HR: 15:05h
AN: V42F-06 INVITED     [PDF]
TI: Formation and Segregation of Melts in the Continental Crust: Time Scales From Trace Element Modeling of Rocks From the Sikkim Himalaya, India
AU: * Chakraborty, S
EM: Sumit.Chakraborty@ruhr-uni-bochum.de
AF: Institut fuer Geologie, Mineralogie und Geophysik, Ruhr-Universitaet Bochum, Bochum, NRW D-44780 Germany
AU: Neogi, S
AF: Institut fuer Geologie, Mineralogie und Geophysik, Ruhr-Universitaet Bochum, Bochum, NRW D-44780 Germany
AU: Bolton, E W
AF: Department of Geology and Geophysics, Yale University, Kline Geology Laboratory, P.O. Box 208109, New Haven, CT 06520-8109 United States
AB: It is difficult to quantify the distribution of trace elements during crustal melting because equilibrium partitioning of these elements is often not attained. But modeling the distribution is crucial for understanding intracrustal differentiation (into upper, middle, and lower crust) as well as the nature and timescale of melt formation and segregation in the crust. To address this problem, we have developed a numerical model to simulate the evolution of trace elements in a system undergoing melting. The model accounts for equilibrium partitioning as well as kinetic processes such as dissolution, crystal growth, and diffusion along arbitrary temperature - time paths. In this project we have field tested the model on rocks from the classic inverted metamorphic sequence in the Sikkim Himalaya. The geological and petrological history of the area has been well constrained in a parallel project. Our study material consists of a sequence of prograde metamorphic rocks with rather similar bulk compositions that have been subjected to progressively increasing pressures and temperatures to ultimately undergo melting by well defined mica dehydration reactions. Using measured protolith and reactant mineral compositions, inferred balanced reactions, petrologically constrained peak temperatures, and known partition and diffusion coefficients in our model we reproduce observed mineral and melt compositions for any one element (e.g. Sr) by varying the thermal history. Once a good fit was observed, the same thermal history was then used to predict other trace element distributions (e.g. Rb, Ba) with a high degree of success. In our simulations so far we have ignored minor phases (e.g. apatite) and back reactions during cooling. Consequently, fits for elements that are known to reside primarily in minor phases (e.g. La) are poor, indicating that our model does indeed capture the essence of processes taking place in nature. Protolith bulk composition, stoichiometry of melting reaction, and partition and diffusion coefficients are found to be the key parameters controlling the distribution of trace elements. Our results so far indicate that the time scales of melting and melt segregation in such a continental collisional setting where dehydration melting reactions are taking place are on the order of 100,000 years.
DE: 1020 Composition of the crust
DE: 3660 Metamorphic petrology
SC: Volcanology, Geochemistry, Petrology [V]
MN: 2003 Fall Meeting