As it turns out, the union of NCAR's community climate model (CCM) and version 4 of the Pennsylvania State University/NCAR mesoscale model (MM4) has been--well, a model marriage. The resulting NCAR regional climate model (RegCM) has proven itself a flexible tool. It's been used for studies of past, present, and future climates over a variety of regions, including the continental United States, Europe, Africa, eastern Asia, Australia, and the Aral Sea. NCAR scientist Linda Mearns has assembled a new group within ESIG to use the RegCM for agricultural and other impacts studies (see below).
"Regional climate modeling is truly something in which NCAR has been a pioneer," says Filippo. "When we started, nobody had done it before. Now most major laboratories, worldwide, and many smaller institutes, including some in developing countries, are doing it." The approach is one of the centerpieces of programs such as START (the Global Change System for Analysis, Research, and Training) that are aimed at fostering climate system research in developing countries.
"There were many questions about running these kinds of models in coupled mode for an extended period of time," says Filippo. "People wondered whether the physics were compatible and whether noise generated at the models' interface [between global and regional scales] would swamp the model solution." Yet the requests for long-term regional information weren't going away.
Back then, Filippo notes, "global climate models [GCMs] were still run at very coarse resolutions." Data points were typically separated by several hundred kilometers. "One reason why Yucca Mountain is so dry is that it is located in the lee [the side sheltered from the wind] of the Sierras. But GCMs didn't even have the Sierras." Instead, Nevada was depicted on the windward side of the model's coarse, generic Rockies, making it act as a relatively wet repository for Pacific moisture rather than the desert it is.
To better simulate Yucca Mountain's topography and climate, Rick Anthes (now UCAR president) and Bob Dickinson (now at the University of Arizona) suggested nesting the MM4 within the CCM. Filippo was hired to join a small group in CGD, including Gary Bates and Ron Errico and headed by Dickinson, that embarked on a six-month pilot study. The group changed some of the radiation and surface-process physics of the MM4 and revised the assimilation scheme at the boundary between regional and global coverage, making for a smoother blend of large-scale and mesoscale information. Although it produced no definitive results, the pilot study was encouraging enough to ensure RegCM's (and Filippo's) future at NCAR. (Some additional work was carried out through fiscal year 1997 by CGD's Starley Thompson and his Global Environmental and Ecological Simulation of Interactive Systems (GENESIS) model. "The overall result was that climate changes resulting from increased CO2 did not appear to pose a hydrological threat to the repository," says Starley. Meanwhile, the project has been on hold for other reasons.)
Since then, CGD has continued to improve RegCM while making it available to the outside community. The model has been used by researchers at over 25 institutions worldwide for a variety of applications in domains as small as Switzerland and as large as eastern Asia.
Some computer time for RegCM was made possible by the Model Evaluation Consortium on Climate Assessment (MECCA). This international collaborative effort, managed by the Electric Power Research Institute, installed a Cray Y-MP at NCAR in the early 1990s expressly for long-term global and regional climate simulations. "The computer time from MECCA provided critical resources for testing and developing RegCM," says Filippo. (A new summary report on MECCA, Assessing Climate Change, has just been released by Gordon and Breach Science Publishers)
Linda, an NCAR scientist since 1988, is one of the most frequent users of RegCM and a regular collaborator with Filippo on regional modeling studies. Her six years of agriculture-related research dealing with the Great Plains bore fruit in 1996 with a bushel of grants for similar work focusing on other locations. Last fall, she and programmer Larry McDaniel moved from CGD to ESIG. The two have since been joined by other new hires to carry out the multiyear grants (see photo).
|The new ESIG group dedicated to regional climate modeling bears flags that represent their regions of birth or heritage. Top row, left to right: Theo Mavromatis (Greece), Linda Mearns (Scotland), and Larry McDaniel (South Dakota). Bottom row, left to right: Elena Tsvetsinskaya (Russia), Wei Gao (China), and Greg Carbone (United States). Other group members: Chris Castro and Carlo Pona.|
The recently ubiquitous El Niño figures heavily in this project. One of El Niño's best-documented effects is to step up winter precipitation in the Gulf states. Linda and colleagues, including Rick Katz of ESIG, want to see how changes in the frequency, persistence, and intensity of El Niño/Southern Oscillation (ENSO) events affect agricultural yields in the Southeast. "We're getting advice from [CGD scientists] Kevin Trenberth and Jim Hurrell and looking at model runs, but the models aren't really telling us much about how ENSO might change in a greenhouse-warmed world," says Linda. Heeding that uncertainty, the group will use past ENSO records along with stochastic models to study a variety of possible ways in which El Niño and La Niña could keep us guessing.
Variability played a key role in Linda's earlier work, too. Focusing on the Great Plains, she used an agricultural model (CERES-Wheat) that took its cues from RegCM. She found that changes in variability--for instance, a trend toward more muted or more mercurial temperature swings--can be just as important as changes in the climatic mean (the average temperature itself). This is especially true in places where the climate is marginal for a given crop, such as wheat on the semiarid plains of western Kansas.
Linda's Great Plains studies confirmed an assumption behind both her work and Filippo's: "The bottom line is, yes, from both a climate and an impacts point of view, it makes a difference whether you use a coarse-resolution climate-change scenario [a GCM alone] or a higher-resolution scenario derived from the coarse one [the RegCM approach]. The latter is more detailed, and that can be quite important. It's not that it just provides a little more refinement--it can provide an entirely different answer."
Linda, Filippo, and colleagues will evaluate changes in land use while including industrial emissions in a new version of RegCM that tracks sulfates and other relevant chemicals. Filippo and Linda are returning from a week in Beijing as the project gets under way. According to Filippo, "Our hope is not to get an answer on the exact future climate but to get a scenario of how rapid urbanization, industrialization, and agricultural practices might interact with the regional climate in a global change context. For example, northern China has been warming at a much faster pace than southern China, and they're not sure why."
RegCM data already indicate that important regional consequences might follow global warming. Since the late 1980s, much of Switzerland has experienced warm winters, with the mildness most pronounced at higher altitudes, accompanied by reduced snowpack. Martin Beniston (Swiss Federal Institute of Technology) has suggested that high-altitude regions should experience more warming than lower elevations in a global-warming scenario for a variety of physical reasons (less snowpack and associated reflectivity, closer coupling with the free troposphere, and the topographic influence on rainfall and snowfall amounts). When the RegCM was run in present-day and doubled-carbon dioxide modes for a region encompassing Switzerland, it supported this notion. A paper by Filippo, Jim Hurrell, Maria Rosaria Marinucci (now in ESIG), and Beniston appears in the February 1997 edition of the Journal of Climate.
RegCM fans will have lots to read next fall. The Journal of Geophysical Research-Atmospheres plans to devote a special section of its November 1998 issue to papers documenting the model and related research. "The nice thing," says Filippo, "is that it'll be the first time we've had all these studies assembled in one place." BH