Predicting species’ distribution changes in the face of global environmental change is fundamental for guiding policy decisions at both local and global scales. In an ideal world, an exhaustive record of distribution changes across the entire range of spatial (local to global) and temporal (from the species’ inception to the present) scales would enable a predictive understanding of these responses. Unfortunately, for most species, we still have inadequate knowledge of their spatiotemporal distributions. Although natural history museum and herbaria collections, photographs, field notes and other historic data sources contribute a great deal to our current state of knowledge, they are unlikely to convey an exhaustive knowledge of spatiotemporal distribution changes because collections were made in an unsystematic and opportunistic manner. Still, these data sources potentially hold essential information on aspects of spatiotemporal changes in species’ distributions (e.g., range limits, phenology, species co-occurrence) which could not be obtained otherwise; how can they be used to provide useful approximations of species’ spatiotemporal distribution changes under global change? This talk will provide an overview of how unsystematic historic data sources have been previously used and misused for examining spatiotemporal changes in species’ distributions, as well as a prospective view of promising approaches which may improve our predictive understanding of future global change impacts.