Integration of deep-time databases: towards building a deep-time digital Earth

The integration of deep-time databases is a key aspect in the creation of a comprehensive digital model of Earth's history, known as the "deep-time digital Earth." This model would enable scientists to better comprehend the intricate processes that have shaped our planet and its life forms over time. Currently, the data are dispersed across various databases and research institutions, making it challenging for scientists to fully utilize the information contained in the data.

We present a virtual integration of deep-time databases, where the data remain in the sources and are accessed as needed at query time. We will demonstrate this integration using the Macrostrat, PBDB, GBDB, GeoLexicon, and the Paleogeographic Atlas Project databases. The first step involves identifying the available information from each data source and deciding on the relevant data attributes, such as lithology, age, and formation name. The next step is aligning the schemata of different data sources to a common mediated schema, allowing attribute names with the same semantics to be merged. For example, Palaeo-block in GBDB is equivalent to geoplate in PBDB. The third step is to create a virtual integration layer that allows users to access and query data from various sources as if they were stored in a single database. This virtual integration layer uses the mediated schema to translate queries and data among different sources and provides tools and interfaces for data visualization and analysis. Our goal is to make deep-time data more easily accessible and usable for a better understanding of Earth's history.