Submarine hydrogeology of the Hawaiian archipelagic apron 2. Numerical simulations of coupled heat transport and fluid flow

We perform numerical simulations of buoyancy-driven, pore fluid flow in the Hawaiian archipelagic apron and underlying oceanic crust in order to deter mine the extent to which heat redistributed by such flow might cause conduc tive heat flow measurements to underrepresent the true mantle heat flux. We also seek an understanding of undulations observed in finely spaced heat f low measurements acquired north of Oahu and Maro Reef with wavelengths of 1 0 to 100 km and amplitudes of 2 to 7 mW m(-2). We find that pore fluid flow can impart significant perturbations to seafloor heat flow from the value expected assuming a constant mantle flux. In the simplest scenario, moat-wi de circulation driven by bathymetric relief associated with the volcanic ed ifice recharges a fluid system over the moat and discharges the geothermall y heated water through the volcanic edifice. The existing heat flow data ar e unable to confirm the existence of such a flow regime, in that it produce s prominent heat flow anomalies only on the steep flanks of the volcano whe re heat flow probes cannot penetrate. However, this flow system does not si gnificantly mask the mantle flux for reasonable permeabilities and flow rat es. Another numerical simulation in which the upper oceanic basement acts a s a aquifer for a flow loop recharged at basement outcrops on the flexural arch and discharged within a permeable volcanic edifice is capable of almos t uniformly depressing conductive heat flow across the entire moat by simil ar to 15%. Large heat flow anomalies (>20 mW m(-2)) are located over the re charge and discharge zones but are beyond the area sampled by our data. Pre sumably finely spaced heat flow measurements over the flexural arch could t est for the existence of the predicted recharge zone. We demonstrate that t he prominent, shorter-wave undulations in heat flow across the Oahu and Mar o Reef moats are too large to be explained solely by relief in the upper oc eanic basement. More likely, shallower large-scale turbidites or debris flo ws also serve as aquifers within the less permeable moat sediments. With ou r limited information on the structural geology of the moat, permeability s tructure of its major geologic units, and their variations in the third dim ension, we are not able to exactly match the spatial distribution of heat f low anomalies in our data, but spectral comparisons look promising.