Developments at GFDL

Leo Donner

WGNE, Tokyo, 20 October 2010

• CM3: Atmosphere-Ocean-Land-Sea Ice Model with Aerosol-Cloud Interactions

• Decadal Prediction Model

• Earth-System Models

• High-Resolution Atmospheric Models

• Parameterization and Climate Processes

Global Model and Parameterization Development

CM3 Integrations

Lead,Larry

Horowitz

CM3 and ESM IPCC AR5 Integrations Underway

CM3: First GFDL Climate Model with Indirect Aerosol Effects

CM2.1 (no indirect effect) 0.66°C

CM3 (includes indirect effect) 0.32°C

GISS Observations 0.52°C

CRU Observations 0.56°C

Globally Averaged Temperature Change(1980-2000) minus (1881-1920)

(5-member ensembles for models)(from Donner et al., 2010, J. Climate, submitted)

CM3 cloud-aerosol interactions from physically based aerosol activation driven by sub-grid PDFs of vertical velocity in shallow cumulus and

stratiform clouds (Ming et al., 2006, JAS; Golaz et al., 2010, JCL, submit).

from Donner et al. (2010, J. Climate, submitted)

from Donner et al. (2010, J. Climate, submitted)

from Donner et al. (2010, J. Climate, submitted)

Taylor diagrams

from Donner et al. (2010, J. Climate,

submitted)

Taylor Diagrams for

TOA Radiation

Fields

from Donner et al. (2010, J. Climate,

submitted)

Coupled Data Assimilation is enabling major improvement in ENSO prediction.

3Dvar

CDA

Anomaly Correlation Coeff norm RMS errors

0.6

1.0

NINO3 SSTA forecast skills

Initial TimeJan Dec Jan Dec

Initial Time

Lead T

ime

Jan

Dec

Dec

Jan

(from Tony Rosati, GFDL)

High-Resolution Atmospheric Modeling:Atlantic Hurricanes and Anthropogenic Warming

Published by AAAS

M. A. Bender et al., Science 327, 454-458 (2010)

Fig. 1 Simulated and observed histograms of maximum surface wind speed (m/s) in the Atlantic basin

Published by AAAS

M. A. Bender et al., Science 327, 454-458 (2010)

Fig. 3 (Left) Tracks for all storms reaching category 4 or 5 intensity, for the control and the warmed 18-model ensemble conditions, as obtained using the GFDL/NWS

hurricane model

Published by AAAS

M. A. Bender et al., Science 327, 454-458 (2010)

Fig. 4 Model tracks for all storms that eventually reached category 4 or 5 intensity, for the control (top) and the warm climate conditions provided by the four individual

CMIP3 models

A CPT for Cloud Parameterization and Aerosol Indirect Effects

Supported by

3 - Close 2 - Select

1 - Prognose

Building a PDF-based parameterizationAdvance prognostic moment equations

Select PDF from functional form to matchmoments

Use PDF to close higher-order moments, buoyancy terms

Diagnose cloud fraction,liquid water, dropletnumber from PDF

Adapted from Golaz et al. 2002a,b (JAS)

from Guo et

al. (2010, Geosci.

Mod. Dev. Disc.)

AM3 Single Column

Model using Multi-Variate Probability

Density Function with

Dynamics, Aerosol

Activation, and Double-

Moment Microphysics from Guo et al.

(2010, Geosci. Model Dev. Discuss.)

NOAA Climate Modeling and Research System (CMRS)

Elements and Schedule of the Subcontract Award to Cray, Inc

May 2010

20

CMRS.1 260 TFCMRS.1 260 TFFY 2010 FY 2011 FY 2012 FY 2013 FY 2014

CMRS.2 720 TF

CMRS.2 720 TF

CMRS.1 UPG 386 TF

CMRS.1 UPG 386 TF

CMRS.2 720 TF

CMRS.2 720 TF

260 TF

980 TF

1.1 PF

CMRS.1Installation

CMRS.1Installation

CMRS.2Installation

CMRS.2Installation

Negotiated Options for both FY2015

and FY2016

Negotiated Options for both FY2015

and FY2016

CMRS.1 Operational NLT 1 Oct 2010

CMRS.2 Operational NLT 1 Oct 2011

NOAA Climate Modeling and Research System (CMRS) Computing Capability Timeline at Oak Ridge National Lab

Summary

• GFDL, in support of IPCC AR5, has developed a new generation of coupled physical and earth-system models. Integrations for AR5 are mostly completed.

• Multiple nesting approach has provided framework for considering tropical cyclones and anthropogenic warming.

• Coupled ocean-atmosphere data assimilation promising for ENSO prediction

• NSF/NOAA Climate Process Teams active• New off-site computing

AM3/LM3: Atmosphere and Land in CM3AM3/LM3: Atmosphere and Land in CM3

23

Interactive chemistry to link emissions to aerosol composition

Sub-grid vertical velocity PDFs for convective and stratiform clouds =>Supersaturation at cloud scale for aerosol activation on sulfate/black

carbon, organic carbon, and sea salt

Stratospheric model for chemistry and possible links to troposphere on multi-year time scales (e.g., Southern Hemisphere Annular Mode)

More realistic land precipitation for land carbon and nitrogen models

Cubed-sphere finite-volume dy-core

Dynamic vegetation model with hydrology and land use

What is NEW in AM3?What is NEW in AM3?

Parameterizations based on sub-grid PDFs of verticalvelocity have been implemented in AM3

Parameterizations based on sub-grid PDFs of verticalvelocity have been implemented in AM3

• Deep convection: Donner et al. (2001, J. Climate), Wilcox and Donner (2007, J. Climate)

• Shallow convection: Bretherton et al. (2004, Mon. Wea. Rev.) implemented by Ming Zhao

• Stratiform: modfication of Tiedtke (1993, Mon. Wea. Rev.) by Chris Golaz, to include w PDF for activation only