Subduction megathrust heterogeneity characterized from 3D seismic data


Megathrust roughness and structural complexity are thought to be controls on earthquake slip at subduction zones because they result in heterogeneity in shear strength and resolved stress. However, because active megathrust faults are difficult to observe, the causes and scales of complexity are largely unknown. Here we measured the in situ properties of the megathrust of the Middle America subduction zone in a three-dimensional seismic reflection volume to determine how fault properties vary. We quantify spatial variability in the megathrust roughness, overburden and rock physical properties. Heterogeneity in the megathrust roughness exists at length scales of a few kilometres because the megathrust is dissected by active lower-plate normal faults, which offset the megathrust and renewed fault roughness. Spatial variations in the rock physical properties at the plate interface are characterized by correlation length scales of hundreds of metres. Frontal prism taper, historical seismicity and the variation in earthquake stress drop values local to the megathrust are all affected by the heterogeneity at these length scales. Both geometric and rheological complexities may therefore control the mechanical behaviour of the subduction plate interface, which includes earthquake rupture characteristics.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Fig. 1: Characteristics of the megathrust.
Fig. 2: Dissection of the megathrust by plate-bending normal faults.
Fig. 3: Analysis of the megathrust reflection.
Fig. 4: Analysis of earthquake source parameters.

Data availability

The 3D prestack depth-migrated seismic reflection data collected as part of Seismic Project MGL1106 Costa Rica Seismogenesis Project (CRISP) are available in the data repository at with identifier The megathrust reflection geometry, depth beneath sea floor and amplitude for the region shown in Fig. 3 are available at

Code availability

The code used to generate the apparent dip (‘dip-steering’) volume and dip-steered median-filtered data can be accessed at


  1. 1.

    Ye, L. L., Kanamori, H. & Lay, T. Global variations of large megathrust earthquake rupture characteristics. Sci. Adv. 4, eaao4915 (2018).

    Google Scholar 

  2. 2.

    Loveless, J. P. & Meade, B. J. Two decades of spatiotemporal variations in subduction zone coupling offshore Japan. Earth Planet. Sci. Lett. 436, 19–30 (2016).

    Google Scholar 

  3. 3.

    Lay, T., Kanamori, H. & Ruff, L. The asperity model and the nature of large subduction zone earthquakes. Earthq. Pred. Res. 1, 3–71 (1982).

    Google Scholar 

  4. 4.

    Noda, H. & Lapusta, N. Three-dimensional earthquake sequence simulations with evolving temperature and pore pressure due to shear heating: effect of heterogeneous hydraulic diffusivity. J. Geophys. Res. Solid Earth 115, B12314 (2010).

    Google Scholar 

  5. 5.

    Fang, Z. J. & Dunham, E. M. Additional shear resistance from fault roughness and stress levels on geometrically complex faults. J. Geophys. Res. Solid Earth 118, 3642–3654 (2013).

    Google Scholar 

  6. 6.

    Shi, Z. Q. & Day, S. M. Rupture dynamics and ground motion from 3-D rough-fault simulations. J. Geophys. Res. Solid Earth 118, 1122–1141 (2013).

    Google Scholar 

  7. 7.

    Yu, H. Y., Liu, Y. J., Yang, H. F. & Ning, J. Y. Modeling earthquake sequences along the Manila subduction zone: effects of three-dimensional fault geometry. Tectonophysics 733, 73–84 (2018).

    Google Scholar 

  8. 8.

    Bletery, Q. et al. Mega-earthquakes rupture flat megathrusts. Science 354, 1027–1031 (2016).

    Google Scholar 

  9. 9.

    Gao, X. & Wang, K. L. Strength of stick–slip and creeping subduction megathrusts from heat flow observations. Science 345, 1038–1041 (2014).

    Google Scholar 

  10. 10.

    Wang, K. L. & Bilek, S. L. Invited review paper: fault creep caused by subduction of rough seafloor relief. Tectonophysics 610, 1–24 (2014).

    Google Scholar 

  11. 11.

    Ikari, M. J., Niemeijer, A. R., Spiers, C. J., Kopf, A. J. & Saffer, D. M. Experimental evidence linking slip instability with seafloor lithology and topography at the Costa Rica convergent margin. Geology 41, 891–894 (2013).

    Google Scholar 

  12. 12.

    Bangs, N. L., McIntosh, K. D., Silver, E. A., Kluesner, J. W. & Ranero, C. R. Fluid accumulation along the Costa Rica subduction thrust and development of the seismogenic zone. J. Geophys. Res. Solid Earth 120, 67–86 (2015).

    Google Scholar 

  13. 13.

    Edwards, J. H. et al. Corrugated megathrust revealed offshore from Costa Rica. Nat. Geosci. 11, 197–202 (2018).

    Google Scholar 

  14. 14.

    Kluesner, J. W. et al. High density of structurally controlled, shallow to deep water fluid seep indicators imaged offshore Costa Rica. Geochem. Geophys. Geosyst. 14, 519–539 (2013).

    Google Scholar 

  15. 15.

    Tobin, H., Vannucchi, P. & Meschede, M. Structure, inferred mechanical properties, and implications for fluid transport in the decollement zone, Costa Rica convergent margin. Geology 29, 907–910 (2001).

    Google Scholar 

  16. 16.

    Rowe, C. D., Moore, J. C., Remitti, F. & Scientist, I. E. T. The thickness of subduction plate boundary faults from the seafloor into the seismogenic zone. Geology 41, 991–994 (2013).

    Google Scholar 

  17. 17.

    Bangs, N. L., McIntosh, K. D., Silver, E. A., Kluesner, J. W. & Ranero, C. R. A recent phase of accretion along the southern Costa Rican subduction zone. Earth Planet. Sci. Lett. 443, 204–215 (2016).

    Google Scholar 

  18. 18.

    Brodsky, E. E., Kirkpatrick, J. D. & Candela, T. Constraints from fault roughness on the scale-dependent strength of rocks. Geology 44, 19–22 (2016).

    Google Scholar 

  19. 19.

    Dascher-Cousineau, K., Kirkpatrick, J. D. & Cooke, M. L. Smoothing of fault slip surfaces by scale-invariant wear. J. Geophys. Res. Solid Earth 123, 7913–7930 (2018).

    Google Scholar 

  20. 20.

    Bangs, N. L. B. et al. Broad, weak regions of the Nankai Megathrust and implications for shallow coseismic slip. Earth Planet. Sci. Lett. 284, 44–49 (2009).

    Google Scholar 

  21. 21.

    Choy, G. L. & Kirby, S. H. Apparent stress, fault maturity and seismic hazard for normal-fault earthquakes at subduction zones. Geophys. J. Int. 159, 991–1012 (2004).

    Google Scholar 

  22. 22.

    Okuwaki, R. & Yagi, Y. Rupture process during the M w 8.1 2017 Chiapas Mexico earthquake: shallow intraplate normal faulting by slab bending. Geophys. Res. Lett. 44, 11816–11823 (2017).

    Google Scholar 

  23. 23.

    Regalla, C., Rowe, C., Harrichhausen, N., Tarling, M. & Singh, J. in Geology and Tectonics of Subduction Zones: A Tribute to Gaku Kimura (eds Byrne, T. et al.) 155–174 (GSA Special Paper 534, Geological Society of America, 2018).

  24. 24.

    Dahlen, F. A. Critical taper model of fold-and-thrust belts and accretionary wedges. Ann. Rev. Earth Planet. Sci. 18, 55–99 (1990).

    Google Scholar 

  25. 25.

    DeShon, H. R. et al. Seismogenic zone structure of the southern Middle America Trench, Costa Rica. J. Geophys. Res. Solid Earth 108, 2491 (2003).

    Google Scholar 

  26. 26.

    Arroyo, I. G., Grevemeyer, I., Ranero, C. R. & von Huene, R. Interplate seismicity at the CRISP drilling site: the 2002 M w 6.4 Osa earthquake at the southeastern end of the Middle America Trench. Geochem. Geophys. Geosyst. 15, 3035–3050 (2014).

    Google Scholar 

  27. 27.

    Candela, T., Renard, F., Bouchon, M., Schmittbuhl, J. & Brodsky, E. E. Stress drop during earthquakes: effect of fault roughness scaling. Bull. Seismol. Soc. Am. 101, 2369–2387 (2011).

    Google Scholar 

  28. 28.

    Denolle, M. A. & Shearer, P. M. New perspectives on self-similarity for shallow thrust earthquakes. J. Geophys. Res. Solid Earth 121, 6533–6565 (2016).

    Google Scholar 

  29. 29.

    Bangs, N. L. et al. Processed 3D Volume of Multi-channel Seismic Data Offshore Western Costa Rica Acquired during the R/V Marcus G. Langseth Expedition MGL1106 (2011) as part of the Costa Rica Seismogenesis Project (CRISP) (Academic Seismic Portal at UTIG, Marine Geoscience Data System, 2018);

  30. 30.

    Knapp, R. W. Vertical resolution of thick beds, thin beds, and thin-bed cyclothems. Geophysics 55, 1183–1190 (1990).

    Google Scholar 

  31. 31.

    Ricker, N. The form and laws of propagation of seismic wavelets. Geophysics 18, 10–40 (1953).

    Google Scholar 

  32. 32.

    Rafaelsen, B. et al. in Glacier-Influenced Sedimentation on High-Latitude Continental Margins (eds Dowdeswell, J. A. & Cofaigh, C. Ó.) 259–276 (Special Publications Vol. 203, Geological Society, 2002).

  33. 33.

    Thomson, D. J. Spectrum estimation and harmonic-analysis. Proc. IEEE 70, 1055–1096 (1982).

    Google Scholar 

  34. 34.

    Brune, J. N. Tectonic stress and spectra of seismic shear waves from earthquakes. J. Geophys. Res. 75, 4997–5009 (1970).

    Google Scholar 

  35. 35.

    Brune, J. N. Seismic sources, fault plane studies and tectonics. Trans. Am. Geophys. Union 52, 178–187 (1971).

    Google Scholar 

  36. 36.

    Boatwright, J. Spectral theory for circular seismic sources—simple estimates of source dimension, dynamic stress drop, and radiated seismic energy. Bull. Seismol. Soc. Am. 70, 1–27 (1980).

    Google Scholar 

  37. 37.

    Aki, K. & Richards, P. G. Quantitative Seismology 2nd edn (University Science Books, 2002).

  38. 38.

    Mori, J. & Frankel, A. Source parameters for small events associated with the 1986 North Palm-Springs, California, earthquake determined using empirical Green functions. Bull. Seismol. Soc. Am. 80, 278–295 (1990).

    Google Scholar 

  39. 39.

    Abercrombie, R. E. Investigating uncertainties in empirical Green’s function analysis of earthquake source parameters. J. Geophys. Res. Solid Earth 120, 4263–4277 (2015).

    Google Scholar 

  40. 40.

    Abercrombie, R. E., Bannister, S., Ristau, J. & Doser, D. Variability of earthquake stress drop in a subduction setting, the Hikurangi Margin, New Zealand. Geophys. J. Int. 208, 306–320 (2017).

    Google Scholar 

  41. 41.

    Viegas, G., Abercrombie, R. E. & Kim, W. Y. The 2002 M5 Au Sable Forks, NY, earthquake sequence: source scaling relationships and energy budget. J. Geophys. Res. Solid Earth 115, B07310 (2010).

    Google Scholar 

  42. 42.

    Eshelby, J. D. The determination of the elastic field of an ellipsoidal inclusion, and related problems. Proc. R. Soc. Lond. A 241, 376–396 (1957).

    Google Scholar 

  43. 43.

    Madariaga, R. Dynamics of an expanding circular fault. Bull. Seismol. Soc. Am. 66, 639–666 (1976).

    Google Scholar 

  44. 44.

    Kitanidis, P. K. Introduction to Geostatistics: Applications in Hydrogeology (Cambridge Univ. Press, 1997).

  45. 45.

    Experimental (Semi-) Variogram (MATLAB Central File Exchange, 2013).

  46. 46.

    Remy, N., Boucher, A. & Wu, J. Applied Geostatistics with SGeMS: a User’s Guide (Cambridge Univ. Press, 2009).

  47. 47.

    Variogramfit (MATLAB Central File Exchange, 2010).

  48. 48.

    Kirkpatrick, J. D., Shervais, K. A. H. & Ronayne, M. J. Spatial variation in the slip zone thickness of a seismogenic fault. Geophys. Res. Lett. 45, 7542–7550 (2018).

    Google Scholar 

  49. 49.

    Lunn, R. J., Shipton, Z. K. & Bright, A. M. in The Internal Structure of Fault Zones: Implications for Mechanical and Fluid-Flow Properties (eds Wibberley, C. A. J. et al.) 231–237 (Special Publications Vol. 299, Geological Society, 2008).

  50. 50.

    Kirkpatrick, J. D. & Brodsky, E. E. Slickenline orientations as a record of fault rock rheology. Earth Planet. Sci. Lett. 408, 24–34 (2014).

    Google Scholar 

  51. 51.

    Ryan, W. B. F. et al. Global multi-resolution topography synthesis. Geochem. Geophys. Geosyst. 10, Q03014 (2009).

    Google Scholar 

Download references


Thanks to N. Bangs, K. McIntosh and the crew of the R/V Marcus G. Langseth for their efforts to acquire and process the data. Thanks also to M. Ikari for constructive comments that substantially improved the manuscript and to J. Conrad for feedback on an early version of the manuscript. This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC), Discovery Grant RGPIN-2016-04677 (J.D.K.) and the National Science Foundation (NSF) grants OCE‐0851529 and OCE‐0851380.

Author information




The study concept was formulated by J.D.K., J.W.K. and R.M.H. J.H.E. applied postprocessing, performed amplitude-driven tracking and extracted the geometric attributes of the megathrust. Geometric data validation and analysis were carried out by J.H.E., J.D.K., J.W.K. and E.A.S. A.V. and R.M.H. completed the source parameter analyses. J.D.K. wrote the manuscript with contributions from all the authors.

Corresponding author

Correspondence to James D. Kirkpatrick.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Primary Handling Editor: Stefan Lachowycz.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Megathrust power spectral density roughness compared to roughness calculated from ground-based LiDAR (Light Detection and Ranging) survey of an exposed ancient fault.

Data are the same as in Fig. 3 in the main text except for the dark grey lines at short wavelength, which are calculated from the Corona Heights Fault50. Red line has a Hurst scaling exponent of 1 for reference.

Extended Data Fig. 2 Analysis of the effect of the seismic volume attributes on roughness estimates.

a, Effect of dip-steered median filter step-outs on roughness measurements in corrugation-parallel direction. Comparison of calculated power spectral density from the shallow, well-corrugated portion of the megathrust horizon extracted from seismic volumes filtered with difference dip-steered median filter step-outs (inline x crossline). b, Same as A but for corrugation-perpendicular direction. c, Comparison of the megathrust power spectral density roughness calculated in different directions to test if the orientation of the megathrust dataset with respect to the coordinate system affected the length scale at which spectra calculated in perpendicular directions converge. Rotating the data around an axis normal to the mean plane through the dataset changes the number of profiles with different lengths, particularly the number of long profiles, and therefore the number of estimates of the PSD at long wavelengths. Roughness was calculated from profiles taken parallel and perpendicular to three reference directions: down dip and across strike; parallel and perpendicular to the corrugations present at shallow depths; perpendicular and parallel to the ridge axes in the middle portion of the megathrust. Differences in PSD at length scales > 5 km result from the smaller number of the longest profiles following rotation of the data around the Z-axis to the different reference directions. These results show the length scale of convergence is approximately the same in each pair of spectra.

Extended Data Fig. 3 Example of an abandoned megathrust horizon extracted from the footwall at around 19 km landward of the trench.

a, Map of the surface topography. Colors correspond to distance from the mean plane fitted through the region shown. Dip direction indicated by arrow. b, Power spectral density roughness of the abandoned megathrust horizon (bold lines) shown in A. This patch shares similar characteristics to the in-situ megathrust, being anisotropic and smoother in the direction of visible corrugations. Spectra for the corrugated and non-corrugated patches from Fig. 3 in the main text are shown for reference. The roughness of the abandoned horizon is intermediate between the corrugated and weakly corrugated portions of the shallow megathrust.

Extended Data Fig. 4 Analysis of 3-D seismic volume spatial resolution.

ac, Amplitude histograms of the megathrust reflection from different depth intervals. Seismic bandwidth decreases (attenuation of high frequencies) with increasing depth. d, Post-migrated Fresnel zone variation with depth. Fresnel zone size increases with depth.

Extended Data Fig. 5 Map and cross section showing earthquakes used for source parameter estimates.

a, Overview map with bathymetry51 of the study region. Circles represent aftershocks occurring between September 1999 and November 199925. Pink circles indicate earthquakes considered here occurring on the plate interface, for which corner frequencies are calculated using the spectral ratio method. Red stars are the earthquakes that occurred between August 20, 1999 and September 24, 1999, for which corner frequencies are calculated using the spectral ratio method. Yellow stars are the mainshock (big star), and earthquakes (small stars) that occurred between September 9, 1996 and September 23, 1999, for which corner frequencies are calculated using the single spectrum method. Green triangles are the CRSEIZE seismometer locations, and black triangles are RSN (SJS) and IRIS/IDA (JTS) station locations. Dashed black line indicates the location of the cross section below. b, Cross section showing earthquake hypocentral depths. Symbols as in a. Inverted triangles indicate locations of seismic stations.

Extended Data Fig. 6 Examples of waveform analyses.

af, Spectral ratio fits for the event pair 4585 (Mw=2.6) and 2554 (Mw=1.9) having CC = 0.8. a. Single spectra amplitude and noise spectra of each event. B. Spectral ratio and the model fit with the final corner frequency estimates. c. and d. Waveforms in the time domain. E. The variance (chi-square misfit) for incremented values of fc1, each value represented by a different color. F. Corresponding fits to the data at different fc1 increments. gl, Spectral ratio fits for the event pair 4016 (Mw=1.8) and 3378 (Mw=1.3) having CC = 0.7. g. Single spectra amplitude and noise spectra of each event. H. Spectral ratio and the model fit with the final corner frequency estimates. i. and j. Waveforms in the time domain. k. The variance (chi-square misfit) for incremented values of fc1, each value represented by a different color. l. Corresponding fits to the data at different fc1 increments.

Extended Data Fig. 7 Geostatistical analysis of the amplitude field from the corrugated subregion of the shallow megathrust.

γ(h) shown here is the same as in Fig. 3d of the main text. The results clearly show that γ(h) attains a constant value for large h (normalized to 1 here). The curves through the results represent model fits to experimental variograms, which have correlation distances, α, of 485 and 445 m in the two perpendicular reference directions.

Extended Data Table 1 Summary of data sources and methods used to estimate the source parameters plotted in the figures in the main text
Extended Data Table 2 Correlation distances obtained from model fits to the experimental variograms shown in Fig. 3 in the main text

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Kirkpatrick, J.D., Edwards, J.H., Verdecchia, A. et al. Subduction megathrust heterogeneity characterized from 3D seismic data. Nat. Geosci. 13, 369–374 (2020).

Download citation

Further reading


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing