3D reconstruction from surface data in complex geological settings: the example of a thrust stack in the Mesozoic cover of the Southern Alps (Italy)

Transcription

1 DOI /s D reconstruction from surface data in complex geological settings: the example of a thrust stack in the Mesozoic cover of the Southern Alps (Italy) Fabrizio Berra & Francesca Salvi & Andrea Zanchi & Alessandro Avaro & Mauro Bonavera & Simone Sterlacchini Received: 15 July 2007 / Revised: 19 May 2008 / Accepted: 23 October 2008 # Springer Science + Business Media, LLC 2008 Abstract A 3D model of a tectonically complex portion of the Southern Alps consisting of deformed sedimentary units has been reconstructed by means of a procedure that has been iteratively applied until the field observations were honoured in detail by the model. The model derives from surface geological data and their elaboration by means of a grid of geological cross-sections, constraining the modelled geological surfaces (both faults and lithological contacts). In order to improve the quality of the final reconstruction, new field data have been collected where the model did not fit with outcrop geometries and, where necessary, new geological sections have been drawn to improve geological data before 3D interpolation. The 3D software (gocad ) demonstrated its efficiency in the interpolation of complex geological data (geological surfaces from geological maps, bedding and fault attitudes, geological cross-sections) through the comparisons of the reconstructed 3D geological surfaces with well-exposed structures (i.e. overturned folds, thrust surfaces, stratigraphic surfaces displaced by faults). The cross-check of the model with field data allowed to correct and thus improve both the geological map and the 3D model, in a virtuous process. The obtained model represents the starting point for further 3D elaborations. Keywords 3D model. Thrust stack. Southern Alps. Geological map. Geological cross sections F. Berra (*) : F. Salvi : A. Avaro : M. Bonavera Dipartimento di Scienze della Terra Ardito Desio, Università degli Studi di Milano, Milan, Italy fabrizio.berra@unimi.it A. Zanchi Dipartimento di Scienze Geologiche e Geotecnologie, Università degli Studi di Milano Bicocca, Milan, Italy S. Sterlacchini CNR IDPA Sezione di Milano, Milan, Italy

2 1 Introduction 3D geological models are usually achieved integrating surface, wells and seismic data. Due to the difficulty of collecting these data, reliable 3D models are generally realised by oil companies for exploration or production and by engineering companies for major infrastructural works, such as dams and tunnels. At present, the investments necessary for the preparation of a well-constrained 3D model with subsurface data are worth being spent only when its expensive and time-consuming construction greatly improves our evaluation of the subsurface geology, allowing to solve geometrical/geological problems which can elude a simple 2D analysis, even if performed by expert geologists. 3D models can therefore improve, at different scale, our capability of understanding the geometry of complex geological settings; they also permit to compute volumes and parameterize geological data for further applications. Recently, the 3D approach to geology has been applied to field data with interesting results [1 6] following different work flows. The construction of a 3D model can require a relative small amount of data in relatively simple geological settings [1 3, 6] but it is more difficult when geology is complex [4, 5]. Nevertheless, in complex cases the 3D approach to geological data is more useful, as it requires a larger amount of data and a deeper understanding of the setting, thus forcing the field geologist to improve the quality and precision of his field work. As a consequence, the definition of a work-flow to build low-cost, precise 3D reconstruction even with only surface data has been increased in the recent years [4, 5], thanks to the availability of effective 3D geological software. Furthermore, the storage of large amount of geological data in GIS database represents an important opportunity for the construction of 3D models using data which are already available in forms which can be easily handled by commercial 3D software. Most of the GIS software easily allows to project the surface geological data on DEM, producing a first visualization of the geology in 3D, but without representing real 3D models as it is not possible to define more z values for each x and y coordinate. Obviously, more data and work with respect to a draped geological map on a DEM are required to obtain real 3D geological models. As a general rule, the production of a 3D model from surface data requires digital maps with a high precision in the position of the geological objects and in the respect of the geometric rules that link geological surfaces with the topography. The topography is stored in digital elevation models (DEM), whose quality could affect the precision of final reconstruction. In order to build a realistic 3D model, surface geological data are sufficient only in simple geological settings [3]. Manifold geological structures have been modelled in 3D using large sets of field measurements of bedding and cleavage in folded areas [2] or using seismic lines [1]. In case of complex geology, subsurface data are not sufficient alone for the interpretation of geological structures at depth; the resulting model can be considered a first order approximation of the real setting, which needs to be tested through further information. In absence of well or seismic data, it is essential to integrate geological maps with other data. The low-cost but efficient way to integrate map data is to produce a few oriented [6] or a net of geological cross-sections [4, 5] with different orientation in order to constrain the surface geology at depth. We followed this approach in the production of the 3D model of the study area from field data. Our work-flow is therefore different from that of [2] and [3] who did not use cross sections and also differs from [1] who used three parallel interpreted seismic lines but rather follows the approach that [4] tested in different geological settings.

3 The 3D reconstruction of complex geological objects can be seen as an iterative process that strictly links the field data with the reconstructed surfaces, honouring basic principles of structural geology and stratigraphy (Fig. 1). The construction of a 3D geological model from surface data alone is definitively not the end of the work-flow. Due to the absence of constraints in depth, the model must be checked in the field in order to verify whether it honours or not the original field data. Only after this stage, still related to field-work, the iterative process can be considered complete. This approach has been followed for the production of a 3D geological model of stratigraphic and tectonic surfaces in a selected area of the central Southern Alps (central Orobic Alps), characterized by a complex tectonic setting. The complexity was further enhanced by the occurrence of widespread Quaternary deposits partially masking the continuity of the bedrock. As a consequence, an interpretation of the geological surfaces below the Quaternary cover, by means of geometric constraints and bed attitudes, has been performed before the construction of the cross-sections used for the 3D model. Our model has been obtained integrating geological data from the 1:10,000 geological GIS database realized by Regione Lombardia (in the frame of the new Geological Map of Italy 1:50,000 CARG project) with new field data specifically collected for the purpose of this work. The model, obtained using only surface data, was iteratively validated through the comparison of the reconstructed geological surfaces with the corresponding structures in the field, in order to evaluate possible misfits and the precision of the model. 2 Geological setting The Orobic Alps, forming the central sector of the Southern Alps, consists of a Permian to Cenozoic sedimentary cover that has been folded and thrust southward during the Alpine orogeny. Alpine tectonics is responsible for a strong shortening of the sedimentary cover Fig. 1 Diagram of the work flow followed for the production of the presented 3-D model

4 that has been recognized since the first regional geologic works (e.g., [7 11]). In the study area, severe shortening of the Mesozoic cover was responsible for the development of an antiformal stacks that consists of up to four tectonic units. The antiformal stacks are welldeveloped in the central part of the Orobic Alps, whereas they evolve into large imbricate thrust sheets to the south. The geometry and thickness of the tectonic units is strongly controlled by the composition of the Mesozoic succession. The presence of two mechanically weak stratigraphic horizons (Carniola di Bovegno, Lower Triassic, and San Giovanni Bianco Formation, Carnian) drove the development of two main detachment surfaces that can be traced all over the Lombardy Basin [11]. The 3D model has been reconstructed for a tectonically complicated area, about 100 km 2 wide (Fig. 2), characterized by thick successions ranging in age from Late Permian (Verrucano Lombardo) to Norian (Dolomia Principale). Toward the North, an important fault system (Valtorta-Valcanale Fault) separates the Triassic from the Permian successions, whereas southward the Clusone Fault separates Carnian from Norian units. Between these faults, the succession is tectonically repeated building up an antiformal stack due to south-vergent thrusting. The stack consists of Lower Triassic to Carnian sedimentary rocks that were detached at the base along the Carniola di Bovegno. Whereas the tectonic setting is complex, no significant facies or thickness changes are observed in the study area; facies changes due to lateral transition only occur in the youngest units (Gorno and Val Sabbia Formations). Fig. 2 Geological map and stratigraphic setting of the study area (modified after CARG data Regione Lombardia). Boxes refer to the position of the views represented in Figs. 5 and 6. The thick lines on the right side of the stratigraphic column correspond to the position of the stratigraphic surfaces that have been reconstructed in the model

5 The lithostratigraphic units exposed in the study area have a total thickness of about 1,500 m, whereas the thickness of the single tectonic units ranges from few tens of meters to about 1,000 m. The oldest unit that outcrops above the basal detachment is represented by sabkha dolostones, vuggy dolostones and siltstones that are referred to the Carniola di Bovegno. This unit rapidly evolves to subtidal and peritidal carbonates (Angolo Limestone and Camorelli Limestone respectively, m thick), covered by basinal marly limestones (Prezzo Lmst., m). Above this basinal unit, carbonate platform facies (Esino Lmst., m) prograde on the basinal areas. The massive Esino Limestone is characterized at the top by an erosional surface covered by peritidal (Breno Fm.) and later subtidal (Calcare Metallifero Bergamasco) limestones ( m). A fine-grained clastic input is recorded by the 150 to 300 m of lagoonal marls and limestones (Gorno Fm.) that both laterally and vertically evolve to deltaic sandstones (Val Sabbia Sandstone). The succession ends with terrigenous-carbonate sabkha facies of the San Giovanni Bianco Fm. (about 100 m preserved). The Valtorta-Valcanale fault dips to the south and is characterized by the occurrence of strongly deformed lenses of the Carniola di Bovegno. The fault is interpreted ([9 11], D Adda, unpubl. thesis) as a triangular zone related to the emplacement of the Orobic Anticline due to out-of-sequence thrusting. The Orobic Anticline, exposed north of the fault consists of Permian rocks capped by the Servino Formation of the beginning of the Triassic, forming an open asymmetric anticline. South of the Orobic Anticline the Middle Triassic to Carnian succession is repeated up to four times along thrust surfaces that define and antiformal stack. The major thrust surfaces (i.e. the thrust at the base of the highest tectonic unit, Fig. 3) show an evident ramp and flat geometry, outlined by the younging of the footwall toward the south. This ramp and flat geometry that can be observed in the field is faithfully represented in the model. This precise reconstruction of this surface was possible because this surface was crossed by cross-sections both perpendicular and parallel to the thrust direction. The antiformal stack has been uplifted in the north due to the emplacement of the Orobic Anticline, so that most of the thrust surfaces gently dip toward the south. As the uplift was more intense in the north, a gentle anticlinal fold with E W axes affected the antiformal stack. The lateral continuity of the antiformal stack is interrupted to the west by the N S trending Grem fault, whereas the described structural units are exposed eastward up to 20 km also outside the study area [9]. A few younger normal faults displace the thrust stack. 3 Data base and methods The administration of Regione Lombardia developed a GIS database for the storage of a geological database surveyed at 1:10,000 scale in the frame of the new Geological Map of Italy 1:50,000 (CARG project). The geological information contained in the GIS database are organized according to a conceptual scheme specifically developed [12, 13]. From this GIS database, a selection of data necessary for the construction of a 3d model (such as geological surfaces, attitude measurement, kinematic indications and so on) has been extracted for the construction of the 3D geological database. The 3D database also includes data collected for the purpose of this work [14]. The DTM was also provided by Regione Lombardia. It is characterised by a m grid and was produced by digitizing the 50 m contour lines from 1:10,000 raster maps (the same maps used as topographic base for the geological field survey)

6 The 3D geological database contains all the geological surfaces (both stratigraphic and tectonic) represented in the map and dip and strike data of stratigraphic and tectonic elements, as well as the traces of axial planes of folds. For some faults, the kinematics and the evaluation of the throw is known. In the light of the high number of boundaries of the stratigraphic units as well as the reduced and constant thickness of some formations, we selected some of the main and significant surfaces to be reconstructed in the 3D model (Fig. 2). In particular, we preferred stratigraphic surfaces bordering thick units or surfaces separating rock bodies with important lithological differences, in order to bound geological volumes with similar lithological properties. As faults and stratigraphic contacts are not continuously exposed in the study area due to the occurrence of Quaternary deposits that cover the bedrock, geological boundaries have been traced below the recent deposits, producing the final data base. Once the 3D geological database has been completed, a regular grid of geological sections (11 and 16 geological cross-sections with respectively N S and E W strike) has been defined. Where the geometry of the structures was particularly complex, some additional sections were drawn (Fig. 3). Section were drawn at 1:10,000 scale and cross checked in order to define geometric constraints along the selected geological surfaces. The total length of the traced cross-sections is about 380 km. The 3D reconstruction was produced with gocad, a software able to integrate different type of data in a 3D model [15]. In gocad the interpolation of the geometric elements is performed through the DSI algorithm (Discrete Smooth Interpolator, [15]), that minimizes the roughness of a triangulated surface honouring imposed constraints. We followed the procedure described by [4]. A pointset extracted from the m pixel-size DTM (Regione Lombardia) was used for the reconstruction of the topographic surface in gocad Fig. 3 Distribution of the geological cross-sections used for the construction of the 3D model. Position of the geological cross-sections used for the 3D model (top left), view of the 3D position of some cross sections in the model (top right), an example of geological section (bottom). Geological boundaries have been projected on the topographic surface. Note that the regular grid of cross-sections has been enriched with several small sections that were used to highlight small or particularly complex geological surfaces (VER: Verrucano Lombardo; SRV: Servino; BOV: Carniola di Bovegno; ANG: Angolo Limestone; ESI: Esino Limestone; Bre: Breno Formation; CMB: Calcare Metallifero Bergamasco; GOR: Gorno Formation)

7 (124,100 points for a km area). Stratigraphic and tectonic boundaries were imported and projected on the topographic surface in order to assign them a z value. The geological cross-sections have been imported and goereferenced in 3D using a special tool [4]. Geological surfaces (faults and stratigraphic contacts) have been reconstructed, interpolating the linear features from the geological map and cross-sections, used as constraints, through the DSI. A relative chronology must be established among different structural surfaces. Thrust and fault surfaces were reconstructed before the stratigraphic horizons. For the reconstruction of folded surfaces, the fold axes and axial planes were also considered as constraints. After the production of a first version of the 3D model, the iterative approach shown in Fig. 1 has been followed. The presence of geometric inconsistencies of the 2D geological interpretation was outlined by the 3D model. These problems often reflected imprecise data in the geological map, with the consequence that, in order to resolve inconsistencies, a check of the field data, of the geological interpretation (map and sections) and of the DTM accuracy was required. In problematic areas of the model, a better definition of the position of the lithological contact and the collection of detailed mesostructural data helped to obtain a 3D model that honoured the field data. At the end of the iterative process, both the 3d model and the geological map were improved. 4 3D model: results and validation process The DSI interpolation of traces of geological boundaries on the topographic surfaces and in the geological cross-sections allowed to reconstruct the 3D geometry of the main lithostratigraphic surfaces and of the faults and thrust surfaces occurring in the study area (Fig. 4). The geometric relationships were geologically coherent. Nevertheless, to verify the quality of the 3D model a final comparison of the obtained surfaces with favourable outcrops has been done. This approach was possible thanks to outcrop conditions, which locally allow to observe large-scale geometry of the rock bodies. In particular, the geometry of well-exposed thrust surfaces (mainly ramps) and deformed stratigraphic surfaces (mainly hangingwall anticlines and footwall synclines) have been compared with the corresponding surfaces created in the model. In detail, we focus our attention on two well-exposed structures. The first one is a footwall syncline, characterized by an E W trending axis and related to the top-to-the-south movement of the tectonic units in the hanging wall (Figs. 4 and 5). The complexity of this surface required the integration of new data from short cross-sections with different orientation in the grid of geological sections. The integration of the data-base allowed to obtain a folded stratigraphic surface honouring field observations. The second structure studied in detail consists of a thrust ramp characterized by the presence of a ramp anticline in its hanging wall (Fig. 6). The structure is perfectly-exposed in the northern part of the study area. This structure was reconstructed using several N S trending cross-sections that cut from north to south side of the highest tectonic unit of the stack. Also the folded stratigraphic surface that defines the ramp anticline is well exposed and has been crossed with several sections. The obtained surface shows an evident ramp and flat geometry, that has been obtained thanks to the grid of geological sections. The described surfaces have been reconstructed after different iterative cycles (Fig. 1). The model has been refined with additional data (both from field survey and from geological sections) that allowed to improve also the quality of the geological map.

8 Fig. 4 View of a selection of the geological surfaces from south-east (top) and detail of the antiformal stack from east (below). The dark (blue in colour) surfaces are stratigraphic boundaries, light (red) surfaces are tectonic boundaries Closer comparisons from field observation and 3D model indicate that the intersection of the geological surfaces with the topographic surface is often irregular. These traces are obtained projecting the geological boundaries from the geological map on the DTM. This problem can be related (a) to the interpolation of the geological contacts in zones with poor exposures; or (b) to a difference between the DTM and 1:10,000 raster maps used for the geological field survey. Despite the problems related to the projection of geological surfaces on the DTM, the comparisons between the 3D model and different field observations was successful also in geometrically complex portions of the study area, confirming that the followed

9 Fig. 5 Reconstruction of the footwall syncline below the highest tectonic unit. Aerial view of the overturned footwall syncline (a) and relative surfaces produced in the 3-D model (b); panoramic view of the structure (c) and 3-D model (d). 1: Folded stratigraphic surface; 2: thrust surface, dotted line: trace of axial surface of the syncline. Abbreviations as in Fig. 3 work flow produced a 3D model that generally respects the constraints given by the field data. 5 Discussion The production of the 3D model of a tectonically complex part of the Southern Alps led to a more coherent interpretation of the geological setting of the study area and to the improvement of the geological map. The construction of the 3D model helps to recognize where the field maps show inconsistencies in the geological interpretation, especially in the first steps of the work flow (Fig. 1). The correction of these inconsistencies, obtained by means of cross check of the geological cross-sections and new field work and/or revision of the available data, increased also the final precision of the geological maps. The obtained model, validated through the comparison of the geometries in the model with those observed in the field, may represent a data base for further characterization of the 3D geological bodies reconstructed in the model. The process of a 3D reconstruction is time-consuming, but it forces the geologist to consider all the relationships among the different surfaces (tectonic and stratigraphic) that are build in the model, because any geometric inconsistency or inaccuracy in interpretation may be outlined by the construction of geological surfaces. This control is possible only when a deep understanding of the geological evolution (both stratigraphic and tectonic) of a region is available. In more detail, cross-cutting and geometric relationships among surfaces

10 Fig. 6 View of a thrust surface with the presence of an evident hangingwall anticline (a, c) compared with the obtained 3-D model (b, d). The ramp and flat geometry of b is well constrained by the geological crosssections that intersect this surface. In d is represented the stratigraphic surface between ANG + PRZ and ESI that is cut by the thrust surface. Note how the thrust surface in the model becomes more irregular close to the topographic surface, due to problems between the representation of the geological object on the map and their projection on the DTM. Abbreviations as in Fig. 3 should be taken into account during the construction of the model. These data may come from field observations and constrain the construction of the geological cross-sections. A correct 3D model must reflect the relative relationships among tectonic and stratigraphic surfaces observed in the field. To obtain a precise 3D model, the geological map must respect some basic requirements. The precision in the position of the outcrops and of geometric measurements is extremely important. A fundamental role is also played by the geometry of the geological boundaries on the topographic map. Any imprecision in the boundaries is reflected by anomalous geometries of the geological boundaries close to the topographic surface. As a consequence, one of the first steps in the elaboration of a 3D model must be a check of the quality of the geological boundaries used in the model. Our experience indicates that, in geologically complex regions, the construction of 3D models from only surface data can suffer the absence of constraints at depth. In these cases, additional data can be obtained by cross-sections which can help to geometrically constrain the position of surfaces at depth. The construction of a net of geological cross-sections increases the time needed for the construction of a 3D model, but guarantees results that are otherwise impossible in complex settings. A further complication is related to regions where outcrops are scarce, because in these situations the reconstruction can be strongly influenced by the geologist s interpretation. The interpolation of surface geology below superficial coverage (Quaternary deposits, lakes etc.) has to be carried out carefully, in order to place in the right position on geological cross-sections the geological surfaces.

11 Another critical aspects regards the selection of the traces of the geological sections for the input of subsurface data in the model. We used a grid of cross sections oriented parallel and perpendicular to the major tectonic structures (in our case, N S and E W). This regular grid has been integrated with selected sections able to characterize small but important structures that are not significantly cut by the regular grid of geological sections. The precision and coherence of the geological data is fundamental, but the quality and the precision of a 3D geological reconstruction heavily depends on the characteristics of the DTM, mainly in terms of (1) resolution, and (2) its relationships with the topographic map used for the field survey. The second aspect is extremely important. In our model, we used a DTM with a m resolution, derived from the manual digitalization of topographic maps (altitude lines every 50 m and points). We locally observed the imperfect position of the altitude lines on the map and the relative DTM, mainly where the altitude lines are not present on the map (steep rocky slopes where altitude lines are missing) and therefore the interpolation suffers from absence of data. As a consequence, the projection of the geological data on the topographic surface locally generated unrealistic irregular lines. This problem can be solved using maps and DTM derived from the same source, as laser scanner survey, able to produce high-resolution topographic databases. Considering the complexity of the construction of 3D geological models, the process can be significantly speeded up by the adoption of a standard procedure that can be easily adapted to different geological settings. The work flow should define operative aspects, such as the check of the geological map, the geometric rules for the drawing of the geologic boundaries below the Quaternary cover, the definition of the surfaces to be modelled, the approximate density and number of geological sections necessary for the model. The reconstruction of 3D models is generally easier if the geological maps have been surveyed considering this application, as in this case the field geologist is forced to be more precise as she/he knows how the data she/he is collecting will be used for. At the end of the process, it is fundamental to verify the quality of the obtained surfaces with a comparison between model and field. This stage of the work flow must represent the final step of the production of a 3D geological model obtained from a geological map, that can not be ignored. 6 Conclusions The construction of 3D geological models from surface data may benefit today by the large amount of existing geological GIS database. The development of a reliable work-flow from field data to 3D models could represent an important goal for the improvement of our geological knowledge in complex geological settings. Our experience indicates that field data integrated by a large number of geological sections may take to reconstruct a generalised geological model also of structurally complex regions. The effort and time required for the reconstruction of a 3d model depends mainly on the available data base, the geological complexity, the detail (i.e. number of geological sections to be drawn) and the skill of the geologist. In detail, with regards to the latter aspect, it is important to stress that a direct knowledge of the local geology significantly speeds up the process of construction of the geological sections and their quality and, thus, the accuracy of the final model. The grid of geological sections can be planned in order to have a regular distribution of the data on the maps, but it is in general necessary to integrate a regular grid with specificallyoriented sections that can depict peculiar geological structures. The 3D model produced by this work flow must be compared with outcrop observations, which represent the only way

12 to check if the field data are honoured by the model. In case of failure, it is necessary to update the geological database with new data and to reconstruct the model, with an iterative process, until it is satisfactory. This final check is, in our opinion, the only way to guarantee the link between data and model and to avoid the realization of models that risk to loose their dependency from field observations. In conclusion, the 3D approach to field geology forces geologists to define the volumes of the rock bodies, a fact that is not explicitly required in 2D geological maps. To obtain this, the geologists are required to define, at least hypothetically, the geometry of surfaces (both tectonic and stratigraphic) that bound the rock bodies. Besides, the construction of the 3D model from surface data and geological cross-sections imposes to check the consistency of the geological map with the 2D cross-section grid. Problems between maps and cross-sections often reveals the existence of errors in the maps or in the position of the geological surfaces. The availability of 3D geological models, strictly constrained by the surface geology, represents an important starting point to characterize the rock units by the mechanical/physical point of view for further quantitative geological elaborations. Acknowledgments This work has been possible thanks to the availability of data form the geological GIS (CARG Project) of Regione Lombardia. We would like therefore to thank the Infrastruttura per l'informazione Territoriale della Lombardia for the use of the data. The first version of this manuscript was improved thanks to the useful suggestions of two anonymous reviewers. References 1. De Donatis M (2001) Three-dimensional visualization of the Neogene structure of an external sector of the northern Apennines, Italy. AAPG Bull 85(3): Fernandez O, Munoz JA, Arbues P, Favilene O, Marzo M (2004) Three-dimensional reconstruction of geological surfaces: an example of growth strata and turbidite systems from the Ainsa basin (Pyrenees, Spain). AAPG Bull 88(8): doi: / Dhont D, Luxey P, Chorowicz J (2005) 3-D modeling of geologic maps from surface data. AAPG Bull 89(11): doi: / Zanchi A, Salvi F, Zanchetta S, Sterlacchini S, Guerra G (2008) 3D reconstruction of complex geological bodies: examples from the Alps. Comput Geosci 35: doi: /j.cageo Bistacchi A, Massironi M, Dal Piaz GV, Dal Piaz G, Monopoli B, Schiavo A (2008) 3D fold and fault reconstruction with an uncertainty model: an example from an Alpine tunnel case study. Comput Geosci 34: doi: /j.cageo Tonini A, Guastaldi E, Massa G, Conti P (2008) 3D geo-mapping based on surface data for preliminary study of underground works: a case study in Val Topina (Central Italy). Eng Geol 99: doi: /j.enggeo De Sitter LU, De Sitter-Koomans CM (1949) Geology of the Bergamasc Alps, Lombardia, Italy. Leidse Geol Meded 14: Gaetani M, Jadoul F (1979) The structure of the Bergamasc Alps. Atti Accad Naz Lincei 66: Forcella F, Jadoul FF (2000) Carta Geologica Della Provincia di Bergamo a Scala 1: (Geological Map of the Bergamo Province, scale 1:50.000). Grafica Monti, Bergamo 10. Schönborn G (1992) Alpine tectonics and kinematic models of the Central Southern Alps. Mem Sci Geol Univ Padova 44: Berra F, Siletto GB (2006) Controllo litologico e stratigrafico sull assetto strutturale delle Alpi meridionali lombarde: il ruolo degli orizzonti di scollamento. Rend Soc Geol It 2:78 80, Nuova Serie 12. Berra F, Liguori F, Piccin A, Mozzi E, Siletto GB (2000) Contenuto informativo e schema concettuale della base dati geologica Regione Lombardia. Arti Grafiche Vertemati, Vimercate (MI), 49 pp

13 13. Brunori CA, Siletto GB, Piccin A, Berra F, Mozzi E (2007) Verso un sistema informativo geologico: l applicativo CARGeo per la banca dati Geologica della Regione Lombardia. Rend Soc Geol It 4: D Adda P, Berra F, Zanchi A (2008) Kinematic and structural meaning of the Grem-Val Vedra fault zone (Orobic Alps, Italy). Rend Online SGI 1: Mallet JL (1997) Discrete modelling for natural objects. Math Geol 29(2): doi: / BF Fabrizio Berra is researcher in stratigraphy at the Dipartimento di Scienze della Terra Ardito Desio, Università degli Studi di Milano, Italy. He works on field mapping of sedimentary succession in orogenic belts and on the stratigraphic interpretation of sedimentary successions. He also worked on the realization of geological GIS. In the last years he worked on successions in the Italian Alps, Sardinia and other countries (Himalaya, Iran). Francesca Salvi is a post-doctoral researcher at the University of Milan, Earth Science Department Ardito Desio. She graduated in Earth Science from the University of Milan in 1999 and received Ph.D in Geotechnologies at the University of Milan-Bicocca in The focus of her research currently lies on the 3D geological model reconstruction of polydeformed and polymethamorphic terrains of the Alpine belt.

14 Andrea Zanchi is full professor in structural geology at the University of Milano-Bicocca in the Department of Geological Sciences and Geotechnologies. His research interests mainly concern the structural analysis of poorly-known Asian Meso-Cenozoic orogenic belts, (Karakoram, Pakistan; Alborz, Iran). Since 1999 he has been working in the frame of the gocad consortium on the development of geometrical techniques for the 3D reconstruction of complex geological bodies. Alessandro Avaro got his MSc degree in 2004 at the Università degli Studi di Milano, Dipartimento di Scienze della Terra Ardito Desio. Work on 3D geomodelling using gocad. He is working as environmental geologist, with particular attention to GIS data management for water monitoring, pollution mapping and land management. Mauro Bonavera got his MSc degree in 2004 at the Università degli Studi di Milano, Italy, Dipartimento di Scienze della Terra Ardito Desio, working on 3D models with gocad. Since 2005 he is a reservoir geologist at the ENI E&P. He works in a R&D group of geomodelling and hydrocarbon field characterization.

15 Simone Sterlacchini obtained his MSc in 1992 and his PhD in Engineering Geology in 1997 from the University of Milan, with a research on Geographic Information Systems for Landslide Hazard Assessment". He joined the National Research Council of Italy (Institute for the Dynamic of Environmental Processes) in 1998 as a researcher, and specialized in the use of Geographic Information Systems for natural hazard and risk assessment and in the use of Decision Support Systems for real-time management of hydrogeological emergency situations in the field of Civil Protection.