Resistivity imaging of Pleistocene alluvial aquifers in a contractional tectonic setting: A case history from the Po plain (Northern Italy)

Journal of Applied Geophysics 93 (2013) 114–126

Contents lists available at SciVerse ScienceDirect

Journal of Applied Geophysics journal homepage: www.elsevier.com/locate/jappgeo

Resistivity imaging of Pleistocene alluvial aquifers in a contractional tectonic setting: A case history from the Po plain (Northern Italy) M. Mele a,⁎, R. Bersezio a, c, M. Giudici a, c, S. Inzoli a, E. Cavalli a, A. Zaja b a b c

Dipartimento Scienze della Terra “A. Desio”, Università degli Studi di Milano, via Mangiagalli 34, 20133 Milano, Italy Dipartimento di Geoscienze di Padova, Italy CNR-IDPA (Consiglio Nazionale delle Ricerche, Istituto per la Dinamica dei Processi Ambientali), via Mario Bianco 9, I-20131 Milano, Italy

a r t i c l e

i n f o

Article history: Received 18 September 2012 Accepted 30 March 2013 Available online 12 April 2013 Keywords: Alluvial aquifers Hydrogeophysics Pleistocene Po basin Vertical Electrical Soundings

a b s t r a c t In this work we present the hydrogeophysical imaging of a key sector of the Quaternary Po foreland basin (northern Italy), focussing on the reconstruction of clastic aquifers and aquitards in a complex tectonosedimentary subsurface architecture. The study area includes the relic reliefs of Casalpusterlengo and Zorlesco, two smooth morphological features involving uplifted and gently folded Pleistocene marine to alluvial sediments, plausibly linked to the buried Northern Apennines thrust and fold belt. The geophysical data include 35 Direct Current Vertical Electrical Soundings collected over a 37 km2 wide area, acquired with Schlumberger array and maximum half-spacing of 500 m. 1-D resistivity-depth profiles were computed for each VES. An integrated hydrostratigraphic approach was applied, to constrain the interpretation of the geophysical data along several cross-sections, including the comparison of resistivity soundings to stratigraphic logs, borehole electric logs and the pore-water properties. The resistivity interfaces, traceable with the same laterally continuous vertical polarity, were used to develop an electrostratigraphic model in order to portray the stacking of electrostratigraphic units down to 200 m below ground surface. Their vertical associations show a general upward increase of electrical resistivity. This assemblage mimics the regional coarsening upwards depositional trend, from the conductive units of the Plio-Pleistocene marine-to-transitional depositional systems to the resistive units of the Middle–Late Pleistocene fluvial and alluvial plain depositional systems. Middle Pleistocene depositional systems host an alternation of North-dipping, high-to-intermediate permeability aquifer systems (70–180 Ωm, thickness of 5–70 m) separated by low permeability aquitards (20–50 Ωm, thickness up to 40 m). These units pinch out against the Casalpusterlengo and Zorlesco relic reliefs, where they cover the uplifted and folded regional aquitard (20–50 Ωm) formed by Pliocene-Lower Pleistocene clays to sandy silts with gravel lenses in agreement with borehole data. In the deepest part of the local stratigraphy, a broad low-resistivity anomaly (b10 Ωm) was clearly mapped through the study area. By comparison with electrical borehole logs in deep oil-wells, it could be interpreted as the fresh–saltwater interface due to the presence of connate waters and brines hosted by the marine-to-transitional shales. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Hydrogeophysical exploration of sedimentary basins is a common tool for hydrostratigraphic and hydrogeological studies (Binley et al., 2010; Bridge and Hyndman, 2004; Hubbard and Rubin, 2005). During the last decade, geo-electrical investigation of alluvial basins has shown the importance of the integration of geoelectrical methods with the (hydro)-stratigraphic reconstruction in order to obtain more realistic and reliable geophysical images of clastic aquifers, aquitards and aquicludes. Until now, the investigation has mainly focussed on multiplescale geoelectrical imaging of the most common sedimentary units ⁎ Corresponding author. Tel.: +39 0250315556; fax: +39 0250315494. E-mail address: [email protected] (M. Mele). 0926-9851/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jappgeo.2013.03.015

encountered in alluvial hydrostratigraphy, like meandering rivers or braided rivers (Baines et al., 2002; Bersezio et al., 2007; Bowling et al., 2005, 2007; Bratus and Santarato, 2009; Mele, 2009; Mele et al., 2012; Sinha et al., 2012; Yadav et al., 2010). These results encouraged a new effort to use geoelectrical methods to investigate the regional architecture of alluvial hydrostratigraphy in a complex tectono-sedimentary setting. This attempt was made in the Lodi alluvial plain, south of Milan, in the central part of the Quaternary Po foreland basin (Northern Italy; Fig. 1A). The hydrostratigraphic structure of this sector of the Po basin formed during and after Plio-Pleistocene thrust-folding and uplift of Middle pro parte (p.p.)–Upper Pleistocene alluvial terraces (aquifer systems) sitting above Upper Pliocene–Middle p.p. Pleistocene marine to littoral sediments (aquitard and aquiclude systems) at the deep intersection between the Alps and Apennines mountain belts. Within this large scale framework, mapping the heterogeneity at a detailed

M. Mele et al. / Journal of Applied Geophysics 93 (2013) 114–126

115

Fig. 1. A) Location of the study area in the Quaternary Po alluvial plain (Northern Italy); B) Main tectonic and geological features at the Alps–Apennines intersection; C) Geological map of the Southern Lodi plain (modified after Baio et al., 2009) showing the position of the study area, the Casalpusterlengo, Zorlesco and Chiesiolo relic reliefs, the San Colombano hill and the main tectonic features.

scale is limited by the uncertainties in the correlation of sparse well log data. In this work, direct current (DC) resistivity methods were used to yield maps of the subsurface electrical resistivity. In clay-poor alluvial permeable sediments, current flow is primarily related to electrolytic conduction, and strongly depends upon the connected porosity, the litho-textural assemblage and the pore-space structure. On the other hand, current flow in clay-rich aquitards and aquicludes is characterized by surface conduction produced at shale–grain interfaces by the excess negative charges of clay particles forming an electrical double layer (Keller and Frischknecht, 1966; Reynolds, 2011; Schön, 2004; Telford et al., 1990): the great specific surface produces smaller electrical resistivity than for clay-free sediments. Moreover, the mutual dependence of electrical resistivity and hydraulic conductivity on the pore-space structure, the shale distribution and the pore water electrical properties (Keller and Frischknecht, 1966; Reynolds, 2011; Schön, 2004; Telford et al., 1990) permits the existence of a relationship between electrical resistivity and hydraulic conductivity, so that geophysical imaging can be used as a proxy to map subsurface hydraulic conductivity structure. Following this premise, the goal of this paper is the application of a methodological approach based on the acquisition of DC data to yield an hydrogeophysical image of (1) the shape and stacking of the hydrostratigraphic units (Maxey, 1964), (2) the heterogeneity of hydraulic

properties and the aquifers connectivity and (3) the relationship of these physical and geometrical properties with the fresh–saltwater interface. The results obtained in the complex tectono-sedimentary setting of the Lodi area provide suggestions about the general applicability of the applied methodology to hydrostratigraphic characterization of alluvial foreland basins. At this purpose, 35 DC Vertical Electrical Soundings (VESs) were collected over a 37 km2 wide area (Figs. 1; 2) and 1-D resistivity inverse modeling was used to define a high-resolution electrostratigraphic model. The concept of “electro-stratigraphic units” (EsUs), as it was redefined by Mele et al. (2012), was adopted to link the vertical and lateral heterogeneities of electrical properties to hydrostratigraphic heterogeneity (external shape, size, lateral continuity and connectivity of sedimentary bodies; textures and fine-to-coarse grained sediments ratios; facies associations and internal structure of the sedimentary bodies; distribution of diagenetic vs. pedogenetic features). Calibration to stratigraphic logs and borehole electric logs allowed correlation of the electrostratigraphic image to the stratigraphy of alluvial units and their marine substratum. 2. Geology and hydrostratigraphy of the southern Lodi plain In the Southern Lodi plain (Fig. 1B), the deposition of the PlioQuaternary marine to continental systems was strongly controlled by

116

M. Mele et al. / Journal of Applied Geophysics 93 (2013) 114–126

Fig. 2. A) Geological map of the study area superimposed on local topography. Circles represent VES stations. Maximum half-spacing of the current dipole and the RMS error of the corresponding 1-D models are also displayed (see text for details); B) Location of VES and TDEM stations and the alignment of the electrostratigraphic cross-sections.

syn-depositional tectonics, in competition with the Pleistocene glacial cycles, by the dynamics of regional vs. local alluvial base-levels and by the isostatic rebound of the Southern Alps orogen (Arca and Beretta, 1985; Bini et al., 2004; Burrato et al., 2003; Carminati et al., 2003; Pellegrini et al., 2003). The marine Upper Pliocene–Middle p.p.-Pleistocene sediments and the Middle p.p.– Upper Pleistocene alluvial hydrostratigraphic systems of the area were deposited above the interference zone between the Southverging buried thrusts of Southern Alps and the northwards advancing Apennine thrusts (Fantoni et al., 2004; Pieri and Groppi, 1981). In the southern part of the Lodi plain under study (Fig. 1B, C), the growth of the WNW–ESE striking Apennine folds produced the uplift and gentle folding of the Pliocene-Lower Pleistocene marine-totransitional shales and sands that form the regional aquiclude of the Lodi plain and contain the interface between freshwater and connate waters (Bersezio et al., 2010; ENI — Regione Lombardia, 2002). The tectonic uplift was progressively accompanied by erosion and terracing of the flanks and hinges of the folds. Subsequent progradation from North to South of coastal, estuarine and deltaic to alluvial depositional systems, fed from the Alpine side, occurred in at least two regressive– transgressive cycles during the Early to Middle Pleistocene. Late Pleistocene alluvial gravels, sands and shales, filled the depocentral areas and smoothed the fold hinges (Bersezio et al., 2010). The combination of Pleistocene tectonic uplift and alluvial terracing led to the formation of the well-known “relic reliefs” (San Colombano, Casalpusterlengo, Zorlesco and Chiesiolo; Fig. 1C) of the Lombardy plain (Benedetti et al., 2003; Bersezio et al., 2010; Cremaschi, 1987; Desio, 1965; Pellegrini et al., 2003).

2.1. Tectonics and stratigraphy The Casalpusterlengo and Zorlesco relic reliefs are located at the southern end of the Lodi alluvial plain (Fig. 1C), which corresponds to the interfluve between the Adda and Lambro rivers, north of the Po Holocene terrace, and belongs to a geomorphological unit called “Livello Fondamentale della Pianura” (Plain base level, LFP; Castiglioni and Pellegrini, 2001). As indicated by deep seismic reflection data (Fantoni et al., 2004; Pieri and Groppi, 1981), the Lodi plain lies above the hinge of the buried Pliocene Southalpine monocline, close to the northernmost buried ramp folds of the Apennines. During late Pliocene–Early Pleistocene, the marine shales and sands at the base of the Quaternary regressive cycle of the central Po plain (Qm Super-synthem; ENI — Regione Emilia Romagna, 1998) were progressively uplifted to the South and downwarped to the North by the growth of the northernmost and youngest, WNW–ESE striking, Apenninic, en-echelon, thrust-related folds (Fig. 1C; San Colombano, Casalpusterlengo structure; Pieri and Groppi, 1981). The Casalpusterlengo and Zorlesco reliefs consist of two nearly elliptical, WNW–ESE elongated morphological rises, much lower than the San Colombano Hill (Fig. 1B, C). They are, respectively, 2-km-long and 700-m-wide and 1-km-long and 500-m-wide, rising above the LFP by 5 m (Fig. 2A). Weathered Middle Pleistocene alluvial units, related to the regional glacio-fluvial cycles (“Mindel” Auct.) with a loess cover (Cremaschi, 1987) and uplifted by the structural culminations of two gentle apenninic anticlines (Bersezio et al., 2010), are exposed. A wide syncline separates these folds from the NW–SE trending San Colombano hill, that exposes Pleistocene to Miocene units (Boni, 1967; Desio, 1965; Pieri and Groppi, 1981) North of the Po river (Fig. 1B, C). The Middle

M. Mele et al. / Journal of Applied Geophysics 93 (2013) 114–126

Pleistocene succession wedges out and partially covers the uplifted and folded Pliocene–Lower Pleistocene marine shales and sands (Bersezio et al., 2010) that form the regional aquiclude of the Lodi plain and which host the fresh–saltwater interface (ENI — Regione Lombardia, 2002). An informal stratigraphic classification of the Early Pleistocene to Post-Glacial units has been proposed by Bersezio et al. (2004, 2010). From the base and upwards the classification includes: • Unit 0 (GU0; Pliocene?–Early Pleistocene: Figs. 1C, 3B) formed by blue, cyan and dark gray clays with mollusc shells and sand gravel/ conglomerate lenses. The unit is interpreted as the marine substratum (“Villafranchiano” Auct.). It does not reach the ground surface in the study area. • Unit 1 (GU1; Early p.p.?–Middle Pleistocene: Figs. 1C, 3B) formed by fining-upward sequences, from gravelly-sands to sands and sandysilty clays. The base of GU1 is an erosion surface, that marks an angular unconformity between GU0 and GU1. The deeply weathered top of GU1 crops out in the Casalpusterlengo relief. GU1 represents the transitional to alluvial succession that unconformably overlays the marine Lower Pleistocene succession. It plausibly corresponds to the Inverino and Cascina Parina units of the San Colombano relief (Pellegrini et al., 2003). • Units 2 and 3 (GU2, GU3; Middle p.p.–Late Pleistocene: Figs. 1C, 2A and 3B) are formed respectively by gravel to silty clay with peat and sandy gravel to sandy silt fining-upward alluvial sequences. These units did not participate to active folding, thus representing the seal of the local Pleistocene Apennine structures. • Post-Glacial unit (PG; Figs. 2A; 3B) includes the alluvial sediments (sands and recycled gravels) of the Lambro, Adda and Po river terraced valleys (Fig. 1C), the most recent sandy fluvial deposits of Brembiolo river. 2.2. Hydrostratigraphy and pore-water properties The regional hydrostratigraphic framework of the central Po plain has been outlined by ENI — Regione Lombardia (2002), that identified four Aquifer Groups (named D to A, in ascending order). Groups A and B are hosted by the Middle–Upper Pleistocene alluvial succession, corresponding in the study area to GU 1, 2 and 3 of Bersezio et al.

117

(2004, 2010). They consist of multilayer freshwater aquifers, including the phreatic and the non-saturated zone, delimited by a water table few meters deep in the LFP. From bibliographic data, electrical conductivity of freshwater extracted from water wells in the Lodi alluvial plain ranges between 500 μS/cm and 620 μS/cm corresponding to an average electrical resistivity of 18 Ωm (Mele et al., 2012). Groups C and D roughly correspond to the lower part of the Middle Pleistocene and Lower Pleistocene marine-to-transitional succession (“Villafranchiano” Auct.), being part of GU1 and GU0 of the same Authors and hosting confined aquifers. The presence of connate waters and brines is a typical feature of the deepest part of the Quaternary fill of the Lombardy alluvial plain. Such high-density, marine waters were trapped at the bottom of the basin during late Messinian (Conti et al., 2000). They occur inside the mostly marine and deepest Group D and rarely rise into the transitional Group C (ENI — Regione Lombardia, 2002). Electrical logs in deep wells drilled for oil exploration (AGIP, 1994; ENI — Regione Lombardia, 2002; Fig. 3A) show a broad low-resistivity anomaly (b10 Ωm) in the deepest part of the local stratigraphy (GU0, GU1) which is interpreted as the signature of the brines. The rising of these waters to a shallow depth has a close spatial relationship with the tectonic features (i.e., thrust faults and folds) related to the buried Appennine fronts (ENI — Regione Lombardia, 2002). 3. Methods 3.1. DC vertical electrical soundings VES is a DC resistivity method which yields the electrical resistivity distribution in the subsurface considering the distortion, produced by subsurface heterogeneity, of an artificial electrical field applied at the ground surface (Dahlin, 2001; Reynolds, 2011; Telford et al., 1990). The method generally requires the use of a four-electrode array, driven into the ground surface along a straight line. Measurements are achieved by applying an electrical field through the external current dipole and measuring the resulting electrical current intensity and the voltage drop between the internal potentiometric dipole. The apparent electrical resistivity of the subsurface (i.e., the electrical resistivity for an equivalent homogeneous medium) is

Fig. 3. A) Electrical resistivity logs (short-normal, dashed line; long-normal, continuous line) at a hydrocarbon exploration well (location in Fig. 2A) compared with the electrostratigraphy along Section 4 (Fig. 10); the fresh-saltwater interface, as revealed by the borehole resistivity logs, is denoted as R2. B) Comparison of stratigraphy (lithology legend in Fig. 6) at borehole C7a3S004 and C7a4S026 and electrostratigraphy at VES V9 and V4 stations (location in Fig. 2A) showing the correspondence between the T1-T2 transition and the vertical stacking of Early Pleistocene to Post-Glacial coarse-grained units (Postglacial, GU1, GU2) of alluvial environment and the Pliocene?–Early Pleistocene marine substratum (GU0).

118

M. Mele et al. / Journal of Applied Geophysics 93 (2013) 114–126

computed using the measured electrical resistance and a geometrical factor which depends on the separation of current and potentiometric dipole. With the Schlumberger inline and symmetric quadripole, consecutive measurements of the apparent electrical resistivity are performed by increasing the separation between the current and the potentiometric dipoles, so that the current flows through an increasing volume of subsurface and, therefore, the apparent resistivity is influenced by the electrical resistivity distribution at increasing depths 4. 3.2. Survey design and data processing 35 VESs were collected with a Schlumberger array over a 37 km 2 wide area (Fig. 2A, B), in order to obtain NNW–SSE and W–E transects, with a station separation ranging from 400 m up to 4 km. Stations were located away from towns, taking advantage of the existing straight roads within farm lands; cables and data-logger were located at the roadsides. A PASI 16 G georesistivity-meter was used to acquire apparent resistivity measurements with current dipole half-spacing varying from 1 m to 500 m (Fig. 4), using an output voltage between current electrodes of 200 V. Apparent resistivity was calculated by stacking voltage and current measurements for two cycles of square wave energizing. When the current electrode separation increased so that the measured voltage fell to values comparable with the noise level, the potential electrode pair was spaced more widely apart; the potential electrode spacing for each VES varied from 0.5 m up to 60 m. This procedure caused a discrete step up or down of apparent resistivity produced by local shallow heterogeneities and by changes in the contact resistance of different potential electrode pairs. The apparent resistivity values obtained with short potential electrode separations were multiplied times a constant correction factor, close to 1, in order to fit the segments to those obtained with larger electrode separations and thus obtain a smooth apparent resistivity curve (Reynolds, 2011). VES data were processed with RES1D software (Loke, 2001) to determine the local electrostratigraphic sequence (i.e., vertical stacking of resistive above conductive electrolayers or vice versa; Mele et al., 2012) at each VES station. The models assume that the subsurface resistivity changes only with depth and no lateral resistivity variations

exist, i.e., 1-D model. The model parameters are the resistivity ρ and the thickness of each layer; the last one is considered as a homogeneous half-space. The 1-D resistivity-depth profiles were computed through an inversion routine with a fixed number of layers. The routine minimizes the root mean square relative error (RMS) between the observed and the modeled apparent resistivity with a damped least-squares optimization method (Lines and Treitel, 1984) and by employing the linear filter method (Koefoed, 1979) for the forward modeling. Highly parameterised models were obtained by using the “automatic inversion” option of RES1D, which yields a 1-D model with a number of electrolayers greater than the number of apparent resistivity data for each VES (Fig. 5). A preliminary smooth 1-D model was then determined from each highly parameterized model by considering the relative peaks in the resistivity-depth profiles. The preliminary model was used to initialize a second inversion step, which yielded the intermediate 1-D resistivity model of the VES. A third inversion step was attempted by adding an electrolayer to the 1-D intermediate model in correspondence to the deepest abrupt decrease in resistivity; the final model obtained in this stage was considered as satisfactory if the RMS reduction with respect to the intermediate model was greater than 0.5%. Each inversion step involved up to 8 iterations of the optimization procedure, with a significantly decreasing RMS error (Fig. 5), and delivered reliable final models with 4 to 8 electrolayers. Values of the RMS error lower than 4.5% (Fig. 2A) were obtained for the final 1-D models. A maximum exploration depth of 200 m below ground surface was qualitatively estimated by considering the thickness of the deepest electrolayer at least equal to that of the electrolayer directly above it and by calculating the cumulative thickness of the electrostratigraphic sequence. The 1-D VES models were also calibrated with eight TDEM soundings (Reynolds, 2011; Telford et al., 1990) collected nearby a few favorable VES locations (Fig. 2B). The position of the TDEM stations was strongly constrained by the agricultural activities and the pervasive presence of man-made structures within the study area. A Zonge NT-20 NanoTEM transmitter with an output current of 6 A at a transmission rate of 32 Hz through a squared loop (100 m side length) was used. The transient voltage produced by the decaying total magnetic field was recorded through a coil with a vertically oriented magnetic dipole, located at the center of the transmitter loop, using a Zonge GDP-32 II receiver. Each individual signal was recorded at logarithmic equally spaced time intervals after switching-off of the applied current in the transmitter loop within a time window extending from 52 μs to 6 ms. A stacking of 90 to 100 signals was applied in order to enhance the signal-to-noise ratio. 1-D resistivity models for TDEM station were computed by applying Occam's inversion (Constable et al., 1987) and then reducing the possible number of electrolayers by comparing and refining the resistivity-depth profiles obtained from the nearby VES. Unfortunately, the presence of both static (pipes, cables, metal fences) and dynamic noise (power lines, oil lines, cathodic protection) could not be avoided. The electromagnetic noise dramatically affected the signal-to-noise ratio at late times in most of the TDEM soundings and only TDEM 4 and TDEM 7 stations yielded reliable 1-D resistivity models (Fig. 7). 3.3. From 1-D VES data to quasi 3-D electrostratigraphy

Fig. 4. Apparent resistivity curves of the VESs collected along cross-section 2 (location in Fig. 2B).

VES models were used to assemble NNW–SSE and W–E oriented cross-sections (Fig. 2B). A two-step correlation procedure was applied (Mele et al., 2012): (1) definition of the 1-D electrostratigraphic sequence; and (2) correlation of the resistivity interfaces between VES electrolayers, according to their polarity, in order to bound the subsurface volumes that are characterized by the minimum internal variation of electrical resistivity. These resistivity interfaces define an assemblage of quasi 3-D electrostratigraphic units (EsUs; Mele, 2009; Mele et al., 2012). EsUs are geophysical entities delimited by

M. Mele et al. / Journal of Applied Geophysics 93 (2013) 114–126

119

Fig. 5. Comparison of observed and computed apparent electrical resistivity (on the left), 1-D multi-layer model and 1-D final model (center) and RMS error inversion error (on the right) for VES V4 and V9 (location in Fig. 2).

two detectable resistivity interfaces with persistent polarity (i.e., resistive above conductive bodies or vice versa) and are characterized by a tight range of electrical resistivity and by a vertical and horizontal extension that can be mapped according to the resolution of the survey. This allows matching of the geo-electrical image of the subsurface (Fig. 6) with the stratigraphic architecture obtained from geological data. In fact, the EsUs are, by definition, comparable to stratigraphic units, that are identified by their boundaries, lithotextural associations, thickness and lateral continuity (Mele et al., 2012). This approach permits to compare and calibrate the geoelectrical images with the stratigraphic data, not only at the 1-D borehole data-points, but also in the 2-D (along cross-sections; Figs. 6, 9) and in quasi 3-D (fence diagram and interpolation of bounding surfaces; Figs. 10, 11). In other words, the dimensionality (1-D to quasi 3-D) and the hierarchic assemblage of electrostratigraphic units, from the lowest rank of the individual electrolayer to the highest rank of the assemblage of EsUs, were taken into account.

4. Results 4.1. Electrostratigraphy and geophysical mapping The results of the DC survey are presented in the form of apparent resistivity maps (Fig. 8), cross sections (Figs. 6, 9) and 3-D fence diagram (Fig. 10). The general features of EsUs are summarized in Table 1, where they are compared to the regional (hydro)-stratigraphy. The cross sections include the individual 1-D VES data, showing the vertical sequence of interpreted electrolayers, the 2-D EsUs obtained after correlation of the 1-D models, and the borehole stratigraphic logs used for local calibration of the neighboring VESs. The combination of apparent resistivity maps with the fence diagram provides the quasi 3-D subsurface image. These results are briefly discussed here, whereas

they are compared to (hydro)-stratigraphy and interpreted in the next two sections. Two major, laterally continuous resistivity interfaces, named R1 and R2, were recognized (Figs. 6, 9 and 10). They separate three associations of EsUs, named, from top to bottom, T1, T2 and S (Table 1) progressively less resistive downwards. The resistivity-depth profiles obtained by TDEM vertical soundings are in good agreement with the electrostratigraphic framework inferred by VES models (Fig. 3B). The resistivity structure across the Casalpusterlengo and Zorlesco relic reliefs is shown in Fig. 6, along the WNW–ESE striking electrostratigraphic cross section 1. From top to bottom, a decrease in electrical resistivity is clearly recognizable. It is quite consistent with apparent resistivity measured at increasing electrode spacing, i.e., increasing exploration depth (Fig. 8). The electrical contrast R1 bounds the base of the association T1, which is up to 140 m thick and is formed by N-dipping, resistive (70–130 Ωm and 130–180 Ωm) and conductive (20–50 Ωm) EsUs (Table 1). T1 thickens southward and terminates above R1 in the vicinity of the Casalpusterlengo and Zorlesco relic reliefs. Here, T1 wedges out and partially covers the culmination of the T2 association of EsUs. The latter is a widespread, conductive unit (20–50 Ωm) containing some lens-shaped resistive EsUs (70–180 Ωm), tens-of-meters thick. The deeper electrical contrast R2, at the base of T2, marks the vertical transition to a dome-shaped, assemblage of high-conductive EsUs (association S: Table 1), characterized by electrical resistivity lower than 10 Ωm. 4.2. Calibration of resistivity data and comparison to (hydro)-stratigraphy Several lithological logs from water wells and boreholes and short and long normal resistivity logs from one deep hydrocarbon well (AGIP, 1994; ENI — Regione Lombardia, 2002) are available in the study area (Fig. 2A). We used these data to calibrate the association

120

M. Mele et al. / Journal of Applied Geophysics 93 (2013) 114–126

Fig. 6. Above: 1-D resistivity-depth profiles along NNW–SSE oriented electrostratigraphic cross-section 1 (location in Fig. 2B) showing the general upwards increasing resistivity trend in the study area. The inferred position of R1 and R2 surfaces are displayed. Below: Interpretation of electrostratigraphic cross-sections 1, showing the architectural component of hydrostratigraphy. Red and blue lines represent R1 and R2 resistivity interfaces interpreted as the resistivity counterpart, respectively, of the boundaries of stratigraphic units (R1) and the fresh–saltwater interface (R2), where trapped connate brines are preserved (volumes colored with light blue). Lithology logs are orthogonally projected from a maximum distance of 300 m; litho-textural classes are modified from the Miall's (1996) classification (G: gravel; S: sand, L: silt: C: clay; f: fine; m: medium, c: coarse).

of EsUs, delimited by the major electric interfaces R1 and R2, with respect (1) to the stratigraphy of the alluvial units and their marine substratum and (2) to the depth of the fresh–saltwater interface, as known for the study area.

Fig. 7. Comparison of observed and computed transient voltage decay at TDEM-4 and TDEM-7 stations (location in Fig. 2B). Corresponding 1-D resistivity-depth profiles are shown in Fig. 3B.

The electrostratigraphy obtained at VES V9 (Section 1; Fig. 6) was compared to the stratigraphy from borehole C7a3S004 (Fig. 3B), located a few meters aside the VES station (Fig. 2A). The depth of R1 inferred from the VES V9 inverse model is in good agreement with the estimate obtained from TDEM 4. It corresponds quite well to the vertical superposition of coarse-grained units of the Early Pleistocene to Post-Glacial (PG, GU1, GU2) above the Pliocene?— Early Pleistocene, mostly fine-grained, marine substratum (GU0). High permeability, freshwater aquifers are assumed to correspond to electrical resistivity greater than 70 Ωm, whereas low-permeability, clay-rich aquitards and aquicludes are indicated by lower electrical resistivity, according to the classification proposed for freshwater saturated sediments of the Lodi alluvial plain by Mele et al., 2012 (Table 1) by considering the dependence of electrical resistivity and hydraulic conductivity on the pore-space structure and the shale distribution (Keller and Frischknecht, 1966; Reynolds, 2011; Schön, 2004; Telford et al., 1990). The sharp decrease of the electrical resistivity observed in the VES and TDEM profiles in correspondence to the transition between alluvial units and the marine substratum is also present in the borehole

M. Mele et al. / Journal of Applied Geophysics 93 (2013) 114–126

121

Fig. 8. Apparent resistivity measured at electrode spacing of the current dipole equal to 8 m, 80 m and 300 m superimposed to the contour plot of the local topography (equidistance 1 m). Black shaded areas correspond to Casalpusterlengo and Zorlesco relic reliefs; black dots represent VES stations. Data were gridded using a Minimum Curvature algorithm.

electrical logs in the previously mentioned deep hydrocarbon well (AGIP, 1994; ENI — Regione Lombardia, 2002) shown in Fig. 3A. The same logs show a broad low-resistivity anomaly (b10 Ωm), which is interpreted as the effect of high-density, marine connate waters, trapped at the bottom of the basin during late Messinian (“brines”; Conti et al., 2000).

5. Discussion The calibration of the electrostratigraphy to available stratigraphic logs, borehole electric logs and the electrical conductivity of groundwater in the area allows to formulate a model of the Plio-Quaternary marine to continental depositional and hydrostratigraphic systems

Fig. 9. Interpretation of electrostratigraphic cross-sections 2 and 3 (location in Fig. 2B); red and blue lines represent R1 and R2 resistivity interfaces. Lithology logs are orthogonally projected from a maximum distance of 300 m; litho-textural classes are modified from the Miall's (1996) classification (G: gravel; S: sand, L: silt: C: clay; f: fine; m: medium, c: coarse).

122

M. Mele et al. / Journal of Applied Geophysics 93 (2013) 114–126

Fig. 10. Fence diagram of electrostratigraphy of the cross-sections 3, 4 and 5 (location in Fig. 2B).

across the subdued relic reliefs of Casalpusterlengo and Zorlesco. The electrostratigraphy depicted from cross-sections shows a general upwards increasing resistivity trend (Fig. 6), which corresponds to the regional, coarsening upwards, depositional trend from the conductive units of the Plio-Pleistocene marine-to-transitional depositional systems to the resistive units of the Middle–Late Pleistocene fluvial and alluvial plain depositional systems (Fig. 3B). The resistivity contrast R1 can be interpreted as the geophysical signature of the top boundary of the Pliocene-Lower Pleistocene marine-to-transitional shales (GU0), which marks the base of the regressive cycle of the central Po plain, (Figs. 6, 9 and 10). The geometry of R1 can be attributed to 1) gentle folding and partial erosion of the marine-transitional shales of GU0 (T2 and S; Table 1) corresponding to the buried Casalpusterlengo and Zorlesco anticlines, and 2) burial by the regressive sands and gravels of GU1 and GU2–3 (T1; Table 1). The EsUs association T1 plausibly corresponds to the Middle Pleistocene to post-glacial fluvial and alluvial plain depositional systems. It hosts the high-to-intermediate permeability aquifer systems separated by intervening low permeability aquitards. In the western sector of the study area (Sections 1, 2, 3; Figs. 6, 9), the aquifers and aquitards are clearly compartmentalized. The intersection of NNW–SSE and E–W electrostratigraphic sections reveals the 3-D geometry of EsUs corresponding to the tabular and undeformed Upper Pleistocene alluvial terraces (GU 2, 3 and

Postglacial) superposed on the wedge-shaped, gently tilted and mildly folded Middle Pleistocene GU1 (Fig. 10). The superposition of resistive above conductive EsUs allows to depict their gentle dip to North, their termination geometry above R1 (corresponding to depositional pinch-outs) and the seal of the Casalpusterlengo fold hinge by a low-resistivity/low-permeability EsU, corresponding to the uppermost part of GU1 (Figs. 9, 10). The described features are consistent with the superposition of stratigraphic units, which lay above the folded ones and are progressively less deformed, moving from the cores of the two relic reliefs towards the external parts of the limbs of the folds (Fig. 1C). This interpretation is in good agreement with the geological reconstruction (Bersezio et al., 2010) that points to syn-depositional growth of the Apennine structure resulting in the lateral termination of the alluvial units above the uplifted marine substratum and their progressively less pronounced tilting moving upsection. In the eastern part of the study area (Section 4; Fig. 10), the alluvial units host high-to-intermediate permeability aquifers but no persistent resistivity interfaces produced by widespread permeability barriers are detectable. Nevertheless, the difference in the electrostratigraphic sequence between VES V32, V31 and V33 suggests lateral terminations of lens-shaped, high-tointermediate permeability aquifers and low permeability aquitards. The precise location and geometry of the terminations is still uncertain (Fig. 10) due to poor VES coverage in the area.

Table 1 Summary of electrostratigraphic features of EsUs and their major associations compared to regional stratigraphy and hydrostratigraphic interpretation (Bersezio et al., 2004, 2010). Permeability classes of EsUs computed from Mele et al. (2012). n.a. = not analyzed. EsUs association

ρ (Ωm)

Thickness range (m)

Geometry & stacking pattern

Depth range (m bgs)

Estimate of permeability classes (Mele et al., 2012)

GU (Bersezio et al., 2010)

Hydrostratigraphy

T1

130–180 70–130 20–50

20–70 5–30 0.5–40

Alternating wedge-shaped resistive EsUs and conductive EsUs

0–140

High Intermediate/high Low

Postglacial, 3, 2, 1

Freshwater aquifers/aquitards

70–180 20–50

0–20 10–100

Resistive lens-shaped EsUs within regional conductive EsUs

10–140

Intermediate/high low

0

Freshwater basal aquitard

b10

–

Dome

>80–160

n.a.

0

Saltwater basal aquitard

R1 surface T2

R2 surface S

EsUs electrostratigraphic units; ρ electrical resistivity; GU geological unit.

M. Mele et al. / Journal of Applied Geophysics 93 (2013) 114–126

The T2-S associations, corresponding to Pliocene-Lower Pleistocene marine-to-transitional depositional systems, represent a widespread basal aquiclude in the Southern Lodi plain. The lens-shaped, resistive anomalies below the R1 surface are interpreted as confined aquifer lenses, plausibly formed by gravel units, in agreement with borehole data (Figs. 3B, 9 and 10). They appear to be confined and separated from the Middle Pleistocene to post-glacial aquifers; nevertheless, their lateral termination and the cross-cut relation with the R1 surface could not be spatially resolved and further high resolution data are necessary to better constrain this interpretation. The top of the unit GU0 (R1 surface) clearly shows the effects of the uplift produced by the growth of the Casalpusterlengo and Zorlesco ramp folds, which resulted in the formation of an hydrogeological divide between the depocentral areas to North and South of them (Figs. 9, 10 and 11). The positive structures, which are elongated in a nearly WNW–ESE direction, are located in correspondence to the Casalpusterlengo (VESs V1, V3, V4, V10) and Zorlesco (VESs V35, V5) relic reliefs. In the western part of the study area, they are separated by a plateau interpreted as a flat erosion surface truncating the top and the flanks of the folds. These are plausibly interpreted as buried alluvial terraces, cut by the deep incision of the post-uplift terrace of GU2. In the central part of the area, a deep incised valley, ENE–WSW oriented and with an apparent drainage direction to ENE, is apparent at the location of VES V7 (Section 2; Fig. 9). Due to the presence of the Casalpusterlengo city area, no further VES could be collected in order to better delineate this structure, so that its orientation is still uncertain. Directly to the North of the Zorlesco reliefs, the top of GU0 shows a

123

sharp step towards the synformal depocenter to North (Sections 1 and 2; Figs. 6, 9 and 11A). The fresh–saltwater interface corresponds to the R2 surface (Fig. 3A). The elevation map of Fig. 11B shows two marked rises of this interface, separated by an intervening WNW–ESE striking depression (VES V1, V7, V24; Fig. 2B). The southern rise is coincident with the location of the Casalpusterlengo relief, in the southwestern sector of the study area (VES V4, V13, V17, V2; Fig. 2B). The northern one (VES V5, V26, V25; Fig. 2B) is slightly shifted South of the Zorlesco relief, but its strike runs parallel to the elongation of the relief itself and of the axial trace of the buried anticline below it. In Fig. 12, the depth of the fresh–saltwater interface in the study area is compared to the depth obtained by Alfano and Mancuso (1996) at the San Colombano hill, that is located a few km to the E (Fig. 1C). This interface is indeed folded, at a local to regional scale, by the Apennine ramps. The connate waters were plausibly squeezed, during shortening, along the San Colombano, Casalpusterlengo and Zorlesco ramp folds and faults.

6. Conclusions This work aimed to determine the subsurface resistivity structure in the Quaternary Southern Lodi plain, a part of the Po foreland basin (Fig. 1), using DC VES soundings with dense data coverage. The method was applied for imaging the hydrostratigraphy of this alluvial plain by integration of stratigraphy and ground-based DC electrical imaging.

Fig. 11. A) Elevation map of R1 interface showing the two positive structures in correspondence of the Casalpusterlengo and Zorlesco relic reliefs (black shaded area; Fig. 2A). Black dots represent VES stations. B) Elevation map of R2 interface (fresh–saltwater interface).

124

M. Mele et al. / Journal of Applied Geophysics 93 (2013) 114–126

Electro-stratigraphic units were identified on the basis of horizontal changes of vertical electrostratigraphic sequences (Mele et al., 2012). The proposed electrostratigraphic model matches the existing geological models and could be considered as a proxy of the hydrostratigraphic architecture of this part of the Po river basin. Two major resistivity interfaces (R1 and R2; Figs. 6, 9 and 10; Table 1) have been identified with the same laterally continuous vertical polarity (resistivity decreasing with depth) in electrical resistivity and traced throughout the study area. They bound three associations of EsUs (S, T2 and T1 in ascending order; Table 1) that show different geometry. The two deepest associations, S (highly conductive EsUs) and T2 (conductive EsUs with resistive lens-shaped enclosed EsUs), show folded boundaries. The lower part of the T1 association (alternating resistive and conductive EsUs) terminates laterally against the S and T2 associations, while its upper part covers the folded shape of R1. The R1 and R2 interfaces were interpreted as the geo-electrical counterpart of, respectively, the boundaries of the stratigraphic units (R1) and interface between the sediments saturated with freshwater and those where trapped connate brines are preserved (R2). A general trend of upwards increasing resistivity from S to T1 is apparent. It mimics the coarsening upwards depositional trend that is typical of the general regressive evolution of the filling of the Po plain foreland basin. Resistivity interfaces have been recognized by other Authors in different sedimentary basins and depositional settings (Singh et al., 2011; Sinha et al., 2012; Tizro et al., 2010; Yadav et al., 2010 among the most recent). In general, these Authors portray the stratigraphic and hydrogeological architectures by directly matching the 1-D electrolayers with the borehole data. In this way, the interpretations of the subsurface geometries of sedimentary bodies are strongly dependent on the correct identification of the electrical properties of the lithotypes in the borehole logs. Instead, the approach applied in this work focuses on deriving images of the (hydro-) stratigraphic architecture from the definition of the electrostratigraphic sequences within a hierarchically stratified sedimentary sequence. These images are almost independent from the lithological interpretation of the individual electrolayers and of their sequence within the single resistivity-depth profiles. The match of the electrostratigraphic model with the geological model of the study area is good. The work clearly depicted the folded geometry of the S and T2 EsUs associations, corresponding to the Plio-Pleistocene marine shales across the subdued relic reliefs of Casalpusterlengo and Zorlesco (Bersezio et al., 2010). The gradual sealing by regressive alluvial sediments, after truncation of the fold hinges and terracing, is portrayed by the shape and internal architecture of the T1 EsUs association, corresponding to the Middle–Upper Pleistocene alluvial stratigraphic units.

The electrostratigraphic model is consistent with the geological evolution of the area. Plio-Pleistocene folding and uplift of two anticlines related to the northernmost thrusts of the Northern Apennines (Casalpusterlengo and Zorlesco; Fig. 1 and Figs. 9 to 12), concurrently with the Middle–Late Pleistocene regression, produced: (1) uplift by approximately 100 m of the sedimentary bodies on horizontal distances of hundreds of meters; (2) the lateral juxtaposition, due to truncation and terracing, of the gently tilted and/or undeformed alluvial units assembled into T1, against the uplifted fine and coarse grained units (marine and transitional sediments) corresponding to S and T2. The correlation between electrostratigraphy and hydrostratigraphy was obtained by relating the electrical resistivity to the average hydrostratigraphic properties (i.e., hydraulic conductivity) of the low-permeability, clay-rich aquitards and aquicludes and of the most permeable sand/gravel aquifers. In keeping with the approaches reported in the literature (de Lima and Niwas, 2000; Hodlur and Dhakate, 2010; Hubbard and Rubin, 2000; Huntley, 1986; Kelly and Frohlich, 1985; Kosinski and Kelly, 1981; Niwas and De Lima, 2003; Ponzini et al., 1984), the average hydraulic properties were estimated according to the classification proposed for freshwater saturated sediments in the area (Mele et al., 2012) by considering the dependence of electrical resistivity and hydraulic conductivity on the pore-space structure and the shale distribution (Keller and Frischknecht, 1966; Reynolds, 2011; Schön, 2004; Telford et al., 1990). The calibration to available stratigraphic logs and borehole electric logs (Figs. 3, 9 and 10) and the electrical conductivity of groundwater in the area lead to a resulting integrated hydro-electrostratigraphic model. The model shows the folded basal aquiclude (EsU associations S and T2), which contains aquifer lenses and an updoming salt–freshwater interface (resistivity contrast R2). The overlying alluvial aquifers and the intervening aquitards are depicted by the EsUs forming association T1. The spatial relationships between them and the basal aquiclude (resistivity contrast R1) are controlled by deformation that led to erosional contacts, both horizontal and vertical, eventually driving mixing of connate brines and freshwaters. Based on these results we can draw some observations about the limits of the electrostratigraphic approach. Firstly, the correlation of EsUs is strictly dependent on the vertical resolution of the DC method. Hence, the physical scale and stratigraphic hierarchy of the detectable geological and hydrostratigraphic features are determined by this resolution. In our case-history the EsUs can be compared to the scale and hierarchy of the stratigraphic formations and of the hydrostratigraphic systems (Maxey, 1964). At this physical scale, the complex geological geometries of the study area (lateral termination of sedimentary bodies, folded geological units and fresh–saltwater interface) could be

Fig. 12. SE view of the Southern Lodi plain showing the 3-D shape of the fresh–saltwater interface in correspondence to the San Colombano hill (modified from Alfano and Mancuso, 1996) and Casalpusterlengo and Zorlesco reliefs (R2 interface); the study area is framed by the dashed rectangle.

M. Mele et al. / Journal of Applied Geophysics 93 (2013) 114–126

recognized and mapped as it is shown in Fig. 8 to Fig. 12. Secondly, the relationship between EsUs and hydrostratigraphic units is strictly dependent on the calibration to the borehole stratigraphy and to the water chemistry. For instance, in the considered case study the regional basal aquiclude below the R1 interface (EsU association T2) was determined and distinguished from the ensemble of aquifer systems above the R1 interface (EsUs association T1). This calibration clarifies without ambiguities the significance of the R2 interface, which is not a boundary between geological units, but is the fresh–saltwater transition. Acknowledgments The Authors would like to acknowledge N. Ceresa, S. Cantinotti, G. Broetto, A. Regorda, Dr. M. Ferrari, Dr. Nicola Praticelli, Dr. Gianpaolo Girardi and Dr. Domenico Lupis of i.Geo Snc for their support during field acquisition. Thanks to D.L. Harry and to an anonymous referee for their critical reading and helpful suggestions. This work was partially supported by the MIUR and the University of Milano through the research projects of national interest “Integrating geophysical and geological data for modelling flow in some aquifer systems of alpine and Apenninic origin between Milano and Bologna” (PRIN 2005) and “Integrated geophysical, geological, petrographical and modelling study of alluvial aquifer complexes characteristic of the Po plain subsurface: relationships between scale of hydrostratigraphic reconstruction and flow models” (PRIN 2007). References AGIP, 1994. Acque dolci sotterranee. Graf 3, Roma. Alfano, L., Mancuso, M., 1996. Sull'applicabilità del metodo dipolare–dipolare continuo nelle ricerche idriche a media profondità in aree di pianura. Acque Sotterranee 13, 61–71. Arca, S., Beretta, G.P., 1985. Prima sintesi geodetico-geologica sui movimenti verticali del suolo nell'Italia settentrionale (1897–1957). Bollettino di Geodesia e Scienze Affini 2, 125–156. Baines, D., Smith, D.G., Froese, D.G., Bauman, P., Nimeck, G., 2002. Electrical resistivity ground imaging (ERGI): a new tool for mapping the lithology and geometry of channel-belts and valley-fills. Sedimentology 49, 441–449. http://dx.doi.org/ 10.1046/j.1365-3091.2002.00453.x. Baio, M., Bersezio, R., Bini, A., Cavalli, E., Cantone, M., Mele, M., Pavia, F., Losi, E., Rigato, V., Rodondi, C., Sommaruga, M., Zembo, I., 2009. Geological and geomorphological map of the Lodi alluvial Plain: the contribution of surface geology to hydrostratigraphic reconstruction. VII Italian Forum of Earth Sciences Geoitalia 2009: Epitome, 3, p. 5 (ISSN 1972–1552). Benedetti, L.C., Tapponnier, P., Gaudemer, Y., Manighetti, I., Van der Woerd, J., 2003. Geomorphic evidence for an emergent active thrust along the edge of the Po Plain: the Broni–Stradella fault. Journal of Geophysical Research-Solid Earth 108, 2238–2258. http://dx.doi.org/10.1029/2001jb001546. Bersezio, R., Pavia, F., Baio, M., Bini, A., Felletti, F., Rodondi, C., 2004. Aquifer architecture of the Quaternary alluvial succession of the southern Lambro basin (Lombardy, Italy). Il Quaternario 17, 361–378. Bersezio, R., Giudici, M., Mele, M., 2007. Combining sedimentological and geophysical data for high-resolution 3-D mapping of fluvial architectural elements in the Quaternary Po plain (Italy). Sedimentary Geology 202, 230–248. http://dx.doi.org/10.1016/ j.sedgeo.2007.05.002. Bersezio, R., Cavalli, E., Cantone, M., 2010. Aquifer building and Apennine tectonics in a Quaternary foreland: the southernmost Lodi plain of Lombardy. In: Bersezio, R., Amanti, M. (Eds.), Proceedings of the Second National Workshop “Multidisciplinary Approach for Porous Aquifer Characterization”, vol. XC. ISPRA, Memorie Descrittive della Carta Geologica d'Italia, pp. 21–30. Bini, A., Strini, A., Violanti, D., Zuccoli, L., 2004. Geologia di sottosuolo dell'alta pianura a NE di Milano. Il Quaternario 17, 343–354. Binley, A., Cassiani, G., Deiana, R., 2010. Hydrogeophysics: opportunities and challenges. Bollettino di Geofisica Teorica ed Applicata 51, 267–284. Boni, A., 1967. Note illustrative della carta geologica d'Italia, foglio 59. Servizio Geologico d'Italia, Pavia. Bowling, J.C., Rodriguez, A.B., Harry, D.L., Zheng, C.M., 2005. Delineating alluvial aquifer heterogeneity using resistivity and GPR data. Ground Water 43, 890–903. http:// dx.doi.org/10.1111/j.1745-6584.2005.00103.x. Bowling, J.C., Harry, D.L., Rodriguez, A.B., Zheng, C.M., 2007. Integrated geophysical and geological investigation of a heterogeneous fluvial aquifer in Columbus Mississippi. Journal of Applied Geophysics 62, 58–73. http://dx.doi.org/10.1016/j.jappgeo.2006. 08.003. Bratus, A., Santarato, G., 2009. The characterisation of aquifers by means of resistivity investigations. Bollettino Di Geofisica Teorica Ed Applicata 50, 15–28. Bridge, J.S., Hyndman, D.S., 2004. Aquifer characterization. In: Bridge, J.S., Hyndman, D.W. (Eds.), Aquifer Characterization: SEPM Spec Publ, 80, pp. 1–2.

125

Burrato, P., Ciucci, F., Valsenise, G., 2003. An inventory of river anomalies in the Po Plain, Northern Italy: evidence for active blind thrust faulting. Annals of Geophysics 46, 865–882. Carminati, E., Martinelli, G., Severi, P., 2003. Influence of glacial cycles and tectonics on natural subsidence in the Po Plain (Northern Italy): insights from C-14 ages. Geochemistry, Geophysics, Geosystems 4, 1082–1096. http://dx.doi.org/10.1029/ 2002gc000481. Castiglioni, G.B., Pellegrini, G.B., 2001. Note illustrative della Carta Geomorfologica della Pianura Padana, vol. IV. Comitato Glaciologico Italiano, Torino. Constable, S.C., Parker, R.L., Constable, C.G., 1987. Occam's inversion — a practical algorithm for generating smooth models from electromagnetic sounding data. Geophysics 52, 289–300. http://dx.doi.org/10.1190/1.1442303. Conti, A., Sacchi, E., Chiarle, M., Martinelli, G., Zuppi, G.M., 2000. Geochemistry of the formation waters in the Po plain (Northern Italy): an overview. Applied Geochemistry 15, 51–65. http://dx.doi.org/10.1016/s0883-2927(99)00016-5. Cremaschi, M., 1987. Paleosols and Vetusols in the Central Po plain (Northern Italy), vol. 28. UNICOPLI, Milano. Dahlin, T., 2001. The development of DC resistivity imaging techniques. Computers & Geosciences 27, 1019–1029. http://dx.doi.org/10.1016/s0098-3004(00)00160-6. de Lima, O.A.L., Niwas, S., 2000. Estimation of hydraulic parameters of shaly sandstone aquifers from geoelectrical measurements. Journal of Hydrology 235, 12–26. http://dx.doi.org/10.1016/s0022-1694(00)00256-0. Desio, A., 1965. I rilievi isolati della pianura Lombarda ed i movimenti tettonici del Quaternario. Rendiconti dell'Istituto Lombardo 99, 881–894. ENI, Regione Emilia-Romagna, 1998. Riserve idriche sotterranee della Regione Emilia– Romagna. SELCA, Firenze. ENI, Regione Lombardia, 2002. Geologia degli acquiferi padani della Regione Lombardia. SELCA, Firenze. Fantoni, R., Bersezio, R., Forcella, F., 2004. Alpine structure and deformation chronology at the Southern Alps–Po Plain border in Lombardy. Bollettino Della Societa Geologica Italiana 123, 463–476. Hodlur, G.K., Dhakate, R., 2010. Correlation of vertical electrical sounding and electrical borehole log data for groundwater exploration. Geophysical Prospecting 58, 485–503. http://dx.doi.org/10.1111/j.1365-2478.2009.00831.x. Hubbard, S.S., Rubin, Y., 2000. Hydrogeological parameter estimation using geophysical data: a review of selected techniques. Journal of Contaminant Hydrology 45, 3–34. http://dx.doi.org/10.1016/s0169-7722(00)00117-0. Hubbard, S.S., Rubin, Y., 2005. Introduction to hydrogeophysics. In: Rubin, Y., Hubbard, S.S. (Eds.), Hydrogeophysics. Springer, Dordrecht, The Netherlands, pp. 3–21. Huntley, D., 1986. Relations between permeability and electrical resistivity in granular aquifers. Ground Water 24, 466–474. http://dx.doi.org/10.1111/j.1745-6584.1986. tb01025.x. Keller, G.V., Frischknecht, F.C., 1966. Electrical Methods in Geophysical Prospecting. Pergamon Press. Kelly, W.E., Frohlich, R.K., 1985. Relations between aquifer electrical and hydraulic properties. Ground Water 23, 182–189. http://dx.doi.org/10.1111/j.1745-6584.1985. tb02791.x. Koefoed, O., 1979. Geosounding Principles, 1: Resistivity Soundings Measurements. Elsevier, Amsterdam. Kosinski, W.K., Kelly, W.E., 1981. Geoelectric soundings for predicting aquifer properties. Ground Water 19, 163–171. http://dx.doi.org/10.1111/j.1745-6584.1981.tb03455.x. Lines, L.R., Treitel, S., 1984. A review of least-squares inversion and its application to geophysical problems. Geophysical Prospecting 32, 159–186. http://dx.doi.org/ 10.1111/j.1365-2478.1984.tb00726.x. Loke, M.H., 2001. RES1D ver. 1.00a beta. 1-D resistivity, IP and SIP modeling. www. geotomosoft.com. Maxey, G.B., 1964. Hydrostratigraphic units. Journal of Hydrology 2, 124–129. Mele, M., 2009. The Architecture of Alluvial Aquifers: An Integrated Geological– Geophysical Methodology for Multiscale Characterization. Univeristà degli Studi of Milan (PhD thesis). Mele, M., Bersezio, R., Giudici, M., 2012. Hydrogeophysical imaging of alluvial aquifers: electrostratigraphic units in the quaternary Po alluvial plain (Italy). Internation Journal of Earth Science 101 (7), 2005–2025. http://dx.doi.org/10.1007/s00531012-0754-7. Miall, A.D., 1996. The geology of fluvial deposits: sedimentary facies, basin analysis and petroleum geology. Springer, Berlin. Niwas, S., De Lima, O.A.L., 2003. Aquifer parameter estimation from surface resistivity data. Ground Water 41, 94–99. http://dx.doi.org/10.1111/j.1745-6584.2003.tb02572.x. Pellegrini, L., Boni, P., Carton, A., 2003. Hydrographic evolution in relation to neotectonics aided by data processing and assessment: some examples from the Northern Apennines (Italy). Quaternary International 101, 211–217. http:// dx.doi.org/10.1016/S1040-6182(02)00103-9. Pieri, M., Groppi, G., 1981. Subsurface Geological Structure of the Po plain, Italy, vol. 411. CNR, Roma. Ponzini, G., Ostroman, A., Molinari, M., 1984. Empirical relation between electrical transverse resistance and hydraulic transmissivity. Geoexploration 22, 1–15. http://dx.doi.org/10.1016/0016-7142(84)90002-4. Reynolds, J.M., 2011. An Introduction to Applied and Environmental Geophysics. WileyBlackwell. Schön, J., 2004. Physical Properties of Rocks: Fundamentals and Principles of Petrophysics, vol. 18. Elsevier. Singh, S.B., Veeraiah, B., Dhar, R.L., Prakash, B.A., Rani, M.T., 2011. Deep resistivity sounding studies for probing deep fresh aquifers in the coastal area of Orissa, India. Hydrogeology Journal 19, 355–366. http://dx.doi.org/10.1007/s10040-010-0697-7. Sinha, R., Yadav, G.S., Gupta, S., Singh, A., Lahiri, S.K., 2012. Geo-electric resistivity evidence for subsurface palaeochannel systems adjacent to Harappan sites in

126

M. Mele et al. / Journal of Applied Geophysics 93 (2013) 114–126

northwest India. Quaternary International. http://dx.doi.org/10.1016/j.quaint.2012. 08.002. Telford, W.M., Geldart, L.P., Sheriff, R.E., 1990. Applied Geophysics. Cambridge University Press. Tizro, A.T., Voudouris, K.S., Salehzade, M., Mashayekhi, H., 2010. Hydrogeological framework and estimation of aquifer hydraulic parameters using geoelectrical

data: a case study from West Iran. Hydrogeology Journal 18, 917–929. http:// dx.doi.org/10.1007/s10040-010-0580-6. Yadav, G.S., Dasgupta, A.S., Sinha, R., Lal, T., Srivastava, K.M., Singh, S.K., 2010. Shallow sub-surface stratigraphy of interfluves inferred from vertical electric soundings in western Ganga plains, India. Quaternary International 227, 104–115. http:// dx.doi.org/10.1016/j.quaint.2010.05.030.