Quantitative vesicle analyses and total CO2 reconstruction in mid-ocean ridge basalts

Abstract

Vesicle textures in submarine lavas have been used to calculate total (pre-eruption) volatile concentrations in mid-ocean ridge basalts (MORB), which provide constraints on upper mantle volatile concentrations and global mid-ocean ridge CO2 flux. In this study, we evaluate vesicle size distributions (VSDs) and volatile concentrations in a suite of 20 MORB samples that span the range of vesicularities and vesicle number densities observed in MORB globally. We provide recommended best practices for quantifying vesicularity, vesicle number densities, and VSDs based on synthetic vesicle populations and comparisons between traditional 2D estimates and X-ray computed micro-tomography results. For 2D measurements, we recommend analyzing multiple polished fragments with a cumulative area >100 times the area of the largest observed vesicle and including >200 vesicles in stereological VSD reconstructions. For 3D measurements, we recommend analyzing sample volumes >0.01 cm3 at resolutions <2.0 μm/pixel for low vesicularity MORB (i.e., <4 vol.%) and sample volumes >0.1 cm3 with resolutions <5 μm/pixel for higher vesicularity samples. Our validation of vesicularity measurements allows reconstructions of total CO2 concentrations in MORB using dissolved volatile concentrations, vesicularities, and equations of state. We assess approaches for estimating the exsolved CO2 concentration in MORB vesicles and find that CO2(g) density is ~40% lower than previously suggested, likely due to melt contraction during quenching. Based on these results, we recommend using sample eruption pressures, magmatic temperatures, and an equation of state that accounts for non-ideality at high temperatures to calculate exsolved CO2 concentrations when independent constraints from Raman spectroscopy or laser ablation are unavailable. Our results suggest that some previous studies may have overestimated MORB CO2 concentrations by as much as 50%, with the greatest differences in samples with the highest vesicularities. These new results imply lower CO2/Ba of undegassed, enriched-MORB and lower integrated global ridge CO2 flux than previously inferred.