Modelling the hydrostratigraphic heterogeneity of a glacio-fluvial scour-pool

Modelling the hydrostratigraphic heterogeneity of a glacio-fluvial scour-pool R. Bersezio1, L. Pessina1, L. Cattaneo2, E. Cavalli1,3, F. Felletti1, M. Giudici2 and M. Mele1 1 Dipartimento Scienze della Terra Università di Milano, via Mangiagalli 34, Milano, ITALIA. riccardo.bersezio@unimi.it 2 Dipartimento Scienze della Terra Università di Milano, via Cicognara 7, Milano. mauro.giudici@unimi.it 3 Now at: CAP Holding., Assago (Milano), ITALIA. emmanuele.cavalli@capholding.it 1. Rationale, aims and methods Realistic representations of the sedimentary heterogeneity of alluvial aquifers are necessary to solve flow and transport problems. The improvements of mathematical tools to simulate the heterogeneity distribution in these sediments permit to handle 3D fields at fine scales. At this purpose, the textural and geometrical components of sedimentary heterogeneity must be reproduced efficiently. Nevertheless, the current mathematical methods are still far from yielding satisfactory results concerning complex geometries and (hydro)-facies distributions (i.e. textures, structures and the related porosity/permeability properties). This is the case of alluvial aquifers that are characterized by: i) complex curve and oblique geometries of the sedimentary bodies bounded by complex truncation surfaces, ii) highly variable and non-stationary facies associations and iii) hierarchic nesting of sedimentary units with contrasting permeability (scour pools, accreting bars, channel fills, terraced valleys and flood plains). Modeling this complexity in the real subsurface aquifer stratigraphy is difficult when fine-scale models are required. The study of exposed aquifer analogues permits to try and improve the geostatistical techniques for simulation of aquifer stratigraphy (pixel or object based methods), using complete sedimentological data-sets. Some recent examples of detailed geometric, sedimentological and geophysical descriptions of different 3-D alluvial aquifer analogue architectures led to reliable mathematical simulations (Zappa et al., 2006; Felletti et al., 2007; Dell’arciprete et al., 2010; Bayer et al., 2011; Comunian et al., 2011). Aiming to combine the best possible reproduction of complex geometries with the accurate simulation of the textural distribution, we investigated an aquifer analogue including the gravel-sand multi-storey fill of a small scour pool. The descriptive approach included the sedimentological, geophysical and petrophysical analyses that yielded a quasi-3-D quantitative model of the analogue. After geostatistical simulation we applied flow and transport models to investigate the behavior of the numerical 3-D “virtual aquifer”. 2. Results The real analogue is a sediment block (7 m  4 m  1.5 m) dug into a gravel-sand quarry that is open into the Upper Pleistocene sandur of the Verbano glacier amphitheater (eastern Piedmont, Northern Italy). The block shows the complex network of cross-cutting, small-scale trough-fills (curve and planar cross-bedded gravels, sandy gravels and sands) filling a scour-pool cut into very fine sand and silt. The complexity of this setting includes also deformation structures due to undermining and failure of the fine-grained pool banks, with gravel fills pinched in between the projecting limbs of the deformed silty-sands. The preliminary geophysical investigation and the subsequent dissection of the block (progressive removal of 50  50  50 cm sediment “voxels”) permitted full inspection of the scour-pool volume. So that, the final descriptive data-set and the sedimentological model include: geo-electrical (VES – ERGI) and GPR images of the block; sedimentological logs, photomosaics and facies maps at cm-scale accuracy along vertical and horizontal intersecting planes; infiltration tests within all the “voxels”, with a spacing of about 50 cm; grain-size distribution measurements on samples; estimates and measurements of hydraulic conductivity on samples of the sandy gravel, sand and, silty-sand facies to parameterize the array of the depositional facies. The descriptive sedimentological and hydrostratigraphic model permitted to check the computed geostatistical model of the volume. The latter was obtained combining the interpolation of the hierarchized stratigraphic boundaries (base of the scour pool, bases of the individual troughs) with the SISIM geostatistical simulation of their sedimentary fill. At the hierarchy of the highest order element (the scour pool and the encasing fine-grained sediments) the comparison of these results with the real block shows that the simulated model respects both the geometry of the sedimentary units and the vertical vs. horizontal assemblage of facies. At the hierarchy of the individual trough fills, a good result was obtained concerning proportions and juxtaposition of (hydro)-facies. Hence, the permeability contrasts were reproduced in their correct locations, respecting the shape of these minor units. Differently, the curve and intersecting geometry of the laminasets was partly lost. The flow and transport model that we applied showed the preferential flow paths and the geometry of plumes of potential solutes, that we virtually injected at different locations. In such a detailed model, the effect of sedimentary heterogeneity in deforming the potential field is quite apparent. Owing to its curvature, many peculiar and unexpected features of the flow paths could be recognized, also highlighting the role of connectivity between the most and least conductive facies. 3. Conclusions In spite of its small size, the studied volume represents a realistic analogue both for comparable sediments in real aquifers and for the large-scale basin fills, that are often formed by concave-up channel fills cut into low-permeable fines; hence the results can be used to critically analyze and improve the simulation and modeling techniques even at this scale. The “virtual aquifer analogue” that we obtained from the real one, represents a numerical laboratory for experiments. The outcomes of these mathematical investigations are ready to compare with the results of field experiments to be worked out into the real analogue. References Bayer, P., Huggenberger,P., Renard, P. and Comunian, A. (2011) Three-dimensional high resolution fluvio-glacial aquifer analog: Part1: Field study. Journal of Hydrology, 405, 1–9. Comunian, A., Renard, A., Straubhaara, J. and Bayer, P. (2011) Three-dimensional high resolution fluvio-glacial aquifer analog–Part2: Geostatistical modeling. Journal of Hydrology, 405, 10–23. Dell’arciprete, D., Felletti, F. and Bersezio, R. (2010) Simulationof fine-scale heterogeneity of meandering river aquifer analogues: comparing different approaches. In: GeoENV VII – Geostatistics for Environmental Applications. (Eds. M. Atkinson and C.D. Lloyd), Quantitative Geology and Geostatistics 16, 127 – 137. Springer. Felletti, F., Bersezio, R. and Giudici, M. (2006) Geostatistical simulation and numerical upscaling, to model ground-water flow in a sandy-gravel, braided river, aquifer analogue. Journal of Sedimentary Research, 76, 1215–1229. Zappa, G., Bersezio, R., Felletti, F. and Giudici, M. (2006) Modeling heterogeneity of gravel-sand, braided stream, alluvial aquifers at the facies scale. Journal of Hydrology, 325, 134 – 153.