Modelling the Formation of Linear Geochemical Trends Using the Magma Chamber Simulator: A Case Study of the Jindabyne Granitoids, Lachlan Fold Belt, Australia

Abstract

Abstract Understanding the origins of major and trace element variations and the isotopic character of granite samples in terms of sources and magmatic processes is, arguably, the core of granite petrology. It is central to attempts to place these rocks in the context of broader geologic processes and continent evolution. For the granites of the Lachlan and New England Fold Belts (LFB and NEFB) of Australia there has been great debate between competing petrogenetic models. The open-system view is that the isotopic variability and within-suite compositional trends can be accounted for by magma mixing, assimilation and fractional crystallisation (FC). In contrast, the restite unmixing model views the isotope compositions of diverse granites as a feature inherited from individual protoliths that underwent partial melting to produce magmas entraining varying proportions of residual material in a felsic melt. Reconciling all aspects of the geochemical data in a mixing model is contingent on a plausible fractionation regime to produce the observed consistently linear (or near-linear) trends on Harker diagrams; however, published FC models lack phase equilibria constraints on the liquidus assemblage and do not account for the likely changes in trace element partitioning across the modelled compositional range. The Magma Chamber Simulator (MCS) can be used to model fractional crystallisation alone (FC) or with assimilation (AFC), constraining phase equilibria and accounting for the thermal budget. Here, this tool was used to conduct a case study of the I-type Jindabyne Suite of granites from the LFB, testing whether thermodynamically feasible geochemical trends matching the observed linear variations can arise through FC (with or without assimilation of supracrustal material). The results of 112 MCS models show: (1) that major element liquid lines of descent (LLDs) may be sensibly linear over limited compositional ranges, (2) that the involvement of assimilation extends the range in which trends are relatively simple and near-linear, and (3) that, despite these observations, neither FC nor AFC are able to correctly reproduce the geochemical evolution of the I-type Jindabyne Suite granitoids as an LLD (contrary to existing models)—instead, these processes persistently produce curved and kinked trends. The output of these simulations were further refined to explore models in which: (1) crystal-bearing magmas evolve via FC or AFC (with chemical isolation assumed to be achieved through crystal zoning) and undergo varying degrees of melt-crystal segregation at different stages to produce the sample compositions, and (2) in situ crystallisation occurs via FC within the crystallisation zone, driving the evolution of a liquid resident magma, which the samples represent. These models are able to reproduce the Jindabyne Suite trends reasonably well. The modelling implies that FC, or some variant thereof, is a viable explanation for the linear trends in Jindabyne; however, tendency for grossly non-linear LLDs highlights that it should not be assumed that FC can generally explain linear trends in granites without careful modelling such as shown here.