资源环境科技发展态势分析平台

Magnetotelluric and seismological studies suggested the presence of partial melts in the middle to lower Himalaya-Tibetan crust. However, the melt fractions inferred by previous work were based on presumed electrical conductivity of melts. We performed measurements on the electrical conductivity of peraluminous granitic melts with 0.16-8.4wt % H2O (the expected compositions in the Tibetan crust) at 600-1,300 degrees C and 0.5-1.0 GPa. Peraluminous melt exhibits lower electrical conductivity than peralkaline melt at dry condition, but this difference diminishes at H2O>2 wt %. With our data, the observed electrical anomalies in the Tibetan crust could be explained by 2-33 vol % of peraluminous granitic melts with H2O>6wt %. Possible reasons for our inferred melt fractions being higher than seismological constraints include the following: (1) The real melts are more Na and H2O rich, (2) the effect of melt reducing seismic velocities was overestimated, and (3) the anomalies at some locations are due to fluids.

Plain Language Summary Whether there are molten zones present in the Tibetan crust is a focus of geophysical and petrological research. Previous interpretations of MT data to infer melt fractions are often based on presumed electrical conductivity values of partial melts. These felsic melts are believed to derive from metapelites and have peraluminous composition, but previous experimental data of electrical conductivity are only for metaluminous and peralkaline melts. In this contribution, we carry out electrical conductivity measurements on anhydrous and hydrous peraluminous granitic melts with 0.16-8.4wt % of H2O at 600-1,300 degrees C and 0.5-1.0 GPa. We find that the electrical conductivity of peraluminous melt is lower than that of peralkaline melt under dry condition, but their difference quickly diminishes at H2O greater than 2 wt %. Based on our experimental results, the melt fractions for various regions in the Himalaya-Tibetan crust are inferred and compared with seismological constraints.