

35.2 -49 40.2

31.6 -49 45.2

26.4 -49 51.1

20.9 -49 57.1

12.9 -50 07.3

51.8 -49 54.7

19.6 -49 39.8

06.6 -50 13.2

39.0 -49 36.6

107 34 111 14 163 7 75 23 325 17

106 30 114 11 165 6 85 22 319 16

105 30 123 11 175 6 96 21 312 16

104 35 125 13 177 7 98 19 313 15

105 40 126 15 179 8 100 17 316 14

109 45 127 16 183 9 108 19 317 13

109 50 129 16 185 10 110 13 322 12

104 50 126 17 184 0 94 16 325 13

102 46 132 16 192 9 102 18 323 14

105 53 130 17 188 1 112 12 328 10

107 28 121 11 173 6 95 23 313 16

274

269

264

265

266

268

271

271

272

275

264

11 137 6 310 5 170

11 137 7 311 4 174

11 137 7 300 4 185

11 137 6 302 5 187

11 137 6 302 5 185

11 137 6 301 5 183

11 137 5 306 5 181

11 137 5 311 6 173

11 137 5 311 5 178

11 137 4 313 6 180

11 137 8 298 4 183


Table 5.1. Summary of sensitivity runs for salinity.

Identifier

(1) Base

(2) High evaporation

(3) Low evaporation

( 4 ) High dispersion

(5) Low dispersion

(6) Short run

(7) Extended domain

(8) Fine grid

Evaporation Rate* Factor

1.0

1.5

0.5

1.0

1.0

1.0

1.0

1.0

Dispersion Coef. ( d / S )

4

4

4

20

0.8

4

4

4

* Evaporation.rate is given by E - f((a+b) + bcos(wt-m)).

Duration of Run (years)

3.0

3.0

3.0

3.0

3.0

1.0

0.8

0.8

Grid Spacing (meters)

900

900

900

900

900

900

900

300

Domain Size

BASE

BASE

BASE

BASE

BASE

BASE

EXTENDED

BASE


I

I I

Figure 1.1 Mop showing existing causeway and proposed extension. (Extension is shown in red.)


I

I Wave Refraction Hydrodynamic

(REFRACT) (S21 HD) Model Model

I I I I I

I I

Transport Sediment Transport Dispersion Model

Analysis (S21TD) Procedure i

Critical Shields Value I Exceeded?

I I Deposition

I I

I Figure 1.2 Flow chart of the modeling procedure. 1


, , , 4 0 ~ 0 0 , , 49'30 , , , 6O:OO I ,60:30 I ,

SAFANIYA 56/57

W MARJAN 28090 Y-

Z Z U L U F OW-1

2 8'3 0 i 28OOO

27'30

27OOO

28'30

28'00

SAFANIYA PIER

I

4WOO 4@30 60'00 50'30

Figure 2.1 Study area map.

2 800 0

27'30

27000

28'30

26000

25'30

I I I I I I I I I I I I I I I I 1 I I Copyright©Saudi Aramco 2009. All rights reserved.

I I I I I I I I I I I I I I 1 I I I I

I 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58-61 64 67 70 97 94 91 a8 85 8i

Figure 2.2 Bathymetry map on 900 m current model grid.


Figure 2.3 Bathymetry map on 300 m current model grid.

I I I I I I I I I I I I~ I

I~ I

I~ I

I1 I I I I I Copyright©Saudi Aramco 2009. All rights reserved.

I

111.00

101.00

01.00

61.00

71.00

61.00

51.00

+l.W

31.00

21.00

11.W

1 .00 I

Figure 2.4 map on 100 m wave model grid;

causeway with proposed extension. causeway,


0 CURRENT METER STATION I TIDE PREDICTION STATION A WATER LEVEL STATION

Figure 2.5 Location of August 1987 tide and current meter stations.

I I I I I I I I I I I I~ I~ I

I' I

I' I 1 I I Copyright©Saudi Aramco 2009. All rights reserved.

I I I I I I I I I I I I I I I I I I I

n z z 0 I- < > W 2 W W 0 U LL

3 v)

Y

-

a

n I 2 0 i= < > W -I W W 0 < L a 3 v)

Y

ABU ALI PIER

0 25 50 75 iee 125

(1/1/87) TIME (HRS)

BERRI PLATFORM

(11 1/87) TIME (HRS)

PREDICTED TIDE HEIGHT

MODELED TIDE HEIGHT

1 e0 l i s

- - - -_----

Figure 2.6 Comparison of modeled and predicted tide height at Abu Ali Pier and Berri Platform.


I CURRENT METER A

230 235 240 245 250

TIME (HRS)

CURRENT METER B CURRENT METER B

235 240 245

TIME (HRS)

Figure 2.7 Comparison of modeled and measured current speed at (a) current meter station A, and (b) current meter station B. Data from Aramco's August 1987 survey.

I I I I I I I I I I I I I I I I I I

MEASURED MODELED _ _ _ - -------



X X

X

X

Y

Lx Figure 3.1 Schematic representation of the Everett twelve-point finite difference scheme.


Figure 3.2 Location of August 1987 salinity measurement stations.

I I I I I I I I. 1

I 1 I I I


I I 1 I 1 E fi I I 1 I I I I I

h

I + z 5

0 W I- < z 0 I- < a 0 a < > w

5 Y

a

-

40

30

20

10 1 1 1 1 I l l

0.0 2.5 5.0 7.5 10.0 12.5

TIME (MNTH)

Figure 3.3 Computed evaporation rate at Abu Safah GOSP for the year prior to Aramco’s August 1987 survey. Dotted line is a cubic spline fit.

5 1 c I Copyright©Saudi Aramco 2009. All rights reserved.

P W

P W

55

50

45

40

35

55

50

45

40

35

35 40 45 50 65

35 40 45 50 55

50.0

47.5

45.0

42.5

40.0

37.5

MEASURED

D=2 M *IS

D-3M2/S

D=4 M 2/ S

Figure 3.4 Predicted vs measured salinities. Data from Aramco’s August 1987 survey. Numbers on middle panel indicate stations (see Fig. 3.2).


WITHOUT EXTENSION

WITH EXTENSION

CONTOUR INTERVAL 0.2mls

Figure 5.1 Contours of the one-year maximum tidal current speed; (a) existing causeway only, and (b) with causeway extension.


Figure 5.2 Positions for which current speed time history plots were prepared in relation to (a) existing causeway, and (b) existing causeway with proposed extension.


57111ON 2

50 100 I t 0 zee a i 0

11nE 1HRS.I

51171011 3

1.5

5 P

D

1

,

E 1.0

; 0 .5

n

0.0

5 P

D < n 7 0.5

E 1.0

I

0 .0

i 50 190 l i e 290 250

1.5

5 P

D I

, I

1.0

n 5 0.5

0 . 0

Figure 5.3a Time history plots of current speed at select locations for the one-year maximum spring/neap cycle; existing causeway only.


lInE 1URS.J

StaTION 4

. - I I I

S P E E D

I R

5 I

I

TlnE lHRS.1

STOTION 5

TIM. lHR5.1

Figure 5.3b Time history plots of current speed at select locations for the one-year maximum spring/neap cycle; existing causeway with proposed extension.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . -

. . . . . . . . . . .

. . . . . . . . . . . . . . . . . . -

r I I I

I

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .,. * . . . . . . -

Figure 5.4 Salinity contours for the "worst case" scenario. Results from the 900 m grid; (a) existing causeway only, and (b) with causeway extension.


I 34 J7 u) 43 16 49 52 55 58 61 64 67 70 73 76 79 61 85 86 P I 94 97

82 79 76 73 70 67 64 61

58 S5 52 49

46

4J 40 37 34 J1 18 2s 12 19

I6 I 3 IO 7 4 1

81 19 76

7J 70 67 66 61

58 55 s2 49 .a 63

40

J7 3+ 31 18 25 22 19 16

IJ IO 7 4 I

I * 7 IO IJ 16 19 22 25 28 JI 34 17 40 4J 46 49 52 55 Y) 61 64 67 70 71 76 79 82 85 G 3 91 94 97

SCALE 1:12 - I J

Figure 5.5 Same as Fig. 5.4 but for results form the 300 m grid; (a) existing causeway only, and (b) with causeway extension.


20.

16

12 AS (PPt)

8

4

C

JAN 82- MODELED NOV 84 AUG 87 900M

I I I

e

e

e

e

o m

e m

m e

e EX I ST I NG

*PROPOSEC

AS = Salinity Difference, inside - outside bay

Figure 5.6 Comparison of the predicted change in bay salinity due to the causeway extension with changes due to the natural variability as determined from Aramco's "bioaccumulation" data set.



1 I 1 I 1 b I i I 1 I 1 I 1 I 1 I I I

1 , . . . . . . \ - - . I . . . . . . . I . .

\ . \.

. . . ... ! , , , , , . , . . . . . ..\ . I , . . . . . . . \ . - \ , I , \ * . > , . % . \ - - \ I , , . . . . . ,

\ \ - . \ . , . , , . . , . . \ . - \ $ I . . I , . , . . . \-\ . . . . . . . . . . I\-.....,..

....... .... < \ . ..... \ \ I . . \ , . \ ....... , I > \.,> \ *

.I ..

Figure 5.8 Vector plot of wave height amplitude and direction from refraction analysis for the ten-year storm: (a) existing causeway only; (b) existing causeway plus extension.

I


I 5 9 13 17 21 25 29 33 37 11 45 4g 53 57 61 65 69 73 77 81 85 UP 93 97 101 105

Figure 5.9 Erosion (a) and deposition (b) under the one-year storm (existing causeway only).


II E

1 I

I 5 9 I 3 17 21 25 29 33 37 41 45 49 53 57 01 65 69 73 77 81 85 89

I ' I

I Contour Interval 12cm1c

1 5 9 13 17 21 25 29 33 37 41 Is 49 53 57 61 65 69 73 77 81 85 89

'- 55

50

Figure 5.10 Same as Figure 5.9 except for existing causeway plus extension.


Figure 5.11 Net changes in bottom topography due to the causeway extension for the one-year storm.

Bay inside

Vertical Scale is

xapgerated 20 Times

Figure 5.12 Net changes in bottom topography due to the causeway extension for the ten-year storm.


40 35 5 A

’ O t 25

20 ::I 5

0 - .

0 2 4 6 8 10 12 MONTH

Figure C.l Ros Tanura evaporation rate (cm/rnonth), 1960-1 977; Aramco formulotion.

9 MEAN-S.D.

* MEAN+S.D.


15

10

5 I 0

* MIN; NO DELTAT

* MEAN+S.D.; NO DELTAT

* MEAN-S.D.; NO DELTAT

a- MEAN; NO DELTAT

MAX; NO DELTAT

2 4 6 a 10 12 0 MONTH

Figure (2.2 Ras Tanura evaporation rate (cm/month), 1960-1 977; Price formulation - air/sea temperature difference set to zero.

25 .I 20

* MIN

* MEAN-S.0

I- MEAN

* MEAN+S.D.

* MAX

6 8 10 12 0 2 4 MONTH

Figure C.3 Ras Tanura evaporation rate (cm/month), 1960-1 977; Price formulation - &/sea temperature difference based on climatological values for Ras Tanura.

I

I

1 I 1 P I 1 I 1 I 1 I

I I I 1

i


I I I I I

1 I 1 I I I I

I 1 I I I

i

i ..

Plate 1 : Satellite Image of the Study Area (Sept. 4, 19721


I I

I I I I

I I I

I

. . . ,

. . F,..

..L. _... .

Plate 2: Satellite Image of the Study Area (Jan. 2 6 , 1973)

I Copyright©Saudi Aramco 2009. All rights reserved.

~ ~~

I I I I I I I I I I u I I I I I I I I

. .

Plate 3: Satellite Image of the Study Area (Oct. 6 , 1978)


I I I I I I I I I I I I I I I I I 1 I

Plate 4: Satellite Image of the Study Area lJune 6, 19791


I 1 I I 1 I I I I I I I I I I I 1 I I

Plate 5: Satellite Image of the Study Area (May 1 1 , 19841



Plate 6: Satellite Image of the Study Area (Apr. 12, 1985)



Plate 7: Satellite Image of the Study Area (May 4, 1987)


SAER-5347

Documents

aramco use

existing causeway

proposed extension

bathymetry map

transportdispersion

salinity model7

vicinity of abu ali

study area map