Fluid venting was recently found on the central Oregon accretionary wedge and evidence of fluid flow along the decollement were obtained in the Barbados wedge. The chemistry of hydrocarbons contained in the migrating fluids is different in these two areas and indicates that the main source is deep for the Barbados decollement and shallow for the Oregon vents. We apply a finite element model, with imposed thrusting kinematics and nonlinearities in the compaction equation, to the toe of both wedges. No deep fluid source is included; thus, the main source is the footwall just below the frontal thrust. The model indicates the following: (1) Pore pressure build up would be essentially controlled by kN, the equivalent permeability in the direction normal to the strata. In spite of very different sediment lithologies, kN is of the order of 10-17 m2 for both Oregon turbidities (Horath, 1989) and Barbados muds (Taylor and Leonard, 1990), and no high pore pressures can be obtained inside the offscraped sediments. (2) Permeability anisotropy greatly modifies fluid flow paths.

A fault with a permeability anisotropy of 103 would channel fluids efficiently but leave the footwall pressure almost unchanged. A stratigraphic permeability anisotropy of 102 (as documented for the Oregon turbidities) would redirect the flow of pressurized fluids in the footwall along strata into the trench. If clogging of the frontal thrust occured, fluids expelled from the footwall would move across the frontal thrust, migrate along permeable strata in the hanging wall and then be vented near the top of the eroded anticline (as observed in Oregon). (3) Comparison of published estimates from data of fluid explusion fluxes with the modeling results suggests that a deeper and larger fluid source is needed at the Barbados site but is not required for the Oregon site. ¿American Geophysical Union 1991