Reconstruction of cyclic Mesozoic–Cenozoic stress development in SE Germany using fault-slip and stylolite inversion

Abstract The Franconian Platform of SE Germany and the underlying Permian and Triassic rocks that developed from latest Permian to Triassic time were affected by multiple compressional and extensional events that created a complex fracture, fault and stylolite network. We reconstructed the spatio-temporal variations of post-Triassic palaeostress fields in the Franconian Platform and Triassic strata using fault-slip and tectonic stylolite inversion. Our highly resolved stress inversion enables us to demonstrate a cyclic stress evolution from the stress regime of normal faulting to thrusting, strike-slip and back to normal faulting. Five main stress fields correlating with two stress cycles are determined for Late Jurassic to Cenozoic time. The first cycle consists of: (SF1) an initially NE–SW-directed horizontal extension during Late Jurassic to Early Cretaceous time; (SF2) NNE–SSW-directed horizontal compression with an early set of tectonic stylolites prior to the development of reverse and thrust faults; and (SF3) a strike-slip-dominated setting with (N)NE–(S)SW horizontal compression representing a first relaxation. The second cycle comprises (SF4) NW–SE-directed horizontal extension during Oligocene–Miocene time; and (SF5) a second strike-slip-dominated regime with WNW–ESE to NW–SE horizontal compression during the Alpine shortening, creating the youngest set of tectonic stylolites. In addition, we consider the transitional stages between thrusting and a strike-slip regime as a snapshot in the process of intraplate tectonics.


Introduction
A fold-and-thrust belt generally comprises thrust-folded sedimentary sequences and can include basement and older faults (Twiss & Moores, 1992). The structural life cycle of a typical fold-and-thrust belt is characterized by a series of deformational phases starting with the burial of sedimentary rocks under extension, followed by the inversion of the former basin in a sequence of strike-slip and thrusting (Tavani et al. 2015;Ferrill et al. 2021). Initial relaxation is represented by another strike-slip regime, followed by the final relaxation due to decreasing shortening and increasing extension (Ferrill et al. 2021). The temporal evolution of a fold-andthrust belt, in particular the changes between the deformation phases and stress regimes, is essential to understand the associated tectonic architecture and stress development in the hinterland.
Intraplate fold-and-thrust belts result from localized shortening of sedimentary basins distal from the plate boundary, such as the Tajik basins (e.g. Stübner et al. 2013) or the Sevier and Bighorn basins in the USA, and in North America (e.g. van der Pluijm et al. 1997;Beaudoin & Lacombe, 2018). A prominent example of intraplate shortening is the European Alpine Foreland (e.g. Ziegler et al. 1995;Cloetingh & van Wees, 2005). In this region, several palaeostress analyses reveal multiple alternating and overlapping phases of contractional and extensional deformation during the Mesozoic and Cenozoic times (Bergerat, 1987;Blés et al. 1989;Hibsch et al. 1995;Peterek et al. 1996;Kley & Voigt, 2008;Sippel et al. 2009Sippel et al. , 2010Coubal et al. 2015;Navabpour et al. 2017). Such phases of deformation were linked to the opening of the Neotethys during the Permian to Early Mesozoic time, the onset of Atlantic rifting, and the development of an active margin between the European and African plates .
Previous palaeostress reconstructions from the Franconian Platform in southern Germany, based on fault-slip analysis and subordinate stylolite stress inversion, reveal regionally and temporally differing stress fields (Peterek et al. 1996. Moreover, geo-and thermochronological studies indicate the existence of several intervening extensional phases during the Late Cretaceous and Cenozoic convergence between Europe and Africa (Abratis et al. 2007;von Eynatten et al. 2019von Eynatten et al. , 2021. So far, however, the stress field evolution was not correlated with neighbouring areas nor was it placed in the context of major tectonic events affecting the area, e.g. Late Cretaceous inversion and subsequent Alpine shortening (Bergerat & Geyssant, 1982;Peterek et al. 1996Peterek et al. , 1997.
We aim to extend the interpretation of the tectonic history of Central Europe by increasing the temporal and spatial resolution of the palaeostress evolution in the Mesozoic sequences west of the Franconian fault zone compared to previous work by using fault-slip and stylolite stress inversion.
With our work we contribute to the understanding of the stress development and the development of intra-continental areas in close vicinity to mountain chains. In addition, the high local resolution allows for identifying local stress perturbations in predicted stress fields.

Geological setting
The Franconian Platform is situated immediately west of the Bohemian Massif (Fig. 1c) and consists of Mesozoic and Cenozoic sedimentary rocks (Freudenberger & Schwerd, 1996). The rocks of the Variscides (Kossmat, 1927) exposed in the Bohemian Massif are buried in the west beneath the Kraichgau Basin fill and the sediments of the Franconian Platform (Paul & Schröder, 2012;Sittig, 2012;Kämmlein et al. 2020;Fazlikhani et al. 2022). The structural framework at the eastern basin margin is dominated by a NW-SE-striking fault network, the most prominent elements of which are the Franconian and the Eisfeld-Kulmbach fault zones.
The Variscan basement was affected by Late Variscan NNW-SSE shortening resulting in final folding and the activation of NE-SW-and NW-SE-striking sinistral and dextral strike-slip systems (Kroner et al. 2007;Büttner, 2012;Stephan et al. 2016). Beginning in the latest Carboniferous, predominant extensional tectonics associated with the incipient break-up of Pangaea (Kroner & Romer, 2013) led to fault reactivation and formation of structures that follow the strike of pre-existent basin structures and faults . These latest to post-Variscan tectonics initiated subsidence and the development of WNW-and Nto NE-striking graben systems in Central Europe .
Triassic deposits in the study area are fairly tabular with local lateral thickness changes and limited faulting, indicating continuous regional tectonic quiescence (Fig. 1d;Fazlikhani et al. 2022).
During the Late Jurassic to Early Cretaceous time (75-55 Ma) the study area was dominated by regional uplift and subsidence to the west of the Bohemian Massif associated with NE-SW-directed horizontal extension (Peterek et al. 1996Scheck-Wenderoth et al. 2008;von Eynatten et al. 2021). Late Jurassic sedimentation is controlled by the transgression of the Tethys Ocean from the south which produced a massive carbonate platform (Meyer & Schmidt-Kaler, 1990). For most parts of Central Europe, including large parts of SE Germany, normal faults bounding NW-SE striking graben systems were active at that time, e.g. grabens of the South German Basin Zulauf & Duyster, 1997;Walter, 2007;Kley & Voigt, 2008;Scheck-Wenderoth et al. 2008;Sippel et al. 2009;Navabpour et al. 2017). According to Navabpour et al. (2017) this phase of normal faulting is also recorded in Middle Triassic rocks in the Thuringian Basin towards the north.
During Late Cretaceous time, the convergence between Iberia-Africa and Europe caused widespread uplift in Central Europe (Kley & Voigt, 2008;Dielforder et al. 2019). Increasing thicknesses and coarsening trends of conglomerates in Lower and Upper Cretaceous units from the (S)W towards (N)E (Hejl et al. 1997;Tanner et al. 1998;Niebuhr et al. 2014) suggest the reactivation of ten to hundreds of kilometres long NW-striking faults (e.g. Pfahl shear zone, Franconian fault zone, Danube shear zone; Schröder, 1987;Scheck-Wenderoth et al. 2008). This intraplate compression, together with lithospheric buckling, i.e. long wavelengths and low amplitudes, affected large parts of the South German Basin . It lasted until Palaeocene time (Voigt et al. 2021) and was accompanied by the inversion of pre-existing grabens (Schröder, 1987;Scheck-Wenderoth et al. 2008) and reactivation of NW-trending fault zones Tanner et al. 1998). Comparing the timing of tectonic activity in individual regions reveals that the inferred relative-chronological order of reverse and strike-slip faulting suggested for Late Cretaceous time does not coincide, even across directly neighbouring areas (Bergerat & Geyssant, 1982;Peterek et al. 1996;Sippel et al. 2009;Navabpour et al. 2017). From the Franconian Platform, for instance, the relative order of thrusting and strike-slip faulting was not possible so far (Peterek et al. 1996) while in the Thuringian Basin farther north, strike-slip movements occurred prior to reverse faulting (Navabpour et al. 2017).
From Eocene to late Oligocene time there is no direct structural evidence for tectonic activity . In Oligocene and early Miocene time, volcanism penetrates the Franconian Platform marked by the 19-24 Ma Oberpfalz volcanics to the east (c. 50 km to the city of Bayreuth) (Todt & Lippolt, 1975) and the c. 31 Ma ultramafic Oberleinleiter volcanics ESE (c. 15 km from the city of Bamberg) (Hofbauer, 2008). Since Eocene time, contemporaneously forming grabens of the European Cenozoic Rift System (ECRIS) record a variety of extensional directions with the WNW-ESE-directed opening of the NNE-SSW-striking Upper Rhine Graben providing the most prominent example (Ziegler, 1992). The NE-SW-striking Eger Rift, however, initially opened NNE-SSW to N-S and then became overprinted by later NW-SE opening (Ziegler, 1992;Schröder et al. 1997;Abratis et al. 2007;Rajchl et al. 2009).
Collision between Europe and Adria-Africa culminated in the formation of the European Alps and triggered intraplate shortening to the north (Rosenbaum et al. 2002;Scheck-Wenderoth et al. 2008;Glotzbach et al. 2010). Dèzes et al. (2004) propose a correlation of contemporaneous volcanism, the opening of the European Cenozoic Rift System (ECRIS) and N-directed shortening induced by the Alpine Orogeny. In the North Alpine Foreland Basin, the increasing tectonic load of the northward-propagating Alpine thrusts led to subsidence and bulging of the lower plate associated with a normal faulting regime in the extensional area of the bulge (von Hartmann et al. 2016). Thus the Franconian Platform was affected by the opening of the Eger Rift towards the east and the Alpine collision towards the south (Adamovič & Coubal, 1999;Reicherter et al. 2008;Rajchl et al. 2009). Towards the east of our study area, historical and recent earthquakes register ongoing seismic activity (Wilde-Piórko et al. 2006), and mofettes and thermal springs imply deep-reaching zones of structural permeability extending from the Eger Rift across the Franconian fault zone (e.g. Heinicke et al. 2019). In the research borehole Lindau 1, near Bayreuth (Fig. 1a), the mean direction of the maximum present-day horizontal stress, derived from borehole breakouts and from hydraulic stimulation experiments, is 135°and 138°(NW-SE), respectively (Röckel & Wonik, 2006). This differs only slightly from results of 2324 S Köhler et al. Reicherter et al. (2008) and Heidbach et al. (2016), who claim that the present-day stress field in our study area is dominated by N(N) W-S(S)E-directed compression.

Methods
In order to reconstruct palaeostress fields, we combine two methods, fault-slip analysis (Angelier, 1984) and stylolite stress inversion (Koehn et al. 2012;Beaudoin et al. 2016). In addition to faults and stylolites (e.g. Figs 2a and 3b), we use related structures such as folds (e.g. Fig. 2c), joints and veins to derive relative age relationships from field studies. We apply the term fracture for brittle deformation structures (e.g. joints) without fault planes on the metre scale in outcrops, whereas the term fault (F) is used for fault planes on metre scale as well as in outcrop scale (50 to 1000 m). We apply the term fault zone (FZ) for fault planes on map scale (1 to 100 km). A fault zone may include a set of sub-parallel faults with similar kinematics. We assume that large single faults that would fit this definition form an exception in our study area. Coordinates, lithologies and stratigraphic positions of all field measurement locations are summarized in Table 1   Platform, with group 4 consisting of layered limestone and group 5 consisting of massive reef dolomite. Group 7 (Nuremberg) is distinguished from the neighbouring locations of groups 6 and 10 (Erlangen and Hersbruck, respectively) due to its location south of the W-E-striking Hersbruck FZ and because it has a high amount of stylolites with a low amount of fault planes (see Section 4a below and Fig. 4). Group 8 is defined by the location of the outcrops in the south of the study area with a large extension in E-W direction. Group 9 includes outcrops of Upper Triassic (Keuper) sedimentary rocks (Fig. 2a). Tectonic stylolites are rough surfaces formed by pressure solution (Park & Schot, 1968). They are associated with layerparallel shortening tangential to the bedding. Stylolite teeth grow parallel to the largest principal stress σ1 (Nitecki, 1962;Koehn et al. 2007). Thus, a stylolite plane with orthogonal teeth is perpendicular to the σ1 direction. However, there is a continuous transition from stylolites associated with layer-parallel shortening through stylolites with tilted planes and oblique teeth and slickolites that develop on fault planes where teeth still preserve the direction of the σ1 (Nitecki, 1962;Koehn et al. 2007Koehn et al. , 2012Toussaint et al. 2018). The intermediate and smallest principal stresses σ2 and σ3 lay in the stylolite plane. Field observations, however, do not allow determination of their exact orientations (Schmittbuhl et al. 2004;Ebner et al. 2010).
Here, we recorded tectonic stylolites as planar features with dip angle and dip direction together with the azimuth of the stylolite teeth as linear features.
Measured faults are categorized into three groups according to the quality of the measurement and the reliability of their sense of movement indicators (sides of steep teeth) (Sippel et al. 2009;Sperner & Zweigel, 2010). We measured a total of 546 faults and 432 tectonic stylolites at 28 locations (Table 1; Figs 5, 6); all data are available online via https://doi.pangaea.de/10.1594/ PANGAEA.929490 (Köhler et al. 2021).
The stress history based on fault-slip inversion (PBT axes method) was reconstructed using the software Tectonics FP (Reiter & Acs, 2020). The PBT method is based on the assumption that fault planes develop at an angle of 30°to σ1 (Anderson, 1972). This is in agreement with the observation that conjugate fault planes form angles of c. 60°. For each fault-slip datum the respective P-(compressional axis, σ1), B-(neutral axis, σ2) and T-(tensional axis, σ3) were constructed graphically (Turner, 1953;Ortner et al. 2002). Following Sippel et al. (2009), we subdivided the calculated data into distinct homogeneous clusters of PBT axes orientations. Following Wallbrecher (1986), Tectonics FP calculates the mean vectors and concentration parameter R% (Ortner et al. 2002). We excluded faults which do not fall into a cluster with at least ten data points and with a minimum R% value of 90 %. We constructed clusters for (i) the entire working area, defining superordinate stress fields, and (ii) the aforementioned outcrop groups to analyse local variations in stress orientation, and compared these to mapped faults. To perform the palaeostress inversion at local resolution, the locations were split into ten groups as explained before (Table 1). A stress field is defined by spatial uniformity or variability of a certain aspect of the stress tensor that persisted over a certain time in the geological past. Here, we use the following terms to describe a stress field: The magnitude of the vertical stress (σ v ) is the integral of the weight of the overburden. Only at the Earth's surface is σ v a principal stress axis of the stress tensor whereas σ v can deviate from a principal stress orientation at greater depths. When σ v is a principal stress axis, the maximum and minimum horizontal stresses (σ Hmax and σ hmin ) are the other two principal stresses. Otherwise, σ Hmax and σ hmin are the projections of the principal stresses into the horizontal plane. The stress regime is considered as the expression of the relative magnitudes of the principal stresses. Tectonic regimes are termed 'normal faulting' when σ v > σ Hmax > σ hmin ; 'thrust faulting' when σ Hmax > σ hmin > σ v ; and 'strike-slip' when σ Hmax > σ v > σ hmin (cf. Zoback & Zoback, 1989). It is worth noting that only when faults are optimally oriented in the stress field does the stress regime coincide with the tectonic regime.
After the clustering into stress fields, we calculated shifts between (i) the superordinate stress fields and (ii) the local stress Fig. 4. Single measurements of fault planes and fault lineation (if applicable) and their relative chronology (based on cross-cuttings and multiple striations), plotted as stereographic projection, lower hemisphere. Kinematic directions of the single measurements can differ from the overall stress fields. Boxes beneath projections and on the right show the schematic direction of the highest (inward arrows) and lowest (outward arrows) horizontal stresses (map view, against north) in their relative chronological order. Colours refer to the stress field. See Table 1 for further details on outcrop groups.
fields. For simplification we assume that two principal stresses of the stress tensor are oriented in a horizontal plane according to Anderson's theory (Anderson, 1972). Variation is then calculated as the azimuthal difference in degree between (i) and (ii), resulting in a clockwise or anticlockwise shift.
The use of fault-slip data for stress inversion is applicable, if the following assumptions outlined by Anderson (1972) and Sperner and Zweigel (2010) apply: (1) measurements of the slip data in small outcrops are representative of the far-field stress; (2) rotation of fault-bounded blocks (100 m to km scale) has not distorted the stress marker significantly; (3) the material is homogeneous enough to allow retrieval of the incremental strain that can then be transferred to the stress through stress inversion; (4) different stress fields can be separated by the data; and (5) the orientation  Table 1 for further details on outcrop groups, individual locations and their coordinates, lithologies and stratigraphy.
of σ1 relative to the fault's dip is 30°at close-to-surface conditions, which is the case in most situations in the field where we can observe conjugate fault sets.
To increase validation and representability (assumption 1), we use a combination of two independent methods, i.e. fault-slip and tectonic stylolites. Using fault-slip inversion alone entails the risk of measuring artefacts of pre-existing fault zones. Those can produce perturbation of the applied stresses and therefore result in a heterogeneous stress pattern in the overall study area (Lacombe, 2012). We will discuss the validation of assumption 2, the negligibility of block rotation, in Section 5.c below. The Upper Jurassic carbonate rocks essentially comprise two lithofacies: (i) the dominant so-called 'normal facies', composed of well-bedded limestones and marls and (ii) the 'reef facies', build-ups formed by sponges and microbial crusts (algal-sponge reefs) and locally also by corals (Munk, 1980). The dominant 'normal facies' rocks are relatively homogeneous in terms of their rheology. We assume that inhomogeneities associated with the visible layering do not have a strong effect at the outcrop or map scale especially where two principal stresses lay in the bedding plane. Thus, assumption 3 is valid, too. The dolomites of the 'reef facies', however, provide an exception (outcrop group 5, Reef Facies) as these are massive and strongly karstificated. Compared to the other groups, fault-slip indicators are poorly preserved in outcrops of this group and therefore the number of measurements is low. Assumption 4, the separation of multiple stress fields, is valid for most of the data that fall clearly into distinct stress fields that can be separated. Data that cannot be statistically clustered are neglected in this study. Assumption 5 is valid because locally slickolites record directly the orientation of σ1 relative to fault planes, the angle between the slickolite and the fault plane is 30°on average and conjugate fault planes form an angle of 60°. We used only faults with Fig. 6. Geological map showing the distribution of tectonic stylolites as density plots (stereographic projection, lower hemisphere) for the individual outcrop groups and for the entire study area (top right). See Figure 1 for key to colours. See Table 1 for further details on outcrop groups. preserved striation and slip direction for our analysis. The risk of missing minor structures or subjective interpretation of slip direction (Sperner & Zweigel, 2010) is minimized by the contributions of four persons to the final dataset (Köhler et al. 2021). All contributors are aware of the importance of distinguishing between slickenfibres, slickolites, stylolites and joints (Lisle, 2013). No outcrop was analysed by one single person.

4.a. Stress fields and their relative age relationships
We identified five different stress field generations, causing brittle deformation, folding and the development of tectonic stylolites in the study area. The relative age relationships of these stress fields were determined through local observations of cross-cutting relations of the respective structures and multiple striations (Figs 3, 4), yielding the following chronology (from old to young) of successive stress fields (SF): We correlate these SFs to the overall SFs (Section 4.c below). However, the observations yield very local information, and thus the orientation of the SFs can differ from the regional observations. Normal and reverse sense striations were preserved on the same W-E-to NW-SE-striking fault planes ( Fig. 4a-b, e1, e2), resulting from N(NE)-S(SW) normal faulting regime (SF1) followed by NE-SWthrusting regime (SF2), respectively. Due to the stratigraphic age of the analysed outcrops (Table 1), the maximum age of SF1 is Late Jurassic (Malm). SF2 is associated with tectonic stylolites indicating NE-SW compression that are rotated by SW-vergent folds (Fig. 2c). We observed cross-cutting relationships between those stylolites and dextral NW-SE-striking faults (Figs 3b, 4g1) that are indicative for a younger strike-slip regime with a roughly N-trending σ1 and Etrending σ3 (SF3). In addition, SF3 strike-slip faults are cutting straight through SF2-related folds. At three locations we observed overprinting of SF3 by N(N)W-S(S)E normal faults associated with SF4. While both SF2 stylolite and SF3 dextral faults are cross-cut by NW-dipping normal faults (SF4) in group 2 (outcrop 2, Kirchleus in Figs 3a, 4g1), SF3 strike-slip faults are reactivated as SF4 normal faults in outcrops 10 and 8b (Figs 3c, 4d, e3).
In addition, SF2 thrusts and reverse faults are overprinted by oblique, NW-striking normal faulting (Fig. 3a) and cross-cut by dextral NE-trending faults (Figs 3b, 4g2), both resulting from a strike-slip regime with an E-trending, horizontal compression (SF3). The younger set of tectonic stylolites show a (W)NW-(E) SE horizontal compression that is related to SF5. These stylolites are cross-cut by dextral WNW-trending fault planes (Fig. 4g3). While these youngest faults are associated with a strike-slip regime under E-W horizontal compression, the younger stylolites could be related to either thrusting (i.e. vertical σ3) or strike-slip (i.e. horizontal σ3; Ebner et al. 2010). The stress field related to the youngest set of stylolites resembles the one associated with the youngest set of strike-slip faults (SF5). However, there is a lack of clear evidence of age relationships between SF4 and SF5. Unambiguous cross-cutting relationship between structures correlating with SF4 and SF5 have not been observed.

4.b. Stylolites and folds
Tectonic stylolites are well developed in Upper Jurassic (Malm) limestones, while their development in dolomites or in Middle Triassic (Muschelkalk) limestones is limited (groups 5 and 1, respectively). In sandstone-dominated lithologies we observed no stylolites at all (group 9). The preferred orientations of the measured stylolites are shown as density plots in Figures 5, 6. As stylolite teeth grow parallel to σ1, a maximum in the density plot corresponds to the direction of σ1. In the Franconian North outcrop group 1, rare tectonic stylolites suggest a bimodal distribution of their orientation, i.e. the coexistence of NNE-SSW-and W-Edirected teeth (Figs 5, 6). Orientation of stylolite teeth in the Kulmbach area varies from bimodal, NNE-SSW-and NW-SEdirected teeth (Kulmbach North, group 2) to unimodal NW-SE-directed teeth (Kulmbach South, group 3) (Figs 5, 6). In the hanging wall of the NE-dipping Eisfeld-Kulmbach FZ (Figs 6 and (further below) 8, outcrop 3b) oblique tectonic stylolites are observed in Muschelkalk limestones of an overturned fold limb (Fig. 2c). Throughout the entire Kulmbach area, fold axes trend NNW-SSE ((further below) Fig. 8). In Central Franconia (group 4) tectonic stylolites show a bimodal distribution in orientation, with the most dominant maximum indicating horizontal compression in the NE-SW direction, and a minor maximum in a NW-SE direction (Fig. 5). Around Erlangen (group 6), tectonic stylolites record NW-SE-and NE-SW-directed horizontal compression (Figs 5, 6, group 6). In Nuremberg (group 7), tectonic stylolites record only a single maximum revealing NNE-SSW horizontal compression (Figs 5, 6). In Franconia South (group 8) the majority of stylolite teeth trend NE-SW and a smaller population records the WNW-ESE horizontal compression (Figs 5, 6). In Hersbruck (group 10), stylolite teeth predominantly trend NNE-SSW, with a minority trending NW-SE.

4.c. Overall stress field and local deviations
For each stress field and each outcrop group we created beachball plots, with the largest (σ1) and the lowest (σ3) principal stress axis in the centre of the white and grey quadrants respectively. The intersection of the quadrants gives the orientation of σ2. To visualize local variations in the respective stress fields we compiled the beachball plots on geological maps for each stress field (Figs 7-11). The local shift of the stress directions with respect to the regional stress orientation is collected in Tables 2-6.

4.c.2. Thrusting regime under NNE-SSW-to NE-SW-directed horizontal compression (SF2)
The second stress field (SF2) is compressive with a NNE-SSW σ Hmax trend as obtained from 20 WNW-ESE-to NW-SE-striking thrust faults. Striations show that fault planes of SF1 were reactivated as reverse faults during SF2 (Fig. 4b, e1-2), e.g. the Hersbruck FZ (Fig. 8). Our stress inversion yields a horizontal orientation of 202/01 for σ1 (i.e. σ1 = σ Hmax ) with an R-value of 97 %, and a subvertical orientation of 299/85 for σ3 (= σ v ) with an R-value of 96 %. We also observed several SW-vergent folds in Muschelkalk units with an overturned, steep limb in the hanging wall of the Eisfeld-Kulmbach FZ (Fig. 2c). These folds rotated some tectonic stylolites, indicating that layer-parallel shortening predated SF2 thrusting and folding. Further west in the study area, several open folds developed in Central Franconia (Fig. 2b). The density distribution of their tectonic stylolite teeth (Fig. 6) shows a maximum at 205°, i.e. parallel to the orientation of σ Hmax derived from fault-slip analysis (Fig. 8). Locally the σ Hmax direction may vary slightly, especially proximal to the Eisfeld-Kulmbach FZ. At Kulmbach South (group 3) for instance, NE-SW-directed tectonic stylolites and NW-SE-striking thrusts are not observed whereas at Kulmbach North (group 2), NW-SE-striking thrusts and NE-SW-trending tectonic stylolite teeth are common (Figs 4-6). SF2 is the most consistent stress field in our study area in terms of its orientation, with a minor scatter of ±4°around the mean orientation of σ Hmax (Table 3).  Figure 1 for key to colours and Table 1 for further details on outcrop groups.

4.c.3. Strike-slip regime under NE-SW horizontal compression and NW-SE horizontal extension (SF3)
SF3 is inferred from 25 strike-slip faults, of which 17 strike NNE-SSW to NE-SW with dextral slip sense and 8 strike E-W with a sinistral sense of slip. With an R-value of 92 %, σ1 is oriented horizontally with an azimuth of 46°(σ1 = σ Hmax ). σ3 is horizontal with an orientation of 315/05 and an R-value of 94 % (σ3 = σ hmin ). Locally, reactivated faults show striations indicating successive normal faulting, reverse faulting and strike-slip. The strike-slip faults associated with this stress field are the most abundant faults (Fig. 3b) in the study area and often extend across whole quarries (100 m to km scale). Some mineralized extensional veins oriented 115/85 and 100/85 were reactivated as dextral strike-slip faults by this stress field. There are also WNW-ESE-striking fault planes with sub-horizontal striations, implying a NNE-SSW-directed horizontal σ1 and an oblique, non-vertical σ2 and σ3, e.g. in group 3 (Fig. 12b, Section 5c below). The orientation of SF3-related faults and the respective σ Hmax and σ hmin orientations vary across the region. For example, at location 10 the strike of sinistral faults is sub-parallel to the E-W-striking Hersbruck FZ (Fig. 9), whereas in outcrop 2 faults strike NE-SW (Figs 9, 12c). In Kulmbach North the horizontal principal stress axes σ1 and σ3 vary with 34°and 33°anticlockwise from the superordinate stress field recording NNE-SSW compression (i.e. σ Hmax ) and WNW-ESE extension (σ hmin ). In contrast, in group  Figure 1 for key to colours and Table 1 (Table 4).

4.c.4. Normal fault regime under NW-SE-directed horizontal extension (SF4)
NW-SE-directed extension is identified from 14 predominantly SE-dipping normal faults. The associated conjugate set of NWdipping faults occurs less frequently and is restricted to the Erlangen outcrop group (group 6; Fig. 5). Fault-slip inversion leads to a (sub-)vertical σ1 orientation of 215/87 (σ1 = σ v ), within an R-value of 90 %. σ3 is oriented horizontally, with 137/01 within an R-value of 90 % (σ3 = σ hmin ). Due to the poor preservation of this stress field over the entire working area, we were only able to calculate local stress fields for four outcrop groups (Fig. 10).
Therefore, the activity of fault zones at larger scales is only assumed. Relative to the overall mean direction, measured SF4 σ hmin directions vary from 15°anticlockwise (ESE-WNW) in group 9, to 10°clockwise ((S)SE-(N)NW) in group 10 (Table 5).

4.c.5. Strike-slip regime under NW-SE horizontal compression and NE-SW horizontal extension (SF5)
Nine ENE-WSW-striking dextral strike-slip faults and 11 NNW-SSE-striking sinistral strike-slip faults are included in subset SF5. σ1 (126/03, σ1 = σ Hmax ) and σ3 (216/00, σ3 = σ hmin ) are horizontal with an R-value of 93 % and 91 %, respectively. This orientation of σ1 is in agreement with the minor maximum of tectonic stylolites (Fig. 5). This configuration indicates a strike-slip regime with horizontal compression in a NW-SE direction of σ Hmax and NE-SW trending σ hmin . As derived from measured strike-slip and oblique  Figure 1 for key to colours and Table 1 for further details on outcrop groups. normal faults (Fig. 3a), SF5 is characterized by a high variability in fault orientations relative to the principal stress axes (Figs 5, 11). In addition, the direction of σ Hmax varies considerably across the study area. For example, at Franconia North (group 1) and Kulmbach North (group 2), σ Hmax deviates clockwise by 21°and 43°from its average orientation and σ hmin varies with 30°and 47°clockwise (Table 5). Both deviations show a strong local shift from the dominating (E)SE-(W)NW-trending σ Hmax towards a more SSE-NNW-directed σ Hmax . Directions of associated tectonic stylolites vary as well (Figs 5,11), with stylolites in Kulmbach South (group 3) indicating horizontal compression in a NW-SE direction, whereas fault-slip inversion shows a horizontal compression in the WNW-ESE direction. In contrast, Kulmbach North (group 2) reveals the opposite trend, with stylolites recording WNW-ESE horizontal shortening, while fault-slip inversion suggests NNW-SSE shortening. This stress field reactivated the Cretaceous strike-slip faults with a reverse slip sense, e.g. faults in the Eisfeld-Kulmbach FZ and in the Hersbruck FZ.

4.d. Oblique stress field
In addition to the main stress fields 1-5, our data indicate the existence of an oblique stress field, i.e. there is no vertical principal stress axis. Data separation for PBT analysis of group 2 (outcrop Kirchleus) led to a homogeneous cluster where σ1 is oriented horizontally along the NE-SW direction, but neither σ2 nor σ3 is vertical (Fig. 12b). Figure 12a and c illustrate clusters of SF2-and SF-3 related data for the same outcrop. Bedding is (sub-)horizontal and σ1 is parallel for all three stress fields, which differ only in the orientations of σ2 and σ3. Fig. 10. Beachball plots illustrating regional variability of SF4 and prevailing NW-SE to N-S extension. Straight black lines show active, dashed lines probably active faults during SF4. Top right: beachball plot and stereographic projection (lower hemisphere) of SF4-related faults with slip sense. See Figure 1 for key to colours and Table 1 for further details on outcrop groups.

Discussion
5.a. Timing of stress fields and correlation with over-regional structures on a Central European scale

5.a.1. Late Jurassic -Late Cretaceous NE-SW extension (SF1)
The oldest recorded stress field corresponds to a normal faulting regime with NE-SW-directed horizontal extension. The maximum age of this stress field is Late Jurassic due to the Malm stratigraphy of the youngest host rocks. The existence of a normal faulting regime prior to the Late Cretaceous inversion is in agreement with palaeostress results of other authors for the same area and adjacent regions (Bergerat & Geyssant, 1982;Bergerat et al. 1992;Peterek et al. 1996Peterek et al. , 1997. NE-SWdirected normal faulting is also recorded from Middle Triassic units in the adjacent Thuringian Basin (Navabpour et al. 2017). Because this stress field is no longer recorded in Upper Cretaceous sequences in the Elbe zone farther north (Coubal et al. 2015), its minimum age is Late Cretaceous. The timing of the stress field coincides with the tectonic evolution of the Central European Basin System (CEBS) that records sediment accumulation and subsidence from Late Permian to Late Cretaceous time Stollhofen et al. 2008;von Eynatten et al. 2021). In the southern part of the CEBS, subsidence and extension have been related to the Fig. 11. Beachball plots illustrating regional variability of SF5 and prevailing NW-SE compression with NE-SW extension. Straight black lines show active, dashed lines probably active faults during SF5. Top right: beachball plot and stereographic projection (lower hemisphere) of SF5-related faults with slip sense. Teeth direction of tectonic stylolites is shown as yellow dashes. Abbreviations: EKFZ -Eisfeld-Kulmbach fault zone; HFZ -Hersbruck fault zone. See Figure 1 for key to colours and Table 1 for further details on outcrop groups.

5.a.2. Late Cretaceous NE-SW compressive phases (SF2 and SF3)
The folding and thrusting regime SF2 reactivated SF1-related structures. (N)NE-(S)SW-directed shortening is observed across the whole of western-central Europe (Kley & Voigt, 2008), and associated with lithospheric folding (e.g. Cloetingh & van Wees, 2005), inversion of Permian-Cretaceous basins and basement exhumation (Thomson & Zeh, 2000;von Eynatten et al. 2019von Eynatten et al. , 2021. Based on stratigraphic and thermochronological constraints, the timing of this compressional phase is bracketed between c. 95 Ma and 75 Ma (Voigt et al. 2021). Viewed in a broader geodynamic context, the thrusting regime in western-central Europe is triggered by the onset of the convergence between Africa-Iberia-Europe (Kley & Voigt, 2008;Dielforder et al. 2019). In our study area strike-slip faults cross-cut stylolites, faults and folds related to SF2. Thereby, the direction of maximum horizontal compression remains constant and the horizontal extension increases as shown by the change from a thrusting to a strike-slip regime under persisting orientation of σ Hmax (Fig. 5). Thus, we assume that SF3 established shortly after SF2 and may represent the final stage of this compressional phase. This compressive phase started in the Franconian Platform with layer-parallel shortening and the development of tectonic stylolites, followed by folding and thrusting and eventually strike-slip faulting.
This chronological order is in agreement with palaeostress analysis from the Elbe Zone in the northeast (Coubal et al. 2015). There, volcanic dykes emplaced at 80-61 Ma (Ulrych et al. 2014) are related to a strike-slip regime postdating thrusting (Coubal et al. 2015). In the Thuringian Basin to the north, however, strike-slip faulting predates the thrusting regime (Navabpour et al. 2017), and in northern Germany the relative chronological order of the strike-slip and thrusting regime is not clearly resolved (Sippel et al. 2009). 5.a.3. Late Palaeogeneearly Neogene NW-SE extension (SF4) Stress field generation SF4 describes a normal faulting regime with σ hmin trending NW-SE. This phase of extension with varying directions of σ hmin is also recorded from the Thuringian Forest (E-W-directed extension; Rauche & Franzke, 1990); from the Thuringian Basin (WNW-ESE-directed extension; Navabpour et al. 2017); from the Bohemian Massif (WNW-ESE-directed extension; Coubal et al. 2015); from Northern Germany (radial extension, i.e. NW-SE-to NE-SW-directed; Sippel et al. 2009); from the Upper Rhine Graben (E-W-directed extension; Bergerat, 1987); and from the Lower Rhine Graben (NE-SWdirected extension; Vandycke, 2002). There are several interpretations of the reason for widespread intraplate extension and associated mafic volcanism in late Palaeogene to early Neogene time, such as asthenospheric melting in response to deeply rooted large-scale upwelling mantle plumes or small-scale diapiric upwelling (e.g. Hoernle et al. 1995;Wilson & Downes, 2006). There is, however, a temporal and genetic link to the onset of continental collisional tectonics, slab detachment and subsequent plate tectonic reconfigurations in the Alpine realm (e.g. Dèzes et al. 2004;Pfänder et al. 2018). Towards the south, normal faulting in the North Alpine Foreland Basin is interpreted to result from bulging, that is a consequence of the increasing thrust load of the Alpine Orogeny (von Hartmann et al. 2016).
As illustrated in Fig. 1a, the Franconian Platform records evidence of Cenozoic intraplate volcanism. A prominent example is the NNE-SSW-striking Heldburg dike swarm system. This system comprises two generations of dikes, the late Eocene to late Oligocene phase (38.0 Ma−25.4 Ma) and the Miocene phase (Abratis et al. 2007;Pfänder et al. 2018). Late Palaeogeneearly Neogene intraplate volcanism is also evidenced by the 19-24 Ma Oberpfalz (Todt & Lippolt, 1975) and the~31 Ma Oberleinleiter volcanism (Hofbauer, 2008;Fig. 1a). To the east of the Franconian fault zone even more pronounced extension-related volcanism is recorded from the NE-SW-striking Eger Graben which initially opened along a NNE-SSW to N-S direction during the late Eocene to latest Oligocene time, followed by NW-SE normal faulting during the early Miocene time (Adamovič & Coubal, 1999;Rajchl et al. 2009). Cenozoic intraplate volcanism was widespread in central Europe to the north of the Alpine front, e.g. the Upper Rhine Graben, Rhön, Vogelsberg (European Cenozoic Rift System; Dèzes et al. 2004). However, the high variability of the direction of σ hmin over the whole of Europe cannot be explained exclusively by one cause. The strike direction of the Heldburg dike swarm differs from the direction of extension for SF4. On the other hand, a link to bulging as observed in the North Alpine Foreland Basin (c. 100 km farther south) requires additional consideration. Thus, the source for the normal faulting regime in the Franconian Platform remains unclear.

5.a.4. NW-SE-directed compressive phase since the Miocene time (SF5)
The youngest stress field is a strike-slip regime associated with NW-SE shortening. However, as compiled in Fig. 4 and stated  in Section 4.a above, the relative chronology between SF4 and SF5 is not finally solved. SF5 is recorded by a high variability in the orientation of the active fault planes and in the orientation of stylolites. This youngest compressional phase where bedding-parallel shortening (e.g. development of stylolites) and thrusting were subsequently replaced by an intracontinental strike-slip regime is also described for the whole of Central Europe (Rosenbaum et al. 2002;Kley & Voigt, 2008;Scheck-Wenderoth et al. 2008;Coubal et al. 2015;Navabpour et al. 2017). Kley and Voigt (2008) correlate this stress field with the Miocene to recent episode of the Alpine Orogeny and associated NW-directed shortening.
According to Heidbach et al. (2016) seismic activity and borehole break-outs imply the persistence of SF5 since the late Miocene time, because the respective stylolites are oriented parallel to the direction of σ Hmax . We cannot exclude overlapping of stress fields 4 and 5. Figure 13 schematically illustrates the successive development from almost undeformed sedimentary rocks during SF1 (Fig. 13a) to a rather fractured upper crust recording the cumulative effects of SF1-5 deformations (Fig. 13e). Most probably fractures and stylolites led to mechanical anisotropies (Baud et al. 2016;Pfänder et al. 2018) and caused local deviation of plate collision induced stress.

5.b. Transitional stress fields
The stress field associated with NE-SW compression observed for Kirchleus (group 2, Franconian North; Section 4d above) is an exotic feature in our study area. Since there is no vertical principal stress axis, it represents a non-Andersonian behaviour (Anderson, 1972). Oblique stress fields could be the result of later, local tilting of preserved, older stress fields (SF2 or SF3) due to, e.g., folding or block rotation (Arboit et al. 2015;Navabpour et al. 2017). Tilting and folding, however, can be neglected due to the observation that the sedimentary layering is still horizontal.
Oblique stress fields can occur under upper crustal conditions as shown by recent studies (Lisle et al. 2006;Sippel et al. 2010;Lacombe, 2012;Beaudoin et al. 2016). For instance, if pre-existent structures or faults are not optimally oriented in the stress field, the reactivation leads to the partitioning of the stress tensor and/or the formation of oblique (transpressional or transtensional) fault kinematics (Sanderson & Marchini, 1984). Local perturbations are also promoted by the vicinity of fault zones (Homberg et al. 1997;Sippel et al. 2010;Lacombe, 2012;Beaudoin et al. 2016). In this case the outcrop might be affected by local stress deflection as it is located close to the EKFZ (e.g. Fig. 8).
There are three stress orientations with approximately the same orientation of σ1 (= σ Hmax ) and (i) a vertical σ3 (SF2, Fig. 12a), (ii) an oblique σ2 and σ3 (Fig. 12b) and (iii) a vertical σ2 (SF3, Fig. 12c) recorded in the same outcrop. We therefore argue that this stress field (ii) most likely records a transitional stage between (i) and (iii), and thus between SF2 and SF3. This suggests that the stress field transition from thrusting to strike-slip is only locally preserved in Franconia. It is worth noting that the transition of stress fields, i.e. the change of the relative magnitudes of the principal stresses, may produce faulting in uniaxial stress geometries when the magnitudes of two principal stress axes become equal (Fig. 12e). The change in the magnitude of σ v is supposed to be caused by thickening due to folding and thrusting (Dalmayrac & Molnar, 1981;Tavani et al. 2015;Ferrill et al. 2021). Thus, this observation can also represent snapshots of the oscillation between such different regimes during the transition as, e.g., shown by Beaudoin et al. (2016). The traces of the principal axis are shown in Fig. 12d.

5.c. Stress field consistency and perturbations
In some areas, e.g. Hersbruck (outcrop group 10), the orientation of local stress fields derived from fault-slip inversion deviates from that of the superordinate stress fields by up to 45° (Fig. 5; Tables 2-6). In those areas, however, the orientation of the maximum horizontal stress parallels the strike of large faults. For instance, σ2 in SF1 and SF2 at locality Hersbruck (group 10) parallels the trace of the E-W-striking Hersbruck FZ, whereas at Kulmbach North (group 2) and Kulmbach South (group 3) σ2 is oriented parallel to the NW-SE-striking Eisfeld-Kulmbach FZ (Figs 7, 8). In the strike-slip regimes of SF3 and SF5 the respective large faults are reactivated. SF3 reactivates the Hersbruck FZ in a sinistral sense of movement and the Eisfeld-Kulmbach FZ in a dextral sense of movement and vice versa for SF5. Another possible explanation for the difference in the stress orientation could be that the area of outcrop group 10 is located in the region of potential interaction between the tips of the PSZ and a minor fault parallel to the NNW-SSE-striking PSZ. According to Homberg et al. (1997Homberg et al. ( , 2004, stress perturbation near fault tips is significantly influenced by the respective fault and the magnitude of the applied stresses. The coincidence of the orientation of σ2 and the strike of nearby faults indicates that stress deflection due to the existence of structural heterogeneities (sensu Hudson & Cooling, 1988;Casas et al. 1992;Zhang et al. 1994;Bell, 1996;de Joussineau et al. 2003;Yale, 2003) plays a significant role in the Franconian Platform. However, the overall stress field is relatively constant. This leads to the validation of assumption 2, made in Section 3 above, the negligibility of block rotation. If perturbations may be induced by block rotation palaeostress studies need to be conducted in the area to the south in order to obtain a larger framework.

5.d. Cyclic stress fields in compressional regimes
Our stress inversion results from the Franconian Platform can be explained in the frame of fold-and-thrust belt cycles sensu Tavani et al. (2015) and Ferrill et al. (2021). For instance, SF1-SF3 represent a complete cycle, initiated by normal faulting due to fault-controlled subsidence, followed by the formation of tectonic stylolites and basin inversion by folding and thrusting. Due to the resulting thickness of the crust during SF2, the vertical stress increases while the orientation of σ Hmax persists (Fig. 12e).
Assuming that the magnitude of σ1 (σ Hmax ) also remained, the exclusive increase of the vertical stress would cause the differential stress to decrease. Since the differential stress needs to be large enough to produce faulting, this scenario would interrupt the brittle deformation (Coulomb, 1776;Mohr, 1900). Strike-slip faulting during SF3 requires the accommodation of faulting exclusively by reactivation of pre-existing faults or fluid overpressure that reduced the mean stress. There is, however, no evidence for both the exclusive reactivation of pre-existing faults and the presence of fluids (e.g. low-T mineralization). Another scenario with fault-parallel, i.e. NW-SE directed, horizontal extension would allow maintaining the differential stress by reducing σ hmin to approach σ3 during an increase of σ v to approach σ2. Nevertheless, this remains an open question since there is (yet) no record for related structures nor have we obtained absolute or relative magnitudes of stresses. An increase of σ1 after the main compressional phase (SF2), however, is in disagreement with the weakened coupling between Iberia and Europe after the collision (Dielforder et al. 2019). This final phase of the tectonic cycle is characterized by the switch of the fault regimes from thrusting to strike-slip and eventually normal faulting (Dalmayrac & Molnar, 1981;Molnar & Chen, 1983). A pause between thrusting and strike-slip faulting probably provides time for another set of tectonic stylolites to grow. A similar chronology containing normal faulting, thrusting, folding and strike-slip faulting is described for the Indochina block in Thailand (Arboit et al. 2015). There, deformation started with extension replaced by layer-parallel shortening in the same direction. Features produced by this compression (stylolites, pressuresolution cleavage) were rotated by folding afterwards.
The detailed analysis of the stylolites to determine the exact orientation of the in-plane principal stresses (σ2 and σ3) (Ebner et al. 2010) is not within the scope of this work. Thus, we cannot distinguish if the stylolites developed in SF2 or before in another strikeslip regime. Both options are in agreement with the concept of Ferrill et al. (2021). The observation of a younger strike-slip regime (SF3) cross-cutting faults and folds of SF2 challenges the concept by Tavani et al. (2015), where no late strike-slip regime is included. In addition, this concept is simplified with respect to stylolites which are assumed to develop in strike-slip regimes, exclusively. This makes a fundamental difference in the stress evolution in the foreland area. However, the switch from thrusting to strike-slip regime is also described for the western Alpine foreland (Smeraglia et al. 2021). The earlier extension is not observed there.
An incipient second stress cycle is documented by a second generation of tectonic stylolites. The apparent coexistence of strikeslip faulting, however, does not allow us to infer the relativechronological order of the stress fields within the new stress cycle. The strike-slip faults can either result from early compression or initial relaxation, or both. The latter case would imply the existence of two generations of strike-slip faults which we cannot distinguish here. Regarding the heterogeneous directions of Late Palaeogene to Neogene normal faulting and the fact that the extension is sub-parallel to the younger compression SF5 (see Section 5.a.3 above) the tectonic cause of SF4 remains unclear. If the extension is linked to foreland bulging induced by the Alpine collision, this stress field SF4 could represent an earlier extensional episode prior to shortening.

Conclusion
We use fault slip analysis and tectonic stylolites to reproduce a highly resolved stress evolution of the Franconian Platform in SE Germany. The analysis allows us to reconstruct the complete tectonic cycle of an inverted basin that originated from initial Late Jurassic -Early Cretaceous extension and subsidence (SF1), followed by successive NE-SW-directed shortening through layer-parallel shortening and the growth of tectonic stylolites and later thrusting and folding (SF2), and a final Late Cretaceous relaxation (SF3). The inversion cycle becomes repeated during a second phase of extension (SF4) and shortening in NW-SE direction (SF5) with a second set of tectonic stylolites and strikeslip to oblique slip faults. During these stress cycles, the main principal stresses switch systematically from Andersonian extensional regime to thrusting, strike-slip and extensional again. The transitions are sometimes preserved in oblique stress fields, with σ1 being the only principal stress in the horizontal plane. Pre-existing faults seem to partly account for small-scale perturbations. Our study provides a crucial insight into the stress development in an intraplate compressional setting where two inversion events developed with their highest horizontal stresses in nearly perpendicular orientation to each other: the first during the Africa-Iberia-Europe convergence followed by a period of extension and then the development of the Alpine orogeny.