A system called Earth

The concepts of environmental sustainability that broke into public awareness in the 1970s took on new overtones two decades later. In its first report, released in 1990, the Intergovernmental Panel on Climate Change (IPCC) made it clear that ever-increasing amounts of human-produced greenhouse gases posed a real risk to a sustainable future.

NCAR was well positioned to take a leading role in efforts to understand this growing threat. As far back as 1972, NCAR’s William Kellogg was among three editors of a workshop-based report entitled Man’s Impact on the Climate. The center’s Community Climate Model (CCM), released in the 1980s, was used by hundreds of scientists at NCAR, universities, and national laboratories, helping to clarify the ways in which greenhouse gases influence world climate. NCAR researchers served tirelessly on the IPCC as well as on National Academies studies and countless committees devoted to unraveling the threat.

To truly understand the future of the atmosphere, though, NCAR realized that something more was needed. While the CCM included oceans as well as atmosphere, the latter component couldn’t influence the former. Modelers had to specify the state of the “slab ocean” rather than allowing it to emerge organically from interactions between sea and atmosphere. Likewise, the model’s depiction of forests, grasslands, and other features lacked interactivity. This meant the model’s vegetation couldn’t spread or decay as it would in real life in response to a changing atmosphere.

In the mid-1990s, NCAR took a risk: it reallocated much of its climate research budget toward creating a wholly new kind of software package. The result was the Climate System Model, produced with support from NSF and the U.S. Department of Energy and released in 1996. Like the CCM, the CSM was free to all interested scientists, and NCAR provided ample online support for users. But the CSM broke free of its predecessor with an innovative architecture that allowed four key parts of the Earth system—the atmosphere, ocean, land, and sea ice—to interact with each other. The switchboard for this virtual conversation was a flux coupler, which monitored and controlled interactions between the four CSM modules.

Photo of two men standing in front of the Flatirons, Boulder, CO
Together with the late Byron Boville (not pictured), Rolando Garcia and Raymond Roble led the development of the Thermosphere Ionosphere Electrodynamics General Circulation Model (TIE-GCM), a model that united upper and lower atmosphere.

One of the immediate benefits of the CSM was the near-elimination of climate “drift,” the tendency for a modeled climate to inch out of balance over time. Since the CSM’s elements could interact much more freely, they served as checks and balances on each other, thus reducing the odds that a simulation would spin into unrealistic territory, even in runs spanning hundreds of years. “The small degree of climate drift in the 300-year CSM integration is a major accomplishment,” wrote NCAR’s John Weatherly and colleagues in a 1998 paper.

The new model proved to be an enormous benefit to university researchers. More than 1,200 scientists at 300 institutions used the CSM in its first few years, many of them collaborating with NCAR on improvements to the various components.

In 2001 those relationships were embedded in the software’s name, as it became the Community Climate System Model. Soon after, with the help of major computing allocations at NCAR, DOE, and elsewhere, the CCSM furnished close to half of all the simulation hours completed in support of the fourth IPCC assessment, released in 2007.

Even as NCAR marshaled its resources in the 1990s to produce the CSM, the center remained a home for other lines of model development. Gerald Meehl and Warren Washington steered the progress of the Parallel Climate Model, a DOE-supported effort to incorporate ocean and atmosphere in software that was tailored for a new generation of computers with hundreds or even thousands of processors operating simultaneously.

Still other modelers extended their perspectives to the very edge of Earth’s atmosphere. Raymond Roble and Cicely Ridley had collaborated with Robert Dickinson in the 1980s to build a general circulation model that linked the thermosphere and ionosphere. The model, which later grew to include coupled electrodynamics and the mesosphere, was eventually paired with NCAR’s global climate model. By the late 1990s, the merger had morphed into the Whole Atmosphere Community Climate Model. With WACCM, for the first time, scientists could follow the evolution of Earth’s climate at all levels of the atmosphere.


Today — A new generation of global models

Photo of James Hurrell

"Decadal climate prediction is a major new focus."

—James Hurrell, NCAR

For scientists hungry to learn about our future climate, each advance brings a new frontier. In the early 2010s, the world’s 20 or so centers of comprehensive climate modeling—including NCAR—have begun studying the kinds of changes that people might expect in their own lifetimes as well as those of their children and grandchildren.

“There is potential forecasting skill that is going untapped,” says James Hurrell, who heads the scientific steering committee for the Community Climate System Model. Now that CCSM and other models can incorporate the evolving state of the ocean—which holds some 90% of the heat already trapped by human-produced greenhouse gases—it’s hoped that the models can capture the movement and release of oceanic heat, and thus predict shifts in climate that can play out over as little as a single decade.

The first such “climate forecast,” issued by scientists from the United Kingdom’s Hadley Centre in 2007, called for global temperatures to hold flat or dip slightly for several years, as they in fact did, then rise to new highs by the mid-2010s. Other modeling groups are now taking a stab at 10- to 30-year forecasts within the Coupled Model Intercomparison Project. The CMIP template provides a common framework for modeling in support of the IPCC. Through CMIP, the products of various models can be compared, apples-to-apples style. Papers on the new work will begin appearing in 2011, with the next major IPCC assessment reports to appear in 2013 and 2014.

Visualization of Earth with white hexagonal grids overlay
Model grids that cover the globe with a hexagonal mesh may yield greater accuracy and efficiency. (Visualization by William Skamarock, NCAR.)

Along with the decadal emphasis, many of the new simulations will not only track the effects of human-produced greenhouse gases but also simulate the uptake and release of atmospheric carbon by plants and oceans over time. Interactive carbon is just one of an array of new features in the latest incarnation of the CCSM, the Community Earth System Model, released in mid-2010. The CESM includes the ability to model direct effects of aerosols (airborne particles) on radiation as well as their indirect effects in helping to form or suppress clouds. “The emerging complexity is fascinating,” says Andrew Gettelman, a leading developer of CESM’s cloud and aerosol components.

A few hints of the next great leap in global modeling, still years away, are already taking shape. Scientists at NCAR, Los Alamos National Laboratory, and the University of Exeter are exploring the use of innovative grids to replace the latitude-longitude boxes traditionally used in global models. A hexagonal structure (see graphic) shows particular promise, as it eliminates problems near the poles and easily accommodates pockets of higher resolution at minimal computing cost.