ill.; maps; Thesis (M.S.)--University of Oregon, 1985; Includes vita and abstract; Includes bibliographical references (leaves 154-158)
i i i An Abstract of the Thesis of Elizabeth M. Prueher for the degree of Master of Science in the Department of Geology to be taken March 1985 Title: THE GEOLOGY AND GEOCHEMISTRY OF THIRTEEN CINDER CONES AT CRATER LAKE NATIONAL PARK, OREGON Major and trace-element variations exhibit an increase in excluded elements and a decrease in included elements as differentiation increases, indicative of crystal fractionation of olivine, clinopyroxene, orthopyroxene, and p 1 a g i o c 1 a s e • L R E E patterns for the rocks are irregular. First, basalt is enriched in LREE relative to other members of the suite. Second, there is a repeated pattern of three pairs of basalt and andesite; with each successive basalt enriched in LREE relative to the preceeding andesite. From consideration of the REE data it is apparent that the cinder cones were derived from more than one partial melt • Compositional variations in the magmas of the cinder cones support the conclusion that they were generated by more than one partial melting event. Partial melting of a iv mantle source region with the composition of peridotite could have produced primitive basaltic magmas. Subsequent mixing and fractional crystallizatio n produced the more differentiated basaltic to andesitic magmas. 129 Summary Petrographic and geochemical evidence suggests that crystal fractionation of plagioclase, magnetite, olivine, a u g i t e , and hypersthene was involved in the genesis of the magmas of the Crater Lake cinder cones. However, crystal fractionation was not the only mechanism. The following observations need to be considered when formulating a model for the origin of the cones. First, the irregular pattern of basalts and andesites observed for the LREE, and the LREE enrichment in basalt relative to other members of the suite. Second, the satisfactory solutions obtained for mass-balance calculations when only major-elements are considered, but the less satisfactory solutions obtained when trace elements are considered, and the observed increase in included elements and the decrease in excluded elements in the daughter products of some of the calculations. Third, the appearance of two trends on many of the variation diagrams, and the observed scatter on the plot of A1203. Fourth, the magmas are not primary. The contents of Ni, Cr, and MgO are too low for the magmas to be in equilibrium with mantle values. From consideration of the LREE data it is apparent that the Crater Lake cinder cones were derived from more than one partial melt, probably a distinct melt for each of no the groups observed. The data require at least four, distinct partial melts, which underwent various degress of mixing and fractional crystallization to produce the Lavas of the cinder cones. If this is true, then the trend is correct for fractional crystallization. In other words, for each pair, the andesite is enriched in LREE relative to the basa1tic andesite as is expected. A separate partial melt is implied for the Williams Crater basalt. Magma mixing and subsequent fractional crystallization may explain the observed patterns in the variation diagrams, and the discrepancies in the mass-balance calculations. The fact that the magmas are not primary indicates that the partial melts underwent differentiation to produce cinder cone magmas. The hypothesis of more than one partial melt would explain the presence of the two trends observed on some variation diagrams. Distinct trends for neither partial melting or fractional crystallization appear on plots of highly excluded versus highly included elements, or on H/M versus H diagrams. Thus, both processes were probably involved in the genesis of the cinder cone lavas. The increase in included elements from parent to daughter, particularly for Ni, Cr, and Sr, can be explained by the addition of olivine, clinopyroxene, and plagioclase, possibly during magma mixing. Plagioclase contamination is 131 further indicated by the large weighted residuals for Sr and Ba, the scatter on the Al2O3 versus SiO2 diagram, the positive Eu anomaly, and the presence of plagioclase macrophenocrysts in the samples.
i i i An Abstract of the Thesis of Elizabeth M. Prueher for the degree of Master of Science in the Department of Geology to be taken March 1985 Title: THE GEOLOGY AND GEOCHEMISTRY OF THIRTEEN CINDER CONES AT CRATER LAKE NATIONAL PARK, OREGON Major and trace-element variations exhibit an increase in excluded elements and a decrease in included elements as differentiation increases, indicative of crystal fractionation of olivine, clinopyroxene, orthopyroxene, and p 1 a g i o c 1 a s e • L R E E patterns for the rocks are irregular. First, basalt is enriched in LREE relative to other members of the suite. Second, there is a repeated pattern of three pairs of basalt and andesite; with each successive basalt enriched in LREE relative to the preceeding andesite. From consideration of the REE data it is apparent that the cinder cones were derived from more than one partial melt • Compositional variations in the magmas of the cinder cones support the conclusion that they were generated by more than one partial melting event. Partial melting of a iv mantle source region with the composition of peridotite could have produced primitive basaltic magmas. Subsequent mixing and fractional crystallizatio n produced the more differentiated basaltic to andesitic magmas. 129 Summary Petrographic and geochemical evidence suggests that crystal fractionation of plagioclase, magnetite, olivine, a u g i t e , and hypersthene was involved in the genesis of the magmas of the Crater Lake cinder cones. However, crystal fractionation was not the only mechanism. The following observations need to be considered when formulating a model for the origin of the cones. First, the irregular pattern of basalts and andesites observed for the LREE, and the LREE enrichment in basalt relative to other members of the suite. Second, the satisfactory solutions obtained for mass-balance calculations when only major-elements are considered, but the less satisfactory solutions obtained when trace elements are considered, and the observed increase in included elements and the decrease in excluded elements in the daughter products of some of the calculations. Third, the appearance of two trends on many of the variation diagrams, and the observed scatter on the plot of A1203. Fourth, the magmas are not primary. The contents of Ni, Cr, and MgO are too low for the magmas to be in equilibrium with mantle values. From consideration of the LREE data it is apparent that the Crater Lake cinder cones were derived from more than one partial melt, probably a distinct melt for each of no the groups observed. The data require at least four, distinct partial melts, which underwent various degress of mixing and fractional crystallization to produce the Lavas of the cinder cones. If this is true, then the trend is correct for fractional crystallization. In other words, for each pair, the andesite is enriched in LREE relative to the basa1tic andesite as is expected. A separate partial melt is implied for the Williams Crater basalt. Magma mixing and subsequent fractional crystallization may explain the observed patterns in the variation diagrams, and the discrepancies in the mass-balance calculations. The fact that the magmas are not primary indicates that the partial melts underwent differentiation to produce cinder cone magmas. The hypothesis of more than one partial melt would explain the presence of the two trends observed on some variation diagrams. Distinct trends for neither partial melting or fractional crystallization appear on plots of highly excluded versus highly included elements, or on H/M versus H diagrams. Thus, both processes were probably involved in the genesis of the cinder cone lavas. The increase in included elements from parent to daughter, particularly for Ni, Cr, and Sr, can be explained by the addition of olivine, clinopyroxene, and plagioclase, possibly during magma mixing. Plagioclase contamination is 131 further indicated by the large weighted residuals for Sr and Ba, the scatter on the Al2O3 versus SiO2 diagram, the positive Eu anomaly, and the presence of plagioclase macrophenocrysts in the samples.