«SECONDARY CIRCULATION IN A SINUOUS COASTAL PLAIN ESTUARY A Dissertation Presented to The Academic Faculty By Susan Anne Elston In Partial Fulfillment ...»
2.1 Discharge Conditions Discharge conditions recorded at the Atkinson, Georgia, U.S. Geological Survey (USGS) gauging station, on the upper Satilla River vary widely on a seasonal, annual, and historical basis. Historically, over the past 70 years, annual freshwater input ranges from between 150 m3 s-1 and 400 m3 s-1. These discharge conditions are punctuated by freshet events with discharges of between 600 and 800 m3 s-1. A record low freshwater discharge measuring 0.42 m3 s-1 was recorded on September 26, 1990, while a record high freshwater discharge measuring 1902.9 m3 s-1 was recorded on April 6, 1948. A histogram of historical discharge conditions suggests that the river discharge in the Satilla is weakly periodic in nature on a 25-year cycle. Annual discharge conditions in the Satilla River are generally higher during fall and winter months due to the influence of hurricanes and nor’easters (Figure 2.3).
Figure 2.3: U.
S. Geological Survey historical daily mean discharge data for the Satilla River at the Atkinson, Georgia, gauging station.
The 1997 Land Margin Ecological Research Project experiment (LMER 4) and the 1999 Satilla SAT1 and SAT 2 experiments occurred, respectively, shortly before and during a period of extended drought for coastal Georgia. Average discharge conditions ranged from near normal at 80 m3 s-1 during the LMER 4 experiment to significantly below normal at 15 m3 s-1 during the SAT 1 and 2 experiments. Freshwater discharge varied substantially (almost 300 m3 s-1) shortly before and during the LMER 4 experiment from 295 m3 s-1 in late February to slightly over 7 m3 s-1 near the end of April. Freshwater input during the SAT 1 and SAT 2 experiments ranged from 118 m3 s-1 in early February 1999 to just below 1 m3 s-1 in early June of 1999 (Blanton et al. 2001). Individual freshet events punctuated the abnormally low discharge seen in the Satilla starting the summer of 1999 and into early 2000. Short-lived severe weather outbreaks and hurricanes were responsible for a majority of the observed freshwater discharge events (Figure 2.4).
Figure 2.4: (a) The Satilla River discharge during the 1997 LMER 4 experiment.
(b) The Satilla River discharge during the 1999 SAT 1 and SAT 2 experiments. Gray boxes denote the period of mooring deployments for each experiment.
The discharge data for the SAT 1 and SAT 2 experiments provides a unique opportunity to diagnose the response and recovery time of the estuary to individual weather events. The 1999 study period was marked by four extreme weather events: a record severe weather outbreak on January 21-22, 1999, Hurricane Dennis (August 30, 1999), Hurricane Floyd (September 15, 1999), and Hurricane Irene (October 15, 1999).
The January severe weather outbreak and Hurricanes Floyd and Irene had the largest impact on the salinity and pressure fields as indicated by the long-term moorings near river kilometers 8 and 40. The January severe weather outbreak and Hurricane Floyd had the largest influence on changes in the freshwater discharge for the Satilla River. In both cases, the response time (or lag time) to a precipitation event exceeding 22 millimeter is about 20 days, resulting in a rapid increase in freshwater discharge and a rapid drop in the subtidal salinity field by approximately 15 PSU throughout the estuary (Figure 2.5). The response of the subtidal salinity field to Hurricane Dennis was less dramatic due to its fast-moving nature and to the predominant up-welling winds associated with this storm.
The recovery period, the length of time to return to pre-freshet conditions, appears to be approximately 70 days, as indicated by the salinity data from the near ocean long-term mooring. These results are in agreement with those of Blanton et al. (2001) and support their hypothesis that in estuaries similar to the Satilla, that there would likely be a rapid response in the salinity field to a freshwater discharge event and a delayed response in the recovery period to pre-freshet conditions.
Figure 2.5: (a) The Satilla River discharge during the SAT 1 and SAT 2 experiments.
(b) The subtidal salinity measured at the moorings 3 – 7 (red, green, cyan, blue, and magenta lines, respectively) in the Satilla River during the 1999 SAT 1 and SAT 2 experiments. The black line represents the salinity field measured at the near ocean longterm mooring. Gray boxes denote the period of mooring deployments for each experiment.
2.2 Salinity Variations The salinity structure in the Satilla River varies widely on the tidal, fortnightly, and seasonal cycles. This variation is due to a combination of factors including synoptic weather patterns, severe storms (hurricanes and nor’easters), tides, and freshwater discharge. From previous studies, it has been found that the largest contributors to changes in the salinity structure of the Satilla River are severe storms which produce down-welling or up-welling winds (trapping water or flushing water out of the estuary), tides, and freshwater discharge (Elston 1998, unpublished data; Dame et al. 2000;
Blanton et al. 2001).
2.2.1 Tidal Variations in Salinity In the Satilla, over a 25-hour tidal cycle, the salinity at a fixed location varies by about 6 PSU at neap tides to about 15 PSU during spring tides. Correspondingly, the vertical structure of tidal salinity varies between stratified conditions at neap tides to nearly well-mixed conditions at spring tides. The actual rate of change in tidal salinity between low water and high water depends largely on the strength of tidal mixing, which is modulated by fortnightly changes in tidal amplitude. Tidal straining, due to the interaction of the sheared axial currents with the longitudinal salinity gradient, can also modulate tidal salinity by amplifying vertical stratification on ebbs and reducing vertical stratification on floods (Simpson et al. 1990).
Determining the tidal excursion, the distance over which a water parcel travels during one phase of the tidal cycle (ebb or flood), is a good way to estimate the effect of tidal advection on the axial salinity field. The tidal excursion is given by the expression
where U s, rms is the root-mean-square (rms) streamwise velocity (m s-1) and ω M 2 is the frequency (s-1) of the dominant tidal constituent (here, M2). In the Satilla River, the average value of lM 2 is 12 kilometers. The tidal excursion also has a fortnightly modulation of approximately 7 kilometers, from 8 kilometers during neap tides when the rms axial velocity is around 0.5 m s-1 to 15 kilometers during spring tides when the rms axial velocity is around 1.0 m s-1. The tidal excursion estimates agree well with the observed changes for the Satilla River in the axial salinity difference observed at a fixed location between low and high water.
2.2.2 Fortnightly Variations in Salinity As with variations in tidal salinity, the fortnightly variations in salinity for the Satilla River are largely affected by changes in tidal amplitude and river discharge. In addition, observed changes in the lateral and vertical salinity structure between spring and neap tides are likely related to the geomorphology of the Satilla River. Similar to ebb dominant estuaries of coastal South Carolina, the Satilla features a sinuous thalweg (deep channel) that is separated from a secondary shallow channel by a series of shoals (Sexton and Hayes 1996). At spring tides, communication is limited between the two channels to times of high water and is virtually non-existent during times of low water.
Based on observations, conditions at neap tide show a vertically stratified system with ( ∆s ) z = 7 PSU (13 to 21 PSU) in the deep channel and ( ∆s ) z = 2 PSU difference over the shoal region (15 to 17 PSU) with a nearly well-mixed lateral structure of about ( ∆s )h = 2 PSU (ranging from 13 to 15 PSU) across the channel. Not surprisingly, due to the slower axial currents associated with neap tides and the sinuous nature of the Satilla, the vertical density gradient is significantly greater than the horizontal density gradient.
This suggests that the currents at neap do not provide enough mixing energy to overcome strong vertical stratification. The well-mixed shoal area is shallow and requires less energy input locally to mix the water column. The salinity in the shoal area is approximately equal to the mean value of the salinity in the deep channel, indicating that there is communication between the two channels during neap tides.
At spring tides, the salinity structure reflects a laterally stratified system with ( ∆s )h = 5 PSU (13 to 18 PSU) from bank-to-bank and a nearly well-mixed vertical structure ranging from ( ∆s ) z = 2 PSU (13 to 15 PSU) in the deep channel to ( ∆s ) z = 0 PSU (vertically well-mixed at 17 PSU) over the shoal region. As anticipated, the mixing power associated with the high velocity spring tide current easily overturns the vertical density gradient and homogenizes the water column. However, the same current increases the horizontal shear and the lateral density gradients as the water flows around the numerous channel bends in the river. Consequently, less dense water is often found in the deep channel. The lighter water is separated from the heavier water which is forced into the shoal region near the channel center. Surface fronts are often observed upriver of the shoal region during an incoming tide where waters of different salinities collide and partially re-mix before entering the next channel bend. In combination with curvature effects, the resultant fortnightly salinity structure of the Satilla River appears to shift from a vertically stratified two layer system at neap tides to a laterally stratified two channel system at spring tides. These observations are discussed further and shown in Chapter 7.
2.2.3 Seasonal Variations in Salinity Seasonal variations in salinity are largely affected by changes in freshwater discharge (Blanton et al. 2001). The seasonal differences in the Satilla River axial salinity distribution can be observed in Figure 2.6. The upper panel shows a synoptic high water survey done in February 1999 under moderate discharge conditions (80 m3 s-1). Similarly, a synoptic high water survey done in December 1999 under low discharge conditions (10 m3 s-1) is shown in the lower panel. For reference, river kilometer 15 is just east of Bailey Cut and river kilometer 20 is just west of Bailey Cut.
The end of Crows Harbor Reach is at about river kilometer 26. Ceylon is located near Satilla River kilometer 32 (cf. Figure 2.6 for river kilometer locations).
In February 1999, the mixing zone, defined here as in Blanton et al. (2001) by a median salinity of 15 PSU, was observed between the mouth of St. Andrews Sound at the ocean (0 kilometer) and Crows Harbor Reach. In December 1999, the mixing zone shifted upriver by approximately 15 kilometers and was located between the west end of Bailey Cut and at a site past Ceylon. Due to logistics, the freshwater boundary was not detected during the December 1999 survey; however, the salinity was found to exceed 10 PSU near Woodbine, Georgia, located near river kilometer 40. Typical salinity values recorded at Woodbine, Georgia, are around 2 PSU as determined from several bi-monthly synoptic surveys and a four month monitoring station.
A series of synoptic field observations suggest that the longitudinal salinity gradient varies from 2 PSU km-1 during high discharge conditions to 0.75 PSU km-1 during low discharge conditions. Between February 1999 and December 1999, it appears that the decrease in freshwater discharge not only reduced the axial salinity gradient, but also shifted the mixing zone upriver by a considerable distance. These results are similar to those found by Blanton et al. (2001) in the Satilla River and by Alexander et al. (2003) in the Ashepoo – Combahee – South Edisto (ACE) River Basin. Both groups found a similar relationship to changes in freshwater discharge, freshwater volume, and the axial salinity gradient. During high flow conditions, these two studies found that the trend in the axial salinity gradient was exponential in nature, wherein the salinity decreased rapidly upriver from the ocean in an asymptotic fashion. Likewise, during low flow conditions, these two studies found that the trend in the axial salinity gradient was quadratic in nature, wherein the salinity decreased slowly near the ocean and then decreased more rapidly closer to the freshwater limit at 0 PSU (Blanton et al. 2001;
Alexander et al. 2003).
Seasonally, the axial salinity distribution in the Satilla River appears to be governed by the classical estuarine (or gravitational) circulation, in which there is a net flow of water seaward in the surface and a net flow of water landward at depth. As discharge decreases, the freshwater volume decreases, and the barotropic pressure head necessary to flush out the estuary also decreases. This allows for more effective ‘salt creep’ in the estuary, a condition that reduces the axial salinity gradient, shifts the mixing zone upriver, and increases the length of salt water intrusion.
Figure 2.6: (a) Synoptic survey of high water salinity in the Satilla River under moderate discharge conditions (February 1999).
(b) Synoptic survey of high water salinity in the Satilla River under low discharge conditions (December 1999).
3. SECONDARY CIRCULATION Secondary circulation is a well known, little-studied physically driven and/or dynamically driven process that significantly affects the lateral and vertical mixing processes in both air and water. A secondary circulation (also known as a transverse circulation, a secondary current, or a secondary flow) is a field of fluid motion considered to be superimposed on the primary field of motion generally through the action of friction or centrifugal acceleration and exists in the vicinity of solid boundaries (Huschke 1989).
Secondary circulation is a broad and often confusing term used to describe several mechanisms whose result is to vertically overturn the water column along the secondary (transverse) axis of the channel. As such, secondary circulation may be composed of one or more lateral cells that can potentially reduce or enhance local gradients. In the field, a physical manifestation of secondary flow is often observed as one or more convergent foam or divergent slick lines along the longitudinal axis of the estuary at the interaction site between two adjacent secondary cells (Figure 3.1).