Invited reviewResetting Southern Tibet: The serious challenge of obtaining primary records of Paleoaltimetry
Introduction
The Himalaya and Tibet together constitute the largest active orogenic system in the world today, so naturally there is significant interest in the history of their development (see Kapp and DeCelles, 2019 for one recent summary). How the absolute elevation of this system evolved through time is a key and still evolving part of the history. Some of the earliest attempts at paleoaltimetric reconstruction using stable isotopes were conducted in the Himalaya (e.g., Garzione et al., 2000), and there have been dozens of studies since then. Our purpose here, however, is not to summarize the tectonic/geodynamic implications of those studies, which has been undertaken elsewhere (e.g., Rowley and Currie, 2006; Deng and Ding, 2015; Ding et al., 2017). Instead, our focus here is more methodological, on the question of the reliability of these reconstructions, many of them our own, using the stable isotopic (δ18O, δ13C, Δ47) composition of secondary carbonates deposited in lakes, in soils, and by ground water. Our study region lies in southern Tibet (Fig. 1) and focuses on the isotope records from where India sutured to Asia (the Indus-Yarlung suture zone, or IYSZ), and its immediate flanks in the Gandese magmatic arc along the southern edge of the Asian plate. Structural and thermochronologic studies show that deformation and exhumation have been much more active inside the IYSZ than outside it (Rohrmann et al., 2012; Carrapa et al., 2014; Laskowski et al., 2018), a difference that may impact patterns of isotopic alteration of marine and non-marine carbonate rocks in these areas.
The stable oxygen isotopic composition (expressed in ‰ using the familiar δ18O notation) of carbonates is a mainstay of many aspects of our understanding of Earth history. However, the serious challenge of obtaining unaltered, primary δ18O values from carbonates after exposure on land was recognized by even the earliest isotope geochemists (Revelle and Fairbridge, 1957; Clayton and Degens, 1959; Keith and Weber, 1964). It has long been recognized that carbonates undergo oxygen exchange, but much less commonly carbon, when in prolonged contact with meteoric water and terrestrial vegetation. The reason for this is that the atomic C/O ratio of typical groundwater is ~10−5, meaning that there is far more O than C in ground water to exchange with carbonates. As a result, marine carbonates with unaltered δ13C values are commonplace back into the Proterozoic, whereas over the same time span unaltered marine δ18O values are only preserved in special circumstances (Veizer et al., 1999; Ghosh et al., 2018), if at all (Degens and Epstein, 1962; Killingley, 1983).
To illustrate these issues for our study region, we compiled the published δ18O (VPDB) and δ13C (VPDB) values (see Fig. 2, caption; Table S1) of mid-Cretaceous to Paleogene (K-Pg) marine limestones from southern Tibet and northern India. These are the youngest marine rocks in the region, and the closest stratigraphically to the non-marine carbonates used for reconstruction of paleoaltitude. We make several observations about this large data set of ~1300 analyses, the first being that all δ18O (VPDB) values from Tibetan/Himalayan limestones, which range from −4.2 to −22‰, are lower than primary marine values (>−4‰) for this time period (Veizer et al., 1999; Kobashi et al., 2001), showing that the δ18O values of virtually all sampled limestones have been reset. The range of “primary” marine values was established by these studies through analysis of thousands of well-preserved, low-Mg calcites precipitated by brachiopods, belemnites and other organisms living in a wide range of marine habitats. Our second observation is that published δ18O values (−13.6 ± 2.7‰, n = 36) obtained from limestones in the IYSZ and Gangdese arc are generally much lower than those from surrounding regions to the north and south (−7.1 ± 2.2‰, n = 620). Lastly, the majority (~1200) of δ13C values fall within the range of normal marine values (>−1.9 to +5‰) for the period, whereas ~20 samples yielded low, nonmarine values of <−2‰ (discussed later).
A chief takeaway from these observations is that all δ18O values of the marine carbonates, as with most older limestones globally (Keith and Weber, 1964; Veizer et al., 1999), have been diagenetically altered due to exchange between marine limestones and meteoric waters, at variable and often elevated diagenetic temperatures (e.g., Hudson, 1977). Paleoelevation reconstruction by isotopic methods is underpinned by the assumption of preservation of primary δ18O and Δ47 values in carbonates deposited in lakes and the vadose zone. And so the regional resetting of carbonate δ18O values calls into question reconstruction paleoelevations using the δ18O values of many equivalent-age (late Eocene and older) nonmarine carbonates in the same region. The same has to be true for Δ47, since it must change if the bulk isotopic composition changes. However, patterns of alteration of δ18O values in marine rocks strongly differ inside and outside the IYSZ (Fig. 2a), a distinction that has major implications for the state of isotopic preservation of non-marine carbonates that overlie these marine rocks in the same basins. To strengthen this test, for our study we more than doubled the sample size of marine limestones from the IYSZ.
The effects of diagenesis should diminish in carbonate samples younger than the late Eocene, (<50 Ma), depending on the local circumstances of burial. We will argue that it is vital to evaluate these circumstances from as many perspectives as possible, a practice that much of the published literature on paleoelevation is only now embracing. Other lines of evidence that assist in understanding burial durations and temperatures include the thickness of overburden, thermochronology (U-Th/He) and fission tracks in apatite and zircon, 40Ar39Ar, petrography, and alteration mapping and mineralogy. In this paper, we offer examples of many of these lines of evidence for our study areas.
Finally, our regional survey of published isotopic composition of marine limestones shows that ~98% of the samples (Fig. 2a) preserve plausibly primary δ13C values of between −2 to +4‰; as defined (95% confidence interval by Veizer et al. (1999) and Kobashi et al. (2001) for K-Pg marine limestones globally. As noted above, this is consistent with the low inorganic carbon content of most natural waters and justifies the use of δ13C values to interpret paleoenvironmental conditions on land in deep time (e.g., Ekart et al., 1999; Breecker et al., 2010). The δ13C value of paleosol carbonate is a valuable and underutilized tool for understanding the evolution of paleoaridity on the Tibetan Plateau.
The record of pre-collision and ongoing collision of Indo-Asia is preserved in four litho-tectonic zones stretching from south to north across northern India and southern Tibet: the Himalayan thrust belt, the India-Asia suture zone, the Xigaze forearc, and the Gangdese magmatic arc (Fig. 1; for reviews see Hodges, 2000; Yin and Harrison, 2000; Kapp and DeCelles, 2019). Prior to initial Indo-Asia collision ca. 60–55 Ma, Neotethyan oceanic lithosphere was subducted northward beneath the southern margin of Asia. A Neo-Tethyan seaway separated northern India and Asia, bounded to the south by Tethyan marine rocks deposited on the northern passive margin of Greater India (the pre-collisional Indian landmass), and to the north by an accretionary prism/ophiolites and the Xigaze forearc basin along the tectonically active southern margin of Asia. Subduction-generated melts fed the Gangdese magmatic arc north of the Xigaze forearc basin. During subduction, this southern part of Asia, underlain by the Lhasa terrane, underwent considerable shortening, and the region may have been raised to high elevations, informally termed the “Lhasaplano” (Kapp et al., 2007; Kapp and DeCelles, 2019).
The collision began in the early to mid-Paleogene (60–55 Ma), suturing the two continents along the IYSZ in southern Tibet (Ding et al., 2005; Najman et al., 2010; Orme et al., 2015; Hu et al., 2016; DeCelles et al., 2018a). Many of the suture zones in central Asia are marked by valleys, and the IYSZ follows the east-west course of the Yarlung-Tsangpo (Brahmaputra) River along much of its >1000 km reach (Fig. 1). In addition to the Xigaze forearc basin, thick and widespread Cenozoic nonmarine strata are exposed within and on both sides of the IYSZ in southern Tibet (DeCelles et al., 2018a) and contain a wealth of information about the collision history and paleoelevation changes.
Our study involves over 300 new isotopic analysis of carbonates from all four major lithotectonic zones in the region, merged with considerable data collected by previous studies. Our study sites are grouped into (Fig. 1): (1) marine carbonates and paleosols from the mid-Cretaceous-Eocene Xigaze forearc sequence, (2) lacustrine carbonates from the lower Eocene Nianbo Formation of the Linzizong Group in the Gangdese retroarc, (3) paleosol carbonates from the lower Miocene (?) Liuqu Conglomerate, and (4) paleosol carbonates and other secondary cements from the lower to mid-Miocene Kailas Formation. We augmented our isotopic analysis with petrographic examination of a subset of samples from most areas.
The mid-Cretaceous-Eocene Xigaze forearc sequence is 5–8 km thick and is exposed for ~550 km along strike of the central IYSZ zone in southern Tibet (Ding et al., 2005; Wang et al., 2012; An et al., 2014). It consists of turbiditic sediments and channelized conglomerates in the lower portion (Xigaze Group, 110–84 Ma; Einsele et al., 1994) and in the upper part of mixed marine limestone, sandstone, and shale grading up to shallow marine deposits mixed with paleosols (Tso-Jiangding Group, 83–51 Ma; Ding et al., 2005; Orme et al., 2015). The sequence is thought to record the Indo-Asian collision sequence, starting with deep-water sedimentation in the Xigaze forearc, shoaling upward to shallow marine and continental deposits at the top as the former forearc was uplifted above sea-level, and continental Indian lithosphere entered the trench (Orme et al., 2015). Our sampling of marine and paleosol carbonates covers most of this sequence and was obtained from two areas: marine carbonates from the Xigaze Group exposed northeast of Lazi (Fig. 1; Orme and Laskowski, 2016), and the younger Tso-Jianding Group exposed northwest of Saga in the Lopu Kangri area (Fig. 1 and Orme et al., 2015). The detailed sedimentology and thermochronology of these sections along with sample locations are described in Orme et al. (2015), Orme and Laskowski (2016) and tabulated in Table S2. These samples have also been the focus of one study of the isotopic composition of the carbonates (Ingalls, 2019), results that we expand upon here with a larger sample set and consideration of the δ13C results.
The Linzhou Basin is located ~30 km north of Lhasa, within the Gangdese magmatic arc. It contains ~3–4 km of intercalated volcanic and volcanoclastic rocks referred to as the Linzizong sequence. The Linzizong sequence consists, from bottom to top, of the Dianzhong, Nianbo, and Pana Formations and is capped by an additional ~3 km of sedimentary rocks due to thrusting (He et al., 2007). It spans the period ~69–44 Ma.
Our sampling focused on the lower Nianbo Formation, dated to ~52 Ma, near the village of Linzizong and in the section pictured by (Huang et al., 2015) . The Nianbo Formation consists of lacustrine marl and siltstone, volcanic tuffs and ash flows, and volcanoclastic sandstone. We analyzed seventeen samples from at least five lacustrine carbonate and marl horizons in the lower 40 m of the section (Table S3). Igneous dikes cut the Nianbo Formation in the area of sampling, and strong hydrothermal alteration is visible along the dikes and in the underlying Dianzhong Formation. Carbonates from the Nianbo and Pana Formations have been previously studied from the perspective of paleoaltimetry by Ding et al. (2014) and Ingalls et al. (2018). Recently, Huang et al., 2015 undertook a detailed study of the thermochronologic and paleomagnetic record of the Linzizong Formation at Linzhou.
The Liuqu Conglomerate is exposed in a narrow (<5 km) belt along a roughly ~150 km length of the IYSZ. It is sandwiched between two major thrust faults in the IYSZ separating ophiolitic mélanges to the south and uplifted Cretaceous forearc deposits to the north, both of Asian plate affinity. The Luiqu Conglomerate is folded into a syncline that exposes up to several kilometers of redbeds, consisting of fluvial conglomerate, sandstone, and floodplain siltstone and paleosols. The tectono-sedimentary context and age of the Liuqu Conglomerate is the subject of ongoing debate. One view is that the Liuqu Conglomerate was deposited in the Paleogene during a suturing event prior to (Davis et al., 2002, Davis et al., 2004) or early in Indo-Asian collision (Ding et al., 2017). A second view is that the Liuqu Conglomerate was deposited during the Early Miocene (Leary et al., 2017). We favor the latter view but recognize that the age of the Liuqu Conglomerate needs more study. Both studies (Ding et al., 2017; Leary et al., 2017) agree that, whatever the age, the Liuqu Conglomerate was deposited at modest elevations well below (≤ 2 kmasl) the current elevation of ≥4 kmasl.
Our samples for this study of the Liuqu Conglomerate come from paleosols exposed in three separate measured sections (Table S4; 1LQ, 4LQ. And 6LQ) which have been described in detail in Leary et al. (2016). Our work builds upon two isotopic studies already performed the Liuqu Conglomerate by Leary et al. (2017) and Ning et al. (2019).
The Kailas Formation (Gansser, 1964; DeCelles et al., 2018b) is a narrow (<10 km), ~1300 km long belt of mainly clastic continental rocks ≤4 km thick found on the northern side of the IYSZ. Where not faulted, Kailas Formation rests in depositional contact nonconformably upon granitic and/or volcanic rocks of the Gangdese arc. The age of the base of the formation appears to decrease eastwards, ranging from ~25 to 26 Ma in the west near Mt. Kailas (E81.3°) to ~23 Ma at Dazhuka (E89.8°), suggesting that the Kailas basin propagated from west to east at ~0.3 m/yr (Leary et al., 2016). The Kailas Formation grades from alluvial fan conglomerates along the paleobasin margins to fine-grained fluvial and lacustrine deposits along the axis of the paleovalley. Deep lacustrine sandy to conglomeratic turbidite facies are also present in the basin center (DeCelles et al., 2018b). Unlike most other Cenozoic basins in the IYSZ, the Kailas basin also contains abundant organic-rich shale and coal. We analyzed carbonates, mostly in paleosols, from the Kailas Formation at seven locations along strike from 81.6–89.8°E (Table S5) and combine these results with those of our previous work in the Mt. Kailas region (E80.9°; DeCelles et al., 2011).
Section snippets
Stable isotope analysis of carbonates
Samples were slabbed and microdrilled for ~0.5 to 1 mg of powder. Carbonate analyzed for δ18O and δ13C values was first heated at 250 °C for 3 h in vacuo to remove volatile contaminants. Laboratory tests show that the carbon and oxygen isotope ratios of inorganic carbonates remain unchanged after roasting at 250 °C (Wierzbowski, 2007). In organic-rich samples, variance in isotopic measurement is reduced when volatiles associated with organic matter are removed by roasting. Stable isotopic
Results
We analyzed twenty-two water samples for δ18O and δD values (Table S7) and ~ 300 carbonate samples, including marine limestones, lacustrine carbonate, paleosol carbonate, and other forms of secondary carbonate, for δ18O and δ13C values (Tables S2–S5). We also performed twenty-two new Δ47 analyses of lacustrine and soil carbonate (Table S6). In some but not all cases, we also described thin sections of the sampled material to assess the state of preservation of the material.
Alteration of marine carbonates and its implications
As shown in Fig. 2b, our study more than doubles the number of published isotopic values of K-Pg limestones from various locations in the IYSZ. These data strengthen the view that isotopic alteration in the IYSZ limestones is much more prevalent than outside it; and that alteration of oxygen isotopes is more pervasive than for carbon. Ninety-six percent of the limestones analyzed retain primary δ13Cls values, whereas none of them preserve primary δ18Ols values. In short, in the case of oxygen,
Summary observations and future directions
In this paper we stress the important role of δ13C ls and δ18Ols values from marine limestones in assessing local and regional patterns of diagenesis, in order to place some constraints on the likelihood of alteration of non-marine carbonates in overlying stratigraphic sequences. Most δ13Cls values are unaltered, allowing us to use the δ13C value of paleosol carbonates to trace the gradual aridification and devegetation the IYSZ sometime after 20 Ma, probably in response (at least in part) to
Acknowledgments
Support for this work came mainly from the NSF Continental Dynamics grant 0907885 to Paul Kapp and others. We thank Brian Currie and David Rowley for their discussion of Tibet paleoaltimetry over the years. We are greatly indebted to Dr. Ding Lin for his logistical support of the project and many scientific interactions with the Arizona tectonics group. Dr. Peter Scholle provided very helpful observations on the carbonate petrography of some of our samples.
References (103)
- et al.
Isotopic composition and origin of precipitation in northern Chile
Appl. Geochem.
(1999) The stable isotopic composition of modern soil carbonate and its relation to climate
Earth Planet. Sci. Lett.
(1984)- et al.
Paleogene island arc collision-related conglomerates, Yarlung-Tsangpo suture zone, Tibet
Sediment. Geol.
(2002) - et al.
High and dry in central Tibet during the late Oligocene
Ear. Planet. Sci. Lett.
(2007) - et al.
Compositional and temperature effects of phosphoric acid fractionation on Δ47 analysis and implications for discrepant calibrations
Chem. Geol.
(2015) - et al.
Review: Implications of vertebrate fossils for paleo-elevations of the Tibetan Plateau
Glob. Planet. Change
(2019) - et al.
Defining an absolute reference frame for “clumped” isotope studies of CO2
Geochim. Cosmochim. Acta
(2011) - et al.
The Xigaze forearc basin; evolution and facies architecture (Cretaceous, Tibet)
Sediment. Geol.
(1994) - et al.
Predicting paleoelevation of Tibet and the Himalaya from δ18O versus altitude gradients in meteoric water across the Nepal Himalaya
Earth Planet. Sci. Lett.
(2000) - et al.
Cretaceous-Tertiary geology of the Gangdese Arc in the Linzhou area, southern Tibet
Tectonophysics
(2007)