"Once you begin that life [as an environmentalist], you can't let anything go by," said Hannon, now a UIUC geography professor spending a half-time sabbatical with NCSA. Hannon's environmental commitment has informed his research since the sixties and is the driving force behind his latest endeavor--participating in a major effort to construct a global ecological economic model.
"Environmental change and economic activity are inextricably combined," says Hannon. Global emissions of CO2, methane, sulfur oxides, and other gases generated by economic enterprise affect the environment in a myriad of ways, according to Hannon, that collectively have profound effects on water quality, crops, and the growth patterns of forests, which in turn have their own effect on the economy and the global climate. "One approach to assessing such complex interconnections is to produce computer models of the world economies, of the world ecosystems, and the interactions between them," Hannon continues. "The effort will require constructing new models, interconnecting existing models, and assimilating existing data bases to simulate the impacts of major economic and environmental changes." A project of this magnitude will require the high-performance computational resources and the networking capabilities of NCSA, which will host the effort involving a large group of ecological and economic modelers from around the world. Contingent on the availability of funding, an invited workshop for the international group is planned.
Duchin's I/O model was used by the United Nations in preparation for the U.N. Conference on the Environment and Development, also known as the "Earth Summit," in Rio de Janeiro in June 1992 to explore strategies for sustainable development over the next several decades. According to the predictions of Duchin's model, the economic objectives laid out by the U.N. of growth in developing and rich nations alike cannot be met while reducing or even leveling off pollutants--"even making explicit assumptions about the rapid spread of much more efficient and cleaner technologies," says Duchin. "In order to meet these kinds of objectives one has to propose more dramatic departures from current practices, or the reliance on fuels or technologies that require some sort of breakthrough."
As a component of the global ecological/economic modeling project, this will be Duchin's first systematic study of the major interactions between the world economy and the ecosystem. "Here I'm thinking of things such as the use of water. . . and the implications of the land use from an economic point of view, as a commodity, as well as from an ecological point of view," says Duchin. She is currently involved in this type of analysis of the economy of Indonesia.
Computer modelof the Atchafalaya delta/Terrabone Parish marshes of southern Louisiana. Increase in suspended sediments from 1960 (left) to 1975 (right) is compared. Courtesy Robert Costanza, University of Maryland, and Fred Sklar, University of South Carolina.
The global ecological model will be constructed in much the same manner, by calibrating the unit models created with STELLA for each ecosystem type, including agricultural and urban "ecosystems." "The model has basically the same structure for all ecosystems, and what differentiates them is the set of parameters you use," says Costanza. "There's a lot of data out there that is not being effectively used, because no one has tried to synthesize it. One of the goals of the modeling project is to locate the existing data and to recommend projects to fill in missing data."
Costanza notes that there is a network of NSF-funded ecological study sites, the Environmental Protection Agency's (EPA) nationwide Environmental Monitoring and Assessment Program, satellite imagery including new imagery available as part of NASA's planned Earth Observation Satellite project, all of which will provide a wealth of data. An EPA-funded Multiscale Experimental Ecosystem Research Center (MEERC) in Maryland will address issues of scaling--using the small-scale information from microcosms and mesocosms to build large-scale models on a global level.
"If I convert my bottomfields from soybeans into prairie grasses, I change the biosphere, but I also change the economy. If more soybean fields are converted to prairie grasses, the price of beans will rise. Do soybeans become too expensive to supple-ment the diet of those in developing countries; do they begin clearcutting another section of rainforest in search of a new protein source?"
One of the challenges of developing a software environment for such a disparate group of people is making it simple to use. "We need to make it easy enough so that the entomologists can function as entomologists and not as computer scientists," says Hardin.
Just this summer, NCSA released the initial versions of a collaborative computational science software suite, called the NCSA Collage series. "This is a set of tools that will allow researchers using IBM PCs, Macintoshes, or any X Window System machine to carry on real-time collaborative work sessions with geographically distant colleagues," Hardin continued. "Researchers will be able to cooperatively construct, analyze, and debug programs on high-performance systems; they can have shared data visualization sessions in a WYSIWIS (what you see is what I see, the collaborative corollary of WYSIWYG) fashion."
For Hardin, the human factors are some of the most fascinating aspects of the collaborative session. "Human computer interactions are of continuing interest to cognitive psychologists. But now we need to be informed by social psychologists and sociologists. How do people interact when the medium of shared reality is a computer screen? What happens when the intelligence on the other side of the screen is not just a computer, but a human in league with a computer?
"We see the NCSA Collage series as being part of the effort to build a national collaboratory, and the global modeling work as being a major test case of the efficacy of such a national collaboratory," says Hardin.
The first workshop will focus on the Canadian Lakes Program, a very specific, highly measured experimental ecosystem. According to Hannon, the program has several lakes with rich databases on the variation in biological activity--and in some cases the geochemical cycles--spanning nearly 20 years.
Hannon will use STELLA as the modeling tool for the workshop. "STELLA allows graphical programming with a minimum amount of mathematics and no prior programming knowledge," says Hannon. "It allows the modeler to focus entirely on the modeling problem with essentially instant feedback on the effects of changes to the model." Because STELLA models can tax the limits of Macintosh PCs, the Software Tools Group is creating a version of STELLA distributed between the Macintosh and the CRAY Y-MP system. The user can develop and test the elements of a complex model on the Macintosh and then run the combined model on the CRAY Y-MP.
"Global economic and ecological modeling must be done and done well," says Hannon, "or we have no chance for rationalizing the degree of intercountry change and impact. It seems to us that, in the absence of good modeling information, effective cooperation between countries is superficial at best and antagonistic at worst.
"In some sense, it's just the scale of the project that's mind-boggling," continues Hannon. "But if we weren't working on boggling things, I don't think it would be very interesting."
__________________________________________________________________
access / July-September 1992 / NCSA / pubs@ncsa.uiuc.edu