Structure of the deep lithosphere between Pamir and Tarim

The Pamir protrudes ~300-km between the Tajik and Tarim lithosphere of central Asia. It overlies a Wadati-Benioff earthquake zone connected to a low velocity zone interpreted as crustal rocks. Together with the mantle lithosphere it constitutes the arc-shaped Asian slab. We use new seismic data to better constrain the lithospheric architecture of the Pamir where it abuts the Tarim block and test competing models of its formation. With complemented local-seismicity and focal-mechanism catalogs and a P-wave velocity model that spans the Pamir and the western Tarim lithosphere, we infer the presence of a high velocity zone, interpreted as an Indian mantle lithosphere indenter, delaminating the Asian slab and overturning it in the eastern Pamir. The indenter bends down in the east under the northwestern Kunlun, where it terminates. The indenter–Tarim lithosphere interface is a compressive transform zone lined by a slice of Pamir Plateau crust. As the largest principal stress at depth parallels surface motion and both are highly oblique to the western Tarim margin, this crustal slice is likely dragged with the indenter and downward underneath the Tarim lithosphere. INTRODUCTION The Pamir—the northwestern prolongation of the Tibetan plateau—is bordered by the Tian Shan, the Tajik basin, and the Tarim basin, in the north, west, and east, respectively; the latter is a cratonic block (e.g., Lu et al., 2008). The northern Pamir and the Kunlun of northwestern Tibet comprise subduction-accretion-arc complexes accreted to and built on Asian continental basement. The central and southern Pamir and the Karakorum and Hindu Kush represent Gondwana-derived microcontinents and subduction-accretion-arc complexes (Fig. 1; Burtman and Molnar, 1993; Schwab et al., 2004). Beneath the Pamir, a band of intermediate24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40


INTRODUCTION
The Pamir-the northwestern prolongation of the Tibetan plateau-is bordered by the Tian Shan, the Tajik basin, and the Tarim basin, in the north, west, and east, respectively; the latter is a cratonic block (e.g., Lu et al., 2008). The northern Pamir and the Kunlun of northwestern Tibet comprise subduction-accretion-arc complexes accreted to and built on Asian continental basement. The central and southern Pamir and the Karakorum and Hindu Kush represent Gondwana-derived microcontinents and subduction-accretion-arc complexes ( Fig. 1; Burtman and Molnar, 1993;Schwab et al., 2004). Beneath the Pamir, a band of intermediate-  Pegler and Das, 1998;Sippl et al., 2013a). Receiver function images, seismic tomography, and the analysis of guided waves show that the earthquakes in the western and central Pamir reside in a 10-15 km thick, E-to S-dipping low velocity zone (LVZ) connected to the Asian lithosphere; seismic velocities indicate that the LVZ represents continental crust, constituting together with the underlying mantle lithosphere the Asian slab (Schneider et al., 2013;Sippl et al., 2013b;Mechie et al., 2019). Beneath the northwestern Kunlun, diffuse seismicity at 100-150 km depth was attributed to Tarim lithosphere underthrusting the Pamir (Fan et al. 1994;Pegler and Das, 1998).
Whether Asian lithosphere subducts as a narrow, back-rolling slab of thinned crust (Burtman and Molnar, 1993;Sobel et al., 2013) or whether forceful subduction/delamination of lower crust and mantle lithosphere due to indentation by cratonic Indian mantle lithosphere (Kufner et al., 2016;Metzger et al., 2017) occurs is debated. This debate impacts on the understanding of processes like continental subduction, indentation, delamination, and rollback.
If an indenter governs the shape of the Pamir orocline, its margins matter. Kufner et al. (2018) argued that a sinistral-oblique transform margin separates indenting cratonic Indian lithosphere beneath the Pamir from subducting Indian continental-margin lithosphere below the Hindu Kush. The most recent subduction model (Sobel et al., 2013) calls for rollback of a narrow Asian slab with thinned continental crust, involving mantle corner flow and a subductiontransform edge propagator fault separating the subducting Asian slab and its hanging wall from the Tarim block. However, geophysical data indicate that the hinterland crust is not thinned (>50 km; Schneider et al., 2019). The delamination model (Kufner et al., 2016;Chapman et al., 2018)  calls for forced Asian slab subduction due to flat-slab underthrusting of a mechanically-strong Indian continental lithospheric mantle indenter, a process recently modeled for the Pamir (Kelly and Beaumont, 2021). The indenter is imaged by refraction seismics and local body wave tomography as a high velocity zone (HVZ) south of the Asian slab (Mechie et al., 2012;Sippl et al., 2013b). Teleseismic body and surface wave tomography shows that it connects with the exposed Indian craton (e.g., Li et al., 2008;Liang et al., 2020); its northern extent has remained unresolved due to the smearing of the HVZ with cratonic Asia.
Herein, intermediate-depth earthquakes, focal-mechanism stress data, and a P-wave velocity model derived from new and published local seismological data illuminate the lithospheric configuration of the central and eastern Pamir and the boundary zone with Tarim.
We characterize the northern tip of an indenter-interpreted as a promontory of Indian mantle lithosphere-and its eastern edge, where it abuts on the lithosphere of the Tarim block. We use seismograms recorded with two new local seismic networks (Yuan et al., 2018a;2018b) and additional regional stations (PMP International, 2005;SEISDMC, 2021) to locate seismicity in the eastern Pamir jointly with an existing catalog from the western and central Pamir (Sippl et al., 2013b), and invert for the 3-dimensional subsurface P-wave velocity structure. The full description of the data and methods ( Fig. S1; Fig. S2), the seismicity catalog, and the velocity structure are presented in the Supplemental Material. Seismicity in segment 2 is less intense compared to segment 1 (Fig. S2). Focal mechanisms of segments 1 and 2 indicate a transpressional stress regime, with the maximum principal stress σ 1 trending N20°W and N12°W, parallel to the surface plate-motion directions, and a vertical σ 3 (Fig. 2). Seismicity in segment 3 forms a continuous, ~ENE-dipping structure at 80-120-km-depth between 37°N and 38°N; it follows the northwestern Kunlun ( Fig. 2; Fig. 3C). Seismic activity is comparably weak (Fig. S2). Focal mechanisms indicate transpression with σ 1 trending N7°W, parallel to the surface plate-motion, and a down-dip σ 3 (Fig. 2).

VELOCITY STRUCTURE
In the shallow crust, northeast of the Main Pamir Thrust System (Fig. 3, overview map), the sediment fill of the Tarim basin forms a LVZ (<5 km/s, TL in Figs. 3B-D). In the middlelower crust, the Tarim basement appears as a discontinuous HVZ (6.5-7.5 km/s, TH in  Fig. 3G). An arcuate crustal LVZ extends below the northern Pamir, the Kongur Extensional System, and the northwestern Kunlun (5-6 km/s, PL in Figs. 3A-C and 3E). It is sandwiched between the Tarim basement HVZ, TH, and another crustal HVZ in the central Pamir (6-7 km/s, PH in Fig. 3A; Fig. 3E).
The seismically active structures are underlain by HVZs (8.5-9.5 km/s, H1, H2, H3 in Figs. 3A-C and 3G) and have the same dip as the LVZs above. In segment 1 and 2, the HVZs H1 and H2 are continuous along strike below ~105 km depth (Fig. 3G). In segment 2, the HVZs H2 and H3 touch, but are separated by seismicity in the same way as the LVZs L2 and L3 ( Fig. 3B; Fig. 3G).

INTERPRETATION AND DISCUSSION
We visualize our interpretation of the lithospheric architecture of the central and eastern zones in the middle crust (Fig. 1, cross section), may represent a heated rock volume, developed by excess radiogenic heat production in the thickened crust. Heating due to asthenospheric inflow in the hanging wall of a S-dipping subduction zone is unlikely, as the tomogram does not show a LVZ south of the seismic zone; in contrast, subcrustal P velocities are >8km/s with HVZs (>8.5 km/s) embedded (e.g., H3), indicating relatively cold and rigid lithospheric mantle south of the Asian slab.
Segment 2 appears to be the eastern continuation of segment 1 of the Asian slab because of the similar depth extent of the seismic zone and the continuity of the underlying HVZ ( Fig. 2; Figs. 3A, 3B, and 3G). The seismically active structure is overturned below ~80 km depth. A tear likely separates segments 1 and 2 because of the short (~40 km) distance across which the slab dip changes and the separating seismicity gap. The Asian slab terminates in a seismicity cluster below the Kashgar-Yecheng Transfer System at 76.2°E (Fig. 2), where it appears torn off Tarim's lithosphere to the east. The dip beyond vertical of segment 2 may be an indication of indentation, because rollback alone cannot steepen the slab to more than vertical (Magni et al., 2013).
For segments 1-3, σ 1 at depth is parallel to the ~NNW-oriented surface velocity of the Pamir crust (e.g., Zubovich et al. 2010, Ischuk et al., 2013, Metzger et al., 2020. The subhorizontal σ 1 indicates that a NNW-SSE compressive stress field governs the deep structure of the Pamir, which favors a pushing indenter. In contrast, N-S extension should occur S of the slab if deformation below the Pamir was governed by a narrow Asian slab rolling back northward. Parallelism of the surface motion with σ 1 at depth implies that the lithospheric mantle is coupled to the crust. For segments 1 and 2 it arises if collision occurs at an indenter tip. In concert with the lack of thinned hinterland crust and the imaging of a HVZ at ~200 km depth below the Pamir-Karakorum that connects with the exposed Indian craton, the following of  (Mechie et al., 2012;Sippl et al., 2013b; this study) that excludes asthenospheric inflow above a S-dipping, back-rolling subduction zone; (2) the overturned dip of the seismic plane of segment 2; (3) the NNW-SSE compressive stress field across the central and eastern Pamir at mantle depth that is coupled with surface motion.
The indenter is most likely cratonic Indian lithosphere, because the lithosphere of the central and southern Pamir terranes would be too weak to transmit enough force to delaminate and overturn the Asian slab (Kelly and Beaumont, 2021). We locate the delamination front at the base of the rheological weak mid-crustal LVZ PL (red line in Fig. 4). The present location and form of the Pamir and the Asian slab is in this interpretation governed by the shape of the indenter. Additional structural complexity, such as the location of slab tears or turn-overs, may be due to lateral changes in the strength of the indented Asian lithosphere or the along-strike variability of the indenter tip (Li et al., 2016;Kelly and Beaumont, 2021). For example, the midcrustal HVZ PH, which overlies a distinctive Moho bulge in segment 1 ( Fig. 3A; , may represent a lithosphere-scale anticline; in segment 1, the top of the indenter appears to rise higher than in segment 2 and in particular in segment 3 (Fig. 4).
In the northwestern Kunlun, the seismicity band of segment 3, the LVZ L3, and the HVZ H3 dip ~ENE, indicating that Pamir crust and indenter mantle lithosphere underthrusts the Asian mantle lithosphere (Fig. 3C). The earthquakes may occur in thickened crust undergoing eclogitization. This crust is dragged to depth between the bulldozing indenter and the margin of the Tarim block. The underthrusting interpretation is supported by a doubled ~E-dipping Moho (Xu et al., 2021). The stress field of the earthquakes inside the underthrusting crust L3 indicates that it moves with the NNW-ward moving indenter. The orientation of σ 1 in segment 3 testifies 184 that underthrusting is highly oblique with respect to Tarim hanging wall. As the tomographic and receiver function Moho both dip ~WSW beneath the northwestern Kunlun east of LVZ L3 ( Fig.   3C; Xu et al., 2021), we infer that Tarim underthrusts the northwestern Kunlun as well, building a stack of Pamir and Kunlun-Tarim crust (Fig. 4C). This excess crust may be responsible for a positive anomaly in the isostatic gravity residual (Balmino et al., 2012;20-mGal-contour in Fig. 2) that flanks the northern edge of the Tibet plateau (Fig. 2, inset), and was interpreted to represent thrusting of Tarim crust under the western and central Kunlun (Wittlinger et al., 2004).
The transpressive stress field of the deep seismic zone (segment 3) outlines a compressive lithospheric transform zone as the deep plate boundary between the Indian indenter and the Tarim block. It changes to a forced subduction/delamination boundary due to indentation under the central Pamir. The tear that separates the Asian slab from Tarim propagates northward with the advancing indenter. Indentation may have caused the capture and dragging along of the crust from the collision system into the transform zone (Fig. 4C). The transform margin likely transitions southeastward into a subduction plate boundary where the Tarim block underthrusts the western Tibet plateau. Our interpretation of the deep structure suggests a strong along-strike segmentation of the northern tip of the Indian plate; it subducts under the Hindu Kush (Kufner et al., 2021), indents in the Pamir (this study), and has variable dip angles and locations in the rest of Tibet (e.g., Zhao et al., 2011).

ACKNOWLEDGMENTS
We thank the drivers and field participants from the Institute of Tibetan Plateau Research, especially Hongbing Liu, who helped to organize the station deployment, and Christian Sippl for sharing code and discussion. Funded by the CaTeNA project of the German Federal Ministry of Science and Education (support codes 03G0878A and 03G0878B). Seismic data was handled using obspy (Krischer et al., 2015). Part of the instruments were provided by GIPP of GFZ Potsdam. Seismic data are archived by GEOFON data center.   (Zubovich et al., 2010;Ischuk et al., 2013), and 20mGal positive isostatic gravity anomaly (Balmino et al., 2012). TJS, Tanymas-Jinsha suture.