Slip‐rate‐dependent melt extraction at oceanic transform faults

Abstract

AbstractCrustal thickness differences between oceanic transform faults and associated mid‐ocean ridges may be explained by melt migration and extraction processes. Slow‐slipping transform faults exhibit more positive gravity anomalies than the adjacent spreading centers, indicating relative thin crust in the transform domain, whereas at intermediate‐spreading and fast‐spreading ridges transform faults are characterized by more negative gravity anomalies than the adjacent spreading centers, indicating thick crust in the transform domain. We present numerical models reproducing these observations and infer that melt can be extracted at fast‐slipping transforms, but not at slow‐slipping ones. Melt extraction is modeled as a three‐step process. (1) Melt moves vertically through buoyancy‐driven porous flow enhanced by subvertical dissolution channels. (2) Melt accumulates in and travels along a decompaction channel lining a low‐permeability barrier at the base of the thermal boundary layer. (3) Melt is extracted to the surface when it enters a melt extraction zone. A melt extraction width of 2–4 km and a melt extraction depth of 15–20 km are needed to fit the tectonic damages associated with oceanic plate boundaries that reach into the upper mantle. Our conclusions are supported by the different degrees of magmatic activities exhibited at fast‐slipping and slow‐slipping transforms as reflected in geological features, geochemical signals, and seismic behaviors. We also constrain that the maximum lateral distance of crust‐level dike propagation is about 50–70 km.