Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Upper-plate controls on subduction zone geometry, hydration and earthquake behaviour

Abstract

Many characteristics of the incoming oceanic lithosphere, such as its age, rigidity, fabric orientation or sediment thickness, are often cited as important properties controlling the geometry, state of stress, dynamics and hazard potential of subduction zones, yet the links between upper-plate structures and subduction zone processes remain poorly understood. Here we report that high forearc wavespeeds (vP greater than 6.6 km s−1) beneath 8,000 km2 of Kii Peninsula are associated with the Kumano pluton. We show that the dense, high-rigidity Kumano pluton generates a large vertical load, which forces the incoming Philippine Sea Plate to subduct with a trajectory that is a factor of two steeper than adjacent regions. Beneath the region of maximum curvature and faulting of the Philippine Sea Plate, reduced mantle velocities (6.5–7.5 km s−1) within a 25-km-thick, 100-km-wide region at 5–30-km sub-Moho depths may reflect serpentinization (more than 40% antigorite) of the subducting mantle and enhanced porosity from bending stresses. We further report that great (larger than Mw 8) earthquakes nucleated from the flanks of the Kumano pluton in 1944 and 1946. Our study demonstrates the profound impact of upper-plate structures on the geometry, hydration state and segmentation of large megathrust earthquakes at subduction zones.

This is a preview of subscription content

Access options

Buy article

Get time limited or full article access on ReadCube.

$32.00

All prices are NET prices.

Fig. 1: Tectonic setting and geophysical data at the Nankai subduction zone in southwestern Japan.
Fig. 2: Impact of the Kumano pluton on subducting plate geometry.
Fig. 3: Impact of plate bending on mantle hydration.
Fig. 4: Spatial extent of mantle hydration, slab geometry and their relationship to the Kumano pluton.

Data availability

All OBS field data, seismic navigation and acquisition logs acquired offshore the Nankai Trough are archived at the Japan Agency for Marine-Earth Science and Technology (JAMSTEC, https://www.jamstec.go.jp/obsmcs_db/e/). All data acquired by the F-net (https://www.fnet.bosai.go.jp/top.php) and Hi-net (https://www.hinet.bosai.go.jp/) networks are archived by the Japanese National Research Institute for Earth Science and Disaster Resilience (NIED). All arrival time data from earthquakes used in this study are archived by the Japan Meteorological Agency (JMA, https://www.data.jma.go.jp/svd/eqev/data/bulletin/deck_e.html). Our final 3D P-wave tomography model of the Nankai Trough46 (https://doi.org/10.26022/IEDA/329655) can be found on the Marine Geoscience Data System.

Code availability

The travel-time tomography code associated with this paper is maintained by A.F.A. at the Institute for Geophysics at the University of Texas at Austin (aarnulf@ig.utexas.edu). Some components are available on request.

References

  1. Garcia, E. S. M., Sandwell, D. T. & Bassett, D. Outer trench slope flexure and faulting at Pacific Basin subduction zones. Geophys. J. Int. 218, 708–728 (2019).

    Article  Google Scholar 

  2. Hunter, J. & Watts, A. Gravity anomalies, flexure and mantle rheology seaward of circum-Pacific trenches. Geophys. J. Int. 207, 288–316 (2016).

    Article  Google Scholar 

  3. Naif, S., Key, K., Constable, S. & Evans, R. L. Water‐rich bending faults at the Middle America Trench. Geochem. Geophys. Geosyst. 16, 2582–2597 (2015).

    Article  Google Scholar 

  4. Ranero, C. R., Morgan, J. P., McIntosh, K. & Reichert, C. Bending-related faulting and mantle serpentinization at the Middle America trench. Nature 425, 367–373 (2003).

    Article  Google Scholar 

  5. Cai, C., Wiens, D. A., Shen, W. & Eimer, M. Water input into the Mariana subduction zone estimated from ocean-bottom seismic data. Nature 563, 389–392 (2018).

    Article  Google Scholar 

  6. Ranero, C. R., Villaseñor, A., Phipps Morgan, J. & Weinrebe, W. Relationship between bend‐faulting at trenches and intermediate‐depth seismicity. Geochem. Geophys. Geosyst. 6, Q12002, https://doi.org/10.1029/2005GC000997 (2005).

  7. Shillington, D. J. et al. Link between plate fabric, hydration and subduction zone seismicity in Alaska. Nat. Geosci. 8, 961–964 (2015).

    Article  Google Scholar 

  8. Seno, T., Stein, S. & Gripp, A. E. A model for the motion of the Philippine Sea Plate consistent with NUVEL‐1 and geological data. J. Geophys. Res. Solid Earth 98, 17941–17948 (1993).

    Article  Google Scholar 

  9. Kodaira, S. et al. A cause of rupture segmentation and synchronization in the Nankai Trough revealed by seismic imaging and numerical simulation. J. Geophys. Res. Solid Earth 111, B09301, https://doi.org/10.1029/2005JB004030 (2006).

  10. Nakanishi, A. et al. Three dimensional plate geometry and P-wave velocity models of the subduction zone in SW Japan, in Geology and Tectonics of Subduction Zones: A Tribute to Gaku Kimura, Vol. 534 (eds Byrne, T. et al.) 69 (Geological Society of America, 2018) https://doi.org/10.1130/2018.2534(04)

  11. Akuhara, T. & Mochizuki, K. Hydrous state of the subducting Philippine Sea Plate inferred from receiver function image using onshore and offshore data. J. Geophys. Res. Solid Earth 120, 8461–8477 (2015).

    Article  Google Scholar 

  12. Ide, S., Shiomi, K., Mochizuki, K., Tonegawa, T. & Kimura, G. Split Philippine Sea plate beneath Japan. Geophys. Res. Lett. 37, L21304, https://doi.org/10.1029/2010GL044585 (2010).

  13. Shelly, D. R., Beroza, G. C., Ide, S. & Nakamula, S. Low-frequency earthquakes in Shikoku, Japan, and their relationship to episodic tremor and slip. Nature 442, 188–191 (2006).

    Article  Google Scholar 

  14. Wang, K., Wada, I. & Ishikawa, Y. Stresses in the subducting slab beneath southwest Japan and relation with plate geometry, tectonic forces, slab dehydration and damaging earthquakes. J. Geophys. Res. Solid Earth 109, B08304, https://doi.org/10.1029/2003JB002888 (2004).

  15. Committee, Earthquake Research. Long‐Term Evaluation of Earthquakes in the Nankai Trough (in Japanese) http://www.jishin.go.jp/main/chousa/01sep_nankai/index.htm (2001).

  16. Arnulf, A. F., Singh, S. C., Harding, A. J., Kent, G. M. & Crawford, W. Strong seismic heterogeneity in layer 2A near hydrothermal vents at the Mid-Atlantic Ridge. Geophys. Res. Lett. 38, L13320 (2011).

    Article  Google Scholar 

  17. Arnulf, A. F., Harding, A. J., Kent, G. & Wilcock, W. Structure, seismicity and accretionary processes at the hot spot-influenced axial seamount on the Juan de Fuca ridge. J. Geophys. Res. Solid Earth 123, 4618–4646 (2018).

    Article  Google Scholar 

  18. Honda, R. & Kono, Y. Buried large block revealed by gravity anomalies in the Tonankai and Nankai earthquakes regions, southwestern Japan. Earth Planets Space 57, e1–e4 (2005).

    Article  Google Scholar 

  19. Yamamoto, Y. et al. Seismic structure off the Kii Peninsula, Japan, deduced from passive-and active-source seismographic data. Earth Planet. Sci. Lett. 461, 163–175 (2017).

    Article  Google Scholar 

  20. Bassett, D. & Watts, A. B. Gravity anomalies, crustal structure and seismicity at subduction zones: 2. Interrelationships between fore-arc structure and seismogenic behavior. Geochem. Geophys. Geosyst. 16, 1541–1576 (2015).

    Article  Google Scholar 

  21. Kimura, G., Hashimoto, Y., Kitamura, Y., Yamaguchi, A. & Koge, H. Middle Miocene swift migration of the TTT triple junction and rapid crustal growth in southwest Japan. Rev. Tecton. 33, 1219–1238 (2014).

    Article  Google Scholar 

  22. Kikuchi, M., Nakamura, M. & Yoshikawa, K. Source rupture processes of the 1944 Tonankai earthquake and the 1945 Mikawa earthquake derived from low-gain seismograms. Earth Planets Space 55, 159–172 (2003).

    Article  Google Scholar 

  23. Hashimoto, M. & Kikuchi, M. Rupture process of the 1946 Nankaido earthquake revealed from seismograms. Chikyu 24, 16–20 (1999).

    Google Scholar 

  24. Kodaira, S., Takahashi, N., Nakanishi, A., Miura, S. & Kaneda, Y. Subducted seamount imaged in the rupture zone of the 1946 Nankaido earthquake. Science 289, 104–106 (2000).

    Article  Google Scholar 

  25. Baba, T. & Cummins, P. R. Contiguous rupture areas of two Nankai Trough earthquakes revealed by high‐resolution tsunami waveform inversion. Geophys. Res. Lett. 32, L08305, https://doi.org/10.1029/2004GL022320 (2005).

  26. Tsuji, T., Kodaira, S., Ashi, J. & Park, J.-O. Widely distributed thrust and strike-slip faults within subducting oceanic crust in the Nankai Trough off the Kii Peninsula, Japan. Tectonophysics 600, 52–62 (2013).

    Article  Google Scholar 

  27. Nakanishi, A. et al. Detailed structural image around splay‐fault branching in the Nankai subduction seismogenic zone: results from a high‐density ocean bottom seismic survey. J. Geophys. Res. Solid Earth 113, B03105, https://doi.org/10.1029/2007JB004974 (2008).

  28. Ivandic, M., Grevemeyer, I., Berhorst, A., Flueh, E. R. & McIntosh, K. Impact of bending related faulting on the seismic properties of the incoming oceanic plate offshore Nicaragua. J. Geophys. Res. Solid Earth 113, B05410 (2008).

    Article  Google Scholar 

  29. Van Avendonk, H. J., Holbrook, W. S., Lizarralde, D. & Denyer, P. Structure and serpentinization of the subducting Cocos plate offshore Nicaragua and Costa Rica. Geochem. Geophys. Geosyst. 12, Q06009, https://doi.org/10.1029/2011GC003592 (2011).

  30. Contreras‐Reyes, E., Grevemeyer, I., Flueh, E. R. & Reichert, C. Upper lithospheric structure of the subduction zone offshore of southern Arauco peninsula, Chile, at ~38° S. J. Geophys. Res. Solid Earth 113, B07303, https://doi.org/10.1029/2007JB005569 (2008).

  31. Ji, Y., Yoshioka, S. & Matsumoto, T. Three‐dimensional numerical modeling of temperature and mantle flow fields associated with subduction of the Philippine Sea plate, southwest Japan. J. Geophys. Res. Solid Earth 121, 4458–4482 (2016).

    Article  Google Scholar 

  32. Peacock, S. Are the lower planes of double seismic zones caused by serpentine dehydration in subducting oceanic mantle? Geology 29, 299–302 (2001).

    Article  Google Scholar 

  33. Ferrand, T. P. Seismicity and mineral destabilizations in the subducting mantle up to 6 GPa, 200-km depth. Lithos 334–335, 205–230 (2019).

    Article  Google Scholar 

  34. Christensen, N. I. Serpentinites, peridotites and seismology. Int. Geol. Rev. 46, 795–816 (2004).

    Article  Google Scholar 

  35. Hirano, N. et al. Volcanism in response to plate flexure. Science 313, 1426–1428 (2006).

    Article  Google Scholar 

  36. Sibson, R., Moore, J. M. M. & Rankin, A. Seismic pumping—a hydrothermal fluid transport mechanism. J. Geol. Soc. 131, 653–659 (1975).

    Article  Google Scholar 

  37. Korenaga, J. Thermal cracking and the deep hydration of oceanic lithosphere: a key to the generation of plate tectonics? J. Geophys. Res. Solid Earth 112, B05408, https://doi.org/10.1029/2006JB004502 (2007).

  38. Phipps Morgan, J. & Holtzman, B. K. Vug waves: a mechanism for coupled rock deformation and fluid migration. Geochem. Geophys. Geosyst. 6, Q08002, https://doi.org/10.1029/2004GC000818 (2005).

  39. Faccenda, M., Gerya, T. V. & Burlini, L. Deep slab hydration induced by bending-related variations in tectonic pressure. Nat. Geosci. 2, 790–793 (2009).

    Article  Google Scholar 

  40. Korenaga, J. On the extent of mantle hydration caused by plate bending. Earth Planet. Sci. Lett. 457, 1–9 (2017).

    Article  Google Scholar 

  41. Muller, R. D., Sdrolias, M., Gaina, C. & Roest, W. R. Age, spreading rates and spreading asymmetry of the world’s ocean crust. Geochem. Geophys. Geosyst. 9, Q04006, https://doi.org/10.1029/2007GC001743 (2008).

  42. Sigloch, K. & Nolet, G. Measuring finite-frequency body-wave amplitudes and traveltimes. Geophys. J. Int. 167, 271–287 (2006).

    Article  Google Scholar 

  43. Arnulf, A. F., Harding, A. J., Kent, G., Singh, S. C. & Crawford, W. C. Constraints on the shallow velocity structure of the Lucky Strike Volcano, Mid-Atlantic Ridge, from downward continued multichannel streamer data. J. Geophys. Res. Solid Earth 119,1119–1144, https://doi.org/10.1002/2013JB010500 (2014).

  44. Van Avendonk, H., Harding, A., Orcutt, J. & McClain, J. Contrast in crustal structure across the Clipperton transform fault from travel time tomography. J. Geophys. Res. Solid Earth 106, 10961–10981 (2001).

    Article  Google Scholar 

  45. Moser, T. Shortest path calculation of seismic rays. Geophysics 56, 59–67 (1991).

    Article  Google Scholar 

  46. Arnulf, F. A., Basset, D. & Harding, A. Three-Dimensional P-Wave Velocity Structure of the Nankai Trough Region, SW Japan (investigators: Adrien F. Arnulf, Dan Bassett, Alistair Harding) (IEDA, 2020); https://doi.org/10.26022/IEDA/329655

Download references

Acknowledgements

This study has been supported by US National Science Foundation grants nos. 1657480 (UTIG for A.F.A.) and 1658010 (SIO for D.B. and A.J.H.). In its early stage, this study, as well as A.F.A. and D.B., were supported by the Green Foundation at the Institute for Geophysics, Scripps Institution of Oceanography, University of California San Diego. We thank the Japan Coastguard for providing seismic data along lines KPr1 and KPr2. Seismic processing and data analysis were conducted using GLOBE Claritas, MatLab and Python. We thank D. Shillington, S. Peacock and L. McNeill for their constructive comments, which greatly improved the quality and impact of this paper.

Author information

Authors and Affiliations

Authors

Contributions

A.F.A. and D.B. conceived the study and analysed the data. S.K., A.N., A.J.H. and G.M. contributed data from Nankai Trough and discussed the results. A.F.A. and A.J.H. wrote the algorithms to process the JMA Earthquake Catalogue data and to conduct the tomographic inversion. A.F.A. and D.B. wrote the manuscript. All authors discussed the results and commented on the manuscript.

Corresponding author

Correspondence to Adrien F. Arnulf.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Geoscience thanks Lisa McNeill, Simon Peacock and Donna Shillington for their contribution(s) to the peer review of this work. Primary Handling Editor: Rebecca Neely, in collaboration with the Nature Geoscience team.

Additional information

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

Extended data

Extended Data Fig. 1 Example of Onshore-Offshore data from Nankai.

a Map showing the location of offshore seismic profiles and onshore seismometers. Dark black lines are seismic lines from the KR10-11 OBS survey lead by JAMSTEC. Red circles and blue hexagons mark Hi-Net and F-Net seismic stations, respectively. Red triangles mark the location of F-Net stations KMT and UMJ. b Data recorded at KMT station from offshore shots fired between A and A’ along seismic profile KR10-11-TK01. c Data recorded at UMJ station from offshore shots fired between B and B’ along seismic profile KR10-11-TK03. Black and green dots mark observed and calculated travel-times through our final 3-D P-wave velocity model.

Extended Data Fig. 2 Azimuthal ray-path density maps provided by the three individual datasets included in our tomographic inversion.

a Azimuthal ray-path density provided by the onshore-offshore seismic data. Yellow circles mark the 188 onshore seismographs for which offshore seismic arrivals have been interpreted. b Azimuthal ray-path density provided by the OBS seismic data. Yellow circles mark the 537 OBSs for which seismic arrivals have been interpreted. c Azimuthal ray-path density provided by naturally occurring earthquake arrivals provided by the JMA earthquake catalogue. Yellow circles mark the 341 Hi-Net stations for which arrival times have been used.

Extended Data Fig. 3 Details about the complete tomographic travel-time dataset used in this study.

a Azimuthal ray-path density map provided by the combined seismic dataset (onshore-offshore, OBS and earthquake data). The individual datasets are presented in Extended Data Fig. 2. b Histogram showing the distribution of source-receiver offsets for the combined (gray) and all individual seismic datasets (OBS: blue, Onshore-Offshore: orange, JMA earthquakes: pink). c Table summarizing the number of arrival times, stations and sources for the seismic datasets analyzed in this study.

Extended Data Fig. 4 Azimuthal ray-path density maps provided by OBS surveys at Nankai Trough.

a JAMSTEC KR08-16 3-D OBS survey and Japan Coast Guard 2-D KPr OBS lines offshore Kyushu. b JAMSTEC KR09-14 3-D OBS survey offshore Shikoku. c JAMSTEC KR10-11 3-D OBS survey offshore Kii channel. d JAMSTEC KR11-09 3-D OBS survey offshore Kii Peninsula. e JAMSTEC KR12-12 3-D OBS survey offshore Kii Peninsula and Shizuoka. Yellow circles mark the location of all interpreted OBSs.

Extended Data Fig. 5 Japan Meteorological Agency (JMA) earthquake data included in our 3-D tomographic inversion.

a Map showing the location of the 156,842 JMA earthquakes that have been included in the inversion. Color and size of events as a function of magnitude. b Histogram showing the distribution of travel-time residuals from earthquake arrivals through our final 3-D P-wave seismic velocity model. Histograms showing the distribution of earthquake magnitude c, number of arrivals per earthquake d, earthquake depth e, earthquake X-position f, and earthquake Y-position g for all events included in our tomographic inversion. Inset tables present important statistical quantities (arithmetic mean, standard deviation, minimum value, 25th percentile, 50th percentile (that is, median), 75th percentile and maximum value).

Extended Data Fig. 6 Kernel density plots of travel-time residual misfits for the combined data and various data subsets.

Inset tables provide important statistical quantities (arithmetic mean, standard deviation, minimum value, 25th percentile, 50th percentile (that is, median), 75th percentile and maximum value). Colors correspond to the various datasets presented in Extended Data Fig. 3b,c. a Travel-time residual misfits for the combined data. b Travel-time residual misfits as a function of source type (active source shots: green, or earthquakes: pink). c Travel-time residual misfits as a function of data type (OBS data: blue, onshore-offshore data: orange, or earthquake data: pink). d Onshore-offshore data travel-time residual misfits as a function of station type (Hi-Net stations: red, F-Net stations: yellow). Quartiles are marked by fine dashed black lines on each violin distribution plot. Travel-time data are well fitted irrespectively of source type, data type or station type. e Travel-time residual misfits as a function of source-receiver offset groups. Quartiles are marked by fine dashed black lines on each violin distribution plot. Count of arrival times within each group (C) and corresponding percentage of the entire dataset (P) are indicated on top of each violin distribution plot. Travel-time data are well fitted irrespectively of source-receiver offsets.

Extended Data Fig. 7 Parallel categories diagram showing the distribution of travel-time residual misfits as a function of source type (1st column), station type (2nd column), location of seismic station within a given Japanese prefecture or offshore (3rd column), source-receiver offset (4th column) and rounded travel-time residuals to the nearest 0.1 second (5th column).

As an example, rounded travel-time residuals in the ‘0 s’ category include all arrivals with an error included in the range: -0.05 s to 0.05 s. Count (C) of arrival times within a given category is labelled in black when possible. Corresponding percentage of the entire dataset is label in blue near each category. Travel time residuals within the range -0.150 s to 0.150 s (categories -0.1, 0 and 0.1 s in column number 5) account for 79.7 % of all arrivals. The travel time tomography data used in this survey are well fitted irrespectively of source type, station type, station location and source-receiver offset.

Extended Data Fig. 8 Results of Nankai checkerboard resolution test. The perturbation imposed was a 50 km x 50 km x 10 km sinusoidal velocity perturbation with maximum amplitude of ±12.5%.

a-b Depth slices showing checkerboard recovery at 25 km and 35 km depth. Black dashed line marks Nankai Trough. Solid black lines mark the position of velocity transects shown in the main manuscript (Figs. 2b-c and 3). c-d Y-slice at Y = -25 km and Y = -75 km showing checkerboard recovery along-strike. e-f X-slices at X = 125 km and 175 km, showing checkerboard recovery down dip. Dashed gray line marks the top of the subducting Philippine Sea Plate. Dotted black lines on vertical sections (c-f) mark the location of the velocity slice shown in Fig. 4 in the main manuscript. Thick black contour lines on all panels mark the isovelocity contour 7.5 km/s and help outline the extent of the low-velocity anomaly within the upper-mantle of the down-going Philippine Sea Plate. Strong recovery of the checkerboard pattern indicates that our velocity model is homogenously well constrained for this given perturbation size and that resolution is maintained in most places to >30 km depth below the top of the subducting Philippine Sea Plate, and to ~25-30 km depth below the Japanese archipelago.

Extended Data Fig. 9 Results of Nankai checkerboard resolution test. The perturbation imposed was a 25 km x 25 km x 5 km sinusoidal velocity perturbation with maximum amplitude of ±12.5%.

a Depth slice showing checkerboard recovery at 27.5 km depth. Black dashed line marks Nankai Trough. Solid black lines mark the position of velocity transects shown in the main manuscript (Figs. 2b-c and 3). b-d Y-slices at Y = -37.5 km, -87.5 km and -112.5 km showing checkerboard recovery along-strike. e-f X-slices at X = 162.5 km and 187.5 km showing checkerboard recovery down dip. Dashed gray line marks the top of the subducting Philippine Sea Plate. Thick black contour lines on all panels mark the isovelocity contour 7.5 km/s and help outline the extent of the low-velocity anomaly within the upper-mantle of the down-going Philippine Sea Plate. Strong recovery of this checkerboard pattern indicates that our velocity model is regionally well constrained above 30 km depth, particularly offshore where ray density is highest. Locally, resolution is maintained to 50 km depth.

Extended Data Fig. 10 3-D perspective views showing the spatial extent of mantle hydration and the geometry of the subducting Philippine Sea Plate in relation to the position of the Kumano Pluton.

a P-wave velocity structure extracted at ~15 km below the Moho of the subducting Philippine Sea Plate. Mesh surface shows the plate interface. Brown surface marks the 5.75 km/s isovelocity surface and reveals the seaward extent of the Kumano Pluton. Note the spatial extent of upper mantle hydration and the geometry of subducting Philippine Sea Plate relative to the position of the Kumano Pluton. Red line marks Nankai trough. b P-wave velocity structure extracted at ~15 km below the Moho of the subducting Philippine Sea Plate, as well as along the Kumano and NT02 seismic transects. Solid black lines on the two vertical transects show the location of sub-crustal reflectors identified from ocean bottom seismic data9,27. These sub-Moho reflectors are particular to the offshore region of the Kii Peninsula and appear well correlated with the top of the upper-mantle low-velocity body. Brown surface marks the 5.75 km/s isovelocity surface above the plate interface (mesh surface). On both panels, residual free-air gravity anomaly20 is draped on top of the topography. High positive anomalies (>100 mGal) are well correlated with high wavespeeds in the overthrusting plate and reveal the spatial extent of the Kumano pluton (heavy dashed black line). Black arrow gives the relative motion of the Philippine Sea Plate with respect to Eurasia8.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Arnulf, A.F., Bassett, D., Harding, A.J. et al. Upper-plate controls on subduction zone geometry, hydration and earthquake behaviour. Nat. Geosci. 15, 143–148 (2022). https://doi.org/10.1038/s41561-021-00879-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41561-021-00879-x

Further reading

Search

Quick links

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