Turn off MathJax
Article Contents

Yiming Luo, Jian Lin, Fan Zhang, Meng Wei. Spreading rate dependence of morphological characteristics in global oceanic transform faults[J]. Acta Oceanologica Sinica. doi: 10.1007/s13131-021-1722-5
Citation: Yiming Luo, Jian Lin, Fan Zhang, Meng Wei. Spreading rate dependence of morphological characteristics in global oceanic transform faults[J]. Acta Oceanologica Sinica. doi: 10.1007/s13131-021-1722-5

Spreading rate dependence of morphological characteristics in global oceanic transform faults

doi: 10.1007/s13131-021-1722-5
Funds:  The Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) under contract No. GML2019ZD0205; National Science Foundation of China under contract Nos 41976064, 41890813, 41976066, 91958211, 4194200104, and 41706056; the China Scholarship Council; the Chinese Academy of Sciences under contract Nos Y4SL021001, QYZDY-SSW-DQC005, and 133244KYSB20180029; the National Key Research and Development Program of China under contract Nos 2018YFC0309800 and 2018YFC0310105; and the China Ocean Mineral Resources Research and Development Association under contract No. DY135-S2-1-04.
More Information
  • Corresponding author: E-mail: zhangfan@scsio.ac.cn
  • Received Date: 2020-02-28
  • Rev Recd Date: 2020-06-29
  • We quantified the systematic variations in global transform fault morphology, revealing a first-order dependence on the spreading rate: (1) The average age offset of both the full transform and transform sub-segments decrease with increasing spreading rate. (2) The average depth of both the transform valley and adjacent ridges are smaller in the fast compared to the slow systems, reflecting possibly density anomalies associated with warmer mantle at the fast systems and rifting at the slow ridges. However, the average depth difference between the transform valley and adjacent ridges is relatively constant from the fast to slow systems. (3) The nodal basin at a ridge-transform intersection is deeper and dominant at the ultraslow and slow systems, possibly reflecting a lower magma supply and stronger viscous resistance to mantle upwelling near a colder transform wall. In contrast, the nodal high, is most prominent in the fast, intermediate, and hotspot-influenced systems, where robust axial volcanic ridges extend toward the ridge-transform intersection. (4) Statistically, the average transform valley is wider at a transform system of larger age offset, reflecting thicker deforming plates flanking the transform fault. (5) The maximum magnitude of the transform earthquakes increases with age offset owing to an increase in the seismogenic area. Individual transform faults also exhibit significant anomalies owing to the complex local tectonic and magmatic processes.
  • †These authors contributed equally to this work
  • 加载中
  • [1] Baines A G, Cheadle M J, Dick H J B, et al. 2003. Mechanism for generating the anomalous uplift of oceanic core complexes: Atlantis Bank, Southwest Indian Ridge. Geology, 31(12): 1105–1108. doi:  10.1130/G19829.1
    [2] Behn M D, Lin J, Zuber M T. 2002. Evidence for weak oceanic transform faults. Geophysical Research Letters, 29(24): 2207. doi:  10.1029/2002GL015612
    [3] Bird P, Kagan Y Y, Jackson D D. 2002. Plate tectonics and earthquake potential of spreading ridges and oceanic transform faults. In: Stein S, Freymueller J T, eds. Plate Boundary Zones. Washington, DC: Geodynamics Series, 203–218
    [4] Boettcher M S, Jordan T H. 2004. Earthquake scaling relations for mid-ocean ridge transform faults. Journal of Geophysical Research: Solid Earth, 109(B12): B12302. doi:  10.1029/2004JB003110
    [5] Bonatti E, Ligi M, Gasperini L, et al. 1994. Transform migration and vertical tectonics at the Romanche fracture zone, equatorial Atlantic. Journal of Geophysical Research: Solid Earth, 99(B11): 21779–21802. doi:  10.1029/94JB01178
    [6] Chen YS, Morgan W J. 1990. Rift valley/no rift valley transition at mid‐ocean ridges. Journal of Geophysical Research: Solid Earth, 95(B11): 17571–17581. doi:  10.1029/JB095iB11p17571
    [7] Croon M B, Cande S C, Stock J M. 2008. Revised pacific‐Antarctic plate motions and geophysics of the Menard fracture zone. Geochemistry, Geophysics, Geosystems, 9(7): Q07001. doi:  10.1029/2008GC002019
    [8] Dick H, Lin J, Schouten H. 2003. An ultraslow-spreading class of ocean ridge. Nature, 426(6965): 405–412. doi:  10.1038/nature02128
    [9] Dziewonski A M, Anderson D L. 1981. Preliminary reference Earth model. Physics of the Earth and Planetary Interiors, 25(4): 297–356. doi:  10.1016/0031-9201(81)90046-7
    [10] Ekström G, Nettles M, Dziewoński A M. 2012. The global CMT project 2004–2010 : Centroid-moment tensors for 13, 017 earthquakes. Physics of the Earth and Planetary Interiors, 200–201: 1–9. doi:  10.1016/j.pepi.2012.04.002
    [11] Embley R W, Wilson D S. 1992. Morphology of the Blanco transform fault zone-NE Pacific: implications for its tectonic evolution. Marine Geophysical Researches, 14(1): 25–45. doi:  10.1007/BF01674064
    [12] Fornari D J, Gallo D G, Edwards M H, et al. 1989. Structure and topography of the Siqueiros transform fault system: Evidence for the development of intra-transform spreading centers. Marine Geophysical Researches, 11(4): 263–299. doi:  10.1007/BF00282579
    [13] Fox P J, Gallo D G. 1984. A tectonic model for ridge-transform-ridge plate boundaries: Implications for the structure of oceanic lithosphere. Tectonophysics, 104(3–4): 205–242. doi:  10.1016/0040-1951(84)90124-0
    [14] Georgen J E, Lin J, Dick H J B. 2001. Evidence from gravity anomalies for interactions of the Marion and Bouvet hotspots with the Southwest Indian Ridge: Effects of transform offsets. Earth and Planetary Science Letters, 187(3–4): 283–300. doi:  10.1016/S0012-821X(01)00293-X
    [15] Gregg P M, Lin J, Smith D K. 2006. Segmentation of transform systems on the East Pacific Rise: Implications for earthquake processes at fast-slipping oceanic transform faults. Geology, 34(4): 289–292. doi:  10.1130/G22212.1
    [16] Gregg P M, Lin J, Behn M D, et al. 2007. Spreading rate dependence of gravity anomalies along oceanic transform faults. Nature, 448(7150): 183–187. doi:  10.1038/nature05962
    [17] Kanamori H. 1977. The energy release in great earthquakes. Journal of Geophysical Research, 82(20): 2981–2987. doi:  10.1029/JB082i020p02981
    [18] Karson J A, Dick H J B. 1983. Tectonics of ridge-transform intersections at the Kane fracture zone. Marine Geophysical Research, 6(1): 51–98. doi:  10.1007/BF00300398
    [19] Kreemer C, Haines J, Holt W E, et al. 2000. On the determination of a global strain rate model. Earth, Planets and Space, 52(10): 765–770. doi:  10.1186/BF03352279
    [20] thickness. Marine Geophysical Researches, 10(3): 205–232, doi: 10.1007/BF00310065
    [21] Ligi M, Bonatti E, Gasperini L, et al. 2002. Oceanic broad multifault transform plate boundaries. Geology, 30(1): 11–14. doi:  10.1130/0091-7613(2002)030<0011:OBMTPB>2.0.CO;2
    [22] Lin J, Parmentier E M. 1989. Mechanisms of lithospheric extension at mid-ocean ridges. Geophysical Journal International, 96(1): 1–22. doi:  10.1111/j.1365-246X.1989.tb05246.x
    [23] Lin J, Purdy G M, Schouten H, et al. 1990. Evidence from gravity data for focusedmagmatic accretion along the Mid-Atlantic Ridge. Nature, 344: 627–632. doi:  10.1038/344627a0
    [24] Lin J, Morgan J P. 1992. The spreading rate dependence of three‐dimensional mid‐ocean ridge gravity structure. Geophysical Research Letters, 19(1): 13–16. doi:  10.1029/91GL03041
    [25] Livermore R A, Tomlinson J S, Woollett R W. 1991. Unusual sea-floor fabric near the Bullard fracture zone imaged by GLORIA sidescan sonar. Nature, 353(6340): 158–161. doi:  10.1038/353158a0
    [26] Lonsdale P. 1994. Structural geomorphology of the Eltanin fault system and adjacent transform faults of the Pacific-Antarctic plate boundary. Marine Geophysical Researches, 16(2): 105–143. doi:  10.1007/BF01224756
    [27] Macdonald K C. 1982. Mid-ocean ridges: Fine scale tectonic, volcanic and hydrothermal processes within the plate boundary zone. Annual Review of Earth and Planetary Sciences, 10: 155–190. doi:  10.1146/annurev.ea.10.050182.001103
    [28] Magde L S, Detrick R S. 1995. Crustal and upper mantle contribution to the axial gravity anomaly at the southern East Pacific Rise. Journal of Geophysical Research: Solid Earth, 100(B3): 3747–3766. doi:  10.1029/94JB02869
    [29] Maia M. 2019. Topographic and morphologic evidences of deformation at oceanic transform faults: Far-field and local-field stresses. In: Duarte J C, ed. Transform Plate Boundaries and Fracture Zones. Amsterdam: Elsevier, 61–87, doi: 10.1016/B978-0-12-812064-4.00003-7
    [30] Morgan J P, Parmentier E M. 1984. Lithospheric stress near a ridge‐transform intersection. Geophysical Research Letters, 11(2): 113–116. doi:  10.1029/GL011i002p00113
    [31] Perfit M R, Fornari D J, Ridley W I, et al. 1996. Recent volcanism in the Siqueiros transform fault: Picritic basalts and implications for MORB magma genesis. Earth and Planetary Science Letters, 141(1–4): 91–108. doi:  10.1016/0012-821X(96)00052-0
    [32] Pockalny R A, Detrick R S, Fox P J. 1988. Morphology and tectonics of the Kane transform from Sea Beam bathymetry data. Journal of Geophysical Research: Solid Earth, 93(B4): 3179–3193. doi:  10.1029/JB093iB04p03179
    [33] Pockalny R A, Fox P J, Fornari D J, et al. 1997. Tectonic reconstruction of the Clipperton and Siqueiros Fracture Zones: Evidence and consequences of plate motion change for the last 3 Myr. Journal of Geophysical Research: Solid Earth, 102(B2): 3167–3181. doi:  10.1029/96JB03391
    [34] Roland E, Behn M D, Hirth G. 2010. Thermal‐mechanical behavior of oceanic transform faults: Implications for the spatial distribution of seismicity. Geochemistry, Geophysics, Geosystems, 11(7): Q07001. doi:  10.1029/2010GC003034
    [35] Royden L H, Horváth F, Burchfiel B C. 1982. Transform faulting, extension, and subduction in the Carpathian Pannonian region. GSA Bulletin, 93(8): 717–725. doi:  10.1130/0016-7606(1982)93<717:TFEASI>2.0.CO;2
    [36] Sandwell D T, Müller R D, Smith W H F, et al. 2014. New global marine gravity model from CryoSat-2 and Jason-1 reveals buried tectonic structure. Science, 346(6205): 65–67. doi:  10.1126/science.1258213
    [37] Sclater J G, Grindlay N R, Madsen J A, et al. 2005. Tectonic interpretation of the Andrew Bain transform fault: Southwest Indian Ocean. Geochemistry, Geophysics, Geosystems, 6(9): Q09K10. doi:  10.1029/2005GC000951
    [38] Searle R C. 1983. Multiple, closely spaced transform faults in fast-slipping fracture zones. Geology, 11(10): 607–610. doi:  10.1130/0091-7613(1983)11<607:MCSTFI>2.0.CO;2
    [39] Searle R C, Thomas M V, Jones E J W. 1994. Morphology and tectonics of the Romanche Transform and its environs. Marine Geophysical Researches, 16(6): 427–453. doi:  10.1007/bf01270518
    [40] Sempéré J C, Lin J, Brown H S, et al. 1993. Segmentation and morphotectonic variations along a slow-spreading center: The Mid-Atlantic Ridge (24°00′ N–30°40′N). Marine Geophysical Researches, 15(3): 153–200. doi:  10.1007/BF01204232
    [41] Shaw W J, Lin J. 1996. Models of ocean ridge lithospheric deformation: Dependence on crustal thickness, spreading rate, and segmentation. Journal of Geophysical Research: Solid Earth, 101(B8): 17977–17993. doi:  10.1029/96JB00949
    [42] Sleep N H. 1969. Sensitivity of heat flow and gravity to the mechanism of sea‐floor spreading. Journal of Geophysical Research, 74(2): 542–549. doi:  10.1029/JB074i002p00542
    [43] Sleep N H, Biehler S. 1970. Topography and tectonics at the intersections of fracture zones with central rifts. Journal of Geophysical Research, 75(14): 2748–2752. doi:  10.1029/JB075i014p02748
    [44] Tapponnier P, Francheteau J. 1978. Necking of the lithosphere and the mechanics of slowly accreting plate boundaries. Journal of Geophysical Research: Solid Earth, 83(B8): 3955–3970. doi:  10.1029/JB083iB08p03955
    [45] Tozer B, Sandwell D T, Smith W H F, et al. 2019. Global bathymetry and topography at 15 arc sec: SRTM15+. Earth and Space Science, 6(10): 1847–1864. doi:  10.1029/2019EA000658
    [46] Tucholke B E, Schouten H. 1988. Kane fracture zone. Marine Geophysical Researches, 10(1–2): 1–39. doi:  10.1007/BF02424659
    [47] Tucholke B E, Lin J. 1994. A geological model for the structure of ridge segments in slow spreading ocean crust. Journal of Geophysical Research: Solid Earth, 99(B6): 11937–11958. doi:  10.1029/94JB00338
    [48] Turcotte D L, Schubert G. 2014. Geodynamics. 3rd ed. Cambridge, United Kingdom: Cambridge University Press, 636
    [49] Wei M. 2019. Seismic behavior on oceanic transform faults at the East Pacific Rise. In Duarte J C, ed. Transform Plate Boundaries and Fracture Zones. Amsterdam: Elsevier, 119–143, doi: 10.1016/B978-0-12-812064-4.00006-2
    [50] Wilson J T. 1965. A new class of faults and their bearing on continental drift. Nature, 207(4995): 343–347. doi:  10.1038/207343a0
    [51] Wolfson-Schwehr M. 2015. The relationship between oceanic transform fault segmentation, seismicity, and thermal structure [dissertation]. Durham: University of New Hampshire, https://scholars.unh.edu/dissertation/2233
    [52] Wolfson-Schwehr M, Boettcher M S. 2019. Global characteristics of oceanic transform fault structure and seismicity. In: Duarte J C, ed. Transform Plate Boundaries and Fracture Zones. Amsterdam: Elsevier, 21–59, doi: 10.1016/B978-0-12-812064-4.00002-5
    [53] Zhang T, Lin J, Gao J Y. 2020. Asymmetric crustal structure of the ultraslow-spreading Mohns Ridge. International Geology Review, 62(5): 568–584. doi:  10.1080/00206814.2019.1627586
  • 加载中
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Figures(17)  / Tables(6)

Article Metrics

Article views(92) PDF downloads(4) Cited by()

Related
Proportional views

Spreading rate dependence of morphological characteristics in global oceanic transform faults

doi: 10.1007/s13131-021-1722-5
Funds:  The Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) under contract No. GML2019ZD0205; National Science Foundation of China under contract Nos 41976064, 41890813, 41976066, 91958211, 4194200104, and 41706056; the China Scholarship Council; the Chinese Academy of Sciences under contract Nos Y4SL021001, QYZDY-SSW-DQC005, and 133244KYSB20180029; the National Key Research and Development Program of China under contract Nos 2018YFC0309800 and 2018YFC0310105; and the China Ocean Mineral Resources Research and Development Association under contract No. DY135-S2-1-04.

Abstract: We quantified the systematic variations in global transform fault morphology, revealing a first-order dependence on the spreading rate: (1) The average age offset of both the full transform and transform sub-segments decrease with increasing spreading rate. (2) The average depth of both the transform valley and adjacent ridges are smaller in the fast compared to the slow systems, reflecting possibly density anomalies associated with warmer mantle at the fast systems and rifting at the slow ridges. However, the average depth difference between the transform valley and adjacent ridges is relatively constant from the fast to slow systems. (3) The nodal basin at a ridge-transform intersection is deeper and dominant at the ultraslow and slow systems, possibly reflecting a lower magma supply and stronger viscous resistance to mantle upwelling near a colder transform wall. In contrast, the nodal high, is most prominent in the fast, intermediate, and hotspot-influenced systems, where robust axial volcanic ridges extend toward the ridge-transform intersection. (4) Statistically, the average transform valley is wider at a transform system of larger age offset, reflecting thicker deforming plates flanking the transform fault. (5) The maximum magnitude of the transform earthquakes increases with age offset owing to an increase in the seismogenic area. Individual transform faults also exhibit significant anomalies owing to the complex local tectonic and magmatic processes.

†These authors contributed equally to this work
Yiming Luo, Jian Lin, Fan Zhang, Meng Wei. Spreading rate dependence of morphological characteristics in global oceanic transform faults[J]. Acta Oceanologica Sinica. doi: 10.1007/s13131-021-1722-5
Citation: Yiming Luo, Jian Lin, Fan Zhang, Meng Wei. Spreading rate dependence of morphological characteristics in global oceanic transform faults[J]. Acta Oceanologica Sinica. doi: 10.1007/s13131-021-1722-5
    • A transform fault is a major plate boundary with predominantly strike-slip motion between two adjacent plates (Wilson, 1965). Oceanic transform faults exhibit significant variability in length and age offset. Globally, the full length of a transform fault ranges from as small as 26 km (Herron transform fault) to as large as 1 099 km (Chile transform fault), and the full transform age offset is in the range of 0.69 Ma (Yaquina transform fault) to 105.8 Ma (Andrew Bain transform fault) (Boettcher and Jordan, 2004; Wolfson-Schwehr, 2015).

      High-resolution bathymetry data of the last four decades have revealed the morphological complexity of oceanic transform systems. Transform faults are commonly composed of multiple sub-segments, separated by fault steps, intra-transform extension basins, or intra-transform spreading centers (e.g., Searle, 1983; Fornari et al., 1989; Ligi et al., 2002; Gregg et al., 2006). This complex transform morphology might be caused by a range of tectonic and magmatic processes in the ridge-transform system (e.g., Royden et al., 1982; Lin et al., 1990; Lin and Morgan, 1992; Sempéré et al., 1993; Perfit et al., 1996; Behn et al., 2002; Gregg et al., 2006, 2007; Wolfson-Schwehr, 2015; Wei, 2019; Wolfson-Schwehr and Boettcher, 2019).

      Transform faults exhibit systematic variations in morphology that contain important information regarding the dynamics of the transforms and adjacent ridges. Almost all transform faults are associated with transform-parallel median valleys, some of which are composed of multiple topographical lows (e.g., Macdonald, 1982; Searle, 1983; Fox and Gallo, 1984; Tucholke and Schouten, 1988; Bonatti et al., 1994; Tucholke and Lin, 1994; Maia, 2019). Transform-parallel topographical highs have been observed in some transform systems either inside the transform valley (forming a median ridge) or flanking the transform fault (forming a transverse range) (Pockalny et al., 1988; 1997; Embley and Wilson, 1992; Maia, 2019). In many transform systems, the greatest depth is located near the ridge-transform intersection, forming noticeable nodal basins (e.g., Karson and Dick, 1983; Fox and Gallo, 1984; Pockalny et al., 1988). On the other hand, distinctive axial volcanic ridges are often observed at the adjacent spreading center axis and at the ridge-transform intersection (i.e., nodal highs). Nodal highs are especially well-developed in systems that have a relatively robust magma supply, such as fast- and intermediate-spreading or hotspot-influenced systems (e.g., Macdonald, 1982; Morgan and Parmentier, 1984).

      The majority of studies to date have focused on investigating the morphology of individual transform faults, while quantitative analysis of the global variability in transform faults and the analysis of the dependence on the spreading rate remain limited. This lack of quantitative analyses hinders our understanding of ridge-transform dynamics.

      In the present study, we quantified the global variability in transform fault morphology and investigated their dependence on the spreading rate. The investigated transform faults included examples from the ultraslow (full spreading rate<20 mm/yr), slow (20–50 mm/yr), intermediate (50–80 mm/yr), and fast (>80 mm/yr) systems (Macdonald, 1982; Dick et al., 2003) (Figs. 1 and A1). The investigated systems included those at the East Pacific Rise (EPR), Pacific-Antarctic Ridge (PAR), Juan de Fuca Ridge (JDF), Southeast Indian Ridge (SEIR), Chile Rise (CR), Aden Ridge (ADEN), Central Indian Ridge (CIR), Mid-Atlantic Ridge (MAR), American-Antarctic Ridge (AAR), and Southwest Indian Ridge (SWIR) (Figs 1 and A1). We first quantified the systematic variations in the key morphological parameters and free-air gravity anomaly (FAA) of the global transform systems, including the length, age offset, depth of transform faults and adjacent ridges, depth of ridge-transform intersection, and transform width. We then analyzed the dependence of the morphological variability on the spreading rate and examined the relationship between transform morphology and earthquake magnitude. Finally, we identified anomalous transform features that deviated from the global systematics and investigated the tectonic and magmatic factors controlling the local variability.

      Figure 1.  Global tectonic map showing plate boundaries. Black lines denote global plate boundaries. Green lines denote the ridge-transform systems. Red lines denote the 78 transform faults with relatively good bathymetric data and were investigated in this study for morphological parameters. JDF: Juan de Fuca Ridge; EPR: East Pacific Rise; CR: Chile Rise; PAR: Pacific-Antarctic Ridge; N. MAR: North Mid-Atlantic Ridge; S. MAR: South Mid-Atlantic Ridge; AAR: American-Antarctic Ridge; SWIR: Southwest Indian Ridge; ADEN: Aden Ridge; CIR: Central Indian Ridge; SEIR: Southeast Indian Ridge.

    • We investigated morphological parameters of transform faults through analysis of seafloor bathymetry and gravity data. We used the latest global bathymetry database with grid spacing of 15′′ (Tozer et al., 2019; SRTM15_PLUS Version 2; http://topex.ucsd.edu/WWW_html/mar_topo.html), which combined the available multibeam bathymetry, shipboard soundings, and satellite altimetry-derived seafloor depth. FAA data at 1′ grid spacing were extracted from a global dataset (Sandwell et al. 2014; V26.1; http://topex.ucsd.edu/marine_grav/mar_grav.html).

    • A total of 137 full transform faults were analyzed for transform length and age offset, based on the dataset of Wolfson-Schwehr (2015) (Fig. 2a1, Table A1). Among the transform faults analyzed, a single transform fault may consist of 1–7 sub-segments according to its morphological structure (Table A1), yielding a total of 201 analyzed transform sub-segments. The transform sub-segment length (LS) was calculated by measuring the distance between its two endpoints following a small circle (Wolfson-Schwehr, 2015). The full transform length (LF) was then obtained by the sum of LS for a given transform fault.

      Figure 2.  Examples of transform faults. a1. Depth and (a2) free-air gravity anomaly (FAA) of the Clipperton transform fault, as an example of a fast-spreading ridge. b1. The depth and b2. FAA of the Romanche transform fault, as an example of a slow-spreading ridge. Profiles of seafloor topography (black) and FAA (blue) across (c) Clipperton and (d) Romanche transform faults. Dashed lines (L1 and L2) denote locations of the across-transform profiles. N.H.: nodal high; N.B.: nodal basin.

      No.NameSegment mid-pointLength/
      km
      Full Rate/
      mm·yr–1
      Age offset/
      Ma
      MwmaxAt/km2
      Latitude/°NLongitude/°E
      1Alula Fartak13.9451.7120318.921.486.62724
      2Amsterdam–36.778.6910862.023.486.2584
      3Andrew Bain A–47.4932.238713.3513.036.4909
      4Andrew Bain B–48.5531.314813.3422.196.42018
      5Andrew Bain C–5129.0747113.3370.676.411464
      6Argo–13.5966.3510233.336.126.0731
      7Ascension A–7.37–13.255829.543.935.6333
      8Ascension B–6.88–12.1420329.5113.766.02180
      9Atlantis30.06–42.356322.45.635.8432
      10Atlantis II–32.7657.0420112.0233.445.83365
      11Balleny–61.43154.8135064.510.856.73340
      12Birubi–49.5127.2614869.624.255.4884
      13Blanco A44.33–129.929451.063.686.2523
      14Blanco B44.05–129.232451.010.946.267
      15Blanco C43.89–128.844150.991.615.4150
      16Blanco D43.45–127.8113550.945.306.4900
      17Blanco E43.08–126.834150.931.616.4151
      18Bode Verde A–12.25–14.595630.023.734.9313
      19Bode Verde B–11.68–13.716229.9810.816.21542
      20Boomerang–37.3678.213562.151.135.8108
      21Bouvet–54.261.9220112.7231.66.63271
      22Bullard A–59.13–17.149512.9914.636.21051
      23Bullard B–58.18–11.4952613.4378.336.813479
      24Chain–1.24–14.5231328.5821.96.84242
      25Challenger A–37–96.627846.563.355.8414
      26Challenger B–37.11–95.726746.582.88329
      27Challenger C–37.25–95.192046.610.8653
      28Challenger D–37.32–94.588246.623.525.4446
      29Charlie Gibbs A52.62–33.2620321.7318.687.12541
      30Charlie Gibbs B52.12–30.8211021.8310.085.81011
      31Chile 38S A–38.33–93.634346.851.845.3169
      32Chile 38S B–38.41–92.986846.862.90336
      33Chile 39S–38.96–92.078446.983.586.1460
      34Chile A–35.14–106.5149346.4221.246.36580
      35Chile B–35.9–102.7918646.468.016.71525
      36Chile C–36.21–99.4242046.4318.096.55174
      37Chiloe–43.03–83.086147.822.555.6282
      38CIR 10S–10.0966.567630.954.915488
      39CIR 12 12–11.8565.9910631.906.655.7791
      40CIR 16S–16.2966.9711035.586.185.6792
      41CIR 1S–1.1967.525029.873.355.8265
      42CIR 5S–4.7368.594931.003.165.3252
      43CIR 6S–6.8368.248931.355.685.4614
      44CIR 7S–7.6168.086230.174.115.4364
      45Clipperton10.22–103.9584106.281.586.6307
      46Conrad–55.71–3.1619814.5127.296.72995
      47Darwin–45.9–76.365348.302.195.9227
      48Discovery A–4.01–104.3536123.550.58680
      49Discovery B–4–104.0127123.480.445.852
      50Discovery II A–43.341.6612412.9119.216.41573
      51Discovery II B–41.8642.5921612.8933.516.73620
      52Doldrums A8.82–40.0210925.568.535.6922
      53Doldrums B8.21–38.7816225.7412.5971664
      54Doldrums C7.72–37.3814925.8911.516.21464
      55Doldrums D7.4–35.6622926.0017.626.52783
      to be continued

      Table A1.  Compilation of global oceanic transform faults with a total of 201 individual fault segments+

      Full spreading rates (UF) for individual transform faults (Table A1) were obtained from the Global Strain Rate Map Project (GSRM, V1.2, Kreemer et al., 2000; Wolfson-Schwehr, 2015). The corresponding full transform age offset (AOF) is given by AOF = 2LF/ UF. Similarly, the transform sub-segment age offset (AOS) is given by AOS = 2LS / UF.

    • We further selected 78 transform faults with relatively good bathymetric data to quantitatively analyze their morphological parameters. The cumulated length of the transform faults analyzed for morphological parameters is 16 473 km. A cumulated length of 9 304 km of the adjacent ridges was also analyzed for comparison with the associated transform faults (Fig. A1, Table A2). For a given transform fault, we first carefully traced the mid-points or the deepest points along the transform valley by visual inspection. We then calculated the average depth of the profile and defined it as the transform fault depth (black lines in Fig. A1, Table A2). We also obtained depth profiles along the axis of the adjacent ridges by visual inspection (red lines in Fig. A1, Table A2). The determination of the ridge axes was also aided by examination of the magnetic anomalies. We defined the average depth of the adjacent ridge axes as the ridge depth (Table A2). At a slow-spreading ridge (e.g., the MAR) (Fig. 2b1), we carefully tracked the deepest points of the axial valley or a small central volcanic high. At a fast-spreading ridge (e.g., the EPR) (Fig. 2a1) or a hotspot-influenced ridge, we carefully tracked the axial topographic high or a small central depression.

      No.+NameDT/kmStd. DT/kmDR/kmDR1/kmStd. DR1/kmDR2/kmStd. DR2/kmDTDR/kmStd. DTDR/km
      1Alula Fartak3.900.773.563.490.463.630.340.341.17
      2Amsterdam2.970.542.412.070.082.760.080.560.62
      3–5Andrew Bain5.540.733.714.320.483.100.631.831.28
      6Argo4.090.473.393.040.203.740.400.700.77
      7–8Ascension3.620.223.663.820.303.500.210.040.48
      9Atlantis4.710.324.244.250.314.230.490.480.72
      10Atlantis II5.320.764.144.170.654.120.461.181.32
      11Balleny2.980.442.412.730.082.080.070.570.51
      12Birubi4.450.503.863.970.313.740.230.600.77
      13–17Blanco3.080.622.802.250.023.350.260.280.76
      20Boomerang2.300.151.991.940.082.050.080.300.23
      21Bouvet4.770.312.061.490.352.630.492.720.73
      22Bullard A4.980.234.314.190.354.440.140.670.47
      23Bullard B5.310.964.284.390.154.180.361.021.21
      to be continued

      Table A2.  Depth of 78 ridge-transform fault systems.

    • We carefully examined the 3D topography of a ridge-transform intersection to determine whether it is associated with a local basin (i.e., nodal basin, Fig. A2a1) or a topographical high (i.e., nodal high, Fig. 2a1). For a local basin, we found the deepest point by visual inspection and defined its depth as the nodal basin depth (Table A3). Similarly, for a local topographic high, we located the shallowest point by visual inspection and defined its depth as the nodal high depth.

      Figure S2.  Examples of transform faults. a1. Seafloor depth and a2. FAA of the Zeehaen transform fault, as an example of an intermediate-spreading ridge; b1. Seafloor depth and b2. FAA of the Du Toit transform fault, as an example of an ultraslow-spreading ridge; c and d. Profiles of seafloor topography (black) and FAA (blue) across the Zeehaenand Du Toit transform faults. Dashed lines (L1, L2) denote locations of across-transform profiles. N.B.: nodal basin.

      Longitude/
      °E
      Latitude/
      °N
      Depth/
      km
      Full rate/
      (mm/a)
      Nodal high/
      Nodal basin
      Longitude/
      °E
      Latitude/
      °N
      Depth/
      km
      Full rate/
      (mm/a)
      Nodal high/
      Nodal basin
      –175.96–65.2693.1156.7Nodal high–12.996–11.5174.529.98Nodal basin
      –174.55–65.9983.4656.7Nodal high–12.979–7.0774.8929.51Nodal basin
      –171.13–64.3412.5156.74Nodal high–12.962–17.7054.2130.28Nodal basin
      –170.2–64.8652.4656.74Nodal high–12.89–22.4283.9930.4Nodal basin
      –166.1–62.1993.2547.58Nodal high–12.886–7.30493.8929.54Nodal basin
      –163.42–63.2552.4947.58Nodal high–12.806–22.7493.3630.4Nodal basin
      –161.91–62.442.4946.71Nodal high–12.637–28.1944.3830.35Nodal basin
      –160.48–63.0752.9546.71Nodal high–12.443–28.8233.730.19Nodal basin
      –158.36–62.8223.1345.98Nodal high–12.266–5.13123.5229.27Nodal basin
      –157.37–63.2413.0445.98Nodal high–11.962–21.3523.7430.34Nodal basin
      –155.76–62.0912.8764.73Nodal high–11.91671.6323.7715.45Nodal basin
      –155.12–62.3822.5964.73Nodal high–11.801–22.24.4330.4Nodal basin
      –151.25–59.462.3468.94Nodal high–11.645–4.9754.0729.27Nodal basin
      –150.38–59.7672.3768.94Nodal high–11.493–21.2173.4730.34Nodal basin
      –148.02–57.4682.4171.9Nodal high–11.236–6.65073.9929.51Nodal basin
      –146.95–57.8152.7571.9Nodal high–10.497–46.9772.727.44Nodal high
      –144.81–55.7943.0674.49Nodal high–9.8428–49.3154.2226.9Nodal basin
      –140.27–57.1034.3574.49Nodal basin–8.0912–48.943.6226.9Nodal basin
      –139.38–56.1083.3175.45Nodal high–7.1415–50.6792.7127Nodal high
      –138.77–56.3392.5975.45Nodal high–6.9874–57.3573.4813.4Nodal high
      –136.97–54.1192.2577.89Nodal high–6.957–57.8954.9913.43Nodal basin
      –135.17–54.6712.4577.89Nodal high–6.619170.9823.7115.45Nodal basin
      –134.47–53.7552.6478.82Nodal high–6.1008–57.2653.4113.4Nodal high
      –130.4244.4442.3851.06Nodal high–6.0138–56.6923.1614Nodal high
      –127.72–55.412.6978.82Nodal high–5.1492–50.2522.9527Nodal high
      –127.4–54.8212.6379.02Nodal high–4.7143–56.6262.9314Nodal high
      –126.6142.9923.5451.06Nodal high–4.6404–55.8224.914.51Nodal basin
      –121.67–56.0883.3679.02Nodal high–2.4183–54.2542.7726.9Nodal high
      to be continued

      Table A3.  Depth of ridge-transform intersections.

    • We calculated the widths of 44 transform faults with relatively simple structure. We first traced the peaks of the two conjugate transform walls. We then calculated the distance between the two conjugate peak points at the mid-point of the transform fault. The measured distance was defined as the width of the transform fault (Fig. 2d, Table 2).

      Transform faultWidth /kmTransform faultWidth /kmTransform faultWidth /kmTransform faultWidth /km
      Alula Fartak29Clipperton25Islas Orcadas32SEIR 120E14
      Amsterdam13Conrad38Kane26SEIR 88E10
      Ascension36Du Toit35MAR 15 2027SEIR 96E18
      Atlantis30Euroka19Marie Celeste28Shaka36
      Atlantis II45Falkland26Menard12Tharp20
      Blanco15Garrett B22Oceanographer27Valdavia17
      Bouvet36Geelvinck20Orozco16Vema37
      Bullard A36Gemino36Rivera23Vema II27
      Bullard B40Hayes27Romanche50Vlamingh36
      Chain36Heezen20SEIR 100E18Wilkes25
      CIR 12°12’30Hillegom’s Hole14SEIR 106E25Zeewolf21

      Table 2.  Transform width at the mid-point of 44 investigated transform faults

    • To the first-order, the oceanic plate adjacent to a transform fault can be approximated as a thermal boundary layer that thickens as the square root of the lithospheric age (Turcotte and Schubert, 2014). Thus, the lithospheric plate thickness (hL) at the mid-point of a full transform fault is shown by hL = 2.32$ \sqrt{\kappa {\tau }_{1/2}} $, where $ \kappa $ = 10−6 m2/s is the thermal diffusivity, and τ1/2=LF/UF is lithospheric age at the mid-point of the transform fault.

    • The maximum earthquake moment magnitudes of global transform faults (Mwmax, Table A1) were obtained by Wolfson-Schwehr (2015) and Wolfson-Schwehr and Boettcher (2019) from the Global Centroid Moment Tensor Project (Dziewonski and Anderson, 1981; Ekström et al., 2012). A 50 km wide rectangular polygon centered on a transform fault was used to define the earthquakes of a given transform fault (Wolfson-Schwehr, 2015).

    • The seismogenic area (At) of a transform fault was calculated by integrating the cross-sectional area of the transform wall above the 600°C isotherm (Table A1, Boettcher and Jordan, 2004; Wolfson-Schwehr, 2015).

    • The oceanic transform faults exhibit common topographical characteristics, including a transform valley, high topography on the flanks (Figs. 2 and A2), and a nodal high (Fig. 2a1) or nodal basin (Fig. A2b1) at the ridge-transform intersection. The seafloor topography remains relatively constant away from the transform fault. Across-transform topographical variations are relatively small (1–2 km) and narrow (20–50 km) in the fast (e.g., Clipperton, Fig. 2a1 and c) and intermediate (e.g., Zeehaen, Fig. A2a1 and c) systems, but are larger (up to ~3.5 km) and wider (up to ~100 km) in the slow (e.g., Romanche, Fig. 2b1 and d) and ultraslow (e.g., Du Toit, Fig. A2b1 and d) systems.

      We also examined the FAA that reflects the integrated gravitational effects of the seafloor topography and density anomalies beneath the seafloor (Figs 2a2-b2 and A2a2–b2). Most of the prominent topographical features, including the transform valley and flanking high, are visible in the FAA (Figs 2c–d and A2c–d).

    • Among the 137 transform faults that we examined for transform length, 23, 64, and 38 are located at the ultraslow, slow, and intermediate systems, respectively, while 12 are at the fast systems (Fig. 3). At the fast systems, the Menard transform fault (208 km) at the PAR has the greatest length (Fig. 4a1). At the intermediate systems, the Tasman transform fault (625 km) at the SEIR has the greatest length (Fig. 4a2), which is composed of 5 sub-segments, with lengths of 90 km, 218 km, 62 km, 173 km, and 82 km, respectively. At the slow systems, the Chile transform fault (1 099 km) at the CR has the greatest length (Fig. 4a3), which consists of 3 sub-segments, with lengths of 493 km, 186 km, and 420 km, respectively. At the ultraslow systems, the Andrew Bain transform fault (706 km) at the SWIR has the greatest length (Fig. 4a4), which consists of 3 sub-segments, with lengths of 87 km, 148 km, and 471 km, respectively (Fig. 4).

      Figure 3.  Distribution of full spreading rate of global transform faults. Colors: red (fast), green (intermediate), magenta (slow), and blue (ultraslow).

      Figure 4.  Distribution of full transform length (a1–a4) and transform sub-segment length (b1–b4). Colors: red (fast), green (intermediate), magenta (slow), and blue (ultraslow).

      Most of the transform faults are within 400 km in full length (Fig 4a1–a4). Ultra-long transform fault systems (full length >400 km) are located at the intermediate, slow, and ultraslow systems (Figs. 4a2–a4), including the George V (414 km), Tasman (625 km), Tharp (462 km), Doldrum (726 km), Romanche (878 km), Saint Paul (589 km), Chile (1 099 km), Valdavia (599 km), Andrew Bain (706 km), and Bullard (526 km). Transform sub-segments with lengths over 400 km include the Tharp, Chile A and C, Romanche, Bullard B, and Andrew Bain C (Figs 4b2–b4, Table A1).

    • The full transform age offset (Figs 5a1–a4) and transform sub-segment age offset (Figs 5b2–b4) increase with decreasing spreading rate. Transform faults with age offset greater than 40 Ma are observed mainly at ultraslow and slow systems, including the Andrew Bain, Bullard, Shackelton, Doldrum, Romanche, Saint Paul, and Chile (Figs 5a3–a4 and 6c). Transform sub-segments with age offsets greater than 40 Ma include the Andrew Bain C, Bullard B, Shackelton, and Romanche (Figs 5b3–b4 and 6d, Table A1). The average full/sub-segment age offset decreases with increasing spreading rate (Figs 6c and d, Table 1). Furthermore, the standard deviation values of the transform length and age offset are in general greater for the ultraslow and slow systems than the intermediate and fast systems (Fig. 6, Table 1).

      Figure 5.  Distribution of full transform age offset (a1–a4) and transform sub-segment age offset (b1–b4). Colors: red (fast), green (intermediate), magenta (slow), and blue (ultraslow).

      Figure 6.  Full transform length (a), transform sub-segment length (b), full transform age offset (c), and transform sub-segment age offset versus full spreading rate (d). Filled circles and black lines show average values and standard deviations for each spreading rate, respectively. A: Andrew Bain; AC: Andrew Bain Segment C; B: Bullard; BB: Bullard Segment B; C: Chile; CA: Chile Segment A; CC: Chile Segment C; D: Doldrums; G: George V; R: Romanche; Sa: Saint Paul; Sh: Shackelton; Ta: Tasman; Th: Tharp; V: Valdavia. Colors: red (fast), green (intermediate), magenta (slow), and blue (ultraslow).

      Average and STD
      Parameter*UltraslowrateSlowrateIntermediaterateFastrate
      LF/km185.3/152.7172.9/201.1158.2/145.5108.1/49.5
      LS/km163.9/118.3122.9/120.4108.6/98.244.7/38.1
      AOF/Ma29.9/26.611.0/12.54.7/4.32.1/1.2
      AOS/Ma26.4/22.27.8/8.23.2/2.80.8/1.0
      DT/km4.60/0.854.27/0.593.53/0.713.42/0.36
      DR/km3.44/0.783.68/0.442.86/0.572.72/0.23
      DTDR/km–1.16/0.60–0.59/0.52–0.66/0.60–0.70/0.39
      DN.H./km3.25/0.643.27/0.602.75/0.443.01/0.24
      DN.B./km4.50/0.733.95/0.573.68/0.633.39/0.14
      Note: * LF: Full transform length; LS: Transform sub-segment length; AOF: Full transform age offset; AOS: Transform sub-segment age offset; DT: Depth of transform fault; DR: Depth of adjacent ridges; DTDR: Depth difference of transform fault and ridge; DN.H.: Depth of nodal high; DN.B.: Depth of nodal basin.

      Table 1.  Average and standard deviation (STD) values of morphological parameters of oceanic transform faults

    • Along the global ridge-transform system, transform faults are consistently deeper than the adjacent ridges (Figs 7a1–k1). Correspondingly, the FAA is consistently more negative at the transform faults than the adjacent ridges (Figs 7a2–k2, A3). Regions affected by hotspots are associated with shallower seafloor and more positive FAA, e.g., the Iceland and Azores (Figs 7h1 and h2). When modeled by Gaussian distribution, the mean depth and FAA of the transform faults are consistently greater than that of the adjacent ridges (Figs 8 and A3).

      Figure 7.  Along ridge-transform profiles of seafloor depth (a1–k1) and FAA (a2–k2) of the global ridge-transform systems. Green lines denote along ridge-transform profiles, using the global bathymetry dataset. Black lines denote the 78 transform faults with relatively good bathymetric data and thus selected for detailed analysis of morphological parameters. Red lines denote ridge adjacent to the 78 analyzed transform faults. AAD: Australian-Antarctic Discordance.

      Figure S3.  Correlation between FAA and depth of the analyzed transform faults (black) and adjacent ridges (red). Multiple regions: EPR (a); PAR (b); JDF (c); SEIR (d); CR (e); ADEN (f); CIR (g); N. MAR (h); S. MAR (i); AAR (j); and SWIR (k). Grey lines represent the best-fitting lines of the ridge values for each region.

      Figure 8.  Frequency distributions of seafloor depth (a1–k1) and FAA (a2–k2) of transform faults (black) and adjacent ridges (red) of the 78 analyzed transform systems.

      The average depth of transform faults decreases with increasing spreading rate (Fig. 9a, Table 1). The average depth of the adjacent ridges also decreases with increasing spreading rate (Fig. 9b, Table 1). However, the average depth difference between the transform faults and adjacent ridges is relatively constant, and is only slightly greater at the ultraslow system (Fig. 9c).

      Figure 9.  Transform fault depth (a); adjacent ridges depth (b); and transform-ridge depth difference (c) versus full spreading rate of the 78 analyzed transform systems. Open circles and thin lines denote depth and standard deviation, respectively. Filled circles and black lines show average depth and standard deviation for each spreading rate, respectively. Color definition is the same as in Fig. 3. Bo: Bouvet; H: Heezen; Jan: Jan Mayen; R: Romanche; Th: Tharp.

      The spreading rate dependence of the transform and ridge depth could be associated with a combination of factors including thermal structure and dynamic topography. First, at a given distance from the ridge axis, the mantle temperature is higher and thus the density is lower at a faster ridge (Turcotte and Schubert, 2014). The overall higher mantle temperature also leads to a greater degree of partial melting at a faster ridge. Both the thermal and partial melting density anomalies contribute to shallower seafloor (Magde and Detrick, 1995). Second, slow-spreading ridges are often associated with an axial rift valley that could be caused by hydraulic head loss in the upwelling mantle (Sleep, 1969; Sleep and Biehler, 1970), tectonic faulting (Shaw and Lin, 1996), and lithospheric necking (e.g., Tapponnier and Francheteau, 1978; Lin and Parmentier, 1989; Chen and Morgan, 1990).

    • The average depth of the global nodal basins (Fig. 10a) is 0.38–1.25 km deeper than those of the nodal highs (Fig. 10b, Table A3). The nodal high is dominant at the fast (Fig. 10c) and intermediate (Fig. 10d) systems. In contrast, nodal basin is prevalent at the slow (Fig. 10e) and ultraslow (Fig. 10f) systems.

      Figure 10.  Depth of nodal high (a) and nodal basin (b) versus full spreading rate. Color definition is the same as in Figs 3cf Frequency distribution of the depth of nodal basin (black) and nodal high (red) for each spreading rate group. Curves denote the best-fitting Gaussian distribution.

      The average depths of nodal highs are relatively constant (3.25–3.27 km) for the slow and ultraslow systems, but slightly smaller (2.75–3.01 km) for fast and intermediate systems (Fig. 10a, Table 1). The average depths of nodal basins decrease with increasing spreading rate (Fig. 10b, Table 1).

      The commonly observed nodal basins at slow and ultraslow systems might reflect a decrease in ridge-axis magma supply (Fox and Gallo, 1984) and viscous resistance to mantle upwelling by the cold transform wall (Sleep, 1969; Sleep and Biehler, 1970). In contrast, nodal highs might appear at ridge systems where robust axial volcanic ridges extend toward ridge-transform intersections (Fig. 2, Table A3).

    • In general, the transform width (w) decreases with increasing spreading rate for the ultraslow (29–45 km), slow (17–50 km), intermediate (10–36 km), and fast (12–25 km) systems (Fig. 11a). Ultra-wide transform systems with widths greater than 40 km include the Romanche (w of 50 km, age offset of 62.1 Ma) of the MAR, Atlantis II (45 km, 33.4 Ma) of the SWIR, and Bullard B (40 km, 78.3 Ma) of the AAR (Table 2). In general, the transform width is systematically greater for systems of larger age offset (Fig. 11a). In a special case when a transform valley is bounded by two inward-dipping normal faults, the transform width (w) is expected to depend on the plate thickness (hL) (Fig. 11b inset). However, other factors might also control the transform fault width, including multiple transform faults and extension episodes.

      Figure 11.  Correlation of width of transform faults with full transform age offset (a), and mid-transform lithospheric thickness (b). Color definition is the same as in Fig. 3. Panel b inset: Gray layer shows a lithospheric plate with a thickness of hL. This corresponds to a special case when a transform valley is bounded by two inward-dipping conjugate normal faults with dip angle (α). The dashed lines show the expected dependence of the transform width (w) on plate thickness, with a dipping angle of 20° and 60°, respectively. Aii: Atlantis II; BB: Bullard B; R: Romanche.

    • While the investigated transform parameters show first-order dependence on spreading rate, the variation within each spreading rate group is large, especially for the slow and ultraslow systems (Figs 6, 9 and 10, Table 1). Furthermore, for each of the investigated parameters, multiple examples of major exceptions were observed. In particular, several transform systems have anomalies in more than one parameter (Table 3). These major anomalies reveal unusually complex local tectonic and magma variability.

      Transform faultRidge systemLF >400 kmLS >400 kmAOF >40 MaAOS >40 MaDepth anomaliesWidth >40 km
      Andrew BainSWIR
      Atlantis IISWIR
      BouvetSWIR
      BullardAAR
      ShackeltonAAR
      DoldrumsMAR
      RomancheMAR
      Saint PaulMAR
      Jan MayenMAR
      ChileCR
      ValdaviaCR
      George VSEIR
      TasmanSEIR
      HeezenPAR
      TharpPAR

      Table 3.  Examples of transform faults with major anomalies

    • The Romanche transform fault (Fig. 2b1) is one of the longest (878 km) and deepest (7.9 km) transform systems on Earth (Bonatti et al., 1994; Wolfson-Schwehr, 2015). It is associated with unusually large values in length (Figs 6a and b), age offset (Figs 6c and d), transform depth (Fig. 9a), and transform width (Fig. 11a). The active transform boundary, previously located in the northern valley, had migrated southward to its present location a few million years ago and formed ridges inside the transform domain. The morphological complexity of the Romanche transform was probably caused by alternating across-transform transtension and transpression induced by plate motion changes (Bonatti et al., 1994; Searle et al., 1994; Ligi et al., 2002).

      For the Andrew Bain transform at the SWIR, Sclater et al. (2005) similarly proposed that its morphological complexity was a response to transtension across the transform system, causing multiple transform sub-segments (Fig. A4a). Geodynamic modeling illustrated the possibility of alternating activation of multiple sub-parallel strike-slip faults within a wide transform domain, such as those observed in the Romanche and Andrew Bain systems (Ligi et al., 2002).

      Figure S4.  Maps of the transform systems with major morphological anomalies: Andrew Bain (a); Atlantis II (b); Bullard (c); Heezen and Tharp (d); Bouvet (e); and Jan Mayen (f). OCC: Oceanic core complex.

      In addition to the Romanche and Andrew Bain systems, there are several examples of major morphological anomalies that might have been caused by across-transform transtension and/or transpression. For example, it was proposed that the Atlantis II transform (Fig. A4b) has experienced transtension for 12 million years due to plate rotation of 10° since ~19.5 Ma (Baines et al., 2003); the Bullard transform (Fig. A4c) was subjected to transpression during the early Miocene period (Livermore et al., 1991); and the Heezen and Tharp transforms at the PAR had undergone transtension since 12 Ma, resulting in the formation of extra deep transform valleys (Fig. A4d) (Lonsdale, 1994; Croon et al., 2008).

    • The Bouvet transform fault at the SWIR is located close to the Bouvet hotspot (Fig. A4e). The average depth of the eastern ridge segment is smaller than that of the conjugate western segment by about 1 km (Fig. 9b). Such a transform-ridge depth difference (–2.7 km) is about 1.6 km greater than the global average value for the corresponding spreading rate (Fig. 9c). Thus, the observed anomalies in the depth difference between ridge and transform fault of the Bouvet system are attributed to excess magma, thicker crust, and shallower seafloor at the ridge axis due to hotspot-ridge interaction (Georgen et al., 2001).

      The Jan Mayen transform fault at the northern MAR connects the Jan Mayen Ridge and is located north of Iceland (Figs A1e and A4f). The depths of the Jan Mayen transform fault (Fig. 9a) and the adjacent ridges (Fig. 9b) are exceptionally shallow and are interpreted as the result of the excess magma supply from the Jan Mayen hotspot (Zhang et al., 2019).

    • The Mwmax in general increases with increasing transform length (Fig. 12a) for length less than 300 km, as well as with increasing age offset (Fig. 12c) for age offset less than 40 Ma. At systems of greater transform length and age offset, the Mwmax increases only slightly.

      Figure 12.  The maximum moment magnitude (Mwmax) of the transform earthquake versus transform sub-segment length (a); full spreading rate (b); transform sub-segment age offset (c); and seismogenic area ((d, At). Open circles denote Mwmax of the global transform faults. Filled circles and black thin lines show the average Mwmax and standard deviations for each spreading rate. Color definition is the same as in Fig. 3. Gray squares in panel b show the empirical corner magnitudes of three subgroups by maximum likelihood. The solid curve denotes the calculated corner magnitude (Bird et al., 2002). Dashed curves in panel d show the calculated Mw dependence on At with slip (S) of 60 cm and 0.5 cm.

      Mwmax increases with increasing At (Fig. 12d), which is consistent with theoretical consideration. The moment magnitude (Mw) of an earthquake is defined as Mw=2/3×log10 (μ×At×S) – 10.7 (Kanamori, 1977), where μ=3×1010 Pa is the rock rigidity, and S is the average slip during an earthquake. Therefore, Mwmax should increase with increasing At and S (Fig. 12d). Roland et al. (2010) proposed that the base of the seismogenic area (600 °C isotherm) is nearly flat, which differs from the half-space cooling model (Boettcher and Jordan, 2004). Therefore, using the half-space cooling model might underestimate At and Mw.

      Mwmax decreases moderately with increasing spreading rate (Fig. 12b). Bird et al. (2002) also found that the corner magnitude of transform earthquakes decreases with increasing spreading rate. Such spreading rate dependence could be due to multiple factors. First, the age offset is in general smaller for fast systems, leading to smaller At and thus Mw. Second, the lizardite mineral phases might exist only at transforms with slow rates, where serpentinization might be present at transform faults. The presence of lizardite in slow-spreading systems might promote unstable sliding along the transform fault, leading to large transform earthquakes (Bird et al., 2002).

    • Our quantitative analysis of the morphological parameters of global oceanic transform faults yielded the following key results:

      (1) The age offset of transform faults and transform sub-segments decrease with increasing spreading rate.

      (2) Both transform faults and adjacent ridges are shallower at the fast-spreading compared to the slow-spreading systems, likely reflecting density anomalies associated with warmer mantle at the fast systems and rifting at the slow ridges. However, the average depth difference between the transform fault and adjacent ridges is relatively constant from the slow- to fast-spreading systems.

      (3) Nodal basins are mostly observed at the slow and ultraslow systems; whereas nodal highs are prevalent at the fast and intermediate systems. The nodal basin is possibly related to reduced ridge-axis magma supply and stronger viscous resistance to mantle upwelling by the cold transform wall, whereas a nodal high is formed where the robust ridge-axis volcanic ridges extend toward a transform fault.

      (4) The average transform width is greater at a larger age offset, possibly reflecting greater effective elastic thickness of the deforming plates flanking the transform valley.

      (5) The average maximum moment magnitude of the transform earthquakes increases with the transform length, age offset, and seismogenic area of the transform faults. The maximum moment magnitude decreases with increasing spreading rate, reflecting relatively small age offset and seismogenic area of fast-spreading systems, as well as the presence of lizardite mineral phases in slow-spreading systems that might promote unstable transform sliding.

    • This work benefited from constructive discussion with Jason Phipps Morgan, Marcia Maia, Hongfeng Yang, Zhiyuan Zhou, and the SCSIO Deep Ocean Geodynamics Group.

Reference (53)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return