their results (e.g., Fisheries Oceanography vol. 10, [Suppl. 1]—“A Sound Ecosystem Assessment [SEA] Synthesis”), most have not. Many studies that have been funded over the past 13 years have yet to be published. Annual reports are not publications and certainly do not qualify as syntheses.
The knowledge gained about Prince William Sound is extensive because of Exxon Valdez oil spill funding. Retrospective analyses have led to new hypotheses and ideas in many instances, not the least of which is the concept of a “regime shift” (Francis and Hare, 1994; Hollowed and Wooster, 1995; Anderson and Piatt, 1999) and the Pacific Decadal Oscillation (Mantua and Hare, in press). However, there is much more to be gained from past studies that should be used to direct the future of GEM. The completion of the third step will lead to the fourth step: assessment of accomplishment of past goals. The synthesis of data and assessment of what has been learned in the recent studies will provide a starting place from which to hone hypotheses needed to direct GEM research.
The generation of new hypotheses will lead to proposals for new work, which in turn will lead to the need for additional synthesis. Synthesis is an iterative process and as such is both the first and last steps. For GEM to continue to be successful, periodic re-synthesis of new data will be needed. A synthesis will assure that there is not a long lag time in publication of results and access to data of other GEM researchers, such as currently experienced under Exxon Valdez oil spill. A periodic synthesis on the scale of the five-year increments will promote comparisons between past and recent conditions. Additionally, scheduled syntheses will ensure evaluation of program direction.
One presumption in a long-term program is that technology will change, providing opportunities for collecting new data types or collecting existing data more efficiently. Another presumption is that users will become more sophisticated, and their needs will change as they become accustomed to the data streams that are produced. Many successful programs incorporate periodic program review to assess how the program should change in response to these new collection opportunities and needs. (Weisberg et al., 2000).
The synthesis will tell whether the science plan and the structure of the program is working.
As GEM is envisioned to be a 100-year plan, we suggest that a time line on a scale longer than five years be included in the GEM plan. We have emphasized that long-term research is the linchpin of this program, and the projected time line should reflect that effort. Within that time line periodic syntheses should figure prominently. Synthesis should be viewed as a key component of the plan and funding for synthesis should be incor-