Western Arctic Shelfbreak Eddies: Formation and Transport

1
Western Arctic Shelfbreak EddiesFormation and
Transport
Michael Spall Robert Pickart Paula
Fratantoni Al Plueddemann Woods Hole
Oceanographic Institution
Chukchi Sea, Sep 2004 (Photo by C.A. Linder)
2
A complex circulation

• Multiple lateral sources
• - Deep warm Atlantic layer
• - Pacific inflow
• - River runoff

• The cold halocline shield

• Halocline origin
• probably the shelves

3
Halocline Ventilation

transport of properties
Shelf-Basin exchange intimately connected to
boundary current dynamics The processes
represented by the big arrow are not well known
4
Pacific water inflow, approx 0.8 Sv passes
through the Chukchi SeaForms a
shelfbreak jet along southern Beaufort
SeaAnticyclonic eddies of shelf water often
found in the interior
5
Beaufort Slope Mooring Array 2002-4
6
Mean boundary current sections Aug 2002Jul 2003
Potential temperature (oC)
Alongstream (eastward) velocity (cm/s)

Very narrow, bottom trapped shelfbreak jet Mean
transport only 0.14 Sv, less than 20 Bering
Strait transport
7
Seasonal water masses
First year time series at center of boundary
current

Winter-transformed Pacific water
Alaskan Coastal water

potential temperature (oC)
Present focus is on Winter-transformed Pacific
water, April-July
8
Upwelling storm event November 2002

Height of storm
Percent ice concentration and 10m winds
9
Upwelling storm event November 2002
Alongstream (eastward) velocity (cm/s)
Potential temperature (oC)

Beginning of storm
Beginning of storm
10
Upwelling storm event November 2002
Alongstream (eastward) velocity (cm/s)
Potential temperature (oC)

Beginning of storm
Height of storm
11
Upwelling storm event November 2002
Potential temperature (oC)
Alongstream (eastward) velocity (cm/s)

Height of storm
Height of storm
Greatly disrupts the structure of the boundary
current We consider only free jet periods
here, time periods of strong wind forcing have
been filtered out (10 events)
12
Mean Sections for the unforced data (AprilJune)

Transport 0.4 Sv maximum velocity O(25
cm/s) Weak stratification and low PV due to
convective origin of waters temperature near
freezing Relative vorticity and twisting terms
are significant, even in the mean Lateral
gradient of PV changes sign with depth,
suggestive of baroclinic instability
13
Winter transformed Bering Water may ventilate
halocline
Winter water (Tlt-1.7) is present 4 months per
year Salinities vary between 32 and 34 spanning
the upper and lower halocline These T/S
properties match those found in anticyclones in
the interior
14
Synoptic jet can be much stronger
Velocities O(50 cm/s) Large volume of winter
water Relative vorticity O(- 0.4 f ) Twisting
vorticity O(- f ) Ertel potential vorticity lt 0
15
Eddy being spawned from the boundary current
Temperature (color) overlaid on density
(contours)
note that it is not near a canyon

16
Detailed surveyof an eddy
Hydrographic Survey September 2004

17
Lateral viewof eddy
Average potential temperature (oC) between 125 m
and 185 m

Grey contours are bottom depth (m)
18
Vertical section through an eddy
Temperature (oC) overlaid on density (kgm-3)
geostrophic velocity (cm/s)

19
3-D view of an Arctic eddy
Bounding density surfaces of the anti-cyclonic
eddy

20
Eddies observed in the western Arctic
Eddies are encountered roughly every 100 km of
drift More prevalent directly offshore of
shelf and west of Barrow Canyon 100-200
estimated to exist at any time A lifetime of 1
year implies an eddy is formed every O(2 days)

Plueddemann and Krishfield (1999)
21
Previous theories for eddy formations

• Boundary current instabilities
• - Hunkins (1974), Hart and Killworth
  (1976)
• 1-D linear theory, considered Alaska
  Coastal Current
• Local buoyancy forcing
• - Chao and Shaw (1996, 1998)
• Some studies focus on canyons as source of
  eddies
• - DAsaro (1988) friction and abrupt
  topography
• - Chao and Shaw (2003), dense outflow and
  canyons
• None of these modeling studies have produced
  large numbers of
• strong, anticyclonic eddies that are consistent
  with observations
• In particular, it seems unlikely that eddies
  could be formed every
• 2 days from only 1 or 2 canyons

22
A simple model of the unforced boundary current

• MITgcm primitive equation model
• Initial basin interior PHC3.0 May Climatology
• Shelf outflow T/S/V -1.8oC / 32.5 / 7.5 cm/s
• initialize at rest, run for 180 days, analysis
  on final 100 days

23
Some model details
Resolution 1 km horizonal 27
levels vertical (10 m upper 220m) Viscous
surface drag to simulate pack-ice Horizontal
viscosity Smagorinsky deformation dependent
Typical values D10-5 to 10-4 s-1 ns1
with 1 km grid spacing Ah
1-10 m2 s-1
24
Advantages to this approach

• No need for open boundary conditions
• Restoring region provides potential and kinetic
  energy to the
• boundary current so that it does not spin
  down
• this allows for long simulations, many eddy
  cycles
• The structure of the boundary current is not
  directly specified
• but instead develops as part of the solution
  important for
• the stability of the boundary current

25
Mean sections

Model Mooring data
26
The boundary current is highly
unstable Numerous low pv, anticyclonic eddies
are formed that carry shelf water into the
interior
27
Evolution of model boundary currenteddy
formation

28
Typical eddy structure anticyclonic, velocity
O( 20 cm/s) radius O(10 km) inner solid body
rotation low PV core
Profile compares well with observations Vmax
19 cm/s Rmax 12 km Observed eddies tend
to be smaller, faster (red line from
Plueddemann and Krishfield, 2007)
29
Energetics of the boundary current
Model

30
Property flux by eddies
Snapshot of the tracer field

31
Property flux by eddies
Snapshot of the tracer field

Black - offshore Red along shore Blue
average 50
32
Eddies transport shelf water into basin interior
Approx 50 of shelf tracer is carried offshore
by eddies
But very little mass transport is diverted into
the basin interior
transport streamfunction
33
Property flux vs. momentum flux
density is carried offshore by eddies
but momentum is left near the boundary

34
Measured boundary current transport

Transport through Bering Strait 0.83 Sv
35
Consider an anticyclonic, basin-scale wind driven
gyre (Beaufort Gyre)
Shelf tracer is now carried further into the
basin and to the west. Offshore flux
increased to 80
36
Consider an anticyclonic, basin-scale wind driven
gyre (Beaufort Gyre)
Shelf tracer is now carried further into the
basin and to the west. Offshore flux
increased to 80 Now mass transport is
also diverted into the basin interior. Eddy
momentum flux is sufficient to get
entrained into westward wind-driven gyre.
37
Boundary current stability is
very sensitive to bottom slope 0.01
offshore flux is 7
0.02 offshore flux is 53
0.03 offshore flux is 93 lateral boundary
condition free-slip produces no eddies
shear layer
enhances baroclinic instability
(Mundt et al. 1995
Berloff
and McWilliams 1999) Smagorinsky viscosity
coefficient 0.75 flux is 11
1.0 flux
is 53
1.5 flux is 91
38
Mean zonal velocity near boundary at mid-depth
Instability increases, eddies increase in size,
with increasing ns Width of the boundary current
increases with ns Maximum mean velocity decreases
with increasing ns Width of the shear layer does
not increase with ns Opposite from what is
expected from linear barotropic theory
39

• Summary
• Model boundary current is consistent with
  observations, both in
• terms of the mean and its variability
• Boundary current is highly unstable, produces
  many anticyclonic eddies
• The energetics indicate baroclinic instability
  in both model and data
• Eddies produced by the model are consistent
  with eddies observed in
• the basin interior
• Eddies transport a significant amount of shelf
  water (Pacific origin)
• into the interior of the Beaufort Sea
• Process is very sensitive to model parameters,
  need for better
• understanding of the instability mechanism

40
Eddy properties
Properties (color) overlaid on density (contours)

Vertical slices through the center of the eddy