Rescuing fragile scientific data is crucial to ensuring that this data can continue to be part of the scientific process. The complementary documentation of data rescue efforts is also essential for ongoing scientific study and future data rescue. This paper reports on a study that examined how best to document and describe data rescue activities. A mixed methods approach included a case study of 20 data rescue initiatives and a content analysis of seven descriptive metadata schemes. The results identified 13 metadata elements that were common across schemes and well suited to describing rescue projects. The findings were used to develop a core metadata scheme for the description of data rescue activities. The study was undertaken as part of the Data-at-Risk Initiative (DARI). Results of the study will help inform future DARI efforts and contribute to the knowledge base on data rescue and project-descriptive metadata.