Surface(Es+mates(of(the(Atlan+c( Overturning(in(Density ...

Surface Es+mates of the Atlan+c Overturning in Density Space in an Eddy-‐Permi=ng Ocean Model

Jeremy Grist1, Simon Josey1

Robert Marsh2

1Na+onal Oceanography Centre, UK 2University of Southampton, UK

Funded by Natural Environment Research Council, UK

Context

Marsh (2000) described (but did not test) a method that might allow ‘the meridional stream function to be largely inferred from surface fluxes alone”. We’ve examined this possibility using output from:

1)  Three IPCC coupled climate models. (100-400 years of GFDL2.1, BCM, HadCM3) (Grist et al. 2009; Josey et al. 2009).

2)  Eddy-permitting (1/4 º) ocean only model (78 years of ORCA-025, ‘NEMO’) (Grist et al., 2012).

Method

Net diapycnal volume flux, G (Θ,ρ) and Diapycnal density fluxes D (Θ,ρ) in an idealized North Atlantic.

Marsh (2000), Walin (1982)

F (θ,ρ)


Marsh (2000), Walin (1982)

F (θ,ρ)

!

G(", #) = F(", #) - $Ddiff (",# )$# + C(", #)

!

F(", #) = $Din (",# )$#

Method


Assuming incompressibility and steady state of water masses, the meridional streamfunc+on then: ψ (Θ,ρ) = G (Θ,ρ)

Marsh (2000), Walin (1982)

F (θ,ρ)

!

F(", #) = $Din (",# )$#

Method

!

G(", #) = F(", #) - $Ddiff (",# )$# + C(", #)


Assuming incompressibility and steady state of water masses, the meridional streamfunc+on then: ψ (Θ,ρ) = G (Θ,ρ)

Marsh (2000), Walin (1982)

F (θ,ρ)

!

G(", #) = F(", #) - $Ddiff (",# )$# + C(", #)

!

F(", #) = $Din (",# )$#

Method

!

G(", #) $ F(", #) = %Din (",# )%#

‘Surface-‐Forced’ Streamfunc+on (Sv)

Latitude (ºN) 20 40 60 80

20 40 60 80 Latitude (ºN)

Latitude (ºN) 20 40 60 80

GFDL2.1 25.5 26.0 26.5 27.0 27.5 28.0

25.5 26.0 26.5 27.0 27.5 28.0

HadCM3

25.5 26.0 26.5 27.0 27.5 28.0

BCM2.0

σ0

σ0

σ0

Grist et al. (2009)

Marsh (2000)

Correla+on of AMOC (48N) vs Surface Forced Index (SFI)

GFDL2.1 HadCM3

BCM2.0

• Year on year SFI ≠ AMOC

•  But significant correlation when SFI leads by a few years in all models.

•  Averaging the SFI over preceding years may give a useful estimate of AMOC variability.

Correla+ons Maximum ‘Surface-‐Forced’ (SFI) & Overturning Stream func+ons

GFDL2.1

HadCM3

BCM2.0

AMOC (48N) solid lines and past average of SFI (dashed) 3 yrs 15 yrs 10 yrs 6 yrs

0.38

0.37 0.38 0.06

0.59 0.66 0.60

0.19

0.31 0.44 0.51 0.56

15-44% of interannual variability

SFI & AMOC in the Coupled Models

We compared our estimate to AMOC (z,θ)

The theory suggests compare With MOC (σ or ρ,θ). (Density as a vertical coordinate!)

Our surface forced streamfunction shows similarities to MOC (σ,θ).

AMOC(z,θ)

MOC(σ,θ)

F(σ,θ)

1000

3000

5000 Dep

th (m

)

21 23 25 27

Pot

. Den

sity

, σ (k

g m

-3)

21 23 25

27 P

ot. D

ensi

ty, σ

(kg

m-3

)

20 30 40 50 60 70

20 30 40 50 60 70

20 30 40 50 60 70 Latitude

Latitude

Latitude

AMOC in ¼º NEMO Model


The theory suggests compare With AMOC (σ or ρ,θ). (Density as a vertical coordinate)

Our surface forced streamfunction shows similarities to MOC (σ,θ).

AMOC(z,θ)

AMOC(σ,θ)

F(σ,θ)

1000

3000

5000 Dep

th (m

)

21 23 25 27

Pot

. Den

sity

, σ (k

g m

-3)

21 23 25

27 P

ot. D

ensi

ty, σ

(kg

m-3

)

20 30 40 50 60 70

20 30 40 50 60 70

20 30 40 50 60 70 Latitude

Latitude

Latitude

AMOC in ¼º NEMO Model


The theory suggests compare With AMOC (σ or ρ,θ). (Density as a vertical coordinate)

Our surface forced streamfunction shows similarities to AMOC (σ,θ).

AMOC in ¼º NEMO Model AMOC(z,θ)

AMOC(σ,θ)

F(σ,θ)

1000

3000

5000 Dep

th (m

)

21 23 25 27

Pot

. Den

sity

, σ (k

g m

-3)

21 23 25

27 P

ot. D

ensi

ty, σ

(kg

m-3

)

20 30 40 50 60 70

20 30 40 50 60 70

20 30 40 50 60 70 Latitude

Latitude

Latitude

20 30 40 50 60 700

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

r2

Latitude

0.9

0.7

0.5

0.3

0.1

r2

20 30 40 50 60 70 Latitude

0.9

0.7

0.5

0.3

0.1

r2

Fraction AMOC explained By Surface Fluxes

AMOC (Z) HadCM3 Climate Model Josey et al. 2009

20 30 40 50 60 700

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

r2

Latitude20 30 40 50 60 70 Latitude

0.9

0.7

0.5

0.3

0.1

r2

AMOC (z) NEMO Ocean Model



20 30 40 50 60 700

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

r2


0.9

0.7

0.5

0.3

0.1

AMOC (σ) NEMO Ocean Model

r2


Fraction AMOC Explained By Surface Fluxes


20 30 40 50 60 700

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

r2


0.9

0.7

0.5

0.3

0.1


r2



20 30 40 50 60 70 800

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

r2

MHT vs AMOC (σ)

MHT vs AMOC (z)


20 30 40 50 60 700

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

r2


0.9

0.7

0.5

0.3

0.1


r2




20 30 40 50 60 700

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

r2


0.9

0.7

0.5

0.3

0.1 AMOC (Z) HadCM3 Climate Model Josey et al. 2009


AMOC (σ) Ocean Model + Ekman wind variability

r2



Theory

Theory + Surface Observations

Theory

-100 -50 0-60

-40

-20

0

20

40

60

80

-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

Theory + Surface Observations SSS SST Wind Stress

Net Heat E-‐P

Updated from Grist et al. (2009)

1960 1970 1980 1990 2000 2010

-4

-2

0

2

4

Surface-Forced Overturning Anomaly (Sv) at 48ºN (solid)

with Ekman Contribution (dashed)


Theory

-100 -50 0-60

-40

-20

0

20

40

60

80

-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

AM

OC

Ano

mal

y (S

v)

420-2 -4

Year


Net Heat E-‐P

1960 1970 1980 1990 2000 2010




Theory

-100 -50 0-60

-40

-20

0

20

40

60

80

-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

1960 1970 1980 1990 2000 2010

-4

-2

0

2

4

1960 1970 1980 1990 2000 2010

-4

-2

0

2

4

AM

OC

Ano

mal

y (S

v)

420-2 -4

1960 1970 1980 1990 2000 2010 Year


Net Heat E-‐P

GECCO

Balmaseda et al (2007)

Bingham & Hughes (2009)

Hobbs & Willis(2012)

Summary •  In ¼º NEMO ocean model, the water mass transformation method

can be used to estimate AMOC variability. •  In sub-polar regions the method explains much more variability in

AMOC (σ0) than AMOC (z). •  The surface density fluxes capture much of the decadal signal

while the additional calculation of Ekman transport allows the higher frequency variability to be captured

•  The method shows greatest potential between 33ºN and 54ºN where 70-84% of the AMOC (σ0) variance is explained.

•  As the method relies only on surface observations, estimate of AMOC variability can be made for the reanalysis era.

•  We seek to determine the spread in time series resulting from the different reanalysis / salinity products & reconcile the surface forced signal with other mid-latitude AMOC estimates.

100yr Mean MOC (Sv) GFDL2.1

HadCM3

MOC (Sv) at 48ºN MOC vs SFOC in Coupled Climate Models

•  BCM2.0

100yr Mean SFOC (Sv)

SFOC Index (Sv)

Dep

th (m

) D

epth

(m)

Dep

th (m

)

σ0

σ0

σ0

Latitude

Latitude

Latitude

Latitude

20N

20N

20N

20N

20N

20N

70N

70N

70N

70N

70N 70N

A

r2 – Surface Forced Estimate vs AMOC (σ0)

20 30 40 50 60 70 La+tude

0.2 0.4 0.6 0.8

0.0

1.0

Grist, Josey, Marsh (JGR-‐Oceans, under review, 2012)

Model AMOC (σ0) (black) & Surface Forced Estimates (red) at 50ºN

AMOC (σ0) Surface Fluxes Surface Fluxes (with Ekman)

Surface Fluxes Surface Fluxes (with Ekman)

0

19 10 20 30 40 50 60 70 Model Year

r2

0 2 4

-‐2 -‐4 A

MO

C A

nom

aly

(Sv)

4. Influence of Surface Fluxes on AMOC Variability

20 30 40 50 60 700

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

r2


0.9

0.7

0.5

0.3

0.1 AMOC (Z) HadCM3 Climate Model

AMOC (σ) Ocean Model

AMOC (σ) Ocean Model + Ekman wind variability

r2




1960 1970 1980 1990 2000 2010

-4

-2

0

2

4

1960 1970 1980 1990 2000 2010

-4

-2

0

2

4




Theory

-100 -50 0-60

-40

-20

0

20

40

60

80

-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

AM

OC

Ano

mal

y (S

v)

420-2 -4

Year


Net Heat E-‐P

GECCO

Balmaseda et al (2007) 1960 1970 1980 1990 2000 2010

Surface(Es+mates(of(the(Atlan+c( Overturning(in(Density ...

Documents

Transcript of Surface(Es+mates(of(the(Atlan+c( Overturning(in(Density ...