1995 Abstracts

Jerad D. Bales, Eric Swain, L.L. DeLong, U.S. Geological Survey, 3916 Sunset Ridge Road, Raleigh, North Carolina 27607, voice: 919-571-4048 fax: 919-571-4041, email: jdbales@dncrlg.er.usgs.gov.

The Natural Systems Model (NSM) was developed in the late 1980's by the South Florida Water Management District (SFWMD) to simulate the hydrology of South Florida for pre-colonization conditions. The NSM includes none of the hydromodifications that have been made in south Florida since European colonization. In addition, the NSM includes estimated topographic and vegetative conditions as they might have been prior to colonization. Because historic climatic data are not available, recent climatic data are applied to simulate the hydrology of the natural system. using algorithms and data from the South Florida Water Management Model (SFWMM).

The SFWMM, initially developed by the SFWMD in the late 1970's and early 1980's, is a regional scale hydrologic model that simulates the hydrology and the highly-managed water system in an approximately 7,600 mi2 (square mile) area of South Florida. Unsteady ground- and surface-water conditions and canal flows are simulated for time-invariant land-use and water management scenarios. The NSM uses the same climatic input data, and similar model algorithms and computational schemes as the SFWMM. Model parameters from the calibrated SFWMM are transferred directly to the NSM. Results from the NSM can be compared with output from the SFWMM for the same set of climatic inputs to estimate the effects of hydromodifications and water management on the natural hydrology.

The NSM cannot be calibrated and tested using traditional approaches. Accurate, detailed information on historic vegetative and topographic conditions, which is required for NSM operation, is largely unavailable. Hydrologic data from the natural system also are unavailable, so model performance cannot be directly tested. The performance of the NSM primarily has been evaluated by using two approaches. First, because the fundamental algorithms used in the NSM are the same as those in the SFWMM, and because the SFWMM appears to perform adequately, it has been assumed that the NSM is properly simulating the important hydrologic processes. Second, a series of sensitivity tests and uncertainty analyses has been performed on the NSM to identify (1) the sensitivity of model output to changes in selected model parameters and (2) geographic areas in which the simulated hydrology is most sensitive to changes in model parameters.

The NSM has been proposed as the "best available tool" for setting hydropattern targets for use in Everglades restoration efforts. Restoration costs may exceed one $1 billion, and water diverted to the Everglades might not be available for a variety of important, competing uses along the southeast coast of Florida, which has a population of almost 4 million. Consequently, decisions made using NSM results could have important and direct implications for the south Florida region.

In July 1995, a study was initiated by the U.S. Geological Survey in cooperation with the Jacksonville District of the U.S. Army Corps of Engineers and with assistance from the South Florida Water Management District. The objective of the study is to determine if the NSM provides a reasonable simulation of South Florida hydrology for pre-colonization conditions, or the natural system, by using recent climatic data. The absence of data from the natural system for model testing requires the application of novel procedures to determine if NSM results are "reasonable." Only selected components and features of the model are to be reviewed because of the limited resources and time available for the review.

Review topics have been formulated as a series of questions addressing hydrologic processes and their numerical representation and calibration issues. In general, topics are being addressed by a combination of (1) development and application of a simplified, but flexible computer code which includes key algorithms of the NSM; (2) application of the new code, NSM, SFWMM, and other appropriate models or analytical solutions for selected conditions, and evaluation of the results; (3) review of scientific literature concerning the formulation of processes simulated by the NSM and (4) documentation of the SFWMM calibration and testing results, and of the findings of experienced NSM users.

Issues related to model representation of hydrologic processes and the numerical approximation of those processes include the following:

· At what spatial and temporal discretization is NSM numerically convergent?

· What is the effect of more refined vegetation information on simulated results?

· What is the effect of spatial variation in rainfall on simulated results?

· How does the delineation of evapotranspiration zones affect model results?

· How are the physical processes of evapotranspiration, channel flow, and ground-water flow represented in NSM?

The specific calibration issues to be addressed are:

· Is the SFWMM calibration unique? If not, what are the implications for the NSM calibration?

· Are there marked differences in the quality of the SFWMM results for low and high flows?

· What are some possible pseudo-methods for evaluating the performance of the NSM?

· Are there geographic areas for which the model appears to perform better than other areas?

This investigation is to provide an independent, scientific review and documentation of key components of the NSM. Results from the study should provide documentation on strengths, limitations, and appropriate applications of the model. Physical processes and (or) computational aspects of the model which should benefit from enhancement, additional research, or improved data also are to be documented.

Boris Galperin, Mark Luther, Meredith Haines, Department of Marine Science, University of South Florida, 140 7th Avenue SouthSt. Petersburg, FL 33701.

Results of modeling circulation and salinity distributions in Florida Bay are discussed. The hydrodynamic model used in this study is an advanced version of the Blumberg - Mellor model in which semi-implicit integration in the horizontal is incorporated. Not only this model affords to increase the computational time step beyond the limit imposed by the CFL condition in the explicit models, but it also accomodates algorithms describing drying and flooding of the coastal areas. These features are very important for Florida Bay because of its shallowness.

The most important diagnostic and prognostic parameters calculated by the model are:

1. Free surface elevation;

2. Horizontal velocity components;

3. Vertical velocity component and its analog in the sigma-coordinate system which is the velocity component normal to sigma-surfaces;

4. Salinity S;

5. Temperature T;

6. Density;

7. Turbulence kinetic energy;

8. Turbulence macroscale;

9. Vertical eddy viscosity;

10. Vertical eddy diffusivity;

11. Two components of the bottom stress vector;

12. Concentration of a conservative tracer C.

The basic equations are cast in a free surface and bottom following, sigma-coordinate system in the vertical direction. In the horizontal, the model uses an orthogonal, curvilinear coordinate system conforming to the coastline. The vertical mixing processes are calculated using a level 2.5 closure scheme by Mellor and Yamada. The subgridscale representation in the horizontal is provided by the Smagorinsky model in which the nonlinear horizontal eddy viscosity depends on grid resolution and the rate of local mean strain.

The model can be executed in the vertically integrated, barotropic mode and fully three-dimensional, baroclinic mode. Despite the use of a semi-implicit horizontal integration scheme, the model is still maximally optimized to efficiently run in a parallel processor environment.

The curvilinear coordinate grid designed for this study is shown in Fig. 1. It contains 72 grid boxes along the east - west direction and 56 points in the north - south direction. On the western side of the model domain an open boundary extends from Cape Sable to Marathon, where forcing functions are prescribed. On the north-eastern side, the grid borders with Barnes Sound. The grid has relatively coarser resolution in its western part adjacent to the west Florida shelf while resolution is increasingly refined towards the north-eastern part where higher gradients are expected due to the topographical constraints, fresh water runoff, etc. The digitized bottom topography for this study was kindly provided by the National Park Service.

Preliminary two- and three-dimensional simulations of circulation in Florida Bay were conducted with the model forced by the tidal elevation along the open boundary using the tidal prediction model. The necessary tidal harmonic constants of 5 tidal constituents, namely M2, S2, N2, K1, and O1 were obtained for Marathon and for the vicinity of Cape Sable (Station M5 of Harbor Branch Oceanographic Institution). Tidal elevations at grids points between the two stations were calculated by linear interpolation. Temporal variations of temperature and salinity data have been prescribed along the open boundary using data at nearest Everglades National Park MMN stations, namely Johnson Key and Peterson Key. Wind data from the nearby Joe Bay station has been collected and used to force the model. Information on freshwater inflows to the Bay was not available so that a constant inflow distributed over 65 grid points along the north-eastern part of the grid has been assigned along the south Florida mainland to determine its impact on the hydrodynamic and salinity regimes of Florida Bay.

Some calibration runs have been made for the period of September 1 to December 31, 1993. Data during this four month period were available for comparison with model results. Generally, simulations in the vertically-integrated mode agreed well with the fully 3D simulation. The best agreement was achieved for free surface elevation. It was found that some parts of Florida Bay are consistently drying and flooding during the tidal cycle; work is being done to better identify and quantify this phenomenon.

At the present time, simulations are being conducted with various values of fresh water runoff, to understand the effect of the runoff on circulation and salinity distributions in Florida Bay. Results of these simulations will be presented at the Florida Bay Science Conference.

Thomas N. Lee, University of Miami/RSMAS, 4600 Rickenbacker Causeway, Miami, Fl 33149; Elizabeth Johns, NOAA/AOML, Miami, Fl 33149.

A new two year observational study of the interaction and exchange of Florida Bay with the connecting coastal waters of the Gulf of Mexico and the Atlantic in the Florida Keys will start­up in early winter 1995 with support from NOAA/COP Florida Bay Program. The research is designed to address several of the key scientific questions presented in the NOAA Florida Bay Implementation Plan as critical to understanding the functioning of the ecosystem and possible future evolution from restoration actions. In particular, the research will address the following questions:

1) To what degree is the circulation of water within Florida Bay coupled to that of the surrounding coastal and oceanic environments?

2) What is the relationship of surface and groundwater flows through the Everglades to the salinity of Florida Bay?

3) Is the quality of the water flowing from the Bay contributing to the degradation of corals along the reef tract of the Florida Keys in the Atlantic Ocean?

Observational methods consist of a combination of synoptic shipboard surveys, in­situ moorings and Lagrangian surface drifters to describe and quantify the circulation within the Bay as related to local forcing and coupling with the waters of the Atlantic and Gulf. These data will be used to determine the rates and pathways of material exchange across the open boundaries to the Bay, which are needed to understand the transport and exchange of planktonic communities with nearby coastal environments, and the recruitment of larvae and young juveniles to the Bay's nursery grounds. These new observations will also provide necessary boundary conditions for future physical and biological models.

George A. Maul, Division of Marine and Environmental Systems, Florida Institute of Technology, 150 West University Boulevard, Melbourne FL 32901-6988.

Numerical models of Florida Bay circulation require the highest possible spatial resolution in order to determine details of the three-dimensional flow. Accordingly, they are nested within lower spatial resolution local- and/or regional-scale circulation models to provide the boundary conditions associated with ocean forcing external to the Bay area. Dynamic consistency of the lower spatial resolution models must be tested by independent observations; satellite measurements of sea surface height (SSH) variability is investigated herein for that purpose.

Altimetry from TOPEX/Poseidon is considered the best means of estimating SSH variability with the current crop of spacecraft. The altimeter measures the range from the orbiting vehicle to the sea, and precise tracking measures the position of the satellite. Differencing these two measurements, and correcting for errors, gives SSH typically within (± 5 cm. However these SSH measurements are the sum of the geoid (which has variability on the order of (± 90 m globally) plus the ocean (which has variability of (± 100 cm globally). Since the geoid, which for this study is fixed in space and time, is not known to the precision required for model verification, only the variability in SSH is conveniently available.

TOPEX/Poseidon data from the Intra-Americas Sea (IAS), that area of the western Atlantic which includes the Gulf of Mexico, the Caribbean Sea, the Straits of Florida and the Bahamas, are being processed to determine SSH variability. Once the three-year mean SSH is computed (1992-1995), the variability for the IAS every 10 days can be summarized. A similar calculation from the University of Miami IAS sigma-coördinates numerical model will then be compared with the SSH, and a difference field prepared. Analysis of the case-by-case and ensemble differences give vital clues as to the ability of the Miami model to replicate variability in the IAS region.

In addition to verifying models, the SSH data per se have intrinsic applicability to Florida Bay science. For example, it is now thought that variability of the Gulf Loop Current, which is the source of the Florida Current, can be traced upstream to perturbations in the Caribbean Current. Thus some of the events in Florida Bay may be associated with happenings external to the Straits of Florida and indeed to the IAS itself. It is these teleconnections that investigation of SSH from TOPEX/Poseidon is expected to clarify as well as (ultimately) providing constraint to numerical models themselves.

  A Mass-Balance Model of Salinity in Florida Bay: A Tool for Research and Management

W.K. Nuttle, J. Fourqurean, Florida International University; B.J. Cosby, J.C. Zieman, University of Virginia.

FATHOM is a 2-D, transient, mass balance model for water and salt in Florida Bay. The

term "model" does not, however, entirely describe the focus of this project. Our goal is not simply to produce a computer code that can simulate present conditions and forecast future conditions in Florida Bay. Rather, toward the broader goal of understanding of the processes controlling water salinity and circulation in Florida Bay, we are working to produce a conceptually simple, computationally efficient code that is accessible to a broad audience of researchers and managers.

Background

A simple accounting of the water budget for Florida Bay reveals the root cause of episodes of high salinity, which have been linked with perceptions of a general ecologic decline. In an average year direct evaporation from the Bay entrains and concentrates sea water from the Gulf of Mexico resulting in generally hypersaline conditions. Evaporation is by far the largest flux in the water budget; the precise figure is not known, but it is believed to be between 150 cm/year and 210 cm/year over the 2.2 109 m2 surface of the Bay. During the period 1980-1989 by comparison, direct precipitation on the Bay was around 120 cm/year, and freshwater runoff via Taylor Slough and the C111 Canal was only about 8 cm/year on average. Precipitation and freshwater runoff may exceed evaporation for short periods within a year, reversing the direction of net advective exchange with the Gulf, flushing hypersaline water from the Bay. In rare wet years, this flushing can maintain salinities below 30 ppt throughout the Bay for extended periods; such an event is currently underway.

At slightly finer temporal and spatial resolution, patterns of salinity in the Bay are determined by hydrodynamic mixing and stochastic variations in freshwater fluxes. Under normal, hypersaline conditions, salinity is determined by the rate of hydrodynamic mixing, which drives a seaward flux of salt by dispersion. This mixing is probably dominated by tides, including residual currents. However the relative contribution of lower frequency processes, i.e. direct and remote wind forcing and sea level fluctuations, has yet to be established. During extended periods of excess freshwater flow over evaporation, the net flux by hydrodynamic mixing reverses to balance advection of salt, now out of the Bay. The occurrence and intensity of these freshet events depends critically on the magnitude of variations in freshwater flows and their timing relative to the variation in evaporation.

Objectives

The main question facing managers is "To what extent have hydrodynamic mixing and freshwater flows been altered by human activities in the Bay and upstream in the Everglades?" A related question is "What will be the effect of future activities?" These questions cannot be answered without some basic research. The objectives for research must be 1) to identify processes that control hydrodynamic mixing and the magnitude and timing of variation in the water balance fluxes in Florida Bay, and 2) to establish the degree to which these processes are sensitive to activities that have occurred and those that may occur in the future. Widely recognized is the need for a model that can be used to interpret data related to mixing and circulation in Florida Bay. A calibrated version of such a model could be used by managers to explore implications of the available data. The model must be physically-based, in order to justify extrapolation from current data, and an uncomplicated model structure is justifiable due to the limited quality and quantity of data describing Florida Bay. Also, a conceptually simple model is more accessible to a broad audience of scientists and managers.

Modeling Approach

FATHOM can be understood as a refinement of the mass balance accounting described above for all of Florida Bay. In the model, the Bay is divided into approximately 40 cells based on the location of prominent shoals and banks. A running account is kept of the mass of water and salt in each cell; precipitation is added, evaporation is lost, and fluxes of water and salt between cells are calculated as described below. The cells are assumed to be well mixed for the purpose of estimating salinity. In addition, freshwater runoff is added to cells along the northern boundary of Florida Bay, and water levels are varied to simulate tides along the Gulf of Mexico and Atlantic Ocean boundaries. Hydrologic fluxes vary monthly following the mean annual cycle for each component, and tides vary harmonically with semi-diurnal, fortnightly and annual cycles represented.

A key assumption of our approach is that mixing is instantaneous within cells and that mixing between cells is by tide-driven advection across the intervening banks, which are also the primary control on water movement. Fluxes of water and salt across a bank are calculated from the instantaneous water level in each cell, the depth distribution along length of the bank and the salinity of the upstream cell. Velocities on either side of the bank are assumed to be zero; therefore the difference in water levels is the potential energy, or head, available to the flow. This head difference is partitioned between energy loss due to bed friction and velocity head, which is dissipated by flow into the downstream cell. If the head difference across the bank is extreme, or if the water level in the downstream cell drops below the top of the bank, the bank will behave as a weir with critical flow occurring at the downstream edge. These conditions are sufficient to calculate a flux per unit length of bank for any combination of upstream and downstream water levels and depth distribution.

FATHOM runs on a moderately endowed PC. An interactive shell provides the user with access to all inputs and parameters to set up and run a simulation, and the same shell allows for selective viewing of simulation results. Bathymetric data for the cells and bank are derived from a GIS file of digitized elevations, which can be updated readily as better data become available. The user can also change the configuration of the cells and the depth distributions along individual banks prior to running simulations. Files of daily values of runoff, precipitation and evaporation can be attached as an alternative to the monthly average values generated internally.

Summary of Results to Date

A series of baseline simulations have been performed of the annual cycle of monthly salinity distributions in Florida Bay. Simulation of an "average" year is based on average monthly hydrology in the region for the period 1980 to 1989 and the "average", predicted tides. "Wet" year and "dry" year simulations are based on varying the precipitation and runoff into the Bay by 50 percent. For comparison to actual, observed conditions, we have identified "average", "wet", and "dry" years in the monthly water quality monitoring data covering the period 1989 to present. Overall, the results are very encouraging. With essentially no calibration, FATHOM reproduces the overall range and the broad spatial pattern of observed salinity.

Outlook For Remaining Work

The existing code is inadequate for the experimental studies for which FATHOM is intended. For example, the current version of the model is not parameterized to include the effect of wind stress on circulation in the Bay; neither has the tide-driven mixing in the model been calibrated against historical data on water levels and salinity. These and other deficiencies were largely a concession to the amount of time allotted for the development of the code and lack of readily available data. The recognized deficiencies must be addressed before FATHOM can be used to investigate the questions that are of direct interest to resource managers.

There are four objectives for further work. First, we propose to address recognized deficiencies in FATHOM by extending and refining the computer code. Second, we propose to use the model to investigate some of the issues relating to the management of water fluxes into Florida Bay. This will entail 1)a systematic exploration of the influence of various factors on salinity in the Bay through a sensitivity study conducted with the model, 2)calibration of the model against the available historical data, and 3)exploration of the impacts of changes in the morphology of the basin and freshwater flows. Third, we propose to integrate FATHOM with an existing program to monitor water quality in Florida Bay. Finally, we foresee further development of the model to allow simulation of nonconservative solutes, i.e. nutrients, in the Florida Bay system.

As part of a monthly water quality monitoring program, FATHOM could be used to "forecast" changes in water quality based on the latest water quality observations in the Bay and data on runoff and climate for the intervening period. The model forecasts would anticipate the onset of critical patterns of salinity and other events of interest, the occurrence of which could be tested by the next set of data collected. Over time, experience accumulated from a number of cycles of model extrapolation and testing by observation will inevitably lead to insights suggesting improvements to both the model and the monitoring program and possibly new areas of research.

Y. Peter Sheng, Justin Davis, Yingfeng Liu, Coastal & Oceanographic Engineering Department, University of Florida, Gainesville, Florida 32611.

A preliminary modeling study on the circulation and transport in Florida Bay has been conducted. The study includes: (i)A preliminary review of the available data from Florida Bay, (ii)A preliminary model simulation of Florida Bay circulation, and (iii)A recommendation on the development of a comprehensive model of Florida Bay which can be used to aid the management/restoration of Florida Bay.

A set of high resolution (20 meters x 20 meters) bathymetric data for Florida Bay were obtained by the Everglades National Park. A comprehensive set of data were also obtained at numerous stations in Florida Bay 1993 and 1994 by the Everglades National Park. These data include water depth, temperature, conductivity, and rainfall. On the contrary, evaporation data were obtained at 1-2 stations only. There were only limited wind and tide data from the offshore waters to the south of Florida Keys. Freshwater discharge data are particularly lacking. Although discharge data are available at 3-4 flow structures upland from the Panhandles, there exists little data to indicate how the freshwater flows into Florida Bay.

Using the data indicated above and a 3-D curvilinear-grid model developed by Sheng (1987, 1989, 1994), we conducted preliminary model simulations of tidal, wind-driven and density-driven circulations in Florida Bay. The simulation results, obtained in the curvilinear grid shown in Figure 1 indicated that the model is capable of simulating many of the observed circulation features in Florida Bay, including tidal amplitude (Figure 2), tidal phase, residual circulation, hypersalinity due to evaporation (Figure 3), and lowering of saline due to freshwater inflow. Two other circulation Natures in Florida Bay: the flooding and drying of mudbanks during periods of high wind and tide and the effect of mangrove vegetation on circulation, can also be simulated by the model Al tough results are not completed for Florida Bay simulations.

To develop a comprehensive circulation model of Florida Bay, it is feasible to include all of the above mentioned model features in the 3-D curvilinear grid model used in this study or another similar circulation model. In addition, it is necessary to incorporate a robust air-sea interaction scheme into the circulation model to improve the estimation of wind stress, heat flux, rainfall, and evaporation flux at the air-sea interface. One possibility is to couple the circulation model with a regional scale atmospheric circulation model which includes a robust marine boundary layer model.

Florida Bay circulation is also influenced by (l)freshwater inflow from the Everglades area to the north of the Bay, (2)tides and circulation in the Western Florida Shelf, and (3)tides and circulation in the Florida Strait. A study, which may include the use of a groundwater flow model, is needed to quantify the freshwater inflow into the Bay. A comprehensive modeling study to incorporate the influences of Western Florida Shelf and Florida Straits on Florida Bay is presented in this report.

The Florida Bay circulation model can be coupled to a water quality model and a seagrass/light attenuation/epiphyte model to study the effect of altered freshwater discharge schemes on the circulation, water quality and seagrass in Florida Bay. Such a coupled model has been developed for other shallow estuaries in Florida. Similar approach may be adapted for the development of a comprehensive modeling system for Florida Bay.

Acknowledgment

The work reported herein has been supported by the National Park Service, Everglades National Park and Dry Tortugas National Park. Robert Brock served as the Project Officer. Dewitt Smith provided all the hydrodynamics and hydrologic data. Jim Fourqurean and David Rudnik provided the water quality data. Ned Smith provided his results of harmonic analysis.

Ned P. Smith, Harbor Branch Oceanographic Institution, 5600 U.S. Highway 1 North, Fort Pierce, Florida 34946.

Historical data obtained since the early 1980s from waters surrounding Florida Bay provide a valuable regional framework for designing studies conducted within the bay itself, and for interpreting data collected from those studies. An eleven-month time series of mid-depth current speeds and directions was obtained at a study site 48 km west of Cape Sable during 1984-5. Results indicated well-defined tidal ellipses with the major axes oriented in an

east-west direction. Low-frequency nontidal flow past the study site was consistently toward the southeast or east-southeast, i.e., directly into Florida Bay. The long-term resultant current speed was approximately 4 cm s-1.

Current meter data from the 81°05'W meridian were collected as part of a Florida Department of Environmental Protection sponsored study to define regions of nontidal inflow and outflow along the western boundary of the bay. During a 12-month period starting in March 1994, the net flow patterns at three locations equally spaced between East Cape and Marathon were quite different. At the northern study site (25°05'N), nontidal outflow was recorded from mid June to early October. This was followed by a period of persistent inflow from mid October through the end of the record in early April. At the central study site (24°57.5'N), a quasi-steady flow into the bay was recorded for the entire 386-day time period. Flow past the southern study site (24°50'N) alternated between south-southwestward and northwestward, but the resultant flow was west-northwestward, out of the bay. Results of this study confirm an active and persistent exchange of water between Florida Bay and shelf waters of the eastern Gulf of Mexico, but the spatial variability cannot be resolved with only three study sites.

Along the eastern and southeastern fringe of Florida Bay, long-term net flow patterns have been established by mooring current meters in the major tidal channels that connect Florida Bay with Hawk Channel on the Atlantic side of the keys. Seven tidal channels in the Upper and Middle Keys have been investigated over the past five years. Results from Tavernier Creek are inconclusive at this time. During a 97-day study period from mid June through mid September 1994, the net flow was into Florida Bay from the start of the study through mid August, then out of the bay through the end of the study period. The data may reveal part of a seasonal pattern, but the net flow past the current meter for this specific time period was nearly zero.

Data from Snake Creek and Whale Harbor Channel collected during the latter part of 1994 and early 1995 reveal an anomalous, though probably local condition in which the long-term nontidal flow is out of Hawk Channel and into Florida Bay. Data from Snake Creek, collected from July 1, 1994 to January 3, 1995 indicate an inflow with a resultant speed of 5.4 cm s-1. Current meter data from Whale Harbor Channel (adjacent to Snake Creek on the Key West side) collected from September 20 to December 13, 1994 show a net inflow with a resultant speed of 7.6 cm s-1.

Data collected within the past several years at all other major tidal channels show a net outflow from Florida Bay. A 118-day time series (September 22, 1994 to January 18, 1995) from Indian Key Channel revealed a resultant 4.4 cm s-1 outflow into Hawk Channel. Similarly, a 178-day time series (January 28 to July 25, 1994) from Channel Two indicated a 4.1 cm s-1 resultant outflow. Data from Channel Five collected between August 3, 1990 and January 3, 1991 showed a resultant outflow of 2.9 cm s-1, but this record appears to include two components of a seasonal cycle. Outflow during the first part of the study period was over three and a half times faster than outflow from late September through the end of the record. The longest time series come from Long Key Channel, and these records are well suited for defining low-frequency variability over seasonal time scales. A 415-day time series (July 1, 1992 to September 9, 1993) showed a resultant outflow of 2.6 cm s-1, but outflow was strongest during winter and early spring months. More recently, a 319-day study from the same location in Long Key Channel indicated a resultant outflow of 4.9 cm s-1, but again the strongest outflow occurred in winter months.

With the above as background, a one-year circulation study was initiated within the Everglades National Park part of Florida Bay. The principal objective of the study was to quantify tidal exchanges and the long-term net transport through three major tidal channels connecting sub-basins in the western part of the bay. Current meters were moored in Conchie Channel (starting July 26, 1994), Iron Pipe Channel (starting August 9, 1994) and Man of War Channel (starting August 30, 1994). Water depth in all three channels is approximately 3 m, and current meters were moored just below mid depth and in mid channel. Current meters record speed, direction and water temperature hourly. The analytical methodology includes harmonic analysis to quantify amplitudes and local phase angles of the principal tidal constituents, the calculation of cumulative net displacements past the current meter, and the calculation of total volume transport through the channel cross-section. Displacement is calculated by multiplying the along-channel current speed by the one-hour time interval it represents. Net displacements are obtained by defining outflow to be positive, and cumulative values are obtained by summing hourly displacements.

The study design calls for leaving the current meters in place until mid September, 1995 to obtain a one-year record from each location. Data collected to date shows a persistent nontidal out-flow at each of the three study sites. Outflow from Conchie Channel (July 26, 1994 to April 5, 1995) averaged 2.3 cm s-1 during this 253-day time period. Low-pass filtered current speeds are generally within the range of 0-5 cm s-1, and low-frequency flow into the bay for more than a day or two at a time is rare. The co-oscillating tidal current superimposed onto the net outflow is substantially stronger. The amplitude of the semidiurnal M2 constituent is 32 cm s-1. Results from Iron Pipe Channel are significantly different only in the sense that both the resultant speed and the amplitude of the M2 tidal current are much higher. During the 239-day period from August 9, 1994 to April 5, 1995, the mean outflow averaged 13 cm s-1. Harmonic analysis of the time series indicates an M2 amplitude of 56 cm s-1. Low-frequency nontidal flow is generally between 10 and 20 cm s-1, and low-frequency inflow appears only five times during the study period. Data from Man of War Channel indicate a weaker resultant outflow of just under 9 cm s-1, but the amplitude of the M2 tidal constituent is 57 cm s-1. Low-frequency nontidal currents are largely between 5 and 15 cm s-1.

The remaining work for this study focuses upon the causes of the observed net outflow through the three tidal channels in the western part of the bay. It is logical to assume that some fraction of the Bay-to-Gulf outflow is a response to regional wind forcing. Similarly, it is likely that at least a portion of the outflow may be traced back to fresh water entering Florida Bay along its northern boundary. A third possibility involves a response to tidal forcing. The tidal "pumping" effect (the difference between transport during the flood tide and transport during the ebb tide) results in a net transport in the direction of the tidal current at the time of high water. Along the western side of Florida Bay, highest water levels coincide closely with strongest flood tide speeds. The magnitude of residual tidal transport is inversely related to water depth, because the frictional force resisting flow at low tide is especially large when the water is shallow. Thus, one would expect a pronounced transport into the bay associated with tidal waves converging into the shallow interior. Tide-induced transport across the shallow mud flats will set up water levels in the interior, and tidal channels will serve as a path of least resistance for the compensating return flow. This process will enhance the ebb tide in tidal channels that connect the interior of the bay with adjacent coasal waters. If the low-frequency outflow through the tidal channels is coherent with the time-varying amplitude of the tide along the western boundary of the bay, this will support the alternative hypothesis that the net outflow may be a local return of water pumped into the bay over each tidal cycle.

In the final weeks of this one-year field study, we will attempt to sort out the relative importance of wind and tidal forcing. Current meter data will be analyzed with wind stress data (calculated from observations made at the C-MAN weather station northwest of Long Key), and the coherence of wind stress and nontidal outflow will be compared with the coherence of tidal forcing and nontidal outflow. Results of the Florida Bay circulation study should put in sharper focus the relative importance of forcing by winds, tides and (as an unexplained residual) surface runoff. The relative importance of alternate forcing mechanisms will be useful in deciding what kind of forcing has to be included in any modeling effort that is to be undertaken to guide management decisions.

S.L. Vargo, J.C. Ogden, Florida Institute of Oceanography, 830 First Street South, St. Petersburg, Florida 33701, USA; J.C. Humphrey, Keys Marine Laboratory, P.O. Box 948, Layton, Florida 33001, USA .

The Florida Institute of Oceanography (FIO) operates six enhanced Coastal Marine Automated Network (C-MAN) SEAKEYS stations designed and installed under a cooperative agreement with the NOAA National Data Buoy Center (NDBC) as part of a major long-term monitoring program encompassing the 220 mile long Florida reef tract. The SEAKEYS stations are located at five sites along the reef tract (Fowey Rocks, Molasses Reef, Sombrero Reef, Sand Key, and the Dry Tortugas) and one site in Florida Bay approximately 2.25 miles NW of Long Key (24° 50' 36" N, 80° 51' 42" W). The Florida Bay station was installed in November 1992 and is located immediately adjacent to the southwest boundary of Everglades National Park. The SEAKEYS stations record meteorological (wind speed, direction, and peak gust; barometric pressure, air temperature, and solar irradiance) and oceanographic (temperature, salinity, and submarine irradiance at 1 and 3 m water depth) data. The Florida Bay station has one set of oceanographic sensors at 1 m depth as the overall water depth is only 2.2 m. The data are transmitted from the stations via GOES satellite, are formatted, verified, and archived by the NDBC and are downloaded by the FIO in near real-time by telephone modem from the Data Collection Automatic Processing System (DAPS) in ASCII machine language and processed with D-Base software. Data are also made available to more than 30 institutions, agencies, and individuals via fax and Internet under a cooperative agreement with the NOAA Atlantic Oceanographic and Meteorological Laboratory (AOML) in Miami.

During the first year of operation (11/6/92 - 12/31/93) data from the Florida Bay station documented a number of environmental events. During November and December 1992, the water clarity index decreased to 10% due to the presence of the algal bloom which developed in Florida Bay after Hurricane Andrew. Water clarity increased rapidly as the bloom moved off the site, then declined rapidly again when the bloom returned on 12/10/93. Cold fronts passing through the area resulted in drops in both water temperature ( 4°C) and salinity ( 4 o/oo). These changes in temperature and salinity persisted for 4-7 days after passage of the front. These rapid declines in water temperature (>5°C over 24 hrs) were again apparent during the "Storm of the Century" on March 13-14, 1993. In the summer of 1993, temperatures ranged from 29 - 33° C and salinity from 33 - 37.5 o/oo. Again wind events were very important. Temperature decreased with increased wind speeds. Salinity was not strongly affected by increased wind speeds. A 4.5 o/oo decrease in salinity in mid-September is probably related to the Mississippi flooding in late summer 1993. Drops in salinity during this period were found at all the SEAKEYS station but the pattern was complex.

Comparing the data from 1992-1993 with the data from January 1, 1995 - August 26, 1995 similar patterns are apparent. Salinities ranged from 28 - 36 o/oo and temperatures from 15 - 33°C. Salinities and temperatures were lowest from January - March 1995. Overall the salinities were lower than those recorded 1993 (range 33 - 39 o/oo). These decreases in salinities may reflect the increased rainfall in south Florida during 1995. It is important to note, however, that despite the increased rainfall salinities were consistently higher (33-36 o/oo) from June - September 1995 during the peak of the rainy season than during the drier period from January - March (28 - 34 o/oo). Due to the areal extend and shallow depths in Florida Bay, evaporation at higher air and water temperatures in the summer may be a major factor influencing salinity. Wind events were also important. Increases in wind speed were followed by decreases in water temperature in both winter and summer. Again there was no apparent effect of wind on salinity.

The data from the SEAKEYS Florida Bay station demonstrates clearly the highly variable physical and biological factors in this shallow water region which must considered in developing monitoring plans to detect environmental changes which may result in management actions.

John D. Wang, Applied Marine Physics, Rosenstiel School of Marine and Atmospheric Science University of Miami; Thomas N. Lee, Meteorology and Physical Oceanography, Rosenstiel School of Marine and Atmospheric Science, University of Miami.

Introduction

It is widely believed that water quality conditions have changed gradually in Florida Bay over the past several decades. The recent extensive sea grass dieoffs and algal blooms are indications of such changes. When seeking causal processes for these changes, reduced runoff from the Everglades and pollutant influx from the west inevitably are raised as significant problems. Resolution of these issues requires that the hydrodynamics circulation response in the Bay be quantified. The transport fluxes across the western boundary of the Bay and through the channels between the Florida Keys are potential major input and output functions and deserve particular attention.

With support from the Florida Department of Environmental Protection, we are presently in the midst of characterizing the current and salinity patterns in western Florida Bay. This two year project began in April 1994.

Objectives

The objective of the study is to collect data on the tidal and longer term transport patterns and fluxes along the western boundary of Florida Bay and in Long Key Channel. The data collection is a cooperative effort with Ned Smith and collaborators of the Harbor Branch Oceanographic Institute (HBOI). HBOI has maintained three recording current meters along 81°05'W and one in Long Key Channel for a year long period. Our contribution is to define the spatial and temporal variability of salinity and currents along the 81°05'W transect and to determine lagrangian trajectories. This additional information will greatly extend the usefulness of the current meter records and will allow transport fluxes to be determined with greater reliability. The results will be made available for calibration and performance evaluation of numerical hydrodynamics models of the Bay.

Methods

The study area is roughly confined to the Bay between 81°05'W and 80°50'W. The shallow water and importance of wind make accurate current meter measurements difficult. Interpolations of current meter records to compute transport is also difficult without additional information on spatial and temporal variability in currents. The 81°05'W transect was selected for installation of current meters, because the relatively deep and smooth bottom topography would more likely be associated with gentle velocity variations, and thus, would allow the current measurements to be interpolated more accurately along the boundary.

To help define spatial and temporal variability, drogues were deployed at a number of stations along the 81°05'W transect during surveys in August - September of 1994, and January - February of 1995 to sample both the summer and the winter season regimes. With vanes set at three different depths, drogues were set free simultaneously and tracked over a 30 minute period using a GPS. The brogue vanes were 25 cm high and the center of the vanes were set at depths of 20 cm, 100 cm, and 200 cm. The drogues were used in place of direct current meter measurements because of the difficulty of deploying these from a small boat in choppy seas.

Satellite tracked drifters were deployed at several locations near the 81°05'W transect and left in the water for extended periods. Of 9 drifter deployments, 6 were picked up by fishermen (and recovered), 1 stranded on Conch Key, and 2 moved offshore through Long Key Channel. The longest undisturbed deployments were for about 1 month.

In recent months we have substituted the cumbersome brogue measurements with current profile transects using an RD Instruments 600 kHz broadband Acoustic Doppler Current Profiler (ADCP). When mounted on a small boat, this instrument can collect water column current profiles while moving along a transect at up to about 4.5 knots (2.2 m/s). Transect location is obtained with a GPS. We have successfully used this instrument in depths as shallow as 1.5 m, obtaining velocities in 25 cm vertical bins about once every minute. The Long Key Channel has been surveyed on two separate days on July 21 and August 31, 1995. The northernmost and southernmost sections of the 81°05'W transect have also been surveyed on July 27-28, and August 10-11, 1995.

Salinity is often an excellent natural tracer. In 1993 lower than usual salinities helped the detection of a plume containing large amounts of water from the great Mississippi flood along the Keys and in Florida Bay. As part of the present project, we have developed an underway sampling system, which measures near surface temperature and salinity. Sampling has effectively been done at speeds up to 18 knots (9 m/s) allowing large areas to be mapped. The sampling system consists of a Sea Bird SBE-21 thermosalinograph through which water is pumped at a rate of about 0.5 liter per second. GPS location data is added directly to the data stream from the SBE-21 using a NMEA interface box also manufactured by Sea Bird. A special bow-mounted water intake system is designed to allow high speed sampling with minimal introduction of air bubbles. The instrument sampling rate is set at 5 sec intervals.

The study area is covered by two almost 100 nautical mile long transects, which are sampled on consecutive days. By repeating these surveys at a few days interval, nearly synoptic maps of salinity are obtained. Net transport velocities can be derived from the variations in salinity patterns, and can be compared with our other observations.

Conclusions and future work

Analysis of the data and interpretation using numerical models is ongoing. Some preliminary results are listed below.

On many occasions a salinity minimum was found between least and Mid Cape (Cape Sable). Although, a number of explanations are possible, the most likely is a local source of freshwater.

The satellite drifter tracks show high visual correlation with wind direction.

Salinity maps are very useful when significant salinity gradients are present. Poor weather conditions during winter cold front passages makes it difficult to collect data closely spaced in time complicating interpretation. Although, very helpful in western Florida Bay, the salinity mapping approach would be even better suited for eastern Florida Bay due to the smaller areas extent, generally calmer sea state, and larger salinity variations.

Drogue velocities are in general agreement with current meter observations, but also indicate vertical shear effects. An additional salinity mapping period is planned for the fall of 95. Additional data quality control and analysis is also planned.

John D. Wang, Applied Marine Physics, Rosenstiel School of Marine and Atmospheric Science, University of Miami; Charles Monjo, Applied Marine Physics, Rosenstiel School of Marine and Atmospheric Science, University of Miami.

Objective

This recently completed study, supported by the Everglades National Park, is aimed at exploring the potential of a 2-dimensional nearly horizontal flow model as a tool for describing the circulation in Florida Bay. Some of the issues that the model should be helpful in addressing are the resultant changes in circulation and salinity to be expected from modifications to the freshwater runoff volumes from the Everglades, and the exchange between the Bay and the Gulf of Mexico or the Atlantic Ocean.

Methods

Several factors are considered in choosing the hydrodynamics model for Florida Bay. The primary factors are the extreme spatial complexity, shallow depths, relatively small freshwater input volumes, and the sparse set of field observations.

The study to be presented intends to define factors and processes that are important for hydrodynamics modeling of the Bay. It emphasizes methodology and data needs, but puts less emphasis on obtaining realistic (accurate) simulations, since there are not sufficient data for prescribing model forcings or for verification of model results.

The model (CAFEX) chosen for adaptation to Florida Bay is a finite element numerical solution of the vertically integrated hydrodynamics equations of motion (Wang and Monjo, 1995).

The bottom topography is obviously a very important model input for the shallow and spatially complex Florida Bay. The bottom topography data base was obtained by digitizing the NOAA/NOS Chart 11451 on a regular 15"x15" grid. The digitization in most cases simply selects the average depth value around the grid point location, but in some cases is adjusted to better represent the effective conveyance in channels between banks or islands.

The model topography is derived from the digitized data base with corrections for datum variations and subgrid scale topography. Consequently, several of the smaller islands are represented as very shallow areas in the model because of their size. We have not attempted to fine tune the model topography to features, which are of the order of the grid point spacing or finer. Such fine tuning will be more productive, when grid resolution can be refined, and when more current observations are available.

The drying and wetting of semi-submerged banks, because of their expansive extent, are likely to be extremely important processes controlling the circulation patterns in Florida Bay. The banks block flow entirely, when the water level is sufficiently low, and bank overflows are subject to strong nonlinear effects due to friction. These processes are incorporated into the model.

Because of the time consuming and data intensive effort to include the actual geometry and topography of the Keys channels, these have simply been included as openings in the model. The size of the openings somewhat reflect the actual width of channels, but is also controlled by the local grid size. A better definition of the channels will require a locally finer grid resolution, detailed current and water height observations, and may even require a separate model component to interface with Atlantic open coast tides.

Seventeen different runs using a grid of 6143 elements and 3205 nodes illustrate the response of the CAFEX model and its sensitivity to input parameters. Many more runs were made to better define realistic parameter ranges. The variable input include the amplitude and phase of the tide on the western boundary, bottom friction, wind, density fore g, between Keys channel flow, and fresh water input from the Everglades.

Numerical results are displayed in several formats. Velocity field plots show the computed velocities in Florida Bay on an approximately 2 km grid (at the centroid of every fifth triangular element), at an instant in time.

Time series plots give a detailed picture of velocity vectors, water depth, and separate (u, v = east, north) components of velocity as a function of times but at a fixed location. This presentation method is particularly well-suited for comparison with data time series. Particle trajectory plots show the track of a water particle over a 30 day period as a simulation of the lagrangian path followed by a water borne substance. Another way of demonstrating the transport and mixing is through a plume simulation obtained as a solution of the advection-diffusion equation.

Results and Conclusions

The results indicate a western zone strongly influenced by Gulf of Mexico tides and with a tidally driven southeasterly mean flow. The central and eastern Bay zones are much less affected by tides and thus more strongly influenced by wind.

The channels through the chain of Florida Keys appear to have important implications on circulation in all parts of the Bay and additionally transmits Atlantic Ocean tides to a region close to the Keys. At least toward the southwest, these channels also serve as relief valves for the Gulf of Mexico tides, thereby reducing their effects in the interior and eastern Bay.

Many model parameters are important to the long and short term dynamic response, including boundary tides, friction, Keys channel transports, and wind. The freshwater runoff from the Taylor Slough area of the Everglades is only moderately important for long term transports and only in eastern Florida Bay and during calm-wind conditions. We have not made simulations with a mean sea surface slope between open boundaries, however, experience has shown that even a small sea level difference, as may be caused by strong ocean current systems, will have a strong effect on mean flows. For accurate model predictions, all these parameters must obviously be reasonably quantified.

The parametric studies showed that the physical processes in the model appear capable of describing the variations in available data on currents and water depths. The main data features that the model describes with reasonable input parameters include current direction and magnitude near the western boundary of the Bay at 81°05'W and tidal height amplitudes throughout the Bay. Other features for which there is qualitative agreement include the flood flow pattern into Johnson Key basin, the tidal flow through Long Key Channels and the southwesterly mean flow into Florida Bay with associated mean transport toward south in the Keys channels. More new observations are needed for additional evaluation of model performance.

For long term transport in Florida Bay, wind appears to have the greatest influence. The direct surface wind stress superposes a downwind velocity on tidal currents in open areas. In the eastern and central regions, local obstacles can cause recirculation cells to form, which tend to trap water masses. In the present work we have only included the direct surface stress from wind, however, many studies have shown that remote forcing and coastal waves can be important too.

A comprehensive set of data are required for model calibration and forcing. For model performance evaluation, information on lagrangian motion is the most useful. There is little hope of being able to use conventional current measurement techniques in the low velocity and complex topography areas of central and eastern Florida Bay. We have suggested mapping of salinity and dye experiments as possible ways of obtaining the required information.

Our investigations with a 2-D hydrodynamics model has not revealed any serious shortcomings in model physics based on available data. For future work, the most important improvements of the model are to obtain better forcing data and to refine the model resolution.

Wang, J.D., and Monjo, C., 1995: A Study to Define Model and Data Needs for Florida Bay. Report, Applied Marine Physics, Rosenstiel School of Marine and Atmospheric Science, University of Miami.

Last updated: 04/23/98
by: Monika Gurnée
gurnee@aoml.noaa.gov