Data provenance for preservation of digital geoscience data

A necessary first step in the preservation of digital scientific data is gathering enough information “about” a scientific outcome or data collection, that it can be discovered and used a decade from now as easily as it is reused next week. Data provenance, or lineage of a collection, can capture how a particular scientific collection was created, when and by whom. Our goal is to devise tools automate the collection of provenance so that this task does not fall onto the researcher, and to efficiently store and represent the provenance data that makes the data more amenable to long term preservation. We demonstrate through application to several projects that automated provenance collection can reach the level of necessary provenance but challenges remain in addressing provenance collection in a non-workflow setting, and in data preservation in cyberinfrastructure architectures.