Episodic trace element and isotopic variations in historic Mauna Loa lavas: Implications for magma and plume dynamics.

J. M. Rhodes, Department of Geosciences, UMass, Amherst

S. R. Hart, Department of Geology and Geophysics, WHOI

Abstract. Over the past 152 years, Mauna Loa volcano has erupted lavas with almost constant major element, and compatible and moderately incompatible trace element abundances at a given MgO content. This uniformity is attributed to continuing replenishment of a shallow magma reservoir. In contrast, incompatible element abundances and ratios, together with Sr, Nd and Pb isotopic ratios, vary systematically with time. The greatest rate of change occurred at a time (1843-1887) when Mauna Loa was vigorously active with high eruption rates, presumably a consequence of a high magma supply rate. Detailed analysis confirms what is evident from the isotopic data: that this open-system magmatism requires two or more parental magmas. One has the compositional attributes of lavas erupted in 1843, the other the characteristics of lavas erupted at the summit early in 1880. All other historical lavas can be considered as mixtures of these two end-members, modified by contemporaneous eruption and olivine crystallization. Both parental magmas have Sr, Pb and Nd isotopic ratios typical of magmas in the Hawaiian tholeiitic array, and intermediate between those of Kilauea and Koolau lavas, the end-members of the array. The 1843 parental magma has incompatible element ratios that are similar to, and overlap with the Koolau and Kilauea data. The inferred 1880 parental magma, however, is more depleted than the 1843 parental magma (and most other Hawaiian lavas), and is also isotopically closer to the Kilauea end-member of the tholeiitic array. The origin of these parental magmas is discussed in terms of melting within a radially heterogeneous plume in which the heterogeneity may develop at the source or through subsequent mantle entrainment. Two models are explored, both depend on the location of Mauna Loa at, or close to the plume margin. In the simplest case the parental magmas are produced by progressive melting of the heterogeneous outer plume. The second model is more dynamic, involving melt production and re-equilibration in a diverging, or inclined, plume.

Al2O3 versus MgO for individual samples (188) of historical Mauna Loa lavas. Only 5 of the 67 samples analyzed from the 1984 eruption [Rhodes, 1988] are included in this plot to avoid domination by the very uniform compositions that characterize this eruption. The regression line fitted to the data (R-squared=0.99) defines an olivine-control trend.

MgO-normalized Sr abundances versus time for historical Mauna Loa lavas. The data have been normalized to a constant MgO content of 7.0% through addition or subtraction of olivine (Fo87). The Sr data are averages of each eruption or sub-unit and the vertical bars reflect the standard deviation (2 sigma) for each unit or sub-unit. These bars are absent when the standard deviation (2 sigma) is less than the symbol size. Olivine-controlled lavas are shown as open squares and picritic and high MgO (>10%) lavas as solid diamonds.