The Longmen Shan, located at the boundary between the Tibetan Plateau and the Sichuan Basin, has received considerable attention following the 2008 Wenchuan earthquake. However, the tectonic history of the southwestern segment of the range has remained poorly constrained. We present zircon fission-track, zircon (U-Th)/He, and apatite (U-Th)/He data from the Baoxing region in the southwestern Longmen Shan that provide the first constraints on the cooling and exhumation history of the region. All of the measured ages are Cenozoic, and the data suggest that exhumation of the Baoxing region was ongoing by ca. 15 Ma. Zircon (U-Th)/He ages from several samples appear to be affected by radiation damage, suggesting that damage may be a concern even in samples with Cenozoic cooling ages. Samples were collected from two bodies of Precambrian crystalline rocks separated by the Wulong fault, and for all three thermochronometers, ages west of the Wulong fault are systematically younger than ages to the east, indicating that the fault has accommodated differential exhumation since 8–10 Ma. The regions east and west of the Wulong fault have experienced 7–13 km and at least 7–10 km of exhumation, respectively. The magnitude of exhumation in the southwestern Longmen Shan is similar to that reported in the central Longmen Shan, indicating consistency along strike. The thermochronology data also suggest that the Erwangmiao fault in the southwestern Longmen Shan is analogous to the Beichuan fault in the central Longmen Shan, and therefore may represent a source of seismic hazard.


The eastern margin of the Tibetan Plateau has experienced a long and complicated history of deformation. The tectonic evolution of this region, particularly its most recent phase of uplift, is poorly understood, but it is of particular interest for the evaluation of proposed uplift and deformation mechanisms of the Tibetan Plateau. Since the May 2008 Mw 7.9 Wenchuan earthquake in western Sichuan Province, considerable attention has been focused on the tectonics of the Longmen Shan region. While several recent studies have discussed the exhumation history of the Pengguan Massif and northeastern Longmen Shan (Godard et al., 2009; Kirby et al., 2002; Wang et al., 2012), the southwestern section of the Longmen Shan has received relatively little attention. We present a data set of apatite and zircon (U-Th)/He and zircon fission-track ages from the Baoxing region that provides the first comprehensive thermochronologic constraints on the Cenozoic tectonic history of the southwestern Longmen Shan.


The Longmen Shan is located on the eastern margin of the Tibetan Plateau and forms an abrupt boundary between the high topography of the plateau, at an average elevation of 3500–4500 m, and the 500–600 m elevations of the Sichuan Basin (Fig. 1). The sharp topographic front of the Longmen Shan is mirrored by an abrupt change in crustal structure, as crustal thickness changes from ∼60 km within the plateau to 40–45 km in the Sichuan Basin (Zhang et al., 2009; C.Y. Wang et al., 2010; Robert et al., 2010). To the northwest, the Longmen Shan is bounded by the Songpan-Ganzi terrane, a package of highly deformed Upper Triassic flysch, while to the southeast, there are largely Mesozoic terrestrial sediments of the Sichuan Basin. The range is made up of primarily Precambrian and Paleozoic rocks that have been faulted and folded along numerous northeast-southwest–trending structures. Precambrian crystalline rocks are exposed in two large anticlines—the Pengguan Massif to the northeast, and the Baoxing Massif to the southwest (Fig. 1).

Mesozoic Deformation

The Cenozoic structures of the Longmen Shan are overlain on a major Mesozoic fold-and-thrust belt. During the closure of the Paleotethys Ocean in Indosinian time, the region was the eastern boundary of southeast-directed folding and thrusting, resulting in intense deformation of the Songpan-Ganzi terrane and the emplacement of a series of nappes in the Longmen Shan thrust-nappe belt (Chen and Wilson, 1996). The nappes are completely exposed north of the Pengguan Massif and as a belt of discontinuous klippen east of the Pengguan and Baoxing Massifs. Mesozoic deformation was accommodated along a series of east-vergent northeast-southwest–trending structures and was accompanied by greenschist-grade metamorphism in the Longmen Shan (Burchfiel et al., 1995; Worley and Wilson, 1996). Shortening during the Mesozoic was accompanied by the development of a deep foreland basin and the deposition of Late Triassic to Cretaceous–Eocene sediment in what is now the Sichuan Basin (Burchfiel et al., 1995).

Deformation in the Songpan-Ganzi terrane to the west appears to have largely ended by the Middle Jurassic, based on dating of undeformed granitic plutons (Roger et al., 2004). In the foreland basin, postshortening deposition of Middle Jurassic sediments constrains the cessation of thrusting in the frontal Longmen Shan (Burchfiel et al., 1995, 2008). Although sedimentation in the Sichuan Basin continued through to the Eocene, deformation within the Longmen Shan between Middle Jurassic and Oligocene time has not been recognized.

Cenozoic Deformation

The northeast-southwest–oriented, northwest-dipping faults of the Mesozoic exerted significant control on the geometry of Cenozoic deformation in the Longmen Shan (Fig. 1). Within the range, Cenozoic shortening largely paralleled and/or reactivated preexisting Mesozoic features (Burchfiel et al., 1995, 2008; Chen and Wilson, 1996). Only to the east, in the Mesozoic foreland basin sediments, did deformation affect previously undeformed rocks. The interior of the Longmen Shan experienced large-scale folding during the Cenozoic, and formation of the anticlines that expose the Pengguan and Baoxing Massifs is attributed to Cenozoic shortening due to the relationship between these anticlines and folds affecting Eocene rocks to the south (Burchfiel et al., 1995, 2008). A series of Cenozoic thrust faults that postdate the folding described here accommodated shortening within the range, and detachments extended well into the Sichuan Basin, accommodating shortening on folds and blind thrusts, some of which involve Late Cretaceous, Eocene, and Pliocene–Quaternary rocks (Hubbard et al., 2010; Jia et al., 2006). The amount of Cenozoic deformation in the region is difficult to constrain due to the scarcity of rocks of Cenozoic age (Fig. 1), but existing estimates of Cenozoic shortening are on the order of 10–40 km (Burchfiel et al., 1995; Hubbard et al., 2010).

In the northeastern and central Longmen Shan, Cenozoic to recent structures are well documented, including the Pengguan/Guanxian fault, the Beichuan fault, and the Wenchuan-Maowen fault (Fig. 1) (Densmore et al., 2007; Xu et al., 2009). The Wenchuan-Maowen fault is a brittle fault superimposed on the Mesozoic Wenchuan-Maowen shear zone, and it has both thrust and right-lateral sense of motion (Xu et al., 2009). The amount of Cenozoic strike-slip deformation along this structure is unknown, but recent thermochronology data suggest a substantial amount of west-side-up displacement during the Cenozoic (Furlong and Kirby, 2013). The Beichuan and Pengguan/Guanxian faults ruptured in the 2008 Wenchuan earthquake. The Beichuan fault accommodates both thrusting and right-lateral slip, with the relative amount of strike-slip motion increasing to the northeast, while slip on the Pengguan/Guanxian fault is dominantly thrust sense (Xu et al., 2009). The Beichuan fault separates reset and unreset low-temperature thermochronometers and is considered to be the primary structure accommodating uplift of the Pengguan Massif (Godard et al., 2009; Arne et al., 1997).

Several studies have investigated the timing of uplift in the Longmen Shan using low-temperature thermochronology from the Pengguan Massif. Kirby et al. (2002) used feldspar 40Ar/39Ar and apatite and zircon (U-Th)/He data to infer that exhumation of the Pengguan Massif initiated ca. 11–12 Ma. Godard et al. (2009) suggested that exhumation began at ca. 8–11 Ma, based on a more extensive data set of zircon (U-Th)/He ages, and Wilson and Fowler (2011) suggested similar timing based on apatite fission-track data. However, a highly detailed data set from Wang et al. (2012) revealed that these earlier studies did not capture the full Cenozoic exhumation history of the Pengguan Massif, and that the 10–12 Ma cooling signal identified by these studies did not mark the onset of exhumation in the Longmen Shan. Instead, Wang et al. (2012) identified two distinct phases of rapid exhumation in the Pengguan Massif at 25–30 Ma (∼800 m/m.y.) and at 10–15 Ma (∼350 m/m.y.), separated by a period of slower exhumation (70–100 m/m.y.), and preceded by another period of prolonged slower exhumation from 50 to 30 Ma (100–125 m/m.y.).

The relationship between the Oligocene exhumation in the Longmen Shan documented by Wang et al. (2012) and the uplift of the plateau to the west is unclear. West of the Longmen Shan, apatite (U-Th)/He data collected from major river valleys cut into the plateau suggest that rapid incision of the Yangtze, Yalong, and Dadu Rivers began at ca. 9–13 Ma (Clark et al., 2005; Ouimet et al., 2010). The authors proposed that the onset of incision was related to regional uplift of the eastern plateau. Xu and Kamp (2000), in contrast, suggested 20 Ma for the onset of exhumation in the eastern plateau, based on an extensive data set of apatite fission-track (AFT) and zircon fission-track (ZFT) ages within the plateau.


The Baoxing region makes up the southwestern section of the Longmen Shan (Fig. 1). To the northeast, the Baoxing Massif is separated from the Pengguan Massif by a zone of Paleozoic and Triassic rocks. To the southwest of Baoxing, the structures of the Longmen Shan end in a complicated zone with both northeast-southwest– and northwest-southeast–oriented thrust faults of unknown age. The relationship between the two sets of structures is unclear. The fold-and-thrust belt is wider in the Baoxing region than in the northeastern Longmen Shan, as folding extends farther into the Sichuan Basin foreland (Fig. 1). Topography in the Baoxing region is more subdued than in the Pengguan region; maximum elevations are lower (∼3500 m in Baoxing vs. ∼5000 m in Pengguan), and slopes are slightly less steep.

The Baoxing region contains two bodies of Precambrian crystalline rocks separated by Paleozoic strata and a series of faults (Figs. 2 and 3). The two Precambrian bodies are associated with different sequences of cover rocks. The western body is overlain by a thick sequence of metamorphosed Sinian rocks, and then Ordovician to Permian schist, marble, phyllite, conglomerate, and volcanic rocks. The eastern body, known as the Baoxing Massif, is overlain by a thin unmetamorphosed succession of Sinian and Ordovician to Lower Triassic (Cambrian rocks are missing) dolomite, shale, limestone, siltstone, and sandstone that becomes increasingly incomplete (lacking Silurian, Ordovician, and Carboniferous) in the central and northern parts of the massif (Figs. 2 and 3). East of the massif, these strata are in depositional contact with the Precambrian rocks, and the sequence is preserved with no faults (Ministry of Geology and Mineral Resources, 1991). Burchfiel et al. (1995) proposed that the western granite and associated cover rocks were part of the thrust sheet emplaced on the Yangtze craton rocks in the Mesozoic, while the eastern granite and associated cover rocks were associated with the Yangtze block.

The metasedimentary strata associated with the western granite are stacked in a series of northwest-dipping thrust faults (Figs. 2 and 3). Metamorphism is largely greenschist grade. Field observations indicate that deformation was largely ductile and included some mylonitization of the western granite body and cover rocks. In a transect along the Xihe River near Wulong village (Fig. 2), we observed several sets of ductile fabrics. On the south side of the river, the C plane of an older S-C fabric strikes 320°, while a younger fabric has a C plane striking 340°–355°, and both indicate a thrust sense of shear, with C dipping less steeply than S. Both fabrics have a component of left-lateral shear, with lineations that plunge to the west or northwest. On the north side of the river, a third S-C fabric shows S striking 10° and dipping 66°W and C striking 40° and dipping 42°N, indicating thrusting with a component of right-lateral shear. This C surface is parallel to the 40°-striking and 40°N-dipping mapped Wulong fault that places the western granite body above Devonian phyllite, and both are oblique to the fabrics on the south side of the river. These observations indicate that early thrusting was accompanied by a component of left-lateral shear, while the later thrusting, which is related to the mapped fault, was accompanied by right-lateral shear.

Cenozoic Structures

The Baoxing region is cut by a series of west-dipping Cenozoic or likely Cenozoic faults. From east to west, these are the Shuangshi/Lingguan fault, the Erwangmiao fault, an unnamed normal fault, and the Wulong fault (Jia et al., 2006; Arne et al., 1997; Burchfiel et al., 1995) (Figs. 2 and 3). The easternmost fault, the Shuangshi/Lingguan fault, cuts the Mesozoic foreland basin sequence, emplacing Upper Triassic Xujiahe Formation above Upper Triassic, Jurassic, and Cretaceous rocks. This fault is associated with folding of the footwall rocks adjacent to the fault (Burchfiel et al., 1995). To the west, the Erwangmiao fault thrusts the Sinian to Lower Triassic cover rocks of the Baoxing Massif above Upper Triassic foreland basin deposits. This fault is located within the klippen belt, obscuring some of the structural relationships (Fig. 3) (Ministry of Geology and Mineral Resources, 1991).

On the western side of the Baoxing Massif, a normal fault of unknown age separates Permian and Triassic rocks from the Precambrian rocks of the massif (Burchfiel et al., 1995). Shear sense indicators, including S-C fabrics in a broad shear zone within the granite, indicate west side down, or normal sense of motion. Shear zones in the granite consist of both ductile (S-C fabric forming) and brittle deformation (Burchfiel et al., 1995). The timing of faulting is unknown, but this fault may be related to a similar zone of normal sense shear on the western side of the Pengguan Massif along the Wenchuan-Maowen fault zone (Burchfiel et al., 1995). Several kilometers farther west, the Wulong thrust places the western body of Precambrian rocks above metamorphosed Silurian and Devonian units. The Wulong fault appears to reactivate an older ductile shear zone, as both the footwall and hanging-wall rocks become increasingly deformed to mylonitic approaching the fault (see previous).

Quaternary–Recent Deformation

The magnitude of active deformation in the southwestern Longmen Shan is unclear. Several recent studies have looked at shortening to the north (Ma et al., 2005; Densmore et al., 2007), but, as the Wenchuan earthquake illustrated, there may be considerable along-strike variation in amounts of shortening and strike-slip motion (Xu et al., 2009), so we hesitate to extrapolate these data south to the Baoxing region.

Global positioning system (GPS) data and associated block models generally do not distinguish variations along the Longmen Shan front and treat the range as a single block. Estimates of slip rate along the range front vary from 1.7 ± 0.9 mm/yr right slip and 1.2 ± 0.1 mm/yr convergence (H. Wang et al., 2010; Wang et al., 2008) to 0.3 mm/yr right slip and 4.1 mm/yr convergence (Gan et al., 2007). Burchfiel et al. (2008) described a model that does separate the Longmen Shan into two blocks; they did not report slip rates for the southern block but did suggest that convergence increases to the northeast.

There is limited field evidence for Quaternary deformation in the southwestern Longmen Shan. Near Wulong village, Yang et al. (1999) reported that a 92 ± 7.2 ka deposit is offset 73 cm by thrust faulting, and that the fault is capped by an undeformed 78.5 ± 6.1 ka deposit. This suggests that the Wulong fault has been active in the Quaternary, but it may not be currently active.

Existing Thermochronology

Few previous thermochronology studies have included the Baoxing region (Fig. 2). In the Baoxing River valley, Arne et al. (1997) reported three AFT ages between 4 and 11 Ma, while Kirby et al. (2002) reported a feldspar Ar/Ar age of 150 Ma. To the east of the Baoxing Massif, in the Mesozoic foreland strata, Arne et al. (1997) reported three Mesozoic AFT ages (93–108 Ma). The small number of data points and the large errors on the AFT ages provide few constraints on the cooling history of the Baoxing region.


In order to constrain the cooling and exhumation history of the Baoxing region, we collected a suite of samples for dating with the apatite (U-Th)/He, zircon (U-Th)/He, and ZFT thermochronometers. The use of several different thermochronometers, with nominal closure temperatures of ∼60 °C (apatite [U-Th]/He), ∼180 °C (zircon [U-Th]/He), and ∼240 °C (ZFT) (Reiners and Brandon, 2006), allows us to determine a more detailed cooling history for each sample. The Baoxing area is characterized by steep, heavily forested slopes, and thus exposure is poor, and access is generally limited to large river valleys. In the Baoxing Massif, we collected a transect of four samples between 960 m and 2400 m elevation just east of Baoxing City, as well as three additional samples in the Baoxing River valley. In the western granite, we were able to collect one sample each from the southern, central, and northern river valleys (Fig. 2). These samples ranged in elevation from 1062 m to 1800 m. In total, 10 samples were analyzed for apatite (U-Th)/He, 7 samples for zircon (U-Th)/He, and 5 samples for AFT (Table 1; Fig. 4).

Samples were crushed, sieved, and washed, and apatite and zircon fractions were isolated using standard mineral separation techniques. For (U-Th)/He thermochronology, individual apatite and zircon grains were handpicked, evaluated for inclusions, measured under a 100× binocular microscope, and packed into platinum (apatite) and niobium (zircon) capsules. Four to five single-grain replicates were analyzed for each sample. Three apatite (U-Th)/He samples (KC05–11, KC05–14, KC05–15) were analyzed at the University of Arizona, using the procedures described in Reiners and Nicolescu (2006). The remaining apatite and all zircon (U-Th)/He samples were analyzed at Arizona State University using the procedures described in Schildgen et al. (2009). Durango apatite and Fish Canyon Tuff zircon were used as standards. We corrected for He loss with the alpha ejection (FT) corrections of Farley et al. (1996) and Farley (2002) for apatite, and Hourigan et al. (2005) and Reiners (2005) for zircon.

For ZFT analyses, grains were processed and analyzed at National Chung-Cheng University following the procedures described by Liu et al. (2001), using the external detector method and the Fish Canyon Tuff standard. Between 13 and 20 grains were counted for each sample. The dosimeter glass NBS-610 was used, with a zeta calibration value of 28 ± 1.

Results and Discussion


Apatite (U-Th)/He ages range from 3.3 to 10.1 Ma (Table 2; Fig. 4A). In several samples, one or more of the replicates are clearly anomalously old (older than the zircon He and ZFT ages), likely due to unseen inclusions in the crystals, since the apatites from this region are generally inclusion rich and of poor quality. These ages were disregarded and are not included in the averages given in Table 1. We obtained at least two consistent replicates for each sample, except for sample KC07–57. For this sample, we discuss the age of the youngest replicate, as the other replicates gave ages older than the zircon He age. The variability in individual grain ages highlights the need for multiple single-grain replicates per sample, particularly when apatite quality is poor.

Zircon (U-Th)/He ages range from 5.7 to 17.3 Ma (Table 3; Fig. 4B). Individual replicates generally agree well for each sample, although there are a few outlier replicates (i.e., KC07–55b), and some samples (i.e., KC07–56) appear to have two different populations. Sample KC07–58d was discarded because the grain was missing a large chip. For the rest of the grains, variations in size or morphology do not explain the variability among the replicates.

ZFT ages range from 10.9 to 14.4 Ma (Table 4; Fig. 4C). Samples KC07–53 and KC07–54 yielded zircons that were not datable. This may be the result of radiation damage (see subsequent section), as high amounts of damage can affect the ability to etch and view tracks (Garver and Kamp, 2002).

Age-Elevation Relationships

A striking feature of the data is the lack of a relationship between age and elevation for the zircon (U-Th)/He data (Fig. 4B; Table 1). In the samples from the western granite, there is no relationship between elevation and cooling age. However, the samples are spaced over a distance of 20 km, and the elevation range is relatively narrow (740 m), so the age distribution could be explained by lateral variations in cooling history.

On the eastern side of the river, a suite of four samples was collected over a short distance (∼6 km) with a 1440 m change in elevation (Fig. 2 inset). These samples also exhibit a poor correlation between age and elevation, particularly for the zircon (U-Th)/He ages. A possible explanation is that the oldest sample, KC07–55, is separated from samples KC07–53 and KC07–54 by a mapped fault that places a narrow belt of Paleozoic rocks on top of the granite (Fig. 2 inset). The fault offsets the depositional contact between Precambrian granite and Sinian carbonate by several kilometers horizontally and 500–1000 m vertically, and stratigraphic relations indicate that the fault must have normal and/or right-lateral sense of slip. Although the fault is not exposed along the sample transect, small shear zones in the granite near the fault dip 45–60°W. Post–15 Ma motion on this fault and the resulting offset in the sample locations could explain the observed ages, particularly if the fault had a significant component of normal sense slip. Because the apatite (U-Th)/He ages are not affected, motion must have ceased by ca. 7 Ma. However, there is no further evidence that the fault has been active in the Cenozoic, and it may instead be a Mesozoic feature.

Possible Effects of Radiation Damage

Alternatively, the variation in ages between these samples could be related to the effects of radiation damage, which is thought to affect the diffusion of helium in zircon (Reiners, 2005; see Appendix). Effective uranium concentration (eU, where eU = 238Uppm + 0.235[232Thppm]) has been used in apatite (U-Th)/He studies (Shuster et al., 2006; Flowers et al., 2009) as a proxy for the amount of accumulated radiation damage in grains with similar thermal histories. For our samples, a plot of eU against measured age shows a correlation between eU and the age of individual crystals, as zircons with higher eU tend to have younger ages (Fig. 5). For older samples, the association of younger ages with higher eU is seen within samples as well as for multiple samples within each tectonic unit. This is particularly apparent in the samples from the Baoxing Massif. The observed relationship between eU and zircon He age suggests that radiation damage may have affected our samples, despite the lack of obvious warning signs for potential radiation damage (high uranium concentrations and old cooling ages).

In order to evaluate the degree of radiation damage, we analyzed zircons from samples KC07–53 and KC07–54 with a Nicolet Almega XR dispersive Raman spectrometer with a 780 nm laser and an Olympus BX51 microscope with a 50× objective. The resulting spectra lack distinct peaks and indicate that the analyzed zircons are highly disordered, suggesting that they have accumulated a significant amount of radiation damage (Fig. 6). Previous studies of the effects of radiation damage on zircon (U-Th)/He ages have focused on zircons with cooling ages of ∼500 m.y. or older; our data suggest that radiation damage may be an issue even in samples with relatively young cooling ages.

The relationship between uranium concentration and zircon (U-Th)/He age may also be influenced by U and Th zoning in the zircons. As the standard alpha ejection correction assumes a homogeneous distribution of U and Th, zircons with enriched rims will give ages that are too young, while zircons that have depleted rims will give ages that are too old (Hourigan et al., 2005). If the zircons in our sample are strongly zoned, and variations in bulk eU are largely due to variations in the eU of enriched zircon rims, then younger ages for zircons with higher eU would be expected. During our ZFT analyses, we did not observe the significant variations in rim and core fission-track densities that strong zoning would produce. To quantify possible zoning, we used laser ablation inductively coupled plasma–mass spectrometry (ICP-MS; with an Agilent 7500cx ICP-MS and a New Wave UP213 laser ablation system) to measure relative uranium concentrations in the rims and cores in 10 zircons from sample KC07–55 and found only minor zoning. The maximum ratio of rim/core concentration was 1.8, and the average ratio was 1.07. Based on Hourigan et al. (2005), rim enrichment of this scale could lead to a maximum age bias of ∼10%, and the relative bias of different grains will be some fraction of that. Further, there is no relationship between grain radius and (U-Th)/He age, as would be predicted by the model of Hourigan et al. (2005). We argue that zoning is not sufficient to explain the observed age-eU relationship, and that the relationship is the result of radiation damage.

The effects of radiation damage on both helium diffusion and the annealing of fission tracks in zircon are not well understood. It has been demonstrated that damage affects zircon He ages in old rocks (Reiners, 2005), but whether this is due to lowering of the effective closure temperature, or to prolonged slow diffusion at very low temperatures is unclear. If radiation damage affects the effective closure temperature of the zircon (U-Th)/He system, then the observed ages could reflect the passage of samples through different isotherms, or, if damage leads to continued diffusion of helium at low temperatures, as suggested by some experiments (Reiners, 2005), the age distribution could reflect later helium loss, and not passage through the closure isotherm. We therefore treat our zircon He ages as minimums, and, for the eastern samples, primarily discuss the age obtained from sample KC11–55, which has the lowest uranium concentration and is likely to be least affected by damage.

For the ZFT system, differences between closure temperatures estimated in laboratory experiments and those based on field constraints have been attributed to radiation damage (Kasuya and Naeser, 1988; Garver and Kamp, 2002; Rahn et al., 2004; Reiners and Brandon, 2006; Bernet, 2009). Studies on zircons thought to be highly damaged suggest closure temperatures of 205 °C (Bernet, 2009) or 210 °C (Zaun and Wagner, 1985) and a partial annealing zone down to 180 °C for ZFT (Zaun and Wagner, 1985; Brandon et al., 1998). Because our samples appear to have a substantial amount of radiation damage, we adopt the closure temperature of 205 °C suggested by Bernet (2009). This is consistent with the existing temperature constraints and allows for a Cenozoic ZFT age in the Baoxing Massif despite the lack of metamorphism in the Paleozoic cover rocks, which indicates a maximum temperature of ∼250 °C (see “Magnitude of Exhumation in Baoxing Region” section).

The strong correlation between zircon (U-Th)/He age and uranium concentration, both among sample replicates and across different samples, suggests that radiation damage has had a significant effect on the measured ages. We conclude that the age-elevation pattern may be due to the effects of radiation damage alone, or due to a combination of radiation damage and slip along the small fault discussed in the “Age-Elevation Relationships” section (Fig. 2 inset); without further data, we can neither confirm nor rule out Cenozoic slip on this fault. Because of this uncertainty, we do not attempt to use the age-elevation relationship to draw any conclusions about the exhumation of the Baoxing Massif.


Although the lack of age-elevation relationships and the potential effect of radiation damage on the zircon (U-Th)/He ages complicate the interpretation of the data, a number of firm conclusions can still be drawn.

Magnitude of Exhumation in Baoxing Region

All of the samples give Miocene to Pliocene ages for all three thermochronometers, indicating that the Baoxing region has cooled from at least ∼205 °C (see previous discussion of radiation damage and ZFT closure temperatures) during the Cenozoic, and suggesting a minimum of 7–10 km of exhumation in Miocene–Pliocene time, assuming a geothermal gradient in the range of 20–30 °C/km.

Several lines of evidence provide constraints on the maximum exhumation in the Baoxing area: (1) The lack of metamorphism in the Paleozoic cover rocks provides the most important constraint on exhumation depth. Sinian dolostone directly overlying the Precambrian basement is completely unmetamorphosed, indicating maximum temperatures of ∼250 °C. This is consistent with the stratigraphy, as the Paleozoic sequence is very thin (ranging from several hundred meters to ∼1 km in thickness) and contains several unconformities (Ministry of Geology and Mineral Resources, 1991), suggesting that the Precambrian rocks of the Baoxing Massif remained at shallow depths from Sinian to Middle Triassic time. (2) The degree of radiation damage in several of the samples requires long residence times at intermediate or low temperatures (likely below ∼400 °C; see Appendix). (3) A single feldspar Ar/Ar analysis by Kirby et al. (2002) suggests that the Baoxing region experienced slow cooling at temperatures around 200 °C from 300 Ma to at least 100 Ma. This cooling history may be oversimplified, since it assumes monotonic cooling during a period that includes events such as the Indosinian orogeny and the eruption of the Emeishan basalts, but the temperatures are consistent with the metamorphic data. These observations together show that the Baoxing Massif experienced limited exhumation over both the Cenozoic and the Mesozoic orogenies. The maximum temperature indicated by the lack of metamorphism suggests a maximum depth of 8–13 km (assuming a 20–30 °C/km geothermal gradient) for the Baoxing Massif.

The western granite body has experienced greater total exhumation. Metamorphism in the sedimentary cover reaches greenschist grade, with temperatures of ∼300–400 °C. The rocks are deformed and stacked in a series of thrust faults, many of which show evidence of ductile deformation. Much of this deformation and exhumation was likely Mesozoic, as the difference in the thermochronology ages suggests only a few kilometers of Cenozoic differential exhumation between the western and eastern rocks (see “The Wulong Fault” section).

Timing of Exhumation in Baoxing Region

The young ZFT ages from our samples indicate that the samples were residing above the ZFT closure temperature (∼205 °C for zircons with radiation damage; see “Possible Effects of Radiation Damage” section) prior to ca. 15–17 Ma. This indicates that our samples are not from an exhumed zircon (U-Th)/He partial retention zone, and that the zircon (U-Th)/He ages record passage through the closure isotherm during exhumation. Therefore, cooling and exhumation of the Baoxing Massif were occurring by ca. 15 Ma. The western granite was cooling rapidly by ca. 11 Ma. Because there is no rollover or older ages, an upper bound on the onset of exhumation is not constrained by these data, and 15–17 Ma represents a minimum age for the onset of exhumation in the region.

Cooling Rates

In the Baoxing Massif, the apatite and zircon He data suggest average cooling rates of ∼8–15 °C/m.y. since ca. 15 Ma. Cooling rates calculated from both the difference between the two thermochronometers, and the age and closure temperature for each system (with an ambient temperature of 10 °C) are generally consistent, providing no evidence for a significant change in cooling rate during that time. The data from the western granite suggest slightly faster cooling rates, and they permit a slight deceleration in cooling from ∼20 °C/m.y. between 10 and 4 Ma to 10–15 °C/m.y. since ca. 4 Ma (disregarding the 5.7 Ma zircon He age of sample KC07–57, which requires extremely rapid cooling of 140 °C/m.y. between 5.7 and 4.9 Ma). For a geothermal gradient between 20 and 30 °C/km, these cooling rates suggest average exhumation rates of ∼0.3–0.8 km/m.y. for the Baoxing Massif, and rates of 0.7–1 km/m.y. from 10 to 4 Ma and 0.3–0.8 km/m.y. since 4 Ma for the western granite.

The Wulong Fault

For all three thermochronometers, the samples on the western side of the Wulong fault give younger ages than the samples to the east (Fig. 4). The difference in ages across the fault is seen most clearly in the zircon He data, where the samples form two very distinct populations (Figs. 2, 4B, and 5). The difference between the eastern and western zircon ages ranges from 2 to 9 m.y. Apatite He ages west of the fault tend to be ∼1–2 m.y. younger than ages east of the fault. The offset in ages across this fault indicates that the fault accommodated differential exhumation in the Cenozoic sometime between at least 10–12 Ma and the present (although it could have started earlier). The difference in the thermochronology ages cannot be directly converted to an offset without knowledge of the duration of faulting, the exhumation rate, or the cooling rate and geothermal gradient. However, if we use a cooling rate suggested by the helium ages of 10–15 °C/m.y., and a geothermal gradient of 20–30 °C/km, a 4 m.y. offset in cooling ages would correspond to a minimum differential exhumation of 1.3 km (10 °C/m.y. and 30 °C/km) and a maximum of 3 km (15 °C/m.y. and 20 °C/km). Thus, Cenozoic uplift along this fault appears limited to several kilometers. The thermochronology data are supported by the observation of Quaternary deformation by Yang et al. (1999). In the field, the Wulong fault is largely unexposed, aside from a single outcrop of gouge near the town of Wulong. The potential for activity on the Wulong fault suggests that this structure should be investigated further.

Frontal Faults

Mesozoic AFT ages were measured by Arne et al. (1997) east of the Baoxing Massif in the foreland basin sedimentary rocks (Fig. 2). The youngest of these ages (93 Ma) was measured in a section of Upper Triassic rocks located between the Erwangmiao and Shuangshi/Lingguan faults, while slightly older ages (105 Ma and 108 Ma) were measured east of the Shuangshi fault. The Erwangmiao fault therefore juxtaposes Cenozoic ZFT ages in the Baoxing Massif with Mesozoic AFT ages in the Sichuan Basin foreland, suggesting that this fault has accommodated differential exhumation. If samples west of the fault were cooled from at least 205 °C, and the samples to the east remained below ∼110 °C (AFT closure temperature; Reiners and Brandon, 2006), at least 85 °C of differential cooling took place. This would suggest a minimum of several kilometers of exhumation along this fault. The Shuangshi fault, in contrast, separates rocks with only minor differences in cooling age, and therefore accounts for a relatively minor amount of differential exhumation.

Relationship between Baoxing Region and Wenchuan Region

The Baoxing and Pengguan Massifs have experienced similar amounts of Cenozoic exhumation. Wang et al. (2012) obtained Cenozoic zircon (U-Th)/He ages and 175–370 Ma ZFT ages from the Pengguan Massif, constraining the maximum amount of cooling since the Mesozoic to less than ∼250 °C, and the amount of cooling in the Cenozoic to between 180 °C and 250 °C. Like the western granite in the Baoxing region, the Xuelongbao granite, west of the Pengguan Massif (Fig. 1), has experienced greater total exhumation, as evidenced by the higher metamorphic grade of the Paleozoic cover (Burchfiel et al., 1995) and Cenozoic ZFT ages (Furlong and Kirby, 2013).

The possibility that ZFT and (U-Th)/He ages in the Pengguan Massif may be affected by radiation damage remains in question. Wang et al. (2012) suggested that the appearance of many of the zircons from the Pengguan Massif is indicative of radiation damage, but their (U-Th)/He ages do not show a strong correlation with uranium concentration. The age-eU relationships for some of the zircon (U-Th)/He data from Godard et al. (2009), in contrast, are suggestive of radiation damage effects, but these are not sufficient to draw any firm conclusions about the prevalence or influence of radiation damage in these zircons.

The timing of exhumation also appears to be similar in the Baoxing and Pengguan Massifs. Wang et al. (2012) showed that rapid cooling in the Pengguan Massif occurred in two phases, starting at 25–30 Ma and 10–15 Ma. Our data are roughly consistent with the upper bound on the second phase of exhumation, with cooling initiating at ca. 15 Ma; earlier exhumation in the Baoxing region remains unconstrained.

Our cooling data also prompt us to consider the relationships between structures in the Baoxing region and the faults in the central and northern Longmen Shan. Many workers assume the fault geometry depicted in Figure 7A, and consider a fault northwest of the western granite to be the southern continuation of the Wenchuan fault, the Wulong fault to be the southern continuation of the Beichuan fault, the Shuangshi fault to be the southern continuation of the Pengguan/Guanxian fault, and the Erwangmiao fault to be unimportant. However, the details of this fault geometry are often not considered. The Wulong fault occupies a different structural position than the Beichuan fault, which separates Precambrian basement from Triassic foreland basin sedimentary rocks that give Mesozoic AFT ages (Arne et al., 1997; Godard et al., 2009; Ministry of Geology and Mineral Resources, 1991, 1:200,000 geologic maps). The Wulong fault, in contrast, juxtaposes two successions of Precambrian and Paleozoic rocks, both with Cenozoic ZFT ages. The thermochronology data indicate that the Wulong fault is not the fault primarily responsible for the exhumation of the basement in this region, and it has likely not accommodated as much uplift as the Beichuan fault. The Wulong and Beichuan faults have had different histories, and if they are connected at present, this connection must have been fairly recent. Structurally, the Wulong fault is in a similar position as the Wenchuan fault, and a link between them is consistent with the findings of Furlong and Kirby (2013), who concluded that, like the Wulong fault, the Wenchuan fault has accommodated differential exhumation in the Cenozoic.

The Erwangmiao fault separates Cenozoic ZFT ages and Mesozoic AFT ages (Arne et al., 1997) and thrusts the Precambrian to Lower Triassic succession above the Upper Triassic foreland basin sediments, putting it in the same structural position as the Beichuan fault. Like the Beichuan fault, the Erwangmiao fault appears to be an important factor in the exhumation of the Precambrian massif to the west. The Shuangshi fault likely accommodates a relatively small amount of shortening in the front of the range, and therefore may be analogous to the Pengguan/Guanxian fault.

The thermochronology data suggest that the fault geometry shown in Figure 7B, rather than the straight through-going faults depicted in Figure 7A, is more accurate over the long-term Cenozoic exhumation of the Longmen Shan. While the geometry may change with time, and it is possible that Figure 7A is more accurate for the present active structures, our data suggest that the modern fault geometry needs to be examined more critically. This has important implications for potential seismic hazard in the southwestern Longmen Shan, as the similar roles of the Erwangmiao fault and the Beichuan fault in accommodating uplift of the Longmen Shan suggest that the Erwangmiao fault, like the Beichuan fault, may be an active fault with the potential for large earthquakes.

The April 2013 Lushan Earthquake

The potential for seismic hazard in the southwestern Longmen Shan was made clear by the magnitude 6.6 earthquake that struck the region on 20 April 2013. Initial reports of the location of the earthquake by the U.S. Geological Survey and the China Earthquake Administration indicate that the hypocenter was located about 16 km southwest of Baoxing at a depth of 14 km (Fig. 2). The moment tensor solutions provided by the U.S. Geological Survey indicate an almost pure dip-slip event with southeast-directed shortening. Aftershocks are concentrated in a zone about 35 km long, and at 15–25 km depth, and roughly define a trend dipping about 40° to the NW (China Earthquake Administration, 2013). While the epicenter of the earthquake is very near the Shuangshi fault, the depth and pattern of aftershocks suggest that the earthquake must have ruptured a deeper and more frontal fault, potentially the Ya’an Thrust (Fig. 2), which was predicted to experience a Coulomb stress increase following the 2008 Wenchuan earthquake (Parsons et al., 2008). The Ya’an Thrust is within Cretaceous rocks for much of its length, but at its southern end Jurassic rocks appear above Cretaceous rocks, and folding in the hanging wall exposes Jurassic strata at the northern end of the fault (Fig. 2). Based on the 1:200,000 geologic map (Ministry of Geology and Mineral Resources, 1991), the central part of the fault is contained within Cretaceous units with a maximum thickness of approximately 1 km, suggesting that the total offset along the Ya’an fault has been relatively limited. While the Ya’an fault is a possible source for the earthquake, the lack of shallow aftershocks suggests that the earthquake may have been on a deeper blind fault, and may not have a surface rupture. The 40° dip outlined by the distribution of aftershocks may indicate that the earthquake ruptured a ramp on a detachment extending into the Sichuan Basin. If the initial location and 14–25 km depth of the aftershocks and epicenter are accurate, they suggest that a deep detachment extends farther into the Sichuan Basin than existing structural models have predicted (Jia et al., 2006; Xu et al., 2009; Hubbard et al., 2010).


From our data set, the first set of cooling ages from the Baoxing Massif, we can draw a number of conclusions about the exhumation and cooling history of the Baoxing region, and about the potential for radiation damage affecting zircon (U-Th)/He ages.

  • (1) Exhumation of the Baoxing region was ongoing by ca. 15 Ma, which is broadly consistent with the 10–15 Ma phase of exhumation proposed by Wang et al. (2012) for the Wenchuan region.

  • (2) Cenozoic exhumation in the Baoxing region can be constrained to a minimum of 7 km and a maximum of 13 km in the Baoxing Massif and a minimum of 7–10 km in the western granite. This is similar to the magnitude of exhumation in the Wenchuan region.

  • (3) The difference in cooling ages between the samples from the Baoxing Massif and the western granite indicates that there has been up to several kilometers of differential exhumation along the Wulong fault since ca. 10–12 Ma.

  • (4) The distribution of cooling ages in the Baoxing region suggests that the Erwangmiao fault is the primary frontal fault in the southwestern Longmen Shan and is analogous to the Beichuan fault in the central Longmen Shan.

  • (5) Our zircon (U-Th)/He data suggest that radiation damage in zircon may affect the ages of zircons with relatively young cooling ages, and that this effect warrants further investigation.


Radiation Damage in Zircons

Studies have recognized that radiation damage can affect the diffusion of helium in both apatite and zircon (Reiners, 2005; Shuster et al., 2006; Flowers et al., 2009). In zircon, a high amount of radiation damage is thought to increase the diffusivity of helium, and therefore lead to younger (U-Th)/He ages and/or lower effective closure temperatures (Reiners, 2005). The amount of radiation damage in a crystal is a function of the dose rate, which depends on the uranium and thorium concentration, and the time over which the damage accumulates, which depends on the age and cooling history of the crystal. Radiation damage is thought to affect grains that have accumulated radiation doses of at least 2e18 α/g, and it has been described in samples with cooling ages of 440 to over 600 Ma (Reiners, 2005).

Raman spectroscopy has been used to evaluate the degree of radiation damage in zircons (Zhang et al., 2000; Nasdala et al., 2001; Palenik et al., 2003). Radiation damage disrupts the crystal structure of zircons, resulting in spectra that have increasingly broad and poorly defined peaks as the degree of damage increases. Crystals with a high amount of damage become largely amorphous and have spectra with broad humps and no distinct peaks (Fig. 6) (Nasdala et al., 2001).

In laboratory experiments, radiation damage in zircons begins to anneal at temperatures above 525 °C (Zhang et al., 2000; Nasdala et al., 2002), but the minimum temperature necessary for annealing is unknown, and the effect of prolonged residence at intermediate temperatures has not been well established. Garver and Kamp (2002) found zircons that retained alpha radiation damage but had reset fission-track ages, and they suggested that these samples come from the zone of temperatures between ∼250 and 400 °C. Reiners (2005) calculated that zircons with eU between 700 and 1000 ppm would require 500–750 m.y. to accumulate a radiation dose of 2e18 α/g. This suggests that the sampled zircons have remained below the temperature required to anneal radiation damage since Precambrian time.

This work was supported in Taiwan by National Science Council (NSC) grant 2811-M-002–092 and in the United States by the National Science Foundation (NSF) Continental Dynamics Program (grant EAR-0003571). We thank Owen Aftreth, Li Junmin, and Huang Sihua for assistance in the field. E. Enkelmann and E. Kirby provided detailed and constructive reviews.