The bottom-simulating refl ector (BSR) is a seismic indicator of the bottom of a gas hydrate stability zone. Its depth can be used to calculate the seafl oor surface heatfl ow. The calculated BSR heatfl ow variations include disturbances from two important factors: (1) seafloor topography, which focuses the heat flow over regions of concave topography and defocuses it over regions of convex topography, and (2) the focused warm fluid flow within the accretionary prism coming from depths deeper than BSR. The focused fluid fl ow can be detected if the contribution of the topography to the BSR heatfl ow is removed. However, the analytical equation cannot solve the topographic effect at complex seafloor regions. We prove that 3D finite element method can model the topographic effect on the regional background heat flow with high accuracy, which can then be used to correct the topographic effect and obtain the BSR heat flow under the condition of perfectly flat topography. By comparing the corrected BSR heatfl ow with the regional background heat fl ow, focusedfl uidfl ow regions can be detected that are originally too small and cannot be detected using present-day equipment. This method was successfully applied to the mid-slope region of northern Cascadia subducting margin. The results suggest that the Cucumber Ridge and its neighboring area are positive heatfl ow anomalies, about 10%-20% higher than the background heatfl ow after 3D topographic correction. Moreover, the seismic imaging associated the positive heatfl ow anomaly areas with seabed fracture-cavity systems. This suggests flow of warm gas-carrying fluids along these high-permeability pathways, which could result in higher gas hydrate concentrations.