Eos, Transactions American Geophysical Union

JGR electronic supplement: A proposal



Scientific journals such as the Journal of Geophysical Research serve a dual purpose: first, as means of communication by which scientists can inform others of their results and, second, as archives—repositories of results for use by others, perhaps decades later. The editorial process of careful reviewing before acceptance is more important for the archiving function because an originating scientist may no longer be available to explain to a user how his results were obtained or may have forgotten some important details.

This archiving function has been squeezed considerably by the economic pressures on journals to limit the number of pages. Limiting is accomplished by such editorial criteria as succinctness of expression, elimination of supporting details, and inferences of interest to a broader audience than those close to an author in technical expertise. The effects of these criteria are generally good, inducing terse writing and increasing the probability that a reader will find something of interest. But they also increase the chances of obscurity, reproducibility, and duplication of effort. And they lead to details of data, modeling, and explanation essential for efficient use of results by others being relegated to unrefereed “gray” literature or other informal means. Modern data collection and analysis involve sampling, instrumentation, and techniques of considerable subtlety and complexity. However, the relevant reports are often poorly distributed and without assurance of future availability. Thus, later generations may find it difficult or impossible to aggregate and interpret the historic data record because critical documentation is lacking. For example, past and future interdecadal and longer variations of climate and other parameters must be determined to separate natural and anthropogenic changes and provide modelers with a means of validating their model calculations via “hindcasts.” Compiling accurate climatologies requires access to details of the individual data collections incorporated.