RESEARCHING CONTINENTAL EVOLUTION WITH COMPUTER-ASSISTED DESCRIPTIONS, COMPARISONS AND SIMILARITY-RANKINGS OF PLUTONS

Plutons are important constituents of the continents, from which information may be gleaned about materials beneath the continents, the processes which formed the continents, and prospective regions for minerals exploration. The classifying of plutons into different classes plays an important role in researching these aspects of plutons, and in making useful generalizations about different kinds of plutons. Since there are many plutons on earth (approximately 190 identified in Nevada alone), and many different ways to classify them, a computer system which can be used to describe and rapidly compare or classify plutons is a useful research tool.

Developing such a system, however, presents challenges in the standardization of geological vocabulary – not only in respect of rock and mineral nomenclature, but also in respect of terms used to describe quantitative attributes such as element concentration, and in respect of frequency attributes, such as how often plutons are expected to be mafic in composition. Once these vocabulary standards are in place, the computer has to be programmed to reason with the standardized vocabulary as closely to human reasoning as possible. For purposes of credibility, it needs to be able to provide to its user explanations which detail how it derived the rankings and classifications it generates.

This paper describes prototype solutions to the vocabulary issues surrounding pluton description, and highlights shortcomings which remain in our geological vocabulary. It illustrates how the prototype vocabularies can be used to great advantage in speeding up the researching and classification of plutons, focusing on the differences between plutons derived from oceanic slab melts and melts derived from partial melting of the mantle wedge.

The pluton classification system was implemented as a free web service named PlutonMatch. Any number of similar systems can be implemented for the description and classification of other entities of interest to the earth sciences. The fact that these systems are likely to use some of the vocabulary used by the PlutonMatch system raises important issues around the international standardistion of geological vocabularies and the global sharing of earth sciences knowledge on the internet.