Felsic magmas from Mt. Baker in the northern Cascade arc: origin and role in andesite production

Dacitic magmas in volcanic arcs play a critical role in the growth and development of felsic continental crust through mixing to form andesite, or to a lesser extent, by directly adding new crustal material through fractionation of mantle derived basalts. Though dacitic erupted lavas are scarce on Mt. Baker, this study discusses their importance in subsurface processes such as mixing with more mafic magmas, and their potential to add directly to the volume of continental crust. A comprehensive data set (including major, trace, and rare earth element abundances, as well as petrography and mineral chemistry) reveals that the most Sirich, Mg-poor dacitic compositions analyzed in this study (dacite of Mazama Lake) can be modeled as liquids derived by crystal fractionation from Mt. Baker high-Mg andesites. These Si-rich compositions are in turn back-mixed with mafic magmas to produce more Sipoor dacites (dacite of Cougar Divide) and andesites (andesite of Mazama Lake). The origin of one enigmatic hornblende-bearing dacite unit (dacite of Nooksack Falls) is unconstrained. None of the dacitic units have geochemical signatures that suggest an origin by melting of a garnet-bearing source such as the subducting slab or the lower crust. The dacite of Mazama Lake (plagioclase, clinopyroxene, orthopyroxene, Fe-Ti oxides) represents a near end-member fractionated composition with only minor contamination from xenocrystic material. Mineral populations commonly lack disequilibrium textures, and exhibit normal zoning. Plagioclase and pyroxene chemistry suggests the majority of the crystal population is original to the dacite of Mazama Lake. Sparse resorbed olivine grains (<1% total crystal population) and weak reverse zoning in some plagioclase and pyroxene grains indicates a minor addition of xenocrystic material. The majority of the Mazama Lake compositions can be reproduced after 44% fractionation (55% remaining liquid) of a high-Mg andesite (the andesite of Glacier Creek), with fractionating phases of 69% plagioclase, 16% orthopyroxene, 11% clinopyroxene, 3% ilmenite, and 1% apatite. Excellent fits of major elements, most trace elements are provided by this model. The dacite of Cougar Divide (plagioclase, clinopyroxene, orthopyroxene, Fe-Ti oxides, olivine) and the andesite of Mazama Lake (plagioclase, clinopyroxene, orthopyroxene, Fe-Ti oxides, olivine) are more Si-poor, and exhibit evidence for magma mixing. The Cougar Divide unit exhibits mingling textures in hand sample and both Si-poor units exhibit mixing textures in thin section, such as calcic normal and sodic reverse zoned plagioclase populations and pyroxene grains with abrupt Mg-rich rims. This suggests that their primary geochemical characteristics come from mixing between more mafic and more felsic magmas. The dacite of Mazama Lake can be used to reasonably reproduce compositions observed in the mixed magmas. Mixing between the high-Mg andesite of Glacier Creek and dacite of Mazama Lake can reproduce an average major and trace element composition from the Cougar Divide unit in mixing proportions of ~60% andesite and ~40% dacite. Major and trace element compositions from the andesite of Mazama Lake can be reproduced by mixing ~30% the high-Mg basaltic andesite Tarn Plateau (a less fractionated parent magma of the andesite of Glacier Creek) and ~70% Mazama Lake dacite. The dacite of Nooksack Falls (plagioclase, hornblende, clinopyroxene, orthopyroxene, Fe-Ti oxides) appears to represent a near-endmember composition, but cannot be reproduced by fractional crystallization of any known parental composition at Mt. Baker. A distinct set of minerals with compositions expected from a basaltic source (such as calcic plagioclase grains, and Mg-rich clinopyroxene grains with high Cr concentrations) suggests the dacite of Nooksack Falls acquired some xenocrystic material. However, removal of this contamination does not permit a fractionation origin from known mafic compositions. One possibility is that the dacite of Nooksack Falls was derived from more mafic magmas that are not currently observed or erupted. These dacites are unlikely to be crustal melts given their high H2O contents. Ultimately, these hypotheses cannot be reconciled without isotopic analysis. The role of dacitic magmas at Mt. Baker is clear; (1) they have the potential to directly contribute to the continental crust through fractionation, and (2) they have a role in mixing, in which andesitic compositions (a common composition at arcs worldwide) are formed.