The GFDL CM3 coupled climate model: Characteristics of the ocean and sea ice simulations

Stephen M. Griffies, Michael Winton, Leo J. Donner, Larry W. Horowitz, Stephanie M. Downes, Riccardo Farneti, Anand Gnanadesikan, William J. Hurlin, Hyun Chul Lee, Zhi Liang, Jaime B. Palter, Bonita L. Samuels, Andrew T. Wittenberg, Bruce L. Wyman, Jianjun Yin, Niki Zadeh

Research output: Contribution to journalArticlepeer-review

237 Scopus citations


This paper documents time mean simulation characteristics from the ocean and sea ice components in a new coupled climate model developed at the NOAA Geophysical Fluid Dynamics Laboratory (GFDL). The GFDL Climate Model version 3 (CM3) is formulated with effectively the same ocean and sea ice components as the earlier CM2.1 yet with extensive developments made to the atmosphere and land model components. Both CM2.1 and CM3 show stable mean climate indices, such as large-scale circulation and sea surface temperatures (SSTs). There are notable improvements in the CM3 climate simulation relative to CM2.1, including a modified SST bias pattern and reduced biases in the Arctic sea ice cover. The authors anticipate SST differences between CM2.1 and CM3 in lower latitudes through analysis of the atmospheric fluxes at the ocean surface in corresponding Atmospheric Model Intercomparison Project (AMIP) simulations. In contrast, SST changes in the high latitudes are dominated by ocean and sea ice effects absent in AMIP simulations The ocean interior simulation in CM3 is generally warmer than in CM2.1, which adversely impacts the interior biases.

Original languageEnglish (US)
Pages (from-to)3520-3544
Number of pages25
JournalJournal of Climate
Issue number13
StatePublished - Jul 2011


  • Atmosphere-ocean interaction
  • Climate models
  • Coupled models
  • Model evaluation/performance
  • Ocean circulation
  • Sea ice
  • Sea surface temperature

ASJC Scopus subject areas

  • Atmospheric Science


Dive into the research topics of 'The GFDL CM3 coupled climate model: Characteristics of the ocean and sea ice simulations'. Together they form a unique fingerprint.

Cite this