Attribution: You must attribute the work in the manner specified by the author or licensor ( but no in any way that suggests that they endorse you or your use of the work).Noncommercial ‒ you may not use this work for commercial purpose.No Derivative works ‒ You may not alter, transform, or build upon this work.Sharing ‒ Individual scientists are hereby granted permission, without fees or further requests to GSA, to use a single figure, a single table, and/or a brief paragraph of text in other subsequent works and to make unlimited photo copies of items in this journal for noncommercial use in classrooms to further education and science.

We welcome the comments by Berlo et al. (hereafter BTBH) and thank them for the opportunity to further discuss lower crustal melting processes in arc settings. Our intention in Dufek and Cooper (2005) was not to explain the U-series disequilibria in all arc magmas through a single model. Instead, because the composition of many arc magmas is the result of polybaric processes, we explored whether U-series disequilibria measured at the surface may reflect some contribution from lower crustal melting where the crust is >30 km thick. We disagree with BTBH that the form of the melting model is insignificant, considering field and experimental evidence for melt segregation at low melt fractions. Here we discuss the three comments made by BTBH and show that significant 226Ra-230Th dis-equilibria can be produced during continuous, incongruent melting, even using the partition coefficients favored by BTBH.

It is well established that models that explicitly consider the duration of melting with low residual porosity (~0.1%) are required to explain observed U-series disequilibria for mid-oceanic-ridge basalts (e.g., Lundstrom, 2003). However, BTBH (2004; and in their Comment) argue that partial melts of lower crust are unlikely to segregate at melt fractions below 20%–625%. As we discussed in our paper, use of a time-dependent melting model is consistent with experimental and field evidence of hydrous melt segregation at low melt fraction (≤10%), often during synchronous deformation (Brown, 2005). Furthermore, while we did advocate small critical porosities relative to batch melting, this is significantly different from stating that we examined only small total melt fractions (i.e., limited progress in the dehydration reaction); in fact, our Figure 1 shows (226Ra)/(230Th) > 3 at up to 0.25 extracted melt fraction, provided that the melting rate is lower than ~102 kg/m3/yr. We agree that zircon, if present in the residue or fractionated from the ascending liquid, will likely act to increase (226Ra)/(230Th). However, this process would only produce excessively large (226Ra)/(230Th) ratios at the surface if 1) lavas at the surface sample these melts directly, without mixing with other melts, and 2) the magma does not stall for any appreciable time in transit. We suggested in our paper that erupted lavas may reflect mixtures of mantle-derived magmas with lower-crustal melts, and the larger Ra excesses predicted in the presence of residual accessory phases would relax the necessary ascent rates and would increase the range of applicability of this model.

BTBH's second point relates to the partition coefficients used in our model. We fully agree that pressure, temperature, and composition must be taken into account when modeling melting processes. We used partition coefficients from the literature that reported P-T-X most like the phases in the Wolf and Wyllie (1994) amphibolite melting experiments that we used as a model melting reaction. We were not able to evaluate the appropriateness of the partition coefficients used by Berlo et al. (2004) because data on the experimental compositions were not available. Here we examine the sensitivity of the continuous melting model to the choice of partition coefficients by performing calculations using those suggested by Berlo et al. (2004), and at similarly high melt fraction (0.35) as their batch melting calculations (Fig. 1). Batch melting is most closely approximated by very fast melting, which produces almost no 226Ra excess. However, at slower and more geologically relevant melting rates, 226Ra excesses develop that are consistent with many continental arc magmas, even with the partition coefficients favored by BTBH. Conduction and even comparatively fast processes such as gas sparging are unlikely to generate average melting rates greater than 10−3 kg/m3/yr (Bachmann and Bergantz, 2006; Dufek and Bergantz, 2005).

We proposed using the trace element signature of the incongruent melting process to identify magmas where the U-series disequilibria have been affected by lower-crustal melting. We agree with BTBH, that based on trace-element criteria, many island arc basalts erupted through a thin (< 30 km) crust are unlikely to have large crustal contributions to their U-series disequilibria. However, there are many settings (especially in continental arcs) where this process remains viable. We reiterate that both the choice of a primitive island arc basalt starting composition and the 50:50 mixing calculation of crustal and mantle melts presented in our paper were intended to be illustrative rather than definitive. Further, provided that melting is slow enough that in-growth effects are important, 226Ra excess can be preserved to high melt fraction over a range of Sr/Y and La/Yb.

In summary, we agree with BTBH that rapid batch melting in the crust is unlikely to produce significant 226Ra-230Th disequilibria. However, field and experimental evidence, combined with numerical models of melting rates within the crust, suggest that melt extraction at relatively small porosities over an extended period of time is likely. Time-dependent melting in the lower crust, especially in the presence of residual garnet, can produce Ra excesses and trace element signatures similar to those observed in a number of arc settings, and this needs to be considered when quantifying magma ascent and storage time scales in these settings.