Evaluating NCAR’s Community Climate System Model Version 3 (CCSM3): Physical Formulations, Multi-Century Stability, and Climate Sensitivity
Observation and Context Global climate models simulate interactions among Earth’s atmosphere, ocean, land, and sea ice to project future climate change. While previous iterations like CCSM2 operated stably without artificial flux adjustments, they suffered from noticeable biases: an overextended cold equatorial tongue, severe cold biases at the tropical tropopause and polar atmosphere, excessive boreal winter land temperatures, and underpredicted sea ice thickness.
Hypothesis Researchers hypothesized that updating the core physical parameterizations—specifically cloud microphysics, aerosol radiative forcing, land canopy interactions, ocean mixed-layer depth, and sea ice transport dynamics—would reduce systematic climate biases and produce stable, drift-free multi-century simulations across multiple spatial resolutions.
Experiment and Methodology The National Center for Atmospheric Research (NCAR) and collaborating institutions developed CCSM3, linking four updated component models via a centralized flux coupler:
- Component Models: The atmosphere (CAM3) and land (CLM3) components ran on an Eulerian spectral T85 grid (about 1.41∘ resolution) with 26 vertical atmospheric levels and 10 subsurface soil layers. The ocean (POP 1.4.3) and sea ice (CSIM5) components ran on a displaced-pole grid at nominal 1∘ resolution with 40 vertical ocean depth levels.
- Physical Updates: Major enhancements included prognostic liquid and ice cloud condensate, direct radiative forcing from multi-species aerosols, water vapor near-infrared absorption, dynamic canopy density resistance over land, chlorophyll-dependent upper-ocean shortwave penetration, and incremental remapping for sea ice dynamics.
- Testing: The high-resolution configuration (T85×1) was evaluated in multi-century present-day control simulations (analyzing years 400–500), stable 1870 pre-industrial runs, and transient 1%-per-year CO2 doubling experiments.
Results and Data
- Atmospheric Temperatures: Thicker cirrus clouds reduced tropical tropopause cold bias by 60% (warming by 2.3 K compared to CCSM2), while polar upper-tropospheric temperatures warmed by 2.3 to 3.9 K, cutting polar cold biases by 33%.
- Radiation Budgets: Increased shortwave absorption by aerosols and water vapor reduced net surface shortwave insolation by 9 W m−2, aligning well with empirical estimates. Top-of-atmosphere shortwave cloud forcing matched ERBE satellite observations to within 0.1 W m−2.
- Ocean and Sea Ice: Central equatorial Pacific cold biases decreased from over 2 K in CCSM2 to under 1 K in CCSM3. North Atlantic ocean heat transport increased by ∼0.3 PW to match observed transects, driven by an overturning circulation of 22 Sv. Central Arctic sea ice thickness reached realistic values of 2.0 to 2.5 m (up from 1.5 m in CCSM2).
- Remaining Deficiencies: CCSM3 retained an overly strong double Intertropical Convergence Zone (ITCZ), a 2-year ENSO periodicity that was narrower than observed, persistent warm coastal sea surface temperature biases up to 7∘C off South America and Africa, and winter cold-region continental warm biases.
Conclusion and Climate Impact CCSM3 produced stable, multi-century simulations without artificial flux adjustments, demonstrating significant advancements in cloud radiative effects, ocean heat transport, and sea ice distribution.
- Climate Impact: When tested under a doubling of atmospheric CO2 (355 to 710 ppmv), the equilibrium climate sensitivity of CCSM3 reached 2.7 K, higher than CCSM2 (2.2 K) due to stronger positive low-cloud feedbacks. CCSM3 provided core simulations for international climate change assessments, including the IPCC Fourth Assessment Report.
Full Citation Collins, W. D., Bitz, C. M., Blackmon, M. L., Bonan, G. B., Bretherton, C. S., Carton, J. A., Chang, P., Doney, S. C., Hack, J. J., Henderson, T. B., Kiehl, J. T., Large, W. G., McKenna, D. S., Santer, B. D., & Smith, R. D. (2006). The Community Climate System Model Version 3 (CCSM3). Journal of Climate, 19(11), 2122–2143.