Major research universities are grappling with their response to the deluge of scientific data in its big data and long tail data forms. The latter consist of many diverse and heterogeneous sets, the data are collected via diverse and specialized methods, and are stored in a variety of formats and places. University libraries and their institutional repositories have traditionally been able to handle scientific output. But long-tail scientific data introduce substantial challenges to a traditional document-based repository through its vast heterogeneity, size, and its demands for meaningful discovery and in the case of large data sets, place-based use. In this presentation we will provide a brief overview of the NSF-funded project "Sustainable Environment - Actionable Data" (SEAD), which addresses the challenges of long-tail scientific data with the focus on sustainability science. We will provide an overview of this project and of its discovery and preservation component, called SEAD Virtual Archive. This component is being developed by the Data to Insight Center team at Indiana University in collaboration with IU and UIUC libraries. We will describe main features and our ongoing work on SEAD Virtual Archive and discuss the value and importance of partnerships between data research centers, such as D2I, and the libraries.