Constraints on in situ stresses in the Nankai Trough, offshore SW Japan from borehole breakouts and laboratory measurements of rock strength

Quantifying in situ stress orientation and magnitude in active tectonic settings is important toward understanding mechanics and absolute strength of faults, yet direct measurements of these quantities are scarce. In boreholes, estimates of stress magnitude can be obtained by combining constraints derived from (1) observation of wellbore failures such as borehole breakouts (BO); and (2) the limiting case of frictional sliding on optimally oriented faults. Here, we use width of BO documented by Ocean Drilling Program (ODP) drilling in the accretionary wedge of the Nankai subduction zone offshore SW Japan to constrain magnitudes of maximum and minimum horizontal stresses (SHmax and Shmin). During ODP Legs 131 and 196, drilling and coring at Site 808, located ~3 km landward of the trench, penetrated the accretionary wedge, plate boundary decollement, and underthrusting sediments. Logging while drilling resistivity-at-the-bit (LWD) images document BO in the uppermost part of the accretionary prism; no BO are observed below 825 mbsf in the lowermost accretionary wedge and subducted sediments. Computed stress magnitudes depend on rock mechanical properties such as unconfined compressive strength (UCS), poissons ratio (n), and coefficients of internal and sliding friction (μi and μ), as well as formation pore pressure (Pf), and annular fluid pressures. Few direct constraints are available for UCS; we estimate a range of values from empirically derived relationships between p-wave velocity (Vp) and UCS. Based on existing laboratory studies, we assume values of μ = 0.4, μi = 0.4, and n = 0.45, and we consider a range of Pf values in our analysis, from hydrostatic to ~75% of lithostatic. Resulting stress magnitudes are most sensitive to UCS, and less sensitive to the choice of μi or n. Pf affects the criteria for generating BO and for slip on existing faults, and thus has a large effect on computed stress magnitude. For the case of hydrostatic Pf, SHmax (referenced to the seafloor) increases from 5.7-9.1 MPa at 364 mbsf, to 11.7-16.0 MPa at 815 mbsf, and is consistent with a thrust or strike-slip faulting regime. For the overpressured case, at several depths the values of SHmax computed from BO widths are not consistent with the limits on stress defined by the failure criteria for slip on existing faults, suggesting that UCS is higher than we estimate, or that Pf cannot be 75% of lithostatic throughout the section. Below 825 mbsf, lack of observed BO indicates that SHmax < 18.5 MPa. Our initial analysis indicates that values of SHmax are not consistent with slip on active thrust faults throughout the accretionary wedge. This could be explained if stress varies through the seismic cycle, such that SHmax in the outer accretionary wedge is transiently increased following large earthquakes to allow active thrusting. However, there is significant uncertainty in values of UCS estimated from Vp; ongoing triaxial tests on core samples will provide direct measurements of UCS to better constrain stress magnitudes.