Embed Size (px)
Citation preview
Eddy Heat Fluxes across the Antarc4c Circumpolar Current in Northern Drake Passage
D. Randolph WaAs1, Karen L. Tracey1, Kathleen A. Donohue1, Teresa K. Chereskin2
1Graduate School of Oceanography, University of Rhode Island, USA (E-‐mail: [email protected]) 2Scripps Ins4tu4on of Oceanography, University of California San Diego, USA hAp://www.cdrake.org
1. Local dynamics array 2. Method to iden4fy divergent eddy heat fluxes
6. Context / Summary
References and Acknowledgments
Overview
5. Time series & case study
Session O21A # 1655
THE
UNIVERSITY OF RHODE ISLAND GRADUATE SCHOOL OF OCEANOGRAPHY
è See also Bishop, S.P. AGU Fall Mee4ng Session OS21A, poster #1657. Bryden, H. (1979) JMR 37, 1-‐22. Lenn, Y-‐D., T.Chereskin, J.Sprintall, and J.McClean (2011) JPO 41, 1385-‐1407. Marshall, J., and G. ShuAs (1981) JPO 11, 1677-‐1680. Phillips, H., and S.Rintoul (2000) JPO 30, 3050-‐3076. è cDrake is funded by US Na4onal Science Founda4on Grants ANT-‐0635437 and ANT-‐0636493.
3. Four-‐year mapped EHF/DEHF fields
For ocean mesoscale eddy heat fluxes, the dynamically important component is the divergent field, because it modifies its environment, whereas the non-‐divergent component just recirculates. Helmholtz showed that vector fields are uniquely separable into non-‐divergent and divergent parts. Divergent eddy heat fluxes are driven by nearly depth-‐independent geostrophic eddy currents that can cross the temperature front.
* mesoscale-‐resolving 24 current and pressure recording inverted echo sounders (CPIES) * 3D maps of temperature and geostrophic current vector, T, u * 4-‐years, 11/07-‐11/11 * in maximum SSH variance region between SAF and PF
dept
h (m
)
−5000
−4000
−3000
−2000
−1000
0SAF
PF
64oW 62oW 60oW
62oS
60oS
58oS
56oS
Marshall and ShuAs (JPO, 1981) discussed the importance of separa4ng non-‐divergent and divergent eddy fluxes. They presented a method to separate them if the mean flow is nearly along temperature contours, sa4sfying approximately, ψ ≈ ψ(T). The non-‐divergent part recirculates EHF around <T’2> = EPE extrema.
φ+ φ-
u = ubc + ubt
ubt
ubc
Absolute geostrophic currents u are the vector sum of a ver4cally-‐aligned baroclinic current, ubc, and a nearly depth-‐independent reference current called the barotropic current, ubt, measured at 4000m depth.
White contours = SSH variance
* CPIES mapped eddy heat fluxes (EHF) { u’T’ = u’bcT’ + u’btT’ } × (ρoCp) * flux divergence is dynamically important ++ modifies the T and EPE fields ++ non-‐divergent part just recirculates
* baroclinic part is non-‐divergent, ∇.<ubcʹ′T’> = 0 ++ ubc carries no heat across front * <u’btT’> contains all dynamically important EHF ++ a major part is s4ll non-‐divergent
20
24
24
1 DEC 2007
20
24
24
7 DEC 2007
20
24
24
13 DEC 2007
20
2024
0.05 ° C m s−10.05 ° C m s−10.05 ° C m s−1
19 DEC 2007
dbar−0.4−0.2 0 0.20.4
7 DEC 2007
20 cm s−1
65°W 63°W 61°W
58°S
57°S
56°S
dbar−0.4−0.2 0 0.2 0.4
4. Ver4cal structure
Upper and deep flow fields. On this day a SAF meander crest and trough had deep HI and LO, offset ~λ/4 downstream. In circled region currents turned with depth as ubt crossed the baroclinic front. Black contours: ϕ baroclinic stream func4on Gray contours: SSH from satellite al4metry Colorbar: P’ at 4000 dbar Vectors: current at 4000 dbar
All five panels: 4-‐yr mean eddy heat flux fields at 400m; DEHF vectors enlarged (boAom right). (All four 1-‐yr mean fields are remarkably similar.)
Top: <uʹ′T’> tends to be high where EKE is high; sum of ‘bc’ and ‘bt’ parts, shown in next row.
Bo7om right: DEHF overlaid on mean baroclinic stream func4on (geopoten4al). Downgradient DEHFs (green shading) systema4cally occur along mean crest-‐to-‐trough segment between SAF and PF; peak values ~ ⅕ of others.
• strongest sites have same sign top-‐to-‐boAom
• peaked in 200-‐400m at 4x10-‐3 oCm/s (15KW/m2)
• 50% of flux below 700m • contrast total EHF at
400m is ~75kW/m2 ~5x larger
DEHFs at six sites chosen to represent strong and weak down-‐ and up-‐gradient fluxes.
Meridional DEHF 4me series at site 2 typifies all six sites. The mean accumulates from many episodic short-‐lived events. The short integral 4me scale ~4 days accounts for why 1-‐yr means are similar.
This case exemplifies rapid growth of a SAF meander crest (A) that jointly steepens with deep an4cyclone. (B) is a canonical example of BC instability with upper crest and trough led (~λ/4) by deep high and low pressure centers; note strong poleward DEHFs. (C) intensifies further, becoming more ver4cally aligned. (D) shows a separated ring, characterized by mature ver4cally-‐aligned current structure and therea�er liAle heat flux.
Poleward DEHFs arise from ~8 southward crests of the warmer SAF steepening jointly with a deep an4cyclone, as illustrated in the case-‐study map sequence . Addi4onal ~3 poleward peaks arise when northward troughs of the colder PF develop.
Case-‐study daily snapshots at 6-‐d interval in December 2007, during poleward peak DEHF . Geopoten4al ( ϕ0 re 4000 dbar) solid contours; proxy SSH for SAF and PF (gray contours); deep reference current (gray vectors) and P4000 (colorbar). Green arrows: DEHF at same six sites.
-‐contrast <ubtʹ′T’> at 400m are ~5 4mes larger, poleward but mainly non-‐divergent -‐contrast <uʹ′T’> have similarly large magnitude, but in recircula4ng direc4ons
Ver4cally integrated poleward DEHF is 6-‐19 MW/m across the front (sites 1,2,3).
Compare Phillips & Rintoul (2000) south of Australia ver4cal average EHF in shear coordinates 11.3kW/m2 (ver4cal integral ~40 MW/m). Compare Bryden (1979) Drake P average EHF = 0.16cal/cm2s = 6.7 kW/m2 at 2700m. Compare Lenn, et al. (2011) Drake P XBT +SADCP transects, poleward EHFs in top 300m up to 290±80 kW/m2 (incl Ekman) calculated rela4ve mean stream coordinates. None separate DEHFs from EHFs.
The mean poleward DEHF peaks are associated with several events per year of strong meander growth; the strongest events result in ring forma4on. The growth is generated by joint interac4on with deep eddies, with phase-‐offset such that their currents cross the baroclinic front and release APE consistent with BC instability. This process drives the eddy variability and drives the poleward DEHF.
1500
600
200
4000
D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N−0.1
−0.05
0
0.05
0.1
2008 2009 2010 2011
Location 2; 400 m
mean = −3.56e−03
v btʹ′Tʹ′ di
v (m s−1
°C)
poleward
equatorward
u!T ! 0.025 ° C m s−1
65°W 63°W 61°W
57.5°S
56.5°S
Surface EKE (m2 s−2)2 2.5 3 3.5
x 10−3
u!
b tT ! 0.025 ° C m s−1
Depth (m)−5000 −4000 −3000 −2000 −1000 0
u!
b cT! 0.025 ° C m s−1
T ! 2 ( !C 2)
0 0.2 0.4 0.6 0.8
u!
b tT! 0.025 ° C m s−1
100 125 150 175 200 225 250 275 300 325
SSH! 2 (cm2)
u!
b tT ! d iv 0.0125 ° C m s−1
Middle: <ubcʹ′T’> (purely nondivergent) recirculates around <T’2> extrema; equatorward in many sectors.
<ubtʹ′T’>, bathymetry, and SSH variance (white contours). <ubtʹ′T’> is high and poleward near SFZ where SSH variance is high. Its non-‐divergent and divergent components shown next row.
Bo7om le=: <ubtʹ′T’>nondiv is the larger component roughly parallel to <T’2>; also resembles mean 4000 m flow (not shown).
A B C D
Observed 4-‐yr mean divergent eddy heat fluxes (DEHF) in the Drake Passage are stable es4mates. The strongest DEHFs occur immediately downstream of Shackleton Fracture Zone (SFZ, topographic ridge) between the Subantarc4c Front (SAF) and Polar Front (PF), where meanders and deep eddies jointly develop rapidly via baroclinic instability, and rings pinch off, and SSH variability is highest.
* Divergent eddy flux must be es4mated as residual a�er OA mapping the non-‐divergent part of EHF
Peak mean downgradient DEHF at 400m is (3.6±1.2)x10-‐3 (oCm/s) ≈ 15 kW/m2
Ver4cally integrated poleward DEHF is 6-‐19 MW/m across the front (sites 1,2,3).
−5 0 5x 10−3
−4000
−3500
−3000
−2500
−2000
−1500
−1000
−500
0
C m s−1
n · u!
b tT!d i v
(!Cms" 1)
123456
1 2
34
5 6
u!
b tT! n ond iv 0.025 ° C m s−1
T ! 2 ( !C 2)
0 0.2 0.4 0.6 0.8
