The thickness of the mushy layer on the floor of the Skaergaard magma chamber at apatite saturation

We present a novel way of constraining the thickness of the crystal mush in fractionated layered intrusions using detailed microstructural analysis. The results are combined with geochemical data to create a snapshot of the crystal mush on the floor of the Skaergaard magma chamber in the period immediately before and after the saturation of the bulk liquid in apatite (the UZa-b boundary). The step-change in the fractional latent heat (that part of the total enthalpy budget associated with crystallization) accompanying the arrival of a new liquidus phase is recorded by a step-change in the median clinopyroxene-plagioclase-plagioclase dihedral angle, Θ_cpp, in fully solidified cumulates. Dihedral angles are formed during the last stages of solidification and hence the change of Θ_cpp associated with apatite-in marks a point close to the base of the mushy layer at the moment the bulk liquid became saturated in apatite, whereas the first appearance of abundant, homogeneously scattered, cumulus apatite crystals in the stratigraphy marks the top of the mushy layer at this moment. Comparison of the offset between these two markers in five widely spaced drill cores through the Skaergaard Layered Series suggests that the mushy layer was only a few metres thick at the UZ_a-b boundary in the centre and east of the floor, whereas it was ~100m thick on the floor near the western margin. There is no correlation between the efficiency of liquid expulsion (as recorded by bulk-rock P₂O₅ concentrations and the stratigraphic distribution of reactive symplectites) and the recorded mush thickness at the moment of apatite saturation, suggesting that existing models of adcumulate formation that depend on mush thickness need to be reconsidered.