Three-Dimensional Hydrodynamic Modeling of the San Francisco EstuaryToward Understanding the Mechanisms Relating Flow to Abundance of Estuarine Biota

Edward S. Gross1 (ed.gross@baymodeling.com), Michael L. MacWilliams2, Daniel J. Schaaf3, Wim Kimmerer4, Renny Talianchich41 Environmental Consultant, 1777 Spruce Street, Berkeley, CA 94709, 2 Environmental Consultant, michael@rivermodeling.com, P.O. Box 225174, San Francisco, CA 94122 3 Schaaf & Wheeler, dschaaf@swsv.com, 100 N. Winchester Blvd. #200, Santa Clara, CA 95050, 4 Romberg Tiburon Center, 3152 Paradise Drive, Tiburon, CA 94920

Model Attributes and Calibration A 200 meter resolution bathymetric grid was developed covering the estuary from the western Delta to the Golden Gate and extending into the coastal ocean. The vertical resolution of the model is 1 meter, resulting in a total of 621,000 active grid cells. The model is forced by observed tides in Monterey and reported dayflow values for Delta outflow. The Sacramento-San Joaquin Delta is represented by two rectangles that were sized to allow approximately correct tidal flow past Rio Vista and Jersey Point compared with UVM data, as shown in the figures below.

The model was calibrated to accurately simulate tidal elevation. In the figure below, the predicted M2 and K1 amplitude and phase are compared with reported M2 and K1 amplitude and phase of observations (e.g., Cheng and Gartner, 1984). The tides propagate accurately in the model. The only notable errors are present in Carquinez Strait and Suisun Bay, where flows can influence harmonic constituents (Lacy, 2000) and, therefore, additional analysis is required. Though the calibration effort is not complete, the results are encouraging.

IntroductionThe purpose of this modeling effort is to investigate the potential mechanisms underlying the relationships of fish abundance to flow (“fish-X2”), which form the basis for the current salinity standard for the estuary. As a first step toward this goal, the three-dimensional TRIM model (Casulli and Cattani, 1994, Gross et al., 1999) has been applied to improve understanding of transport mechanisms in San Francisco Bay. This poster focuses on Carquinez Strait and Suisun Bay where observations indicate vertical velocity shear, frequent stratification and large variability in salinity during the tidal cycle. Furthermore currents and salinity in shoal regions can be substantially different than properties in the channels. Therefore a three-dimensional modeling approach that captures this variability in space and time was applied. The model is being calibrated against hydrodynamic and salinity data and applied to investigate salt transport and movement of biota (see poster by Talianchich et al.).

Salinity SimulationSalinity in San Francisco Bay was predicted during spring and summer of 1994, a relatively dry year, to coincide with the Entrapment Zone Study in Suisun Bay (Burau et al., 1998). The time series graphs to the right compare the predicted salinity at 3 locations in San Francisco Bay to USGS observations (Buchanan et al., 1996). At each station observed salinity at the lower sensor is compared to predicted salinity at the same location during a spring-neap cycle of the Suisun Bay Entrapment Zone study. The tidal time scale variability in salinity is predicted accurately. At Point San Pablo the trends in salinity are most complex due to complex local bathymetry.

The predicted salinity along the axis of the San Francisco Estuary is compared with USGS pilot RMP salinity observations (Edmunds et al., 1995) during four cruises in spring to summer of 1994. The beginning of the simulation corresponds to cruise on March 16, 1994 and observed salinity from that cruise was used to specify the initial salinity in the model. As seen in the figures below, both the seasonal trend of gradually increasing salinity and the degree of stratification are predicted accurately.

Carquinez Mothball Channel Honker Bay

Delta Outflow

Comparison to the 1994 Entrapment Zone Study Observations

Preliminary Analysis of Tidally-averaged Velocity and Salt Transport

Literature Cited

Buchanan, P.A., Schoellhamer, D.H., and Sheipline, R.C., 1996, Summary of suspended-solids concentration data, Central and South San Francisco Bays, California, water year 1994: U.S. Geological Survey Open-File Report 95-776, 48 p.

Burau, J.R., Gartner, J.W., and Stacey, M., 1998. Results for the hydrodynamic element of the 1994 Entrapment Zone Study in Suisun Bay. In Report of the 1994 Entrapment Zone Study, edited by Wim Kimmerer, 13-53. Interagency Ecological Program for the San Francisco Bay/Delta Estuary.

Casulli, V., and Cattani, E. 1994. Stability, accuracy and efficiency of a semi-implicit method for three-dimensional shallow water flow. Computers Math. Applic. 27(4), 99-112.

Cheng, R.T. and Gartner, J.W., 1984. Harmonic analysis of tides and tidal currents in South San Francisco Bay, California. Results of measurements 1979-1980. Open-File Report 84-4339, U.S. Geological Survey, Menlo Park, CA, 1984.

Edmunds, J.L., Cole, B.E., Cloern, J.E., Caffrey, J.M., and Jassby, A.D., 1995, Studies of the San Francisco Bay, California, Estuarine Ecosystem. Pilot Regional Monitoring Program Results, 1994: U.S. Geological Survey Open-File Report 95-378, 436 p.

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

Gross, E.S., Koseff, J.R., and Monismith, S.G., 1999b. Three-dimensional salinity simulations of South San Francisco Bay. Journal of Hydraulic Engineering 125 (11): 1199-1209.

Lacy, J. R., 2000. Circulation and Transport in a Semi-Enclosed Estuarine Subembayment, Ph.D. Thesis, Stanford University.

Oltmann, R.N., 1998, Indirect measurement of Delta outflow using ultrasonic velocity meters and comparison with mass-balance calculated outflow. Interagency Ecological Program for the Sacramento-San Joaquin Estuary, 11(1).

Comparison to the 1994 Entrapment Zone Study ObservationsThe predicted near-bottom salinity is compared with the observed salinity at four continuous monitoring locations on the map above. The model predicts salinity accurately at all locations over this spring-neap cycle.

The eight salinity transects collected during April 27th and April 28th of the Entrapment Zone Study (spring tide) consist of vertical “profiles” of salinity at each of the 10 synoptic sampling stations, numbered 0 through 9 on the map above. The time periods of synoptic data collection are indicated by the gray areas in the graph of observed Martinez tidal elevation. The eight figures below the graph compare the predicted salinity with the synoptic observations for the 8 transects. Both the tidal cycle variability in salinity and the degree of stratification are predicted accurately.

Below the salinity transect figures, predicted velocity is compared with observed velocity at the Martinez station. The observed velocity was measured by an ADCP with a vertical bin spacing of 50 centimeters (Burau et al., 1998), while the model results were available at 1 meter vertical spacing. The predicted velocity matches the observed velocity well in terms of magnitude, direction and phase.

Flux Through Cross-sectionsThe lateral and vertical structure of predicted longitudinal residual velocity is shown, viewed from the east looking west, for three cross-sections: Carquinez Bridge, Mothball Fleet Channel and Honker Bay. The Carquinez Bridge velocity cross-section indicates a gravitational circulation pattern across the width of the section. The Mothball Fleet Channel cross-section shows gravitational circulation across most of the width. In the Honker Bay section, a gravitational circulation pattern is present in the main channel but the residual velocities are landward in Honker Bay, as noted in observations (Lacy, 2000). In all three sections a thin layer is present near the surface in which residual velocities are directed landward. This pattern, known as Stokes Drift, is present because more of the period above mean sea level occurs during flood tide than during ebb tide resulting in net up-estuary flow above mean sea level.

The predicted tidally-averaged salt fluxes roughly correspond to the patterns seen in the velocity with a net up-estuary flux near the bed and in the thin layer above mean sea level of the Carquinez Bridge and Mothball Fleet Channel sections. In the Honker Bay section, gravitational circulation is not an important mechanism for transport of salt during this simulation period. Instead, the analysis has shown that the salinity during flood tides is typically higher than the salinity during ebb tides, leading to landward transport of salt. This pattern is associated with “tidal trapping,” which results from the interaction of oscillatory tides with complex bathymetry and can be particularly strong in areas with side embaymentssuch as Honker Bay and Grizzly Bay (Fischer et al., 1979). Tidal trapping is an important transport mechanism in all of the cross-section locations examined.

The model was applied to provide insight into transport of water and salt. An idealized tide with realistic M2 and K1 amplitudes was specified in the coastal ocean, with a modified M2 period of 12.0 hours so that the same tide pattern repeats each day. A steady Delta outflow of 166 m3/s was used in the analysis. These idealized inflow and tidal conditions allow the model to reach a daily-averaged steady-state, therefore simplifying analysis of model results.

When a steady daily-averaged salinity field was reached, the daily-averaged transport of salt through each cross-section in the model is precisely zero. At this point in time the seaward advective transport of salt resulting from the Delta outflow is precisely balanced by the dispersive landward transport of salt. Under these conditions the velocity and salinity at each grid cell repeat the same pattern each day. The velocity and salinity were averaged over the tidal cycle at each of the grid cells in the model to obtain the daily-averaged velocity and salinity. The top figure shows the vertical structure of the longitudinal residual velocity along the axis of the estuary from the Golden Gate to Rio Vista. From the Golden Gate to the Benicia Bridge, the residual velocity is landward (positive) at the bed with magnitude of up to 20 cm/s and seaward (negative) near the water surface with magnitude of up to 30 cm/s. The gravitational circulation pattern results from longitudinal density gradients and is partially responsible for the presence of zone of increased abundance of estuarine biota, known as the entrapment zone (Burau et al., 1998).

The depth-averaged residual velocity map below indicates that the horizontal structure of the residual velocity is quite complex in SuisunBay, which may be expected expected due to the complex bathymetry in this region.

Implications for ManagementSeveral lessons learned during this work, and related work, may be of interest to resource managers:• Three-dimensional hydrodynamic simulations of the entire San Francisco Bay are feasible on average desktop and laptop computers.• Three-dimensional models resolve important transport processes and therefore require much less empirical information (“tuning parameters”) than lower dimensional models. This

makes them reliable over a larger range of flow conditions and for altered bathymetry.• The strength of various transport mechanisms and their effect on movement of organisms can be explored using a three-dimensional model.• Estuarine circulation is present in San Pablo Bay and Carquinez Strait even during low Delta outflow. This circulation affects movement of organisms and salinity in the Delta.• Channel bathymetry has a large effect on salinity. Therefore, accurate representation of bathymetry is necessary for model accuracy.

Conclusions The model accurately predicts tidal amplitude and phase in the San Francisco Estuary and the tidal flows into the Sacramento River and San Joaquin River match the observed tidal flows measured at the USGS UVM stations (Oltmann, 1998). Ongoing comparisons to tidal velocity observations are also encouraging. The model accurately predicts both tidal time-scale and seasonal trends in salinity in the San Francisco Estuary. Additional comparisons will be made to observations in the 1994 Entrapment Zone Study and during periods of moderate to high Delta outflow. Higher resolution simulations may also be performed to determine if improved resolution of bathymetric features allows significant increases in model accuracy.

A preliminary analysis of tidally-averaged transport was conducted. The analysis indicates that, even during low Delta outflow conditions, gravitation circulation is present in the San Francisco Estuary up to Carquinez Strait. Tidal trapping is also believed to be important for salt transport in all locations in the Estuary during these low Delta outflow conditions.

Implications for modeling transport processes the San Francisco Estuary include: 1) Three-dimensional models have substantial advantages over depth-averaged models for simulating intertidal transport.2) Adequate resolution of bathymetry is important.

Acknowledgements

Funding was provided by the CALFED Ecosystem Restoration Program. Thanks to Jon Burau, Pete Smith, Cathy Ruhl, Vincenzo Casulli, Richard Smith, Mark Stacey, Stephen Monismith, Oswald Lanz, Dave Schoellhamer, Greg Shellenbarger, Dave Ralston, Larry Smith, Ralph Cheng, Jeff Gartner, Jeff Koseff and Lisa Lucas.

Tidal Circulation

N ^

Elevation

legend

Current Station

Tide Station

UVM Station

Bathymetric Grid and Hydrodynamic Data Stations

Sacramento River

Threemile Slough

Dutch Slough

San Joaquin River

Longitudinal Residual Velocity

Depth-Averaged Residual Velocity

Elevation (m NGVD)

Elevation (m NGVD)