Climate Modeling at GFDL: The Scientific Challenges V. Ramaswamy NOAA/ Geophysical Fluid Dynamics Laboratory November 12, 2008.

Climate Modeling at GFDL:The Scientific Challenges

V. Ramaswamy

NOAA/ Geophysical Fluid Dynamics LaboratoryNovember 12, 2008

Be a world leader for the production of timely and reliable knowledge and assessments on natural climate variability and anthropogenic changes and in the development of the required earth system models.

Work cooperatively in NOAA to advance its expert assessments of changes in national and global climate through research, improved models, and products.

 GFDL MissionDirectly supports the DOCAnd NOAA Strategic Goals

One of 2 Climate Modeling Centers called for in the US Climate Change Science Program [CCSP]

Atmospheric, Ocean, Land, Coupled, Earth System Model Developments

Matrix Managed

AM2,-----------

LM2, SIS,OM3

(MOM4)

AM3, LM3, SIS,OM3 (MOM4, GOLD)

2001-2004 2005 2006 2007

CM2.0 CM2.1

2008

AR4, WMO/UNEP

GFDL’s Recent Major Climate Model Developments

CCSPs, NARCCAP

B-grid FV core

FV, CS cores

Hi-resAM2

CM 2.4ESM 2.1w/ OM3{M,G}

CM3

Figure SPM.2

Figure SPM.4

Figure SPM.5

Figure SPM.6

Figure SPM.7

Anthro. RF > 0 (v. high conf.)

20th Cent. continental warminglikely due to human activity

Projected global warming

Projected patternof rainfall

changes in 21st Cent. Projected warming patternin early and late 21st Cent.

NOAA/ GFDLmodel simulations

contributed toIPCC AR4

AR4 conclusions

Global decreases in sulfate aerosolwarmer U.S. summers in 2100

CCSP 1.1

CCSP 3.2

(sfc - tropos): Models vs. Obs.

CCSP 2.4

CCSP 3.3

NOAA/ GFDL

contributionto

CCSPreports

OUTLINE

• Understanding present climate; quantifying the causal factors and attribution of past climate change; and projections of future climate changes.

• Challenges and progress in modeling the Atmosphere, Ocean, Coupled Atmosphere-Ocean, and Biosphere to address the key issues.

Schwarzkopf and Ramaswamy (2008)

The World Has Warmed

Globally averaged, the planet is about 0.75°C warmer than it was in 1860, based upon dozens of high-quality long records using thermometers worldwide, including land and ocean.

Eleven of the last 12 years are among 12 warmest since 1850 in the global average.

Globally averaged, the planet is ~0.75°C warmer than it was in 1860, based upon dozens of high-quality long records, including land and ocean.

Eleven of the last 12 years are among 12 warmest since 1850 in the global average.

IPCC AR4

Nat = Natural ForcingAnth = Anthropogenic ForcingAllForc = (Nat + Anth) ForcingsCRU = Observations

GFDL Climate Model CM2.1

1950 2000

IPCC AR4 simulation

Human and Natural Drivers of Climate Change

IPCC (2007)

GFDL CM 2.1

Anthropogenic forcings and response[1950-2000]

1860

2000

1860

2000

Uncertainties

Aerosol microphysics

Aerosol-Cloud interaction

Sulfate AODoverestimated

(Europe)

AOD overestimated

(East coast)

Biomassemissions

underestimated(S Africa)

Biomassemissions

underestimated(S America)

AOD GFDL CM2.1 (1996-2000)

Aerosol Optical Depths from GFDL Coupled Model 2.1 (CM2.1), AVHRR, and MODIS

The Future

What will be the impacts of changes in Greenhouse Gases and Aerosols?

How might future changes in aerosols affect climate?

HISTORICAL and FUTURE SCENARIOS

CO2 concentrations

pp

mv

Emissions of Short-lived Gases and Aerosols (A1B)

1880 1920 1960 2000 2040 2080

NOx (Tg N yr-1)

SO2 (Tg SO2 yr-1)

BC (Tg C yr-1)25201510 5 0

250

200

150

100

50

605040302010

Horowitz, JGR, 2006

Large uncertainty in future emissiontrajectories for short-lived species

Pollution controls

A1B

IPCC, 2001

Up to 40% of U.S. warming in summer (2090s-2000s) from short-lived species

From changing well-mixed greenhouse gases +short-lived species

From changing only short-lived species

Warming from increases in BC + decreases in sulfate;depends critically on highly uncertain future emission trajectories

Results from GFDL Climate Model [Levy et al., 2008]

Change in Summer Temperature 2090s-2000s (°C)

New Science Questions for Next-Generation Model

• What are the roles of aerosol-cloud interactions in climate and climate change?

• How will land and ocean carbon cycles interact with climate change?

• To what extent is decadal prediction possible?

• What are the dominant chemistry-climate feedbacks?

Atmospheric Model Developments to Address the New Questions

• Interactive chemistry to link emissions to aerosol composition

• Aerosol activation requires super-saturation at cloud scale => Sub-grid PDFs of vertical velocity for convective and stratiform clouds

• Sufficiently realistic tropical land precipitation for land carbon model

• Stratospheric model for chemistry and links to troposphere, including those on multi-year scales relevant to decadal prediction

Model – satellite difference spectrum

Unit: W m-2OLR Window band

Total sky Clear sky Total sky Clear sky

CERES 241.73 275.87 66.94 83.28

GCM 240.63 263.43 73.99 87.56

GCM-CERES

-1.10 -12.44 7.05 4.28

OverestimationUnderestimation

H2O vib-rotWindow

[Huang et al. 2007 GRL]

Total-sky MODEL-AIRS radiance difference

Water vapor band radiance error budget

i

ii

i

iai

a

ss

OHOH

R

TT

R

TT

RR

22

Clean/Maritime

Polluted/Continental

Aerosol Indirect Effects (1st and 2nd)

Ramanathan et al. (2001)

Aerosol vs. Dynamics

T = 288 Kp = 850 hPaAerosol mass = { 0.5, 0.5, 0.5 } x 10-12 kg

CCN activation is a non-linear function of vertical velocity

from Ming et al. (2006,

JAS)

updraft: activation

downdraft: evaporation

~ 12.9 km

Large Eddy Simulation shows small-scale activation.

simulation by Chris Golaz

Large-scale CCN activation

Layer-averaged activation:

Because N* is non-linear

However,

To use satellites to evaluate GCMs, the GCM must be sampled like the satellites

OAR/CDC

Cloud drop radius (µm)