Proceedings of the 7th International Conference on HydroScience and Engineering

Philadelphia, USA September 10-13, 2006 (ICHE 2006)

ISBN: 0977447405

Drexel University

College of Engineering

Drexel E-Repository and Archive (iDEA) http://idea.library.drexel.edu/

Drexel University Libraries www.library.drexel.edu

The following item is made available as a courtesy to scholars by the author(s) and Drexel University Library and may contain materials and content, including computer code and tags, artwork, text, graphics, images, and illustrations (Material) which may be protected by copyright law. Unless otherwise noted, the Material is made available for non profit and educational purposes, such as research, teaching and private study. For these limited purposes, you may reproduce (print, download or make copies) the Material without prior permission. All copies must include any copyright notice originally included with the Material. You must seek permission from the authors or copyright owners for all uses that are not allowed by fair use and other provisions of the U.S. Copyright Law. The responsibility for making an independent legal assessment and securing any necessary permission rests with persons desiring to reproduce or use the Material.

Please direct questions to [email protected]

The 7th Int. Conf. on Hydroscience and Engineering (ICHE-2006), Sep10 –Sep13, Philadelphia, USA

METEOROLOGICAL AND SUBSURFACE FACTORS AFFECTING ESTUARINE CONDITIONS WITHIN IN THE ST JOHNS RIVER, .

Joseph Stewart1, Peter Sucsy2, and John Hendrickson3

ABSTRACT

Lake George is a flow-through lake located in the St Johns River (SJR), an elongated shallow river estuary. Tide propagates upstream as far as the lake (190 km) where it is dampened out. The filling and draining of Lake George is dominated by subtidal variability of Atlantic Ocean waterlevel (Morris, 1995). Summer cyanobacteria concentrations are often high with chlorophyll-a levels regularly exceeding 100 μg L-1. Such high levels of cyanobacteria cause undesirable shifts in higher tropic levels. The cyanobacteria add approximately 1400 MT yr-1 of nitrogen by N-fixation, further contributing to eutrophication of the downstream marine portion of the estuary. Because of the importance of mixing and circulation processes in Lake George to understanding phytoplankton dynamics, a 3-D hydrodynamic model (EFDC) was applied to the lake. Salinity (chloride) entering the lake through springs along the western shore was used as a conservative tracer for verification of the model’s mixing and transport processes. Simulated dye tracers were used to determine flushing rates and delineate volume sources under varying meteorological conditions. Average turnover rate of the lake was 84 days, but during low-flow periods turnover rate ranged to 180 days. The estimated GPP of the lake was 800 gC m-2yr-1. These two results indicate that peak algal biomass is partially controlled by flushing. Under these conditions model tests show that water quality is significantly influenced by local groundwater sources entering by springs. The enters the St Johns 8 km downstream from Lake George. Reverse flows can push water entering the SJR from the Ocklawaha up into lake George. For the period studied, this accounted for 5% of the total volume entering lake. This study is a continuation of similar work underway downstream (Sucsy, 2002). This paper will summarize an evaluation of the subsurface and meteorological processes that influence the water quality and estuarine character of the lake.

1. STUDY AREA DESCRIPTION and RESOURCE ISSUES

1.1 Study Area Description

------1 Engineer Scientist, Saint Johns River Water Management District, Palatka, FL 32178, USA ([email protected]) 2 Supervising Engineer Scientist, Saint Johns River Water Management District, Palatka, FL 32178, USA ([email protected]) 3 Environmental Scientist V, Saint Johns River Water Management District, Palatka, FL 32178, USA ([email protected])

Centered about 190 km upstream from the Atlantic Ocean, Lake George is Florida's 2nd largest lake, with an area of 189 km2 and a volume of 0.493 km3. It is approximately 10 km wide by 15 km in length. The mean depth of the lake is 2.5 m. To investigate Lake George, boundaries were set on the SJR to the south at Astor (SR 40) and to the north at Buffalo Bluff Rodman (Figure 1). A record of daily discharge and waterlevel Dam existed at these locations for the study period (1995 - 2005). Temperature and specific conductivity are Black collected continuously at Buffalo Bluff. Three features Point that are referenced in the study: (big) and Hog Island (small), on the north end of the lake; and Black Point, north of Drayton Island where the river returns to a single channel course. The largest tributary of the SJR, the Ocklawaha River, enters the system 8 km downstream of Lake George. A range of water quality parameters are regularly collected at both boundary locations, within the lake, and from within the Ocklawaha. Lake George is unaffected by marine salinity (the greatest upstream encroachment of marine salinity is 112 km) (Morris, 1995), but can reach appreciable levels of dissolved solids due to the influence of in-lake and upstream brackish springs. This imparts marine characteristics into the lake. Tides propagate upstream as far as the lake where they are dampened out. Figure 1: Lake George Study Area Large-scale processes in the Atlantic can translate upstream well beyond lake George, where the presence of local source of salt allows a large assemblage of marine species to reside in the lake. The continuous supply of salt may have had a similar role in the migration of marine species well upstream of Lake George for thousands of years (Odum, 1953). 97 species of fish are listed for the SJR around the lake, with approximately 41 marine species, including stripped bass and mullet (McLane, 1955). It was one of the most productive fresh water fisheries in the late 19th century, but began to decline in the mid 1900s. Shrimp migrate into the area under ideal conditions, and the lake still supports a local blue crab industry. H.T. Odum’s method for estimating primary production was developed using research that included local spring runs (Odum, 1956). Florida owes is geologic structure to the variability of sea level over millions of years. During the Tertiary Period, reefs that grew when the entire region was part of a shallow sea were then covered with sand as the state emerged from the water (White, 1970). This pattern repeated itself several times, creating shorelines that are echoed in the terraces of Florida today. The springs of the SJR in the region around Lake George also owe their existence to this process. Water flows through underground passages of relic limestone reefs. The presence of this constant artesian source has also helped to preserve the present course of the SJR (Pirkle, 1971).

1.2 Resource Issues

Pollution Load Reduction Goals (PLRGs) and Total Maximum Daily Loads (TMDLs) being developed in the Lower St John River (LSJ) basin downstream, and the Middle (MSJ) and Upper (USJ) St Johns River basins upstream. The upstream boundary for the study area was set at Astor,

Florida. The drainage basin upstream of Astor is 8624 km2. A large portion of the USJ flood plain was converted to farming during the 20th century. Agricultural practices continue, but most of the flood plain has been bought back by the state of Florida and managed for flood control, to improve water quality, and environmental benefits. The MSJ, including the Orlando metropolitan area, is undergoing rapid urbanization. Nutrient loading from urban runoff finds its way into the SJR through surface runoff, and other sources such as septic systems and wastewater treatment plants. There is also an increase in nutrient concentration coming into the system from springs. The low gradient over the last 200 km of the river increases the residence time considerably in comparison to the upstream 300 km of the St Johns River. Historically, vast expanses of aquatic vegetation, primarily water lettuce, covered areas of the St Johns River. In the late 1800s, water hyacinth was introduced to the system. By 1900 it had out competed water lettuce. Whole sections of the River would be occasionally blocked to navigation. (Figure 2). Considered an impediment to navigation and commercial use of the river, the Army Corp of Engineers sought ways to eradicate it. For the first half of the 20th century, mechanical means were used with minimal effect. Intensive spraying of herbicides that started in the late 1940s was successful in bringing floating vegetation under control (Simberloff, 97). Figure 2: Water hyacinth on the SJR, pre 1900 However, removing the vegetation had negative consequences. The abundant vegetation in the river shaded the water column from direct sunlight, without it, more light and heat became available to algae production. The growth of aquatic vegetation removes nutrients from the water column. The application of herbicide results in the release of nutrients back into the water column as a plant decays (Moody, 1970). By one estimate, up to 600 tons of N are fixed in Lake George per year, as estimated from work in 1999 (Hendrickson, pers comm.). However, phosphorus is the limiting nutrient in the system. While a portion of the phosphorus that is loaded to the system is anthropogenic in origin, groundwater in Florida can have high natural concentrations of dissolved phosphorus due to large deposits of phosphate rock around the state (Odum, 1952). Phosphorus-laden discharge entering the Lake George from the middle and upper basin has time to attenuate due to long residence times under low-flow conditions. Management scenarios are currently being Figure 3: Algae surface scum near Lake George evaluated to determine the best course to take in reducing the likelihood of harmful algae blooms in Lake George and downstream. For the LSJ, one key to controlling the development of harmful algal blooms is to limit the amount of

phosphorus entering the lower portion of the river from the lake. However, if natural levels of phosphorus are high, removing the anthropogenic sources, while beneficial, may not be adequate. Another management strategy being evaluated is to allow greater coverage of aquatic vegetation to shade the water and take up nutrients. Plants, and thus nutrients could then be removed from the system by mechanical harvesting. Other potential strategies for phosphorus removal include, but are not limited to: • Gizzard shad harvesting • Alum treatment at key points upstream • Septic system replacement with sewers • Agricultural BMPs • Wastewater treatment system improvements • Urban retrofits

Another resource issue concerns the removal of Rodman Dam on the Ocklawaha River (Figure 1). The dam was built as part of the now defunct Cross-Florida Barge Canal. Currently, nutrients are attenuated within the reservoir behind the dam. If it is removed, this attenuation will be lost and the loading of nutrients to the SJR will increase. In a reverse flow event, this additional nutrient load can be driven upstream into Lake George, making it available to algal processes. Any consideration for the removal of the Rodman Dam would have to include an assessment of the resulting effect on Lake George.

2. OBSERVED DISCHARGE and STAGE DATA for the STUDY AREA

Discharge from all terrestrial sources was compared and a summary of daily discharge for 1995 to 2005 compiled (Table 1). The majority of water flowed into the lake from Astor, but due to frequent reverse flow events and the large capacity at Buffalo Bluff, 8.2% of the water that entered the study area came from downstream sources. Within the 10-year record, the daily averaged Astor discharge exceeded Buffalo Bluff 27.4% and the combined discharge from Astor and the Ocklawaha River exceeded Buffalo Bluff 38.8% of the time. A positive value in Table 1 indicates flow directed downstream, a negative value indicates reverse flow.

Table 1: Summary of discharge from terrestrial sources, 1995 to 2005 Discharge (m3s-1) to the study area Astor Ocklawaha Local sources Buffalo Bluff Mean 97.3 32.9 12.7 145.5 Max 469.3 264.0 77.3 744.5 Min -250.7 0.0 5.4 -631.1 Stdev 84.6 24.5 9.6 153.0

2.1 Astor and Buffalo Bluff

The largest monthly flow from 1995-2000, and the second largest from 1995-2005 occurred during the El Nino winter (Jan-Mar) of 1998. During this time, the SJR was relatively unaffected by tropical cyclonic activity. When reverse flow events are analyzed two distinct peaks in discharge occur in May (avg -287 m3s-1) and September (avg -489 m3s-1). The average duration of a reverse

flow event was 5 days. In May when stage in the SJR is typically at its lowest, the river is sensitive to small baroclinic fluctuations. The influence of tropical cyclone activity on the SJR tends to peak in September in the past decade. The effect of lake George on dampening out tidal forces is evident when evaluating daily discharge at the study area boundaries (Figure 4). 800 Buffalo Bluff 600 Astor ) -1 s

3 400

200

0 Discharge (m -200

-400 Jul-95 Jul-96 Jul-97 Jul-98 Jul-99 Jan-95 Jan-96 Jan-97 Jan-98 Jan-99 Jan-00

Figure 4: Mean daily discharge, St Johns River, 1995-1999

Conditions in the Atlantic Ocean are transferred into the study area across the downstream boundary at Buffalo Bluff. A comparison of mean daily waterlevel at Mayport with the two St Johns River boundary locations shows that there were frequent occurrences of waterlevel at Mayport that was at or higher that those at the boundaries (Figure 5). When this occurs, the river changes direction. The three locations show many similarities as the St Johns reacts to conditions in the Atlantic Ocean that are transferred into the River past Mayport. Deviations between stage of Astor and Buffalo Bluff in comparison to Mayport are primarily due to seasonal variability of precipitation. 0.8 Buffalo 0.6

.. Astor 0.4 Mayport

0.2

0.0

-0.2 Stage (m-NAVD88) -0.4

-0.6 Jul-95 Jul-96 Jul-97 Jul-98 Jul-99 Jan-95 Jan-96 Jan-97 Jan-98 Jan-99 Jan-00

Figure 5: Mean Daily Waterlevel, St Johns River 1995 - 1999

Variations in waterlevel in the Atlantic Ocean that are the result of large-scale processes (i.e. hurricanes or frontal boundaries) move water into the river and then upstream, sometimes beyond Astor. When the mean stage between Astor (0.054 m) and Mayport (-0.132 m) is calculated for 1995-1999 (Figure 5), there is a drop in the water surface of approximately 19 cm from Astor (202 km) to the mouth at Mayport. This comes to about 1cm/10km of surface slope within the river.

2.2 The Ocklawaha River

The Ocklawaha River enters the St Johns 8 km downstream from Lake George. It is the largest tributary to the SJR (avg 32.9 m3s-1). It is similar to the St Johns in that its baseflow is almost entirely provided by groundwater discharge. The upstream contributing watershed (7114 km2) includes several large lakes and Silver Springs. For the Ocklawaha River, a continuous time series was available for the study period from the outfall. While the local contributing watersheds might drop to zero during dry periods, and the minimum discharge from the Ocklawaha River typically remains between 10 and 20 m3s-1. This baseflow comes almost entirely from Silver Springs and other springs in the Ocklawaha River. The moderating effect of the Ocklawaha River passing through the Rodman reservoir outfall structure can be clearly seen (Figure 6). Discharge from Rodman reservoir can drop to zero (Table 1). This occurs when the outfall structure is closed to fill the man-made lake. For this study, analysis of the influence of the Ocklawaha on the SJR uses the existing discharge time-series. 800 Buffalo Bluff 600 Astor

) Ocklawaha -1 s

3 400

200

0 Discharge (m -200

-400 Jul-00 Jul-01 Jul-02 Jul-03 Jul-04 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05

Figure 6: Plot of Astor and Buffalo Bluff (SJR), and the Ocklawaha River, Jan-03 to Jun-05.

2.3 Springs and Local Watersheds

Common to the north-central region of Florida are substrates of a marine origin that yield water with high conductivities (Odom, 1957). The age and composition of waters emerging from a spring lends a unique chemical signature that can act as a conservative tracer. Typically, water from a spring comes from a myriad of sources as indicated by water quality samples collected from different vents in a spring for isotope studies (Toth, 1999). The eleven named springs in the study

Figure 7: Relief map of region around study area, including Silver and Blue Springs area (Figure 7) owe their existence to the succession of marine and terrestrial process mentioned earlier. The focus of this discussion will be the four largest contributors of spring water to the study area: Silver Glen, Salt, and Croaker Hole, and Juniper Creek. Three notable springs discharge into the creek before it enters the southwest side of Lake George. The smaller springs in the study area will only be mentioned briefly. The salt that passes through the bridge at Astor comes from many sources. Ambient salinity in the lake typically ranges from 0.3 – 0.8 ppt as a function of discharge from upstream (compare Figure 6 and Figure 8). In developing a time-series for the salinity condition at Astor, it was necessary to explore the sources of the salt, since we were analyzing spring water within the lake. While there are many notable springs in the MSJ, salt is also introduced to the system from the dissolution of minerals from lakes and salt marshes in the MSJ and USJ (Steward, 1984; McLane, 1955). At times it isn’t clear if specific conductivity as measured at Astor doesn’t owe some of its character to downstream sources through upstream advection of salt under low-flow conditions. Data collected from within the lake, and three locations upstream (Figure 8) show the spatial and temporal variability of salt within the river (Kroenig, 2004). Data was derived from field measurements and continuous sensor readings. The maximum salinity in the five-year period occurred upstream in , 265 km upstream from the ocean. Total dissolved solids

concentrations as high as 1.2 mg/L have been measured when upstream and surface runoff is low, evapoconcentration is high, and artesian spring inputs are large in relative contribution. 1.4 1.2 Sanford (265km) Deland (233km) Astor (202km) Lk George (185km)

.. 1.0 0.8 0.6

Salinity (ppt) 0.4 0.2 0.0 Jul-00 Jul-01 Jul-02 Jul-03 Jul-04 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05

Figure 8: Ambient Salinity at 4 locations on the St Johns River, Jan-00 to Jan-05. Distance is km from the mouth of the river.

Three major springs on the west side of Lake George, with a combined discharge of about 6 m3s-1, discharge groundwater at salinities of 1 – 4 ppt, setting a horizontal gradient of decreasing salinity moving away from shore. Elevated salinity due to the local spring sources can extend several hundreds of meters during high river flows to several kilometers during low flows. The periods when springs contribute the greatest proportional salt load to the lake coincide with periods when conditions are optimal for cyanobacteria production, making the springs an ideal natural conservative tracer for development of a water quality model of Lake George. To develop discharge time-series for the springs that flow directly into the study area, first the available spring data was compiled. Observed spring data was intermittent and extended back about 70 years. Observations at spring locations were sparse during the 95-05 data period. There were typically 3 or 4 measurements per year. Given this, it was decided that the resolution would be a monthly estimation of discharge, and constant salinity would be suitable. There is a question of whether there is a daily variability of spring discharge and salinity coming from a spring. This is a valid assertion, since the water issuing forth from a spring is typically comprised from multiple sources. Ancient water from deeper aquifers with a constant composition may mix with more variable surficial aquifer with a shorter residence time underground (Toth, 1999). To develop monthly time-series for spring discharge, comparisons were made to the stage in two wells and to Blue springs upstream (Figure 9). Several springs were compared to Sharpes Ferry well and well V-0510 (Figure 7). To initially validate this approach, a comparison was made between Silver springs discharge and Sharpes Ferry well stage. The close relationship between these Figure 9: Comparison of Juniper springs locations is well documented. (Ferguson, 1947). discharge and stage in two wells

The two wells were selected since daily stage has been measured in them going back at least 70 years. To compare spring to well, the well stage was extracted for the date of a spring measurement. There was a decent correlation with Juniper (Figure 9), Fern Hammock, and Sweetwater springs. Blue Springs is a first magnitude spring (greater than 100 MGD) with an average discharge of 4.4 m3s-1. Blue Springs and the majority of the springs in the study area discharge into a spring run or the river bottom and are directly affected by variations in the waterlevel of the SJR. Based on this and a the availability of a nearly complete monthly data set of Blue Springs discharge for the study period, Blue Springs data was used to fill in missing values. The average discharge from each spring in the study area and Blue Springs was determined and the ratio of Blue Springs and local springs was used as a correction factor to adjust the Blue Springs discharge for application to the local spring. If observed data was available for a spring in a given month between 1995 and 2005, it was used. Summarized in Figure 10 is the average discharge from the study period and the salinity for the 4 main ‘springs’ in the study area. The Juniper creek springs are summarized together. Silver Glen is nearly a first magnitude springs. A close second is Croaker Hole, followed by Salt springs. Salinity does not follow the same trend. The big contributor in terms of salt mass is Salt springs with a discharge of 2.26 m3s-1. Salt, Silver Glen, and Juniper Creek discharge directly into Lake George. Croaker Hole discharges south of the Ocklawaha River into Little Lake George (Figure 1). Figure 10. Average discharge and salinity from selected springs

The two smallest springs in terms of discharge, Satsuma and Nashua put salinity into the system at above 3.0 ppt (Figure 7). While these local springs may make an imperceptible contribution in terms of the water budget in the system, the influence of salinity locally within the St Johns river may be important in providing continuity in the salt balance of the system and so are considered. Croaker Hole (at 1.4 ppt) may play a similar role in moderating the effect of large inflows of surface runoff (freshwater) from the Ocklawaha. H.T. Odum studied primary production in Beecher Springs (Odum, 1957). Surface basin inflows and estimations of nutrient loading to the SJR from local watersheds were derived using the Pollutant Load Screening Model (PLSM) with daily proportioning to Haw Creek, a local gauged watershed (Mundy, 1998). The surface area of local contributing watersheds (423 km2) is only 10% larger than the surface area of Lake George. Springs on the west side supplement the discharge coming from local watersheds, so even during the driest periods there is some inflow to the lake. On the east side during dry periods, discharge may drop to zero. The average combined discharge from the 4 main spring sources and all local watersheds was 8.7 m3s-1 and 5.7 m3s-1, respectively. The maximum spring discharge (9.94 m3s-1) and local surface discharge

(38.16 m3s-1) indicates that although much lower than the peaks for Astor or the Ocklawaha (Table 1) at times local watersheds may be a significant contributor of nutrients to the system, at least locally.

3. WATER QUALITY CHARACTERISTICS IN THE STUDY AREA

Lake George is eutropic and, as a run-of-the-river lake, typically exhibits a seasonal oscillation in lotic and lacustrine character. Algal blooms can occur starting in March, and typically persist through August (Figure 11). Typical annual maximum chlorophyll-a is between 100 – 150 mg/m3. Diatom algae predominate early in the season, but are quickly supplanted by cyanobacteria when silica becomes limiting, usually by April (Phlips and Cichra, 2002). Nitrogen is the principle limiting macro-nutrient (Paerl, 2005) and nitrogen-fixing cyanobacteria predominate in the lake phytoplankton during spring and summer blooms, during which they can double the lake nitrogen

160

140

120

100

80

60

Chlorophyll-a, mg/m3 Chlorophyll-a, 40

20

0 0 30 61 91 122 152 182 213 243 274 304 334 365 Julian Day

Figure 11: Annual trend in phytoplankton chlorophyll-a at the Outlet of Lake George, 1995-2005 concentration. Lake phosphorus concentration was reported as 0.044 mg/L in 1951 by Odum (1953), but by the late 1960’s had risen to between 0.110 to 0.130 mg/L (Moody, 1970). Presently (2000-2005), in-lake P concentration average is 0.081 mg/L, and has exhibited a slight upward trend since the beginning of recent monitoring in 1989. Mean nutrient concentrations for lake inputs are listed in Table 2. Table 2: Mean nitrogen and phosphorus form concentrations for major inputs to Lake George, 2001-2005. Values in mg/L

Fraction of Lake Location NOx NH3 TON PO4 TP Volume SJR at Astor (Upstream) 0.89 0.084 0.048 1.29 0.050 0.087 Artesian Springs 0.06 0.079 0 0.067 0.021 0.028 Local Watershed Runoff 0.05 0.024 0.024 0.662 0.043 0.066 L. George Outlet N/A 0.037 0.037 1.67 0.010 0.067

Several trips to the lake were made to collect data on the horizontal and vertical distribution of basic water quality parameters. The findings from these trips are summarized in the Results.

4. PRECIPITATION, EVAPORATION, AND WIND

Daily rainfall estimates derived from NEXRAD Doppler radar were available for the 1995 – 2005 period. This data is compared to ground stations for accuracy. Using this radar data, a mean estimate of precipitation to the study area was obtained. Prior to 2002, pan evaporation data, corrected to open water was used. More recent evaporation data came from an open-water station north of the study area in the LSJ. Evaporation from this station was derived in the same way as a station on the Lagoon (Sumner, 2005). The LSJ data is 0.40 suitable to this study area since it 0.35 is in the same temperate zone and 0.30

comes from data collected ) 3 0.25 downstream over open water similar to Lake George. 0.20 The cumulative difference 0.15

between average precipitation and Volume (km 0.10 evaporation from 1995 to 2005 0.05 was determined (Figure 12). Over the 10-year record, Lake George 0.00 has gained about 0.7% of its -0.05 volume from atmospheric exchange versus terrestrial inflows (compared to lake volume of 0.493 Jan-95 Jan-96 Jan-97 Jan-98 Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 km3). The plot is useful in Figure 12: Cumulative atmospheric assessing drought condition in the exchange study area. The peak in 1998 is attributed to the excessive rainfall of an El Nino winter. Likewise, the rapid increase in 2004 of associated with record rainfall from several tropical cyclones that affected the St Johns basin. The effect of drought can be seen in the downward trend between late 1998 and 2002. In these years, more water was lost atmospherically than gained. In all other years, precipitation exceeded evaporation. 1999-2000 was exceptionally dry, and this is noteworthy for the algae blooms that occurred that year in the lake in late spring - early summer.

Wind data used in the study area came primarily from a NOAA station at Jacksonville Naval Air Station 50 km to the north of the study area. The location of the wind gauge is geographically similar to the study area. The data is reliable and continuous. Lake George is shallow relative to its size, and a thermocline never forms in the lake. Given this, the wind data was deemed adequate. Certainly larger scale processes due to wind such as mixing and resulting circulation patterns in the river are important but the primary concern in modeling the lake was diffusive mixing in the water column. The average wind speed was 3.11 m/s with a standard deviation of 1.57 m/s. A wind gauge was installed on the lake in June 2004. Data from this location was used as available.

5. HYDRODYNAMIC MODELING

A bathymetric relief of the study area was developed using NOS data for the bottom elevation of the SJR combined with a shoreline coverage set at 0 m NAVD88. Depths in the study area ranged from 0.0 to 16.0 m with an average depth of about 3 m. The first evaluation of system dynamics involved developing a mass balance for Lake George. The approach assumed that there is no mixing in the lake. By using only monthly averaged data, reverse flows were negated and the system could be treated as a one-way, flow-through system. A stage-area-volume relationship for Lake George was developed from the bathymetric relief of the study area. This was used to determine the turn over rate of the lake under various flow conditions. To analyze flow and hydrodynamic characteristics within this portion of the SJR, a hydrodynamic model was developed using EFDC, the environmental fluid dynamics code. It was selected since it has been used extensively in the Lower St Johns (Sucsy, 99), Future work will include integrating the two models, as well as extending modeling efforts upstream. EFDC used a curvilinear, boundary- fitted grid to which depth data is applied.

200 Figure 13: Study area bathymetry and EFDC model grid

) -1 s

3 0 Figure 13 shows an overlay of the Lake George circulation grid on top of the study area bathymetry. The grid has 672 cells in the horizontal, and 6 cells in -200 the vertical to provide resolution to the water column.

Discharge (m With an average depth of approximately 3 meters in the Observed system, the resolution of a vertical cell is (on average) Simulated 0.5 meters. The horizontal cell resolution ranges from -400 0.016 to 0.64 km2 (mean 0.37 km2), with the entire grid encompassing 209 km2. At 0.0 m NAVD88 the surface 6:00 8:00

10:00 12:00 14:00 16:00 18:00 area differed by 1% and volume differed by less than Figure 14: ADCP flow 5%, respectively, when comparing the model grid and comparison, Georgetown East the bathymetric relief between the boundaries of Astor and Buffalo Bluff.

Salinity data became available starting in June 2004 from a location centered in the lake. Observed and modeled salinity matched at this location and appears to be primarily due to the salinity entering the upstream boundary at Astor (Figure 8). At Buffalo Bluff, The M2 tide is about 14 cm. Tide continues to propagate upstream until it is dampened out by Lake George. The hydrodynamic model was calibrated for tide between Buffalo Bluff and Georgetown at the north end of Lake George. ADCP flow measurements were collected in May 2002 near Fruitland and at Georgetown on the east side of Drayton Island, the island just to the north of Lake George (Figure 14). Getting a good match at these locations primarily involved getting the cross sectional area correct in the model grid. Both measurements were taken on the same day and for the same period to approximate half a tidal cycle (12 hours). Both included occurrences of reverse flow that are represented as a negative number in the plot.

6. RESULTS

6.1 Comparison using discharge data and indicators of algal production

The variation in rainfall in Florida (Figure 12) is clearly discernable when summarizing seasonal flow into Lake George (Figure 15). The El Nino winter of 1998 is second in total input only to the hurricane season of 2004. The drought period between 1999 and 2002 is also easy to delineate. Chlorophyll-a production is tied to the seasonal variability of inflow to the lake (Figure 11).

3.0 Surface 2.5 Springs Astor

) 2.0 3

1.5

Volume (km 1.0

0.5

0.0 96s1 96s2 96s3 97s1 97s2 97s3 98s1 98s2 98s3 99s1 99s2 99s3 00s1 00s2 00s3 01s1 01s2 01s3 02s1 02s2 02s3 03s1 03s2 03s3 04s1 04s2 04s3

Figure 15: Lake George Water Budget, by Season, from 1996 – 99. Annual flows are divided into 3 seasons: 1) Dec – March, typically moderate precip. supplied in frontal systems, and low evaporation; 2) April – July, hot and typically dry; 3) August – November, warm and wet with flow driven by convective and tropical storms.

Periods of maximum chlorophyll-a concentrations coincide with minimum discharge (Figure 16). Crashes in phytoplankton populations are somewhat associated with increases in inflow to the lake, but internal population dynamics appear to also be at play. To gain a better

understanding of the importance of residence time in phytoplankton processes, the turn over rate of the lake was evaluated.

Figure 16: Comparison of chlorophyll-a and Buffalo Bluff discharge 1996-2005

To calculate the turnover rate in days, the outflow volume in m3s-1 was converted to a total for the month. This volume was then compared to the "instantaneous volume" of the lake derived from the stage-volume relationship. The lake volume divided by the monthly outfall volume times the number of days in the month is the number of days it would take to turn over the total lake volume. The average turnover rate over the study period was 84 days. This correlates with an average inflow from Astor of 100 m3s-1. Under peak flow conditions this dropped to 24 days and low flow conditions increased to 180 days. The relationship between Astor inflow and Lake George turnover rate was used to further investigate system dynamics.

LkGeo Springs LkGeo Surface Astor Turnover (days) 120

100 ) and

-1 80 s 3 60

40 Turnover (day)

Discharge (m 20

0 dec-mar apr-jul aug-nov Figure 17: Seasonal variations in Lake George inflow versus turnover rate

Another way to look at the results of the mass balance approach was to determine seasonal variations in Lake George inflow versus turnover rate (Figure 17). The seasons were described in the water budget for Lake George (Figure 16). The turnover rate of the lake is considerably higher in the dry season than the other two. This is important in regards to phytoplankton growth. A longer turnover rate means a longer residence time in the lake. This is the time of year when the ambient temperature is increasing towards the optimum condition for algae growth. Also, there is a lower inflow from Astor to the lake, and the local watersheds are at their minimum, so the contribution from spring discharge is proportionally larger. Thus, the water coming into the lake is clearer, and this increase water clarity in the lake (at least along the western shore) increasing the size and depth of the photic zone. For the study period the overall spring contribution when looking at monthly average discharge was 8.1% and the surface 2.3%, leaving the remainder (89.6 %) coming in from Astor. The difference for the total study period versus just 2001-2005 (Table 2) is due to the reduction of inflow from local watersheds and Astor during the drought period (late 1998 to early 2001, Figure 15).

Figure 18: Comparison of Buffalo Bluff discharge and downstream production

The lowest discharge at Astor typically occurs in the months of May to June. It is also during this time that conditions become optimal for algal production. The best condition for algal production appears to coincide when inflow drops below 40 m3s-1 and other factors are aligned (Figure 18). Conditions at Dancy Point are similar to Lake George, so these production estimates are similar to values for the lake. An animation of the salinity characteristics of the system yield insights in to circulation patterns in Lake George, the relative abundance of spring water with its moderating effect on the west side of the lake, and the ambient conditions under varying flow conditions. An excerpt from this animation of a low flow and high flow scenario provides a visualization of the ambient conditions effecting production (Figure 19).

6.2 Comparison using model simulations and field observations

Model analysis using a numerical tracer was used to:

• Determine the portion of lake water from various sources under varying flow conditions • Estimates of turnover rates under varying flow conditions • Delineate the portion of water within the lake attributed to the downstream Ocklawaha River

Figure 19: Snapshot of salinity animation, May 27 2004 and Oct 20 2004

Under the low flow condition (May 27, 2004), the high salinity coming in from upstream (Figure 8) obscures the signature from the springs (Figure 19, right). To investigate this relationship further, a numerical tracer was applied to the springs to better determine the extent and distribution of spring water in the lake. Though a snapshot is not presented here, the result of that study was similar to the seasonal water budget analysis (Figure 15), in that is resolved the relative contribution of spring water to the lake versus what comes in from upstream. Under high flow conditions (Oct 20, 2004) like those that existed after several hurricanes dumped record rainfall on the SJR watershed, the deeper part of the lake acts more like a conveyance, leaving some disparity in residence time between the sides and the deeper middle (Figure 19, left). However, it is the low flow condition that is of greatest concern in regards to suitable conditions for algal production. The snapshot of low flow condition does indicate a much higher residence time overall in the system, not just in Lake George, since the dissolved solids concentration has time to steadily increase (Figure 8). A comparison to the two different conditions does support the idea that there is a spatial variability of salt within the lake due to the presence of springs, indicating that at times horizontal mixing within the lake is incomplete.

To test the horizontal spatial variability of residence time within Lake George, a numerical dye was evaluated under different flow conditions (Figure 20). The primary consideration in doing this was to determine if any part of the lake may act as a quiescent zone where residence time was longer than, or varied significantly from the over all system, or if the system could be treated as homogeneous in regards to turnover rate. Three periods of observed data were selected when

low flow condition < 50 m3s-1 (4/1/2001 to 6/24/2001) 100

80

60

40

20

0 % Tracer remaining in cell

-Apr -Apr -May -Jun 1 8 5-Apr 2-Apr 9-Apr 6 3-May 0-May 7-May 3 0-Jun 7-Jun 4-Jun 1 2 2 1 2 2 1 1 2

medium flow condition ~100 m3s-1 (9/1/1998 to 11/24/1998) 100

80

60

40

20

% tracer remaining in cell 0

-Sep -Sep -Oct -Nov 1 8 5-Sep 2-Sep 9-Sep 6 3-Oct 0-Oct 7-Oct 3 0-Nov 7-Nov 4-Nov 1 2 2 1 2 2 1 1 2

high flow condition > 200 m3s-1 (9/1/2002 to 11/24/2002) 100 zone1-sw 80 zone2-se zone3-swcentr 60 zone4-secentr zone5-nwcentr 40 zone6-necentr 20 zone7-nw zone8-ne

% tracer remaining in cell 0

/1/2002 /8/2002 9 9 /15/2002 /22/2002 /29/2002 0/6/2002 1/3/2002 9 9 9 1 0/13/2002 0/20/2002 0/27/2002 1 1/10/2002 1/17/2002 1/24/2002 1 1 1 1 1 1

Figure 20: Comparison of residence time in 8 zones of Lake George conditions were comparable to the discussion of turnover rate and seasonality described in Section 6.1. Observed discharge at Astor for the low flow and high flow scenarios can be seen in Figure 6 and observed discharge for the medium flow condition can be see in Figure 4. An initial dye

concentration was set at 100 % and each scenario was run for 84 days, the average turnover rate as discussed earlier. For each plot, two zones are singled out in the southeast section of the lake (zone 2) and the northwest section of the lake (zone 7) for comparative purposes. For the low flow condition, the sides of the lake tended to attenuate dye longer than the center of the lake, but swings in dye concentration are evident from the plot. For the medium flow condition, the dye percentage was reduced by 90% at the end of the 84-day run, but was still present at most locations. For the high flow condition, the dye was reduced to less than 10 % by day 24 at most locations. This analysis supported the “instantaneous volume” approach discussed in section 6.1. It is important for the consideration of algal processes to note the low and medium flow condition scenarios indicate the mixing that occurs within the lake, with this mixing being due primarily to local wind conditions.

The lake was visited several times between July 2004 and May 2005; basic water parameters were collected using a YSI data collection sonde (Figure 21). The purpose for these field visits was to get a better feel for the horizontal and vertical variability with in Lake George and to collect information for comparison to modeling efforts. The results of these trips are presented here for two trips in September and November, and two trips in May 2006. Salinity, pH, dissolved oxygen (DO), and are presented for comparison to other data in this study of the effect on ambient conditions due to flow variability. All data is vertically averaged from the collected profiles.

Figure 21: Summary of synoptic survey of Lake George for September and November of 2005 (high flow condition), and May 2006 (low flow condition).

In general, the system appeared to be vertically well mixed. Stratification of salinity, temperature, and DO does occur at certain locations under the right conditions. Especially at the mouths of, and just outside of spring runs, where stratification inversions induced by salinity and temperature differences between the inflow and resident lake can occur. On one occasion, a horizontal gradient (increasing east to west) and vertical stratification was observed at Black Point. This was also produced by the model (Figure 19). For the low flow situation (May 2006), salinity is generally higher throughout the lake. The average discharge from Apr-Jun 2006 was 25 m3s-1. Under the high flow condition (Sept, Nov 05), higher salinities are restricted to the west shore and adjacent springs. The average discharge during this time was 200 m3s-1. Salinity data collected around the mouth of springs indicated that an assumption of constant salinity for modeling was valid, and that our values were reasonable. The snapshots of pH and DO clearly show an increase due to algal production. Surface waters in Florida area generally well buffered (pH of 7-8) but in May 06 we see shifts to high pH indicating that available carbon in the water column is being rapidly consumed. Supersaturated DO in May is also an indication that increased production was taking place in the lake, especially in the shallow area to the northwest of Drayton Island, where pH approached 10 and DO 15 mg/l. Under calm conditions (i.e. no wind and resulting wave action), stratification occurred in deeper areas. For example, DO would range from supersaturated in the top 1 m down to saturated at depth, and water temperature would vary by as much as 5 oC within the top meter of the water column.

To analyze the effect of the Ocklawaha River on Lake George under reverse flow conditions, a numerical tracer was applied and a model run from January 2003 to Jan 2005. A visual summary of this scenario is depicted for seven days from Jan 9 to Jan 15, 2004 (Figure 22). The period was selected for the series of reverse flow events coinciding with appreciable Ocklawaha inflows to the study area (Figure 6).

Figure 23: Intrusion of Ocklawaha River water into Lake George Jan 9 – Jan 15, 2004

In the first frame (9-Jan), Ocklawaha River is filling up the area around its mouth in Little Lake George. In the second frame, reverse flow is sending the packet of Ocklawaha water upstream. By the third frame (13-Jan), the packet has entered the lake, but by 15-Jan, the flow has returned to normal so that as a portion of the Ocklawaha water propagates upstream into the lake, the main body is exiting the system downstream.

Qavg Buffalo Qmin Buffalo %Ocklawaha 500 6.0 passing Black Point Percent Ocklawaha ) -1

s 250 5.0 3 0 4.0 -250 3.0 -500 2.0 -750 1.0 Discharge (m -1000 0.0 Jul-03 Jul-04 Jan-03 Jan-04 Mar-03 Mar-04 Sep-03 Nov-03 Sep-04 Nov-04 May-03 May-04 Figure 23: Comparison of Buffalo Bluff discharge and Ocklawaha intrusion to Lake George, Jan03 to Jan05

Rather than report a seasonal estimate of the amount of Ocklawaha water that is resident in Lake George at any given time it seemed more prudent to provide the model result for the amount of water that passed Black Point upstream of the lake (Figure 23). This is the point north of the lake where the system returns to a single channel. At this location for the peak of the reverse flow event, 5 % of the water flow upstream came from the Ocklawaha. For the model period of Jan 2003 to Jan 2004, there appears to be a relationship between a succession of reverse flow events and the amount of water that can be transported upstream from the Ocklawaha. In prohibiting downstream conveyance, these events allow water entering the St Johns River to build up in and around Little Lake George before it is sent upstream. For this study, analysis of the influence of the Ocklawaha on the SJR uses the existing discharge time-series, but a simulated hydrograph representing an unstructured Ocklawaha is necessary to assess the affect of removing the Rodman Dam will have on water quality in the SJR.

7. CONCLUSION

Lake George is a flow-through lake located in the St Johns River (SJR), an elongated shallow river estuary. Because of the importance of mixing and circulation processes in Lake George to understanding phytoplankton dynamics, a 3-D hydrodynamic model (EFDC) was applied to the lake. Results from modeling were compared to existing data, as well as synoptic data collected in the field. Ambient salinity in the lake is dominated by upstream conditions, with local effects due to springs and other terrestrial sources. An analysis of the salinity characteristics of the lake indicate that while it may be possible to delineate spring water under high-flow conditions, under low-flow, the ambient salinity due to upstream sources obscures the salt signature of the local springs. A promising aspect of the modeling of salt is the snapshot it yields of the ambient condition in the lake and its relation to flow regime. In general, the higher the salt concentration in Lake George, the longer the water has been there. A salinity gradient observed and modeled at Black Point, may indicate a bifurcated flow sets up when salt-laden water from the west side of Lake George (especially salt springs) combines with water leaving the lake on the east side of Drayton Island. Croaker Hole springs, just 3 km north and also on the west side of the river, may also influence the development of this gradient. Interestingly, this is a common occurrence where spring water enters Florida’s smaller black water rivers. This could be an indication that salt laden discharge from Salt Springs tends to attenuate in the shallow area north east of Drayton Island around Hog Island, also indicated in the model results (Figure 19).

A water budget approach resulted in a reasonable estimate of the amount of water within Lake George due to spring discharge, it is clear that spring water tends to favor the west side of the lake in a gradient of decreasing salinity out from the shoreline that varies with flow regime. An isotope study that tracks the movement of water from a particular spring may be one way to clarify the problem. The influence of spring water in setting up a gradient in the lake, and the resulting effect algae population dynamics and distribution is important and will continue to be explored. Wave action on the lake may be another factor affecting the likelihood that algae bloom will occur. Periods of extended calm or minimal wind allow stratification to intensify. Wave set up of >1m is well documented with above average winds, on a lake with a mean depth of 2.5m. An average wind can kick up a .3 to .6 m chop. The lake is unusual in Florida among the big lakes in that the deep section (avg 3.35m, ~150 km2) is sandy and uniformly flat. A wind station was established on the lake in June 2004, as a dataset is built, the situation would be reexamined. Future evaluation of wind dynamics will also include steady-state tests to examine the mixing response of the lake. The mean residence time of the lake (84 days) was calculated from monthly averages to negate the effect of reverse flows. Obviously, reverse flows increase residence time by prohibiting the draining of, and to some extent backfilling, the lake. This can occur frequently when conditions are optimal for algae growth. Future work will consider more closely these issues. While reverse flows can push water entering the SJR from the Ocklawaha up into Lake George, it is likely that the contribution to the lake of nutrients under the current hydrologic regime of the Ocklawaha is small. However, if Rodman dam is removed and nutrient attenuation in the lower reach of the Ocklawaha is lost, its influence on Lake George water quality will have to be re-evaluated, using a discharge hydrograph for the river that represents a natural flow condition.

REFERENCES

Durako, M.J., et al. (1998). Assessment of Fisheries Habitat: Northeast Florida. Florida Marine Research Publications. July,1988 Ferguson, G.E., et al. (1947). Springs of Florida. Geological Bulletin No. 31. Florida Geological Survey. Kroening, S.E. (2004). Streamflow and Water-Quality Characteristics at Selected Sites of the St Johns River in Central Florida, 1933 to 2002. USGS Scientific Investigations Report 2004-5177. McLane, W.M. (1955). The Fishes of the St Johns River System. PhD. Dissertation. University of Florida, August, 1955. Moody, H.L. (1970). Factors in the Decline of the Fishery of the St Johns River, Florida. Florida Game and Fresh Water Fish Commission. Morris, F.W. (1995). Lower St Johns River Basin Reconaissance. Technical Publication SJ95-9, St Johns River Water Management District. Mundy, C., Bergman, M. (1998). The Pollution Load Screening Model: A Tool for the 1995 District Water Management Plan and the 1996 Local Government Water Resources Atlas. Technical Memorandum No. 29. St Johns River Water Management District. Odum, HT. (1953). Factors Controlling Marine Invasion into Florida Freshwaters. p135-155. Bulletin of Marine Science, Vol. 3(2) 1953. pp 135-155. ------, (1953b). Dissolved Phosphorus in Florida Waters. Report of Investigations No9, Florida Geological Survey, State of Florida. Vol. p 1-40. ------, (1956). Primary Production in Flowing Waters. Limnology and Oceanography, Vol. 1(2). pp102-117. ------, (1957). Primary Production Measurements in Eleven Florida Springs and a Marine turtle- grass Community. Limnology and Oceanography, Vol. 2(2)1957. pp 85-97.

Paerl, H.W. et al. (2003). The Effects of Changing Nutrient Conditions on the Structure and Function of the Phytoplankton of the Lower St Johns River. Final Report to the St Johns River Water Management District. 9 December 2003. Phlips, EJ. et al. (2000). Light Availabilty in Phytoplankton Standing Crops in a Nutrient-rich Blackwater river. Limnology and Oceanography, Vol. 45(4), pp 916-929. Phlips, E.J, Cichra, M. (2002). Spatial and Temporal Patterns in the Plankton Community of the Lower St Johns River. Report to the St Johns River Water Management District. 30-Aug-02 Pirkle, W.A. (1971). The Offset Course of the St Johns River, Florida. Southeastern Geology, Vol. (13). pp. 39-59. Steward, J. (1984). Water Quality of the Southern reach of the Middle St Johns River: A Focus on the Drought of 1980 to 1981. Technical Publication SJ84-8, St Johns River Water Management District. Sumner, D.M. and Belaineh, G. (2005). Evaporation, precipitation and associated salinity changes at a humid, subtropical, estuary. Estuaries, Vol. 28(6). pp. 844-855. Sucsy, P.V, Morris, F.W. (2001). Salinity Intrusion in the St Johns River, Florida, Estuarine and Coastal Modeling, Conference Proceedings, ECM9. pp 120-139. ------(2002). Calibration of a 3D circulation and mixing model of the Lower St Johns River. Draft Technical Memorandum. St Johns River Water Management District, Palatka, FL. Toth, DJ. (1999). Water Quality and Isotope Concentrations from Selected Springs in the St Johns River Water Management District. SJRWMD, Palatka, Florida. White, W.A. (1970). The Geomorphology of the Florida Peninsula. Geological Bulletin No. 51. Florida Bureau of Geology.