Mantle heterogeneity controls on small-volume basaltic volcanism: COMMENT

A recent article published in Geology by McGee et al. (2015) investigates the relationship between chemical characteristics and erupted volumes in the Auckland Volcanic Field (AVF), New Zealand. In the study, five eruptions from a possible 53 were selected from the field to provide insight into mantle source properties and processes. Based on this sample suite, a strong correlation is hypothesized between eruption volumes and chemical parameters used to draw a number of conclusions on the timing and nature of mantle melting. We here wish to question the selection of the small sample suite and the robustness of the volumecorrelation conclusions compared to using a more comprehensive AVF dataset available in the literature. Rare earth element (REE) data are available for 14/53 eruptions with major element data available for 41/53 eruptions in the AVF (Heming and Barnet, 1986; Huang et al., 1997; McGee et al., 2013; Needham et al., 2011; Searle, 1960; Smith et al., 2008). From this suite, we show here that P2O5 and REE concentrations correlate extremely well (R = 0.96; Figs. 1A and 1B) and that P2O5 is thus a robust proxy for REE abundances for the eruptions without trace element data. When the most primitive sample for each eruption is plotted (highest MgO content, which is >10wt% for most eruptions), any fractional crystallization effect is minimized. Using this proxy, the conclusions of McGee et al. (2015) can be tested on 41/53 eruptions in the field, rather than five eruptions selected based on unclear criteria.