Skip to content



Integrating the Infrastructure for the Advance of Scientific Understanding

Sid Karin, NPACI Director

I n the last ENVISION, I invited readers to join me in supporting the recommendations of the President's Information Technology Advisory Committee (PITAC). The efforts we and others have made to give life to these recommendations have now borne fruit.

At the AAAS Annual Meeting in Anaheim in January, the administration proposed a $366 million dollar initiative, to be called Information Technology for the Twenty-First Century (abbreviated IT2). Representing a 28 percent increase over previous spending in this area, IT2 would support long-term information technology research, advances in computation, and research on the economic and social implications of the information revolution. Federal agencies sharing in the task of funding this research include NSF, DARPA, DOE, NASA, NIH, and NOAA.

This IT2 initiative is certainly a direct result of the PITAC report, containing the recommendations of the nation's most distinguished information technology researchers. It is a welcome endorsement of the call by Ken Kennedy, Bill Joy, and the rest of the PITAC members for a balanced program of long-range, basic research. Advances in computer science that will be enabled by the initiative are urgently needed by the research community.



NPACI All-Hands Meeting group photoAll-hands Meeting 1999
NPACI partners met on the UC San Diego campus in January to discuss NPACI's progress and plans for the next year and where NPACI will be in the year 2002.



There are many ways in which our community can enhance American science and engineering. The PITAC call for long-term research was not unique. There have been several calls for more intensive research programs issued by science policy and academic groups. In connection with this issue of ENVISION, I want to mention the statement on global climate change issued by the American Geophysical Union (AGU) in December 1998. The AGU points to the significant uncertainties in the processes by which current climate variables are modeled and their future course predicted.

"Nonetheless," the AGU says, "scientific understanding based on peer-reviewed research must be central to informed decision-making. AGU calls for an enhancement of research to improve the quantification of anthropogenic influences on climate."

Climate change and changes in the quality of the environment have indeed engendered impassioned debates. Such debates are, perhaps, inevitable. But the strongest force in the effort to respond to change in a way that benefits the earth and its inhabitants is the dispassion of advancing knowledge, which will enable us to predict, anticipate, and mitigate disasters both sudden and subtle. The AGU's call is correct.

Top | Contents | Next


The global climate models do not exist in a vacuum. They obtain their predictive force from the accuracy of representations of their parts: regions in which strong interactions take place between air and land or sea. This issue of ENVISION concentrates on current work in NPACI's Earth Systems Science thrust area at the subglobal scales.

Whether the topic is crisis management in bays and estuaries, restoration of biodiversity, or chemical interactions in the atmosphere, the ongoing work requires high-performance computational resources--not just for each model or module in itself, but to facilitate the interconnection and integration of models into accurate pictures of a rapidly changing world.

The theme of the 1999 NPACI All-Hands Meeting was integration, and this ENVISION emphasizes the myriad ways in which regional models may be integrated into the supra-regional fabric of a global science. Parallel techniques and the active linkage of models to models and data are creating bottom-up answers to the "top-down" questions posed by the global models. We no longer live in a universe of isolated models whose assumptions differ so from one to the next that interconnection seems a dream.

Factors never thought critical, or even important, are proving their subtle power. It is through the integration of our most accurate models--in what Edward O. Wilson (echoing William Whewell) would call their consilience--that our confidence in their predictions will be built. Confidence requires the effort in information technology that has been proposed, and it requires the zealous attention of us all.

Top | Contents | Next
Top | Contents | Next