Abstract. The interaction between mineral reaction and mass transport in a rock can lead to reaction front instability and the development of channel-like voids. This phenomenon is studied with a two-dimensional model accounting for the nonlinear feedback between flow, reaction, and matrix porosity-permeability evolution. In our model we calculate the flow field in both the porous medium and the reaction-induced voids, using the Brinkman equation. While a linear analysis cannot determine the length scale of the channels which can develop in a typical geological system, our simulations indicate that the channel size is actually unique and well characterized. While the onset of instability is favored at a preexisting heterogeneity, the channel growth and orientation is governed by the global flow pattern, even in an initially heterogeneous system.

No abstract

S U M M A R YFan-type sandstone deposits within a sedimentary basin generally present an anisotropic structure with maximum permeability in the paleo-horizontal direction. Modelling of 3-D circulations in inclined isotropic porous layers suggests the following patterns of fluid motion: for subcritical Rayleigh numbers Ra the fluid circulates along the steepest slope: this is the well known basic flow; at low inclinations, the supercritical convection consists in polyhedric cells and beyond a critical slope of 32", longitudinal coils (i.e. with axes parallel to the steepest slope) are the preferential mode of convection. We extend these results to anisotropic media, with anisotropies up to 1000, by means of 3-D numerical modelling. From a systematic linear study near the critical Rayleigh number Ra, it turns out that the critical slope lowers to 12.7", 4.2", 1.4" for anisotropy ratios of 10, 100, 1000 respectively. A complete non-linear study at twice the critical Rayleigh number and for a slope of 2" presents only longitudinal rolls; their width is lower than four when the anisotropy does not exceed 1000. Moreover we observe that the circulation of the fluid particles inside the coil follows a loop stretched in the slope direction with a length of six and 60 times the thickness of the layer, for anisotropies of 100 and 1000 respectively. We suggest that this offers a possibility of very far transport of minerals in the direction of the slope.

We present the results of a numerical simulation study of the thenno-convective circulation in a sedimentary pile, in the distal part of the Bengal Fan. The occurrence of faults and the variability of surface heat flow data led previous workers to suggest hydrothermal circulation in this area. According to the purpose of Ocean Drilling Program survey Leg 116, data were collected which are adequate to constrain a model for the comprehension and the quantification of the relationship between ocean basin heat flow and pore fluid circulation. In this paper the sedimentary pile is modeled as an anisotropic and inclined porous layer interrupted by faults with a periodic distribution. An important observation the model tries to reproduce is the shift of approximately 2-3 km between the surface expression of the heat flow maximum and the fault zone near site 718. The influence of permeability anisotropy and of permeability within those fault zones is analyzed. When the maximum vertical permeability in the fault zones is 5 times greater than the vertical permeability of the porous block, the circulation consists of longitudinal rolls (i.e., rolls with their axis oriented downslope), with more vigorous convection in fault zones. When the maximum permeability of the fault zone is 10 or more times greater than the vertical permeability of the porous layer, convection is primarily restricted to the fault plane. In all cases maximum heat flow values are located directly above faults. This paper shows that the best agreement between those data suggests a 1 m yr 4 forced convection and a permeability of the sediments of 0.5 darcy (5 x 1043 m 2) and 5 mdarcy respectively, in the horizontal and vertical derictions. for deformation within the Central Indian Ocean crust includes 1Temporarily at (1) broad scale (100-300 km) undulations of the Moho with amplitudes ranging from 1 to 3 km [Weissel et al., 1980; Geller et al., 1983]; (2) smaller-scale (5-20 km) high-angle faulting of oceanic crust and overlying sediments revealed by seismic reflection data [Weissel et al., 1980; Bull and Scrutton, 1990]; and (3) focal mechanisms of intraplate earthquakes showing a dominant north-south compressive stress [Stein and Okal, 1978; Levchenko, 1989]. The existence of faults extending through the sediments into oceanic crust has led various workers to suggest hydrothermal fluid circulation in the Bengal Fan [Cochran et al., 1989; Corrigan, 1991]. Heat flow in the Central Indian Ocean has been extensively surveyed [Von Herzen and Vacquier, 1966; Langseth and Taylor, 1967; Anderson et al., 1977; Kutas et al., 1979; Geller et al., 1983]. The heat flow data display two main characteristics: they are extremely variable over the Bengal Fan area and their average value of 71 mW m -2 is 20-30 mW m -2 greater than Parson and Sclater's [1977] theoretical value for a 72-82 Ma old oceanic lithosphere. Several models have been proposed to explain this heat flow anomaly. Weissel et al. [1980] suggest that excess heat is generated by friction 8083 8084 ORMOND ...

No abstract