Chronological datasets that span large periods of time are the focus of study in several diverse disciplines: climatology, geology, biology, history, and anthropology, to name just a few. While these datasets share strong links to time scales, they exhibit a great diversity in most other aspects. For example, some datasets are derived from ice cores, while others arise from archaeological artifacts, instrumental records, and numerical modeling. In each discipline, researchers have spent considerable effort to carefully developing a time scale for each such dataset, and to interpreting the resulting temporal data, in turn leading to scientific insights. Such chronological datasets are growing rapidly in both size and diversity, due to not only increasing activity by scientists worldwide but also improving technology. The integration and analysis of this large and diverse collection of datasets promises to yield significant scientific breakthroughs. However, such integration and analysis currently requires a very significant human effort, which severely restricts our ability to realize its potential value. Most of the analysis techniques developed in the past do not scale up to the size and diversity of data we must now analyze. The goal of this project is to break this barrier to scientific breakthroughs by developing methods to solve the problems arising from the diversity and volume of data, and to implement, test, and apply these methods in an interactive workbench.
Developing the above workbench requires work in multiple cyberinfrastructure research topics, which by themselves have received considerable research attention. By their nature, however, these topics do not admit many universal solutions and there is currently a large gap separating the generalized results available in the informatics literature and their realization in operational systems that are ready for use by domain scientists. Our effort will focus on bridging this gap by adapting and specializing known methods, and developing new ones when needed, to the characteristics of the chronological data of interest. This work is not limited to the task of implementing cyberinfrastructure for the described domains based on known methods. Rather, the work requires research effort to map concepts and techniques from computer science to various scientific domains, and vice versa, carefully adapting prior results and developing new ones. This work requires not only expertise on both the cyberinfrastructure and domain sciences sides of the gap, but also a close- knit collaboration that can overcome traditional inter-domain barriers.
By enabling effective use of the large and diverse collection of chronological data, our workbench opens the door to scientific breakthroughs in multiple disciplines. As just one specific example, our workbench will improve our understanding of one of the most important scientific and societal problems: the operation of the global climate system, past, present, and future. In addition to enhancing the effectiveness of the scientists specializing in each discipline, our work will allow researchers from other disciplines, and people with a more casual interest, to easily access and, importantly, process and interpret the same data from which key research results and forecasts emerge. This aspect is particularly important as it demystifies the related research that may otherwise appear arbitrary or opinionated. Shared examples will be traceable by others all the way back to raw data. Our results will be disseminated not only through the traditional professional channels but also through outreach to the broader community.
