Fault activity controls the formation of knickpoint
In a previous study27, Schmidt et al proposed that the Gyaca knickpoint resulted from upstream migration of erosional waves starting from the Yarlung Tsangpo Grand Canyon (Fig. 1b) in response to uplift of the southeastern Tibet prior to ∼10 Ma27. We confirmed that only headward incision of the Yarlung River since ~ 12 Ma due to uplift of the southeastern Tibetan Plateau27 or intensified monsoon precipitation35–36 cannot fit the remarkable young AHe and AFT ages in the Gyaca gorge (Fig. 3e). In addition, the rock outcrops in the Gyaca gorge and its downstream valley are mainly granitoids (Fig. 1c), thus lithological variation is not a major controlling factor for the formation of the Gyaca gorge. In contrast, a high-angle normal fault is required as indicated by the well-constrained Pecube model (Scenario B), to fit with an increased rock cooling rate at the Gyaca gorge since ~ 7 Ma (Fig. 3g). Therefore, we suggest that activity of the Woca fault has controlled the formation of the Gyaca knickpoint since ~ 7 Ma, and that a similar mechanism could also explain the formation of other knickpoints in the middle and upper reaches of the Yarlung River (Fig. 1b).
Mechanisms For Impeding River Incision In The Late Miocene
Based on QTQt and Pecube modeling results, a relatively high incision rate (~ 1.73 ± 0.07 km/Ma) of the eastern Yarlung River initiated at ∼12 Ma, with a slightly decreasing trend in thermochronological ages from east to west along the Yarlung River. Published thermochronometric data from the externally drained portion of the eastern10,37−38 and central39–40 Lhasa terrane, as well as from large rivers in the southeastern Tibetan Plateau36,41, demonstrate that rapid exhumation rates (> 1 km/Ma) were pervasive across the southern and southeastern plateau between ∼17 and 10 Ma. These spatially large-scale synchronous rapid incision events most likely reflect enhanced Asian summer monsoon precipitation in the mid-Miocene35–36 that promoted the headward erosion of the Yarlung River channels (Fig. 4a). However, the incision rate upstream and downstream of the Gyaca gorge has decreased dramatically with the rapid activity of the Woca fault since 7 Ma (Fig. 2g and Supplementary Table 6).
Thermochronometric data from rifts in the Tibetan Plateau42–43 suggest that southern Tibet has experienced rapid late Miocene to Pliocene rift acceleration28 (see Fig. S7). This rapid rift activity controls knickpoints, such as the Gyaca knickpoint, in the plateau interior. Further, the high rate of rift extension in south Tibet also facilitates thinning of the upper crust while its lower crust is thickened by ongoing compression28,44−45. This could be a main driving force accelerating eastward crustal flow46. The accelerated crustal flow at the Eastern Himalayan Syntaxis (EHS) drives the localized deformation and uplift47–49, leading to the active coupling between crustal rock advection and river erosion in the EHS50 since ~ 7 Ma (Fig. 4b). A stationary knickpoint in the EHS develops initially. In fact, Zeitler et al. suggested a sustained high-elevation base level for the Yarlung River in the eastern syntaxis since ∼10 Ma15. Although Wang et al. argued for a rapid uplift of the EHS at ∼2.5 Ma based on sediment fill immediately upstream of the Yarlung Zangbo gorge11, we consider this as only one of several stages of EHS uplift, rather than the initial uplift episode. Furthermore, detrital thermochronological data from foreland basin sediments downstream of the EHS suggested an efficient coupling between tectonics uplift and erosion starting at 8 Ma50, 7 − 5 Ma51 and/or 6 − 4 Ma52.
A current paradigm for the initiation of rapid exhumation in the EHS is the Tectonic Aneurysm model, in which spatially focused surface erosion driven by the Yarlung River locally accelerates rock uplift and exhumation of hot and weak crust at the syntaxes15–16, 53. Evidence from sediments in the Himalaya foreland19, Bengal basins20 and our thermochronometric data (Fig. 2) show that the Yarlung river was definitely set in course before ~ 12 Ma. This does not specifically address the role of the Yarlung River in driving initial rapid exhumation of the EHS, which is thought to have begun 8 − 6 Ma50–52. In contrast, we find that a synchronous exhumation pulse in the rifts (Fig. S7) and EHS50 since the late Miocene. This synchronous rapid exhumation implies that the tectonic system with the accelerated late Miocene extension of southern Tibet drives regional fault activity to control rapid exhumation and formation of knickpoints in the rifts and EHS.
With the stabilization of the knickpoints, river gradients above the knickpoints are commonly so low that fluvial incision may generally have difficulty in keeping up with increase in tectonic uplift rates. As a result, channel slope and stream power would decrease transiently above these Yarlung River knickpoints (Fig. 4b). Simultaneously, the reduced river gradient and sediment flux could promote upstream aggradation, burial of bedrock valley floors13,54−55, and valley widening (Fig. 4b), which could have caused widespread backwater aggradation, forming broad valley trains occupied today by braided river systems (Figs. 1d and 4b). This braided river only incised when the stripping of deposited alluvial material exposes bedrock to processes of abrasion, weathering and plucking55–56.
Hence, coeval tectonic systems (rifting upstream and uplift and exhumation of the EHS downstream) with the accelerated extension of southern Tibet drove the onset of knickpoints formation in the Gyaca and EHS since the late Miocene. These were subsequently adjusted and reduced the regional river gradient, which retarded the headward incision of the knickpoints, thereby facilitating stabilization of southern Tibetan Plateau topography since the late Miocene.
Actually, many other active orogenic belts grow first to a certain height, and then experienced laterally outward propagation57. This expanding uplift generally increases the steepness of rivers, and promotes upstream erosion24–26. But this expansion is not a simple one-dimensional process, it is also accompanied by strong regional tectonic deformation of the upper crust58–59 (e.g., rifting, subduction, and strike-slip faulting) and lower crustal flow46. Lower crustal flow in its turn will promote and further drive regional tectonic deformation of the upper crust. Thus, we propose that the activity of the local fault systems limited the migration of these river knickpoints as that on the eastern Yarlung River, which are still located in the location of regional tectonic deformation (Fig. 5). This process subsequently reduced fluvial incision in the upstream of these knickpoints, stabilized the topography of the Plateau and protected high-elevation, low-relief surfaces over geologic time in active orogenic belts.
Our finding suggests that tectonic activity does not always necessarily enhance fluvial incision, and that the diverse activity can also decrease these rates by the adjustment of fault systems in active orogenic belts. This discovery also provides a new mechanism for explaining the universal stability of long-term topography and high plateau of orogenic belts worldwide.