Atlantic Meridional Overturning Circulation Slowdown Causes Widespread Cooling In The Atlantic

Stuart A. CunninghamScottish Association for Marine Science

Stuart.Cunningham@sams.ac.uk

Christopher D. Roberts3, Eleanor Frajka-Williams2, William E. Johns4, Will Hobbs5, Matthew D. Palmer3, Darren Rayner1, David A. Smeed1, Gerard McCarthy1

1National Oceanography Centre Southampton,2University of Southampton, 3Met Office, Exeter, 4RSMAS, University of Miami, 5IMAS, Hobart.

OHC error ~=0.2x1022 J

A cold subtropical North Atlantic

Spatial and Temporal Pattern of Ocean Heat Content in the North Atlantic

(1991-2010 seasonal cycle removed & 0 to 2000 m)

Enhanced Ocean Data Assimilation and Climate Prediction EN3 v2a gridded objective analysis of quality-controlled sub-surface temperature observations (Ingleby and Huddleston 2007, http://www.metoffice.gov.uk/hadobs/en3/).

OHC (1022 J) : Feb 2004

OHC (1022 J) : Aug 2004

OHC (1022 J) : April 2005

OHC (1022 J) : Aug 2005

OHC (1022 J) : Sept 2006

OHC (1022 J) : May 2007

OHC (1022 J) : Oct 2009

OHC (1022 J) : March 2010

OHC (1022 J) : Jan 2011

OHC (1022 J) : Sept 2011

Observing the AMOC and Associated Heat Flux

41°N

Johns, W. E., et al., (2011). "Continuous, array-based estimates of Atlantic Ocean heat transport at 26.5°N." J. Clim. 24(10): 2429-2449.

Hobbs, W. R. and J. K. Willis (2012). "Midlatitude North Atlantic heat transport: A time series based on satellite and drifter data." J. Geophys. Res. 117(C01008): doi:10.1029/2011JC007039.

Subtropical Atlantic Heat Budget (0-2000m)

S’26.5°N N’41.0°N

F’surface flux

0 m

2000 m

Meridional Heat Transport and

Divergence1/4/04 to 31/3/09 Mean SD

26.5°N 1.26 0.11

41°N 0.48 0.11

41-26.5°N -0.77 0.2

1/4/08 to 31/3/10 Mean SD

26.5°N 0.85 0.26

41°N 0.35 0.14

41-26.5°N -0.5 0.23

Ocean Heat Content

ERA-Interim surface flux anomalies

Relative Heat Content Change

Palmer & Haines (2009): Estimating oceanic heat content change using isotherms, J.Clim, 22,

What generates the MLD (q>14°C) temperature anomalies?

Summary

1. Sustained cooling in upper 2 km of subtropical Atlantic between 2010-2012.2. OHC change partitioned equally between the seasonal mixed layer >14°C and deep

ocean.3. Reduced AMOC at 26.5°N is the largest contributor to reduced MHT divergence.4. In seasonal mixed layer heat loss is due to atmospheric heat loss (60%) and MHT

divergence (40%).5. Results emphasise the role for the ocean in the North Atlantic climate system on

seasonal to interannual timescales and suggest a role for the AMOC in setting sub-surface temperature anomalies.

6. These anomalies have previously been linked to re-emerging SST patterns and subsequent NAO anomalies.

Need to do a bit about re-emergence/SST patterns and link to NAO (Taws).Then say we identify the OHC change due to divergence as responsible.

OHC Errors

Ekman and Geostrophic Heat Transport Variability

Interannual Variability

What happened to the MOC in 2009-10?

Table of annual changes (Std of annual means):Layer 2004-2008 2009 (change)

Mean [Sv] Std [Sv] Mean [Sv]

MOC 18.1 0.88 12.2 (-5.9)

Gulf Stream 31.8 0.26 30.7 (-1.1)

Ekman 2.9 0.36 1.1 (-1.8)

Upper mid-ocean -16.6 0.95 -19.6 (-3.2)

UNADW (1100-3000m) -11.9 0.18 -10.7 (1.2)

LNADW (3000-5000m) -7.8 0.69 -3.8 (4.0)

RAPID MOC: 2009-2010, Ekman constant

What happened to the MOC in 2009-10, not directly due to Ekman?

Table of annual changes (Ekman fixed):

Layer 2004-2008 2009 (change)

Mean [Sv] Std [Sv] Mean [Sv]

MOC 18.7 0.89 14.3 (-4.4)

Gulf Stream 31.8 0.26 30.7 (-1.1)

Ekman 3.6 0.0 3.6 (0)

Upper mid-ocean -16.7 0.91 -19.9 (+3.2)

UNADW (1100-3000m) -12.2 0.32 -11.5 (-0.7)

LNADW (3000-5000m) -8.0 0.66 -4.7 (-3.3)

Longer duration (18-month) slowdown of the MOC:seen in the Gulf Stream, upper mid-ocean and LNADW.

• MOC timeseries and related data products are available from www.noc.soton.ac.uk/rpdmoc• Data from individual instruments are available from www.bodc.ac.uk

Gulf Stream, MOC, Ekman & Upper Mid-Ocean Transports(10-day & 3-month, low-pass filtered)

Gulf Stream

MOC

Ekman

Upper Mid-Ocean

LNADW (3-5km)

UNADW (1.1-3km)

Meridional Heat Transport Time Series

Johns, et al. (2011), Continuous array-based estimates of Atlantic Ocean Heat Transport at 26.5N, JClim

Year Eh (PW)

2004 1.41

2005 1.37

2006 1.36

2007 1.32

2008 1.35

2009 1.09

2010 1.10

Meridional Heat Transport at 26.5°N

Contribution to Qnet by spatially correlated v,T variability across interior from argo data..

QGS → Cable voltage calibrated for temperature transport, (Shoosmith et al., 2005) r = 0.94, σ = 0.1 PWQEk → ECMWF ERA Interim wind stress (daily) • Reynolds SST (weekly)QWBW → Directly calculated from moored CM’s/thermistors in Abaco WB arrayQINT → Zonally-averaged interior transport profile from endpoint geostrophic moorings • Seasonally-varying interior hyrdographic climatology (Hydrobase, R. Curry) merged with argo data.

Ocean Heat Content

RAPID @ 26.5°N1.26±0.11 PW 04-09

Hobbs & Willis @ 41°N0.48±0.11 PW 04-09TF08 error bars

Coupled models (CM2.1, CCSM4)

Radiation balance residual (NCEP,ECMWF,TF08)lGlobal hydro inverse (Ganachaud)Air-sea flux climatology (Large)RAPID & 41N: Direct

Residual Radiation Balance, Climatologies and Direct Estimates

Atlantic Ocean Heat Transport Estimates

Overturning and Gyre Heat Transport

Atlantic Meridional Overturning Circulation Slowdown Causes Widespread Cooling In The Atlantic

Stuart A. CunninghamScottish Association for Marine Science, Oban

Stuart.Cunningham@sams.ac.uk

Christopher D. Roberts3, Eleanor Frajka-Williams2, William E. Johns4, Will Hobbs5, Matthew D. Palmer3, Darren Rayner1, David A. Smeed1, Gerard McCarthy1

1National Oceanography Centre Southampton,2University of Southampton, 3Met Office, Exeter, 4RSMAS, University of Miami, 5IMAS, Hobart.