ricanes, tornadoes, floods, earthquakes, and volcanic eruptions. Besides the societal impacts associated with climate, natural hazards, and natural resources, there are numerous man-made hazards that are coupled with natural phenomena that are the subject of Earth science research. Examples include the prediction and mitigation of pollution plumes in ground water or the atmosphere (e.g., chemicals or radioactive materials), the factors involved in stratospheric ozone depletion, and the monitoring of treaties that ban underground nuclear testing (e.g., in support of the Comprehensive Test Ban Treaty).
The physical, chemical, and biological processes that shape the world in which we live are complex and interdependent. To understand them requires observations with sufficient spatial and temporal resolution and coverage to characterize the phenomena of interest and to constrain theoretical predictions that are based on conceptual or quantitative models. Therefore, the lifeblood of research in most of the Earth sciences is observational data, sometimes global in coverage, and taken repeatedly over time. Many of these data also must be integrated with data from experimental manipulations, or from other disciplines.
An example is atmospheric circulation, which controls weather over the entire Earth with significant variations on time scales ranging from hours to decades or longer, and spatial scales ranging from less than 1 km to thousands of kilometers. Weather forecasts for more than a day at a time require the rapid and repeated acquisition, processing, and interpretation of very large amounts of synoptic observations on at least a continental scale. Satellite systems that gather the necessary data have been and are being developed, but timely access to the data gathered by different organizations or countries is a major concern. Climate studies require many of the same observations as for weather prediction, but also data on the oceans, land surface, and cryosphere for the entire Earth. Therefore, international sharing of very large volumes of global atmospheric circulation data is essential for meaningful scientific investigation of past and present climates.
Scientific knowledge in the various subdisciplines of the Earth sciences has advanced to the point where important, multidisciplinary global-scale problems can be tackled with insight and scientific rigor, provided that high-quality global observations are available and that computational resources are adequate to process and interpret large and diverse data sets. Major examples of interdisciplinary and integrating research programs in the Earth sciences are the World Climate Research Program of the World Meteorological Organization (WMO) and the International Council of Scientific Unions (ICSU), the International Geosphere-Biosphere Programme organized under the auspices of ICSU, and, nationally, the U.S. Global Change Research Program.10 These are major initiatives, begun in the 1980s to understand the driving mechanisms (both natural and human) that cause significant changes in the Earth system. These efforts involve collecting and analyzing massive data sets from Earth-observing satellites and integrating them with multiple-area or site-specific data all over the Earth, including developing countries. Significant progress in these types of complex