Mineral Informatics: A Key to Deep-Time Data-Driven Discovery in Earth and Planetary Sciences

Abstract

Minerals and the rocks that hold them are the oldest objects that we can hold in our hands. Each specimen is an information-rich time capsule, waiting to be opened. Every sample contains scores of attributes, including trace and minor elements, stable and radiogenic isotopes, solid and fluid inclusions, optical and magnetic properties, structural defects, twinning, exsolution, zoning, and more. These diagnostic attributes are enhanced by information on the ages, associations, and geological contexts of specimens. Collectively, these characteristics of minerals tell stories of the origins and evolution of planets and moons.Mineral informatics[1], a field championed by the Deep-time Digital Earth Program, advances understanding of planetary evolution by pursuing three complementary objectives. First is to develop comprehensive open-access data resources for minerals and rocks. Several vital platforms, including mindat.org, earthchem.org, and rruff.info, provide large and growing resources for researchers. Ongoing work will build these essential research tools, while advancing a culture of FAIR data.A second objective of our work is to establish a new mineral classification system based on mineral informatics that highlights formation processes and evolutionary stages of minerals [2-6]. Traditional approaches to classifying minerals ignore this history. The International Mineralogical Association has catalogued >6000 mineral species, each with a unique combination of idealized chemical composition and crystal structure. This essential scheme facilitates identification of different condensed crystalline building blocks of planets and moons. However, the IMA system lacks contexts of time and process. We have introduced, and are now completing, a new and complementary approach to mineral classification called the “evolutionary system of mineralogy.” Our system differs from that of the IMA in three important ways. (1) We split many IMA species based on their varied modes of formation and age of earliest occurrence. For example, diamond (carbon in the diamond crystal structure in the IMA system) occurs in at least 5 distinct mineral “kinds” in the evolutionary system, including stellar diamond and impact diamond. (2) We lump varied IMA species that form continuous solid solutions through the same process, for example, combining different species of the tourmaline group into a single kind. (3) We include  amorphous  solids, such as obsidian and limonite, which are important in crustal processes. A third goal of the mineral informatics program is to develop and apply advanced methods of data analysis and visualization to better characterize evolving mineral systems through more than 4.5 billion years of planetary history. To this end, we have incorporated network analysis, cluster and analysis and community detection, association analysis, and other methods to quantify the changing diversity and distribution of minerals through deep time, while estimating total mineral diversity and predicting new localities of critical mineral resources.