ocean circulation and sea ice in the mpi-m ipcc experiments · 2005. 12. 15. · atmospheric and...

Ocean circulation and sea ice in the MPI-M IPCC experiments

Johann JungclausMax- Planck- Institut für Meteorologie

Hamburg

+ M. Esch, H. Haak, F. Landerer, M. Boettinger, E. Roeckner ++

Introduction: The role of the ocean in the climate system

The oceans store and transporthuge amounts of heat

The ocean provide the slowcomponent of variability and its inertia may mask global climate change

The ocean provides also a storage and transportmeans for carbon. (25% of anthropogenic carbonemission are taken up bythe oceans)

European temperature March 2005

Atmospheric and oceanic heat transportstotal

atmos

ocean

90S 60S 30S EQ 30N 60N 90N

6

-4

4

0

-6

-2

2

Atmosphere and oceantransport about 6 PW (1 PW=1015 W) heat from theequator to high and middlelatitudes

The global oceanaccounts for 1/3 (ca. 2 PW) of this. The Atlantic heat transport is about 1.3 PW

The Thermohaline Circulation (THC)

Evidence from the past

Abrupt climate changes in the past are associated withdisturbances of the Atlanic Thermohaline Circulation caused byexcessive fresh water input into the North Atlantic

Effects of a possible sea change

Eiszeit oder Treibhaus?

PM September 2004 Titel

Sea ice in the climate systemSea ice effectively decouples ocean and atmosphere. Owing to its high albedo itreflects sun light back to space. Loss of seaice will enhance oceanic heat uptake (icealbedo feedback).

Therefore the sea ice covered regions arevey sensitive to global change

Sea ice stores huge amounts of fresh water. Its release under warming condition mayinfluence the THC.

The presence of sea ice influencesatmospheric circulation.

Sea ice changes will influence the marineand terrestrial ecosystems, and high latitude societies and economics.

The coupled ocean atmosphere ModelECHam5: MPI atmosphere model (Roeckner et

al., 2003), interactive runoff and glacier calving scheme.

Resolutions: T63L31 (IPCC)

OASIS 3.0 PRISM coupler

MPI-OM (C-HOPE) (Marsland et al., 2003)C-Grid, z-level, partial cells, BBL

parameterizationHibler-type sea ice model incl. snow and

fractional ice coverConformal mapping: 1.5° with refinement in

grid pole regions

NO FLUX ADJUSTMENT

ECHam5

OASIS

MPI-OM

Ocean model grid

Every 5th grid line is shown here; in the vertical, there are40 unevenly spaced depth levels

Ocean model grid

How well does the model reproduce the presentclimate ?

SST is in general agreement with observations but biases > 1º C are common owing to unresolved processes/features.

These errors are not a unique MPI-OM/ECHAM feature.Jungclaus et al., 2005

Precipitation (modeled – observed)

[mm/day]Precipitation bias is most pronounced in equatorial region(Double ITCZ)

Jungclaus et al., 2005

Northern hemisphere sea ice conc.Chapman & Walsh 1996 Model (T63/GR1.5 ctrl.)

March

September

Jungclaus et al., 2005

Southern hemisphere sea ice conc.Model (T63/GR1.5)Chapman & Walsh 1996

March

September

The Atlantic Meridional Overturning Circulation

Jungclaus et al., 2005

Meridional heat transport in the Atlantic

ECHAMT63/L31, MPI-OM 1.5

Jungclaus et al., 2005

Results from the IPCCscenario experiments

Atm. CO2 concentration Glob. Atm. Surface air temperature

Figure by M. Boettinger, DKRZ

IPCC Experiments A1B: SST difference end of 21th century minus 20th century

Widespread warming, North Atlantic shows effect of weaker THC

A1B: Sea surface salinity difference end of 21th century minus 20th century

What is the fate of the MOC?In a warmer climate, the heating of the upper ocean is

accompanied by freshening in high latitudes and salinification in low latitudes (acceleration of thehydrological cycle). Surface waters in the deep waterformation regions become lighter and the watercolumn is more stably stratified.

This will affect the sinking of deep water and the MOC.

Feedbacks such as salt advection by the ocean maycounteract or enhance the decrease of the MOC.

The relative role of individual feedbacks appears to behighly model dependent.

Estimates from different models (Third Assessment Report (TAR))

Figure from IPCC, 2001

Projections for the Atlantic MOC for the 21st C

Global Atm. Surfacetemperature[ºC]

Atlantic Meridional overturningcirculation

[Sv, 1Sv=106m3s-1]

and beyond...

Global Atm. Surfacetemperature[ºC]

Atlantic Meridional overturningcirculation

[Sv, 1Sv=106m3s-1]

Temporal evolution of Atlantic upper oceansalinity

Difference betweenzonally averagedsalinity in the A1B experiment and the 20th

century mean

Positive salinityanomalies in the North Atlantic evolve as a resultof northward salttransport.

Temporal evolution of zonally averagedsurface air temperature

A1B

Future evolution of Arctic sea ice

Figure by M. Boettinger, DKRZ

Animation by M. Boettinger, DKRZ

Animation by M. Boettinger, DKRZ

Impact of Arctic climate change: New seatransportation routes?

Figure by BBC

Open water in the North East Passage (in % of the total area of the coastal corridor)

Time (yr)

% o

f ope

nw

ater

September August October March

Great future for Arctic shipping but at whichcosts for the environment?

Ship emissions as observed from space;

will we see a similar track around Siberia?

What are climate model results good for?ACIA, the Arctic Climate Impact Assessment

An international project of the Arctic Council and the International Arctic Science Committee (IASC), to evaluate and synthesize knowledge on climate variability, climate change, and increased ultraviolet radiation and their consequences.

The model results were mainly basedon the IPCC TAR (2001) experiments.

ACIA International Scientific Symposium held in Reykjavik, Iceland in November 2004.

http://www.acia.uaf.edu

ConclusionsThe new coupled model simulates a reasonable climate withoutflux adjustment.Improvements in comparison with the IPCC TAR models areseen in ocean heat transport, sea ice, and other featuresIn the IPCC scenarios, the Atlantic MOC decreases over the 21st

century but slowly recovers thereafter; this is also the case ifthere is an additional fresh water input from a melting Greenlandice sheet. (This may be model dependent)The model show a considerable and rapid loss of Arctic sea icein the 21st century. While winter sea ice cover is only slightlyreduced the Arctic ocean is projected to become ice free in summer in the 2nd half of the 21st century

Data issues

Data are available on original grid (256x220x40 grid points) and (partly) on IPCC standard 1º grid (360x180x40).

Data format is „EXTRA“ (a fortran binary) or NetCDF

10 yr monthly mean of a 3d- data field requires 1.1 GB on original (EXTRA) and 1.25 GB (NetCDF) on regular grid.

10 yr monthly mean of a 2d- data field requires 27 MB on original and 32 MB on regular grid.

Owing to technical problems not all data have beentransferred fo PCMDI.

Data will also be stored in M&D CERA data base

Thank You….

No total breakdown of the MOC....What if we include the meltdown of Greenland ?

The present simulations do not include ice sheet dynamics. Icesheet model projections show a considerable loss of Greenlandice mass (with a high range of uncertainty); estimates rangefrom 0.01 to 0.1 Sv for various scenarios.

From the IPCC experiments one can deduce a melting rate of about 0.03 Sv for the A1B scenario at the end of the 21st

century.

A sensitivity experiment under A1B forcing is carried out where an additional fresh water inflow is prescribed around Greenland. The flux is ramped up from 0 Sv in 2000 to 0.09 Sv in 2100.

Effect of additional melt water input fromGreenland on the MOC