Late-stage foreland growth of China’s largest orogens (Qinling, Tibet): Evidence from the Hannan-Micang crystalline massifs and the northern Sichuan Basin, central China

The Triassic–Jurassic Qilian–Qinling–Dabie–Sulu (short Qinling) orogen of central China involved subduction of cratonal crust to and exhumation from mantle depths (e.g., Hacker et al., 2004). Dates are now available for the onset of continental subduction (ca. 255 Ma; e.g., Cheng et al., 2011), the polyphase ultrahigh-pressure metamorphism (ca. 245 Ma and ca. 227 Ma; e.g., Hacker et al., 2006; D. Liu et al., 2006; X. Liu et al., 2012), and the exhumation of the ultrahigh-pressure rocks to the surface (≤190 Ma; Grimmer et al., 2003; Wang et al., 2009). However, dating of foreland propagation, and in particular of the emplacement of the thrust front of the Qinling orogen, has been hindered by the overprint of the foreland by deformation related to the exceptionally wide Pacific plate boundary (e.g., Ratschbacher et al., 2003; Li and Li, 2007). The Sichuan Basin, floored by relatively rigid cratonic basement and characterized by fairly high Mesozoic subsidence rates (e.g., Guo et al., 1996; Meng et al., 2005) escaped strong shortening. Its northern part, however, exposes in its bounding ranges and marginal sedimentary rocks the leading edge of the Qinling fold-and-thrust belt and thus allows examination of the terminal stages of the Qinling orogeny.

The Cenozoic Tibetan Plateau is Earth’s most spectacular evidence of upper-plate thickening in the aftermath of a continent-continent collision (e.g., Argand, 1924). In the north, the plateau has overstepped the Tarim Basin, causing the formation of the Tien Shan (e.g., Avouac et al., 1993; Zubovich et al., 2010), and it is currently growing northeastward into the Hexi Corridor along the eastern end of the Altyn Tagh fault (e.g., Zheng et al., 2006; Palumbo et al., 2009). Along its eastern margin, the plateau terminates abruptly against the Sichuan Basin, forming spectacular topographic relief; there, the plateau appears to be growing mostly vertically (e.g., Kirby et al., 2008; Z. Shen et al., 2009). It has been suggested that lateral plateau spreading diverges around the Sichuan Basin (Clark and Royden, 2000), expanding southeastward toward southeastern Asia (e.g., Ouimet et al., 2010) and northeastward into the Qinling orogen (e.g., Enkelmann et al., 2006; Clark et al., 2010).

The northern Sichuan Basin and the Hannan-Micang crystalline massifs occupy the junction between the WNW-trending Qin Mountains (Qinling), part of the Qinling orogen, and the NE-trending Longmen Shan, the northeastern margin of the Tibetan Plateau (Fig. 1; e.g., Burchfiel et al., 1995). Thus, the Proterozoic crystalline basement core of the Hannan-Micang crystalline massifs and the sedimentary rocks of the northern Sichuan Basin provide an excellent natural laboratory for studies of the foreland propagation of both the Qinling and the Tibetan Plateau orogens. The principal questions this paper addresses are: (1) When was the frontal part (the leading edge) of the Qinling fold-and-thrust belt emplaced? The Qinling orogen terminates in the fold belt of the northern Sichuan Basin with a blind thrust that dips N and NE underneath the Hannan-Micang crystalline massifs and the Daba Shan (Fig. 2; e.g., Shi et al., 2012); we date deformation at the deformation front by estimating cooling- and exhumation-rate changes in the thrust hanging wall (the Hannan-Micang crystalline massifs) and the fold belt (the northern Sichuan Basin). (2) When were the Hannan-Micang crystalline massifs and northern Sichuan Basin incorporated into the eastward growth of the Tibetan Plateau? We examine the late-stage cooling history in the region east of the Longmen Shan and study whether the plateau is overstepping the Sichuan Basin.

Our analysis of the late-stage growth of the Qinling and Tibetan Plateau orogens integrates new and published zircon and apatite fission-track and (U-Th)/He thermochronology applied to the Hannan-Micang crystalline massifs and Sichuan Basin rocks and the depositional history of the northern Sichuan Basin. We argue that enhanced cooling and exhumation from the late Middle Jurassic to the early Late Cretaceous signify propagation of the Qinling orogen into its leading foreland, and that accelerated cooling starting in the late Miocene marks the incorporation of the Hannan-Micang crystalline massifs and the northern Sichuan Basin into the eastward-growing Tibetan Plateau.

The Longmen Shan, Qinling, Daba Shan, and northern Sichuan Basin bound the Hannan-Micang crystalline massifs in the west, north, east, and south, respectively (Figs. 1 and 2A). The Longmen Shan separates the eastern Tibetan Plateau from the Sichuan Basin; it formed during the Late Triassic, Cretaceous, and Cenozoic (Burchfiel et al., 1995; Liu et al., 1995; Arne et al., 1997; Kirby et al., 2000, 2002; Li et al., 2012). The Qinling was assembled by collision of the North and South China cratons and intervening accretionary wedges and magmatic arcs; collision occurred in the Triassic–Jurassic, and reactivation by wrenching and rifting ensued in the Cretaceous and Cenozoic (Meng and Zhang, 1999; G.W. Zhang et al., 2001; Ratschbacher et al., 2003, 2006; Dong et al., 2011a). The southwesterly convex Daba Shan orocline is part of the foreland fold-and-thrust belt of the Qinling orogen; it formed in the Jurassic–Early Cretaceous (Shi et al., 2012; Hu et al., 2012). The Sichuan Basin comprises a nearly complete Paleozoic–Cenozoic stratigraphic sequence that rests on Proterozoic Yangtze (South China) craton basement that is comparable to the crystalline rocks of the Hannan-Micang crystalline massifs (Guo et al., 1996; Richardson et al., 2008). The Hannan-Micang crystalline massifs consist of the northern Hannan massif, central Huijunba syncline, and southern Micang massif (Fig. 2B); they expose Proterozoic metavolcanosedimentary strata, voluminous Neoproterozoic granitoids, and, above an unconformity, Sinian to Silurian clastic and carbonate cover rocks (e.g., G.W. Zhang et al., 2001; Dong et al., 2011a). In the northern Sichuan Basin, Upper Triassic molasse unconformably covers Permian to Middle Triassic passive-margin limestones. Throughout the Jurassic and Early Cretaceous, the basin continued to serve as a clastic foreland depocenter to the Qinling orogen. The Huijunba syncline was part of the northern Sichuan Basin from the Late Neoproterozoic to the Early–Middle Jurassic; its sequence contains the Late Triassic unconformity but lacks Upper Jurassic and Cretaceous strata (Fig. 2; S.B.G.M.R, 1991; Guo et al., 1996; Meng et al., 2005).

The major map-scale Cenozoic faults of the Hannan-Micang crystalline massifs are the dextral transpressive Qingchuan fault (F1), and the Mujiaba (F2) and Guanba (F3) faults, which have unknown kinematics (Fig. 2A). The late Cenozoic Hanzhong and Xixiang basins (Fig. 2A) appear to have been controlled by these faults; together with the established late Cenozoic stress field of the West Qinling (compilation in Enkelmann et al., 2006), these basins allow speculations on the nature of the Mujiaba and Guanba faults. Accordingly, the Hanzhong basin may have formed along a releasing bend of the Qingchuan fault, and the Xixiang basin may have formed along the eastern transtensional horse tail of the Guanba fault (Fig. 2A). The kinematics of the Cenozoic faults within the Hannan-Micang crystalline massifs, the Mujiaba and Guanba faults, would therefore also have been dextral transpressive.

Little has been published about the pre-Cenozoic structures in the Hannan-Micang crystalline massifs and the northern Sichuan Basin (Du et al., 1998). Likely, the Cenozoic faults reactivate older structures. All faults of the Hannan-Micang crystalline massifs and the eastern abutting Daba Shan, however, root in a basal detachment that underlies the Hannan-Micang crystalline massifs and the entire Daba Shan fold-and-thrust belt and connects with the Qinling suture >100 km to the north (e.g., Shi et al., 2012, and references therein).

Previous thermochronologic studies have covered the central and south Longmen Shan (Arne et al., 1997; Kirby et al., 2002; Clark et al., 2005; Godard et al., 2009; Ouimet et al., 2010; Li et al., 2012), the western Qinling (Enkelmann et al., 2006; Zheng et al., 2006; J.H. Liu et al., 2012), and the northeastern Tibetan Plateau and adjacent Sichuan Basin (Richardson et al., 2008; Li et al., 2012); the northern Longmen Shan is less studied (Li et al., 2012; Fig. 1 for data compilation). These studies have defined the age and mechanism of Tibetan Plateau growth. U/Th-Pb zircon and ⁴⁰Ar/³⁹Ar and Rb/Sr biotite dates for the Hannan-Micang crystalline massifs have documented rapid Late Proterozoic postintrusive cooling (e.g., Z.Q. Zhang et al., 2001). Three zircon (U-Th)/He ages indicate involvement of the Hannan massif into Qinling orogen foreland deformation since the Late Jurassic (Xu et al., 2010), and apatite fission-track and (U-Th)/He ages trace terminal southward propagation of the Qinling orogen by rapid pre–Late Cretaceous cooling (>100–90 Ma) in the northern Sichuan Basin (data compilation in Fig. 2A; Chang et al., 2010; Xu et al., 2010; Tian et al., 2012). The growth of the Tibetan Plateau has involved the western Qinling since 9–4 Ma (Enkelmann et al., 2006) and the Hannan-Micang crystalline massifs since ca. 16 Ma (Tian et al., 2012; Fig. 2A).

We sampled granite, diorite, gabbro, and Upper Triassic to Lower Cretaceous sandstones from the Hannan-Micang crystalline massifs and the northern Sichuan Basin (Fig. 2A). We selected 11 samples for apatite fission-track (AFT), six for zircon fission-track (ZFT), and two samples for apatite (U-Th)/He (AHe) dating. Appendix A details the methodology used at the Freiberg track laboratory and the Tübingen (U-Th)/He laboratory. Tables 1–3 locate the samples and present the AFT, ZFT, and AHe data; Figure 2A shows the age distribution.

The bedrock ZFT ages cover 481 ± 34 Ma to 246 ± 19 Ma with no apparent correlation with geographic and geologic position. All the ZFT samples passed the χ² test, and thus all crystals are concordant within the statistical uncertainty. The Hannan (samples HN17, HN19, HN03) and Micang massif (MC01, MC02, MC25) ages span 364 ± 26–246 ± 19 Ma and 481 ± 34–270 ± 25 Ma, respectively. Field-based studies suggest a ZFT partial annealing zone (ZPAZ) at 240 ± 30 °C (Zaun and Wagner, 1985; Hurford, 1986; Brandon et al., 1998; Bernet, 2009), although the variation of ZFT annealing with radiation damage (Rahn et al., 2004) may cause considerable shifts. The U content is commonly used to assess the radiation damage in zircon; in general, it accounts for 80%–90% of the radiation damage (e.g., Garver and Kamp 2002; Garver et al., 2005). The U content of our zircons ranges from 129 to 68 ppm, representing low to average values (e.g., Garver and Kamp, 2002); this range suggests that there is little difference in radiation damage between our samples. Therefore, we interpret the broad ZFT age range to reflect slow cooling of the Hannan-Micang crystalline massif samples through the ZPAZ.

Our new AFT ages span 114.4 ± 7.4–60.3 ± 2.3 Ma. All pooled ages are younger than the intrusion or deposition ages of the sampled rocks. All bedrock samples except MC12 (P [χ²] = 3.1%) passed the χ² test (P [χ²] > 5%), but three Jurassic and Lower Cretaceous sedimentary rock samples failed the test. A plot of the single AFT ages against the host-rock stratigraphic age indicates that some apatites of these three samples were not fully reset (Fig. 3). C.B. Shen et al.’s (2007) vitrinite data show that paleotemperatures decreased from ∼250 °C to 130 °C from the Upper Triassic to Middle Jurassic strata (Fig. 3). Up-section projection of this temperature gradient places the bottom of the AFT partial annealing zone (APAZ; 110 ± 10 °C) in the Upper Jurassic–Lower Cretaceous strata; this allows for partial preservation of detrital grain ages in these strata. As only a few grain ages overlap the depositional ages, we used all single-grain ages to calculate the pooled ages (also for those three sedimentary rocks samples) (Fig. 3; Table 1).

Mean AFT confined track lengths vary between 11.5 ± 0.1 μm and 13.8 ± 0.1 μm; the corresponding c-axis projected mean lengths are 13.2 ± 0.1–14.9 ± 0.1 μm (Table 1). The track-length distributions show negative skewness. Mean etch pit size parallel to the c-axis, Dpar (Donelick, 1993), ranges from 1.9 ± 0.2 μm to 2.4 ± 0.2 μm; sample HN11 has a 2.9 ± 0.3 μm value and also has the longest mean track length of 13.81 ± 0.1 μm (14.9 ± 0.1 μm after c-axis projection). Most Dpar values are close to those of Durango apatite (1.83 μm; Ketcham et al., 1999), and the small variation implies little chemical variability and similar annealing kinematics among our samples.

Figure 4 displays track-length distributions and presents temperature-time (T-t) path models for 10 samples with 120–261 confined tracks (see Appendix A for our methodology to increase the number of etchable confined tracks and the modeling parameters). Two groups are apparent: Group 1 samples (HN03, HN12, HN11, MC01, MC02, MC25) entered the APAZ between ca. 150 and 110 Ma, cooled through the PAZ before ca. 90 Ma, and stayed for a prolonged time at low temperatures (≤60 °C). Group 2, with a broader track-length distribution, includes three samples from the northern Sichuan Basin strata (MC03, MC05, MC11) and one sample (HN17) from the Hannan-Micang crystalline massifs. Group 2 samples entered the APAZ between ca. 110 and 80 Ma, resided for a prolonged time within the upper part of the APAZ (90–60 °C), and experienced accelerated cooling since 20–5 Ma. We re-modeled four of our previously published AFT data (Enkelmann et al., 2006), located north of the Hanzhong basin (Q55, Q65, Q67, Q69; Fig. 2A; Appendix B, Fig. B1). Enkelmann et al. (2006) modeled these samples with Laslett et al.’s (1987) annealing equation and without c-axis projection. These samples belong to our group 2 data. They entered the APAZ before 100 Ma, stayed until 15–5 Ma in its upper part, and cooled to surface temperatures thereafter (Fig. B1). In the last diagram of Figure 4 (bottom center), we overlapped all good-fit (GOF > 0.5) T-t path envelops of group 2 samples (including those re-modeled from Enkelmann et al., 2006; Fig. B1); these envelopes likely confine the extreme cooling-history range of each sample. Assuming that all group 2 models stem from rocks that experienced the same cooling history, but that the distinct cooling models were influenced by various parameters (e.g., natural ones, such as variable rock type, or laboratory induced ones, such as counting efficiency), causing their variability, the overlap defined by these envelopes yields an “average” of all calculated T-t histories derived in this study. Taking conservative brackets on this overlap (the “average”), we suggest that rapid cooling started at 13–8 Ma.

In total, seven apatite grains from two samples (MC01 and HN17) were used for (U-Th)/He (AHe) dating (Table 3). The mean AHe ages are younger than the AFT ages of the same samples and consistent with the AFT-derived T-t paths (Fig. 4). However, the single-grain ages vary considerably; we speculate that among the many reasons that may contribute to such an age variation, e.g., undetected U- and Th-rich inclusions, disparate crystal size, zonation, slow cooling (e.g., Fitzgerald et al., 2006; Shuster et al., 2006), the latter, i.e., the long residence of our samples in the apatite He partial retention zone (APRZ), is the main reason for the variation in the singe-crystal ages. This slow cooling through the APRZ is also indicated by AHe and AFT age-elevation trends (see following) and the T-t path models (Fig. 4).

Here, we integrate our data with previously published data and derive several independent parameters pertinent to the cooling and exhumation history as a foundation for a regional interpretation of the Qinling and Tibetan Plateau forelands. Our and Tian et al.’s (2012) AFT ages correlate with sample elevation in the Micang massif and the northern Sichuan Basin (Fig. 5A): Older ages are associated with higher elevations. The calculated best-fit linear trend to the AFT age-elevation data in the Micang massif exhibits a break-in-slope at ∼900 m, with exhumation rates declining from ∼0.04 mm/yr prior to ca. 95 Ma to <0.02 mm/yr thereafter (Fig. 5A); both trends are poorly defined, and we will discuss the significance of the break-in-slope later herein. We interpret the higher-elevation trend as recording an exhumation event that brought the rocks from >110 °C through a pre-Cretaceous APAZ; the lower elevation age-elevation trend may trace the slow passing of the samples through this APAZ, with ages representing a mixture between the higher-elevation event and a younger one. In the Hannan massif, our and Xu et al.’s (2010) AFT ages range from 70.1 ± 4.1 Ma to 133.1 ± 6.0 Ma, with most ages older than 90 Ma. These ages may also follow a weak age-elevation trend (Fig. 5B). However, the younger dates, HN17 (70.1 ± 4.1 Ma) and D7613 (77 ± 6 Ma), and also MC12 (68.9 ± 3.5 Ma), by far the youngest high-elevation sample from the Micang massif, are all located along the Guanba fault (Fig. 2A); these three samples will be discussed separately in the following.

Plots of our and Tian et al.’s (2012) AFT mean track lengths against elevation reveal a similar pattern as the age-elevation data for the Micang massif and northern Sichuan Basin (compare Figs. 5A and 6A): Higher-elevation samples, with longer tracks and less mean track-length variation (12.7 ± 0.2–13.2 ± 0.1 μm), record more rapid exhumation. The lower-elevation samples, with shorter tracks and larger variation in mean track length (11.5 ± 0.1–12.5 ± 0.2 μm), probably represent a pre-Cretaceous upper APAZ. The Hannan massif track length–elevation trend is as inconclusive as its age-elevation relationship (compare Figs. 5B and 6B): Mean track lengths range from 11.5 ± 0.1 μm to 13.8 ± 0.1 μm. However, the Hannan massif track lengths are all longer than 14 μm after c-axis projection (Table 1), except HN17, which has the shortest value of 13.1 ± 0.1 μm. Apatite from HN11, with the longest mean track length of 13.8 ± 0.1 μm (14.9 ± 0.1 μm after c-axis projection) and the largest Dpar (2.9 ± 0.3 μm), appears more resistant to annealing.

Plotting the AFT ages against mean track lengths for all Hannan-Micang crystalline massif samples reveals two groups (Fig. 6C): Group 1 apatites are older (123.5 ± 6.0–90.3 ± 6.8 Ma), have longer mean track lengths (>12.5 ± 0.1 μm; except sample HN19, with only 34 confined tracks, and one sample from Tian et al. [2012], which has a mean track length of 11.4 ± 0.3 μm), and have a narrower track-length distribution than group 2 (see following; Fig. 6D). All group 1 apatites fall into the “undisturbed basement” field (Fig. 6D), which likely represents monotonic cooling from temperatures where tracks fade to ambient temperatures (Gleadow et al., 1986); thus, group 1 apatites may stem from above a fossil APAZ. Group 2 apatites are younger (60.3 ± 2.3–82.9 ± 7.8 Ma; mostly younger than 70 Ma), have shorter mean track lengths (11.5 ± 0.1–12.5 ± 0.2 μm; except for one sample from Tian et al. [2012] with a mean of 12.7 ± 0.2 μm), and have a wider track-length distribution than group 1 (Fig. 6C). Group 2 apatites occupy the “mixed” field and the transition zone between “mixed” and “undisturbed basement” (Gleadow et al., 1986; Fig. 6D). Thus, group 2 samples, mainly from the Micang massif and the northern Sichuan Basin, define a pre-Cretaceous APAZ based on both their ages and track lengths. In Figure 6C, longer lengths correlate with narrower distributions for older ages, mean track lengths vary less in samples older than 90 Ma, and mean track lengths decrease with decreasing ages in samples younger than 90 Ma; this pattern defines a “half” boomerang, and thus only the upper part of a fossil APAZ is currently exposed in the Hannan-Micang crystalline massifs.

Plotting our (sample MC01), Chang et al.’s (2010), and Tian et al.’s (2012) Micang massif and northern Sichuan Basin AHe ages against sample elevation defines a linear trend that suggests slow exhumation at ∼0.02 mm/yr between 92.2 ± 1.8 Ma and 31.9 ± 9.9 Ma; this rate is similar to, albeit slightly higher than, the weakly defined one calculated from the younger than 95 Ma part of the AFT age-elevation data (Fig. 5A). Chang et al.’s (2010) Hannan massif AHe age-elevation data suggest a poorly defined exhumation rate of ∼0.38 mm/yr (correlation coefficient R² = 0.04); sample elevations vary from 1508 m to 613 m within a narrow age spread (123.5 ± 2.7–107.4 ± 3. 8 Ma; Fig. 5B). This rate would be ∼10× higher than the exhumation rate calculated from the AFT age-elevation trend of the Micang massif and northern Sichuan Basin here, covering a comparable period (123.5 ± 6–93.9 ± 4.1 Ma); enhanced exhumation would, however, be consistent with coeval enhanced cooling, as derived from the Hannan massif’s AFT T-t paths (Figs. 4 and 6B). Two of our AFT ages (HN11, HN19) are younger than Chang et al.’s (2010) AHe ages at the same elevation, further casting doubt on the significance of the AHe age-elevation relationship. The 43.4 ± 16.8 Ma AHe age of sample HN17 is much younger than other Hannan massif AHe ages; this sample also has the youngest AFT age (Fig. 5B; Table 1), and it is located along the Guanba fault (Fig. 2A; see discussion).

The thermochronologic data analysis suggests more rapid exhumation and cooling during the Early Cretaceous than during the Late Cretaceous–Tertiary. The onset of this accelerated exhumation/cooling period, however, cannot be quantified by AFT and AHe data due to the inability of these thermochronometers to provide age information above temperatures of ∼110–80 °C. Unfortunately, our ZFT ages (ca. 481–246 Ma) and Xu et al.’s (2010) zircon (U-Th)/He ages (ca. 184–153 Ma) vary strongly and are possibly part of fossil partial annealing/retention zones. Here, we approximate the age of onset of deformation around the Hannan-Micang crystalline massifs, which we relate to the onset of enhanced exhumation/cooling, from a correlation of newly measured Jurassic stratigraphic strata in the Huijunba syncline (northern section A–A′; for previous data, see S.B.G.M.R., 1989; Guo et al., 1996) with data from the northern Sichuan Basin (southern sections B–B′ and C–C′; Fig. 2A; see also Meng et al., 2005; Burchfiel et al., 1995). We assume that disparities in the strata reflect north to south propagation of Qinling foreland deformation.

Figure 7 and Appendix C detail the evolution of the Jurassic–Cretaceous strata; the Huijunba syncline and the northern Sichuan Basin sections are identical, at least up to the Middle Jurassic (165 ± 5 Ma; ICS, 2010) basal section of the Shaximiao Formation. No younger Jurassic strata are preserved in Huijunba syncline north of the Micang massif. The absence of strata younger than Middle Jurassic in the Huijunba syncline may be due to later (for example, Cenozoic) erosion of potential Upper Jurassic–Cretaceous deposits, but it may also be due to a switch from deposition to erosion at ca. 165 Ma. The identical stratigraphy in the Huijunba syncline and the northern Sichuan Basin suggests that the onset of accelerated exhumation did not start earlier than ca. 165 Ma, and we speculate that deposition terminated in the Middle Jurassic in the Huijunba syncline due to migration of foreland deformation from the Hannan-Micang crystalline massifs to the northern Sichuan Basin, where the deposition of thick deposits and the onset of coarse clastic input suggest onset of deformation during the late Middle–Late Jurassic (Fig. 7).

Isopach maps of Mesozoic strata within the Sichuan Basin (Guo et al., 1996; Li et al., 2003; S.G. Liu et al., 2006) indicate migration of depocenters throughout the Late Triassic and Jurassic; this has been interpreted as reflecting foreland propagation of deformation in the Qinling orogeny. The Late Triassic (Norian–Rhaetian) depocenter was in front of the Longmen Shan, with strata thickness ≥3 km (Li et al., 2003) and an average sediment accumulation rate of ∼0.19 km/m.y. Lower Jurassic sediment thickness is <400 m, which corresponds to an abrupt decrease of the sediment accumulation rate to ∼0.02 km/m.y.; sediment thickness varied little throughout the Sichuan Basin, which suggests low erosion rates in the adjacent drainages. In the late Middle Jurassic, a depocenter initiated in the northern Sichuan Basin, where the <2-km-thick (see Appendix C, section C–C′) Shaximiao Formation was deposited; the average accumulation rate was ∼0.2 km/m.y. Sediment accumulation rates >0.1 km/m.y. persisted throughout the Late Jurassic (<1830-m-thick strata, section C–C′) and the Early Cretaceous. From the Late Cretaceous onward, most of the northern Sichuan Basin was an area of erosion.

Published thermochronology (Z.Q. Zhang et al., 2001; Zhou et al., 2002; Ling et al., 2006; Zhao et al., 2006; Chang et al., 2010; Xu et al., 2010; Dong et al., 2011b; Tian et al., 2012) and our new geothermochronology define the Neoproterozoic to Holocene thermal evolution of the Hannan-Micang crystalline massif rocks (Fig. 8). This history consists of five stages: (1) Neoproterozoic postmagmatic cooling; (2) thermal steady state within a Paleozoic Yangtze block passive continental margin; (3) Middle Jurassic–Early Cretaceous cooling due to foreland propagation of the Qinling orogeny; (4) thermal steady state with Late Cretaceous formation and Tertiary persistence of a regional low-relief landscape; and (5) late Cenozoic cooling due to eastward growth of the Tibetan Plateau. We discuss stages 2 to 5 in the following.

Our ZFT ages (481 ± 34–246 ± 19 Ma) indicate that the Hannan-Micang crystalline massif rocks remained in the ZPAZ (240 ± 30 °C) for a prolonged period. Consistently, biotite ⁴⁰Ar/³⁹Ar (closure temperature ∼300 °C) and zircon (U-Th)/He ages (closure temperature ∼180 °C) are at 796 ± 20 Ma (Z.Q. Zhang et al., 2001) and 184.3 ± 13.4–152.6 ± 2.9 Ma (Xu et al., 2010), respectively, and sedimentary strata in the Hannan-Micang crystalline massifs and northern Sichuan Basin record a Paleozoic–Middle Triassic passive-margin evolution with Devonian and Carboniferous hiatuses (Guo et al., 1996). The Triassic North China–South China collision (Meng and Zhang, 1999; Ratschbacher et al., 2003, 2006) transferred the Hannan-Micang crystalline massif area from a passive margin into a southern foreland basin to the Qinling orogeny; clastic sediments prevailed from the Late Triassic onward. In the Huijunba syncline, foreland uplift is recorded by the Late Triassic hiatus.

Several parameters, such as AFT age, AFT track length–elevation relationships (Figs. 5A and 6A), and T-t paths (Fig. 4), indicate Early Cretaceous accelerated cooling and exhumation of the Hannan-Micang crystalline massif rocks. The Micang massif and northern Sichuan Basin AFT age–elevation data constrain this event between ca. 125 and ca. 95 Ma, with an exhumation rate of ∼0.04 mm/yr; assuming a geothermal gradient of 30 ± 5 °C/km yields cooling rates of 1.2–1.6 °C/m.y. T-t paths show that group 1 samples pass through the APAZ in 20–30 m.y. before 120–90 Ma; the modeled cooling rates are 1.5–2.5 °C/m.y., and the exhumation rates are between 0.04 and 0.10 mm/yr, i.e., close to the results from the age-elevation data. The inconclusive Hannan massif AFT age-elevation relationship (Fig. 5B) is probably due to the relative low relief of the Hannan massif (Fig. 1), prohibiting sampling outside the APAZ.

Though both T-t paths and AFT age-elevation data outline accelerated cooling and exhumation during the Early Cretaceous, they do not reveal its onset. Our ZFT ages suggest that the Hannan-Micang crystalline massifs stayed in a thermally stable passive continental margin setting until the Middle Triassic (Fig. 8). After the Late Triassic hiatus in the Huijunba syncline and deposition of coarse clastics in the northern Sichuan Basin, which likely record the North China–South China collision (see previous), the stratigraphic sequence is identical up to the Early–Middle Jurassic on both sides of the Micang massif, i.e., in the Huijunba syncline and the northern Sichuan Basin (Fig. 7); thereafter, Late Jurassic–Early Cretaceous deposition likely occurred only south of the Micang massif. The regional variation in sediment thickness and the depocenter migration in the Sichuan Basin indicate that the Hannan-Micang crystalline massifs were a source of detritus to the Sichuan Basin and that orogenic propagation induced basin subsidence in the northern Sichuan Basin beginning in the late Middle Jurassic. Thus, the sedimentologic evidence suggests that the Qinling foreland deformation migrated into the Hannan-Micang crystalline massif area and the northern Sichuan Basin during the late Middle Jurassic (−165 Ma); in the light of these inferences, Xu et al.’s (2010) Hannan massif 184–153 Ma zircon (U-Th)/He dates can be interpreted to provide the thermochronologic record for the incorporation of the Hannan-Micang crystalline massifs into the foreland growth of the Qinling orogen.

The AFT age-elevation data (Fig. 5A) indicate initiation of a prolonged phase of slow exhumation at ca. 95 Ma (<0.02 mm/yr; cooling rates of <0.5–0.7 °C/m.y.). Consistently, the AHe age-elevation data (Fig. 5A) document slow exhumation of ∼0.02 mm/yr (cooling rate of 0.5–0.7 °C/m.y.) from at least 92.2 ± 1.8 Ma to 31.9 ± 9.9 Ma. Furthermore, group 1 T-t paths (Fig. 4) indicate passage of most Hannan-Micang crystalline massif rocks through the APAZ before ca. 90 Ma, subsequent to which, they stayed at temperatures <60 °C; the average cooling rate was ≤0.44 °C/m.y. Group 2 T-t paths (Fig. 6) suggest prolonged residence of other Hannan-Micang crystalline massif rocks within the upper part of the APAZ after ca. 90 Ma, followed by enhanced cooling since 13–8 Ma. The cooling rate between ca. 90 and ca. 10 Ma was very slow, ∼0.38 °C/m.y. We infer that the waning phase of the enhanced cooling and exhumation stage (ca. 95 Ma; see previous) was synchronous with topographic decay and the establishment of a regional low-relief landscape (or peneplain, which we regard as a featureless land surface of considerable area; e.g., Davis, 1899; Hetzel et al., 2011). Previous thermochronologic and geomorphologic research supports regional slow cooling and exhumation throughout the Late Cretaceous–Tertiary and the possible existence of a low-relief landscape in eastern Tibet and central China. For example, based on AFT T-t path models, Arne et al. (1997) recorded slow cooling from the Cretaceous to ca. 20 Ma for the Longmen Shan. Enkelmann et al. (2006) documented continuous slow cooling (∼1.2 °C/m.y.) from 100–70 Ma to 9–4 Ma across the western Qinling. Kirby et al. (2002) reported slow cooling (∼1 °C/m.y.) along the eastern margin of the Tibetan Plateau from the Jurassic until the late Miocene or early Pliocene, using ⁴⁰Ar/³⁹Ar K-feldspar and (U-Th)/He zircon and apatite thermochronology. Sedimentation in the Sichuan Basin also reflects this stage: Deposition withdrew into its southwestern corner in the Late Cretaceous–Eocene. Also other basins in central China show hiatuses in their sedimentation history; for example, the Ordos Basin of northern Central China lacks sedimentation in the Late Cretaceous–middle Miocene (S.B.G.M.R., 1989; C.Y. Liu et al., 2006).

Group 2 AFT T-t paths, based on the youngest ages, shortest confined track lengths, highest length deviations, and lowest elevation samples, suggest accelerated cooling from the upper APAZ (70–60 °C) to surface temperatures since 13–8 Ma (Fig. 4), with cooling rates of ≤5 °C/m.y.; assuming a geothermal gradient of 30 ± 5 °C/km and a surface temperature at 20 °C yields exhumation rates of 0.14–0.20 mm/yr, corresponding to 1.4–2.0 km of overburden removal. Our modeling results correspond to those of Tian et al. (2012) and the re-modeling of Enkelmann et al.’s (2006) AFT data north of the Hannan-Micang crystalline massifs. Enhanced exhumation since 13–8 Ma is consistent with sedimentation in the Hanzhong and Xixiang basins (Fig. 2A; S.B.G.M.R., 1989): Pliocene mammal fossils in the middle to upper section of the basin (Tang and Zong, 1987) suggest an earlier, probably Upper Miocene basal sequence. These basins are located along NE-striking faults (Fig. 2A) and probably formed at releasing sites within an overall dextrally transpressive late Cenozoic stress field (Burchfiel et al., 1995; Enkelmann et al., 2006; Fan et al., 2008; see previous sections). AFT samples MC12, HN17, and D7613 were sampled along the Guanba fault (Fig. 2A). Sample MC12, at an elevation of 978 m, has a 68.9 ± 3.5 Ma age, which is inconsistent with the Micang–northern Sichuan Basin age-elevation relationship (Fig. 5A). The T-t paths of sample HN17 indicate accelerated cooling from 70 to 20 °C in the past ∼10 m.y. (Fig. 4, group 2 samples). These samples may trace thermal advection in the near-field of the active Guanba fault. This possible thermal effect and the association of the Qingchuan and Guanba faults with the late Cenozoic Hanzhong and Xixiang basins also suggest that the Hannan-Micang crystalline massifs were incorporated into the deformation along the eastern Tibetan Plateau margin in the late Miocene.

Late Cenozoic enhanced cooling and exhumation have also been documented west and north of the Hannan-Micang crystalline massifs (see compilation in Fig. 9); consequently, these areas have been interpreted as being incorporated into the growth of the Tibetan Plateau. For example, Zheng et al. (2006) proposed ca. 8 Ma initiation of shortening along the Liupan Shan, forming the western margin to the Weihe graben (Figs. 1 and 9). To the north of the Hannan-Micang crystalline massifs, Enkelmann et al. (2006) documented accelerated exhumation since 9–4 Ma in the West Qinling. Two AFT age-elevation profiles across the Qinling range adjacent to the Weihe graben documented ca. 10 Ma onset of accelerated exhumation (Fig. 1; J.H. Liu et al., 2012). To the south and west of the Hannan-Micang crystalline massifs, Kirby et al. (2002) reported onset of rapid cooling at ca. 12–5 Ma in the Longmen Shan and at 5–3 Ma in the Min Shan (Fig. 1). The central Longmen Shan shows ∼0.65 mm/yr exhumation since ca. 11–8 Ma (Godard et al., 2009, 2010). Finally, AFT thermochronology traced regional exhumation since ca. 10 Ma in both the northern to southern Longmen Shan (Li et al. 2012). In this framework, we suggest that at 13–8 Ma, the Tibetan Plateau started to grow across the Sichuan Basin, progressively incorporating the northern Sichuan Basin and the Hannan-Micang crystalline massifs.

The available data on cooling and exhumation in the foreland of China’s two most prominent orogens, the Qinling and the Tibetan Plateau, constrain the mechanisms by which and the dates at which these orogens accommodated their final growth. The principal results from the Hannan–Micang massifs and the northern Sichuan Basin, where this late-stage orogenic growth can be studied in a combined approach, are as follows:

• (1) No or slow Paleozoic to Middle Triassic cooling in South China’s northern passive margin was terminated in the Middle–Late Triassic by the Qinling orogeny.

• (2) The leading foreland edge, represented by the Hannan-Micang crystalline massifs and northern Sichuan Basin, was incorporated into the Qinling orogeny in the late Middle Jurassic (ca. 165 Ma); tectonic activity persisted up to the Early Cretaceous (ca. 95 Ma). This period shows cooling at 1.2–2.5 °C/m.y. and exhumation at 0.04–0.10 mm/yr.

• (3) Negligible cooling and exhumation throughout the Late Cretaceous–Tertiary (ca. 95–10 Ma; 0.38–0.70 °C/m.y.; <0.2 mm/yr) indicate the formation of a regional low-relief landscape in the waning stage of foreland growth (see item 2 above) and its regional persistence across the eastern Tibetan Plateau and much of central China.

• (4) Overall, the Hannan-Micang crystalline massifs experienced <8 km of denudation since the Paleozoic and no more than 4–5 km of denudation from the early Mesozoic to the Cretaceous.

• (5) Enhanced cooling and exhumation, associated with fault activity since ca. 13–8 Ma and ≤2 km of denudation, mark the eastward growth of the Tibetan Plateau into the Hannan-Micang crystalline massifs and northern Sichuan Basin. Future work must identify whether areas east of the Sichuan Basin, e.g., the Daba Shan, are currently being involved in the overstepping of the Sichuan Basin by the growth of the Tibetan Plateau.

Zircons were mounted in Teflon, ground, and polished. The fossil tracks in the zircons were etched in a eutectic mixture of KOH and NaOH at 228 °C for 9–30 h. Zircon samples were covered with 50-μm-thick, uranium-free muscovite external detectors and packed between three mounts of uranium glass (IRMM-541), inserted at the top, middle, and bottom of the irradiation cans. The cans were irradiated in the hydraulic channel of the FRM-II reactor, Munich, Germany. The muscovite external detectors were etched in 40% HF for 30 min at room temperature. All ages were determined using the zeta approach (Hurford and Green, 1982, 1983), employing the IRMM-541 monitor; the zircon zeta values were calibrated by counting Fish Canyon tuff zircon age standards, which were irradiated together with samples (Table 2). We performed track counting on prismatic zircons with a Zeiss Axioplan microscope at ×1500 magnification in transmitted light; the corresponding muscovite external detectors were counted using an Autoscan (Autoscan Systems Pty. Ltd., Brighton, Australia) system.

We divided the apatites into two separates, one used for age dating (age group) and the other used for confined track-length measurements (length group). Apatites were mounted in epoxy, ground, and polished. The age-group apatites were etched for 15 s in 23% HNO₃ at 25 °C, and the muscovite external detectors were etched in 40% HF for 30 min at room temperature. Ages were determined using the zeta approach, employing the IRMM-540R uranium glass; the apatite zeta values stem from independent calibrations of three persons by counting several Durango and Fish Canyon tuff apatite age standards (Table 1). We performed the track counting on prismatic apatite surfaces with a Zeiss Axioplan microscope at ×625 magnification in transmitted light. The muscovite external detectors were repositioned, trackside down, on the apatite mounts in the same position as during irradiation; fossil tracks were counted by focusing on the underside of the external detector without moving the microscope stage (Jonckheere et al., 2003). Where possible, we counted at least 20 crystals of each sample.

The length-group apatites were etched for 20 s at 21 °C in 5.5 N HNO₃ (Donelick et al., 1999). The track-length measurements used ×1250 magnification on a Zeiss microscope equipped with the Autoscan system. All suitable confined tracks parallel to the prismatic surfaces were measured. Twelve samples were irradiated with heavy ions at GSI (Helmholtzzentrum fuer Schwerionenforschung GmbH) Darmstadt to increase the number of etchable confined tracks (Jonckheere et al., 2007). T-t paths for each sample were derived through inverse Monte Carlo modeling using the HeFTy software (version 1.6.7; Ketcham, 2005; Ketcham et al., 2009) and employing Ketcham et al.’s (1999) annealing model. We used a fixed l₀ value of 15.9 μm instead of an initial mean track length of 16.3 μm (Donelick et al., 1999), based on personal calibrations of the induced confined track length in Durango apatite. C-axis projection, using the method of Donelick et al. (1999) and Ketcham et al. (2007), was applied to account for the variation of track length with angle to the crystallographic c-axis. We used the monotonic-variable path setting to allow for both cooling and heating histories. The Kolmogorov-Smirnov test was employed to assess the fit between modeled and measured track-length distributions, with merit values of 0.5 and 0.05 for good and acceptable fits, respectively. For each sample, 100.000 paths were calculated; we obtained at least several hundred (except sample HN11, which had <100) to up to 1800 “good” T-t paths. We did not include the apatite (U-Th)/He ages as parameters in the models; instead, we plotted the (U-Th)/He ages in the T-t paths to test the validity of our models.

Clear and unbroken apatite grains without inclusions were selected using a binocular microscope. The grain dimensions were measured for the calculation of the alpha-correction factor (Farley et al., 1996). After that, we packed the single apatite grains in Nb-tubes for (U-Th)/He analysis. We analyzed 3–5 aliquots per sample. The samples were analyzed in the Patterson helium-extraction line at the University of Tübingen, which is equipped with a diode laser to extract the helium gas. Apatite grains were heated for 5 min at 11 Amps. Each grain was heated again and analyzed to make sure that the grain was degassed entirely in the first step. The residues generally were <1% of the first signal. After He analysis, the grain packages were sent to the University of Arizona at Tucson for inductively coupled plasma–mass spectrometry, U, Th, and Sm analysis. The analytical errors of the mass-spectrometer measurements were generally very low and did not exceed 2%; the reproducibility of the sample age reveals a much higher external error. We therefore report the mean (U-Th)/He age and the standard deviation of the measured aliquots as the sample error.

Re-modeling results for four of our previously published AFT data (Enkelmann et al., 2006) from north of the Hanzhong basin are presented in Figure B1. These samples were initially modeled with Laslett et al.’s (1987) fanning linear annealing equation and without c-axis correction. Here, we used Ketcham et al.’s (1999) annealing model, an l₀ value of 15.9 μm, and c-axis projection (Donelick et al., 1999; Ketcham et al., 2007; see Appendix A). The new parameters result in well-defined cooling paths over the past 20 m.y., and thus the re-modeling secured Enkelmann et al.’s (2006) inference of a phase of late, rapid cooling (our group 2 samples) after a prolonged period of track accumulation at temperatures of accelerated track annealing. However, it did not yield a more precise estimate of the onset of this rapid terminal cooling, previously derived from a linear relationship between the fission-track age and the number of tracks formed above and below 60 °C; Enkelmann et al. (2006) suggested ca. 9–4 Ma, and the re-modeling yielded 15–5 Ma for the onset.

In profiles B–B′ and C–C′, the Baitianba Formation conformably covers Upper Triassic rocks. It consists of basal conglomerate and fine-grained sandstone and mudstone interlayered with coal beds in its middle to upper portions; it is ∼380 m thick in B–B′ and ∼300 m thick in C–C′. Conglomerate occurs together with coarse-grained sandstone and consists predominantly of well-rounded quartzite pebbles of 2–5 cm in B–B′ and 5–10 cm in C–C′. The basal Baitianba Formation in A–A′ is ∼330 m thick and pseudoconformably overlies Middle Triassic rocks. It consists of basal conglomerate but otherwise is mainly composed of coarse-grained quartz sandstone; its middle and upper portions consist of grayish-green, fine-grained sandstone and mudstone with coal layers.

In the ∼330- and ∼523-m-thick sections B–B′ and C–C′, the Qianfoyan Formation consists of basal conglomerate (B–B′) and medium- and coarse-grained sandstone (C–C′). The upper section is composed of grayish-green, fine-grained sandstone and greenish siltstone and mudstone. Section A–A′ (∼738 m thick) shows similar composition but contains more siltstone and mudstone.

In sections B–B′ (∼2020 m) and C–C′ (∼1810 m), the Shaximiao Formation conformably covers the Qianfoyan Formation; it is the thickest part of the Middle Jurassic succession, and it is mainly composed of purple fine-grained sandstone, siltstone, and mudstone with calc-concretions. Section B–B′ has thick marls. In section A–A′, the upper portion of the Jurassic is absent; there, the ∼400-m-thick Shaximiao Formation consists of purple siltstones and mudstones that are interbedded with medium- and coarse-grained sandstone; the mudstones contain calc-concretions.

The Suining Formation (∼430 m in C–C′ and ∼500 m in B–B′) conformably covers the Shaximiao Formation; it contains reddish, coarse- and medium-grained sandstone interbedded with reddish siltstone and mudstone. Sandstone is mainly arkose and arkosic quartz arenite and has erosive basal contacts with mudstone.

Section B–B′ is incomplete. The Penglaizhen Formation is ∼1381 m thick in C–C′, consisting of basal conglomerate and coarse-grained sandstone, and sandstone, mudstone, and siltstone in the upper portion. Pebbles are mostly well-rounded quartzite. Gravelly sandstone and sandy conglomerate facies are common and often display alternation of planar-bedded conglomeratic and sandy layers. Fine-grained sediments are more pronounced upward in the section.

This work was supported by China Scholarship Council (CSC), Deutscher Akademischer Austausch Dienst (DAAD), Open Research Fund of Key Laboratory of Tectonics and Petroleum Resources (China University of Geosciences Wuhan), and Ministry of Education (TPR-2011-28,TPR-2012-25) grants to Yang, Deutsche Forschungsgemeinschaft (DFG) grants Ra442/25 and 27 and Robert-Bosch Foundation grants to Ratschbacher, an Austrian government Schrödinger Foundation scholarship to Wiesinger, and grants from the National Natural Science Foundation of China (grants 41190074, 41225008, and 40902038) to Dong and Shen. Richard Gloaguen is thanked for help with Figure 1. We thank Brian Horton and an anonymous colleague for very helpful reviews and Eric Kirby for clear directions for manuscript improvement.