HR: 0800h
AN: V21A-0595    [Abstracts]
TI: Consequences of adiabatic decompression melting on magmatic channeling instabilities
AU: * Fang, Y
EM: yf91@columbia.edu
AF: Dept. Applied Physics/Applied Math, Columbia University, 500 w 120th st, New York, NY 10027 United States
AU: Spiegelman, M
EM: mspieg@ldeo.columbia.edu
AF: Dept. Applied Physics/Applied Math, Columbia University, 500 w 120th st, New York, NY 10027 United States
AU: Spiegelman, M
EM: mspieg@ldeo.columbia.edu
AF: Lamont-Doherty Earth Observatory, Rt 9W, Palisades, NY 10964 United States
AB: In many partially molten regions of the earth (e.g. beneath mid-ocean ridges), the solid mantle upwells and melts adiabatically due to decompression. Observations of replacive dunites suggest that the melts react with the upwelling solid to form dissolution ``channels'' which may have widths and spacings of 0.1--100 m or larger (e.g. Braun and Kelemen, G-cubed 2002). Previous work (Spiegelman et. al, 2001) has explored reactive channeling instabilities in a static framework. Here we extend the problem to consider how including solid upwelling may change these results. We explore a simplified 2-D column model for highly reactive systems that includes mantle upwelling and both adiabatic and reactive melting. We present both the linear stability analysis for this solution as well as numerical results for fully non-linear channel formation. Linear analysis provides information on the growth rate and spacing of the channels as a function of two principal parameters, $w_0/W_0$, the melt velocity relative to the solid upwelling velocity and $h/\delta$, the height of the system in compaction lengths. The compaction length $\delta$ is controlled principally by the solid viscosity. In particular, we show that there exists a critical value of both parameters beyond which no channels grow. Using the linear analysis, we produce a phase diagram of numerical solutions of the full non-linear problem and show that they agree well with the predictions of the linear analysis but saturate to a steady state. We then derive an approximate analytic solution for the steady-state that allows us to estimate the strength of channeling as well as the depth and width of the channels as a function of the fundamental parameters. Both the linear analysis and the numerical model suggest that this instability could be a viable mechanism for flow localization in the earth. Moreover, these results suggest that, all else being equal, this mechanism is more likely at fast spreading ridges than slow spreading ridges.
DE: 8434 Magma migration
DE: 5114 Permeability and porosity
DE: 3035 Midocean ridge processes
DE: 3230 Numerical solutions
SC: Volcanology, Geochemistry, Petrology [V]
MN: 2004 AGU Fall Meeting