Provenance in scientific research involves documenting the history of data, focusing on the changes it undergoes and the relationships between various data points throughout its entire lifecycle. By capturing detailed information on how data is processed and transformed, provenance enables researchers to trace, validate and ensure the integrity of their results. This process becomes particularly crucial in complex workflows and high-performance computing environments, where managing intricate tasks and data relationships is essential for maintaining scientific rigor and fostering collaboration. This paper highlights the importance of supporting data reliability and traceability of scientific workflows, providing valuable insights and tools to tracking provenance in Workflow Management Systems.

Open Science is a vital part in the current and future research agenda worldwide. In order to meet Open Science goals, it is of paramount importance to fully support the research process, which includes also properly addressing provenance and reproducibility of scientific experiments. Indeed, provenance and reproducibility are two key requirements for analytics workflows in Open Science contexts. Handling provenance at different levels of granularity and during the entire experiment lifecycle becomes key to properly and flexibly managing lineage information related to large-scale experiments as well as enabling reproducibility scenarios. To this end, this work introduces the micro-provenance concept, and it provides an in-depth description of its design, implementation and exploitation in the context of a multi-model climate analytics workflow.