Provenance makes it possible to describe information about the steps involved in the production of a piece of data and allows an assessment of its quality, reliability, or credibility. When it comes to scientific workflows, provenance establishes the relationships between the artifacts associated with a given set of simulations and can be used to enable: (i) their sharing with the scientific community, (ii) the reproducibility of the results, or (iii) the evaluation of erroneous outputs. This paper presents the work in progress towards building a graphical provenance modeling tool conforming to the W3C PROV standard and following Model-Driven Engineering (MDE) concepts. The modeling tool can be used to model scientific workflows' provenance, enabling, for instance, their visual representation, reproducibility, and sharing.