Description: Marine heat flow data, sampled by a 22-channel violin bow probe offshore Japan in 2016. Data are unfiltered, that is, they are measured field data, not necessarily refecting geothermal heat flux. The data are correct in a technical way but they are influenced by mud volcanism, gas hydrates and bottom water currents.

Description: The location of the seaward tip of a subduction thrust controls material transfer at convergent plate margins, and hence global mass balances. At approximately half of those margins, the material of the subducting plate is completely underthrust so that no accretion or even subduction erosion takes place. Along the remaining margins, material is scraped off the subducting plate and added to the upper plate by frontal accretion. We here examine the physical properties of subducting sediments off Costa Rica and Nankai, type examples for an erosional and an accretionary margin, to investigate which parameters control the level where the frontal thrust cuts into the incoming sediment pile. A series of rotary-shear experiments to measure the frictional strength of the various lithologies entering the two subduction zones were carried out. Results include the following findings: (1) At Costa Rica, clay-rich strata at the top of the incoming succession have the lowest strength (µres = 0.19) while underlying calcareous ooze, chalk and diatomite are strong (up to µres = 0.43; µpeak = 0.56). Hence the entire sediment package is underthrust. (2) Off Japan, clay-rich deposits within the lower Shikoku Basin inventory are weakest (µres = 0.13–0.19) and favour the frontal proto-thrust to migrate into one particular horizon between sandy, competent turbidites below and ash-bearing mud above. (3) Taking in situ data and earlier geotechnical testing into account, it is suggested that mineralogical composition rather than pore-pressure defines the position of the frontal thrust, which locates in the weakest, clay mineral-rich (up to 85 wt.%) materials. (4) Smectite, the dominant clay mineral phase at either margin, shows rate strengthening and stable sliding in the frontal 50 km of the subduction thrust (0.0001–0.1 mm/s, 0.5–25 MPa effective normal stress). (5) Progressive illitization of smectite cannot explain seismogenesis, because illite-rich samples also show velocity strengthening at the conditions tested.

Description: Temperature gradients and in-situ thermal conductivity, measured with the Bremen heat flow probe. There are 21 channels, spaced 0.26 cm.