Nonlinear nature of the unavoidable long-lived isotopic, trace and major element contamination of a developing magma chamber

Abstract

Contamination of magmas by crustal materials is a possible alternative explanation of many of the trace element and isotopic features of basalts which have been advanced as evidence of mantle heterogeneity. The space problems associated with the emplacement of large masses of plutonic cumulates in most central volcanoes, and with the formation of the gabbro-dunite portion of the oceanic crust, indicates that at least some digestion of pre-existing crustal rocks has occurred. Contamination is therefore unavoidable. In the ocean basins, contamination is not due directly to incorporation of sea water, but to the digestion of previously erupted basalts which may have been hydrothermally altered by circulating sea water, plus some sediment. In a magma chamber that is periodically replenished with parental magma and periodically tapped to provide lava flows during fractional crystallization, contamination produces changes in the composition of the erupted product which are not, in general, linear between the initial and final states of the system. Moreover, the change in composition produced need not lead towards the composition of the contaminant. Provided that the periodic additions and subtractions of magma are small relative to the size of the magma chamber, the effects of short-term variations in the amount of contamination are smoothed out, while the effects on incompatible element concentrations are persistent long after active contamination has ceased. The effects on major components that enter the crystallizing phases are constrained by the need for the remaining liquid to conform to the relevant low-pressure phase equilibria.