Gao and Wang (2014) inverted surface heat-flow observations with two-dimensional thermal models of ten subduction margins and concluded that seismogenic, stick-slip megathrusts are weak (apparent friction μ′\mu' of roughly 0.025 for the Japan Trench) while creeping megathrusts are stronger (roughly 0.13 for the northern Hikurangi margin), with μ′\mu' anti-correlating with the maximum earthquake magnitude across margins. The authors themselves flagged that this inference is confounded by sampling-biased heat-flow coverage, spatially variable hydrothermal circulation, transient coseismic dynamic weakening, incompleteness of the maximum-magnitude proxy, and an unknown mechanism linking seafloor ruggedness to integrated fault strength. The open problem is to determine robustly the long-term strength contrast between stick-slip and creeping megathrusts with these confounders bounded; to identify the mechanism that sets it; and to reconcile the heat-flow inference with independent indicators — earthquake stress orientations, exhumed-rock thermal records, borehole heat flux, and geodesy — some of which imply the opposite ranking. These questions remain unresolved in the literature.

Background

A subduction megathrust is the huge thrust fault along which an oceanic plate dives beneath another plate; it accommodates plate convergence either by sudden earthquakes (stick-slip or locked behavior) or by nearly continuous aseismic sliding (creeping behavior), with many margins mixing the two. A long-standing way to probe the strength of such faults is frictional (shear) heating: if the fault resists sliding with shear stress τ\tau, mechanical work τ\tau times slip rate is dissipated as heat, warming the overriding plate and raising surface heat flow (the outward flow of geothermal energy per unit area, measured with seafloor probes or inferred from bottom-simulating reflectors (BSR), a seismic marker of the temperature-controlled base of gas hydrate stability). Because laboratory rock friction is high (coefficients around 0.6-0.85, the Byerlee range), the long-standing heat-flow paradox on the San Andreas fault — measurements showed too little heat for such strong friction — motivated the idea that major faults are weak. Gao and Wang (2014) extended this logic to subduction in a Science paper. They quantified strength by the apparent coefficient of friction μ′=τ/σn′\mu' = \tau/\sigma'_n, the ratio of shear stress to effective normal stress (normal stress reduced by pore-fluid pressure), and fitted two-dimensional finite-element thermal models of ten margins to surface heat-flow data. Their headline results: the Japan Trench, a smooth-incoming-seafloor margin that produced the 2011 MwM_w (moment magnitude) 9.0 Tohoku earthquake, requires μ′≈0.025\mu' \approx 0.025, whereas the northern Hikurangi margin off New Zealand, whose rough incoming seafloor creeps and hosts slow-slip events (transient episodes of aseismic slip, slower than earthquakes but faster than steady creep, recorded by geodetic monitoring of surface deformation), requires μ′≈0.13\mu' \approx 0.13; across all ten margins μ′\mu' anti-correlates with MmaxM_{max}, the largest earthquake magnitude recorded on each margin. They concluded that smooth megathrusts producing great earthquakes are weaker and dissipate less heat than rough, creeping ones. The paper's own discussion, however, enumerates unresolved issues: heat-flow coverage is biased toward highly seismogenic margins; Costa Rica's apparent friction was poorly constrained and the Kermadec margin's present creeping-versus-locking state was undetermined from geodesy; hydrothermal circulation (groundwater flow through the volcanic crust that advects and redistributes heat, decoupling surface heat flow from deep heat generation) is not quantified everywhere; transient dynamic weakening during earthquake slip (rapid stress drops during rupture) means that low frictional heat does not immediately imply low static strength; MmaxM_{max} is an incomplete proxy for long-term seismic slip; and no quantitative mechanism connects seafloor ruggedness and damage-zone structure (the architecture of the fractured, permeable rock volume surrounding the principal slip surface) to integrated μ′\mu'. The problem is sharpened by an independent observable that appears to contradict the heat-flow ranking: earthquake focal-mechanism stress orientations (the orientation of the ambient stress field inverted from the radiation patterns and slip senses of earthquakes) have been argued to show that creeping subduction zones are weaker than locked ones, with low apparent friction in creeping margins attributed to pore-fluid overpressure (fluid pressure above hydrostatic, which reduces the effective normal stress and hence frictional resistance) and low intrinsic friction, and stress-rotation modeling at Cascadia likewise requires a weak plate boundary at seismogenic depth. Later heat-flow work also disagrees about absolute stress levels: borehole heat flux and exhumed-rock thermal modeling have been used both to require non-negligible shear heating and to infer shear stresses around 100 MPa at depth beneath northern Honshu, far above what μ′≈0.025\mu' \approx 0.025 on a lithostatically loaded fault would suggest, and theoretical rupture dynamics shows that pulse-like earthquake ruptures can radiate little heat while absolute stresses remain high. Determining the true long-term strength of seismogenic versus creeping megathrusts, the mechanism that sets it, and the consistency of the different strength indicators is therefore an open, actively contested problem in fault mechanics and subduction geophysics.

Problem Statement

Determine, robustly and quantitatively, the long-term integrated strength of seismogenic stick-slip subduction megathrusts versus creeping subduction megathrusts — expressed as the apparent friction μ′=τ/σn′\mu' = \tau/\sigma'_n (fault-parallel shear stress over effective normal stress) or equivalently as absolute shear stress resolved on the plate interface — improving on the heat-flow-inverted estimates reported by Gao and Wang (2014) (Japan Trench μ′≈0.025\mu' \approx 0.025; values up to μ′≈0.13\mu' \approx 0.13 for geometrically rough, creeping margins including northern Hikurangi; with μ′\mu' anti-correlating with MmaxM_{max}, the largest earthquake magnitude recorded on each margin). Determining this requires (a) bounding, rather than assuming away, the confounders identified in the source work (Gao and Wang 2014): spatially variable hydrothermal circulation that decouples surface heat flow from frictional heating, transient coseismic dynamic weakening that decouples frictional heat from static strength, sampling-biased heat-flow coverage that under-constrains creeping and intermediate margins, and historical incompleteness of the MmaxM_{max} proxy for long-term seismic slip; (b) identifying the physical mechanism — incoming-seafloor ruggedness, damage-zone architecture, pore-fluid overpressure, or pressure-solution creep (deformation by dissolution and re-precipitation of minerals at stressed grain contacts) coupled with frictional-viscous flow — that sets the inferred strength or strength contrast, quantitatively enough to predict which margins are weak versus strong; and (c) reconciling the heat-flow-based inference with independent strength indicators — illustratively, earthquake stress-orientation data (which have been argued to imply the opposite ranking, that creeping zones are weaker than locked ones), exhumed-rock pressure-temperature records (thermal histories of rocks now exposed at the surface that once sat at depth on the plate interface) requiring shear heating, borehole heat-flux measurements implying higher absolute stresses, and geodetic locking estimates (degree of plate-interface locking inferred from GPS-observed surface deformation); any claim that two or more indicator classes cannot be jointly satisfied must rest on observationally justified assumptions consistent with the confounder bounds required in (a). These three components are jointly necessary for one scientific objective: an internally consistent determination of how strong subduction megathrusts actually are and why, reported as the long-term integrated μ′\mu' for each margin class with its uncertainty and the sign of the contrast between the classes.

References

  1. Gao, X. & Wang, K. (2014). Strength of stick-slip and creeping subduction megathrusts from heat flow observations. Science 345(6200), 1038-1041. doi:10.1126/science.1255487 https://doi.org/10.1126/science.1255487
  2. Hardebeck, J. L. (2015). Stress orientations in subduction zones and the strength of subduction megathrust faults. Science 349(6253), 1213-1216. doi:10.1126/science.aac5625 https://doi.org/10.1126/science.aac5625
  3. Hardebeck, J. L. & Loveless, J. P. (2017). Creeping subduction zones are weaker than locked subduction zones. Nature Geoscience 11(1), 60-64. doi:10.1038/s41561-017-0032-1 https://doi.org/10.1038/s41561-017-0032-1
  4. Li, D., McGuire, J. J., Liu, Y. & Hardebeck, J. L. (2018). Stress rotation across the Cascadia megathrust requires a weak subduction plate boundary at seismogenic depths. Earth and Planetary Science Letters 485, 55-64. doi:10.1016/j.epsl.2018.01.002 https://doi.org/10.1016/j.epsl.2018.01.002
  5. England, P., Matsumoto, T. & Wallis, S. (2024). Implications of surface heat flux for shear stress and temperature on the plate interface beneath northern Honshu. Geochemistry, Geophysics, Geosystems 25(2), e2023GC011285. doi:10.1029/2023gc011285 https://doi.org/10.1029/2023gc011285
  6. Kohn, M. J., Castro, A. E., Kerswell, B. C., Ranero, C. R. & Spear, F. S. (2018). Shear heating reconciles thermal models with the metamorphic rock record of subduction. Proceedings of the National Academy of Sciences 115(46), 11706-11711. doi:10.1073/pnas.1809962115 https://doi.org/10.1073/pnas.1809962115
  7. Hass, B. & Harris, R. N. (2016). Heat flow along the Costa Rica Seismogenesis Project drilling transect: Implications for hydrothermal and seismic processes. Geochemistry, Geophysics, Geosystems 17(6), 2110-2127. doi:10.1002/2016gc006314 https://doi.org/10.1002/2016gc006314
  8. Antriasian, A., Harris, R. N., Trehu, A. M., Henrys, S. A., Phrampus, B. J., Lauer, R., Gorman, A. R., Pecher, I. A. & Barker, D. (2018). Thermal regime of the northern Hikurangi Margin, New Zealand. Geophysical Journal International 216, ggy450. doi:10.1093/gji/ggy450 https://doi.org/10.1093/gji/ggy450
  9. Fulton, P. M. & Fresonke, M. (2026). Low heat flow near the deformation front of the northern Hikurangi Margin revealed by deep borehole observatories. Geochemistry, Geophysics, Geosystems 27(6), e2025GC012750. doi:10.1029/2025gc012750 https://doi.org/10.1029/2025gc012750
  10. Ye, L., Hu, Y., Xia, T., Lay, T., Sang, Y., Chen, X., Kanamori, H., Romano, F., Lorito, S. & Gui, Z. (2025). The 2021 Mw 8.1 Kermadec earthquake sequence: great earthquake rupture along the mantle/slab contact. Journal of Geophysical Research: Solid Earth 130(4), e2024JB030926. doi:10.1029/2024jb030926 https://doi.org/10.1029/2024jb030926
  11. Zeng, H., Wei, S. & Yang, H. (2025). The 2021 Mw8.1 Kermadec megathrust earthquake: an event rupturing the slab-mantle interface with up-dip high-frequency seismic energy radiation. Geophysical Research Letters 52(4), e2024GL113015. doi:10.1029/2024gl113015 https://doi.org/10.1029/2024gl113015
  12. Lambert, V., Lapusta, N. & Perry, S. (2021). Propagation of large earthquakes as self-healing pulses or mild cracks. Nature 591(7849), 252-258. doi:10.1038/s41586-021-03248-1 https://doi.org/10.1038/s41586-021-03248-1
  13. Lambert, V. & Lapusta, N. (2023). Absolute stress levels in models of low-heat faults: links to geophysical observables and differences for crack-like ruptures and self-healing pulses. Earth and Planetary Science Letters 618, 118277. doi:10.1016/j.epsl.2023.118277 https://doi.org/10.1016/j.epsl.2023.118277
  14. Uchida, N. & Bürgmann, R. (2021). A decade of lessons learned from the 2011 Tohoku-oki earthquake. Reviews of Geophysics 59(2), e2020RG000713. doi:10.1029/2020rg000713 https://doi.org/10.1029/2020rg000713
  15. Copley, A. (2017). The strength of earthquake-generating faults. Journal of the Geological Society 175(1), 1-12. doi:10.1144/jgs2017-037 https://doi.org/10.1144/jgs2017-037
  16. Collettini, C., Tesei, T., Scuderi, M. M., Carpenter, B. M. & Viti, C. (2019). Beyond Byerlee friction, weak faults and implications for slip behavior. Earth and Planetary Science Letters 519, 245-263. doi:10.1016/j.epsl.2019.05.011 https://doi.org/10.1016/j.epsl.2019.05.011
  17. Harris, R. N., Spinelli, G. A. & Fisher, A. T. (2017). Hydrothermal circulation and the thermal structure of shallow subduction zones. Geosphere 13(5), 1425-1444. doi:10.1130/ges01498.1 https://doi.org/10.1130/ges01498.1
  18. Fagereng, A. & den Hartog, S. A. M. (2016). Subduction megathrust creep governed by pressure solution and frictional-viscous flow. Nature Geoscience 10(1), 51-57. doi:10.1038/ngeo2857 https://doi.org/10.1038/ngeo2857
  19. Tulley, C. J., Fagereng, A. & Ujiie, K. (2020). Hydrous oceanic crust hosts megathrust creep at low shear stresses. Science Advances 6(22), eaba1529. doi:10.1126/sciadv.aba1529 https://doi.org/10.1126/sciadv.aba1529
  20. Bürgmann, R. (2018). The geophysics, geology and mechanics of slow fault slip. Earth and Planetary Science Letters 495, 112-134. doi:10.1016/j.epsl.2018.04.062 https://doi.org/10.1016/j.epsl.2018.04.062
  21. Koyama, Y., Wallis, S. R. & Nagaya, T. (2024). Subduction plate interface shear stress associated with rapid subduction at deep slow earthquake depths: example from the Sanbagawa belt, southwestern Japan. Solid Earth 15(2), 143-166. doi:10.5194/se-15-143-2024 https://doi.org/10.5194/se-15-143-2024
  22. Bletery, Q., Thomas, A. M., Rempel, A. W. & Hardebeck, J. L. (2017). Imaging shear strength along subduction faults. Geophysical Research Letters 44(22), 11,114-11,122. doi:10.1002/2017gl075501 https://doi.org/10.1002/2017gl075501