CHARACTERIZATION AND MODELLING OF COMPLEX GEOLOGICAL ARCHITECTURES: THE QUATERNARY FILL OF THE PO BASIN AT THE PO PLAIN-APENNINES BORDER (LOMBARDY, ITALY)

Object of this work is the 3-D modelling of complex geological architectures in the Quaternary Po Basin (Lombardy, Italy). Reliable surface and subsurface models in Quaternary alluvial basins are important for several applications, including groundwater research and management, geohazard evaluation, exploitation and protection of other natural resources. The study area is the Po Plain-Apennine border in Lombardy (Italy), in a peculiar sector where the structural culminations of the buried Emilian Arc salient of the Northern Apennines determine the location of isolated reliefs in the Po Plain (i.e. San Colombano Hill and Casalpusterlengo – Zorlesco subtle relic reliefs). This area was selected because it permits to improve the 3-D modelling procedure in a complex tectono-stratigraphic-geomorphological setting, which is of interest for both the still controversial geological reconstructions of the Po Basin and the relevant issues in groundwater management and geothermal energy exploitation. The Quaternary sedimentary fill of the southern margin of the Po Basin in Lombardy records the complex interplay between active Apennine thrusting to the South, rebound and isostatic response to deglaciations at the flexed Alpine margin to the North and the dynamics induced by Quaternary glacial cycles. All of these factors produced the assemblage of nested stratigraphic, structural and geomorphological complexities which are the object of this work. Reliable 3-D models must account for multiple ranks and scales of sedimentary heterogeneity. To obtain such a result, this works attempts to compute 3-D models, constrained not only by the traditional explicit geological “hard” surface and subsurface data, but also by the implicit “soft” data represented by the increments of the geological evolution of the basin. At present, none of the available modelling methods incorporates geological evolution, hierarchy of stratigraphic and structural components of geological heterogeneity and uncertainty as formal rules of 3-D model building in a straightforward manner. Aim of the work is to propose an integrated, multidisciplinary methodology to combine both explicit and implicit geological knowledge as constraint for 3-D (4-D) architectural geological modelling of the study area. Specific aims of this work are: i) to reconstruct the surface and subsurface Quaternary geology of the study area at different scales; ii) derive the increments and the autogenic vs. allogenic controlling factors on the geological evolution; iii) develop alternative 3-D (4-D) models of the Quaternary sedimentary infill of the area, honoring the new maps and subsurface reconstructions and accounting for the incremental geological evolution; iv) contribute to improve and implement a method that combines explicit geological data with the implicit hierarchic and evolutionary constraints for 3-D geological modelling. A multidisciplinary methodology has been set-up. It integrates i) classical geological, sedimentological, stratigraphic, geopedological, geomorphological and structural field surveys; ii) subsurface reconstruction based on stratigraphic correlation of borehole logs and geophysical images, along a fence of 2-D cross-sections over an area of 400 km2 and a maximum investigation depth of 150 m b.g.s.; iii) 3-D geological modelling based on integration of the GIS management of the multiple data-sets and the GeoModeller® 3-D modelling software. GeoModeller® was chosen for the feasibility to deal with the bounding surfaces, which is the key-concept to describe hierarchic frameworks and the key to introduce the genetic interpretation of the basin history (4th dimension) into 3-D representations. To do that, new software routines and novel concepts for the modelling rules were set-up and implemented in the commercial code. Results of the work include: i) a new geological and geomorphological map of the San Colombano hill at 1:10.000 scale over an area of about 60 km2; ii) a hierarchic stratigraphic scheme of the surface-subsurface Quaternary succession of the southern Po Basin, integrated to the iii) incremental tectono-depositional evolution of the Po Basin-Apennine border, that relates the ranking and the significance of the stratigraphic and morphological boundaries to the hierarchy of the Quaternary increments of the geological evolution; iv) the conceptualization of the implicit hierarchic rules to be introduced into 3-D model building, and the procedure to progressively include the explicit and implicit geological rules within multi-scale realizations; v) some new computing routines which let GeoModeller® to manage the new rules and vi) alternative 4-D geological models accounting for different interpretations of the geological evolution. Six high-rank increments of the geological evolution (“stages”) punctuated by low-rank steps have been described in this work. During stages 1 and 2, N-ward thrusting along the blind Emilian Arc originated the Zanclean and the Gelasian Unconformities. On the San Colombano hill, the Calabrian shallow marine San Colombano Fm. (PL4 highest-rank succession) unconformably overlies the truncated deep-marine Miocene formations, up-thrusted during Mio-Pliocene. At stage 3, Early to Middle Pleistocene increments of thrust-folding at the northernmost buried reaches of the Emilian Arc induced erosion of the intra-Calabrian unconformity (U1) and separated local depocentres related to the San Colombano and Casalpusterlengo – Zorlesco structures. These were filled by transitional and alluvial units (PS1 highest–rank succession). These regressive deposits, lap onto the uplifting structures of San Colombano and Casalpusterlengo - Zorlesco, suggesting the onset of their structural separation. At stage 4, these latter two structures were separated from the San Colombano thrust, since the Middle Pleistocene, by means of a newly interpreted dextral lateral ramp (San Colombano lateral ramp), as testified by the delayed migration of the depocentres of the Middle Pleistocene glacio-fluvial units and by the time-shift of the onlaps onto the different structures. After folding of U1, at the base of these units, the Early-Middle Pleistocene unconformity U2 was carved, bounding the base of the PS2 alluvial and glacio-fluvial high-rank succession. During stage 5, Late Pleistocene alluvial and glacio-fluvial units (PS3 highest-rank succession, correlative to late Besnate and Cantù Alpine glaciations) covered, through the Late Pleistocene unconformity (U3), the older glacio-fluvial succession in the subsurface of Casalpusterlengo and Zorlesco areas, while they terraced the deformed marine succession in the San Colombano area, both on the uplifted hilltop and on the surrounding “Plain Main Level” (Castiglioni and Pellegrini, 2001). Syndepositional normal faulting, related to dextral wrenching regime, occurred during this stage. Fault-related offset of Late Pleistocene units, stratigraphic and morpho- structural evidences (facets, relic surfaces and drainage patterns), document ongoing transtension, at stage 6 (Latest Pleistocene – Holocene; U4 unconformity), plausibly relating to the NNW-wards thrusting and related wrenching along the Pavia-Casteggio lateral ramp (Benedetti et al., 2003). Field evidences suggest to propose a link between the entrenchment and the anomalies of the post-glacial river network at the southern margin of the Po Plain to this tectonic stage. This reconstruction links the origin of the highest-rank unconformable stratigraphic boundaries to the Quaternary tectonic stages of Apennine thrusting, wrenching and extension. The intermediate- and low-rank unconformities relate to both minor tectonic increments and to the climatic-driven glacial cycles, because the bases of the glacio-fluvial units are nested within the highest-rank tectonic-induced unconformities. On the isolated reliefs, in situ paleosols testify the preservation of non-erosional surfaces, i.e. morphological surfaces, related to sites of morphological stability. These became the sites for loess aggradation during the Late Pleistocene, that means when the isolated reliefs had been already uplifted and the main controlling factor on deposition was climatic. The recognition of unconformable stratigraphic boundaries vs. conformable “morphological” boundaries permits to unravel the different chronostratigraphic significance of these two surface types (respectively time-transgressive and almost isochronous) and to use them to constrain the reconstruction of the chronological evolution of the basin and the 4-D model to be computed. A novel approach in the use of GeoModeller® is proposed by implementing a model building procedure based on coded ‘hierarchic rules’, at present not encompassed in the modelling suite. A rigorous routine is proposed to apply these rules to obtain at least three ranks of visualization of the 3-D geological architecture of the study area. The ordering of the geological units in the stratigraphic pile, combined with the set of the reference surface (top/bottom) and the nature of the interpolation for each surface (erode/onlap) conceptualized the hierarchic rules valid to represent complex stratigraphic architectures at each scale. 1) The isopotentials of GeoModeller® (i.e. the lowest rank surfaces which can be computed and represented by this software) describe well the morphological surfaces, i.e. surfaces stable through time. Using the orientation of the morphological surfaces as reference top boundary for model computation means to constrain the isopotentials to the deformation history of the area. This concept strongly impacts on the 3-D model application to the simulation of internal facies, as it would be necessary to simulate the distribution of hydrostratigraphic parameters. 2) Since crossing the isopotential, the erode stratigraphic boundaries bring the significance of the time-transgressive unconformable surfaces, in accordance with the geological evolution. 3) By attributing erode nature to the high-rank surfaces, and onlap rules and reverse ordering in the stratigraphic pile to the intermediate-rank ones, the resulting 3-D model displays the high-rank surfaces as composite stratigraphic unconformities, like they have been described by the geological model, since they collect the minor increments of deformation, deposition and erosion through the geological time. As a result, the proposed 3-D models are multiscale and honour the explicit geological observations and the implicit geological evolution at each scale of observation. The intermediate-rank boundaries and sediment volumes represent the result of the intermediate-rank evolutionary increments. On larger spatial and temporal scales, they can be grouped and visualized into higher-rank boundaries (‘U’ unconformities) and volumes, related to the major tectono-depositional stages. The relationship between geological history and geometrical features, with the possibility to upscale and downscale the model according to its hierarchic configuration in view of any specific application, is one novelty of the modelling results here presented. The uncertainties derived from the interpretation of the geological evolution gave rise to two alternative geological models of the San Colombano hill area. Both honour the input explicit data and differ on the interpretation of the extent of the conjugate fault systems that involved the Late Quaternary stratigraphy. The final visualization of the 3-D, ranked stratigraphic units and surfaces highlights the basic role of consistent 4-D geological models as the best synthesis of heterogeneous and multi-scale datasets, that represent the base for several applications at different scale. The adopted approach yields a model that can be easily updated, as soon as new knowledge gets available and modified, and permits to test different hypotheses accounting for any new implicit geological constraints.