Mean and seasonal surface current patterns in South Florida coastal seas from
drifter trajectories
Elizabeth Williams, Thomas Lee, and Villy Kourafalou
Rosenstiel School of Marine and Atmospheric Science, University of Miami
During the last seven years, over forty nearsurface CODE-type drifters have been
released on bimonthly intervals in the Shark River Plume off the southwestern tip
of the Florida mainland (25.35 N, 81.23 W). The drifter data have been employed
to study the mean flows and the seasonal variability in the coastal seas adjacent to
Florida Bay. The trajectories show that there is a strong link between the South
Florida coastal waters (southwest Florida shelf, western Florida Bay, the Florida
Keys coastal zone and the Dry Tortugas), as the preferred pathways generally
follow a southeastward route through western Florida Bay and the passages
between the Keys, then westward along the reef tract to the Tortugas. The route
through western Florida Bay is driven primarily by local wind forcing and by a
mean sea level slope between the Gulf of Mexico and the Atlantic. The westward
route along the reef tract is induced by the prevailing westward component in the
local wind and by recirculating gyres and eddies north of the Florida Current. A
clear seasonal pattern in surface trajectories emerges with flow toward the
southeast in winter/spring, northwest in summer, and southwest in fall. This
pattern follows the seasonal cycle of local wind forcing. The most direct pathways
to the Tortugas occur during the strong northeasterlies in fall, while the longest
pathways to the Tortugas occur in summer, due to the southeasterlies that,
although weak, effectively cause a northward surface drift reaching up to 27 N.
However, multiple pathways can be seen in any season, reflecting the variability
in wind forcing.
Wind and moored current measurements are employed to further elucidate the
underlying shelf dynamics. It is found that seasonal changes in the regional wind
forcing produce seasonal differences in the strength and variability of surface
currents on the southwest Florida shelf that are similar to the seasonal changes
shown by the drifters. Current amplitudes are greater in winter than in summer,
following the enhanced wind stress. Moored currents are more southward in the
fall, winter, and spring seasons, changing to northward in the summer, due to the
shift of summer winds to southeasterly. An example of the winter/spring
southeastward pathway is shown in Figure 1. The time series of drifter derived
currents and wind from the Molasses CMAN station indicate a strong wind
influence for the flow along the drifter tract (Figure 1). This is particularly evident
in the v-component (north to south), partially due to flow alignment with the
isobaths, but also in the u-component (east to west).
A simple multiple regression model was employed to analyze the relationship of
the wind components to the drifter derived current components. A cubic spline
was fitted to the drifter positions and interpolated to 6-hourly time intervals,
followed by filtering of the tidal components. The current components were
regressed separately against individual wind components and combined u- and vwind components (multiple regression). The analysis was focused on the coastal
areas which are not directly influenced by strong large-scale flows of the Loop
Current and the Florida Current. We found that approximately 70 to 80% of the
subtidal variance of drifter derived currents on the southwest Florida shelf and
western Florida Bay is due to local wind forcing. Figure 2 displays the model
computed currents versus the drifter derived currents for the period of the
trajectory shown in Figure 1. The agreement is quite good, especially in the vcomponent, which has a high regression coefficient of 0.78. The residual current
is also shown, as the difference between model and drifter currents. The residual
current represents the part of the flow that cannot be attributed to the local wind
forcing.
Elizabeth Williams, University of Miami, RSMAS/MPO, 4600 Rickenbacker
Causeway, Miami, Fl, 33149, Phone: 305-361-4070, Fax: 305-361-4696,
ewilliams@rsmas.miami.edu, Question 1
Figure 1: Time series of drifter derived currents from a sample trajectory (drifter
21014, left panel) and local wind: u-components (upper right panel) and vcomponents (lower right panel).
Figure 2: Upper panels: time series of currents derived from the drifter trajectory
(thick solid line) and from the multiple regression model (thin solid line); the
difference between the two time series (dashed line) is the residual flow (the part
of the flow that is not driven by the local winds). Lower panels: Scatter plots and
linear correlation between drifter and model currents.