Journal of Petrology | Volume 39 | Number 5 | Pages 1039-1061 | 1998
© Oxford University Press 1998
Reactive Melt Transport in the Mantle and Geochemical Signatures of Mantle-derived Magmas
Department of Geological Sciences, University of Washington Seattle, WA 98195, USA
Received April 17, 1997; Revised typescript accepted January 9, 1998
 |
Abstract |
|---|
Modeling of equilibrium reactive melt transport in the mantle, including the effects of mineralogic reactions and changes in porosity in the solid matrix, provides a series of robust predictions of the consequences of reactive melt transport on magma trace element compositions. The composition of the leading melt batch through a column of reactive mantle (the melt front) will be shifted towards that of an incipient partial melt of the mantle matrix. Successive melt batches migrating through and emerging from the column will show a temporal–compositional trend reflecting exhaustion of the reactive capacity of the mantle, and will eventually return to the original input melt composition. In cases where the melt source and column matrix are similar in composition, the melt front will be enriched in incompatible elements, and the temporal–compositional trend will be one of decreasing incompatible elements in erupted melt batches with time. Cogenetic melt batches should show some type of chromatographic decoupling in trace element and/or isotopic variations. However, mineralogic reaction in the mantle column, in the form of changing matrix mode, can smooth or mask chromatographic effects on trace element abundances, though not on isotopic compositions. Thus lack of chromatographic decoupling in magma trace element abundances alone does not preclude significant melt–mantle reaction. Mineralogic reactions within the column may also impart distinctive trace element variations to melts. In particular, lherzolite-to-dunite reaction in the matrix produces large variations in heavy rare earth elements in emerging melts, whereas lherzolite-to-pyroxenite reaction produces a series of subparallel rare earth element patterns with decreasing overall abundances with time. Although these models assume end-member conditions and maximum extents of melt–mantle reaction, if reactive melt transport is an important petrogenetic process that strongly influences magma trace element compositions these chemical effects should be observed to some extent in carefully chosen sample suites. Certain magma types show chemical characteristics that are broadly similar to these predicted effects, but more complete sample suites yielding detailed temporal–compositional variations of clearly cogenetic lavas are needed to test the petrogenetic significance of reactive melt transport on erupted magmas.
KEY WORDS: melt migration; trace elements; melt–mantle reaction
* Present address: Division of Geological and Planetary Sciences, MS 170–25, California Institute of Technology, Pasadena, CA 91125, USA. Telephone: (626) 395 6177. Fax: (626) 586 0935. e-mail: reiners{at}gps.caltech.edu