With the relatively recent awareness of the fragility of the natural world environment (Rio Conference "Earth Summit" in 1992), Information Systems (IS) trying to identify and help to manage biodiversity have proliferated. Given the various data that they contain, they are complex applications, hardly accessible to the general public. At the same time, new communication systems allow a more natural assimilation of information: three-dimensional representations of immersive virtual worlds, also known as metaverses. By putting users in a reality reproduction built from IS, all information can be represented in a consistent virtual environment. But turning IS's data in a coherent metaverse is not a trivial process. Depending the IS's structure and the availability of metadata, several virtual worlds, representing several level of details, can be generated. This paper is a contribution to the analysis of the different possibilities to generate metaverses representing the information stored in the IS dedicated to biodiversity. It presents the generic process of worlds generation, the requirements, the different models of metaverses' structure, and the metadata that could be used to refine the level of detail.