GFDL - Geophysical Fluid Dynamics Laboratory

Topic: New models – new ‘parameter space’

Info associated with a Journal of Climate paper (Delworth et al., 2011, in press) and a poster presented at the October 2011 WCRP Open Science Conference (Dixon et al, 2011).

[thumbnail of GFDL CM2.6 Arctic SSS]

Sustained model development efforts and the availability of enhanced computer resources have allowed researchers at NOAA?s Geophysical Dynamics Laboratory (GFDL) to construct a pair of new, higher resolution, global models of the coupled physical climate system. Known as CM2.5 and CM2.6, these models are being applied to problems spanning seasonal-to-interannual up to decadal-to-century time scales. A goal of this development path is to explore a new ?parameter space? of global climate models at GFDL ? one that includes very energetic ocean flows. In the ocean component, higher spatial resolution and model configuration choices together allow sharper gradients to be maintained than in prior models. We plan to use this suite of models to study topics including the role of ocean eddies in climate and climate change.

As indicated in the chart below, the new CM2.5 and CM2.6 models represent one branch of model development efforts that have been pursued at GFDL following the successful CM2.0 and CM2.1 models (circa 2005). We continue to evaluate the extent to which this development stream produced models that have simulation characteristics that differ from GFDL’s previous workhorse climate models, CM2.0 and CM2.1. The improved representation of some smaller-scale processes in the climate system appear to improve substantially the simulation of many key aspects of climate – something that will lead us to perform numerous studies to better understand the mechanisms at play in the models and the real world processes they simulate. The present works builds on, and is complementary to, efforts at other institutions (such as NCAR in the US,¬† the Hadley Centre in the UK, and JAMSTEC in Japan) to build high-resolution coupled climate models.

One should note that the model simulations described here are from free-running climate model experiments. In other words, after initialization, no observational data from the atmosphere, ocean, or land are ingested by the model to force the models prognostic variables toward observations (i.e., there is no data assimilation nor nudging the model towards observations). The models generate their own internal variability and mean flows, and their own ‘climate drift’.¬† Carbon dioxide and other anthropogenic greenhouse gas levels are prescribed in these models – they are not prognostics variables.

Going forward, we expect the CM2.5 model will be a workhorse model at GFDL for one class of global climate model experiments. The higher resolution, and thus more computationally expensive CM2.6 model will be used for a more limited set of experiments in the near-term.

[Chart of GFDL model development streams in 2011]

  • Delworth, T.L., et al., 2011: Simulated climate and climate change in the GFDL CM2.5 high resolution coupled climate model. Journal of Climate, (in press).