GARNET: A COMPREHENSIVE, STANDARDIZED DATABASE OF GARNET GEOCHEMICAL ANALYSES INTEGRATING PROVENANCE AND PARAGENESIS

Integrating mineralogy with data science is critical in order to modernize the geosciences. Garnet samples were compiled from a variety of sources ranging from large databases (EarthChem, RRUFF, MetPetDB) to individual peer-reviewed literature. It is important to include papers published prior to 1980 in order to prevent the loss of “dark data” and expand the availability of mineralogical data. Since garnets are commonly used as kinematic indicators of formation environments, which directly correlates with geochemical data, they were an ideal subject for the creation of an extensive database. This standardized database consists of over 95,000 garnet sample analyses and incorporates locality information, paragenetic modes, age, temperature, pressure and geochemical information. The major oxides are reported in weight percent (wt%), while trace elements and rare earth elements (REEs) are recorded in parts per million (ppm). Paragenetic and locality attributes for each sample were classified by increasing resolution from generic material type (i.e. Igneous, Metamorphic, Sedimentary, etc.) to more specific rock type names (i.e. Eclogite, Diorite, Skarn, etc.). A similar system was adopted for locality information, increasing specificity by examining the Continent, Country, Area, Context, Longitude, and Latitude. An additional 277 original electron microprobe (EMP) analyses, performed at the Carnegie Institution for Science's Geophysical Laboratory, examined 9 garnet samples across common garnet species including: pyrope, andradite, almandine, spessartine, grossular, and uvarovite. Initial results from ternary diagrams display a clear distribution between garnet types, skewed by a large number of igneous and metamorphic samples. Preliminary cluster analysis using k-means reveals 2 distinct clusters dependent on the concentration of major oxides. This data resource will be examined further in future cluster analysis to determine if there is a correlation between composition, locality, and paragenesis. This comprehensive database of garnet information is intended to be an open access resource for future mineralogical studies. We encourage scientists to contribute to vast databases of mineralogical information that are valuable to advancing the process of scientific discovery.