Time scales of magma generation, differentiation, and storage: Constraints from uranium-238-thorium-230-radium-226 disequilibria

Time scales of magma generation and differentiation are important variables in understanding magmatic systems, but are difficult to measure. I present a new approach to determining minimum magmatic residence times by dating phenocrysts using 226Ra-230Th disequilibria. When fractionation of Ra from Ba during crystal growth is considered along with the effects of inclusions in the minerals, data for the early 1955 east rift eruption of Kilauea Volcano are consistent with crystallization at a mean age of $1000+300-400$ y. Unless more than ∼30% of the crystals are xenocrystic the minimum magmatic residence time must be ∼550 y, considerably longer than most previous estimates of storage time at Kilauea. By applying the same method to <∼2 ka Mount St. Helens lavas, data for all but one of the samples are consistent with closed-system crystallization, unlike the previous interpretation that the data require a late-stage addition of 226Ra to the magma after crystal formation. Incorporation of old xenocrystic material combined with additional crystal growth during ascent could provide a mechanism for decoupling of the 226Ra-230Th and 230 Th-238U systems. If the model crystallization ages largely reflect closed-system crystallization, the implied magma storage times would be ∼2–3 ka. 238U-230Th- 226Ra disequilibria measured in minerals separated from a tholeiitic cumulate xenolith erupted at Hualalai Volcano, Hawaii are inconsistent with fractionation during crystallization during the tholeiitic shield-building stage. Possible mechanisms of recent Th-U-Ra fractionation include crystallization of tholeiitic magma, or interaction with a hydrous/CO₂-rich fluid.; I present 238U-230Th data for trachyandesites from the Ashikol Basin of northwestern Tibet. The large excesses of 230Th compared to 238U require garnet in the source of the melts together with slow melting rates and/or very small residual porosities during melting. The low (230Th)/(232Th) ratios and high 208Pb/206Pb ratios measured in Ashikol Basin lavas correspond to high present-day and time-integrated 232 Th/238U ratios in the source(s) of the Ashikol magmas, consistent with Nd model ages for source enrichment of ∼1.2–3.5 Ga. The 230Th-238U disequilibria are more consistent with enrichment of the mantle by silicate melts rather than hydrous or carbonic fluids, and are not consistent with melt generation by intracontinental subduction. Melts could potentially be produced during shear heating of the uppermost lithospheric mantle or by convective removal of the lower lithosphere and heating of the remaining lithospheric mantle.