hi dick and ken,

i've attached below a series of descriptions of several specific drifter
events we found in the 95-99 10-m data.  in particular, there are five
events described (A - G) and several other ideas (H -  ):

A: Quick series of strong SE, E storms: 1/20-2/9, 1995
D: Series of strong SE, NE, S storms; 2/25-3/10, 1996
E: Strong S storm: 3/27 - 4/8, 1997
F: Moderate SW storm: 6/2-6/11, 1997
G: Warm-core ring entrainment: June-July, 1999
H: Dispersion in the Great South Channel: 1999  (idea)

i've also written a bunch of matlab mfiles to analyze and plot the
drifter data, and plan to have both dick and claudia bring back copies
of the entire directory to whoi so you can reproduce all the event
figures.  for each of the above events, there is a single mfile (e.g.,
gb_plotA.m) that will make the figure for that event.  if you want to
plot all drifters for say 1998, use gb_plot98(t1,t2) and it will step
through time in units of 5 days between t1 and t2, my stop-action
version of dick's movie.

a note on my codes. i started with ken's gb10mall.mat file, and added an
8th column containing a simple sequential drifter id number from 1 to
180, and called this gb10mallA.mat  i used this new file in all
analysis, finding it handy to be able to single out individual drifters
by id. none of the event figures show the id's.  for 1999, dick gave me
the file of the 40-m drogued drifter data.  i also made it into a 8-
column file with id number, gb40m99.mat

i will spend a little more time looking for interesting events, but i
think these events should be ok. i have several impressions from looking
at dick's movies and the gb_plot95 through gb_plot99 sequences that may
be useful:

1) how very strong wind stresses can organize the 10-m flow over most of
the bank, with the drifters moving first downwind and then turning
clockwise with time.

2) during the winter months, the drifters do get moved around  over the
bank by these storms, most of which have a southward component, so there
must be some onbank nearsurface flow over the northern flank by
continuity,

3)during the spring/summer when wind driving is much weaker and
stratification stronger, the drifters appear to move more quickly around
the bank and along the southern flank  -- however, there are plenty of
times when one or more drifter will just stop, or reverse course, or
stop and spend a number of days more or less in a limited area on the
shelf, while other drifters continue to flow along the bank.  the idea
that the flow is always organized, or more organized in summer, or
mostly laminar-like seems to me to be more fiction that true.  i don't
know how to quantify this impression.  ken, does the ratio of mean
current to principle axes change with season on the southern flank?
looking, i think i see sufficient drifter stopage along the southern
flank that the mean southwest flow - 2 x standard deiviation should be
close to 0.

4) drifters tend to take two paths when approaching little georges shoal
(near 41.6, 68.4) from the southwest, most pass to the north while some
pass south. i think none pass over it. my guess is that those that pass
south are wind-driven but have not analyzed this.  i think little
georges shoal is the shallowest part of the bank, which separates the
northeast end of the great south channel and the northern flank of the
bank.  one would expect it to exert topographic control over the flow
from the great south channel region, and from scopex ctds, it does seem
to force the alongbank flow to narrow to go around the shoal and speed
up.  the 60 to 100 m isobaths converge sharply approaching the shoals.

5) the exchange of water over the northern flank has always been a
puzzle.  some drifters appear to leave the jet in this area and wander
north in the gulf, later rejoining the current, others just stay in the
jet the entire length.  it would be good to see if there was some
seasonality to this behavior. i would guess that the jet meanders and
drifters get ejected and injected (as in the gulf stream).

hope this helps. i can work more on these descriptions or other things
you think i need to do. also, feel free to edit these or redo the
figures, etc.

cheers,
bob

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Drifter Event A.

The first set of five drifters were deployed within the 60-m isobath on
Georges Bank on January 14-15, 1995. Although these drifters moved
towards the southwest with speeds of 15-20 cm/s on the 18th in response
to a short (1-day) but strong (0.4 N/m^2) southwest wind event, all five
drifters remained within the 60-m isobath for the next 9 days (Figure
A1). Sustained winds (0.2-0.4 N/m^2) towards the southeast during the
24-30th drove all the drifters southward at speeds of 5-15 cm/s, with
three drifters crossing the 60-m isobath on the southern flank.  During
the next five days, the wind stress was more variable in direction and
the drifters moved generally westward along the southern flank. A gale
hit the bank on Febuary 4, with sustained winds (0.6 N/m^2) towards the
east- southeast for the next 4 days.  The drifters responsed quickly and
moved towards the east and then turning toward south at speeds of 18-28
cm/s. By the 9th, four of the five drifters had moved to or crossed the
100-m isobath. The drifters continued to be driven southward into deeper
water over the next four days by two short but strong eastward wind
events (0.2-0.4 N/m^2) with the result that four became entrained into a
Gulf Stream warm-core ring by Febuary 14, with the fifth drifter
following the next day (Figure A2).

This example demonstrates one response of 10-m drifters over the bank to
strong wind forcing during very weak stratification.  In particular, the
drifters tended to move down wind initially and then veer clockwise
during very strong, persistent winds. The near-surface water surrounding
the drifters must have moved southward too, so that there must be on-
bank flow on the northern side of the bank by continuity.  While other
examples of similar wind-forced drifter motion will be described, this
is the only example during the GLOBEC field program that we observed all
drifters located within the 60-m isobath being driven southward and off
the bank in roughly 20 days.  We think this is due to the rapid
succession of strong east to southeast wind events over this 20-day
period.  None of the winter storms observed in GLOBEC were able
individually to drive a drifter from the crest of the bank to off the
southern flank.  The succesion of storms in January-February 1995
provided sufficiently persistant and strong east to southeast winds to
move these five drifters off the bank.

Fig. A1. Drifter tracks and wind stress for January 20 - February 9,
1995. Each panel shows the tracks of five drifters over a 5-day period,
starting on January 20.  The 5-day average wind stress vector is shown
in the upper right of each panel. An eastward 0.2 N/m^2 wind stress
vector is shown in the lower righr corner for comparison. [gb_plotA.m]

Fig. A2.  The track of drifter 5 for the period January 15 to February
16, when it has been entrained into a warm-core ring. The 2-day vector-
averaged wind stress is plotted every two days along the track.  An
eastward 0.2 N/m^2 wind stress vector is shown in the upper right for
comparison. [gb_plot_track(5,15,47,1995).m]

=======================================================================
Drifter Event D.

A succession of three strong winter storms passed Georges Bank during
February 25 though March 10, 1996.  The first storm caused sustained
winds (0.2-0.6 N/m^2) towards the southeast during January 25-28, and
was followed by two days of weaker southeast winds (0.1-0.4 N/m^2).  The
second storm produced sustained winds (0.2-0.8 N/m^2) towards the
eastnortheast during March 2-6.  The third storm produced southwest to
southeast winds (0.2-0.6 N/m^2) during March 7-10.

During this 14-day period, seven drifters initally located within the
60-m isobath on Georges Bank were advected to the south and southeast
into deeper water over the outer shelf and slope on the southern flank
(Figure D1).  An eigth drifter starting on the northwest flank was
carried across the southwestern crest of the bank to almost the 60-m
isobath on the southern flank. Seven of these eight drifters continued
towards the southwest along the southern flank, while the drifter that
started at the 60-m isobath on the southern flank was carried offbank
and shortly entrained in a warm-core ring.  The drifter starting on the
Northeast Peak was also carried eastward offbank into the slope water
and lost to the bank.

This example illustrates three features of the near-surface Lagrangian
flow on the bank during weak stratification.  First, a short sequence of
strong southeast and eastward wind events can move water over the crest
to the southern flank (deeper than the 60-m isobath), with some of the
crest water being carried offbank into the slope water.  (This example
is similar to the January-February 1995 already described.)  Second,
during strong sustained winds, the near-surface flow over much of the
bank is spatially coherent, i.e., the flow is roughly in the same
direction with similar speeds during the strongest wind forcing.  Third,
the near-surface flow over the southern flank can be reversed, i.e,
directed towards the east and northeast during strong eastward wind
forcing.

Fig. D1. Drifter tracks and wind stress for February 25 - March 16,
1996. Each panel shows drifter tracks over a 5-day period, starting on
February 25.  The 5-day average wind stress vector is shown in the upper
right of each panel. For comparison, a 0.2 N/m^2 east wind stress vector
is shown in the lower right of each panel.
[use gb_plotD to make figure]

=======================================================================
Drifter Event E

On March 31, 1997, a strong low pressure system passed to the south of
Georges Bank, bringing winds above 15-17 m/s (0.5-0.8 N/m^2) first
towards the southwest then veering to south for the next three days
through April 3. Winds on the bank were significantly weaker (less than
5-10 m/s) both before and after this storm, so that this strong
sustained southward wind stress appears as an isolated forcing event.

When the storm first struck the bank, 15 drifters were located either on
the bank or in the Gulf of Maine near the northwest flank of the bank
(Figure E1). Prior to the storm, these drifters moved relatively slowly
(with mean speeds less than ~10 cm/s except for one moving with a mean
speed of ~ 17 cm/s on the eastern tip of the bank), and exhibited little
spatial coherence, due in part to the weak wind forcing during this pre-
storm period.

During the four-day storm (March 31 to April 4), all drifters moved
towards the southwest, with mean speeds that varied from a minimum of ~
10 cm/s for the drifter in Wilkenson Basin to a maximum of ~ 35 cm/s for
the drifter that crossed the Great South Channel and moved westward
south of Nanucket Shoals.  This last drifter started in the eastern side
of the Great South Channel, very quickly accelerated to a peak speed of
~73 cm/s towards the southwest on late April 1, then slowed as it turned
west.  The five drifters initially located on the southern and eastern
flank moved along the southern flank at mean speeds ~18 to 34 cm/s, with
peak speeds from ~24 to 57 cm/s.

During the next four days, with little to no wind forcing, the drifters
located along the northeastern and northern flank of the bank reverse
their storm-driven motion and move towards the northeast, in the
direction of the normal clockwise circulation around the bank.  One of
these drifters started on the crest and was carried west into the
current that flows along the northern side of Great South Channel and
clockwise around Little Georges Shoal.  The five drifters on the
southern flank continued to move westward along the shelf but at
different speeds.  The two drifters furthest off the shelf south of the
Great South Channel continued to move rapidly down the shelf, with mean
speeds of ~32 and 38 cm/s and peak speeds of ~52 and 46 cm/s.  The two
drifters near the 100-m isobath moved more slowly, with mean speeds of
~15 and 19 cm/s and peak speeds of ~35 and 36 cm/s.  The one mid-shelf
drifter was even slower (mean speed ~11 cm/s) and started to reverse
direction on April 7.

This example shows that a sufficiently strong and persistent southward
wind stress can drive the near-surface Lagrangian flow towards the
southwest over the entire bank.  In particular, the normally
northeastward along-bank flow on the northern flank can be reversed and
the normally southwest flow on the southern flank can be enhanced.
During this particular storm, the three drifters initially located on
the southern flank were advected west to the New England shelf in
roughly 8 days.  It is not clear why these three drifters continued to
move rapidly after the storm passed, except that it seems likely that
the two outermost drifters became entrained in the shelfbreak jet. [[is
there any satellite avhrr imagery that would confirm that drifter 89 is
still in shelfwater]]

Fig. E1. Drifter tracks and wind stress for March 27 - April 8, 1997.
Each panel shows drifter tracks over a 4-day period, starting on March
27.  The 4-day average wind stress vector is shown in the upper right of
each panel, and a 0.2 N/m^2 wind stress scale vector is shown in the
lower right for comparison.
[use gb_plotE to make figure]

=======================================================================
Drifter Event F

On June 2, 1997, a second strong isolated storm passed to the south of
Georges Bank, bringing moderate winds to the southwest (0.1-0.2 N/m^2)
until the 5th when the winds veered to south and the stress increased to
0.4-0.8 N/m^2 for one day.  The winds then relaxed back to 0.1-0.2 N/m^2
towards the southsouthwest for the next 3 days.

Ten drifters were on the bank prior to the storm. During the storm,
these drifters continued to move generally towards the southwest along
the local topography, but with differing speeds (Figure F1).  The three
drifters on the eastern flank moved southwest with mean speeds of only
~7-9 cm/s (peak speeds ~18-23 cm/s), while the two drifters at the
eastern end of the southern flank moved more westward at a mean speed of
~15 cm/s (peak speed ~26-28 cm/s).  The three more southern drifters on
the southern flank moved along the shelf more rapidly, with mean speeds
of ~17-31 cm/s (peak speeds ~28-45 cm/s). The two drifters on the crest
moved slowly west. (The path of the one drifter caught in an eddy in the
Northeast Channel is intriguing. A south-southwest wind stress should
cause flow into the Gulf, with perhaps a compensating outflow through
the Northeast Channel. The rapid advection of the drifter out the
channel along its southern side may reflect such a flow.--pure
speculation)

During June 7-9, despite moderate winds (0.1-0.2 N/m^2) towards the
south, all drifters along the southern flank basically stopped, and some
even began to move south and east. Why this occurred is unclear.
Starting first on the eastern flank, these drifters start to move
southwestward along the southern flank.  The one drifter south of the
Great South Channel on June 9 remained in that area for more than 10
days before moving west.

The flow response to this southward wind stress event differs in several
aspects from that observed during the March 31 storm.  First, this
second storm was weaker, and the near-surface wind-driven flow was
weaker and only marginally coherent over the bank.  Second, the mean
wind stress during this second storm was oriented more to the west than
the March 31 storm, which may explain why seven of the eight drifters on
the southern flank moved onbank to shallower water during the storm.

Six of the southern flank drifters were working for the 17-day period
prior to the storm when there was little wind forcing.  Their paths
(panel 4, Figure F1) exhibit the classic around-bank flow, with mean
speeds varying from ~8 cm/s for the shallowest drifter on the Northeast
Peak to up to ~15-20 cm/s for the more offbank drifters. Note that four
of these drifters move off to close to the 2000-m isobath, yet remain
part of the shelf flow.  The three drifters that make the clockwise arc
offbank near 67W are thought to be part of the shelfbreak jet and
frontal system.

What happened during June 7-9 to stop the westward flow along the
southern flank of the bank remains unknown.  It is also unclear why the
around-bank flow appeared to return to normal first on the eastern end
of the bank and last south of the Great South Channel.

Figure F1. Drifter tracks and wind stress for May 29 -June 13, 1997.
Each panel shows drifter tracks over a 5-day period, starting on May 29.
The 5-day average wind stress vector is shown in the upper right of each
panel, and a 0.2 N/m^2 wind stress scale vector is shown in the lower
right for comparison.  The lower right panel shows the trajectories of
six of the drifters shown in the other panels for the period May 16-June
11.  [ gb_plotF to make figure]

=======================================================================
Drifter Event G

During late June 1999, a large Gulf Stream warm-core ring was located
southeast of Georges Bank, centered near 40.8N, 65W. As this ring moved
southwest along the southern flank of the bank, AVHRR imagery shows the
ring begin to pull off cooler shelf water, forming a large clockwise
streamer of shelf water around the northeastern half of the ring (Figure
G1).  During the first half of July, several drifters deployed on the
bank were entrained in the developing streamer.  The wind stress on the
bank during late June through July was generally weak (less than 0.05-
0.1 N/m^2) and oriented towards the northnortheast (Figure G2). As a
result, the drifter motions observed during this period were not
noticably influenced by wind-forcing, and provide some insight into the
Lagrangian kinematics of streamer formation.

In order to investigate the vertical structure of the Lagrangian flow on
the bank, some drifters drogued at 40 m were deployed during 1999 along
with the standard drifters drogued at 10 m.  On June 21, three pairs of
10-m and 40-m drifters were launched in a line across the northern flank
near 42.1N, 66.9W (panel 1, Figure G2).  These drifters joined six
additional 10-m drifters and three 40-m drifters already located on the
southern flank.  During the next 3-5 days, two of the three 10-m
drifters deployed on the northern flank moved quickly (~20-25 cm/s) off
the bank to the east and did not return.  The remaining 10-m drifter
from the northern flank moved more slowly (~9-12 cm/s) south over the
next 8 days, turning to the southwest on June 28 between 66.5-67W. The
three 40-m drifters moved slightly faster (~13-16 cm/s) towards the
southeast over the eastern tip of the bank.  During this period, the 10-
m drifter near the 60-m isobath continued to move southwest alongbank,
and the three 40-m drifters also moved towards the southwest until about
June 25-26, when two reversed direction, and the net southwest movement
slowed. the two 10-m drifters just inside the the 60-isobth on the
southwest corner moved very little.

During June 29 - July 4 (panel 2, Figure G1), the three 40-m drifters on
the eastern flank turned and moved southwestward.  The two shallower
drifters moved at similar speeds (~12 cm/s) initially and then slowed on
the 4th, while the outermost drifter followed the 100-m isobath at a
faster speed (~20 cm/s) until the 4th when it accelerated and turned
counterclockwise to the southeast and crossed the 2000-m. One of the
three nearby 10-m drifters moved to near the 60-m isobath and moved
southwest (~12 cm/s) while the other two moved south across the shelf at
similar speeds, with the eastern most reaching the 100-m isobath on the
4th.  The three 10-m drifters near the 60-m isobath to the southwest
continued to move generally towards the southwest slowly (vector speed
~3-8 cm/s). The three 40-m drifters near 67.6W moved in a complex
clockwise pattern, with the more onbank drifter turning south on July 1
and accelerating (to ~35 cm/s) as it moved more southwestward, with the
other drifters turning east over the next two days.

During July 4 - July 9 (panel 3, Figure G2), one 40-m and two 10-m
drifters were carried off the shelf in the streamer near 66.5W. The
group of three 40-m drifters initially near 67.6W moved eastward until
the 6th when the shallowest drifter turned west, the next drifter turned
north, and the deepest drifter continued to accelerate eastward,
crossing the 2000-m isobath on the 7th at ~70 cm/s as it entered the
streamer.  The 10-m drifter near the 60-m isobath near 68.1W turned
southeast on the 5th, and moved rapidly to the 100-m isobath and turned
east at speeds above 50 cm/s to cross the 2000-m isobath into the
streamer on the 9th.  The 10-m drifter already near the 100-m isobtah
near 66.5W continued to accelerate towards the southeast as it moved
into the streamer.  The 40-m drifter near the 2000-m isobath on the 4th
continued to move southeast and began a small cyclonic path centered
near 40.05N, 65.95W.

During July 9 - July 19 (panel 4, Figure G2), the one 10-m drifter
starting near 40.95N,66.8W moves slowly (~5-10 cm/s) southeast across
the upper slope until rapidly accelerating eastward on the 11th. While
this 10-m drifter passed within 10 km of a 40-m drifter, the one 10-m
and four 40-m drifters remaining on the southern flank continue to move
along bank towards the southwest, with all but one of the 40-m drifters
eventually turning north through the Great South Channel. The one 40-m
drifter caught in a small cyclonic eddy (rough estimate of v/rf ~ 0.3)
made two circuits before joining the streamer.

The drifter data taken during this streamer formation and warm-core ring
entrainment event suggests the following interpretation.  Near-surface
water may been drawn from as shallow as the 60-m isobath into a
streamer, while the source region for subsurface water near 40-m depth
appears to be restricted to the outer shelf, perhaps as deep as the 100-
m isobath.  Several other entrainment events on the southern flank were
captured by drifters during the 1995-99 field effort.  These events
exhibit similar behavior, although none were sampled as well as this
July 1999 event.  Analysis of hydrographic data collected during the
1983 Warm Core Ring Experiment suggests that the shelf water found in
streamers originates over the outer shelf (Schlitz, 2001).  Additional
measurements are needed to see if the source region for entrained water
is such a sharp function of depth.

Figure G1.  AVHRR image of event. perhaps sequence.

Figure G2.  Drifter tracks for June 19 - July 19, 1999. Each panel shows
drifter tracks over a 10-day or 5-day 5-d period. The 10-m drifter
tracks are shown in blue, with a black spot at the head; the 40-m
drifter tracks in red, with a black spot at the head. The panel-average
wind stress vector is plotted at the center of the bank, with a 0.1
N/m^2 scale vector shown in the upper left for comparison.
[use gb_plotG to make figure]

=======================================================================
event H:

During May 1999, a cluster of 5 10-m drifters moved southwest along the
60-m isobath during a period with relatively weak wind forcing (maximum
stress ~0.3 N/m^2 for a few hours, mean stress less than 0.03 N/m^2)
(Figure H1). Around May 30, these drifters entered the Great South
Channel, and four of the five continued west along the shelf after about
10 days.  The other drifter turned north and appeared to join two other
drifters that had be relatively stationary in the GSC since early May.
These drifters then moved north through the GSC.  During this May-June
period, one 40-m drifter moved along the southern flank and followed the
10-m drifters west along the shelf.

Two of the original cluster of 5 drifters started 35-km apart on the
mid-shelf near 41.3N, 66.8W on April 30, were driven to near the 60-m
isobath by a short westward wind stress, then moved along the 60-m
isobath until reaching the Great South Channel where they moved west to
the 60-m isobath on Nantucket Shoals on June 10.  During this 35-day
period, these two drifters had a mean separation of ~12 km (they were
within 1 km twice within 5 days) and a mean speed ~6 cm/s. On June 11,
these two drifters finally diverged, with one continuing westward along
the outer shelf, and the other moving northwest onto Nantucket Shoals.

This case provides one example of drifters remaining clustered as they
move along the 60-m isobath along the southern flank, perhaps a
consequence of surface convergence associated with the tidal mixing
front that is located near the 60-m isobath. (Note that there are plenty
of counter-examples, where drifters do not remail clustered).  It also
illustrates the fickle nature of flow in the Great South Channel. As
noted by previous drifter (Limeburner,..) and model studies (Riddenrink
and Loder,..) and this study, drifters located near each other can end
up following quite different paths out of the GSC area.

Figure H.   gb_plotH.m