Cyclicity of multistage anatexis of deeply subducted continental crust during the North Qaidam orogeny: Tracing the source, timescale, and evolution of pulsed melts

The most significant mass transfer processes at a convergent plate boundary are tectonic accretion and fluids/melts released from sites of generation to sites of accumulation. However, some crucial questions remain with regards to the source, timescale, and evolution of such anatectic processes, for example, single or multistage anatexis of deeply subducted continental crust. To better understand the processes involved in anatexis, we have quantified the timescale and nature of formation and evolution processes of multistage felsic veins within retrograde eclogite using zircon, monazite, and xenotime geochemistry and geochronology, whole-rock composition, and Sr-Nd-Hf isotope analysis from the Lüliangshan, North Qaidam orogen. The U–Pb dating of coexisting zircon, monazite, and xenotime gives three groups of ages at ca. 441 to 435, ca. 425, and ca. 413 to 409 Ma, respectively, which record at least three episodes of pulsed melts. The first stage of zircon formation is characterized by not only the absence of oscillatory zoning of different zircons in cathodoluminescence images but also some of them with flat HREE and without Eu anomalies, indicating they may form in anatectic melts under eclogite-facies conditions. The second and third phases of melts may have occurred under granulite- and amphibole-facies conditions during exhumation. Furthermore, two classes of felsic veins within the eclogite show wide variations of whole-rock composition, (87Sr/86Sr)i, εNd (t), and εHf (t) values, which demonstrate that they were derived from fluid-present dehydration partial melting of different proportions of subducted gneiss and eclogite during different periods. These findings show that melts have systematic differences in chemical and isotopic signatures resulting from lithological diversity and depth of partial melting. Thus, small-scale melts released from the source can excellently explain the variability in whole-rock composition, accessory mineral growth zoning, and prominent isotope variability in syn-collisional heterogeneous granites. An additional implication is that as these melts escape their adjacent area of formation, they migrate and mix along channelized melt pathways, resulting in melt-rock and crust-mantle interaction and the triggering of syn- and post-collisional magmatism.