Climate feedback on the marine carbon cycle in CarboOcean Earth System Models J. Segschneider 1, E. Maier-Reimer 1 L. Bopp 2, J. Orr 2 1 Max-Planck-Institute.

Climate feedback on the marine carbon cycle in CarboOcean Earth System Models

J. Segschneider1, E. Maier-Reimer1

L. Bopp2, J. Orr2

1Max-Planck-Institute for Meteorology, Hamburg, Germany2Institut Pierre-Simon Laplace, Gif-sur-Yvette, France

EU FP6 IP 511176 (GOCE)

Background:

Main goal of core theme 5 within CarboOcean: Provide coupled climate carbon cycle simulations

to

• provide predictions of oceanic carbon sources and sinks

• identify and possibly quantify the feedback processes between climate and the oceanic carbon cycle

Atmospheric pCO2 rises because of anthropogenic emissions

Atmospheric temperature rises and hence circulation is changed

Ocean state (temperature and circulation) are influenced by atmospheric conditions, internal dynamics may amplify perturbations

Physical conditions impact on marine carbon cycle and hence oceanic CO2 uptake, which impacts on atmospheric pCO2

Controlling mechanisms for climate feedback on marine CO2 cycle

Possible marine feedback mechanisms

physical pump: - higher SST/reduced solubility - reduced deep water formation/transport to abyss + less ice cover/increased gas exchange at high latitudes ? wind/gas exchange

biological pump: - weakening MOC, less nutrients at surface, less export

alkalinity pump: - acidifcation/less calcification + less silicate/more calcification

Sign indicates impact on oceanic CO2 uptake

Instrument

Earth System Models CM4, COSMOS1 comprising

- Atmosphere dynamics and chemistry LSCE: LMDZ-4, MPIM: ECHAM5

- Ocean dynamics and biogeochemistry LSCE: ORCA/PISCES, MPIM: MPIOM/HAMOCC5.1

- Land biosphere LSCE: ORCHIDEE, MPIM: JSBACH

To be added: NCAR CSM1.4 (some results, as used by Bern group)

BCM (Bergen climate model)

Forcing protocol (same as C4MIP):

historical 20C3M CO2 emissions 1860 -1999 (418GtC)

future SRES A2 CO2 emissions 2000-2100 (1770GtC)

two experiments: one with climate feedback

one without climate feedback

Quantifying sensitivity: New results from CarboOcean

ocean ocean

•IPSL •MPI•C4MIP

IPSL-CM2C IPSL-CM4 MPIM NCAR

o (PgC / °C) -30 -16 -22 -17

Quantifying climate impact

Mechanisms for climate impact:

• Increasing Sea Surface Temperature decreases CO2 solubility• Decreased mixing with sub-surface and deep- waters prevents the penetration of anthropogenic carbon into deep ocean• Decrease in biological production reduces the amount of carbon transported to depth.

Identifying mechanisms for climate impact

NCAR

IPSL CM2 IPSL CM4 MPIM NCAR

SST [oC]

Export [PgC/yr]

Identifying mechanisms:

NCAR

IPSL CM2 IPSL CM4 MPIM NCAR

SST [oC]

Identifying mechanisms:

NCAR

IPSL CM2 IPSL CM4 MPIM NCAR

mixed layer depth [m]

IPSL CM2 IPSL CM4 MPIM MOC [Sv]

NCAR

SST [oC]

Identifying mechanisms:

NCAR

IPSL CM2 IPSL CM4 MPIM NCAR

IPSL CM2 IPSL CM4 MPIM

mixed layer depth [m]

MOC [Sv]

Identifying mechanisms:

0

0.5

1

1.5

2

2.5

3

-40 -30 -20 -10 0

Gamma (PgC/°C)

SS

T (°

C) IPSL-CM2C

IPSL-CM4

MPI

NCAR

-2

-1.5

-1

-0.5

0

-40 -30 -20 -10 0

Gamma (PgC/°C)

Exp

ort

(Pg

C/y

r)

IPSL-CM2C

IPSL-CM4

MPI

NCAR

Identifying/Quantifying climate impact:

SST

Export

-10

-8

-6

-4

-2

0

-40 -30 -20 -10 0

Gamma (PgC/°C)

TH

C (S

v) IPSL-CM2C

IPSL-CM4

MPI

-60

-50

-40

-30

-20

-10

0

-40 -30 -20 -10 0

Gamma (PgC/°C)

Mix

ed L

ayer

Dep

th (m

)

IPSL-CM2C

IPSL-CM4

MPI

NCAR

Identifying/Quantifying climate impact:

MOC

MLD

Regionalisation

Regionalisation

Identifying mechanisms:

- 10%

opex90

- 13%

Identifying mechanisms:

caex90

+ 20%

Identifying mechanisms:

Conclusions -identifying mechanisms-

CO2 induced warming and weaker MOC weaken physical pump

Decrease in export production weakens biological pump

Increase in calcite export strengthens alkalinity counter pump

All this points to weakening potential for oceanic uptake

Conclusions -quantifying feedback-

CarboOcean models range from -16 to -30 GtC/K

indication for linear relationship with :

nearest for MLD

less clear for export (if CM4 omitted),

inverse if anything for MOC (?)

scattered for SST

MPIIPSLSabine et al. 2004

Anthropogenic DIC

MPI-OM grid