The structural analysis aimed to define the geometric framework of both ductile and brittle structures within the study area. Remote sensing data were used to compile 2D foliation, folds, and brittle shear zones, as well as a lineament interpretation of linear brittle features in the bedrock. Aeromagnetic maps were used to identify 2D ductile form lines (Geological Survey of Finland), supported by field observations, and ground-surface elevation LiDAR (National Land Survey of Finland, 2026) for LiDAR lineaments. LiDAR lineaments ranged from approximately 50 to 500 m in length, thus providing larger-scale support for fracture measurements obtained during field mapping.

Fractures are commonly divided into sets based on different criteria (e.g. Peacock and Sanderson, 2018), which guides analyses of the fracture networks, and helps us to determine potential inheritance relationships with fractures and ductile structures (e.g. Crider, 2015). During the field mapping, we investigated 152 horizontal and vertical bedrock exposures and collected structural measurements of the foliation and fractures (including outcrop walls). Fractures, observed both as traces and exposed fracture-controlled outcrop walls, were divided into major (R1) and minor (R2) groups based on the length of the fractures (e.g. Hamdani and Baali, 2019). R1 fractures represented the most continuous fracture orientation per outcrop relative to the size of the outcrop exposure, whereas other fracture orientations were classified as R2. The number of subhorizontal fractures was limited due to the scattered locations of the vertical outcrops. Field measurements and remote sensing data (LiDAR, aeromagnetic maps) were used to generate two separate thematic interpretations that emphasised the characteristic ductile trends (form lines) and brittle discontinuities (LiDAR lineaments) in the bedrock.

The 2D-photogrammetry model was generated to analyse fracture orientation and topology and to compare the results with the overall fracture pattern and orientation of the bedrock depression within the study area, with a particular focus on the bedrock depression along the Aura River (Figs. 1c, 2). Aerial photographs were acquired using a UAV at a flight elevation of 20 m, with a minimum overlap of 60 % to 70 % between photographs. Agisoft Metashape software was used to generate an orthomosaic from the aerial photographs, with eight ground control points. The compiled ortomosaic covered an area of ca. 100 × 100 m. A total of 1185 fracture traces were manually digitised on the 2D-photogrammetry model, and nodes of the fracture topology were interpreted by using the NetworkGT plugin (Nyberg et al., 2018) in the QGIS software. The same plugin was used to create the rose plots.

Data on bedrock surface elevation were based on geotechnical drillings in the covered areas and on bedrock outcrop elevation points derived from ground-surface LiDAR, as described in detail by Ahlqvist et al. (2026). To account for the impact of the 2D form lines on the bedrock surface in the sea area, control points (Fig. 2a) were introduced to adjust the elevation of the interpreted high bedrock surface where geotechnical drilling data were unavailable. The elevation of the control points was set to -40 m a.s.l. (meter above sea level) at sea and -15 m a.s.l. in a small area of land along the ASZ. Additional control points with elevation between -15 and -16 m a.s.l. were added to fill data gaps in the SE part of the study area (Fig. 2). The elevations of the control points were determined on the basis of the estimated thickness of the sediment deposits in the bedrock depressions within the study area.

Three bedrock surface models were generated (Fig. 2), including (i) the bedrock-DEM showing the bedrock surface elevation within the whole study area, and two cross-section DEMs across (ii) the central part of the study area (T1-DEM), and (iii) the SE part of the study area (T2-DEM). The bedrock-DEM and the cross-section DEMs were generated in MOVE Suite (3D-modelling software provided by PE Limited) using the Ordinary Kriging (OK) interpolation method. The OK interpolation was guided by the semivariogram (Fig. 2c, d, e), which measures the average degree of dissimilarity between unsampled values and nearby values (Oliver and Webster, 1990).

The bedrock-DEM was interpolated onto a 30 m grid using an exponential experimental model. In the kriging semivariogram, an experimental model was fitted to the variogram, representing our data with high accuracy (Fig. 2c). Furthermore, we used the honour points method in MOVE, which allows more accurate guidance of the interpolation by data points relative to the grid size. This approach increased accuracy to a 10 m grid in areas where bedrock surface elevations were derived from ground-surface LiDAR data (Fig. 2b).

The cross-section DEMs (T1-DEM & T2-DEM) were interpolated using a 10 m grid size, with increased accuracy of 2 to 5 m in areas where bedrock elevation was detected in ground-surface LiDAR data. Bedrock surface elevation and bedrock structures data were selected within 750 m of Trace 1 for the T1-DEM. In contrast, the data for T2-DEM covered an area of ca. 3000 × 5000 m in the SE part of the study area (Fig. 2). Best-fit interpolation was achieved using an exponential experimental model in both cross-section DEMs, although semivariogram settings were adjusted separately due to the different datasets (Fig. 2d, e). The cross-section of the T1-DEM focused on the steepness of the foliation and its interaction with the geometry of the brittle shear zones and fracturing. A cross-section of the T2-DEM emphasised the association between data coverage and the analysis of brittle structures to generate geologically realistic bedrock surface morphology. As the final step of the analysis, a structural synthesis was compiled to illustrate how bedrock structures influence bedrock surface morphology. The characterisation of the bedrock structural framework was accomplished by integrating brittle shear zones, LiDAR lineaments, and fracture data. The elongated bedrock depressions were interpreted from the bedrock-DEM and compiled with the integrated setting of the brittle bedrock structures to improve the identification of major structural features of the bedrock, including previously undetected brittle shear zones. This analysis was subsequently conducted at a more detailed scale using cross-section DEMs, in which the relationships between detailed bedrock surface morphology and LiDAR lineaments and fractures were analysed near major structural features. Lastly, based on the compilation of bedrock structures and detailed analyses of the cross-section DEMs, a final structural synthesis was generated, revealing a complex network of brittle shear zones and associated brittle structures within the damage zones that control the bedrock surface morphology.

The ductile structural fabric of the study area is dominated by a subvertical foliation throughout. Biotite paragneisses have stronger foliation than the granites, which are weakly foliated in general. However, foliation becomes moderately developed in the vicinity of the brittle shear zones (Fig. 3a). Subhorizontal foliation occurs within an open upright synform-antiform structure in the central part of the study area (Figs. 3b, 4). Foliation shows a prevailing ENE strike, which shifts to more E-striking in the southern part of the study area (Fig. 4). Other foliation orientations are sporadic and localised in the proximity of the brittle shear zones or fold structures. shear zones

Five brittle shear zones were identified within the study area (Fig. 4). Aura River Shear Zone (ARSZ; Ib in Fig. 1c), SZ1 (Ia in Fig. 1c) and SZ2 have a subparallel strike with the regional foliation, whereas Jaaninoja Shear Zone (JaSZ; IIA in Fig. 1c) and SZ3 (III in Fig. 1c) strike obliquely to the foliation (Fig. 4). The ENE-striking ARSZ and SZ1 are the most distinct brittle shear zones with a continuation across the whole study area. These brittle shear zones are subparallel with the lithological boundaries of the bedrock, which is dominated by granitic rocks in the south and biotite paragneisses in the north. The discontinuous SZ2 strikes E–W, thus corresponding to the shift of the foliation strike from ENE to E in the southern part of the study area. SZ3 strikes NW and extends beyond the study area in the north. In the central part of the study area, SZ3 intersect ARSZ and SZ1, and likely terminates on the western side of the open-folded synform structure (Fig. 4). JaSZ is an NE-striking zone that intersects ARSZ and SZ1 in the northeastern part and SZ2 in the southwestern part.

The fracture pattern observed on a typical outcrop comprises one major subvertical fracture orientation (R1), occurring either as an individual fracture or as a set of parallel fractures, often accompanied by at least one minor (R2) fracture set (Fig. 5). While R1-fractures are continuous and regularly occurring on the outcrops, R2-fractures are short and they often abut to the R1-fractures (Figs. 5b, 6b). The N-striking subvertical fractures are the most distinct R1 orientation, while the ENE-, NW- and E-striking subvertical fractures are more often categorised as R2-fractures (Fig. 7b–d).

The 2D photogrammetry model reveals fracture patterns in which the major (R1) and minor (R2) fractures, along with their relationships, are readily distinguished (Fig. 6a, b). The NNE-striking fractures (R1) reach up to +100 m in length and are closely spaced, with approximately two fractures per metre. The minor fractures (R2) include NW-striking fractures throughout the outcrop, whereas N-striking fractures are restricted to the northeastern and eastern parts of the outcrop. Topology analysis shows that NW-striking fractures often abut the NNE-striking ones (y-nodes in Fig. 6b) and, less frequently, intersect them (x-nodes in Fig. 6b). The isolated i-nodes commonly occur at the outcrop boundaries, where the bedrock surface disappears beneath the vegetation. This suggests that many of these fractures likely continue beyond the exposed bedrock and may abut or intersect other fractures that are not visible within the mapped area.

The LiDAR lineaments comprise three main orientation sets that strike ENE, NNE and NW in order of abundance (Fig. 7a, e). When calculating the number of lineaments in these sets, they account for 46 % of all interpreted lineaments (two highest columns included per set; Fig. 7e). The ENE-striking lineaments characterise the central part of the study area, whereas E-striking lineaments prevail in the southern part, both showing sporadic appearance elsewhere (Fig. 7a). NNE- and NW-striking lineaments are spatially distributed throughout the study area (Fig. 7a). The NNE-striking lineaments are prevailing in the southeastern part of the study area. The NW-striking lineaments are evident in the northwestern part of the study area, which hosts a more variable and densely interpreted lineament system than other parts.

The bedrock surface elevation varies from 60 m below to 70 m above sea level (-60 to 70 m a.s.l.), showing abrupt changes due to elongated bedrock depressions with varying strikes throughout the bedrock-DEM (Fig. 8). Bedrock depressions show four orientation sets with ENE, NNE to NE, NW and E strikes. The ENE-striking depressions are the predominant morphological features of the bedrock surface. Extending through the central and northern parts of the study area, these depressions form prominent linear features and produce pronounced contrasts in bedrock surface elevation. The most common orientation of the bedrock depressions is NNE to NE (Fig. 8), but these are narrow and discontinuous compared to the ENE depressions. The NW-striking bedrock depressions are mainly scattered in the north and south. However, one extends through the study area, separating the mainland from the archipelago and becoming more N-striking in the north. E-striking bedrock depressions are almost exclusively found in the southern part of the study area.

The network of brittle shear zones suggests that large-scale brittle structures do not uniformly align with the ductile framework within the study area. Brittle shear zones that conform to the ductile framework of the bedrock are primarily concentrated in the central and southern regions of the study area, exhibiting strike directions of ENE and E, respectively, as exemplified by ARSZ. Additionally, NNE- to NE-striking shear zones, such as JaSZ, characterise the setting of large-scale brittle structures and are extensively distributed throughout the study area. N–S-striking R1 fractures, which show separate NNE- and NNW-striking sets in the T1-DEM and T2-DEM, indicate that these fractures form in response to movement within the brittle shear zones. Cross-section DEMs illustrate that ARSZ and JaSZ control the spatial occurrence of these fractures by bordering and enhancing the expression of the NNW- to NNE-striking fractures within the damage zones. The findings challenge the notion that pre-existing ductile shear zones preferentially localise brittle deformation, as proposed, e.g., by Mancktelow and Pennacchioni (2005) and Skyttä and Torvela (2018).

Instead, the localisation of brittle deformation appears to have been influenced by the orientation of the stress field relative to the pre-existing structures, as also suggested by Saintot et al. (2011) and Viola et al. (2016). This implies that the reactivation potential of pre-existing structures was controlled not only by their inherited mechanical weakness but also by their geometric relationship to the prevailing stress regimes. Based on the prevailing NNW- and NNE-oriented compressional stress regime in southwestern Finland at ca. 1.75–1.65 Ga (Mattila and Viola, 2014), we suggest that NNE and NNW lineaments and fractures represent tensional fractures formed in different stress regimes. That interpretation, and a well-developed system of Riedel fractures, suggests sinistral kinematics for both JaSZ and ARSZ (Fig. 11).

The ARSZ and JaSZ, however, differ in both their fracture characteristics and orientations. In the ARSZ, NNE-striking tension fractures and NNW-striking Riedel' fractures are prevailing, whereas in the JaSZ, NNE-striking tension fractures are prevailing. Therefore, we suggest that brittle deformation within the ARSZ and JASZ reflects distinct events rather than a single regionally uniform deformation event. Sinistral JaSZ shows a geometry similar to that of the Fault System I distinguished by Nordbäck et al. (2022) during NNW compression. Whereas the fracture setting in the ARSZ corresponds to NNE compression, suggesting that this reactivation of the ARSZ generated the characteristic fracture pattern in the ARSZ. In addition, the JaSZ has cross-cutting relationships with the ARSZ (Figs. 4, 12), suggesting that the JaSZ is the result of a younger event. The reported later brittle events and reactivations (Nordbäck et al., 2023) seem to have a limited effect on the brittle structures within the ARSZ and JaSZ, or their potential effect is confined to the unexposed core domains.

Our results indicate that the large-scale brittle structures (lineaments and shear zones), particularly the main ENE-striking brittle shear zones, closely correlate with elongated topographic depressions along the eroded bedrock surface identified in the bedrock-DEM. However, the slopes of the bedrock depressions exhibit morphological irregularities at a detailed scale due to prevailing NNE- to NNW-striking tension fractures within the brittle shear zones. Figure 12a shows the overall relationship between the LiDAR lineaments, bedrock depressions, and brittle shear zones. This structural relationship is further illustrated in Fig. 12b, which shows the comprehensive network of brittle shear zones (black solid lines) alongside fracture measurements (red lines). In addition, the figure includes distinctive bedrock depressions that were not interpreted as shear zones (black dashed lines) due to their weaker contrast in bedrock elevation and/or discontinuous character.

The ductile framework of bedrock in southwestern Finland is characterised by ∼ NE–SW-striking foliation (Väisänen and Hölttä, 1999), and ∼ E- and ∼ NE-striking shear zone patterns (Nordbäck et al., 2023; Torvela and Ehlers, 2010; Väisänen and Skyttä, 2007). The bedrock-DEM reveals that the bedrock depressions along the ENE- and NW-striking brittle shear zones are wider than the narrow, sharply defined depressions along the NNE- to NE-striking zones. This observation suggests that brittle deformation localised along pre-existing ductile shear zones was more distributed than deformation along the brittle shear zones lacking such precursors. Consequently, reactivation of inherited ductile fabrics produced wider damage zones with more extensively fragmented bedrock. The NNE to NE-striking bedrock depressions tend to terminate against the continuous ENE-striking depressions. The deepest parts of the bedrock elevation are at the intersections of two or more bedrock depressions and are underlain by brittle shear zones.

The strikes of the elongate bedrock depressions align with the orientation of the brittle shear zones, whereas the oblique NNW- and NNE-striking fractures control the morphology of the slopes of the bedrock depressions identified near the ARSZ and JaSZ in the T1-DEM and T2-DEM, respectively. Bedrock elevation, particularly along the ARSZ in the T1-DEM, does not increase gradually away from the shear zone but instead forms zigzag patterns on the slopes of the depressions, controlled by the intense NNE-striking fractures (Fig. 13a, b, c). In a cross-sectional/vertical perspective, the bedrock depression slope is characterised by abrupt breaks when the controlling fracture set is subvertical or dips in the opposite direction to the slope (Fig. 13d, e, f). The interpolated bedrock surface in the cross-sections also shows that the SSE-side of the depression wall along the ARSZ and the ESE-side depression wall along the JaSZ are steeper than the respective opposite slopes.

Glacial ripping and quarrying are primary erosion mechanisms that fragment bedrock along fractures in shear zones, creating abrupt surface breaks (Dühnforth et al., 2010; Lane et al., 2015; Bukhari et al., 2021; Skyttä et al., 2023). In Turku, bedrock depressions are wider and deeper along NW-striking shear zones due to their favourable orientation to NW-SE glacial flow, which enhances erosion. ENE-striking depressions, even perpendicular to flow, show deep features possibly from lee-side plucking (Krabbendam and Bradwell, 2011).

The combined influence of pre-existing bedrock structures and glacial erosion processes produces a complex bedrock surface morphology. The interpolation method applied in this study yields more precise results in areas of high-elevation bedrock surface, but similar accuracy cannot be achieved along the elongated bedrock depression due to insufficient data points. Therefore, multi-scale structural analysis provides important, albeit often indirect, constraints on the morphology of the bedrock surface within these depressions. Integrating structural geological information with interpolation results enables the construction of a more geologically and structurally realistic bedrock-DEM, particularly in areas where direct observations are sparse.