Magmatic evolution of translithospheric diapirs: a thermodynamic modelling

Abstract:

In the current model of generation of magmatism in continental arcs, subducted sediments mix with basalts from the subducting slab, forming diapirs that ascend through the mantle wedge and undergo melting due to heating and decompression. These diapirs subsequently relaminate to the base of the continental crust, and segregation of magmas from this relaminated body can form Cordilleran granitoids^[1]. Consequently, it is relevant to determine the composition and amount of melt that can form in the diapirs, and particularly the influence that the sediment:basalt proportion in the diapir has in those melt variables. Thermodynamic calculations using the Perple_X software^[2], the hp633ver.dat thermodynamic database^[3] and several solution models^[4] have been made following a P-T path from 500 °C and 3.0 GPa to 900 °C and 1.0 GPa^[5]. The results show that the melt composition and amount at the latter P-T conditions is fairly homogeneous irrespective of the initial diapir composition (SiO₂: 70.25-72.09 wt.%; Al₂O₃: 12.69-13.68 wt.%; alkalis: 8.46-9.35 wt.%; FeO_t+MgO: 4.10-4.53 wt.%; A/CNK: 0.89-0.90; amount: 21.6-24.8 vol.%), in agreement with experimental studies^[1]. The melt amount in the relaminated diapir is high enough to form an interconnected network, and thus melt can segregate from the diapir and ascend through the continental crust. However, the modelled melts are slightly depleted in Al₂O₃ and CaO and enriched in K₂O compared to Cordilleran granitoids, causing slightly lower values for the A/CNK. Co-segregation of plagioclase with the melt (i.e. restite entrainment) could improve the fit of the magmas with the granitoids. Subsequent processes like fractional crystallization can modify these compositions to produce the compositional and lithologic variability observed in continental arc magmatic bodies.

This work was funded by the Universidad de Los Andes, “Fondo de Ayuda para Profesores Asistentes”, number INV-2019-63-1701.

[1] Castro et al. (2010) JPet. 51, 1267-1295. [2] Connolly (2009) G³, Q10014. [3] Holland et al. (2011) JMG 29, 333-383. [4] Holland et al. (2018) JPet 59, 881-900. [5] Vogt et al. (2012) PEPI 192-193, 1-20.