Evaluating NOAA’s GFDL CM2 Global Coupled Climate Models: Formulations, Drift, and Climate Simulations
Observation and Context Computer models must simulate Earth’s physical climate system realistically across timescales ranging from daily weather to multi-century climate shifts. Previous coupled models at the Geophysical Fluid Dynamics Laboratory (GFDL) often relied on artificial “flux adjustments” (rigid mathematical corrections to heat and water exchanges) to keep the simulated climate stable and prevent it from drifting unrealistically far from real-world observations.
Hypothesis Scientists hypothesized that coupling high-resolution atmosphere, land, ocean, and sea ice modules without artificial flux adjustments would produce realistic, stable multi-century climate simulations. Furthermore, they expected that updating the atmospheric dynamical core to a finite-volume scheme alongside targeted adjustments to clouds, land hydrology, and ocean viscosity would significantly reduce global temperature and wind pattern biases.
Experiment and Methodology The researchers constructed and tested two coupled climate models: CM2.0 and CM2.1.
- Model Design: Both models use an atmospheric and land grid resolution of 2∘ latitude×2.5∘ longitude (with 24 vertical atmospheric levels) and an ocean resolution of 1∘×1∘ that sharpens to 1/3∘ near the equator (with 50 vertical ocean levels). A tripolar ocean grid avoids polar calculation errors.
- Key Differences: CM2.0 uses a standard B-grid atmospheric core. CM2.1 implements a finite-volume (FV) dynamical core, reduced cloud-drop threshold sizes (increasing shortwave solar radiation reaching Earth’s surface), a land module that blocks evaporation over frozen soil, and lower horizontal ocean viscosity in the extratropics.
- Testing: The team executed 300-year control runs under fixed 1990 greenhouse gas conditions, as well as multi-century runs (500 to 1,000+ years) under pre-industrial 1860 conditions.
Results and Data
- Sea Surface Temperature (SST): CM2.1 outperformed CM2.0, reducing the root-mean-square error of global SST from 1.54 K in CM2.0 down to 1.16 K in CM2.1.
- Wind Stress & Gyres: CM2.0 suffered from an equatorward drift of midlatitude westerly winds, producing an excessive cold bias. CM2.1 pushed westerly winds and the zero-wind-stress-curl line poleward, strengthening oceanic subtropical gyres and warming the midlatitudes.
- Ocean Circulation: The lower ocean viscosity and improved winds in CM2.1 strengthened the North Atlantic thermohaline circulation (THC) and the Antarctic Circumpolar Current (ACC). This enhanced subpolar gyre heat transport into the Labrador and Greenland seas, significantly diminishing excessive North Atlantic sea ice.
- Remaining Biases: Both models displayed persistent biases, including a “double ITCZ” rain belt in the tropical Pacific, a lack of precipitation over the Amazon Basin, and overly thin Arctic sea ice.
Conclusion and Climate Impact The team successfully built stable, multi-century global climate models operating without artificial flux adjustments. Incorporating a finite-volume dynamical core, tuned cloud radiation, and refined ocean viscosity (CM2.1) significantly minimized temperature, salinity, and wind biases.
- Climate Impact: When tested with a doubling of atmospheric CO2 coupled to a slab ocean, the equilibrium climate sensitivity was determined to be 2.9 K for CM2.0 and 3.4 K for CM2.1. Both models successfully reproduce twentieth-century warming trends, providing vital data and projections for the 2007 Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report.
Full Citation Delworth, T. L., Broccoli, A. J., Rosati, A., Stouffer, R. J., Balaji, V., Beesley, J. A., Cooke, W. F., Dixon, K. W., Dunne, J., Dunne, K. A., Durachta, J. W., Findell, K. L., Ginoux, P., Gnanadesikan, A., Gordon, C. T., Griffies, S. M., Gudgel, R., Harrison, M. J., Held, I. M., Hemler, R. S., Horowitz, L. W., Klein, S. A., Knutson, T. R., Kushner, P. J., Langenhorst, A. R., Lee, H.-C., Lin, S.-J., Lu, J., Malyshev, S. L., Milly, P. C. D., Ramaswamy, V., Russell, J., Schwarzkopf, M. D., Shevliakova, E., Sirutis, J. J., Spelman, M. J., Stern, W. F., Winton, M., Wittenberg, A. T., Wyman, B., Zeng, F., & Zhang, R. (2006). GFDL’s CM2 global coupled climate models. Part I: Formulation and simulation characteristics. Journal of Climate, 19(5), 643–674. https://doi.org/10.1175/JCLI3629.1