Insights into the tectonic evolution of the North China Craton through comparative tectonic analysis: A record of outward growth of Precambrian continents

a State Key Lab for Geological Processes and Mineral Resources, China University of Geosciences, Wuhan 430074, China b Center for Global Tectonics, School of Earth Sciences, China University of Geosciences, Wuhan 430074, China c Department of Earth and Environmental Sciences, University of Windsor, Ontario, Canada d Department of Geology, University of Leicester, U.K e Department of Geosciences, University of Houston, TX, USA f University of California Davis, University College, Oxford, United Kingdom g Department of Atmospheric and Environmental Sciences, University at Albany, USA h Earth-Life Institute, Tokyo Institute of Technology, Japan


Introduction
It is not easy to recognize suture zones and records of past Wilson Cycles that may be preserved in ancient high-grade metamorphic rocks that lack fossils, and from which it is not possible to obtain reliable paleomagnetic data. However, if methods of tectonic analysis such as delineation of regional tectonic zonation, sedimentology, structural analysis, geochronology, geochemistry of magmatic rocks, and metamorphic petrology are mutually combined, a robust tectonic analysis can ensue. Studies that rely solely on one or two methodologies are unlikely to generate sufficient fundamental data to be able to produce meaningful results. In this work we apply all these methods of tectonic analysis to the North China Craton (NCC hereafter) in order to document the presence of sutures, the nature of different units that were mutually sutured, the geometry of subduction that led to the suturing, and we examine how these processes may or may not have changed with time from the Archean to the present and from the low-grade upper crust to the high-grade lower crust. For the NCC there are considerable data, but also many controversies about the tectonic sub-divisions, the location and ages of sutures and orogens, the polarity of interpreted subduction zones, the presence of collisional orogens and sutures, and the meaning of metamorphic trajectories. We shed light on these problematic interpretations through copious use of global comparisons between Archean examples and Phanerozoic analogues. This major craton preserves a record of 3.8 billion years of geological activity. We use the methods of tectonic analysis and comparative tectonics to test whether or not there has been any secular change in tectonic style during its long history.
The NCC is divisible into several different tectonic units (Fig. 1), the boundaries of which and the timing and significance of their formation are under lively discussion. The Eastern Block consists of Neoarchean tonalite-trondhjemite-granodiorite (TTG gneisses, granitoids), and

ACCEPTED MANUSCRIPT
8 is typically characterized by structurally complex, thrust-imbricated rocks such as mélanges, dismembered ophiolites and ocean plate stratigraphy that were scraped off the intervening oceanic plate during convergence and collision of the two tectonic blocks (Fig. 2). A fundamental current problem is that early work on sutures was in upper crustal rocks such as the Appalachians and Alps (e.g., Dewey, 1969;Bird and Dewey, 1970;Windley, 1995) for which there are modern analogues, but much later work has been in lower crustal rocks, but few deep crustal levels of modern orogens are exposed today. Thus the more deeply buried and highly metamorphosed a suture zone becomes, the harder it is to recognize it as a suture, until at some point -typically at granulite facies levela suture becomes cryptic, and might only be recognized by "a few specks of fuchsite" along a shear zone (Burke et al., 1976). For instance, boundaries between different tectonic blocks in the Archean craton of West Greenland are marked by up to 200 meter wide mylonitic, amphibolite facies volcanic and sedimentary rocks and serpentinites. These have been interpreted to represent the roots of Archean suture zones (Nutman and Friend, 2007;Windley and Garde, 2009;Polat et al., 2015). One of the goals of this paper is to use modern understanding of suture zones based on young orogens such as shown in Figure 2 to test whether or not we can recognize any similar patterns in the ancient North China Craton, and then compare this with other Archean terranes world-wide to define criteria about how to recognize sutures in ancient high-grade rocks.

Fig. 2 near here, cross-section of orogen
Suture zones are commonly overprinted by later events, and thus become even harder to recognize. For example, and compared with the NCC, in the Appalachians (Fig. 3) the Ordovician Taconic suture between the North American passive margin and an accreted island A C C E P T E D M A N U S C R I P T

ACCEPTED MANUSCRIPT
12 in this geological / tectonic scenario. The suture marks the zone that separates rocks that were deposited on one continent from an accretionary prism and arc-related rocks on the other.

Fig. 3 near here map of NE Appalachians
After the arc collides with the continent, the trench is consumed, and the subduction zone typically steps outboard towards the ocean, dipping under the continent and accreted arc terrane (e.g., Clift et al., 2003;Deng et al., 2013;. This is because convergence still continues between the continent and the remaining open part of the ocean, and the back of the arc is the weakest part of the system. Therefore, the simple tectonic zonation described above begins to be overprinted, and the arc may even become dismembered into smaller fragments as in the Timor-Australian collision zone (Rutherford et al., 2001). New continental-margin arc magmas intrude the collision-modified continental margin, and most importantly, another arc or continent will eventually collide with this margin. So now there are two accretionary wedges in this accretionary orogen, which become overprinted by processes associated with the conversion from an accretionary orogen to a collisional orogen. In the case of the Appalachian orogen, the younger accretionary wedge(s) related to closure of the Acadian Ocean is preserved in the Merrimack Synclinorium (Fig. 3), whereas the older accretionary wedge related to closure of the Iapetus Ocean during the Taconic orogeny is preserved farther inboard. Thus, two sutures can be drawn on the map of New England (Bradley, 1983). During continent/continent collision, deformation and metamorphism can be intense, typically up to granulite facies, and new plutons of crustal melt granites intrude the structurally complex package. Acadian plutons are widespread throughout the northern Appalachians (Fig. 3), whereas Taconic-aged plutons are rare. The highgrade metamorphism resulting from the terminal continent-continent collision overprints and A C C E P T E D M A N U S C R I P T

ACCEPTED MANUSCRIPT
13 typically obliterates much of the lower-grade metamorphism of the earlier Taconic arc-continent collision, which is why most metamorphic studies in the Appalachians reveal just Devonian metamorphic peaks and P-T-t paths, and only very detailed structural/metamorphic study together with geochronology can reveal the earlier events (e.g., Tremblay et al., 2000;Castonguay et al., 2012;de Souza et al., 2014). In following sections we use a holistic tectonic approach, as described above, to define Precambrian suture zones in the NCC, and to test for records of Wilson Cycles and the operation of plate tectonics during the craton's long evolution (eg. Kusky et al., 2014).

Geological Zonation of the NCC at 2.5 Ga: Tracing an Archean Suture
In this section we attempt to explain the clear differences in Archean geology between the Eastern Block, the COB, and the Western Block ( Fig. 1) (using terminology of Kusky and Li, 2003;Kusky et al., 2007a;Kusky and Santosh, 2009;Kusky, 2011a,b). We next examine the geology of the boundary between the Eastern Block of the NCC and what we interpret as an accreted Archean arc terrane (Wutai/Fuping arc) within the COB, to determine the nature of rock units, structures, and metamorphism on either side, to determine the time of suturing of these two terranes, and to trace this suture across the NCC. Note that the cross-strike scale of the COB is similar to that of the Taconic orogen in eastern North America (compare Fig. 1 with Fig. 3), but the original length of the COB can only be estimated as greater than 1,300 km.
In a direct application of the Burke et al. (1976) model of microblock accretion for the Archean, it has been proposed that there are several different microblocks within the Eastern Block that amalgamated between 2.7 and 2.6 Ga (Zhai et al., 2000Zhai, 2014; M A N U S C R I P T ACCEPTED MANUSCRIPT 14 Santosh, 2011), but the boundaries of these blocks are not well-defined (if they even exist), and the geometry, timing, and mechanism of these proposed events are not yet well-established , and it is not clear if they represent primary microblocks, or a single terrane latter dismembered by younger tectonics, as in the case of Indonesia which has developed its present short arcs since the Australian collision with Timor about 18 Ma ago (Rutherford et al., 2001). In addition, the purported "Neoarchean ophiolite" located near one of these boundaries (Santosh et al., 2016) is only about 100 meters thick, and consists only of very altered rocks interpreted as hornblende norite, hornblende-OPX-lherzolite, gabbroic hornblendite, hornblendite, hornblende-gabbro, and granite. It is not laterally extensive, lacks deep-water sedimentary deposits, has no pillow lavas, no sheeted dike complex, no layered gabbros or cumulates, nor harrzburgite tectonites; it only consists of a suite of deformed maficultramafic rocks of uncertain origin within a contemporaneous pluton. It is not clear if this is a "micro-ophiolite", an ophiorag (sensu Sengör and Natal'in, 2004), or just a mafic-ultramafic phase of the enclosing pluton. Therefore, we refrain from speculation on these purported microblocks and the nature of their boundaries, and begin our analysis at the time when the western margin of the Eastern Block was covered by a thick passive margin (from > 2.6-2.5 Ga).

Zanhuang Massif
The Zanhuang Massif (Figs. 1, 4) is located in the south-eastern COB along its border with the Eastern Block of the NCC. It consists of three main tectonic zones (Trap et al., 2009(Trap et al., , 2012, which can each be subdivided into Domains (Wang, J.P. et al., 2013). The Eastern Domain consists of TTG gneiss and migmatite of the Eastern Block of the NCC, and is overlain in the west by a sequence of metasandstone, marble, and metapelite, grading up into a metagraywacke-pelite unit, then more younger sediments. This zone is interpreted as the older

ACCEPTED MANUSCRIPT
15 continental crust of the Eastern Block overlain by a passive margin sequence, then by a foreland basin with flysch sediments followed by a superimposed retroarc basin sequence (Kusky and Li, 2003;Li and Kusky, 2007;Wang, J.P. et al., 2013Wang, J.P. et al., , 2016. The Western Domain of the Zanhuang Complex consists of tonalitic gneisses with ages of 2692+/-12 Ma (Yang CH et al., 2013), as well as a suite of hornblende-bearing plutons dated at 2511+/-36 Ma and 2528+/-18 Ma (Wang J.P. et al., in press). The Western Domain was correlated with the Wutai/Fuping Arc (Fuping Block) in the COB by Kusky and Li (2003), Trapp et al (2009,2012), Wang, J.P. et al. (2013) and Deng, H. et al. (2013), and is considered to be an island arc with magmatic ages of circa 2.7-2.5 Ga. The Central Domain of the Zanhuang Massif consists of a complex mixture of metapelites, metapsammites, metabasalts, metagabbros, and rare ultramafic rocks, forming a structurally complex mélange (Wang, J.P. et al., 2013(Wang, J.P. et al., , 2016Figs. 5, 6). The mélange belt shows consistent sense of shear indicators of thrusting from the NW to the SE, but the mélange shows at least two generations of similarly-oriented fault sets, suggesting a complex tectonic history (Figs. 5, 6) (Trap et al., 2009a). We interpret this zone to be a subduction/ accretion/ collisional mélange (Wang, J.P. et al., 2016), related to the closure of the ocean between the Eastern Block and the Wutai/Fuping arc in the COB, and final collision of the Western Zanhuang Domain (the Wutai/Fuping arc) with the passive continental margin of the Eastern Block. It thus represents the suture between the Eastern Block and an arc that collided with it. The clear tectonic zonation within the Zanhuang Domain fits the formal criteria for defining a suture zone, as discussed in section I. The timing of suturing of the Wutai/Fuping arc with the Eastern Block has been controversial. Kusky and Li (2003), Kusky et al. (2007a), Kusky (2011a,b), Polat et al. (2005Polat et al. ( , 2006 have all suggested that this collision occurred around 2.5 Ga, because of the overwhelming geological and geochronological evidence for accretionary and collisional tectonic events at that time. However, Zhao et al. (2001Zhao et al. ( , 2009 have consistently argued for a circa 1.85 Ga collision, based on their interpretation of P-T-t paths and recrystallized zircons from metamorphic rocks. The problem was solved by Wang, J.P. et al. (2013) and Deng, H. et al. (2013), who reported zircon ages of circa 2.5 Ga on granitic plutons and pegmatites that cut the fabrics in the mélange, clearly showing that the accretion and collision took place before 2.5 Ga. The Zanhuang (or Taihanghshan) suture thus formed at circa 2.5 Ga or earlier. This does not mean, however, that rocks of the COB failed to experience younger sedimentation, deformation, and metamorphic events; just as the Taconic orogen was overprinted by the stronger effects of the Acadian orogeny in New England, as discussed above. So the COB was affected by younger and stronger (early Paleoproterozoic) events. Deng, H. et al. (2013) reported geochemical data from two suites of mafic rocks from the Zanhuang massif, including older blocks in mélange, and a younger suite of cross-cutting but deformed dikes. These dikes are in turn cut by undeformed 2.5 Ga granite dikes and pegmatites.
To explain this, Deng, H. et al. (2013) and Wang, J.P. et al. (2015Wang, J.P. et al. ( , 2016 suggested that after the circa 2.5 (or slightly older) arc/continent collision, the arc polarity was reversed from westwarddipping, to eastward dipping under the newly collision-modified margin of the Eastern Block, converting this margin for a short time into an Andean-type arc. In consequence, subsequent sedimentation, deformation, and metamorphic events in the COB could be related to deformation and magmatism in this arc, collision of outboard terranes, closure of the ocean on the outboard side of the accreted Wutai/Fuping arc (along the Trans-North China suture of Trap et al., 2012), or more regional events such as the postulated collision of the amalgamated NCC with the Columbia (Nuna) Continent along the northern margin of the craton at circa 1.85 Ga.

Fig. 5, 6 near here, Zanhuang
We next present structural details of this suture zone as they are not commonly discussed, and yet it is important to work out the different types of structures, their kinematics, and their orientations, because they provide key information about the geometry of accretionary and collisional events. Trap et al. (2012) 7). Wang, J.P. et al. (2013Wang, J.P. et al. ( , 2016 recognized an earlier fabric than the D1 of Trap et al. (2012), characterized by a scaly low-grade mélange fabric and intense imbrication of many structural slices in the Central Zanhuang mélange belt, and suggest that the D1 structures of Trap et al. (2012) are later cross-cutting high-grade shear zones. The D1 of Trap et al. (2012) therefore corresponds to D2 or D3 of Wang, J.P. et al. (2013Wang, J.P. et al. ( , 2016. D2 of Trap et al. (2012) is recognized as folds with NW-dipping axial surfaces and an S2 cleavage in the Hutuo Group, which unconformably overlies the Wutai and Fuping Complexes and is only weakly metamorphosed, so must be relatively young. D3 is attributed to normal A C C E P T E D M A N U S C R I P T

ACCEPTED MANUSCRIPT
18 shearing with localized S3 foliation developed along normal-sense shear zones such as the Pinshan low-angle normal shear zone. The S3 foliation is associated with an L3 lineation; these are considered to be late extensional structures formed after crustal thickening. D4 is attributed to late strike slip shearing, best exemplified by the km-scale, EW-striking, sinistral Zhujiafang Shear Zone (Fig 7). The structural history of Trap et al. (2012) is internally consistent, but did not consider the earlier events described by Kusky and Li (2003), Wang, J.P. et al. (2013Wang, J.P. et al. ( , 2015Wang, J.P. et al. ( , 2016 and Deng, H. et al. (2013Deng, H. et al. ( , 2014. Trap et al. (2012) constructed a quantitative P-T-t-D path with their D1-D4 representing the second major tectonic event. Their D1 metamorphic assemblage of qz+bi+mu+st+g+ky yields P-T conditions between 6.8 and 7.8 kb (with a range of 7-9.2 kb) and 650-660°C (however, they excluded the core 18% of the garnet from the analysis, which naturally would be expected to contain the earliest metamorphic assemblages), with a U-Th/Pb EPMA age from unzoned monazite of 1887+/-4 Ma . Interestingly, the D1 -M1 assemblage of Trap et al. Thus, it is clear that the D1 event of Trap et al. (2012) corresponds to D2 or D3 highergrade events of Wang, J.P. et al. (2013Wang, J.P. et al. ( , 2016, and the M1 event of Trap et al. (2012) corresponds to the higher grade and younger M2 event of Xiao, L.L. et al. (2014). Thus, the well-constrained CW P-T-t-D path of Trap et al. (2012) corresponds to the second major tectonic event in the socalled TNCO, and does not include the earlier, perhaps more significant, accretionary events related to the collision of the Wutai/Fuping Arc with the Eastern Block of the NCC.
The D1 event is associated with low-grade greenschist facies metamorphism shown best by inclusion trails in garnet porphyroblasts , but this M1 event remains undated in the Wutai Complex. In contrast, in the Zanhuang Complex the M1 event is dated at 2507 Ma (Xiao et al., 2014). Thus, although Zhao et al. (1999) (Li, J.H. and Quan, X.L., 1991;Kusky and Li, 2003), or is the Fuping Complex part of the arc that collided with the Eastern Block at 2.5 Ga? We test these possibilities using structural geology, sedimentology, geochronology, and geophysical profiles.
The Fuping Complex mostly comprises amphibolite-facies TTG gneisses with inclusions of mafic granulites, and it is characterized by multiple phases of deformation forming fold interference patterns (e.g., Zhang, J. et al., 2009Zhang, J. et al., , 2012Li S.Z. et al., 2010). It is intruded by circa 2077-2024 Ma monzogranites and granodiorites (Zhao et al., 2002c). Zhang J. et al. (2009) A C C E P T E D M A N U S C R I P T

ACCEPTED MANUSCRIPT
21 documented three phases of deformation in the Fuping Complex. D1 deformation includes rarelypreserved tight-to-isoclinal folds with associated axial planar fabrics and mineral lineations in mafic to pelitic rocks, which have variable orientations because of later overprinting deformations. L1 lineations are defined by syn-kinematic aggregates of clinopyroxene or hornblende and, although overprinted, indicate a NW-SE sense of thrusting and shearing on the S1 planes (Zhang et al., 2009). D2 strongly overprints D1 structures and formed meter-to kilometer-scale tight to isoclinal folds with an associated axial planar fabric. The folds are asymmetric and overturned, and associated with thrust faults, with SSE-to-E vergence and have hinges that plunge SW-NE, indicating NW-SE shortening (Zhang J. et al., 2009 (Liu, S.W. and Liang, H.H. 1997;Zhao et al., 2002;Kusky and Li, 2003). These rocks appear superficially similar to the shelf sequence deposited on the western edge of the Eastern Block, as described from the Zanhuang Complex. If true, then the Fuping Complex could represent a piece of the Eastern Block, repeated along major thrusts, with the Longquanguan Thrust representing a repeat of the Zanhuang (Taihangshan) Suture (e.g., Kusky and Li, 2003). However, Zhao et al. (2002b, c) and Zhang, J. et al. (2009) reported that a zoned zircon from a metapelite from the Wanzi assemblage has a near-concordant U-Pb zircon age of 2.11 Ga, which shows that the Wanzi sequence is much younger than the shelf sequence deposited on the western margin of the Eastern Block. Thus, we retain the interpretation that the Fuping Complex represents a deeper arc root to the 2.7-2.5 Ga arc, and that the Wutai volcanic and plutonic rocks represent higher levels of this same arc. This may be an entirely intra-oceanic arc, or it may represent an older microcontinent rifted from an unknown continent on the other side of the Luliang Ocean of Faure et al. (2007). Future geochronological studies should be able to resolve this issue. The young circa 2.11 Ga zircon crystal from the overlying Wanzi assemblage must be related to a younger event, just as in the Appalachians (see Fig. 3

), where
Devonian sediments overlie Ordovician arc rocks, and both were metamorphosed together in the Devonian Acadian orogeny. The Longquanguan shear zone may be an intra-arc structure or a post-collisional thrust formed during the 2.5, 2.4, 2.1, or 1.85 Ga tectonic events, or could be a hint that the Wutai/Fuping Arc is compound with different inter-arc elements sutured along this zone.

Eastern Hubei: Zunhua -Structural belt / suture/ Qinglong foreland basin fold belt
The eastern Hebei area contains a well-exposed cross-section of Archean crust that changes from a fore-arc accretionary complex containing ophiolitic mélanges and slivers, through a foreland fold-thrust belt, to a little deformed foreland basin that is cut by 2.4 Ga granitoids. A

ACCEPTED MANUSCRIPT
23 late Archean suture is preserved in the Zunhua mélange belt that separates the fore-arc accretionary complex from gneisses of the late Archean Taipingzai enderbiticcharnockitic gneiss complex (Fig. 8). The late Archean rocks of this belt are referred to as the Zunhua-Qinglong Structural Belt (Li, J.H. et al., 2002 a,b), or more simply as the Zunhua Structural Belt (ZSB: Kusky and Li, 2010). The ZSB comprises highly-strained metasedimentary gneiss, numerous tectonic slices of 2.6-2.5 Ga greenstones (mostly amphibolite facies metabasalts, gabbros, and ultramafic rocks with minor andesite and dacite), banded iron formations, and ophiolitic mélanges with metamorphosed blocks of basalt, gabbro, ultramafic rocks including harzburgite tectonite, dunite, and podiform chromite-bearing serpentinites (Li, J.H. et al., 2002;Huang et al., 2004;, Kusky, 2011a. Algoma-type banded iron formations (BIFs) that are interpreted to have formed in a fore-arc environment at 2541 +/-21 Ma to 2553 +/-31 Ma contain zircons with metamorphic rims yielding ages of 2512+/-13 and 2510 +/-10 Ma (Zhang., L.C. et al., 2012). The ZSB exhibits many east-vergent folds with west-dipping axial surfaces, sliced by numerous NE-striking shear zones. The belt is intruded by numerous 2.6-2.5 Ga tonalite-trondhjemite-granodiorite rocks that are now gneisses, by 2.5 Ga granites, and is transected by numerous ductile shear zones. The whole complex is thrust over the Taipingzhai gneiss complex, and the linear structural patterns in the ZSB are clearly discordant with the more domal-style structural fabric of the early Archean granulitegneiss dome of the Taipingzhai complex (Fig. 8). The Paleoproterozoic Chengde-Hengshan high-pressure granulite belt overprints the northwestern part of the belt, and is cut by numerous circa 300 Ma plutons.
The ZSB is noteworthy for two remarkable features: two large circa 2.5 Ga ophiolitic slices (NW Belt and Shangyin slices of the Dongwanzi Ophiolite, DWO), and the 2.5-2.6 Ga Zunhua podiform chromite bodies (Figs. 8,9) in an ophiolitic mélange (Kusky et al., 2001;Li et al., 2002a;Huang et al., 2004). Since the original definintion of the DWO in 2001, the Central M A N U S C R I P T

24
Belt has been shown to consist mostly of circa 300 Ma Paleozoic plutonics (Kusky et al., 2004;Kusky and Zhai, 2012) with a few rafts of the older Archean and Proterozoic basement, so we drop the former, now defunct, Central Belt from the DWO, and just include the SE belt (the Shangyin ophiolitic sheet) and the NW belt in our classification of the DWO.
However, it must be empahsized that even though the DWO is cut by some younger plutonic rocks, the the NW Belt and the Shangyin ophiolitic sheets have Archean ages, and represent wellpreserved relicts of a dismembered and metamorphosed Neoarchean ophiolite (Kusky and Zhai, 2012).

Fig 8 near here -Eastern Hebei map
The Shangyin ophiolitic sheet has a preserved basal thrust zone, which includes an ophiolitic mélange along the base that grades up to a harzburgite tectonite and a mantle transition zone with a well-exposed Moho (Julian Pearce, pers. comm., 2014). The mantle transition zone consists of a circa 2-km thick interlayered harzburgite, mafic and ultramafic cumulates, and gabbro. This in turn grades up into gabbro, then high level gabbros with local dike complexes, and these are in structural contact with several tens of meters of well-preserved pillow lavas, with rare inter-pillow cherts, and with fault slices of BIF (Fig. 10;Kusky et al., 2001Kusky et al., , 2004. Southwestwards the Shangyin ophiolitic sheet is imbricated by shear zones that continue into the Zunhua ophiolitic mélange belt (Figs. 8,9). In the southern part of the ZSB near Zunhua, the ophiolitic mélange contains blocks of harzburgite tectonite, dunite, podiform chromite, cumulate gabbro, isotropic gabbro, and lenticular amphibolitic units that were likely original pillow lavas and/or dike complexes. In the Shangying ophiolitic sheet these units are better preserved with magmatic transitions from the cumulate ultramafic rocks to gabbro, to gabbro A C C E P T E D M A N U S C R I P T

ACCEPTED MANUSCRIPT
25 intruded by contemporaneous dikes, to amphibolite facies metabasites including basaltic flows and pillows. Small pods and beds of silica (Fig. 10) are interpreted as interpillow cherts, and beds of BIF as volcanogenic exhalative deposits (e.g., Li. J.H. et al., 2004). The podiform chromites (Fig. 11) are unique because they preserve some of the best nodular and orbicular textures in any Archean ophiolite and are very similar to those in ophiolites such as Semail in Oman, Troodos in Cyprus, and Josephine in the California Coast Ranges (Li, J.H. et al., 2002a).

Figs, 9, 10, 11 near here
The ages of the Shangyin ophiolitic sheet and associated Zunhua podiform chromites are well-constrained (Fig. 12). U-Pb ages from gabbros from the Shangyang ophiolitic sheet yield ages for the gabbro section of 2505+/-2 Ma (Kusky et al., 2001), and Re-Os ages on chromites from the Zunhua podiform deposits yield ages of 2.5-2.6 Ga (Kusky et al, 2007c). Peridotites from the base of the Shangyang sheet and Zunhua also yield Lu-Hf ages of 2.55 Ga (Polat et al, 2006), showing that the mantle and crustal sections of the DWO and Zunhua podiform chromitites are contemporaneous. Claims that the DWO cannot be an ophiolite, because it is cut by circa 300 Ma mafic to felsic magmatic intrusive rocks , are not supported by the exposed field relations and high-precision isotopic dates of Archean age on wellcharacterized samples from the Shangyin ophiolitic sheet (Kusky et al., 2001Kusky and Li, 2008;Polat et al., 2006), and it can be clearly shown that every other Precambrian unit in the region is also intruded by such younger magmatic rocks (Fig. 8). The chemistry of the chromites in the Zunhu Structural Belt has also been debated. Li et al. (2002) reported orbicilar and nodular textures in podiform chromites (Figs. 11,12), and noted that these texture are only known from ophiolites of any age on the planet. Zhang et al. ( , 2004 reported that the chemistry of the A C C E P T E D M A N U S C R I P T

ACCEPTED MANUSCRIPT
26 chromites in the Zunhua-Structural Belt was more like that of a continental intrusion rather than an Alpine-type peridotite. However, analyses of chromites from the Zunhua belt by Polat et al. (2006), which excluded chromites of uncertain age from the younger pluton in the central belt of Dongwanzi, and did not include portions of the chromites altered to ferrit-chromite, all plot within the characteristic fields of chromites and spinels in ophiolites. The samples of Zhang et al. ( , 2004 have no specified locations, no associated field studies, and appear to have analytical defects since they scatter widely over the Cr/(Cr=Al) vs. Fe2+/(Fe2+=Mg) discrimination plots (Polat et al., 2006, Fig. 7). The podiform chromites from Zunhua have nodular and orbicular textures (Fig. 11). The podiform chromites from Zunhua have nodular and orbicular textures (Fig. 11) only found in ophiolitic chromitites, and their ages are the same as those in the crustal section. The Eastern Hebei area (and its extensions to Liaoning to the north and Wutai Shan to the south) is also host to large banded iron formations (BIFs; Zhai and Windley, 1989

ACCEPTED MANUSCRIPT
27 structurally underlain by the Luxiang Group (Wu, C.H. et al., 1998), which contains a lower unit of metabasalt and tuff interpreted to be rift-related, succeeded by shallow water sedimentary strata including quartz-mica schist, marble, sandstone, and banded iron formation that Kusky and Li (2003) and Li and Kusky (2007) interpreted as a remnant of the 2.7-2.5 Ga passive margin developed on the western margin of the Eastern Block. The main sedimentary fill of the Qinglong Basin is called the Qinglong Group, the lower part of which consists of interbedded meta-shales and graywackes with well-preserved Bouma sequences, siltstones, and BIF, and an upper part of coarser-grained conglomerates and sandstones. Pebbles in the upper unit consist of mafic volcanics, vein-quartz, granodiorite, andesite, quartz diorite, and metasediments (Qi, H.L. et al., 1999) and the conglomerate is several meters to 400 meters thick. Pebbles are coarser in the western side of the basin, suggesting derivation from the west, and sedimentary analysis suggests that the basin was bound by thrusts in the west, and deepened in that direction (Bai, J., et al., 1996). This sequence was interpreted as a flysch-to-molasse transition by Kusky and Li (2003) and Li, J.H. and Kusky (2007), and alternatively as a rift sequence by Lv, B. et al., (2012). Kusky and Li (2003) and Li and Kusky (2007) correlated the Qinglong foreland basin with rocks of the Hutuo Group in Wutaishan, but more recent work (Wilde et al., 2004; has shown that the Hutuo Group is much younger, so we abandon this correlation and relate the Hutuo Group to younger events, discussed below. The Qinglong Basin (Fig. 8) is deformed by asymmetric, tight to isoclinal east-vergent folds with a penetrative axial planar cleavage, is cut by two major west-dipping thrusts (Qi, H.L., et al., 1999), and pebble-elongation lineations as well as stretching lineations plunge westwards, all suggesting that this basin is a foreland basin derived from erosion of the orogenic belt to the west. Sedimentary rocks of the Qinglong Basin are cut by abundant circa 2.4-2.5 Ga diorites (Li, J.H. and Kusky, 2007), and are overthrust by the 2.54 -2.64 Ga Shuangshanzi Group, the 2.51

ACCEPTED MANUSCRIPT
28 Ga Dongwanzi ophiolite, and the 2.55-2.51 Ga Zunhua Mélange to the west. The Qinglong Basin therefore records the history of rifting of the western margin of the Eastern Block sometime between 2.7 and 2.5 Ga, the development of a thin passive margin sequence, then deposition of a foreland basin with a flysch-to -molasse transition by 2.5 Ga during emplacement of a fore-arc accretionary wedge bearing ophiolitic slivers and mélanges. The boundary between the foreland basin (and underlying shelf) and the accretionary wedge marks the suture between the Eastern Block and the COB, and the entire sequence represents a classic record of a Wilson Cycle preserved in an Archean suture zone. Geological relationships in Eastern Hebei are similar to those farther south in the Zanhuang massif, so we correlate this Zunhua suture with the Zanhuang suture (Taihang suture of Trap et al., 2012), and recognize it as the leading edge of the late Archean arc-continent collision between the Eastern Block of the NCC and the Wutai/Fuping Arc terrane in the COB.

Fig 14 near here.
Based on the rock types, depositional and intrusion ages, and the timing and grade of metamorphism, we suggest that the Jianping Complex is comparable to the Wutai and Dengfeng Complexes, and represents part of the arc complex that collided with the Eastern Block of the NCC at circa 2.5 Ga. The 2.5 Ga suture therefore lies farther to the east of the Jianping Complex.
The presence of podiform chromites in blocks of dunite/harzburgite in a metasedimentary mélange belt in the Jianping Complex (Li JH et al., 2002a) suggests that this belt is similar to the Zunhua Structural Belt, and that the suture is not far away to the east in the subsurface.

South: Denfeng Complex
The Dengfeng Complex in the southern part of the COB contains a mixture of TTG gneisses, metamorphosed Archean diorites, meta-sandstones and metapelites (

Tracing the 2.5 Ga Zunhua-Zanhuang (ZZ) suture in the NCC
Using traditional geological relationships, we are able to use the data discussed above to trace the circa 2.5 Ga suture between the Eastern Block and the accreted Wutai/Fuping Arc in the Central Orogenic Belt (COB) for more than 1,300 km across the NCC. This suture, which we previously named the Zunhua-Zanhuang Suture (e.g., Kusky, 2011;Kusky et al., 2012), and Trap et al. (2012) alternatelynamed the Taihangshan suture, has been widely agreed to separate a circa 1,300 km long arc terrane (the Wutai/Fuping Arc) in the COB (with an east-vergent accretionary wedge attached to its eastern side) from the Eastern Block of the NCC (e.g., Kusky, 2011a,b;Kusky et al., 2001Kusky et al., , 2004Kusky et al., , 2007aKusky et al., ,b,c, 2013Li et al., 2002;Li and Kusky, 2007;Kusky and Li, 2003;Polat et al., 2005Polat et al., , 2006Deng et al., 2013Deng et al., , 2014Deng et al., . 2016Wang W. et al., 2013Wang W. et al., , 2015. In their petrological study Wang W. et al. (2015) report that these rocks formed in an evolving intra-oceanic arc system with five different magmatic suites with affinities to MORBs, island arc tholeiites (IAT), calc-alkaline basalts (CAB), high-magnesium andesites (HMA), and adakites. They relate this to an evolving intra-oceanic arc system that was initiated by partial melting of depleted to slightly enriched asthenospheric mantle at a spreading center that generated the MORB. Incipient intra-oceanic subduction at ~2550 Ma metasomatized the mantle and generated the IAT, CAB, HMA and adakites (Fig. 18), and high-Mg TTGs from

ACCEPTED MANUSCRIPT
34 2550 to 2506 Ma, and partial melting of the arc root by underplated mafic melts generated a suite of low-Mg TTGs . Collision of the arc with the Eastern Block at 2490 Ma caused the regional granulite facies metamorphism and also generated K-rich granitoids from crustal anataxis . Following the models of Kusky and Li (2003) and Kusky (2011, and references therein),  correlated the arc-rocks in this Fuxin greenstone belt with others in the COB, all the way south to the Wutaishan belt, and suggested that this was a large intra-oceanic arc system that collided with the Eastern Block above a west (or NW)-dipping subduction zone (e.g., Fig. 18).
There are hints that there may be significant along-strike variations in the accreted  (2002b) reported xenocrystic zircons with ages of 2.7 Ga from the Longquanguan gneiss also in the Wutai Greenstone Belt, suggesting that there may be an older basement in the root of the arc, or that zircon-bearing sediments eroded from an older orogen or craton were deposited in the trench, subducted, and incorporated into the arc magmas. Thus, it is likely that the Wutai/Fuping Arc system in the COB had significant along-strike variations in basement character, or that the COB contains a more complex system of amalgamated arcs of different origins. Alternatively it is possible that this arc system resembled the present-day Aleutians, where an oceanic arc merges along strike into an arc built on older accreted continental crust (e.g., Kay et al., 1982).
In the other direction towards the hinterland of the orogen, the accretionary wedge rocks pass, through structural imbrication, into the allochthonous arc sequence (e.g., compare Figs. 2 and 18). greenschist to granulite facies metamorphism (with CCW P-T-t paths) at 2.6-2.5 Ga, whereas rocks of their "TNCO" only experienced regional metamorphism at circa 1.85 Ga (with CW P-Tt paths). Zhao and Zhai (2013) suggested that this is because both the Eastern and Western blocks were located above separate mantle plumes at circa 2.6-2.5 Ga, whereas the "TNCO" was not metamorphosed until 1.85 Ga during a continent/continent collision between the two separate, plume-impinged blocks at that time. Here, we list data that show that rocks of the "TNCO" (roughly corresponding to the COB) did indeed experience regional high-grade metamorphism at circa 2.5 Ga, which can be related to an arc/continent collision at that time. Descriptions of the widespread circa 2.5 Ga event in the Eastern Block are abundant and can be found in Pidgeon, Jahn, B.M., and Zhang, Z.Q., (1984), Jahn, B.M., et al. (1988Jahn, B.M., et al. ( , 2008, Wan, Y.S., et al. (2001Wan, Y.S., et al. ( , 2005 in the south, a distance of more than 1,300 km (see Supplementary Data Table 1 for details).
Based on the few, but well-documented ages of 2.5 Ga for the M1 metamorphism in the COB, we discount the many other descriptions where M1 is undated and yet used to suggest a prograde section of a 1.85 Ga CW P-T-t path in order to support a tectonic model of a continental collision between the Eastern and Western Blocks at that time (e.g. Zhao et al., 1999Zhao et al., , 2001Zhao et al., , 2005Zhao, 2009;2011;Zhao and Zhai, 2013 Thus, the metamorphism in the COB is more complex than most workers on the NCC have reported, because there are at least two or more different P-T-t paths needed for the different orogenic events. It is inappropriate to force all the data into a single P-T-t path that fits a predetermined tectonic history. Collisions are not associated with 700 Ma + P-T-t histories (e.g., Dewey, 2005). The so-called paths are, we suggest, different points representing the approximate peak metamorphic conditions of a 2.5 Ga orogenic event, a 1.8 Ga orogenic event, and either a younger event or the retrograde path of the M2 assemblage.

Reversal of Subduction Polarity?
It has recently been proposed that, after the Wutai/Fuping Arc collided with the Eastern Block at ca. 2.5 Ga, the subduction polarity was quickly reversed from west-dipping to east- (Kusky, 2011;Deng, H. et al., 2013;Wang, J.P. et al., 2013Wang, J.P. et al., , 2015 with a new subduction zone developed on the back side (western margin) of the Wutai/Fuping Arc (Fig. 18b).
An active example of arc-polarity reversal is taking place where the eastern Sunda Arc is colliding with the NW Australian continental margin, and a back-thrust (Flores Thrust) has formed in the back-arc region, marking the initiation of subduction-polarity reversal (Fig. 19), which is propagating westward as the collision progresses, and is happening only a few millions to tens of millions of years after the initial collision of the Sunda Arc with Australia (e.g., Reed et al., 1987;Rutherford et al., 2001). On Fig. 19, note the remarkable similarity of scale and geometry between the Sunda Arc/Australia collision and arc-polarity reversal, and that proposed here for the Wutai/Fuping Arc collision with the NCC and its arc-polarity reversal.
This subduction-polarity reversal event released slab-derived melts beneath the collisionmodified margin of the Eastern Block, initially generating mafic dike swarms (Deng, H. et al., 2013(Deng, H. et al., , 2014, which in turn added heat to the base of the crust, which led to a suite of granitoids across the Eastern Block (Wang, J.P. et al., 2015). The heat from this mafic underplating and the granitoids associated with it, are here suggested to be responsible for the widespread HT -LP metamorphism, (with CCW paths) in many places in the Eastern Block (see references above). Lu et al. (2008), Wu et al. (2012),  suggested that the heat for this metamorphism was from a mafic underplate related to a mantle plume, but we suggest that it was from a mafic underplate derived from fluids that induced partial melting of the underlying metasomatized mantle wedge (Fig. 18b), or from slab melting, and thus a consequence of the arc-

ACCEPTED MANUSCRIPT
39 polarity reversal. In either case, the resulting PT-t paths would be similar, showing a CCW trend, so it is difficult to differentiate between these models using only PT-t paths and without a regional tectonic synthesis. Whilst there was an ocean behind the accreted Wutai arc (named the Luliang Ocean by Trap et al. (2009)), the western margin of the Eastern Block was an active continental margin, or an Andean-style arc, as envisioned by Zhao et al. (2001, etc.). However, there is still considerable debate and uncertainty about how long this ocean remained open, whether until 1.85 Ga (Zhao et al., 2001, etc.), 2.4, 2.3 or 2.1 Ga (Trap et al., , 2008(Trap et al., , 2009a(Trap et al., , b, 2012Wang, Y.J. et al., 2004;Wang Z.H., et al.;2010;Wang, Z.H., 2009), or only for a short time. We discuss this in the next section on the "Trans North China Suture".  Peng, P. et al. (2015) to suggest that this sequence evolved above a "mantle wedge-absent hot subduction zone, in which the ultramafic-mafic rocks originated from undifferentiated mantle, the rhyolite was derived from eclogite-facies crust of the overriding plate, the quartz diorite resulted from mixing of mantle melts and the overriding crust, the TTG suite was derived from partial melting of the subducting slab at amphibolite to amphibole-bearing eclogite facies conditions, and the quartz monzodiorite was generated by melting of the overriding midlower crust. The whole history of arc magmatism lasted from 2570 Ma to 2480 Ma, from arc initiation, through maturation, to exhumation presumably during collision, and the magmatic sequence could have formed in a continental arc with special Archean conditions, in which there was slab melting instead of slab dehydration, and hydrous minerals were preserved in the slab through the eclogite facies. If the subducting oceanic slab was thick and had limited dehydration, there would be no modern-style hydrous mantle wedge (c.f. Fig 18b, where we retain the hydrous mantle wedge, but not beneath the accreted arc, as suggested by , and a 'hot-subduction system" would result, producing the magmatic series documented from the Qingyuan Belt. We suggest that this short-lived Andean-style arc may also have formed after the subduction polarity reversal event (which was slightly earlier in the north than in the center of the COB), and that is the reason why the entire history of the arc, from formation to exhumation lasted only <80 million years.

Where is the next suture, the Trans North China Suture, and how old is it?
Little is known about when the remaining open part of the "Luliang" Ocean (Trap et al., 2012) closed, since most workers in China have simply assumed that the borders of the TNCO as defined by Zhao (2001) represent the borders along which the eastern and western blocks collided at 1.85 Ga. However, a few studies have looked deeper into this "second suture", and found several features that indicate that the closure time was, may be, at 2.5-2.4, 2.3, 2.1, or 1.85 Ga. Faure et al. (2007) and Trap et al. (2007Trap et al. ( , 2008Trap et al. ( , 2009aTrap et al. ( , b, 2012 provided detailed structural documentation of a suture that they termed the "Trans North China Suture" along the western side of the accreted Wutai/Fuping Arc terrane. This suture is best exposed in the Luliang Massif (LL on Fig. 17) where it crops out as greenschist facies metasedimentary rocks mixed with mafic and ultramafic rocks, which have an oceanic affinity (Trap et al., 2009b;2011;Polat et al., 2005). Trap et al. (2011) included the flysch, mafic rocks, pillow basalts and other sedimentary rocks of the Wutai Greenstone Belt in an allochthonous unit (called the Low Grade Mafic Unit, LGMU in their terminology) that was extruded from the Trans North China Suture and thrust over TTGs and migmatites of the Fuping Block (Fig. 7). Interestingly, the pillow basalts, gabbros, and felsic volcanic rocks in this suite have ages of 2.530-2.515 Ga  whereas the underlying TTG gneisses of the Fuping Complex have ages of 2.560-2.515 Ga . This in turn implies that the Luliang Ocean closed not too long after 2.5 Ga, since it is unusual for oceanic-affinity rocks in sutures to be older than the ocean formation age, and most are obducted soon after they form (Burke et al., 1977;Dewey, 1977). Strangely, Trap et al. (2012) interpreted these rocks to have formed in a basin that rifted from a previously amalgamated NCC at 2.3-2.2 Ga, and therefore the Luliang Ocean closed at circa 1880 Ma.
It is clear that there were significant magmatic and partial melting events between 2.3 and 2.1 Ga in the Fuping, Wutai, and Hengshan areas, and also along the northern margin of the NCC, Ordos Block, Alxa Block, and Eastern Block (Fig. 17 Trap et al. (2012) interpreted these to reflect a regional anatectic melting event within the TNCO, and Zhao et al. (2008, etc.) interpreted them as products of Andean-type arc magmatism related to eastward subduction beneath the TNCO. In contrast, Kusky and Li (2003) interpreted these intrusions to form a wide zone of magmatism related to a convergent margin and Andean arc-related activities stemming from subduction under the northern margin of the craton, 200 km to the north, and stretching 1,600 km EW along the northern margin of the craton.
Note that magmatism associated with the present-day Andean arc extends some 500 km from the convergent boundary, so this is not an unusually large distance (Fig. 17). In recent papers  and Zhang, C.L., et al. (2015) reported U-Pb ages and chemical data from circa 2.2-2.0 Ga granitic gneisses from boreholes in the Ordos basement, and interpreted them to be part of a continental margin arc. This is in perfect agreement with the interpretation that the northern margin of the craton was an Andean-style arc at this time (Kusky andLi, 2003, Kusky, 2011a), but it would have been impossible to produce, if the TNCO had not closed by then, and the Andean arc was located above an east-dipping subduction zone beneath the eastern NCC (e.g., Zhao et al., 2001, etc H. et al., 2014). We therefore suggest, based on this limited information, that the Luliang Ocean closed by 2435 Ma, and the Eastern and Western Blocks were amalgamated at this time (Fig. 20).

43
The magmatic gap from 2435-2300 Ma represents the time between closure of the "Luliang" Ocean, and when the Andean arc was set up along the northern margin of the craton.
The nature of the basement to the Western Block is enigmatic, since it is largely covered by late Archean to modern sedimentary rocks. However, Kusky and Mooney (2015) synthesized geophysical, geological, and geochronological data on the nature of the Ordos (part of the Western Block) basement, and suggested that it is likely an oceanic plateau that was accreted (Fig. 20a,b), and experienced several periods of later differentiation during younger subduction and collision events along the northern margin of the craton.

Post-orogenic extension and rifting at 2.4 Ga
There has been much recent work along the northern margin of the NCC, focusing on the ages and PT conditions of HP and UHT metamorphism (e.g., Guo, J.H., et al., 2006;Santosh and Sajeev, 2006;Santosh et al., 2007a, b;Wan, Y.S., et al., 2009;Santosh and Kusky, 2010;Li et al., 2011;Zhai and Santosh, 2011 and references therein;Guo et al., 2012;Yin, C.Q et al., 2009Yin, C.Q et al., , 2011Wan, B. et al., 2015), but a paucity of geological mapping for tectonic discrimination. Kusky and Li (2003) and Zhai and Santosh (2011) suggested that, following the amalgamation of the East and West Blocks through accretion of the Ordos Oceanic Plateau  (Kusky et al., 2007a, b;Kusky and Santosh, 2009). Zhao and Wilde (2002) also recognized this belt, but suggested that it was older than the collision of the Eastern and Western Blocks. Condie (1992) referred to a belt of granulite facies meta-pelitic rocks intruded by S-type granites south of this belt as the "Khondalite Belt", also taken up by Zhao et al. (2009).
Rocks in parts of the IMNHO have also been referred to as the "Yinshan Block" Zhai and Santosh, 2011), separated from the Ordos Block by the Khondalite Belt (which has also been called the "Inner Mongolia Suture Zone" (Santosh, 2010).
The northern part of the IMNHO is a strongly tectonized metasedimentary belt that consists of deformed shallow water sedimentary rocks, to the south of which is a predominantly plutonic belt including TTG-quartz diorite plutons, and younger granodiorite, metamorphosed to greenschist through amphibolite facies (Li, J.H. et al., 2000a,b;Kusky and Li, 2003;Kusky et al., 2007b). South of this is another metasedimentary belt that is intruded by gabbro and diorite complexes, metamorphosed to amphibolite facies. Kusky and Li (2003) (Li N., et al., 2015), and like other metapelites across the craton, were metamorphosed to granulite facies between 1.95 and 1.85 Ga. conclude that "the present dataset seems to support the idea that the Alxa Block is part of the Paleoproterozoic IMNHO", and the similarities between the Beidashan Complex and the Guyang and Wuchang Complexes also suggest that the Yinshan Ribbon Micro-Continent extends all the way to the Alxa (Alashan) Block.

The Khondalite Belt
The Khondalite Belt (Condie et al., 1992) forms the southern margin of the IMNHO, but the southern margin of the Khondalite Belt is currently unknown because its buried equivalents extend at least half way across the Ordos Basin Wang W et al., 2014;Zhang CL et al., 2015). A > 1.90 Ga molasse basin lies along the northern margin of the Khondalite Belt, separating it from the Yinshan Ribbon Micro-Continent. Granulites of the Khondalite Belt are best-exposed in the Jining, Liangcheng, Fenzhen, Daqingshan, Wulashan,

ACCEPTED MANUSCRIPT
50 and Helanshan Massifs (Fig. 17), and include assemblages of khondalite, charnockite, and metagabbro, intruded by S-type granites which have yielded U-Pb ages between 1.97-1.83 Ga (Guo et al., 1999;Kusky and Santosh, 2009 We suggest therefore that the Khondalite Belt and its equivalents to the east preserve several noted that the Halaqin volcano-sedimentary sequence is highly deformed, including ductile thrusts, isoclinal folds and sheath folds, and that the whole succession is a thick tectonic pile thrust from the SE to the NW over the Yinshan Ribbon Micro-Continent, and that it likely represents an accretionary wedge in the Khondalite Belt located between the accreted Yinshan Block and the amalgamated Eastern Block/Wutai/Fuping Arc (Huai'an terrane). They emphasized, however, that no modern structural studies have been undertaken on these complex rocks.

Fig 22 Comparison of NCC with Andes near here 4.3.4 Andean arc-related magmatism
Andean-style magmatism affected much of the northern, central, western (including the Alxa Block), and northern part of the eastern NCC from 2.3-1.88 Ga (see Zhao et al., 2005Zhao et al., , 2008Zhao et al., , 2010Zhao and Zhai et al., 2005, 2010, Peng et al., 2012bZhai and Santosh, 2013;, with a strong magmatic pulse that affected regions as far south as the central  Wilde et al., 2003;Yu et al., 1997;Zhang et al., 2010;Zhao et al., 2008b;. This suite of rocks has a continental arc geochemical signature , including variable Hf isotopic compositions similar to continental margin arc rocks in the Central Zone and Eastern Block. Their presence throughout the Western Block does not support the tectonic model (Zhao et al., 2001 that the "TNCO" closed by eastward subduction of an oceanic plate attached to the Western Block beneath an Andean arc developed on the Eastern Block, since the arc-related magmas are now known to extend in a broad E-W arc that perpendicularly crosses the proposed 1.85 Ga suture of the TNCO (Fig. 22c).
Some of these are deformed into granitic gneisses, whereas others are relatively undeformed. The dikes are tholeiitic in composition and were interpreted by Peng et al. (2012b) to have been derived from the Archean sub-continental lithospheric mantle, perhaps in a rift setting. However, A-type granites and the other igneous rocks in this suite can also form in Andean-arc settings, especially in places where the arc is fairly mature and the granites are derived from crust that has already produced a suite of orogenic granites, leading to the formation of A-type granites (e.g., dikes intruded, so the crust had already produced numerous melts. From different Andean settings "from Antarctica to Alaska", Kay and Rapela, (1990) have shown that it is difficult to assign A-type granites and other subduction-related magmas to specific tectonic settings based on geochemistry, since magmatism in one setting, such as a volcanic arc, rift, continental collision, etc., may overlap in composition considerably and many different types may occur in a single subduction-related setting based on the relative proportion of crustal material involved in the magma genesis.
The Inner Mongolia -Northern Hebei Orogen also contains an assortment of TTG to dioritic gneisses, 2.2-1.9 Ga granites, metasedimentary and metavolcanic rocks, and rare gabbroic to ultramafic intrusions (Kusky and Li, 2003;. Deformation is characterized by roughly EW-striking shear zones and folds that extend at least as far south as the Hengshan, which is cut by the E-W striking Zhujifang and Datong-Chengde Shear Zones (Fig. 17) which disappear under young cover of the Ordos Basin. Geophysical studies reveal a series of NEstriking faults in the basement beneath the cover of the Ordos Basin, and it is possible (although speculative) that some of these may be reactivated extensions of the Zhujifang fault (Figs. 17,22) or the Datong -Chengde Shear Zones.
In the Jiao-Liao Ji Belt in the Eastern Block (Fig. 22), a group of magnetite monzogranites intruded at circa 2176-2166 Ma, and a suite of hornblende-biotite monzogranites was emplaced between 2150 and 2143 Ma . Nd isotope geochemistry reveals that the monzogranites have model ages (TDM) of 2.4-2.6 Ga (Li S.Z. et al., 2006) and show evidence of derivation from a source that includes partial melting of the Archean crustal basement. The magmatism is associated with a group of volcaniclastic sediments, pelites, and flood basalts, , that we relate to processes in a retro-arc foreland basin in the next section.

A C C E P T E D M A N U S C R I P T
ACCEPTED MANUSCRIPT 54

Tectonics
The northern part of the IMNHO was thrust to the south and SE over the NCC (Kusky and Li, 2003;Peng et al., 2011), forming widespread south-vergent folds and thrusts in the Khondalite Belt and intrusion of numerous S-type granites between 2.2 and 1.90 Ga (Kusky and Li, 2003). In the current interpretation, the Yinshan Ribbon Micro-Continent is part of the accretionary orogen of the IMNHO, and the accretionary wedge and arc built on the microcontinent were thrust over the shelf sequence, shedding flysch, later to become the so-called Khondalite Belt (Kusky and Mooney, 2015). Zhai and Peng (2007) suggested that events in this time period (previously referred to as the Luliang Movement) can be divided into a Wilson Cycle sequence of events from rifting at 2350 Ma, followed by subduction and collision by 1970 Ma, then a regional high-grade metamorphic event at 1950-1820. The former phases of Zhai and Peng's orogenic cycle correspond to the events proposed by Kusky and Li (2003), and Kusky (2011a, b), and the regional high-grade event likely records the incorporation of the NCC into the Columbia (Nuna) Continent (Kusky and Santosh, 2009;Wan et al., 2015).

Events further inboard from the Andean margin
In Wutaishan, the Hutuo Group has historically been regarded as Archean in age, but more recent geochronological studies have shown that it is much younger and corresponds in age to the magmatism associated with the Andean arc on the northern margin of the craton. The Hutuo Group (Fig. 23)   interpreted through non-seismic geophysical methods as a fault . A N-S residual gravity profile Fig. 24) shows that the basin thickens to the north and shows obvious N-over-S thrust structures, but  concluded that the basin- It is interesting to compare the scales of deformation, volcanism, and plutonism in the present-day Andes to the North China Craton in the interval between 2.3-1.95 Ga, when we propose that the northern margin of the craton was an Andean arc with related tectonic belts.

ACCEPTED MANUSCRIPT
58 22a shows a simplified map of the central Andes, with an outline map at the same scale of the NCC plotted over it so that the proposed Andean margins are parallel. Note that in the Andes the magmatism is discontinuous, the deformation front lies 600-1,000 km from the trench, and that much of the region under the Altiplano is characterized by double-thickness crust. Thus, if this crust were isostatically eroded to 35 km normal thickness, we would be looking at the remnants of a widespread granulite facies metamorphic event, affecting an area roughly the size of the whole NCC. Since the width of the area affected by Andean-related tectonism ranges from 600-1,000 km, we show two lines, 600 and 1,000 km away from the northern margin of the craton on

59
The 1.92 UHT metamorphism is best documented locally in the Alxa area (Wan, B. et al., 2015), and in the Jining complex (in the Bao'an, Hongshaba, Xumayao and Tuguiwula areas) where it appears to be associated with contact metamorphism near large gabbronorite intrusions ,which in turn may be associated with a ridge subduction event that preceded a major continent-continent collision (Santosh and Kusky, 2010;Peng et al., 2011;and Wang W. et al., 2013). If so, that would mean the Andean arc on the northern margin of the craton was still active at circa 1.92 Ga.
The circa 1.85-1.80 Ga event has been interpreted by most workers in China, following Zhao et al. (2001Zhao et al. ( , 2005Zhao et al. ( , 2009, to be the result of the "final" amalgamation of the Eastern and Western Blocks along the TNCO. However, this has been debated for years by Kusky (Kusky, 2011;Kusky andLi, 2003, 2010;Kusky et al., 2007a, b, c;Kusky and Santosh, 2009;Kusky and Zhai, 2012;Polat et al., 2005), and more recently by Peng P. et al., (2014) who demonstrated that the metamorphic data (from which this purported orogen is defined) are so similar inside and outside the TNCO that they provide no basis for its definition as a separate orogenic belt. We elaborate on this below, and then provide an alternative, actualistic interpretation of the existing data.
In detailed analyses of Paleoproterozoic metamorphic events across the entire NCC, Kusky (2011) and Peng, P. et al. (2014) concluded that there is no evidence for the existence of the so-called TNCO as a Paleoproterozoic orogenic belt. They noted that the spatial distribution of circa 1950-1800 Ma metamorphic events are widely distributed across the craton, but concentrated, and at higher grade, along the northern margin of the Craton, and along the southern margin (Fig 24). Peng, P. et al. (2014) used statistical analysis of multiple data sets on all reported metamorphic events in this time-frame, building on earlier claims of Kusky and Li (2003), , Kusky and Santosh, 2009), Zhai and Santosh (2011), and Kusky boundaries defined as the TNCO (Zhao et al., 2001, etc.). In the Kusky et al (2007a) interpretation, the metamorphism of this age was related to a continent-continent collision along the northern margin of the craton when the already-amalgamated NCC joined the Columbia/Nuna Continent. In the new Peng, P. et al. (2014) interpretation, the metamorphism in this time-frame was related to both a collision along the northern margin of the craton, and another along the southern margin of the craton, suggesting perhaps that the NCC was located in a more interior part of the Columbia Continent than in the reconstructions of Kusky et al. (2007a), Kusky andKusky (2011a, b). The above detailed data analysis showed that there is nothing unique in terms of metamorphic history about the so-called TNCO, and that it does not exist as a Paleoproterozoic orogen. The TNCO was defined as on orogen based on the claim that rocks in a N-S striking zone, bound by Mesozoic faults, contained different circa 1.85 Ga CW P-T-t paths than areas outside of these Mesozoic faults, which were claimed to show CCW paths at that time. Even ignoring the fact that orogenic belt boundaries cannot be defined by structures that are 1.7 billion years younger than the deformation and metamorphism, it is now clear that there is nothing distinct about the metamorphic paths, or their timing, in locations within the socalled TNCO and outside it. The circa 1.88-1.79 metamorphic event was a "pan-North China Craton" event that affected the whole craton, and is recognizable almost everywhere that rocks of appropriate age are exposed (Zhai, M.G. 2014). Metamorphic grades are higher in the north, and EW-striking structures place higher P assemblages to the N over lower-grade assemblages to the S (Fig. 25). The delineation of the "TNCO" was based solely on this metamorphic interpretation and recrystallized zircon ages, and did not include any regional analysis of tectonic zonations, structural history, or other types of data used for tectonic analysis and definition of suture zones between different, once-widely-separated terranes. Thus, there is no evidence that a N-S striking

An actualistic interpretation of the circa 1.85 Ga Pan-NCC metamorphic event
The main metamorphic event in the NCC saw granulite facies conditions across much of the craton at 1.85-1.80 Ga, with HP granulites and garnet websterites (2.5 GPa) in the north (Wan, B. et al. 2015), and medium-pressure granulites in the rest of the craton (with the exception of amphibolite facies assemblages preserved in a few locations). Kusky et al., (2007a, b) related this to a continent-continent collision, with the outboard continent representing the Columbia (Nuna) Continent. The scale of this event is immense, but of the same magnitude as the current postcollisional zone in Central Asia that resulted from the India -Asia collision (Fig. 25a) so arguments that "the metamorphism in Hengshan and Wutai could not be caused by a continentcontinent collision in the north, because it is 200-300 km away" (Trap et al., 2012) are open to discussion. Several hundred km north of the India-Asia suture today, we find ourselves in Tibet (Hodges, 2000), with a double-thickened crust, partial melting at mid-crustal levels (e.g, Chung SL et al., 2003), and intense deformation with lower crustal flow similar to that documented for the high-pressure Datong-Chengde Granulite Belt (Trap et al., 2011) causing divergent directions of thrusting at the surface (e.g., Royden et al., 1997;Clark and Royden, 2000;Hubbard and Shaw, 2009). For instance, the strike of the Longmenshan, which is being uplifted as a result of the India-Asia collision, is roughly parallel to the motion direction of India into Asia, and the thrusting is at right angles to the convergence direction (e.g. Burchfiel et al., 1995). Thus, having a 10 angle between the northern margin of the craton, and predicted perpendicular thrusting

ACCEPTED MANUSCRIPT
62 directions in some places such as Wutai or Zanhuang are not unexpected. The same argument applies to much of the deformation across Asia north of Tibet (e.g., Cunningham, 2015). Note that the Tibetan Plateau, underlain by granulites and zones of partial melting, extends for more than 500 km from the India-Asia suture zone in India (Fig. 25b). If the scale of the continent-continent collision that juxtaposed the NCC against the Columbia Continent was comparable to the India-Asia collision, then the entire NCC would have been involved, as is demonstrated by the metamorphic data. It is not confined to a narrow ~200 km wide belt called the TNCO. Note also that the time-scales of the deep metamorphism, melting, and deformation are similar. The India/Asia collision that is uplifting Tibet began more than 50 Ma ago (e.g., Rowley, 1996;Zhu et al., 2005;Ding et al., 2005), and is still on-going to this day (Harrison et al., 1992;Kirby et al., 2003). Thus, the roughly 100 Ma long group of metamorphic events from 1.88 Ga to 1.79 Ga can all be related to different stages of the ancient continent-continent collision initiated along the northern margin of the NCC.
The N-S collision between the NCC and the Columbia/Nuna Continent is supported by a study of Re-Os isotopes from 99 peridotite xenoliths from the Central NCC. Liu JG et al., (2011) reported that peridotites from the northern part of the craton are more fertile than those from the south, and that the peridotites in the north have Os model ages (T RD ) of ~ 1.8 Ga, suggesting that the lithospheric mantle in the north is significantly younger than the overlying Archean crust. In contrast, in the south, the (T RD ) ages are ~2.1-2.5 Ga, consistent with the collision of the Eastern Block and arcs in the COB in the late Archean-early Proterozoic. Moreover, the lithosphere in the north seems to have been replaced at circa 1.8 Ga, consistent with a "major north-south continent-continent collision that occurred during assembly of the Columbia Supercontinent at ~ 1.8-1.9 Ga" (Liu JG et al., 2011).

Fig. 25 near here. Tibet -NCC comparison
The 1.85 Ga event is recorded in the many metamorphic P-T-t paths so widely reported and interpreted to be a result of continent-continent collision in the TNCO. Unfortunately, nearly all these works (with the exception of Trap et al., 2009aTrap et al., ,b, 2012 and Faure et al. (2007) and Zhang et al. ( , 2009Zhang et al. ( , 2012 lack accompanying detailed field and structural studies, and led only to construction of P-T-t paths. We concur that the P-T-t paths are indicative of a continentcontinent collision, but think that collision took place along the northern margin of the Craton (e.g., Wan, B. et al., 2015), and not along the defunct TNCO.
The circa 1.9-1.85 Ga continent-continent collision is associated with a suite of leucogranites that are distributed across the craton, particularly in its northern half (Fig. 25d, Fig.   26). Many of these are leucocratic veins that are associated with crustal anataxis and cut most of the high-grade metamorphic rocks as in Fig Ma (Zhao et al., 2008). In North Korea, massive porphyritic post-tectonic monzogranites formed at 1865-1843 Ma . We relate these abundant post-tectonic anatectic granitoids to crustal thickening following collision (e.g., Dewey and Burke, 1973) of the NCC with the Columbia/Nuna Continent (Fig. 22d; Fig. 26), and suggest that they are analogous to the Himalayan and Tibetan leucogranites forming today in response to the India-Asia collision (e.g., Le Fort et al., 1987;Yin and Harrison, 2000;Galliard et al., 2004).

High-Pressure granulites, eclogites, and orogen-parallel lower crustal flow in the Hengshan and the northern margin of the NCC: an analog to modern day Tibet
The Hengshan Complex to the north of the Wutai Complex (Figs. 7, 17, 25, 26) contains a suite of circa 2550-2450 Ma tonalitic, trondhjemitic, and granodioritic (TTG) gneisses and granitic gneisses, metamorphosed to granulite facies in the Paleoproterozoic that include numerous boudins of high-pressure mafic granulites, retrogressed eclogites, and high-grade metasedimentary rocks (Li and Qian, 1991;Li, J.H. et al., 2000a,b,c;Wilde et al., 2002;Wilde and Zhao, 2005;Zhao et al., ,b,c, 1999Zhao et al., , 1999cZhao et al., , 2005Zhao et al., , 2006Kröner et al., 2005Kröner et al., a,b, 2006O'Brien et al., 2005). A major EW-striking ductile shear zone, the Zhujifang Shear Zone cuts the complex in two, with high-pressure granulites confined to north of the shear zone, and medium-pressure granulites and amphibolites only to the south of the shear zone (O'Brien et al., 2005;Kröner et al., 2006;Trap et al., 2011).The high-pressure belt in the north strikes ENE for about 400 km before it disappears under younger cover, and is about 150 km wide in an N-S direction (Figs. 7 and 17).  documented a complex deformation history of the Hengshan Complex.
Early D1 (maybe not the earliest event) structures include a compositional fabric with small isoclinal rootless intrafolial folds with axial planar transposing intrafolial foliations and mineral lineations. Zhao et al. (2001b) and  noted that this deformation event may be associated with the earliest metamorphic assemblage preserved in these rocks that includes quartz and rutile inclusions in garnet, and symplectic clinopyroxene-plagioclase intergrowths. There are no constraints on the age or PT conditions of this event, other than that it pre-dates D2.
Overprinting D2   . This stage of deformation is also associated with a series of low-angle detachment faults, leading  to suggest that this last phase of deformation was related to exhumation of the complex. Support for this interpretation comes from the near-isothermal decompressional symplectites and coronas that surround embayed garnet grains that formed during this late exhumation stage of the Hengshan.  presented evidence that the entire sequence of events from the high-grade metamorphism during D2 to the exhumation during D5 lasted from 1880 Ma to 1820 Ma. It is not known when the M1 metamorphism occurred, so we do not relate it to the same single tectonic event and PT-t path as , but rather, we just say it predates the circa 1880-1820 Ma event. Trap et al. (2011) presented a comprehensive structural analysis of the entire exposed part of the High Pressure Belt, focusing on the role of partial melting interacting with the deformation.
They divided the complex into units of diatexite and metatexite (sensu Sawyer, 2008). They show that the overall geometry of the HPB is that of a 400 km long (ENE) and 100 km wide antiform with a gentle north-dipping northern limb bound by the newly-defined Datong-Chengde shear zone, and a southern limb that dips steeply south, bound by the Zhujifang shear zone (Figs. 7,17,25,26). The antiformal hinge curves from E-W to NE-SW, and plunges W in the W, and E in the E. Based on their analysis Trap et al. (2011) defined four main stages of deformation related to the establishment of this geometry.
D1 is an amphibolite-facies gneissic foliation preserved in paleosomes with an associated L1 mineral lineation, which becomes mingled with S2 in metatexites. Outside the zone of partial melting (PMZ) S1 remains the dominant fabric as an E-W to NE-SW gneissic foliation with a NW/SE-trending sillimanite, biotite or amphibole mineral lineation. Top-to-the-SE shear parallel to L1 is indicated by sigma-type porphyroclasts (Trap et al., 2011). Kröner et al. (2005aKröner et al. ( , 2006 A C C E P T E D M A N U S C R I P T

ACCEPTED MANUSCRIPT
67 and Guo et al. (2005) obtained SHRIMP U-Pb ages of 1881 +/-8 Ma and 1872 +/-16 Ma in migmatites and high-pressure granulites, which Trap et al. (2007Trap et al. ( , 2011 interpreted as a phase of crustal thickening during the D1 event leading to the major partial melting and crustal anatectic event at circa 1850 Ma. The most significant structure within the PMZ is D2, which formed during amphibolite to granulite facies metamorphism accompanied by in situ partial melting (Trap et al., 2011).
Numerous conventional multigrain and SHRIMP U-Pb ages from migmatites and HP granulites date this event rather precisely at 1850 +/-10 Ma , 2005Faure et al., 2007;. The S2 foliation is defined by leucocratic material in the veins that is parallel to S1, forming a composite S1-S2 fabric. Trap   The consistent E-W to NE-SW orogenic flow (see Fig 7a) is remarkably similar to that of modern orogens, such as the Tethysides, where orogenic flow parallel to the orogenic strike at mid-to-lower crustal levels is widely thought to partly accommodate orogenic escape and collapse of orogenic highlands. The near-parallelism of the Zhujifang and Datong-Chengde Shear Zones, which are at the heart of the zone of partial melting within the northern margin of the craton is geometrically analogous to the flow to the east away from the extant India -Asia collision (e.g. Tapponier and Molnar, 1976;Molnar, 1988;Molnar and Tapponnier, 1975). Thus,

ACCEPTED MANUSCRIPT
68 we relate this exposed Paleoproterozoic mid-lower crustal flow at circa 1.85 Ga to the proposed 1.85 Ga continent-continent collision on the northern margin of the craton, a few hundred km to the north. Such orogen-parallel flow during continent-continent collision is common (England and McKenzie, 1982;Houseman and England, 1993;Royden et al., 1997;Klemperer, 2006;Clark and Royden, 2000) and matches the geometry of the NCC system, so we so do not call on oroclinal bending of a supposed NS-striking TNCO to explain this phenomena (c.f. Trap et al., 2011), but instead interpret the flow in terms of the present geometry of the orogen (Figs. 7, 17, 23, 26).
D3 was a subsolidus deformation developed at the granulite-amphibolite facies transition, but the E-W flow is still recognized by subsolidus structures such as pressure shadows and porphyroclast systems, indicating co-axial flow (Trap et al., 2011). The main Datong-Chengde Shear Zone formed during D3 as a km-scale ductile normal-sense shear zone. S3 is a mylonitic foliation that shows top-to-the-NW senses of movement, such that the DCSZ has a normalsinistral sense of movement (present coordinates).
D4 is represented by the formation of the late Zhujiafang strike-slip shear zones. The ZSZ has a sub-vertical to steep south-dipping mylonitic to ultramylonitic foliation that shows sinistral kinematics , although Kröner (2005b) reported dextral shear in places. Early vertical motion may be explained by low-strain zones that still preserve a steep lineation .
The P-T evolution of the HP rocks in the Hengshan has been the subject of considerable research and the near-isothermal decompressional part of the PT-t path is well established. The "peak" high-pressure granulite assemblages (M1 of Trap et al., 2011, M2 of Zhang et al. (2007 consist of garnet + clinopyroxene +/-quartz, and locally an eclogite facies assemblage of garnet + quartz + omphacite pseudomorphs (Zhao et al., 2001a, Zhang et al., 2006. This assemblage is

ACCEPTED MANUSCRIPT
69 succeeded by a medium-pressure assemblage of garnet + plagioclase + clinopyroxene + orthopyroxene +/-quartz, a low-pressure (M3) granulite facies assemblage of orthopyroxene + clinopyroxene _ plagioclase +/-quartz, then finally an amphibolite-facies assemblage of hornblende + plagioclase (Trap et al., 2011). M1 (M2 of  took place at conditions of 800-850 C -14-16 kbar, M2 under conditions of 800-825 -10 kbar, M3 at 800 C -7 kbar, and M4 at 650 C -5 kbar Zhao et al., 2000Zhao et al., , 2001aO'Brien et al., 2005). These data are consistent with an isothermal decompression path from HP granulite facies to LP granulite facies conditions, followed by slow cooling through the amphibolite facies. Big questions remain, however, about the early prograde path. Are the M1 qtz-rutile assemblages reported by  part of the same PT path, or related to an earlier tectonothermal event? Is the PT path truly clockwise as widely reported, or, as the data suggests, do we only have sufficient data to constrain the near-isothermal decompression path? Should all of the data from the early pre-granulite events be included in the same PT path, or, as the structural data suggest, might there be an earlier tectonothermal event that is much older than the circa 1.85 Ga granulite facies event, which the above data are recording?
There has been much discussion of the origin of the circa 1.85 Ga metamorphism that formed the granulitic-eclogitic assemblages (now retrogressed) in the mafic boudins of the Hengshan Complex, with most workers suggesting that these demonstrate that the Eastern and Western Blocks collided at circa 1.85 Ga and that the high-grade metamorphism is related to this hypothetical continent-continent collision (Zhao et al., ,b,c, 1999(Zhao et al., , 1999cKröner et al., 2005Kröner et al., a,b, 2006O'Brien et al., 2005). However, this is not necessarily the case. The eclogitic metamorphism simply tells the time at which the mafic rocks were at the appropriate PT conditions to transform into eclogite. The regional structural patterns suggest a different origin for the high-grade metamorphism. The eclogites and HP granulites are all located north of the

ACCEPTED MANUSCRIPT
70 EW-striking Zhujiafang ductile shear zone, and rocks to the south of this major tectonic structure are all medium-grade granulites. It is interesting, and no coincidence, that the EW-striking Zhujiafang shear zone and the Datong-Chengde shear zone are parallel to the EW-striking Northern Hebei Orogen on the northern margin of the Craton, and that the Zhujiafang shear zone marks a major crustal boundary, which placed high-pressure granulites over medium-pressure granulites at circa 1.85 Ga. That is one reason why Kusky and Li (2003), Kusky et al (2007a, b), Kusky and Santosh (2009), and Kusky (2011)  which we think did not, and does not, exist.

Post collisional extension.
Recent studies have demonstrated that the continentcontinent collision was likely terminated by 1750Ma. Yu J.H. et al. (1993, 1996, Ramo et al.  F. et al. (2015) synthesized data on the AMCG suite, and suggested that they intruded in a post-orogenic extensional setting during post-collisional collapse of the orogen that was followed by the formation of rift and graben structures and mafic dike swarms that propagated across the whole craton, such as the 1780 Ma Taihang dyke swarm that emanates from the Xiong'er plume centre of the >0.1 M km 2 Large Igneous Province (Peng et al., 2015b).Eventually this led to the break-out of the NCC from the Columbia/Nuna Continent, and the beginning of a stable phase of evolution of the NCC that would last until the Phanerozoic (Li, Q.L. et al., 2007;Kusky et al., 2007a;Jiang N. et al., 2011;Wang W. et al., 2013).

Assessment of Precambrian tectonic stylesin the North China Craton
Suture zones and orogens are defined by using a combination of structural, stratigraphic, geochronologic, metamorphic, paleontologic, paleomagnetic, and paleoclimatic data. In old highgrade Precambrian terranes such as the NCC some of these tools are not available, but sutures between different tectonic units must still be defined using a multi-disciplinary tectonic analysis.
Suturing of different small tectonic units or large cratons is not a simple process whereby different terranes just "amalgamate" or "dock" or instantaneously bang into each other and stop, but involves complex and protracted geological processes (structural, metamorphic, magmatic, geochemical, temporal, erosional, depositional and others). An integrated or holistic assessment

ACCEPTED MANUSCRIPT
72 of these processes in the NCC to search for evidence of the operation of the Wilson Cycle, and hence plate tectonics in the Precambrian is presented in this work.
From 2.5 Ga to 1.8 Ga, the NCC grew by outward accretion of island arcs, accretionary wedges, oceanic plateaus, and ribbon micro-continents, progressively from the Eastern Block, to younger orogens to the west, northwest, then north (Fig. 17). The Eastern Block has been proposed to have formed by amalgamation of 'microblocks" between 3.8 Ga and ~ 2.6 Ga, with a peak between 2.6 and 2.7 Ga (Zhai M.G. et al., 2005(Zhai M.G. et al., , 2010Zhai M.G. and Santosh, 2011 Ga (Diwu C.R. et al., 2010(Diwu C.R. et al., , 2013 suggesting that there may yet be other regions of very ancient crust to be discovered in the NCC. While recognizing these older different components of the  (2007) suggested that the style of tectonics on Earth changed at about 3.0 Ga, from a "plumedominated system" to a "plate-dominated system." While our observations about the apparent change in tectonic style between 3.0 and 2.5 Ga in the NCC are consistent with this hypothesis, the style of pre-3.0 Ga tectonics in the NCC has yet to be rigorously tested, and remains a matter of discussion between the authors (e.g., see Dewey, 2007;Kusky et al., 2013a, b).
The style of accretion in the NCC is similar to that of the Superior Province (e.g., Card, 1990;Percival et al., 2006Percival et al., , 2012, in which progressively younger arcs, accretionary prisms, and microcontinents were added to the outboard portions of a core microcontinent (in a general sense), to form a large craton at the end of the Archean. This in turn has led some to propose the existence of a large end-Archean supercontinent, Kenorland (e.g., Hoffman, 1991;Aspler and Chiarenzelli, 1998;Bleeker, 2003;). Without any paleontological or rigorous paleomagnetic data it is difficult to test such a hypothesis, but the global data from Archean cratons do suggest a major amalgamation event at the end of the Archean. In the case of the North China Craton, the evidence does suggest that the style of accretion changed from small arclike landmasses between 3.8 and 2-7-2.6 Ga, to accretion of larger arc terranes to an amalgamated continental landmass at the end of the Archean, followed by progressive addition of arcs and reworking of existing crust until the record terminated at circa 1.7 Ga. This style of tectonism is also consistent with modern orogens such as the Carpathians in which the slab hinge converges relative to the upper plate (e.g., Doglioni et al., 2007).

Orogenic styles in Archean vs. Phanerozoic orogens as inferred from map patterns
One of the common ideas about Archean orogens is that they have fundamentally different characteristics and map patterns from Phanerozoic orogens, but we have shown that this is not the case for the late Archean of the NCC. Some geologists and geodynamic modelers claim that Phanerozoic orogens exhibit linear patterns, whereas Archean orogens are characterized by basins and domes. They then use this statement to claim that the crustal and geothermal gradients in the Archean were higher, and that the Archean tectonic style was dominated by vertical, rather than horizontal, movements (e.g., Choukroune et al., 1995;Hamilton, 2003Hamilton, , 2007Rey et al., 2003;Van Kranendonk et al., 2004;Bedard, 2006;Cagnard et al., 2006;Rey and Houseman, 2006;Gapais et al., 2009;Bedard et al., 2013;Debaille et al., 2013;Lin et al., 2013). In spite of the fact it is possible to construct numerical geodynamic models in the laboratory to explain this scenario, the basic observations and interpretations of such a fundamental difference between Archean and Phanerozoic orogens are mostly invalid (e.g., Polat et al., 2014) (for reviews of differences of opinion about Early Archean tectonics, see Van Kranendonk et al. 2004;Dewey, 2007;Kusky et al., 2013). In the NCC it is possible to find both "basin and dome" map patterns

ACCEPTED MANUSCRIPT
75 (e.g., the Taipingzhai gneiss terrane in Fig. 8) and linear map patterns (e.g., the 1,300 km long COB in Fig. 17). Thus, we briefly compare other terranes of Archean and younger ages globally to see if these different map patterns reflect a secular change in tectonic pattern, or just different tectonic environments.  Likewise, some workers claim that linear tectonic belts that characterize Phanerozoic orogens are absent from the Archean record (e.g., Hamilton, 2007). This is also untrue, especially for the Neoarchean, as exemplified by examples from the NCC in this paper. Fig. 28-a-b compares the map pattern in the Paleozoic Appalachians of Newfoundland with that of the Archean Yilgarn Craton. Note again that the styles and scales of the linear tectonic belts as well as their constituent rock types are similar in these two orogens of contrasting age. Fig. 28 Percival et al., 2012;Sengör et al., 2014). In fact, comparison of the two figures shows that the older Superior craton, the largest surviving fragment of an Archean craton on the planet, exhibits greater linearity than the equivalent Paleozoic CAOB. There are differences, however, in that the Superior Craton shows a fairly regular outward growth from the oldest "core" of the craton in the north (Percival et al., 2012), whereas the CAOB shows progressive outward growth of accretionary orogens to the south (present coordinates) from the Siberian Craton, and to the north (present coordinates) from the North China Craton, with the two orogens separated by a giant shear zone (see review by . Thus, the notion that Archean belts are dominantly characterized by dome-and-basin shaped outcrop patterns, and that Phanerozoic orogens are all characterized by linear outcrop patterns is a myth and should be dismissed. There is as much variation in Archean terranes, especially in the Neoarchean, as there is in young orogens, and similar map patterns can be found in both in different environments.
Thus, this notion cannot be used to suggest that Archean tectonic styles were different from those in the Phanerozoic, and should not be used as input for numerical models.
In summary, we emphasize this fundamental point about orogenic style, because failure to appreciate the importance of the correct interpretation of map patterns can lead to erroneous interpretations of the geology, geochemistry, geochronology, metamorphic patterns, and eventually to wrong conclusions about the role of plate tectonics in the early Earth. For example, the long-misunderstood interpretation and theoretical modelling (sagduction and diapirism) of the dome-and-basin map pattern of East Pilbara (Fig. 27b) is resolved by field-based evidence that the so-called 12 km-thick intact volcanic pile is actually broken into 5 units by at least 4 thin, but major, thrusts along which the mafic-ultramafic lavas of each unit are capped by cherts and

ACCEPTED MANUSCRIPT
77 marked by shear fabrics, and this scenario is constrained and confirmed by multiple U-Pb zircon dates that increase upwards in the volcanic pile (see Fig.29

Conclusions
The North China Craton (NCC) consists of a number of discrete tectonic units that can be interpreted coherently using the paradigm of plate tectonics from at least 2.7 Ga into younger times, and from understanding the geological effects of those plate tectonic processes in the preserved geology. From that perspective, we reach the following conclusions: 1. At about 2.5 Ga the tectonic style in the NCC underwent an apparent change from accretion of microcontinents and arc-type archipelagos (characteristic during the interval 3.8-2.7/2.6 Ga; Burke et al., 1976), to accretion of long linear orogenic units (Fig. 30). Whether this reflects a true change in the length-scale of accreted tectonic elements, or dismemberment of a larger arc to 1.9 Ga (and followed by sutures of the Central Asian orogenic belt in the Paleozoic; Fig. 30).
This type of progressive accretion away from an older nucleus is similar to that of the Superior Province of North America (Fig. 25), reflecting the amalgamation of smaller tectonic units into larger continental landmasses at the end of the Archean (perhaps leading to the formation of the Kenorland Continent) and into the Paleoproterozoic with the formation of the Columbia (Nuna) Continent.
2. A circa 2.5 Ga suture zone can be traced for ~1,300 km from north to south through a series of exposed Archean massifs in the North China Craton (Fig. 30). The suture separates the late Archean Wutai/Fuping arc in the Central Orogenic Belt on the west from the Eastern Block in the east. The subduction zone dipped to the west under the arc, and several accretionary prism fragments with fore-arc ophiolites and ophiolitic mélanges were obducted over the Eastern Block during the collision (Fig. 18). The Eastern Block consists of a series of smaller "microblocks" that may represent a tectonic collage of microcontinents and a SW-Pacific arc-like archipelago that contains small relicts of ancient crust up to 3.8 Ga old, and underwent a major accretion and crustal growth event at 2.7-2.6 Ga. A thick passive margin of shelf sediments, which formed on the western edge of this Eastern Block by 2.5 Ga, was involved in and imbricated with the arc and fore-arc ophiolitic mélanges in the Central Orogenic Belt during its collision with the Eastern Block at 2.5 Ga.

Fig. 30 near here
and generated a suite of mafic dikes and granites, with associated regional metamorphism, from underplated mafic magmas. This period of eastward subduction ended about 70 Ma later, at 2.43 Ga, when the Western Block collided with the arc-modified margin of the composite Eastern Block, shutting off that subduction system (Fig. 20).
4. Soon after this second collision, the composite North China Craton underwent rifting, and a fragment drifted off its northern margin, leaving a failed rift arm striking through the center of the Craton. Passive margin sediments were deposited over the rift facies sediments, and were affected by the collision of an arc along the northern margin of the craton, which took place at circa 2.3 Ga (Fig. 30). This arc was built on older basement, and soon after this collision the northern margin of the craton was modified to become an Andean-style arc (possibly through reversal of subduction polarity), and the entire craton was affected by Andean-type tectonics from 2.3 Ga to 1.9 Ga (Fig. 26). Features related to this significant interval in the development of the NCC include suites of continental-margin arc magmatic rocks that form a swath several hundred km wide that strikes E-W across the craton. Along the northern margin there was UHT metamorphism related to a ridge subduction event, deposition of volcanic and volcaniclastic rocks in retro-arc foreland basins several hundred km from the active margin front, and deposition of thick clastic sediments in an apron adjacent to the arc. The scale of these tectonic units is the same as that in the present-day Andes.

ACCEPTED MANUSCRIPT
80 5. From about 1.88-1.79 Ga, the entire NCC underwent a high-grade granulite facies event with high-P granulites and eclogites from this event now exposed in the north, and medium-P granulites now exposed in the center and south parts of the craton. Crustal anataxis is locally associated with this metamorphic event and large-scale lower-crustal flow accommodated escape parallel to the orogen. This craton-wide event was not associated with the addition of any new juvenile crust, just the re-working of older crust, and we relate this to continent-continent collision along the northern margin of the craton (Figs. 26, 30   WS Western Shandong; WT -Wutai; XH -Xuanhua; ZH -Zanhuang; ZT -Zhongtiao.
(b) Simplified cross section across the Denfeng complex. We interpret this region to represent the transition from the arc to the accretionary prism. Eastern Block at this time , Deng et al., 2014 and is postulated to be the cause of the widespread CCW metamorphism of rocks in the Eastern Block at circa 2.5 Ga.    in C compiled from , and Gong et al. 2011 see Supplementary Data Table 2 for details). Note that most magmatism and deformation is concentrated within 600 km of the active Andean margin (see "600 km front" line), but can extend as far as 1000 km (1000 km "Andean front" line). as in the present day Andes.   Map (a) and cross-section (b) redrawn from Nelson et al., 1996. Metamorphic data compiled from Peng et al., 2014;Zhang et al., 2015;Wan et al., 2013 (Hamilton, 2003(Hamilton, , 2007Dewey, 2007;Van Kranendonk et al., 2004), but the rock types, structures, and scales are remarkably similar. We suggest that domal granitoids reflect more the specific tectonic environment rather than a global change in tectonic style. Panel a after Hildebrand (2013), panel b compiled from Geological Survey of Western Australia (1990) and Hamilton (2007).