Strain Localized Deformation Variation of a Small-Scale Ductile Shear Zone

A continental-scale strike-slip shear zone frequently presents a long-lasting deformation and physical expression of strain localization in a middle to lower crustal level. However, the deformation evolution of strain localization at a small-scale remains unclear. This study investigated <10 cm wide shear zones developing in undeformed granodiorites exposed at the boundary of the continental-scale Gaoligong strike-slip shear zone. The small-scale ductile shear zones exhibit a typical transition from protomylonite, mylonite to extremely deformed ultramylonite, and decreasing mineral size from coarse-grained aggregates to extremely fine-grained mixed phases. Shearing sense indicators such as hornblende and feldspar porphyroclasts in the shear zone are the more significantly low-strain zone of mylonite. The microstructure and EBSD results revealed that the small-scale shear zone experienced ductile deformation under medium-high temperature conditions. Quartz aggregates suggested a consistent temperature with an irregular feature, exhibiting a dominated high-temperature prism  slip system. Additionally, coarse-grained aggregates in the mylonite of the shear zone were deformed predominantly by dislocation creep, while ultra-plastic flow by viscous grain boundary sliding was an essential deformation process in the extremely fine-grained (∼50 µ m) mixed-phases in the ultramylonite. Microstructural-derived strain rates calculated from quartz paleopiezometry were on the order of 10−15 to 10−13 s−1 from low-strain mylonite to high strained ultramylonite. The localization and strain rate-limited process was fluid-assisted precipitation presenting transitions of compositions as hydrous retrogression of hornblende to mica during increasing deformation and exhumation. Furthermore, the potential occurrence of the small-scale shear zone was initiated at a middle-deep crust seated crustal condition dominated by the temperature-controlled formation and rheological weakening.

Experiments and models on the deformation of rocks have ward to join the Sagaing fault zone. It serves as the tectonic boundary between the Tengchong and Baoshan blocks (Fig. 1b;Tang et al., 2020;Dong et al., 2019;Liu et al., 2017;Zhang et al., 2012a, b;Ji et al., 2000a). The Cambrian gneiss and the Neoproterozoic metamorphic units, named the Gaoligong metamorphic complex, represent the basement units in this area and evolve into the Gaoligong strike-slip shear zone along after the reactivation in Cenozoic ( Fig. 1b; Dong et al., 2019;Zhu et al., 2017;Zhang et al., 2012b;Wang W et al., 2008;Wang Y J et al., 2006). The main rock types are mylonitic gneisses (granitic gneisses and migmatitic gneisses) and schists, as well as amphibolites and marbles. Along the GLG-SZ, a considerable number of Mesozoic and Cenozoic granitic rocks intrude into the Gaoligong metamorphic complex (Dong et al., 2019;Zhang et al., 2018;Wang et al., 2006) (Fig. 1b). Recent zircon U-Pb and 39 Ar/ 40 Ar chronological data revealed that both of the unfoliated and foliated granitic intrusions in the northwest part of the GLG-SZ has the emplaced ages of 112-125 Ma (Early Cretaceous) during the collision of the Lhasa and the Qiangtang blocks, post-magmatic melting timing of ca. 35 Ma (Early Oligocene), and subsequent cooling during the Middle Miocene (ca. 13 Ma) (Dong et al., 2021;Zhu et al., 2017;Xu et al., 2012). Two-stage tectonothermal evolutions since the Late Cretaceous have also been proposed. Around 76 -74 Ma, earlier regional metamorphism occurs in the high-pressure granulite facies owing to crustal thickening and magmatism. Around 24-23 Ma, the later stage was defined by amphibolite-greenschist facies conditions in connection with shearing deformation (Song et al., 2010;Ji et al., 2000b). The analysis of geochronological data from the Tengchong area suggests that the dome uplift and deep crustal material were exhumed during 32 -10 Ma in the south of the GLG-SZ (Dong et al., 2019;Zhang et al., 2017;Xu et al., 2015).
This study emphasizes the small-scale shear zones newly observed within the unfoliated granidiorite at the western part of the . The magmatic fabric of granidiorite is deficient in solid-state ductile deformation features and presents randomly arranged feldspar phenocrysts (Fig. 2). The small-scale shear zones have a thickness of approximately 10 -3 -10 -1 m. Most structures are steeply dipping, and the orientations are relatively disorderly (Fig. 1c).

Microscopy and Cathodoluminescence
Microstructure and petrology in the small-scale shear zone and granidiorite rock were investigated in thin sections by optical and SEM, cathodoluminescence (CL) imaging, and electron backscatter diffraction (EBSD). A Sigma 300VP field emission scanning electron microscope (FEG-SEM) and BII CLF-2 Cathodoluminescence (CL) are employed in the China University of Geosciences (Wuhan). CL operated at a voltage of 15 KV, power consumption of 150 W, a current of 300 A, and a beam current of 1 mA with a diameter of 30 μm.

Electron Backscatter Diffraction (EBSD)
The Sigma 300VP FEG-SEM with a Symmetry EBSD detector in China University of Geosciences (Wuhan) was applied to obtain the mineral CPO. The highly polished thin sections with conductive tape attached to the surface were put in the SEM chamber and rotated at a 70° tilt angle. Electron backscatter patterns were acquired using the automatic mapping mode under the conditions of low vacuum, with a detector distance of 193.1 mm, an acceleration voltage of 20 kV, and a beam working distance of 15.6 mm. Indexing is considered acceptable when at least six detected kikuchi bands correspond to those in the analyzed mineral phases' standard reflector file. Following the completion of the test, the electron backscatter pattern analysis was performed using the Aztec Crystal and HKL Channel 5. The pole figure of representative CPO in samples was plotted in equal-area stereographic diagrams using the lower hemisphere projection and the base circle represents the X-Z plane parallel to the lineation and vertical to the foliation. Automated orientation maps revealed systematic not-indexing, and such data were replaced with zero solution pixels.

EPMA Methodology
Compositional data of unfoliated granitoids and foliated granitic rocks were measured on a JEOL electron microprobe (JXA-8600) with a wavelength dispersive system at the Department of Geography and Geology, University of Salzburg. Measuring conditions using a focused electron beam involved a 15 kV acceleration voltage and a 40 nA sample current. The calibration of the microprobe was performed based on natural silicates and synthetic oxides standards. The matrix correction for quantitative analysis was conducted by the ZAF oxide method for most silicate minerals. The detection limits (2σ) are 0.06 wt.% and 0.04 wt.% for Si and Al, respectively, and are 0.025 wt.% for Na, K, Mg, Mn, and Fe.

DEFORMED CHARACTERISTICS OF GRANODIO-RITES AND SMALL-SCALE SHEAR ZONES 3.1 Mesoscale Structures of Granodiorites and Small-Scale Shear Zones
As mentioned above, the GLG-SZ exposed widespread granitic intrusions of various ages (Fig. 1b;Dong et al., 2021;Tang et al., 2020;Chiu et al., 2018;Zhang J Y et al., 2018;Zhang B et al., 2017;Zhu et al., 2017). Most granitic intrusions underwent strong mylonitization within the GLG-SZ. Notably, the unfoliated granodiorites were exposed at the western boundary of the GLG-SZ. The major body of the studied grano-diorites exhibits little macroscopic evidence of solid-state deformation-metamorphism, and igneous relationships are well preserved .
The granodiorites present strain localization on a network of different types of small-scale shear zones with a thickness of approximately 10 -3 -10 -1 m. They are invariably localized on approximately planar structural and compositional heterogeneities within the protolith (Figs. 1b,. The small-scale shear zones exhibit significant ductile and/or brittle deformation characteristics. Regarding centimeter-scale or decimeterscale shear zones, the strain is strongly localized in a narrow band, and minerals are elongated directionally in the outcrop scale (Figs. 2c,2d). Most structural and compositional heterogeneities demonstrate a nearly horizontal stretching lineation and extremely fine-grained minerals of ductile shearing. The enclaves occur in the granodiorites crosscut by isolated, knifesharp (<1-3 mm wide) brittle fractures, and may have a strike length of many tens of meters (Fig. 2g). The brittle fractures are typically identified by a dark biotite-rich slit. Most of these reflect displacement discontinuities in the outcrop scale. For example, cross-cut markers (e.g., mafic enclaves) are severely truncated and displaced by the shear zones without any dragging effect (Figs. 2g,2h). Some display bands and several centimeters wide of a sigmoidal-shaped foliation of ductile shearing, implying a dextral sense of shear (Figs. 2f, 2h).

Microstructures of Unfoliated Granodiorite
The unfoliated granodioritic host rocks are composed of quartz (16%-19%), K-feldspar (17%-19%), plagioclase (~59%-63%), hornblende (~7%), and biotite (2%-3%). The unfoliated granodioritic exhibits little evidence of ductile deformation. Feldspar and hornblende grains both primarily consist of euhedral to subhedral coarse grains and form with microfracture ( Fig.  3a). The coarse-grained feldspar grains are dominated by plagioclase and tiny amounts of K-feldspar. The plagioclase (up to several millimeters in length) has significant polysynthetic twinning and ring or zoning magma structure that grows fine-grained inclusions of biotite and quartz grains (Fig. 3a). The crystal sizes of hornblende are about 0.5-5 mm, and the grains develop two groups of cleavages (Fig. 3b). Quartz grains present polycrystal aggregates, which are mostly xenomorphic around the plagioclase grains. The biotite grains demonstrate the features of an undeformed or only slightly bent shape with magmatic phase (Fig. 3c).

Microstructures of the Small-Scale Shear Zone
Under the microscope, the small-scale shear zone developing in the granodiorite has dramatical shearing banding, reflecting a transition deformation characteristic from protolith to ultramylonite. The mineral grain size gradually decreases from rim to center, with a strong strain gradient. According to the different grains size, three distinct deformed microstructure zones can be recognized (Fig. 3): Zone A with relatively close to the outer/rim low strain portions (Fig. 3d), high fine-grained Zone B of traversing into the shear zones, and strong finegrained Zone C in the center of the shear zone.
Zone A in the small-scale shear zone is of the rim outer relative low strain portions, composed of plagioclase, K-feldspar, biotite, quartz, and a small amount of apatite. It presents the characteristics of protomylonite. The coarse-grained plagioclase, K-feldspar, and hornblende are well preserved. The quartz grains form polycrystal aggregate ribbons. The borders of quartz grains exhibit various morphologies, ranging from slightly curved to serrated (Fig. 3c), suggesting characteristics of dynamic recrystallization by grain boundary migration (Ahanger and Jeelani, 2022;Mansard et al., 2018;Passchier and Trouw, 2005;Stipp et al., 2002). The large plagioclase grains develop mechanical twins and fractures locally. Some of the fractures of plagioclase crosscutting the grains are filled by quartz (Fig. 3c).
Under the SEM observation, the BSE images reveal a characteristic core-mantle structure in the K-feldspar porphyroclasts surrounded by fine grains or subgrains (Fig. 4a, average length 760 μm). The long axes of porphyroclasts are parallel or oblique to the shear zone. Myrmekites develop at the rim of the K-feldspar porphyroclasts. Neocrystallization quartz grains (average length 45 μm, Fig. 4b) within myrmekites are elongated vertical to the long axis of K-feldspar porphyroclasts. Neocrys-

413
Lefan Zhan, Shuyun Cao, Yanlong Dong and Wenyuan Li tallization plagioclase grains (average length of ~60 μm) within the myrmekites are equiaxed. The fine-grained quartz and plagioclase grains nucleate around K-feldspar porphyroclasts (Figs. 4a,4b). The mica grains (average length 10 μm) are long-prismatic and plate-prismatic, slenderer than these in the main body of granodiorite (Figs. 3c,3d). Several mica pieces are parallel to each other, and they cut across quartz aggregates or locate in the edge of aggregates (Figs. 3c,3d). Some newly formed micas precipitate in the fractures of plagioclase grains and the grain boundaries of quartz aggregates (Fig. 4a).
The deformed Zone B is the most prominent characteristic of the mylonites with the porphyroclasts (feldspar and hornblende) embedded in a fine-grained matrix (Figs. 3d, 3e). The feldspar porphyroclasts are smaller in size compared to Zone A, with the features of elongated and lenticular, as well as ir-regular and serrated grain boundaries. Inhomogeneous extinctions of porphyroclastic feldspar grains are apparent, indicating plastic deformation. The elongated hornblende porphyroclasts form the mineral fish fabrics, presenting a dextral sense of shear (Figs. 3d, 3g). Locally, the quartz grains form typical polygonal aggregates, with the long axis parallel to or subparallel to the major stretching lineation in the small-scale shear zone. The matrix consists of plagioclase, K-feldspar, quartz, and biotite, containing a minor content of apatite. The mineral phases in the fine-grained matrix are not homogeneously mixed. Instead, the layering of compositions can be observed (Fig. 3e). Those layered aggregates of minerals exhibit an orientation roughly parallel to the mylonitic foliation.
The BSE images imply that the K-feldspar layers are composed of many small aggregates of fine-grained quartz and pla-

414
Strain Localized Deformation Variation of a Small-Scale Ductile Shear Zone gioclase (Figs. 4d, 4e). The same situation can be observed in the plagioclase layers (Figs. 4d,4f). Additionally, some small grains of mica are distributed in the plagioclase grain boundaries as rod-like cross-sections (Fig. 4f). In contrast to the quartzrich aggregates in Zone A, the quartz aggregates in Zone B are not bulk but layered (Fig. 3e). The grain boundary of quartz is irregular. Moreover, many relatively small, recrystallized quartz grains are mixed with K-feldspar, plagioclase, and biotite in the matrix at the edge of the aggregates (Fig. 4d). In the quartz-rich layers, isolated K-feldspar grains exist at triple junctions of quartz grains, and some small grains of biotite are distributed in the grain boundary (Fig. 4d). Except for the small grains in the quartz grain boundary, most biotite grains are highly elongated subparallel to the foliation and form biotite-rich layers (Fig. 3g).
Zone C presents the dominated characteristics of ultramylonites composed of extreme fine-grained matrix and only a few feldspar porphyroclasts with irregular grain boundaries. The hornblende disappears. Zone C is microstructurally more homogeneous than the other two zones and consists of fine- grained K-feldspar and plagioclase grains (Figs. 3d, 3f). The fine-grained grains of feldspar, quartz, and mica are slightly elongate or sub-equant. The feldspar porphyroclasts have disappeared. Quartz grains are disseminated throughout the matrix, and the residual quartz-quartz grain boundaries are more straight compared with those in the quartz aggregates in Zone A or Zone B (Fig. 4c). Phase boundaries between quartz and plagioclase or K-feldspar are frequently extensively curved. The biotite grains distribute homogeneously in the matrix and are extremely elongated subparallel to the foliation (Figs. 3f, 4c).

Mineral Grain Sizes of the Small-Scale Shear Zone
The grain size significantly decreases as the strain increases from Zone A to Zone C. The software Image J is adopted to count the size of grains by manual operation. In Zone A, the quartz grains gather together to form the quartz-rich aggregates (Fig. 5a). The mean and median grain sizes are 186 and 173 μm, respectively (Figs. 5b, 5c). In Zone B, the quartz aggregates are elongated and closely aligned with the foliation, which form the deformed quartz ribbons. The quartz ribbons are recrystallized and the neocrystallized grains are mixed into the matrix (Fig. 5a). The mean and median grain sizes are 88 and 80 μm, respectively (Figs. 5b, 5c). In Zone A, a relatively broad distribution of grain size can be observed, while a rather narrow distribution of grain size is observed in Zone B (Fig. 5c). In Zone A and Zone B, larger grain

416
Strain Localized Deformation Variation of a Small-Scale Ductile Shear Zone sizes correspond to the grains without recrystallization or recrystallization relict in the quartz-rich aggregates, and the smaller grain sizes correspond to the small, neocrystallized quartz grains in the quartz-rich aggregates or matrix at the edge of the aggregates. In Zone C, the quartz aggregates are fully recrystallized and mixed into the matrix (Fig. 5a). The mean diameter of quartz grains is reduced to 44 μm, and it has the narrowest distribution among the three zones (Figs. 5b, 5c).
Generally, the feldspar grain size is larger than the quartz grain size in all three zones. In Zone A, a considerable amount of plagioclase porphyroclasts appear with a size corresponding to the magmatic grain. Some neocrystallized plagioclase grains within the myrmekites are very small (Figs. 4a,4b,5a). Thus, the distribution of plagioclase grain size has a wide range from 80 to 1 000 in Zone A (Fig. 5c). The mean and median values of plagioclase grain size are 225 and 169 μm, respectively (Figs. 5b, 5c). In the box plot, the outliers represent large feldspar porphyroclasts (Fig. 5c). In Zone B, the plagioclase porphyroclasts are fully recrystallized (Fig. 5a), so the grain size decreases dramatically. The mean and median values of plagioclase grain size are 103 and 101 μm in Zone B, respectively (Figs. 5b, 5c). The distribution of plagioclase grain size in Zone B is much narrower than the distribution of plagioclase grain size in Zone A. The outliers in the box plot suggest that the amount of plagioclase porphyroclasts significantly decreases in Zone B (Fig.  5c). In Zone C, the mean and median diameters of plagioclase grains are reduced to 65 and 63 μm, respectively. Meanwhile, it has the narrowest distribution in the three zones, and there are hardly any plagioclase porphyroclasts (Figs. 5b, 5c).
The microstructure of K-feldspar grains is similar with the plagioclase grains in our sample (Fig. 5a). The mean and median values of K-feldspar grain size are 165 μm and is 145 μm in Zone A, respectively (Figs. 5b, 5c). The distribution of K-feldspar grain size is similar to the distribution of quartz grain size in Zone A. In Zone B, the mean and median values of K-feldspar grain size are all 97 μm. The distribution of K-feldspar grain size in Zone B is narrower than the distribution of K-feldspar grain size in Zone A. The outliers in the box plot reflect that the amount of K-feldspar porphyroclasts significantly decreases in Zone B (Figs. 5b, 5c). In Zone C, it has the narrowest distribution in the three zones, and there are hardly any K-feldspar porphyroclasts. The mean and median values of K-feldspar grain size are 60 and 59 μm, respectively (Figs. 5b, 5c).

MINERAL EBSD ANALYSIS IN THE SMALL-SCALE SHEAR ZONE
The CPOs (crystal preferred orientation) of quartz and feldspar were investigated mainly on the three zones (zones A, B, and C) of the micro-shear zone to further constrain the deformation conditions of the small-scale shear zones. The results are illustrated in equal-area lower-hemisphere pole figures.

Quartz and Feldspar Aggregates in Zone A
In Zone A, the quartz grains mainly formed irregular polycrystalline aggregates, and Dauphiné twins are occasionally observed in it (Fig. 6a). The pole figure of c-<0001> axis exhibits  -(11-20) planes present a weaker girdle close to the XZ plane (Fig. 6b). In the sample coordinate system (SCS), the low angle (2°-15° in this article) rotation axes demonstrate high spatial density close to the Y-axis, consistent with the pole figure of the c-axis (Fig. 6c). The low angle rotation axes indicate high spatial density close to the caxis in the crystal coordinate system (CCS; Fig. 6c). In the misorientation angle distribution histogram, the relative frequency of misorientation angles less than 15° is around 0.08, and the relative frequency of misorientation angle of 60° is around 0.13 in the corrected pairs (Fig. 7a). The misorientation angle distribution of the uncorrected pairs exhibits an irrelevance with the calculated random distribution curve (Fig. 7a).
The K-feldspar porphyroclasts gather into aggregates and are surrounded by large quantities of small quartz and plagioclase grains in Zone A (Fig. 6a). The pole figure of K-feldspar reveals a low maximum value of MUD, which is 2.69. The (100) plane forms two maxima in the z-axis, while the pole figures of (010) and (001) planes form a point maximum in the direction with a low angle to the X-axis (Fig. 6b). Although the rotation axes present a point maximum between the X-and Zaxis in the SCS and a point maximum close to the <001>-axis in the CCS, a clear clustering is not observed in the low angle rotation axis distributions of K-feldspar (Fig. 6c). In the misorientation angle distribution histogram, the distribution of misorientation angles of corrected pairs is uniform except <15° and 180°, whose relative frequencies are much high than other angles. The misorientation angle distribution of the uncorrected pairs reveals a positive correlation with the calculated random distribution curve (Fig. 7b).
The shape of plagioclase grains is regular, and Albite twins are common in the plagioclase grains in Zone A (Fig. 6a). The pole figure of plagioclase suggests a low maximum value of MUD, which is 2.23. The plagioclase and K-feldspar have similar crystallographic orientations in the (001) plane. The (010) plane forms a maximum between the X-and Z-axis, and the (100) plane presents high spatial density near the Y direction (Fig. 6b). In the SCS, the low angle rotation axes demonstrate high spatial density in the position between the X-and Z-axis, and the low angle rotation axes reflect high spatial density close to the <100>-axis in the CCS. The low angle rotation axis distributions of plagioclase are also disorderly similar to K-feldspar's

418
Strain Localized Deformation Variation of a Small-Scale Ductile Shear Zone (Fig. 6c). In the misorientation angle distribution histogram, the misorientation angle of 180° has the highest relative frequency of around 0.15 in the corrected pairs. The misorientation angle distribution of the uncorrected pairs exhibits a positive correlation with the calculated random distribution curve. The difference is that the random distribution curve rises linearly (Fig. 7c).

Quartz Ribbons and Feldspar Layers in Zone B
In Zone B, the quartz irregular polycrystalline aggregates have disintegrated, and the content of quartz in the matrix is higher than those in Zone A (Fig. 8a). The pole figure of the caxis reveals a well-developed point maximum in the Y-axis, with the maximum value of MUD of ~5.33. The pole figures of m-and a-plane show a weaker girdle in the XZ plane (Fig. 8b). In the SCS, the low angle rotation axes suggest high spatial density close to the Y-axis, similar to the pole figure of the c-axis (Fig. 6h). The low angle rotation axes exhibit high spatial density close to the c-axis in the CCS (Fig. 8c). In the misorientation angle distribution histogram, the relative frequency of the corrected pairs' misorientation angles (<15°) is also around 0.08, while the relative frequency of misorientation angles (60°) is around 0.12, which is less than the value in the Zone A. The misorientation angle distribution of the uncorrected pairs presents a negative correlation with the calculated random distribution curve (Fig. 7a).
In Zone B, the K-feldspar porphyroclasts are hardly observed, and the smaller grains gather into K-feldspar layers (Fig. 8a). The pole figure of K-feldspar reveals a low maximum value of MUD of ~2.14. The pole figure of the (100) plane forms a point maximum near the X-axis. The pole figures of (010) and (001) planes present weak patterns (Fig. 8b). The low angle rotation distributions indicate very scattered data, and the quantity of data points reduces to 70% compared to Zone A (Fig. 8c). The distribution of misorientation angles is also in line with those in Zone A, while the relative frequency of <15° and 180° is lower in the corrected pairs. The misorientation angle distribution of the uncorrected pairs exhibits a positive correlation with the calculated random distribution curve (Fig. 7b).
The pole figure of plagioclase reveals a low maximum value of MUD of ~1.90. The pole figure of the (001) plane forms maxima between the X-and Z-axis, and the pole figures of (100) and (010) planes present weak patterns (Fig. 8b). The low angle rotation distributions also indicate very scattered data, and the quantity of data points decreases to 40% compared with that of Zone A (Fig. 8c). The distribution of misorientation angles is consistent with those in Zone A, while the relative frequency of 180° is lower in the corrected pairs. The misorientation angle distribution of the uncorrected pairs demonstrates a positive correlation with the calculated random distribution curve. Besides, the random distribution curve rises linearly (Fig. 7c).

Mixed Matrix of Quartz and Feldspar in Zone C
In Zone C, the quartz grains are completely disseminated among the other matrix phases (Fig. 9a). The pole figures suggest a low maximum value of MUD of ~1.50. A clear clustering is not observed in the pole figure of the c-axis. The pole figures of m-and a-plane present a weak and wide girdle in the XZ plane (Fig. 9b). The low angle rotation axes exhibit a maximum close to the c-axis in the CCS, while it is much weaker compared with Zones A and B (Fig. 9c). In the misorientation angle distribution histogram, the relative frequency of the corrected pairs' misorientation angles (<15°) drops to 0.04, and the relative frequency of misorientation angles (60°) decreases to 0.08. The misorientation angle distribution of the uncorrected pairs displays a positive correlation with the calculated random distribution curve (Fig. 7a).
The feldspar grains are mixed with quartz grains in Zone C (Fig. 9a). The pole figure of K-feldspar reveals a low maximum value of MUD of ~1.67 (Fig. 9b). The low angle rotation distributions also suggest very scattered data without a clear clustering (Fig. 9c). In the misorientation angle distribution histogram of corrected pairs, the relative frequency of misorientation angles (180°) is low. The misorientation angle distribution of the uncorrected pairs exhibits a positive correlation with the calculated random distribution curve (Fig. 7b). The pole figure of plagioclase demonstrates a low maximum value of MUD of ~2.37. The feature of the pole figure of the (001) plane is consistent with that in Zone B (Fig. 9b). However, the low angle rotation distributions reflect few scattered data (Fig. 9c). The relative frequency of misorientation angles (180°) is around 0.10 in the misorientation angle distribution of corrected pairs. The uncorrected pairs present a positive correlation with the calculated random distribution curve. Additionally, the random distribution curve also rises linearly (Fig. 7c).

THERMOBAROMETRY OF THE GRANODIORITE 5.1 P-T Estimation
Hornblende-plagioclase thermometry is frequently utilized in granites and gneisses with coexisting hornblende and plagioclase to estimate temperature and pressure in magma crystallization or subsequent metamorphism (Dong et al., 2021;Ridolfi and Renzulli, 2011;Popp et al., 1995;Schmidt, 1992). Mineral chemistry was determined on unfoliated granodiorite at the GLG-SZ boundary and on foliated granitic rocks within the GLG-SZ. Experiments were conducted on hornblende grains  Strain Localized Deformation Variation of a Small-Scale Ductile Shear Zone combined with quartz or plagioclase under the premises for Al-in-hornblende barometry application (Ridolfi and Renzulli, 2011;Stein and Dietl, 2001;Popp et al., 1995;Schmidt, 1992;Hollister et al., 1987). In this study, the hornblende-plagioclase geothermometer designed by Holland and Blundy (1994) is employed to determine the temperature. These calculations were based on hornblende solid-solution models and wellconstrained natural systems. The temperatures and pressures of the unfoliated granitoids are T = 641-730 °C with an average of T = 673 °C, P = 4.0 -5.9 kbar with an average of P = 5.1 kbar (Fig. 10a). The crystallization P-T values for the foliated granitic rocks are T = 658-736 °C with an average of T = 710 °C, P = 2.1-2.9 kbar with an average of P = 2.7 kbar (Fig. 10b).

Emplacement Depth
The crystallization pressures for the investigated granodiorite are calculated by the method developed by Anderson and Smith (1995). It is possible to estimate the pressure with an error of about 0.6 kbar using the method proposed by Popp et al. (1995). This error corresponds to about 2.10 km in depth. The density assumption of 2.8 g/cm 3 is used for the GLG-SZ in our calculations to convert the pressures measured to emplacement depths. After the temperature adjustment, the calculated unfoliated granitoids pressures range from 4.0 ± 0.6 to 5.8 ± 0.6 kbar, implying that emplacement depths range from 14.3 ± 2.1 to 20.7 ± 2.1 km, and the average depth is 17.5 km (Fig. 10a). The calculated pressures in the foliated granitic samples change from 2.2 ± 0.6 to 2.9 ± 0.6 kbar, suggesting that emplacement depths vary from 7.8 ± 2.1 to 10.3 ± 2.1 km, and the average depth is 9.0 km (Fig. 10b).

PALEOPIEZOMETRY 6.1 Flow Stress Estimate from Recrystallized Quartz Grains
The grain size of dynamically recrystallized quartz varies with differential stress in plastic deformation and is used as a method for calibrating the magnitude of paleostress (Boutonnet et al., 2013;Koch, 1983;Twiss, 1980Twiss, , 1977Mercier et al., 1977). This study only considered dynamically recrystallized quartz grain sizes from the small-scale shear zones for estimating paleostress. The optical size of recrystallized grains was determined with standard petrographic microscopy. Measurements were performed with each grain perpendicular to macroscopic foliation and parallel to macroscopic lineation (Behrmann and Seckel, 2007). The results of the analysis are listed in Table 1. The standard error of the differential stress estimates is less than 15%.
The differential stress is estimated through piezometer calibration (Stipp and Tullis, 2003), followed by a calibration corrected by Holyoke and Kronenberg (2010). The low-strain domain (Zone A) average size is 165 μm; the differential flow stress is 12 MPa (Stipp and Tullis, 2003) and 9 MPa (Holyoke and Kronenberg, 2010). The medium-strain domain (Zone B) average size is 78 μm; the differential flow stress is 21 MPa (Stipp and Tullis, 2003) and 15 MPa (Holyoke and Kronenberg, 2010). The high-strain domain (Zone C) average size is 44 μm; the differential flow stress is 33 MPa (Stipp and Tullis, 2003) and 24 MPa (Holyoke and Kronenberg, 2010). However, uncertainties remain in the estimation of flow stress from the mineral grain sizes that can be affected by the presence of other phases and notably fluid during deformation.

Flow Stress Estimate from Recrystallized Grain Size
Estimating the strain rate is a critical step in comprehending deformation processes. The relationships between temperature, microstructures, and CPO patterns corresponding to the dominant slip systems indicate deformation under amphibolitefacies conditions at temperatures of ca. 400 -700 °C in the small-scale shear zones (Stipp et al., 2002;Hirth and Tullis, 1992). In this study, strain rates of quartz grains at temperatures of ca. 550 °C were constructed. Ductile creep curves were calculated for strain rates from 10 -10 to 10 -16 S -1 following the flow law from Luan and Peterson (1992) and the wet quartzite flow law of Hirth et al. (2001). In the calculation, the effect of water fugacity was considered, though the dependence of strain rate on water fugacity was not determined in the original paper. When it was applied to the differential stress estimates from dynamically recrystallized grain sizes of quartz by the piezometer (Holyoke and Kronenberg, 2010;Stipp and Tullis, 2003), strain rates estimated in Zone A are 4.75 × 10 -16 S -1 to 1.17 × 10 -14 S -1 , strain rates estimated in Zone B are 5.13 × 10 -15 S -1 to 1.26 × 10 -13 S -1 , and strain rates estimated in Zone C are 3.16 × 10 -14 S -1 to 7.75 × 10 -13 S -1 . The average strain rate estimated in Zone A is 4.29 × 10 -15 S -1 , the average strain rate estimated in Zone B is 4.62 × 10 -14 S -1 , and the average strain rate estimated in Zone C is 2.85 × 10 -13 S -1 .

DISCUSSION 7.1 Significance of Quartz CPOs within the Small-Scale Shear Zone
Before this study, the detailed characteristics and conditions of deformation and CPOs of minerals (quartz and feldspar) in the small-scale shear zone were largely undocumented and discussed, though numerous data on structures, microfabrics, and geochronology have been published from the GLG-SZ show Cenozoic high-temperature ductile deformation conditions (Dong et al., 2019;Xu et al., 2015;Zhang et al., 2012b). The small-scale shear zone developing in the unfoliated granodiorite presents a significant decrease in grain size from the rim (Zone A) to center (Zone C) with increasing strain. In the lowstrain domains of zones A and B, the quartz polycrystalline aggregated ribbons are characterized by grain boundary migration recrystallization, revealing a medium-high temperature plasticdeformation condition (Figs. 3, 6 and 7; Ahanger and Jeelani, 2022;Dong et al., 2019;Cavalcante et al., 2018;Hansen et al., 2013;Holyoke and Tullis, 2006;Passchier and Trouw, 2005;Stipp et al., 2002;Hippertt et al., 2001).
The deformation conditions can be recorded by developed dominated slip systems of deformed quartz grains, which are normally temperature-sensitive (Stipp et al., 2002). Prism <a> slip occurs frequently in high-grade metamorphic rocks (Cao et al., 2013a(Cao et al., , b, 2011b, while basal <a> slip appears in low-grade or overprinted metamorphic rocks (Cheng et al., 2018;Cao et al., 2011bCao et al., , 2010Toy et al., 2008). The low angle rotation axis distribution and the c-axis patterns of quartz grains from the three zones in the small-scale shear zone display dominated the high-temperature prism <a> slip system (Figs. 6, 8). However, the quartz c-axis patterns in zones A and B are more intensive, with the Max between 5.33 and 6.42. The misorientation angle distribution of uncorrelated grain pairs do not conform to the random curve (Fig. 8a).
All these results suggest that the quartz grains in the smallscale shear zone have undergone significant high temperature dislocation creep deformation, similar to the GLG-SZ (Dong et al., 2019). However, the quartz c-axes pattern in Zone C demon-strates the weaker intensive of 1.50 (Fig. 8b). The effects of intragranular deformation are dramatically reduced (Fig. 17a). Certain minerals from the high-strain Zone C are completely transformed into ultramylonites generated by extremely fine grains. Ultra-plastic flow is an essential process of quartz deformation in the high-strain domain within the shear zone.

Mechanism of Feldspar Deformation and Changed CPO Patterns
Studies have demonstrated that feldspars have different deformation behaviors and mechanisms, including brittle fracturing and cataclastic flow in the shallow crustal low-temperature conditions and dynamic recrystallization associated grain-size reduction under the high-temperature conditions (Dong et al., 2019;Mansard et al., 2018;Dang et al., 2017;Mancktelow and Pennacchioni, 2004;Wintscha and Yi, 2002;Yund, 1991, 1987;Kohlstedt, 1985, 1984). In the studied small-scale shear zone, the feldspar grains present the well-marked variation of compositions and grain sizes from the undeformed magmatic texture to typical crystal plastic flow deformation, revealed by the microstructure, CL, and CPO properties (Fig. 3). In the low-strain domain of Zone A, undulatory and inhomogeneous extinction are common in the porphyroclastic feldspar grains. Occasionally, the feldspar grains exhibit irregular and sharpened grain boundaries. Most K-feldspar porphyroclasts display elongation and grain-size reduction by dynamic recrystallization. The dynamic recrystallization grains occur in the asymmetric porphyroclast (e.g., hornblende and feldspar) tails with neocrystallization aggerates, which extend to the shear zone s mylonitic foliation/lineation. The K-feldspar porphyroclasts are surrounded by quartz, plagioclase, or K-feldspar fine grains, establishing a typical coremantle structure (Figs. 4a, 4b). With an increase in the strain, the feldspar porphyroclasts disintegrate, and the fine feldspar grains gather to layered aggregates with an orientation nearly parallel to the mylonitic foliation (Fig. 4e). Fractures or mechanical twins can be observed in a small amount of large Kfeldspar porphyroclasts (Fig. 3d).  Stress estimated using differential piezometer, ST-2003-Stipp and Tullis, 2003and SH-2010-Koch, 1983. H-2001 flow law is referenced to Hirth et al. (2001), LP-1992 flow law is referenced to Luan and Peterson (1992).

Deformation Associated Fluid of the Small-Scale Shear Zone
Recognizing the evolution of small-scale ductile shear zone is also particularly valuable for the understanding of the processes of shear localization in the middle and lower crusts (Pennacchioni and Mancktelow, 2018;Pen-nacchioni, 2013, 2005;Kilian et al., 2011;, as well as interpreting shear zone history regarding P-Tfluid evolution along strain gradients (Cao and Neubauer, 2016;Bestmann and Pennacchioni, 2015). Hydrolytic weakening has been demonstrated to be a major process facilitating strain localization (Cao et al., 2017;Finch et al., 2016). Fluid can weaken rocks or minerals in several methods (Cheng et al., 2018;Cao et al., 2017;Finch et al., 2016;Oliot et al., 2014;Kilian et al., 2011;Kohlstedt, 2006;Mancktelow and Pennacchioni, 2004;Sibson, 1977). This is in that the fluid in crystals can weaken the mechanical strength of crystals by decreasing the strength of Si-O bonds. It allows easier glide of dislocations (Kohlstedt, 2006) and diffusion at lower temperatures (Sibson, 1977). Intergranular fluid results in nucleation of new grains in cavities by mass transfer, contributing to accelerating grain boundary sliding and reducing the intercrystalline rock strength (Kilian et al., 2011;Mancktelow and Pennacchioni, 2004;Kronenberg, 1994;Chen and Argon, 1979). The relevant evidence from low-strain (Zone A) or medium-strain domains (Zone B) reveals that quartz and mica grains of extremely small size (ca. 20 μm) occur at the fine-grained plagioclase aggregates or at the junctions of K-feldspar grains (Fig. 11). This kind of feature can be explained by fluid-accommodated grain boundary sliding. The incomplete displacement along grain boundaries by grain boundary sliding can trigger the opening of the cavities and leads to the ingress and diffusion of the material (Mansard et al., 2018;Precigout et al., 2017;Finch et al., 2016;Platt, 2015;Kilian et al., 2011;Passchier and Trouw, 2005;Fliervoet et al., 1997). The neocrystallization grains pin in cavities restrains grain growth by impeding grain boundary migration and arresting the original grain size at the approximate size of the dynamic recrystallization new grains. The grain size reduction and the increase of phase mixing can further weaken rock and strain localization in the small-scale shear zones.
Intergranular fluid also reduces the intercrystalline rock strength by metamorphic reaction (Liu, 2017;Spruzeniece and Piazolo, 2015;Oliot et al., 2014;Hippertt, 1998;White and Knipe, 1978). For example, the retrogressive metamorphism of hornblende and involved water can produce the weaker phase as biotite and quartz (Fig. 4c;Liu, 2017). The studied smallscale shear zones reveal the distinct evidence of strain localization accompanied by hydrous retrogression of hornblende to interconnected weaker mica parallel to the main ultramylonitic foliation. That mica presents the visible appearance from undeformed magmatic phases to ductile deformed phases. The biotite occurs as undeformed or slightly bent magmatic phases in the studied unfoliated granodiorite (Fig. 3), generally regarded as the weakest phase (Tullis and Wenk, 1994). As the strain increases, the micas in the low-strain domain break into small pieces, and several mica pieces are parallel to each other which cut across quartz-rich layers. With further localization, a network of monophase biotite layers starts to form in the mediumstrain domain of Zone B (Fig. 3g). The interconnected mica layers can be frequently formed during crystal plastic deformation. Additionally, a portion of biotite appears at the quartz and plagioclase grain boundaries (Fig. 4). In the high-strain domain, the network of micas is destroyed, and the mica grains 423 Lefan Zhan, Shuyun Cao, Yanlong Dong and Wenyuan Li are disseminated in the matrix (Figs. 4c,11c). This presents similar distributed features to diffusion creep (Herwegh and Jenni, 2001;Fliervoet et al., 1997). The microstructure characteristics imply that the formation of biotite can soften the rock's matrix, resulting in increased deformation intensity (Mansard et al., 2018;Fossen and Cavalcante, 2017;Mancktelow, 2008).

Formation Conditions and Processes of the Small-Scale Shear Zone
Microstructural analyses of the small-scale shear zones reveal ductile deformation processes in the deep-seated crust. The unfoliated granitoids have an average emplacement depth of 18 km. Macro-and micro-structures apparently reveal progressive plastic deformation behaviors of the major mineral phases (quartz, feldspar, hornblende, and mica) in response to a progressive ductile deformation history of the small-scale shear zone (Fig. 11). The results demonstrate that the smallscale shear zone has experienced high-temperature deformation conditions at least amphibolite facies during the dominant ductile shearing. The small-scale shear zones and the foliated granodiorite within the GLG-SZ exhibit identical synkinematic metamorphic assemblages and microstructures ( Fig. 3; Dong et al., 2021Dong et al., , 2019. In other words, they formed under similar metamorphic and deformation conditions. The temperature conditions of foliated granodiorite are determined based on hornblende-plagioclase thermometry to be 670 -735 °C (Fig.  10b). Similar temperature conditions are inferred in the smallscale shear zones. The development of myrmekite and the recrystallization of quartz also can verify the high-temperature metamorphic conditions in the small-scale shear zones (Figs. 3 and 4; Ceccato et al., 2018;Tribe and D'Lemos, 1996;Wirth and Voll, 1987). The initiation of prism <a> slip and the features of grain boundary migration recrystallization of quartz grains confirm that the small-scale shear zones formed at the medium-high temperature condition ( Fig. 6; Xia and Liu, 2011;Gibert and Mainprice, 2009;Toy et al., 2008;Passchier and Trouw, 2005;Stipp and Tullis, 2003;Mainprice et al., 1986;Hobbs, 1985).
Generally, rheological weakening mechanisms play a significant role in the localization of shear zones, as shear zones typically form in the weakest zone (Pennacchioni and Mancktelow, 2018;Fossen and Cavalcante, 2017;Liu, 2017;Cao and Neubauer, 2016;Yamasaki et al., 2014;Dayem et al., 2009;Rosenberg, 2004;Imber et al., 1997;Schmid et al., 1996). Besides temperature and pressure, several other factors such as mineralogy, strain rate, microstructure, texture, and fluid that can weaken mechanisms are activated during strain localization (Pennacchioni and Mancktelow, 2018;Fossen and Cavalcante, 2017;Liu, 2017;Cao and Neubauer, 2016;Finch et al., 2016;Oliot et al., 2014;Mancktelow and Pennacchioni, 2004). Localization may be caused by an external inhomogeneity such as a precursor fracture or joint, according to the macroscopic shear zone distribution and orientation similar to joint orientations in the same rock (Menegon and Pennacchioni, 2009). In this study, microstructural observations imply that the switch of deformation characteristics from wall rocks to the highstrain domain cannot be explained by the "precursor effect" (Fig. 3d). This inference cannot be confirmed by the varying orientation of small-scale shear zones (Fig. 1c). The hightemperature deformation conditions suggest that the initiation of the small-scale shear zones is at depth, where temperaturecontrolled rheological weakening mechanisms play an essential role in localizing future shear zones. Thermal heterogeneities of the lithosphere can lead to shear concentration along hot-to-cold contacts ascribed to thermally enhanced rheological weakening. Hence, rheological weakening by heterogeneities may induce strain localization and increase strain rates within magmatic rocks bearing shear zones (Fossen and Cavalcante, 2017;Liu, 2017;Cao and Neubauer, 2016). The fast strain rate within the high-strain domain of the small-scale shear zone is the consistent geological evidence (Table 1). Thus, the deformation in the small-scale shear zones would be localized to a narrow site through thermal-enhanced rheological weakening mechanisms when GLG-SZ is deformed.
Interestingly, the strain rates gradually decrease from the high-strain (2.85 × 10 -13 S -1 ) to the low-strain domain (4.29 × 10 -15 S -1 ) in the small-scale shear zone. It can be explained by the model of shear zone widening during shear deformation (Oliot et al., 2014) after the influence of temperature is eliminated. The center of a shear zone generally presents extreme grain-size reduction, involving fluid-assisted granular flow deformation. The GBS can facilitate fluid migration through shear zones by causing cavities to open and closure, which in turn induces fluids to be pumped through the rock (Mansard et al., 2018;Finch et al., 2016;Kilian et al., 2011;Passchier and Trouw, 2005;Fliervoet et al., 1997). This created a pressure gradient, which expelled fluids, leading to hydraulic microfracturing, metasomatism, host rock weakening, and shear zone widening (Finch et al., 2016;Oliot et al., 2014Oliot et al., , 2010. In the small-scale shear zone, the localized deformation provoked the release of intracrystalline water to grain boundaries and the migration of water to less deformed rocks, widening the shear zones. Fluid content could restrict weak rheology in the widened shear zone and decrease strain rates. This is also the reason for the grain size stratification of smallscale shear zones (Fig. 11). Field observations and microstructural observations reveal that the kinematic directions of the small-scale shear zone and the continental-scale GLG-SZ are consistent (Fig. 3), reflecting that small-scale shear zones are controlled by the GLG-SZ during progressive deformation and exhumation. The geothermal data demonstrate that the foliated granitic rocks have occurred at least 9 km of vertical displacement in GLG-SZ. The intrusion depth to the shearing depth may be related to the GLG-SZ tectonic shearing and exhumation (Fig. 10). It also produces a lower temperature condition in granitoids where brittle deformation can occur. Thus, the small-scale shear zone activates as a fault and makes the mafic enclaves cut in the outcrop-scale (Fig. 2g).

CONCLUSIONS
The analyses of meso-and micro-structural, EBSD texture, paleopiezometry, and thermobarometry lead to the following conclusions (1) The small-scale shear zones at the boundary of GLG-SZ have experienced vibrations deformation, mineral composition, and fabric transition from the rim zone of protomylonite to the center zone of ultramylonites, accompanied by a significant grain-size reduction and progressive phase mixing of minerals with an increase in the strain.
(2) The progressive development of microstructures suggests that the ductile deformation of small-scale shear zone is at least amphibolite facies conditions. Rheological weakening due to thermal heterogeneities induces strain localization, resulting in the initiation of the small-scale shear zone.
(3) The deformation mechanism of the coarse-grained aggregate zone in the shear zone is dominated by dislocation creep, while the polyphase fine-grained mixed zone possesses the dominant mechanism of viscous grain boundary sliding. Fluid-assisted deformation plays a crucial role in the hydrous retrogression and subsequent flow rheological weakening of the shear-zone.
(4) The deformation of the small-scale shear zone within the unfoliated granodiorite is controlled by the continental GLG-SZ. The small-scale shear zones experience the same kinematic direction of ductile shearing at depth and during exhumation, as well as the GLG-SZ.