have to be maintained, so that new directions and opportunities can be pursued as our understanding is improved and research directions are refined. In order to achieve its goal, the U.S. Dec-Cen Program must represent a balance between the following elements:
1. A long-term, stable observing system to constantly monitor, with sufficient accuracy and resolution, a subset of crucial earth-system variables (i.e., key state variables and primary forcings) on dec-cen time scales. A modeling activity should be an integral part of this system, to add value to the observations by assimilating them into suitable models. By more clearly delineating the actual nature of decadal variability in the climate system, the capability for detecting anthropogenic climate change will be improved, thereby making an important contribution to the IPCC's goals. The observing system can be built incrementally, but must be initiated immediately, because a long, sustained climate record is required to document the wide range of climate variability and change over dec-cen time scales. (Specific details are outlined in Chapter 6.)
2. A hierarchical program of modeling studies that has the ultimate goal of simulating and predicting the entire Earth system on dec-cen time scales. This element requires an infrastructure that makes powerful computational and communications resources readily available to the full modeling community. (Again, specific details are outlined in Chapter 6.)
3. Process studies that address those geophysical, chemical, and biological processes which are so poorly understood that they hinder our ability to define, understand, and predict dec-cen variability. These studies should include analytic, model-based, and observational ones, as well as combinations of the three. (Fundamental processes requiring particular attention are listed following the issues in each subsection of Chapter 5.)
4. Producing and disseminating long-term proxy and instrumental datasets for use in the study of dec-cen variability. These datasets provide the only means for immediately securing a reliable, albeit preliminary, assessment of past dec-cen variability, and for establishing the pre-anthropogenic-effect baseline of natural variability. Because the sources of some of the most well-established proxy records are at risk of permanent destruction in the very near future (e.g., alpine glaciers and long-lived stable tropical corals), it is imperative that these records be secured while it is still possible to do so.
The recommended U.S. Dec-Cen Program would be part of the U.S. contribution to the DecCen and ACC (Anthropogenic Climate Change) components of the Climate Variability and Predictability (CLIVAR) Programme of the WCRP. It would also contribute to the International Geosphere-Bio-sphere Programme (IGBP). The scope of the U.S. program would be broader than that of either the WCRP or IGBP alone; the research objectives would deal in an integrated manner with the biological, terrestrial, oceanic, and atmospheric aspects of climate variability on dec-cen time scales. It would thus draw on the contributions already being made by such programs and projects as the IGBP's PAGES, the WCRP's GEWEX, SPARC, ACSYS, and WOCE, and GCOS, GOOS, and GTOS.
The U.S. Dec-Cen Program must also be well coordinated with the U.S. and WCRP seasonal-to-interannual program, the Global Ocean-Atmosphere-Land System (GOALS) Program, because the dec-cen influence on short-term variability and change may aggravate the difficulties of short-term climate prediction. Indeed, several components of the proposed Dec-Cen Program cannot be separated from those of the GOALS Program (e.g., determining the influence of dec-cen variability on the frequency, intensity, and duration of E1 Niño events).
The ultimate research objective of a U.S. Dec-Cen Program would be to define, understand, and model dec-cen climate variability and changenatural and anthropogenicso that the extent to which they are predictable can be reliably determined. If it can be shown that they are predictable, the ultimate practical aim of the Dec-Cen Program would be to design and implement a system that predicts the various aspects of climate at the temporal range permitted by their predictability. This system would build on the emerging seasonal-to-interannual prediction systems now being constructed, predict future decadal-to-centennial variations to the extent possible, and learn to use these predictions for the benefit of all.
The dec-cen prediction system would have to investigate the future responses to whatever external-forcing scenarios are imposed, both naturally and anthropogenically, upon the Earth's climate system. It would provide a basis for distinguishing natural variability from anthropogenic change, to the extent theoretically and practically possible. The understanding of this distinction will yield a strategy for the detection and attribution of any anthropogenic climate change that is superimposed on the background of natural variability.
Decadal-to-centennial climate variability is a subtle phenomenon. It proceeds too slowly to be perceptible to our senses, but its cumulative effects ultimately define the life prospects of future generations. Informed stewardship of the Earth's resources for the generations to come must draw on the insights afforded by model predictions and model-aided extrapolation of observations that can give us a glimpse of the future. A U.S. Dec-Cen Program will be the first step toward assuming this responsibility.