THE ROLE OF RESIDUAL MELT MIGRATION IN PRODUCING COMPOSITIONAL DIVERSITY IN A SUITE OF GRANITIC-ROCKS

We present a quantitative model for in situ crystallization within a s olidification zone (or boundary layer) based on the trace and major el ement compositions of plutonic igneous rocks and related geochemical d ata. We developed the model to account for the characteristics of a su ite of granitic rocks: nearly uniform mineral compositions in rocks th at range widely in bulk composition (e.g., 58-76 wt.% SiO2); linear va riation and correlation of all major and trace elements analyzed excep t Ba; large and apparently random variations in Ba concentration. Thes e characteristics cannot be explained by any standard petrogenetic mod el involving fractionation, mixing, or restite unmixing, but are succe ssfully reproduced by our model of residual melt migration. We do not attempt to model the entire crystallization history of a pluton but, r ather, only that interval during which melt migration processes have r ecognizable geochemical effects. For the granitic suite, the chemical signature of residual melt migration resulted from the change in Ba co mpatibility with the onset of orthoclase crystallization at a granite minimum. Our results demonstrate that plutonic rocks can develop large compositional variations, comparable to those expected to result from extreme differentiation within a large magma body, over short distanc es by melt migration under conditions of high permeability or slow cry stallization rate within the solidification zone. Melt migration is pr obably a common process that could be overlooked in large plutons if s ampling is sparse and if variations in some components that appear ran dom are not considered. Our model equations yield estimates for parame ters that describe the proportions of residual melt that crystallize w ithin and migrate out of the solidification zone. Values of these para meters can be used to infer information about permeability and melt mo bility within the solidification zone. The model parameters derived fo r the granitic suite have a roughly concentric spatial pattern in the pluton, suggesting that residual melt was trapped near the margins, ac cumulated in the interior, with a zone of enhanced melt mobility and p ossibly compositional convection in between. Our model may be of gener al usefulness because it requires no assumptions about magma chamber g eometry or magma dynamics, it is applicable to magmas of any compositi on, and the equations could be formulated to include those variables b est constrained by a particular suite of plutonic igneous rocks.