Computer Modeling of Major and Trace Element Variations of Hole 504B Diabase and Basalt

Four computer modeling techniques were used to examine the intersample variation in Hole 504B basalt and diabase: (1) a liquid-line-of-descent calculation (the CHAOS program, Nielsen [1985,1990]); (2) Pearce element ratio analysis (Pearce, 1968);(3) a least-squares-fit parent-daughter calculation (Bryan et al., 1969); and (4) correlation coefficient pattern analysis (the DARWIN program, Cox and Clifford [1982]).The first two techniques require few input assumptions, and can be used to identify what phases are fractionating, the compositions of those phases, and the relative proportions of those phases.The other two techniques require a considerable amount of input assumptions, and are primarily useful for testing petrologic models.The liquid-line-of-descent calculation suggests that fractional crystallization of plagioclase An 744 , augite Wo 41 7 En 48 6 Fs 9 7 , and olivine Fo g5 3 from an initial mid-ocean-ridge basalt parent could explain the range of compositions observed in the 504B data set.The complementary gabbro would have the cumulus mineral assemblage 57.5% plagioclase, 23.3% augite, and 19.3% olivine plus an unknown amount of intercumulus liquid.Pearce element ratio analysis suggests that fractional crystallization of plagioclase An 79 _ 86 , augite Mg# g2 _ 83 , and olivine Fo 78 _ 79 could explain the range of compositions observed in the 504B data set.The complementary gabbro would have a cumulus mineral assemblage of 62% to 67% plagioclase, 23% to 27% augite, and 8% to 12% olivine plus an unknown amount of intercumulus liquid.All four computer models are consistent with a differentiation scheme for 504B magmas in which an initial high Mg# parent fractionates plagioclase, augite, and olivine to produce a series of daughter magmas with lower Mg#s.None of the models require significant amounts of assimilation or more than one parent magma composition to explain the variation observed in the 504B data.Fraction of the observed 1726 m of basalt and diabase at Site 504B would require a minimum thickness of 573 m of gabbro if we assume 0% trapped liquid, or 1042 m of gabbro if we assume 45% trapped liquid.The lower crustal section at Hole 504B is probably thicker and more primitive than that calculated.