Brahmaputra Sediment Flux Dominated by Highly Localized Rapid Erosion
Total Page:16
File Type:pdf, Size:1020Kb
Brahmaputra sediment fl ux dominated by highly localized rapid erosion from the easternmost Himalaya R.J. Stewart1, B. Hallet*1, P.K. Zeitler2, M.A. Malloy2, C.M. Allen3, D. Trippett4 1Quaternary Research Center and Department of Earth and Space Sciences, University of Washington, Seattle, Washington 98195, USA 2Department of Earth and Environmental Sciences, Lehigh University, Bethlehem, Pennsylvania 18015, USA 3Research School of Earth Sciences, Australian National University, Canberra, 0200 ACT, Australia 4Quaternary Research Center, University of Washington, Seattle, Washington 98195, USA ABSTRACT The Brahmaputra River slices an exceptionally deep canyon through the eastern Hima- laya. Fission-track and laser-ablation U-Pb ages of detrital zircon grains from the river docu- ment very rapid erosion from this region and its impact on sediment fl uxes downstream in the Brahmaputra. Downstream from the canyon, 47% of the detrital zircons in the river’s modern sediment load comprise a fi ssion-track age population averaging only 0.6 Ma. Equally young cooling ages are reported from bedrock in the canyon through the Namche Barwa–Gyala Peri massif but are absent from riverbank sands of major tributaries upstream. Simple mixing models of U-Pb ages on detrital zircons from samples taken above and below this massif inde- pendently suggest that 45% of the downstream detrital zircons are derived from the basement gneisses extensively exposed in the massif. Constraints on the extent of the source area pro- vided by bedrock cooling ages together with sediment-fl ux estimates at Pasighat, India, sug- gest exhumation rates averaging 7–21 mm yr–1 in an area of ~3300 km2 centered on the massif. This rapid exhumation, which is consistent with the very young cooling ages of the detrital zircons from this area, produces so much sediment that ~50% of the vast accumulation in the Brahmaputra system at the front of the Himalaya comes from only ~2% of its drainage. This extreme localization of rapid erosion, sediment evacuation, and bedrock cooling bear on (1) common assumptions in geodynamic and geochemical studies of the Himalaya about sources of sediment, and (2) plans for hydroelectric development and fl ood management in southeastern Tibet and the heavily populated areas of eastern India. Keywords: Himalaya, Yarlung-Tsangpo River, Siang River, Brahmaputra River, fi ssion-track dat- ing, U-Pb ICP-MS dating, detrital-mineral thermochronology. INTRODUCTION while crossing an active antiform developing in the canyon, and at Pasighat, India (sample 301; We report fi ssion-track and U-Pb ages of Protero zoic rocks that were deformed and meta- Fig. DR1; Table DR1), 180 km downstream detrital zircon grains in sediments of the Brahma- morphosed in the Pleisto cene (Burg et al., 1997, from the canyon, where the Brahmaputra putra River that refl ect both the areal extent and 1998; Ding et al., 2001; Zeitler et al., 2001) emerges from the Himalayan foothills on its the tempo of erosion for an erosional “hot spot” in (Fig. 1). The antiform has been exhuming at rates way to the Bay of Bengal (Fig. DR1). The this drainage basin, which is among the top three of 3–5 mm yr–1 over the past 5–10 m.y. as shown Pasighat site is of particular interest because sediment producers on the planet (Summerfi eld by petrological data and U-Pb dating of acces- the sediment fl ux has been measured there, and and Hulton, 1994). Herein, we use “Brahma- sory phases (Booth et al., 2004, 2008), and even we will examine the corresponding erosion putra” to refer liberally to the main stem of this higher rates up to ~10 mm yr–1 have been esti- rates in the source area. large river system where it traverses the eastern- mated for more recent intervals (Burg et al., 1997, Details of the samples collected, the meth- most Himalaya. We focus on the “Big Bend” 1998). Together with the ubiquity of steep slopes, ods and standard assumptions used, and the region, where the river changes direction 180° landsliding (Bunn et al., 2004), and high relief data reported in this paper are given in the after running east for 1300 km in southern Tibet. (Finnegan et al., 2008), this long history of rapid GSA Data Repository. For fi ssion-track dat- Here, it cuts one of the world’s deepest canyons erosion suggests that suffi cient fl uvial incision ing, data for 296 grains (Fig. 2; Table DR1; (>5000 m) between the peaks of Namche Barwa has occurred to bring most portions of the canyon GSA Data Repository) were analyzed using (7782 m) and Gyala Peri (7151 m) (Fig. 1; landscape close to steady state, with long-term the BINOMFIT peak-fi tting routine (Ehlers Fig. DR1 in the GSA Data Repository1) and erosion everywhere occurring at the same rate. et al., 2005; Stewart and Brandon, 2004) plummets 2000 m over a spectacular knickpoint This simplifi es our interpretation because detrital which searches for the best-fi t set of signifi - zircons are likely to originate from throughout cant components in the grain-age distribution. 1GSA Data Repository item 2008179, detrital the landscape, and their age distribution will not U-Pb analyses of 312 detrital zircons from zircons from the easternmost Himalaya, is available be signifi cantly impacted by erosional transients. samples 301 (downstream of the canyon, at online at www.geosociety.org/pubs/ft2008.htm, or Pasighat) and 302 (upstream, near Pai) were on request from [email protected] or Docu- ments Secretary, GSA, P.O. Box 9140, Boulder, CO METHODS conducted at the Research School of Earth 80301, USA. We separated zircon grains from fl uvial sands Sciences, Australian National University, collected from the banks of the Brahmaputra using laser-ablation ICP-MS (Tables DR2, *E-mail: [email protected]. and its tributaries at seven sites upstream from DR4a and DR4b). © 2008 The Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or [email protected]. GEOLOGY,Geology, September September 2008; 2008 v. 36; no. 9; p. 711–714; doi: 10.1130/G24890A.1; 2 fi gures; Data Repository item 2008179. 711 6 36 Tethyan metasediments 95°E 11 Tungmai 5 Lhasa Block rocks 11 0.4 N 4 Basement gneiss 308 15 1.7 13 Fault 1.8 1.5 1.3 0.7 3 12 20 km 1.4 1.6 14 303 0.8 0.6 304 309 30°N 30°N 1500 m 2.8 Parlung River 2 Gyala Peri 8 2.0 7151 m 16 310 2.5 1.3 1 1.0 2.3 9 12 14 1.7 16 3 0.5 14 0.3 0.7 0.6 10 0 1.8 1.7 Biotite 5 40Ar/39Ar 305 0.1 1 10 100 Nyang River Knickpoint <2 Ma 8 Bayi 2500 m 25 6 2900 m 2.5 Namche 0.6 5 18 18 2.5 Barwa 2.1 7782 m 6 Pai 1.3 500 m 1.2 2.7 15 9 10 5 5 4 2.9 2.4 7 36 4 302 4 5 5 Tsangpo gorge Zircon 12 302 er (U-Th)/He Detrital sand sample 2 <2 Ma 21 10 6 Zircon (U-Th)/He sample 0 Medog 11 40 39 Brahmaputra Riv 95°E Biotite Ar/ Ar sample 0.1 1 10 100 Probability density (% per delta z = 0.1) 10 Figure 1. Geologic sketch map of Namche Barwa–Gyala Peri massif and the Brahmaputra canyon (Pan et al., 2004, with modifi cations courtesy of W.S.F. Kidd). Black squares are 8 locations of detrital sand samples. Circles represent bedrock zircon (U-Th)/He cooling ages; diamonds represent bedrock 40Ar/39Ar biotite cooling ages (Malloy, 2004). Green line is con- 6 tour for zircon (U-Th)/He ages younger than 2 Ma, and red line is contour for 40Ar/39Ar biotite ages younger than 2 Ma. 4 2 RESULTS FROM DETRITAL ZIRCON 2.2 Ma. The absence of young grains entering 0 FISSION-TRACK DATING the canyon, coupled with the absence of major 0.1 1 10 100 The fi ssion-track grain-age distribution tributaries to the Brahmaputra between the Fission-track grain age (Ma) of 101 zircon crystals from the composite canyon and Pasighat, suggests that the canyon Pasighat sample (Fig. DR1) contains fi ve com- region through the Namche Barwa–Gyala Peri Figure 2. Grain frequencies, probability den- sity distributions (PDD), and best-fi t peaks ponents with “peaks” at 0.6 ± 0.1, 4.7 ± 1, massif is the source of the very young grains (Ehlers et al., 2005) for detrital zircon fi ssion- 10 ± 2, 18 ± 3, and 37 ± 15 Ma (Tables DR1and observed at Pasighat (Fig. DR1). The fraction track grain-age distributions for three river- DR3; Fig. 2A). The 0.6 Ma peak consists of of grains in the 0.6 Ma peak at Pasighat also bank samples. A: From the Brahmaputra 48 individual grain ages ranging between 0.1 suggests that ~50% of the suspended sediment at Pasighat, India, well downstream of the canyon (Fig. DR1). B: From the Brahmaputra and ca. 2.2 Ma, and it is robust, remarkably in the Brahmaputra downstream of the canyon upstream of the canyon, at the confl uence young, and large, as it comprises 47% of the is derived from source areas having similarly with the Nyang River (Fig. 1). C: Composite of total population. For comparison, the 0.6 Ma young bedrock zircon fi ssion-track ages. This samples 303, 304, 305, 308, 309, and 310, from zircon population at Pasighat is three times result is valid, making the reasonable assump- the Parlung River and its tributaries before it younger than the youngest coherent popula- tion (see GSA Data Repository) that the gneisses enters Brahmaputra canyon (Fig.