The importance of data mining for conservation science: a case study on the wolverine

Abstract

Assessing the scale of ecological changes that have occurred since the onset of the Anthropocene is challenging. One major problem is that of shifting baselines, whereby the norms we set for judging the state of species, populations, or ecosystems change over time due to incomplete information. Here we show how data mining can be used to fill some of the information gaps fueling shifting baselines. We used as example an elusive species, the wolverine (Gulo gulo), given that information gaps are so prevailing for such species. We applied the concept of data mining to search documents hosted on publicly accessible online repositories and found information about the historical occurrence of wolverines that allowed us to revise their historical range in eastern North America. We found 12 historical accounts attesting the presence of wolverines in various parts of the Maritime Provinces of Canada, of which 11 were new to contemporary science. According to our results, the eastern limit of the historical range of the wolverine should be extended to include the current jurisdictions of New Brunswick and Nova Scotia. Biological change is the central paradigm of species status assessments. We show that online repositories of public domain literature can now be critical sources of information to assess biological change, including in the case of elusive species. Data mining constitutes a productive tool to uncover useful knowledge hidden in a sea of digitized historical information, and should thus allow researchers and conservationists to more effectively mitigate the problem of shifting baselines.