Late to post-Variscan differential uplift exhumation and basement segmentation along the SW Bohemian Massif, Central Europe

. The exposed Variscan basement in Central Europe is well-known for its complex structural and lithological archi-10 tecture resulting from multiple deformation phases. We study the southwestern margin of the Bohemian Massif, which is characterized by major and long-lived shear zones, such as the Pfahl and Danube shear zones, extending over >100 km and initiated during Variscan tectonics. We integrate Bouguer gravity anomaly and LiDAR topographic data analyses and combine our results with available data and observations from low-temperature thermochronology, metamorphic grades, and the exposed granite intrusion depthsinventory to detect patterns of basement block segmentation and differential upliftexhumation. 15 Three NW-SE striking basement blocks are bordered by the Runding, Pfahl, and Danube shear zones from the northeast to the southwest. Basement block boundaries are indicated by abrupt changes in measured gravity patterns and metamorphic grades. By applying high-pass filters to gravity data in combination with lineament analysis, we identified a new NNW-SSE striking tectonic structure (Cham Fault), which further segments known basement blocks. Basement blocks that are segmented by the Cham Fault differ in the abundance and spatial distribution of exposed late Variscan granites and are further characterized by 20 variations of apparent thermochronological age data. Based on our observations and analyses, a differential exhumation and tectonic tilt model is proposed to explain the juxtaposition of different crustal levels exposed at the surface. Block segmentation along the NW-SE striking Pfahl and Runding shear zones most likely occurred prior, during, and after late -orogenic granite emplacement at ca. 320±10 Ma, as some of the granites are cross-cut by the shear zones while others utilized these structures during magma ascent and emplacement. In contrast, activity and block segmentation along the Cham Fault occurred after 25 granite emplacement as the fault sharply truncates the granite inventory. Our study provides evidence for intense and continuous fault activity during late and early post-orogenic times and highlights the importance of tectonic structures in the exhu-mation and the juxtaposition of different crustal levels and the creation of complex lithological patterns in orogenic terrains.

The outlined relatively simple succession of tectonic events does not reflect the real complexity of the Bohemian Massif's structure, juxtaposing high-grade metamorphic domains (e.g., the Moldanubian Unit) against low-grade domains (e.g., the Teplá-Barrandian Unit, Krohe, 1996;Cymerman et al., 1997;Kroner et al., 2008). At the local scale, complex lithological and metamorphic patterns can also be observed within each metamorphic unit (Krohe, 1996), which is especially evident in the 50 Moldanubian Unit along the southwestern Bohemian Massif (Fig. 1b). Several studies have contributed to the deciphering of the geochronological and geochemical character of that area, especially focusing on the magmatic evolution of the granitic intrusions during the late Variscan tectonothermal event (Finger and Clemens, 1995;Chen et al., 2003;Chen and Siebel, 2004;Dietl et al., 2005;Siebel et al., 2006b;Siebel et al., 2008;Klein et al., 2008;Galadí-Enríquez et al., 2010;Finger et al., 2010).
However, the causes for the observed juxtaposition of various metamorphic units and the role of fault reactivation and associ-55 ated upper crustal vertical movements have not yet been investigated in detail.
In this paper, we apply an integrated methodological approach, combining the analysis of filtered gravity anomaly data, highresolution Digital Elevation Models (DEMs), and published thermochronological data to reveal the spatial distribution of exposed Variscan units and their boundaries along the southwestern Bohemian Massif. In this context, we also discuss the role of upper crustal fault zones in the exposure exhumation of different crustal levels responsible forand the observed juxtaposition 60 of varying lithological domains, the latter being one of the most characteristic features of the entire Variscan Orogen (e.g., Krohe, 1996).

Origin and characteristics of the Moldanubian basement units
During the ceasing Cadomian Orogeny in the late Late Neoproterozoic/early Early Paleozoic, the area of the nowadays south-80 western Bohemian Massif was situated along the northern margin of Gondwana (e.g., Rohrmüller et al., 1996;Linnemann et al., 2004;Fatka and Mergl, 2009;Žák and Sláma, 2018). A succession of mainly pelitic greywackes, intercalated with magmatitesigneous rocks, are thereby considered as the protoliths of the so-called "Monotonous Group". The latter is characterized by a series of biotite-plagioclase-bearing para-and orthogneisses with an only a minor abundance of quartzites, marbles, graphitic schists, amphibolites, and granitic gneisses (Rohrmüller et al., 1996;Franke, 2000;Kroner et al., 2008). On the other 85 hand, volcano-sedimentary successions, probably formed in response to continental rifting and associated volcanism, are documented within the so-called "Varied Group". This group is characterized by paragneisses with higher abundances of quartzites, marbles, graphitic schists, and amphibolites (Rohrmüller et al., 1996;Kroner et al., 2008). In the southeastern part of the Bavarian Forest, the diatexites and their enclaves are interpreted as derivatives of a series of dacites, andesites, and basalts, which indicates the presence of an ancient island arc (Propach et al., 2008). 90 Late Neoproterozoic/early Early Paleozoic units of the southwestern Bohemian Massif were overprinted by HT/LP metamorphism during the late stages of the Variscan Orogeny. Pressures of up to 7 kbar (i.e., ca. 20 km burial depth) and temperatures of >800 °C led to partial melting and the formation of vast migmatite complexes (Grauert et al., 1974;Kalt et al., 1999;Kalt et al., 2000), the anatectic grade of which is increasing from the northeastNNE to the southwestSSW (Blümel, 1972;Teipel et al., 2008;Galadí-Enríquez et al., 2009b). Whereas this metamorphic event has been dated to ca. 335 to 340 Ma in the southern 95 central Bohemian Massif, younger thermal overprint and partial anatexis are documented for the Bavarian part of the Moldanubian Unit, with metamorphic ages progressively decreasing towards the southwest (ca. 320 to 315 Ma in the area to the southwest of the Pfahl Shear Zone, Kalt et al., 2000;Propach et al., 2000;Gerdes et al., 2006;Finger et al., 2007;Siebel et al., 2012;compiled by Teipel et al., 2008). Contemporaneously, voluminous granite bodies intruded large parts of the southwestern Bohemian Massif (Table 1). Estimates of granite emplacement depths in the study area are restricted to the southeastern Ba-100 varian Forest ( Fig. 1b and Table 1). They vary between 14-15 km for the Saldenburg Granite, which is part of the Fürstenstein Composite Massif (Dietl et al., 2005), and 16-18 km for the Hauzenberg Granite II, which is part of the Hauzenberg Composite Massif (Klein et al., 2008). Erosional products of these late -orogenic granites were deposited in the adjacent Permian basin (Naab Trough sensu Schröder, 1988), indicating their rapid exhumation and erosion shortly after their Upper Late Carboniferous to early Early Permian emplacement Galadí-Enríquez et al., 2009a). 105 Table 1 Ages and minimum emplacement depths for granites along the southwestern Bohemian Massif. All ages have been measured on zircons. For granite locations, see Fig. 3. a Klein et al. (2008), b Siebel (2004), c Dietl et al. (2005), d Siebel et al. (2008), e Siebel et al. (2006a), f Siebel et al. (2010), g Chen et al. (2003). ND not determined.

Structural characteristics
The tectonic configuration of the southwestern Bohemian Massif during the late stages of the Variscan Orogeny is interpreted 110 as a conjugate shear system related to N to NNW directed shortening (Wallbrecher et al., 1991;Brandmayr et al., 1995;Peterek et al., 1996;Galadí-Enríquez et al., 2010). The NW-SE trending Pfahl Shear Zone strikes over 150 km from northern Austria to southeast Germany; it is the best-known and one of the most prominent examples among the resulting dextral strike-slip faults (e.g., Brandmayr et al., 1995; (Teipel et al., 2008;Galadí-Enríquez et al., 2009b;Fig. 1b). Changes in anatectic grades and variations of the granite geochemistry across the Pfahl Shear Zone are interpreted as reflecting differential uplift exhumation with deeper crustal levels exposed to the southwest of the shear zone (Grauert et al., 1974;Beer, 1981;Finger and Clemens, 1995;Siebel et al., 2008;Finger and Rene, 2009). In addition, analysis of quartz mineralizations along the Pfahl Shear Zone gives an indication for the exposure of deeper crustal levels along the southeastern segment of the shear zone 120 . Two contrasting interpretations of the origin of the Pfahl Shear Zone exist, i.e., (I) the shear zone represents a former suture zone along which two different basement terranes amalgamated (Siebel et al., 2008;Siebel et al., 2009) or (II) the shear zone just intersects the otherwise continuous Moldanubian Unit (Finger et al., 2007;Finger and Rene, 2009;Finger et al., 2010).
Sub-parallel and ca. 10 km north of the Pfahl Shear Zone, the Runding Shear Zone represents another important tectonic 125 lineament in the study area (Fig. 1b, Table 2). Similar to the Pfahl Shear Zone, the Runding Shear Zone marks a pronounced change in anatectic grades, separating the metatectic-dominated domain of the northeastern Bavarian Forest from higher-grade diatectic rocks in between the Pfahl and Runding shear zones (Teipel et al., 2008;c.f., Fig. 3Fig. 1b).
In the southwest, the exposed Moldanubian basement is delimited by the Danube Shear Zone (Fig. 1b). Similar to the Pfahl Shear Zone, the Danube Shear Zone is characterized by multiple deformation phases lasting from the Llate Paleozoic until the 130 Cenozoic ( Table 2). The structural grain patterns in between these two major shear zones is are interpreted as a consequence of to be related to a dextral Riedel-type shear system that developed in response to a "rift-and-wrench" tectonic phase under a NNW-SSE directed compressional stress field (Zeitlhöfler, 2007).

Data and methodology
This study uses an integrated approach combining the analysis of Bouguer gravity anomaly data and Digital Elevation Models (DEMs) to unravel the lithological and structural architecture of the exposed Variscan crust in the southwestern Bohemian 140 Massif.

Gravity analysis
The Bouguer gravity dataset is part of a pool of ca. 350,000 reprocessed and merged gravity data points in Germany and surrounding areas with a mean point spacing of 2-3 km and an overall accuracy of ± 100 µGal (Leibniz-Institut für Angewandte Geophysik, 2010;Skiba, 2011). In this study, we compare surface geology with unfiltered and filtered Bouguer gravity anom-145 aly data to identify potential granites in the subsurface and to reveal the spatial relationship to their exposed counterparts. We applied 20 and 30 km wavelength high-pass filters to the Bouguer gravity anomaly data (using Oasis Montaj, ©Seequent) to identify regional and local anomaly sources in the subsurface. The 20 km high-pass filter mainly includes the gravity signature of causative bodies located in the shallower subsurface, while a 30 km high-pass filter also considers somewhat deeper crustal bodies (e.g., Lowrie, 2007). In areas with exposed crystalline rocks, local circular and semi-circular negative gravity anomalies 150 are often related to exposed and buried granites due to the density contrast between granites and the higher-density metamorphic country rocks (e.g., Trzebski et al., 1997;Siebel et al., 1997;Sedlák et al., 2007;Sedlák et al., 2009). To confirm the density difference between granites and surrounding metamorphic rocks, we collected density data from exposed granites and compared these to published metamorphic rock densities. Granite densities were measured by applying a buoyancy technique in isopropanol (Archimedes' principle). Samples were either collected as loose rocks or, if applicable, as drilled cores of two 155 to five centimeters in diameter to ensure "fresh" samples without major cracks.

Topographic analysis
The DEM used in this study is based on LiDAR (Light Detection and Ranging) point clouds acquired by the Bavarian Agency for Digitisation, High-Speed Internet and Surveying. Only the last returned laser pulses have been considered in the dataset so that the final DEM represents the landscape's elevation without being affected by the vegetational cover. Each data point has 160 been georeferenced, referred to the Gauss-Krüger Coordinate System (zone 4) and the German Combined Quasigeoid 2011 (GCG2011) for precise horizontal and vertical positioning, achieving horizontal accuracies of ± 0.5 m and vertical accuracies of ± 0.2 m. DEM data were resampled in ArcGIS Pro (©Esri) to a spatial resolution of 10 m, which ensures efficient data processing with sufficient detail to perform lineament mapping and analysis in the study area.

165
Traditional elevation profiles along single lines can only image a small fraction of all topographic characteristics, which is related to their small footprint. In contrast, topographic swath profiles are able to illustrate even complex landscapes without neglecting the spatial distribution of important but spatially limited morphological features such as local peaks and troughs (e.g., Telbisz et al., 2013). We used topographic swath profiles to illustrate the overall topographic expression of the study area. Swath profiles are constructed by projecting the elevation data within a rectangular swath onto its longitudinal axis. 170 Depending on the spatial extent of the studied area, the swath can have widths ranging from 100s of meters to 10s of kilometers.
Swath profile data were obtained using the ArcGIS Add-in "SwathProfiler" (Pérez-Peña et al., 2017). We used a fixed swath width of 10 km for all of the profiles. This value is large enough to summarize the present topographic variations adequately while small enough to avoid mixing of morphologies from very different landscapes. Elevation data were sampled along 50 parallel profiles with a step size of 15 m (i.e., 1.5 times the DEM resolution). Important statistical parameters such as the 175 maximum, minimum, and mean elevation can be illustrated simultaneously, providing information about both the general spatial distribution of elevations within an area but also of discrete topographic features such as local peaks and troughs, paleosurfaces, and incised valleys.

Lineament analysis
The analysis of topographic lineaments yields valuable information on the structural patterns of large areas, especially if remote 180 or inaccessible. In cases where active tectonic processes prevail over erosional processes and leveling of the landscape by sediment mobilization and deposition, fault slip is often immediately transferred to the Earth's surface, forming distinct topographic lineaments (i.e., fault scarps) that can be easily traced and interpreted using aerial photographs, DEMs, or field mapping (e.g., Stewart and Hancock, 1990;Keller and Pinter, 2002;Burbank and Anderson, 2012). If the time between fault rupture is long or tectonic activity ceases, however, the scarp will degrade and geomorphic processes will modify the hillslopes 185 and channels of the landscape towards a new equilibrium. In such a case, tectonic structures will no longer be visible as welldefined fault scarps but, because they often induce gradients in rock erodibility, as linear to curvilinear river valleys, ridgelines, or slope breaks (e.g., Jordan et al., 2005; Fig. 2). Strike-slip faults thereby tend to form symmetrical morphologies such as river valleys or ridgelines (e.g., Fürst et al., 1978;Fig. 2). High-angle normal and reverse faults typically form linear slope breaks ( Fig. 2), whereas low-angle thrust faults tend to appear somewhat irregular in topography (Fürst et al., 1978;Prost, 190 1994; Drury, 1987;Goldsworthy and Jackson, 2000).
Lineaments were mapped detected using DEM-derivatives such as hillshade, slope, curvature, and aspect maps as enhancement tools to identify linear features in topography. A shortcoming of traditional hillshade maps is that they induce directional biases due to their heterogeneous way of illuminating the landscape, which is mainly a result of the fixed azimuth of the artificial light source (e.g., Scheiber et al., 2015). To reduce this bias, we used multi-directional hillshades to map topographic 195 lineaments, illuminating landscapes in a more homogenous way.
Calculation of DEM derivatives and the mapping procedure were carried out in ArcGIS Pro (©Esri). Statistical analysis was accomplished by using a modified routine in MATLAB (©Mathworks) that partly relies on the FracPaQ toolbox (Healy et al., 2017).

Basement lithological configuration of the southwestern Bohemian Massif and its relation to the structural architecture
Three main basement domains in the study area are defined based on their dominant metamorphic grades (Fig. 3). Basement 210 domains A, B, and C are bounded by the essentially NW-SE striking Pfahl, Danube, and Runding shear zones (Fig. 3b). By considering the spatial distribution of exposed late Variscan granites, defined basement domains are further subdivided into (1) domains A1, B, and C1, where late Variscan granites are less exposed and are preferentially aligned with NW-SE striking shear zones, and (2) domains A2 and C2, where late Variscan granites are abundant and not necessarily aligned with major shear zones (Fig. 3b).

Domain A: Diatexite-dominated
Domain A exposes rocks that have experienced an advanced stage of anatexis, with diatexites forming the predominant rock type (ca. 45 % at the present erosional level, Fig. 3a). In the central part, gneissic rocks are intercalated with diatexites ( Fig.   3b). Lower-grade paragneisses are restricted to the so-called "Donauleiten-Serie" in the very southeast of domain A (Daurer, 1976;Fig. 3b). Late Variscan granites in domain A are extensively exposed in the northwest, whereas in the central part, a 220 very limited number of granitic bodies are exposed, which are restricted to the traces of the Pfahl and Danube shear zones (e.g., the Metten Massif and the Patersdorf Stock, Fig. 3b). Based on the abundance and distribution of exposed late Variscan granites, a distinct boundary subdivides domain A into domain A2, i.e., the northwestern part with abundant granites, and A1, i.e., the central-southeastern part, where granites are solely exposed along the traces of major fault zones (Fig. 3b). This boundary also defines the southeastern border of the Cretaceous Bodenwöhr Trough and marks the western limit of the ca. 10 km 225 wide, NNW-SSE striking Stallwang Fault Zone, indicating a tectonic origin of the boundary (Troll, 1967). Towards the very southeast of domain A1, the abundance of granites gradually increases again and their exposure is not aligned with the Pfahl and Danube shear zones (Fig. 3b). Here, the diatexites and their enclaves are interpreted as derivatives of a series of dacites, andesites, and basalts, indicating the presence of an ancient island arc (Propach et al., 2008). Despite the different lithological character of the southeastern part of domain A1 (i.e., ortho-anatexites and extensively exposed granites), a distinct boundary 230 separating this part from the northwestern part of domain A1 (i.e., para-anatexites and fault-related granites) is not observed.
Hence, we interpret the southeastern part of domain A1 as a lithologically different portion of the same (fault-bounded) domain.

Figure 3
Statistical and spatial distribution of exposed lithologies. Five fault-bounded domains with characteristic rock inventories can be identified (A1, A2, B, C1, and C2). (a) Stacked bar plots illustrating the relative occupied area of the exposed lithologies in each of the  Fig. 4). The densities of exposed granites in the study area range from 2640 to 2680 kg/m³, whereas the metamorphic country rocks show densities > 2690 kg/m³ (Table 3). Distinct local and circular to semi-circular gravity lows along the western Bohemian Massif that are not associated with exposed granites are generally interpreted as buried granite bodies (e.g., Bott and Smithson, 1967;Behr et al., 1989;Trzebski et al., 1997;Sedlák et al., 2009;Petkovic, 2014;de Wall et al., 2019). Applying 20 and 30 km high-pass filters confirms the presence of granitic 270 bodies (circular and semi-circular gravity lows) at different crustal levels (Figs. 4c and 4d). Some of these subsurface granites trend subparallel to major fault zones, similar to their exposed counterparts ( Fig. 4a-d). The presence and distribution of buried granites, as evidenced by filtered gravity anomaly maps, suggest a rather homogenous distribution of granites in the subsurface of the study area ( Fig. 4c-d).
275 Figure 4 Compilation of unfiltered and filtered gravity data. (a) Gray-scaled geological map of the southwestern Bohemian Massif with exposed late Variscan granites highlighted in red (modified from Freudenberger and Schwerd, 1996;Toloczyki et al., 2006;Teipel et al., 2008;Galadí-Enríquez et al., 2009b). (b) Unfiltered Bouguer anomaly map showing the total gravity signal of the study area and its surroundings (data source: Leibniz-Institut für Angewandte Geophysik, 2010; Skiba, 2011). (c)-(d) High-pass filtered gravity data (20 km and 30 km wavelength, respectively) depicting the upper crustal configuration of the study area. The data are color-coded with respect to their 280 standard deviation (σ) to highlight relative gravity highs and lows. Traces of major fault zones and the interpreted Cham Fault, as well as the locations of exposed and interpreted granite bodies, are shown. The scientific color maps "roma" (b) and "vik" (c-d) (Crameri, 2021) are used to prevent visual distortion of the data and exclusion of readers with color-vision deficiencies (Crameri et al., 2020 Klominský et al. (2010). TB Teplá-Barrandian.
The metamorphic rocks surrounding the granites show different gravity signatures . Areas dominated by gneissic rocks generally show less pronounced negative Bouguer anomalies (domain C, Fig. 4b-d). In contrast, rocks 290 of higher metamorphic grades (i.e., diatexites) show more pronounced negative anomalies (domain A, Fig. 4b-d). This relationship can be explained with by the density contrasts of the present rock types, with higher-grade metamorphic rocks tending to be less dense compared to lower-grade rocks (Table 3).

Topographic analysis
In this chapter, we present the morphological analysis results using swath profiles and the mapping of topographic lineaments to confirm and characterize basement domains identified previously based on the spatial distribution of metamorphic rocks and late Variscan granites.

300
Two swath profiles illustrate across-and along-strike topographic variations of the southwestern Bohemian Massif (Fig. 5). A mountainous landscape marks the northwestern parts of domain A1 with peaks up to ca. 1000 m a.s.l. (Fig. 5a-b). This elevated area shows an increased abundance of gneissic rocks (Fig. 5b). In contrast, the southeastern parts of domain A1, which is dominated by higher-grade diatexites, show a low-lying landscape with a reduced relief and elevations down to ca. 350 m a.s.l.. In domain A2, the Variscan basement is predominantly exposed to the south of the Regen River and is characterized by an undulating relief with elevations in between ca. 400 and 750 m a.s.l. (Fig. 5a). The very southeast of that area is marked by a 315 distinctly elevated plateau that gently dips towards the northwest (Figs. 5a-b). To the north of the Regen River, the Mesozoic to Cenozoic sediments of the Bodenwöhr Trough evoke a flat relief (Fig. 5a). The Stallwang Fault Zone, which is defined as the area in between the Cham Fault and the Rattenberg Fault Zone, borders domain A2 to the southeast and is characterized by sub-parallel, NNW-SSE-trending erosional furrows indicating a highly deformed area that favors erosion (Figs. 5a-b). Domain B, with its high-grade metamorphic rocks (diatectic gneisses), shows lower elevations compared to adjacent areas and 320 catches the drainages of its surroundings (Figs. 5a and 5c).
A very diverse topography characterizes the metatectic-dominated area of domain C1, with its eastern part comprising a highrelief mountainous landscape (Figs. 5a and c). In contrast, in the northwest, a heterogeneous landscape with basins (e.g., the Rötz Basin) that are surrounded by peaks up to 900 m a.s.l. high is developed (c.f. Figs. 6a and 7). Domain C2, which is also dominated by metatectic rocks but comprises a much higher number of exposed granites compared to C1, is characterized by 325 low to intermediate elevations of ca. 350 to 750 m a.s.l.. In its center, the landscape is dissected by the broad Naab River valley, draining to the south (Fig. 5a). To the west of the Naab River, the Naab Mountains form a local fault-bounded topographic high (Fig. 5a).

Topographic lineament analysis
In total, 1315 topographic lineaments have been identified in the study area (Fig. 6). Linear river valleys count for ca. 50 %, 330 ridgelines count for ca. 30 %, and sudden breaks in slope count for ca. 10 % of the identified lineaments. The remaining ca.
10 % of the identified lineaments are related to a combination of valleys, ridgelines, and/or slope breaks. Known faults are thereby often associated with distinct lineaments, suggesting a close relationship between tectonic and topographic features.

Topographic signatures of major fault zones
Major fault zones are all well-visible in topography (Fig. 6a). The Pfahl Shear Zone is characterized by two different topo-335 graphic expressions. Along its northwestern and southeastern segments, the shear zone is expressed by pronounced slope breaks ( Fig. 6a and profiles 1 and 3 in Fig. 6b). In the northwest, the Pfahl Shear Zone defines the northern border of the Bodenwöhr Trough, whereas its southeastern segment separates a mountainous landscape in the north from a moderate-relief landscape in the south (Fig. 6a). Along its central segment, the Pfahl Shear Zone forms a narrow but distinct morphological ridge (Fig. 6b, profile 2). This morphological ridge is attributed to the pervasive quartz mineralization along the Pfahl Shear 340 Zone (Priehäusser, 1961;Lehrberger et al., 2003). The Runding Shear Zone is predominantly characterized by deeply incised river valleys ( Fig. 6a and profile 4 in Fig. 6b). To the northwest, the Runding Shear Zone terminates against the Cham Fault and borders the Cham Basin to the north (Fig. 6a). 350 This contrasts to the southeast, where the Runding Shear Zone is expressed as a distinct slope break terminating against the Pfahl Shear Zone (Fig. 6a and profile 5 in Fig. 6c). Here, elevations of up to 1000 m a.s.l. occur to the northeast of the Runding Shear Zone, whereas generally lower elevations of less than 650 m a.s.l. prevail to the southwest. Interestingly, our data also provide evidence for the presence of several lineaments splitting off the Runding Shear Zone towards the east and southeast.
These lineaments rotate from an overall NW-SE into an E-W orientation, indicating the presence of a horsetail splay originating 355 from the terminating Runding Shear Zone (Fig. 6a).
The Danube Shear Zone and Keilberg Fault Zone show distinct slope breaks in topography (Fig. 6a). In the case of the Danube Shear Zone, however, this slope break is not expressed uniformly along-strike. Along the northwestern segment of the Danube Shear Zone, the slope break occurs very abruptly ( Fig. 6a and profile 6 in Fig. 6c),. In contrast, whereas the central segment is largely buried beneath the Cenozoic sediments of the North Alpine Foreland Basin (profile 7 in Fig. 6c). HereAt this location, 360 the topography is very irregular, with numerous embayments of Cenozoic sediments occurring across the Danube Shear Zone and reaching far into the crystalline basement in the northeast (Fig. 6a). The boundary between the topographically wellexpressed northwestern and the sediment-covered central segment of the Danube Shear Zone is defined by the Cham Fault (Fig. 6a). The southeasternmost segment of the Danube Shear Zone (locally also called "Aicha-Halser-Nebenpfahl") deviates from the course of the Danube River and separates a local basement high in the south from domain A1 in the north (Fig. 1b   365 and profile 8 in Fig. 6c).
The Cham Fault is also characterized by a well-defined topographic lineament (Fig. 7).  (Fig. 7). shear zones strike NW-SE/WNW-ESE and only represent a minor peak in the rose diagram (Fig. 8b). This The latter observation can be related to the technical representation of major fault zones in the analyzed geological maps, being expressed as wide zones of fault rocks rather than single lines (Teipel et al., 2008;Galadí-Enríquez et al., 2009b). Therefore, individual structural elements of the shear zones could not be included in the directional statistical analysis, which slightly biases the 385 relative proportions of fault orientations.   (Davis, 2002). Only lineaments outside of a 2 km buffer along the domain boundaries were considered in the analysis to avoid biases related to the trends and lengths of the major fault zones and associated lineaments. For comparison, 410 the rose diagram of all lineaments in the entire area is shown (not buffered). Note that the Bodenwöhr Trough in domain A2 and the Cham Basin in domain B (c.f., Fig. 56) have been excluded from the analysis due to their sedimentary infill. All rose diagrams have been lengthweighted and calculated using a bin size of 10°. Lineaments and faults have been segmented prior to statistical analysis.

Observations from low-temperature thermochronological data in the study area
We compiled thermochronological data from the literature that were obtained on apatite and zircon reflecting and reflect the 415 low-temperature history of the study area (Fig. 109). On either side of the Cham Fault, two Zircon Fission Track (ZFT) data points, located close to the village of Winklarn and ca. 2 km apart (Fig. 10b9b), record an apparent age gap of ca. 45 Myrs (ca. 260 Ma and 215 Ma, respectively, data from C.W. Naeser in Gebauer, 1984 andWagner et al., 1997). Similar ages of ca. 250 Ma have been reported from a quarry to the south of the Luhe Line (northeastern part of domain C2) and a sample located in the central-eastern part of domain A2 (data from C.W. Naeser in Gebauer, 1984). In contrast, younger apparent ZFT ages of 420 ca. 215 Ma are known from the Passau Forest, located close to the southern end margin of domain A1 outside the study area, and from the area adjacent to the Teplá-Barrandian Unit (Domain C1, data from C.W. Naeser in Gebauer, 1984).

Figure 9
Compilation of Apatite Fission Track (AFT) data from the literature (Wagner et al., 1989;Vercoutere, 1994;Siebel et al., 2010;Vamvaka et al., 2014).  Low-temperature Apatite Fission Track (AFT) data also show significant spatial differences in apparent ages across the Cham Fault, even among samples obtained at similar altitudes (Wagner et al., 1989;Vercoutere, 1994;Siebel et al., 2010;Vamvaka et al., 2014;Fig. 10a9a). An apparent age gap of ca. 40 to 50 Myrs is recorded between two clusters separated by the newly defined Cham Fault. In the area to the northwest west of the Cham Fault (domains A2 and C2), AFT apparent ages mainly cluster between ca. 150 and 200 Ma, whereas to the southeast east of the fault (domains A1, B, and C1), most AFT data record 435 apparent ages below ca. 100 Ma (Fig. 10a9a). No significant correlation between sampling elevation and ages is observed in the compiled AFT dataset. In addition, the data do not show a clear age gap across the Pfahl and Runding shear zones.

Discussion
By analyzing Bouguer gravity anomaly, topographic, and geological data, we identified three main basement domains in the southwestern Bohemian Massif. These domains are interpreted as individual basement blocks that were differentially uplifted 440 exhumed during and after the Llate Paleozoic Variscan Orogeny. Furthermore, the heterogeneous distribution of exposed granites and available thermochronological data show a second segmentation by the previously unknown crustal-scale Cham Fault. Below we discuss the timing and succession of the observed block segmentation and its implications for the tectonic framework along the southwestern Bohemian Massif.

445
The observed metamorphic units across the study area are interpreted as being part of a genetic sequence of thermally overprinted rocks, starting in domain C from intermediate-grade biotite-garnet-chlorite schists adjacent to the southern tip of the Teplá-Barrandian Unit over higher-grade biotite-plagioclase gneissesmica gneisses to partially molten (metatectic) gneissesmetatectic cordierite-sillimanite-K-feldspar gneisses in the southeast of domain C (Blümel, 1972). Hence, a NNE to SSW progressive Buchan-type metamorphic zoning from upper greenschist facies to anatexis under low to intermediate pres- 450 sures becomes apparent towards the north of the Pfahl and Runding shear zones (Read, 1952;Grauert et al., 1974;Winter, 2010). Consequently, the diatexites of domain A can be considered as successors in this sequence, where advanced partial melting of the metatectic progenitor precessors has led to the formation of nebulitic and schlieren structures or, in parts, even to a completely homogenized, granite-like rock texture (Brown, 1973;Rohrmüller et al., 1996;Wimmenauer and Bryhni, 2007;Chen and Grapes, 2007 Kyanite has not yet been observed in the anatectic rocks of the study area, thus bracketing pressure conditions to the andalusitesillimanite stability field. This is in line with estimates of pressure and temperature conditions, which, unfortunately, are largely restricted to the area north of the Pfahl Shear Zone. Here, early studies estimated metamorphic conditions of 2-4 kbar and 650-730 °C (Schreyer et al., 1964;Schreyer and Blümel, 1974;Blümel andSchreyer, 1976, 1977). Higher P/T conditions of up to 460 5-7 kbar and 800-850 °C are suggested for the cordierite-bearing migmatites in the central part of the study area (Kalt et al., 1999).
A strong correlation between metamorphic grades and rock densities is observed, resulting in distinct variations of Bouguer anomalies between the identified domains. Higher Higher-grade, diatectic domains thereby show more pronounced negative Bouguer anomalies compared to lower-grade, gneissic domains, an observation that is in line with preliminary results of 465 Schaarschmidt et al. (2019). Based on this relationship, domain B likely comprises rocks with a higher mean metamorphic grade ("diatectic gneisses") compared to the central part of domain A ("diatexites and gneisses"), important information that is not directly assessable from recent geological maps (Fig. 3). The distribution of diatectic gneisses in domain B and the presence of a distinct topographic lineament suggests a previously unknown extension of the southeastern tip of the Runding Shear Zone towards the south, terminating against the Pfahl Shear Zone (Figs. 3 and 6). 470 We suggest differential uplift vertical motion and exhumation of the crust along distinct, basement block-bounding fault zones as a the cause of the observed differences in exhumation and metamorphic grades (Fig. 1110). The generally higher metamorphic grades in domain A indicate a higher total amount of uplift exhumation southwest of the Pfahl Shear Zone, which is supported by variations in granite geochemistry across the fault (Grauert et al., 1974;Beer, 1981;Finger and Clemens, 1995;Siebel et al., 2008;Finger and Rene, 2009). In contrast, domain C, dominated by metatectic gneisses, is interpreted to have  A sharp contrast in the spatial distribution of exposed late Variscan granites from the northwest to the southeast is used as evidence for another level of segmentation along the Cham Fault. This faultThe Cham Fault separates areas exposing shallower crustal levels with less abundant granite exposures in the southeast (domains A1, B, and C1) from areas exposing deeper crustal levels and a higher number of granites in the northwest (domains A2 and C2, Fig. 3). From the Cham Fault to the southeast, 490 granite exposure gradually increases. This observation is interpreted as a counterclockwise block rotation to the east of the Cham Fault (domains A1, B, and C1, Fig. 1110). This rotation resulted in the exposure of deeper crustal levels in the very southeast of the study area, where both metamorphic grades and the amount of exposed late Variscan granites are similar to those observed in domains A2 and C2, respectively (e.g., diatexites in between the Fürstenstein and Hauzenberg composite massifs, Figs. 3 and 1110). 495 Exposed late Variscan granites in the center of the study area are aligned with the Pfahl, Runding, and Danube shear zones.
These tectonic structures appear to have guided magma ascent in this area, acting as low-pressure zones enabling magma transport to upper crustal levels as it is also observed, e.g., in the Alps and northeastern Brazil (Rosenberg, 2004;Weinberg et al., 2004). In the deeper subsurface, a nearly uniform distribution of granites is shown by the filtered Bouguer gravity anomaly data (Fig. 4), supporting early assumptions of Stettner (1975).  Fig. 3) can be explained by block rotation and differential upliftexhumation.

505
The topographic analysis highlights a close relationship between metamorphic grades and topographic relief. Higher-grade metamorphic rocks (i.e., diatexites) thereby correlate with lower elevations, whereas lower-grade metamorphic rocks (i.e., gneisses) generally correlate with higher elevations (Fig. 5). This correlation is especially evident in the southeastern part of domain A1, where the occurrence of granites and diatexites is associated with a pronounced decrease in altitude (Fig. 5b).
Hence, the observed correlation between topography and exposed rock type indicates, with that higher-grade metamorphic 510 rocks being are generally more prone to erosion compared to lower-grade metamorphic rocks. (Fig. 5). This relationship supports our interpretation of domain B comprising higher-grade metamorphic rocks compared to the central part of domain A (i.e., the northwestern part of domain A1), as it shows distinctly lower elevations compared to adjacent areas and thus catches the drainages of its surroundings. Nevertheless, it must be noted that also younger, late to post-Mesozoic differential tectonic block motion may represent an important factor controlling the recent distribution of elevated areas, as significant reactivation 515 of Variscan structures has taken place during this time (e.g., Kley and Voigt, 2008).
Lineaments mapped detected based on topographic data along the southwestern Bohemian Massif greatly exceed both numbers and lengths of faults known from geological maps. This suggests the presence of a considerable number of yet unidentified faults or fault segments. From the statistical analysis, two main lineament directions (I) NNW-SSE/N-S and (II) NW-SE/WNW-ESE have been identified (Fig. 8a). These results are in accordance with previous studies of the lineament inventory 520 along the southwestern Bohemian Massif (Lehrberger et al., 2003;Zeitlhöfler, 2007;Zeitlhöfler et al., 2015). These The observed directions correspond to faults known from geological maps and are in line with the prevailing trend of the large, late-Variscan fault zones, e.g., the Pfahl Shear Zone, Danube Shear Zone, and Keilberg Fault Zone (Fig. 8b). Therefore, a similar origin regarding timing and tectonic framework can be assumed for most of the mapped detected structures. A third, ca. E-W oriented direction in the lineament inventory originates from the proposed horsetail splay located in the southeast of the study 525 area. Such an E-W orientation has not yet been only rarely been observed in mapped faults.
The tectonic configuration of the southwestern Bohemian Massif during the late stages of the Variscan Orogeny is interpreted as a conjugate shear system related to N to -NNW directed shortening (Wallbrecher et al., 1991;Brandmayr et al., 1995;Peterek et al., 1996;Galadí-Enríquez et al., 2010). Under this tectonic regime, a Riedel-type fault interaction likely initiated the observed lineament patterns (Zeitlhöfler, 2007;Fig. 1211). In such a model, NW-SE to NNW-SSE striking lineaments On a regional scale, the general structural configuration of the entire study area is also interpreted to reflect Riedel-type fault interactions, which most likely initiated during late to post-Variscan wrench tectonics (Peterek et al., 1996;Peterek et al., 1997;Zeitlhöfler, 2007). The presence of topographic lineaments crossing Variscan tectonic structures, however, indicates a younger (re)activation phase along the southwestern Bohemian Massif, which was most likely initiated in the course of Cretaceous to 550 Cenozoic tectonic processes (Lehrberger et al., 2003).  (Vercoutere, 1994). 560 An even more complex thermotectonic record is suggested for the Bavarian Forest in the central and southeastern part of the study area (Vamvaka et al., 2014). Here, thermal models indicate a first reheating during Middle to Late Jurassic times (ca.

Thermochronological evidence for block faulting
160-140 Ma), which was followed by exhumation during the Early Cretaceous (ca. 140-120 Ma). After a phase of stagnation, sedimentation recurred during the Late Cretaceous (ca. 95-85 Ma), which caused reheating of marginal parts of the Bavarian Forest. The latter phase is especially depicted by the "Grub" sample, which is the only sample in the study of Vamvaka et al. 565 (2014) that is located to the west of the newly proposed Cham Fault (Fig. 9). Similar to the tectonic record of the Naab Mountains (Vercoutere, 1994), the final uplift phase of the Bavarian Forest was initiated in the Late Cretaceous (Vamvaka et al., 2014), probably in the course of inversion tectonics related either to the Alpine collision (e.g., Ziegler, 1987;Ziegler et al., 1995) or Africa-Iberia-Europe convergence (Kley and Voigt, 2008).
Hence, from this record, a complex regional thermotectonic evolution of the study area is proposed based on AFT data. Nev-570 ertheless, a significant difference in the thermal evolution between the sector to the west of the Cham Fault compared to the eastern sector is undoubtedly present, as evidenced by two pronounced age clusters at either side of the fault (Jurassic vs. Late Cretaceous, Fig. 9). Vamvaka et al. (2014) attributed this change to the presence of a fault zone in between the sample of Grub and the remaining samples to the east (i.e., the Cham Fault), accommodating a vertical displacement of at least 1 km. This interpretation is in line with Gebauer (1984), who proposes the presence of a significant tectonic structure close to the village 575 of Winklarn between domains C1 and C2, based on two ZFT data points located only 2 km apart that record an apparent age gap of ca. 45 Myrs (ca. 260 Ma and 215 Ma, respectively), which is again in accordance with the newly proposed Cham Fault.
A possible cause for the apparent differences in timing between the western sector (ZFT: Late Permian, AFT: Jurassic) and the eastern sector (ZFT: Late Triassic, AFT: Late Cretaceous) might be attributed to the NW-SE prograding Late Permian to Mesozoic depositional system (e.g., Meyer, 1989;Peterek et al., 1997;Schröder et al., 1997). Based on the above-mentioned 580 gap of ca. 45 Myrs between two ZFT data points close to the village of Winklarn, Meyer (1989) placed the basin margin during the Triassic close to this locality. From this interpretation, we conclude that large parts of the southeastern margin of the Mesozoic basin most likely have been controlled by the Cham Fault, which, in turn, resulted in a larger sedimentary cover and the partial resetting particularly of AFT ages towards the west of the Cham Fault. This interpretation is in accordance with the AFT data towards the east of the Cham Fault, where most sample sites are thought to have been covered only with an insig-585 nificant (<200 m) or, in parts, even completely absent sedimentary cover (Vamvaka et al., 2014). In fact, even the formation of syn-tectonic Permian basins along the northwestern segments of the Pfahl and Danube shear zones (Schröder, 1988;Peterek et al., 1996) might have been guided by the Cham Fault.
AFT data obtained to the northwest of the fault are thought to represent post-Variscan cooling ages that were partially reset during Mesozoic burial and re-heating (Vercoutere, 1994;Wagner et al., 1989;Hejl et al., 1997;Vamvaka et al., 2014). In 590 contrast, most of the AFT data southeast of the Cham Fault suggest ongoing, slow exhumation during the Jurassic and Cretaceous, although slight re-heating might also have occurred (Vamvaka et al., 2014).
Regardless of the true nature of the compiled FT ages (i.e., cooling ages or mixed ages), the fact that samples located on both AFT ages in post-Cretaceous times (Vamvaka et al., 2014).

Relative timing and succession of block segmentation events
ZFT data indicate unroofing of the western Bohemian Massif during Permian and Triassic times . This is supported by Permian syn-tectonic sedimentation in the Donaustauf Basin bounded by the northwestern segment of the Danube Shear Zone (Siebel et al., 2010). K-Ar and Rb-Sr illite ages thereby bracket uplift and erosion of the Variscan basement and 605 tectonic extension to a period prior to 255-266 Ma (Siebel et al., 2010).  Fig. 3). Together with a distinct mylonite zone within the Patersdorf Stock, this indicates an activity along the Pfahl and Runding shear zones postdating the emplacement of late Variscan granites (Siebel et al., 2006a;Büttner, 2007). This is further supported by quartz mineralizations within the mylonite zone ("Pfahl Quartz") of the Pfahl Shear Zone, indicating a Triassic reactivation of the shear zone (Horn et al., 1986). In contrast, numerous granites appear to have utilized the NW-SE Erosional products of late -orogenic granites deposited in adjacent Permian basins indicate rapid exhumation and 630 erosion during the Cham Phase shortly after Upper Late Carboniferous to early Permian granite emplacement Galadí-Enríquez et al., 2009a).
As it becomes apparent from the timing indicators for fault activity and block segmentation mentioned above, the Pfahl Phase cannot be considered as a single event but rather as a series of events, which, in total, resulted in a higher total amount of uplift 635 exhumation towards the southeast southwest of the Pfahl Shear Zone (domain A). Total vertical upliftexhumation during the Pfahl Phase must have amounted to less than the mean 14-18 km emplacement depth of the granites (Table 1). Otherwise, a homogeneous distribution of granites, as observed in domains A2 and C2 and in the deeper subsurface, would be expected across the surface of the entire study area (Fig. 4). In contrast, the Cham Phase is considered as post-magmatic and involved higher amounts of uplift exhumation in the northwest (domains A2 and C2) and, likely due to block rotation, in the southeast 640 of the study area. This segmentation phase resulted in the area-wide exposure of granite bodies at the Earth's surface, with total uplift exhumation and denudation during both segmentation phases (i.e., the Pfahl and Cham phases) exceeding the estimated granite emplacement depths of 14-18 km (Table 1).

Conclusions
Our integrated study using gravity anomaly and topographic data provides insights into the crustal architecture of the south-645 western Bohemian Massif. Domains of different metamorphic grades and/or exposed granite inventories are interpreted as individual crustal blocks that are bordered by distinct tectonic structures. With the Cham Fault, we introduce a previously unknown NNW-SSE striking tectonic structure that is of similar importance to other major fault zones, such as the Pfahl and Danube shear zones. We propose a model of differential exhumation and block rotation that led to the juxtaposition of contrasting lithological domains. Increasing amounts of uplift exhumation evoked the exposure of deeper crustal levels, as evi-650 denced by the presence of higher metamorphic grades and a higher percentage of granite bodies at the surface. Gravity anomaly filterings thereby indicate a rather homogeneous distribution of granites in the subsurface. This observation contrasts with the heterogeneous exposure of granites at the surface, suggesting that an important phase of segmentation and differential uplift exhumation must have occurred after granite emplacement. A post-Variscan activity of the Cham Fault is evidenced by abrupt changes of apparent ZFT and AFT ages across the tectonic structure. The fact that the Cham Fault also forms tectonic bound-655 aries of Cretaceous to Cenozoic geological features, such as the Bodenwöhr Trough, implies that it played a significant role in the tectonic evolution of the southwestern Bohemian Massif even during the younger geological past.
Our model of block segmentation and differential uplift exhumation along the southwestern Bohemian Massif emphasizes the significance of vertical displacements along distinct tectonic structures to explain the observed complex lithological configuration in the study area. The example of the newly discovered Cham Fault thereby suggests that potentially several more yet 660 unidentified tectonic structures might exist in that area. To precisely reconstruct the timing and succession of block segmentation events along the discussed faults and to quantify the amounts of uplift exhumation of the five outlined basement blocks, however, additional data on granite intrusion depths, P-T metamorphic conditions, and thermochronology are needed.

665
Data availability. The high-resolution topographic data used in this study are available by request at the Bavarian Agency for Digitisation, High-Speed Internet and Surveying. The Bouguer anomaly data are provided by the Leibniz Institute for Applied Geophysics (LIAG).

670
Author contribution. AE, HF, WB, and HS designed the study. AE prepared the manuscript and performed the analysis of the topographic, filtered gravity, granite density, and literature data. HF carried out the filtering of the gravity data and supervised the study. HS and HdW supervised and acquired the financial support for the project leading to this publication. GG provided the gravity data and supported their interpretation. All authors contributed to the reviewing of the manuscript and the discussion of the results.