Development of a large volume of eruptible mush in the upper Wooley Creek batholith, Klamath Mountains, California: evidence from bulk rock, mineral analyses and textural observations

The modalities of development of large volumes of mush in the middle to upper crust capable of erupting have been debated over the past few years. The existence of crystal-rich ignimbrites in the volcanic record indicate that eruptive products do not necessarily correspond to evacuation of the residual magma but that the mush itself can be drained during eruptive events. In this study we present a plutonic example of a large magma batch that evolved by fractional crystallization at a hundred km3 scale: the upper zone of the Wooley Creek batholith (WCb). The WCb is an intrusive complex emplaced over less than 3 m.y. (Kevin Chamberlain, personal communication). The upper zone grades upward from quartz diorite (53 wt% SiO2) to granite (70 wt% SiO2). Hornblende from the central and upper zone have rare earth element patterns that are parallel to one another and with REE concentrations and negative Eu anomalies that decrease from core to rim. The similarities of hornblende REE patterns throughout both the central and upper zones of the system (160 km2 of exposed area) suggest that hornblende crystallized from a magma batch of fairly homogeneous composition. Thus, upward changes in bulk composition between rocks at the bottom and the top of this unit result from varying mineral proportions, with more subhedral plagioclase and hornblende at the bottom and more anhedral to euhedral quartz and interstitial to poikilitic K-feldspar at the top. Two possible explanations are considered: 1) more felsic batches of magma were emplaced at the top of the system and more mafic ones were restricted to the bottom, 2) the upper zone acquired its upward compositional zoning through melt percolation, with the less dense felsic melt ponding at the roof of the system. In the first case, the similarity of hornblende REE patterns throughout the upper zone cannot be explained. Therefore, we favor the second explanation, which is also supported by the lack of sharp contacts in the upper zone. Individual magma batches in the central zone contain hornblende of similar composition as in the upper zone and are interpreted as a preserved part of the feeder system of the latter. Therefore the magma in both the central and upper WCb was already fairly homogeneous when it arrived at the level of emplacement. Dacitic to rhyodacitic roof dikes with30-40% phenocrysts of hornblende and plagioclase with compositions similar to those found in the central and upper zones indicate that the mush was once eruptible. The presence of quartz phenocrysts, which are only found in the uppermost portion of the upper zone, show that 'eruption' occurred after the development of the broad zoning of the upper zone and after more evolved melt had collected at the top of an underlying mush. This study introduces new tools to study magmatic reservoir evolution. The combination of bulk rock and mineral data allows assessment of the extent of mineral-melt separation and identification of the composition of a potential parental magma(s). These data can ideally be used to delimit the size of magma batches and constrain the scale of their chemical/physical connectivity.