secular change in lifetime of granitic crust and the continental growth a new view from detrital zircon ages of sandstones

Accepted Manuscript Secular change in lifetime of granitic crust and the continental growth: a new view from detrital zircon ages of sandstones Hikaru Sawada, Yukio Isozaki, Shuhei Sakata, Takafumi Hirata, Shigenori Maruyama PII: S1674-9871(16)30208-0 DOI: 10.1016/j.gsf.2016.11.010 Reference: GSF 515 To appear in: Geoscience Frontiers Received Date: 29 February 2016 Revised Date: November 2016 Accepted Date: 25 November 2016 Please cite this article as: Sawada, H., Isozaki, Y., Sakata, S., Hirata, T., Maruyama, S., Secular change in lifetime of granitic crust and the continental growth: a new view from detrital zircon ages of sandstones, Geoscience Frontiers (2017), doi: 10.1016/j.gsf.2016.11.010 This is a PDF file of an unedited manuscript that has been accepted for publication As a service to our customers we are providing this early version of the manuscript The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT Secular change in lifetime of granitic crust and the continental growth: a new view from detrital zircon ages of sandstones RI PT Hikaru Sawada Maruyama d a,* , Yukio Isozakia, Shuhei Sakatab, Takafumi Hirata c, Shigenori M AN U SC a Department of General System Studies, the University of Tokyo 10 b Earth-Life Science Institute (ELSI), Tokyo Institute of Technology 11 c Department of Chemistry, Gakushuin University 12 d Geochemical Research Center, the University of Tokyo 15 16 17 EP 14 AC C 13 TE D * Corresponding author: hsawada@ea.c.u-tokyo.ac.jp (H Sawada) 18 ACCEPTED MANUSCRIPT 20 21 22 23 Highlights Detrital zircon U-Pb dating for Archean–Proterozoic sandstones from Australia, N America, and Africa RI PT 19 Zircon age compilation of 2.9, 2.6, 2.3, 1.0 and 0.6 Ga sandstones revealed history of continental growth Rapid production/recycle of continental crusts in the Neoarchean–Paleoproterozoic 25 Net growth of continents occurred after 2.0 Ga, whereas net decrease after 1.0 Ga M AN U SC 24 26 Abstract 28 U-Pb ages of detrital zircons were newly dated for Archean sandstones from the Pilbara 29 craton in Australia, Wyoming craton in North America, and Kaapvaal craton in Africa 30 By using the present results with previously published data, we compiled the age spectra 31 of detrital zircons for 2.9, 2.6, 2.3, 1.0, and 0.6 Ga sandstones and modern river sands in 32 order to document the secular change in age structure of continental crusts through time 33 The results demonstrated the following episodes in the history of continental crust: (1) 34 low growth rate of the continents due to the short cycle in production/destruction of 35 granitic crust during the Neoarchean to Paleoproterozoic (2.9–2.3 Ga), (2) net increase in 36 volume of the continents during Paleo- to Mesoproterozoic (2.3–1.0 Ga), and (3) net 37 decrease in volume of the continents during the Neoproterozoic and Phanerozoic (after 38 1.0 Ga) In the Archean and Paleoproterozoic, the embryonic continents were smaller 39 than the modern continents, probably owing to the relatively rapid production and AC C EP TE D 27 ACCEPTED MANUSCRIPT destruction of continental crust This is indeed reflected in the heterogeneous crustal age 41 structure of modern continents that usually have relatively small amount of Archean 42 crusts with respect to the post-Archean ones During the Mesoproterozoic, plural 43 continents amalgamated into larger ones comparable to modern continental blocks in size 44 Relatively older crusts were preserved in continental interiors, whereas younger crusts 45 were accreted along continental peripheries In addition to continental arc magmatism, 46 the direct accretion of intra-oceanic island arc around continental peripheries also became 47 important for net continental growth Since 1.0 Ga, total volume of continents has 48 decreased, and this appears consistent with on-going phenomena along modern active 49 arc-trench system with dominant tectonic erosion and/or arc subduction Subduction of 50 a huge amount of granitic crusts into the mantle through time is suggested, and this 51 requires re-consideration of the mantle composition and heterogeneity 52 TE D M AN U SC RI PT 40 Keywords: detrital zircon, U-Pb age, continental growth, subduction erosion, 54 preservation bias AC C 55 EP 53 ACCEPTED MANUSCRIPT 56 Introduction The vast occurrence of granitic continental crust, as well as the existence of life, is 58 one of the unique features of the Earth, in remarkable contrast to other planets of the solar 59 system Nevertheless, the origin and history of continent has been the main topic of 60 discussion for years but not yet fully understood Fig 1A shows geotectonic map of all 61 extant continents and age proportion of continental crusts, from the Archean to Cenozoic 62 Conventionally, the gradual accumulation of continental (sial) crust through time was 63 assumed on the basis of its relatively large buoyancy with respect to oceanic (sima) crust 64 (e.g., Hurley and Rand, 1969) In contrast, considerations on overall thermal history of 65 the Earth drove some researchers to imagine the vigorous formation of continental crust 66 particularly during the earliest history of the planet; the total volume of early continents 67 exceeded even over that of the present continents (e.g., Fyfe, 1978; Armstrong and 68 Harmon, 1981; Fig 1B) TE D M AN U SC RI PT 57 During the 1980s–1990s, geochemical analyses provided other lines of evidence to 70 assume the relatively steady-state growth of total continental volume (e.g., McLennan 71 and Taylor, 1982; McCulloch and Bennett, 1994; Fig 1B) In addition, anatomy of major 72 continents with distinct age composition became much clearer than before, particularly in 73 North America, by detailed field mapping and geochronological studies (e.g., Hoffman, 74 1988; Bowring et al., 1998) The initiation/operation of plate tectonics in the early Earth 75 was practically proved by concrete geological lines of evidence of oceanic 76 subduction-related rock units (accretionary complex and arc garnitoid) and horizontal 77 layer-parallel shortening (duplex) structures (Sleep and Windley 1982; Maruyama et al., 78 1991; Komiya et al 1999, 2015; Komiya and Maruyama, 2011; Kusky et al., 2013) They AC C EP 69 ACCEPTED MANUSCRIPT explained that plate tectonics started during the early Archean or even in the Hadean 80 Strongly opposing this view, some researchers insist for no operation of plate tectonics 81 during the Archean; by estimating relatively low density with thicker basaltic crust, ca 82 times thicker than the present, for example, Davies (1992, 1995) argued that Archean 83 oceanic crust was thus too buoyant to be subducted Some geophysical models also insist 84 that plate subduction was highly limited with assumed warm Archean mantle, much 85 warmer than today because of highly depleted peridotitic mantle lithosphere (e.g., O’Neil 86 et al., 2007; Korenaga, 2008) These claims can be reasonably refuted by considering 87 slab melting and mineral-phase change in deeper mantle (Komiya et al., 2004) Recently, 88 more actualistic geophysical models with respect to mineral physics (e.g., Ogawa, 2007, 89 2014; van Hunen et al., 2008; Sizova et al., 2010; Fischer et al., 2016) suggested that a 90 certain kind of plate subduction, not necessarily the same as modern one, has operated 91 during the Archean, and probably started much earlier already in the Hadean TE D M AN U SC RI PT 79 In the 1990s–2000s, a totally new input of information was given by the introduction 93 of detrital zircon chronology (e.g Gehrels and Dickinson, 1995; Gehrels et al., 1995) 94 Detrital zircons from the Archean Narryer complex (Yilgarn Craton, W Australia) 95 positively suggested that the production of felsic continental crust has started already in 96 the mid-Hadean (Mojzsis et al., 2001; Harrison et al., 2005, 2008; Ushikubo et al., 2008; 97 Carley et al., 2015) This notion is contradictory with the conventional view on the 98 Hadean crust (e,g, McLennan and Taylor, 1982; McCulloch and Bennett, 1994), as many 99 researchers still considered that ancient crusts were originally komatiitic/basaltic without 100 any granitoid produced by arc magmatism (Griffin et al., 2014; Nebel et al., 2014; 101 Reimenik et al., 2014; Kamber, 2015; Gaschnig et al., 2016) AC C EP 92 ACCEPTED MANUSCRIPT On the other hand, Rino et al (2004, 2008) analyzed age spectra of detrital zircons in 103 river sands from extant modern continents, in particular, U-Pb age spectra of detrital 104 zircons of deltaic river sands from the Mississippi River, and compared the results with 105 the surface crustal age distribution of North American Craton with lesser Precambrian 106 sedimentary covers Consequently they demonstrated that the river sand composition 107 faithfully reflect the crustal composition of the provenance, regardless of orogenic 108 disturbance by the Rocky mountains and/or terrigenous noise by the Quaternary 109 glacier-interglacial cycles This result confirmed that crustal composition of hinterland is 110 by and large reflected in river sands deposited at lower streams of a major river with large 111 drainage system, when the crustal basement is extensively exposed Essentially similar 112 results were obtained also from other continents (Rino et al., 2008) Although almost 113 parallel to each other, the narrow gap space (shaded area in Fig 1B) existing between the 114 two cumulative curves by Ustunomiya et al (2007) and Rino et al (2008) suggests that 115 the corresponding amount of continental crusts has disappeared by sedimentary recycling, 116 particularly during the last one billion years They indicate that continental crusts older 117 than 2.3 Ga (i.e the first half of the Earth history) merely occupy no more than 20% of all 118 continents, and those of > Ga are quite rare (Fig 1B), in good agreement with actual age 119 proportion of crust on extant continents (Fig 1A) AC C EP TE D M AN U SC RI PT 102 120 By assuming that oceanic subduction has produced granitic crust continuously since 121 the Archean, a huge amount of buoyant continental crust is expected to have formed and 122 accumulated on the planet’s surface according to the elapsed time Nonetheless, this is 123 not the case that we observe on extant continents, as mentioned above This 124 disagreement between the long elapsed time and smaller remnants of older crusts can be ACCEPTED MANUSCRIPT 125 reasonably explained only when older continental crusts disappeared from the surface 126 secondarily As observed in modern Earth, granitic continental crust is formed under the 128 operation of plate tectonics, in particular, along active subduction zones; on the other 129 hand, plate subduction can cause significant volume reduction of continental crust 130 through subduction erosion, sediment subduction, and island arc subduction (von Huene 131 and Lallemand, 1990; Scholl and von Huene, 2007; 2009; Clift et al., 2009; Yamamoto et 132 al., 2009; Isozaki et al., 2010) Recent estimates on the global volume of global crust 133 generation and destruction along modern subduction zones show that the rate of 134 destruction equals or even exceeds the production (e.g Clift et al., 2009; Stern, 2011) 135 Recent seismic tomography data also suggest the occurrence of large amounts of recycled 136 silicic crustal material in the mantle (Kawai et al., 2009, 2013; Ichikawa et al., 2013; 137 Garnero et al., 2016) TE D M AN U SC RI PT 127 Recent compilations of large dataset (ten thousands of ages from multiple sources) 139 of detrital zircon ages, for minimizing local bias, recognized some peaks in zircon age 140 (Condie et al., 2009; Belousova et al., 2010; Voice et al 2011; Roberts and Spencer, 141 2015) These distinct peaks apparently correspond to the timings of supercontinent 142 amalgamation during the Proterozoic and Phanerozoic (Rino et al., 2004, 2008; Condie et 143 al., 2009) Some researchers proposed that episodic production of juvenile crust caused 144 by mantle plumes related to break-up of supercontinents (Rino et al., 2004; Condie et al., 145 2009) Other researchers pointed out that the continental crust would obtain preservation 146 potential through shielding continental inboard from subduction zones during 147 supercontinental periods (Hawkesworth et al., 2009; Roberts, 2012) These discussions AC C EP 138 ACCEPTED MANUSCRIPT highlighted contrasting views on the process of construction of continental crust, i.e 149 episodic production versus preservation potential To date, diverse models have been 150 proposed for the secular change in total continental volume through time (Fig 1B); 151 however, it is too crude to interpret the accumulated age spectra without checking 152 depositional ages and/or settings of host sandstones because of significant destruction of 153 older continental crust through time As such a major obstacle exists in reconstructing 154 precise volume of ancient continents, the secular change in age structure of continental 155 crust through time has not yet been clearly demonstrated M AN U SC RI PT 148 Age structure is one of the fundamental parameters for population dynamics 157 (Veizer and Jansen, 1985) Without any recycling at all, an accumulated age curve of 158 modern river sand can easily lead/reconstruct those for any given time in the past (Fig 159 1C); however, we must admit that such an ideal case is extremely rare The evolution of 160 continental crusts indeed occurred not in steady-state but with volatile changes of 161 production and destruction through the Earth’s history TE D 156 In this study, first we analyzed age spectra of detrital zircons from Archean 163 sandstones collected from above-unconformity horizons in Australia, North America, 164 and Africa On the basis of the present data, together with previously reported data, we 165 compiled the age structure of detrital zircons for several time-bins subdivided by 166 depositional ages of host sandstones This compilation led us detect a contrasting aspect 167 in age spectra of detrital zircons between pre-2.3 Ga interval and post-1.0 Ga one, which 168 likely suggests that a major change in age structure of continents has occurred sometime 169 in mid-Precambrian time This article discusses the secular change in age structure of the AC C EP 162 ACCEPTED MANUSCRIPT 1241 Figure 5: Comparison in cumulative curves of averaged detrital zircon age spectra for 1243 distinct time bins, i.e ca 2.9, 2.6 Ga (Archean), ca 2.3, 1.0, 0.6 Ga (Proterozoic), and 1244 Ga (Phanerozoic) (see main text for original references) 1245 Vertical axis show age structure of continents in the form of cumulative age frequency 1246 distribution with taking 100 % on a continental volume at each period (A) cumulative 1247 curves of raw data of averaged age spectra; (B) cumulative lines fitted for polygonal line 1248 function with original curves of age structure M AN U SC RI PT 1242 1249 Figure 6: Patterns of possible cumulative curves (fitted cumulative lines) and 1251 interpretations in terms of net growth (production minus destruction) of continental 1252 crusts and their preservation bias 1253 (A) When crust production is constant without destruction, the cumulative line changes 1254 its slope toward gentle, and the age span becomes wider along time (from t0 to t2) (B) 1255 Three options (1 to 3) may occur according to a balance between the production and 1256 destruction of continental crust; i.e., (1) production exceeding destruction, (2) balanced, 1257 and (3) destruction exceeds production The mutual distance (age) between lines and 1258 their slopes are different among the three options (C) Six options (1 to 6) may occur 1259 according to the production/destruction balance and also to preservation bias The 2.9, 1260 2.6, and 2.3 Ga cumulative curves correspond to the option 2, whereas those of 1.0, 0.6 1261 and Ga to the option AC C EP TE D 1250 1262 56 ACCEPTED MANUSCRIPT Figure 7: Estimated growth history of continents after 3.0 Ga, on the basis of the present 1264 compilation of detrital zircon age spectra of sandstones in time bins and their 1265 interpretation 1266 Note that the main trend in continental growth with rapid increase during ca 2.0–1.0 Ga 1267 and sharp decrease after 1.0 Ga The first large continent (Nuna/Columbia) appeared as 1268 soon as the main increase started, whereas the onset of modern-style cold subduction 1269 terminated the unidirectional continental growth, instead, the decrease of total 1270 continental mass M AN U SC RI PT 1263 1271 Figure 8: Schematic models of continental growth and island arc accretion since the 1273 Hadean in map view and in profile, and the corresponding world map 1274 (A) During ca 4.4–3.2 Ga, numerous island arcs formed during this time interval (right 1275 map) and accretion/subduction of island arcs occurred frequently In the case of parallel 1276 collision of two island arcs (left), they easily amalgamated to each other and grew into 1277 minor land masses (arc accretion) In contrast, in the case of perpendicular collision of 1278 one arc to the other, the crust of the colliding arc likely subducted smoothly into the 1279 mantle (arc subduction; middle) Moreover, the subduction erosion also occurred to 1280 destruct pre-existing arc crusts Consequently the preservation potential of the 1281 continental crust was very small 1282 arcs (left) emerged as embryonic continents, which were larger than individual island 1283 arcs but smaller than that of modern continents without having significant amount of 1284 older crust (right) AC C EP TE D 1272 (B) During ca 3.2–1.8 Ga, some collided composite Tectonic recycling occurred along active continental margins (left, 57 ACCEPTED MANUSCRIPT 1285 middle) to suppress the net growth of continental crusts 1286 plural embryonic continents amalgamated to build larger continents comparable to 1287 modern ones (right) Pre-existing crusts in the interiors were protected from the 1288 subduction-related tectonism along the active margins (left, middle), thus the 1289 preferential preservation of older crusts started (left, middle) For example, around ca 1290 1.8–1.7 Ga the first supercontinent Nuna/Columbia developed On the other hand, the 1291 accretion of island arcs along the peripheries were effective to increase the total mass of 1292 continental crusts M AN U SC RI PT (C) During ca 1.8–1.0 Ga, 1293 1294 Figure 9: 1295 view with respect to the cumulative curves of detrital zircon ages from time bins and 1296 to assumed mantle potential temperature 1297 (A) World map with continental crusts in three time intervals; i.e., 4.5–3.2 Ga, 3.2–1.8 1298 Ga, and 1.8–1.0 Ga (simplified from Fig 8) 1299 continents (Fig 7) 1300 (Komiya, 2004; Herzberg et al., 2010) Note that the mode change in continental entity 1301 through time occurred in accordance with the general cooling trend of the planet, and 1302 this is reflected in detrital zircon age patterns TE D (B) A speculative growth history of EP (C) Assumed secular change in potential mantle temperature AC C 1303 Schematic images showing the secular change in continental growth in map 1304 Figure 10: Speculative diagram showing the secular change in total production of 1305 continental crust (light blue), in total subduction of continental crust into the mantle 1306 (light purple), and the resultant growth pattern of continents (orange) through time 58 ACCEPTED MANUSCRIPT The production/subduction of old continental crusts, older than the Mesoproterozoic, 1308 was huge in magnitude with respect to that in the Phanerzooic This suggests the burial 1309 of great amount of granitic crustal material into the mantle in the earlier half of the 1310 Earth’s history RI PT 1307 1311 SC 1312 M AN U 1313 1314 Tables 1315 Table Fitting for polygonal line function and continental average cycle calculated 1316 from the result of fitting EP 1319 AC C 1318 TE D 1317 59 ACCEPTED Table Fitting for polygonal line function and continental average cycle calculated from the result of fitting Age T[Ma] 2900 2600 2300 100 600 767 748 1287 1938 2869 TE D M AN U SC RI PT 678 EP Cycle = X-T [Myr.] 3578 3367 3048 2287 2538 2869 AC C X [Ma] MANUSCRIPT AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT A ca 4.4-3.2 Ga arc subduction collision perpendicular collision A A B M AN U very small preservation B A SC B B A RI PT arc accretion oceanic island arcs B ca 3.2-1.8 Ga A TE D embryonic continent (composite arc) small preservation A B AC C EP B C ca 1.8-1.0 Ga embryonic continents preserved old crust stable large continent large preservation A A B B arc accretion stable large continents 5000 km subduction orogen trench mid oceanic ridge AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT ...AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT Secular change in lifetime of granitic crust and the continental growth: a new view from detrital zircon ages of sandstones. .. Archean sandstones from the three representative Archena cratons; i.e., 183 the Pilbara craton in Western Australia, Kaapvaal craton in southern Africa, and 184 Wyoming craton in North America (Fig... groups The mean and 237 standard error of the measured ratios among each eight 91500 and/ or Plešovice zircon 238 standard data bracketing unknown sample groups were calculated, and the mean and