Knowledge of the stress state in subduction zones is crucial for understanding the mechanics of geological and geophysical processes occurring in subduction zones, such as earthquake generation (e.g., Moore & Saffer, 2001), stress changes during the earthquake cycle (e.g., Wang & Hu, 2006) and accretionary wedge formation (e.g., Dahlen, 1990). Rock strength, which must be known to determine the stress state, is often determined by combining a failure criterion with the relationship between P-wave velocity (V p ) and unconfined compressive strength (C 0 ) (e.g., Chang et al., 2010;Song & Chang, 2017;Zoback, 2007). For example, to estimate the possible ranges of in situ horizontal principal stresses, measurements of breakout width, an indicator of rock failure, are commonly combined with estimates of rock strength (C 0 ) derived from V p measurements. However, because relatively few C 0 data are available from laboratory experiments on drill cores, C 0 is estimated instead from logging-while-drilling (LWD) V p data and laboratory-derived V p -C 0 relationships (e.g., Chang et al., 2010). Thus, when LWD V p data are available, the V p -C 0 relationship can be used for estimating rock strength and stresses at depth in subduction zones (e.g., Chang et al., 2006). Among published V p -C 0 relationships, those derived from measurements of sedimentary rocks, in particular, vary depending on the tectonic field and the rock type and age (e.g.,