In this research, the electrical resistivity tomography (ERT) and time domain electromagnetic (TDEM) methods were applied to a hydrogeological study in Taubaté basin, São Paulo State, Brazil (Figure 1), where the Taubaté aquifer is an important water supply for the region.
Each method has advantages and limitations in terms of investigation depth and resolution. The ERT method is normally used in shallow investigations (tens of meters) due to logistical limitation because it requires a long aperture to achieve greater depths. On the other hand, the TDEM method has a low resolution for the shallow layers; it is used in deep investigations (hundreds of meters) and can define conductive layers   .
Both methods are widely applied in hydrogeophysical studies, because of rapid data acquisition, relatively low cost, and reliability. They have the potential to identify aquifers, aquicludes, and aquitards, which normally, in Brazil, are more conductive than the surrounding layers. They are also sensitive to geological properties, like clay content.
Electrical resistivity method     and TDEM  -  have been successfully applied individually and also jointly      for hydrogeophysical investigations in many places around the world, including Brazil         . Their applications also include mining, geotechnical and environmental studies   , among others.
The study aims to characterize the geoelectrical stratigraphy of the subsurface to locate the contact of the Quaternary and Tertiary sediments and identify the Taubaté aquifer. Furthermore, it aims to contribute to the geoelectrical methods interpretation in a hydrogeological context like Taubaté basin and to collaborate with the hydrogeophysical studies in São Paulo State.
2. Study Area
The studied area is in Taubaté city, São Paulo State, Brazil, which is located on the central portion of the Taubaté Basin (Figure 1(a)). Taubaté is a medium-sized city with about 308,000 inhabitants, where the demand for water has been increasing.
The data acquisitions were conducted on a cattle farm around 6 km from the city center. The terrain was practically flat, which has the advantage of not requiting topography correction on data processing.
Figure 1. Study area location. (a) Taubaté Basin and Taubaté city, São Paulo state, Brazil; (b) Study area and wells (40); (c) ERT line and TDEM loop.
2.1. Geological Setting
According to  , Taubaté basin is the largest basin of the Continental Rift of Southeast Brazil (CRSB) with 170 km length and 20 km width. The basin is elongated in NE-SW direction and presents normal faults in NW-SE direction. The basin is related to Tertiary extensional tectonics and it is characterized by a series of half-grabens with 850 m of maximum sediments thickness  .
Figure 2 presents a geological map and the stratigraphic chart of the basin. It was developed over Precambrian gneisses and granites  . The sedimentary fill is basically continental and can be divided in two phases: the first, syntectonic to the rift, with the deposition of the Taubaté Group; and the second, posterior to diastrophic tectonics, with the deposition of the Pindamonhangaba Formation and alluvial and colluvial deposits  .
The main sedimentary package is the Taubaté Group, which is subdivided by the Resende, Tremembé and São Paulo formations. Taubaté Group is formed mainly by shales and sandstones. Resende Formation is the most abundant package of the Taubaté Group  . Pindamonhangaba Formation was deposited in the Neogene in a fluvial meandering environment. Finally, in the Quaternary, there are alluvial, colluvial, colluvial-alluvial and talus deposits.
Figure 2. Geological map and stratigraphic chart of Taubaté basin (adapted from  ).
2.2. Taubaté Aquifer
Taubaté Aquifer occurs mainly in two areas of the basin, in the southwest and in the northeast. Between these two regions, where Taubaté city is located, there is a compartment filled predominantly by argillites and shales with low permeability, which presents aquiclude characteristics  . As a result of its depositional environments, the aquifer is a multilayer type, with alternation of sandy or aquifer layers, associated with fluvial facies, and clayey or confining layers, associated with lacustrine or floodplain facies  . In Taubaté city, the saturated thickness of the aquifer varies from 200 m to 300 m. According to  , the central region of the basin, where Taubaté is located, presents unfavorable characteristics for groundwater exploration, with flow rates lower than 10 m3/h.
There are 40 wells in Taubaté city available in the Groundwater Information System (SIAGAS) of the Geological Service of Brazil (CPRM)  database, which include the coordinates, depth, water use and a geological profile for each well. Figure 1(b) shows the location of the wells. All wells are tubular and the water is mainly used for domestic or industrial supply. Some of them present the water level and flow rate, which in general are lower than 10 m3/h, being consistent with  . Six wells reached the basement depth, which is formed by gneisses.
The closest well is around 750 m distant and the deepest (W40) (that reaches 650 m depth) is around 9.5 km distant from the study area. Table 1 shows W40 lithological description. Down to 11 m depth represents the Quaternary sediments. The sandstone between 11 m and 16 m depth probably represents the Pindamonhangaba formation. The rocks between 16 and 510 m depth represent the Taubaté Group, which is formed mainly by shales, and bellow 510 m depth is the basement formed by gneisses.
Table 1. Lithological description of the deepest well (W40) in Taubaté city  .
In general, from all wells lithological descriptions, the Taubaté Group is formed mainly by shales intercalated with thin layers of sandstones. Therefore, in the Taubaté Group, the shales form aquicludes and the sandstones form the aquifer.
3. Geophysical Methods
In this research were applied two geophysical methods: Electrical Resistivity Tomography (ERT) and Time Domain Electromagnetic (TDEM). Both methods investigate the same physical property of subsurface materials, the electrical resistivity (ρ).
In the ERT method, the investigation is done through the injection of electric currents and, in the TDEM method, through the electromagnetic induction in conductive materials in the subsurface. Both methods use artificial sources and the response is measured at the surface. Thus, it is possible to relate the electrical resistivity distribution with the geology or variations in the lithological composition, for example, the presence of water, fractures, and mineralogy.
3.1. Electrical Resistivity Tomography (ERT)
The ERT method consists of injecting an electric current (I) into the ground through metallic electrodes and measuring the resulting potential (ΔV) by other electrode pairs. In this way, it is possible to obtain the electrical resistivity distribution of the subsurface.
Electrical resistivity values can be estimated because the spatial arrangement of the electrodes is known. According to  , the apparent electrical resistivity ( ) is given by:
where K is the geometric factor, which depends on the electrodes array. The investigation depth depends on the subsurface resistivity and the electrodes separation, i.e., greater depths are achieved by increasing the array size.
The ERT is used to map the lateral variation of the resistivity as a function of depth (2D). There are several types of electrode arrays, however, in this research was only used the Pole-Dipole array with six investigation levels.
The ERT data was acquired in June 2016 with the Syscal Pro (Iris Instruments) equipment. The imaging line was 400 m length (Figure 1(c)) and with 20 m of electrodes spacing. The data were inverted with the RES2DINV software  , which uses the field data to automatically determine a two-dimensional model of the subsurface resistivity.
3.2. Time Domain Electromagnetic (TDEM)
The TDEM method is based on the electromagnetic induction principle and it is used to estimate the resistivity variation as a function of depth   . The investigation is performed by measuring the decay of an induced secondary magnetic field in the subsurface due to the variation of a primary magnetic field generated on the surface.
According to  , the relationship between the secondary magnetic field variation ( ) and the apparent resistivity ( ) is given by:
where I (A) is current, a (m) the loop radius, the vacuum magnetic permeability and t (s) the time.
There are several acquisition arrays and field procedures. The array configuration, size, and other parameters depend on the purpose of the research. In this research, the central loop array (Figure 1(c)) was used, where the receiver coil is placed in the transmitter loop center. It is one of the most popular TDEM arrays, having the advantage of a good signal-to-noise ratio.
The TDEM method is sensitive to conductive layers, because of the current flows that are induced in these layers. The investigation depth depends on the subsurface resistivity, the more resistive the medium, the faster the secondary field diffuses and vice versa  . The investigation depth also depends on the magnetic dipole moment ( ) of the transmitter loop. Therefore, increasing M, it is possible to reach greater investigation depths. In this research, the acquisition was done with and 200 × 200 m square transmitter loop (Figure 1(c)).
This method has an insignificant influence from natural sources noise. The main noise sources are man-made artifacts such as power transmission lines, cables and buried pipes and metal fences near the data acquisition area  . The transmitter induces electric currents in these conductors which, as consequence, interfere in the induced secondary currents in subsurface materials. No noise source influenced the acquisition, because the electric power lines and metal fences were more than 100 m away, which is considered a safe distance  .
The data were acquired in April 2016 with the PROTEM 57-MK2 D  equipment, which consists of a transmitter that is connected to a power generator to produce the primary electromagnetic field; and a 3D receiver coil of about 1 m diameter and 200 m2 of effective area that is connected to a computer to record the signal of the secondary electromagnetic field induced in the subsurface. The main acquisition parameters were: repetition rates of 30 Hz, 7.5 Hz and 3 Hz, integration time of 30 s and average of three points measured for each repetition rate.
The data were inverted with the Curupira software  . The inversion process consists of determining the electrical resistivity and thickness of the subsurface layers from the measured data.
4. Results Analysis
Figure 3 shows the ERT inversion result, where: a) the measured apparent resistivity pseudosection, b) the calculated apparent resistivity pseudosection and c) the inverted geoelectric model. The RMS error is 1.59% after 6 iterations. The black line in the inverted model (Figure 3(c)) represents the contact between the Quaternary and Tertiary sediments. The Quaternary sediments are formed by soil, alluvial and colluvial deposits going down to maximum around 30 m depth in the left side; it is a very resistive zone, with the resistivity varying from more than 200 Ω∙m to 80 Ω∙m. Below that, the Pindamonhangaba Formation is formed by sandstones and presents resistivity between the 80 Ω∙m and 30 Ω∙m. The white line represents the interface between the Pindamonhangaba formation and the Taubaté Group, which is very conductive, between 30 Ω∙m and 10 Ω∙m.
Figure 4 shows the TDEM inversion result. Figure 4(a) presents the apparent resistivity (Ω∙m) versus time (ms) showing the measured data and the adjusted curve; and Figure 4(b), the inverted geoelectric model, depth (m) versus resistivity (Ω∙m). The adjustment error is 1.7% after 1000 iterations.
Figure 3. ERT inversion result. (a) Measured apparent resistivity pseudosection; (b) Calculated apparent resistivity pseudosection; c) Inverted geoelectric model.
Figure 4. TDEM inversion result. (a) Feild data points with the adjusted curve on Curupira software; (b) The inverted geoelectric model.
In the geoelectric model, there are 3 geoelectrical layers down to 300 m depth. The first layer down to 13 m is resistive, with more than 200 Ω∙m, representing the Quaternary sediments. The second layer, from 13 m to 30 m depth, has a resistivity around 100 Ω∙m, which represent the Pindamonhangaba formation. The third layer, from 30 m to around 300 m is very conductive, around 10 Ω∙m, representing the Taubaté Group. The TDEM result is consistent with the ERT result.
The Taubaté Group is very conductive, therefore the signal concentrates and dissipates in this layer, not being able to induce an electromagnetic field in the layer below, which is the resistive basement. The basement is formed by gneisses, which is a very resistive rock, with resistivity values in the order of thousands of Ω∙m  .
Table 2 summarizes the ERT and TDEM results interpretation. The first layer is interpreted as the Quaternary sediments, the second layer is the Pindamonhangaba formation and the third layer represents the Taubate Group. The resistivity values were estimated based on the ERT (Figure 3(c)) and the TDEM (Figure 4(b)) inverted geoelectric models. These results are coherent and they agree with wells information.
The results have shown a great potential for the application of the combined ERT and TDEM methods to the geoelectrical characterization of the Taubaté Basin, making it possible to identify the contact between the Quaternary and Tertiary sediments and to define the Taubaté Group, which stores the Taubaté aquifer.
Both data inversions obtained a low RMS error (<2%), resulting in a geoelectrical model coherent with the known geology.
Table 2. Interpretation of the ERT and TDEM inverted geoelectric models.
The ERT was able to image the shallow sedimentary layers until 50 m depth and the TDEM provided information until about 300 m depth, allowing defining the conductive layer represented by the Taubaté Group.
The Taubaté Group is approximately 270 m thick, mainly formed by shale and thin sandstone layers. The shales form the aquiclude and the sandstones form the aquifer. Both are very conductive, so the TDEM method identified them as a single conductive layer.
The Taubaté Group is a very conductive layer and, below it, the basement is very resistive. Therefore, in this research, the TDEM method was not able to pass through the sedimentary conductive layer; however, based on the well information, the basement top is probably around 300 m depth, but more studies are needed.
Both geophysical methods have the advantage of rapid data acquisition, relatively low cost, and they are reliable to identify aquifers and aquicludes. Other geophysical studies have already been done in Taubaté basin. For example, Cogné et al.  have reinterpreted 11 seismic profiles of the basin to study its tectonic setting; and Padilha et al.  have applied the magnetotelluric method across the basin to obtain a deep (20 km) 2D geoelectrical model, which identified the conductive sediments of Taubaté Group with low resolution. However, the present research is the first in terms of a hydrogeophysical study in Taubaté basin.
RCR and LRH thank CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior) for providing the research scholarships. JLP thanks to FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo) (grants: 2009/08466-3 and 2012/15338-4) and CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) (grants: 301692/2013-0 and 406653/2013-5), both for providing financial support to develop this research. CAB thanks to CNPq for the Postdoctoral scholarship (150230/2016-8). We thank IAG/USP for providing the infrastructure support. We thank Ernande Costa Santos, Marcelo Cesar Stangari and the students for helping in geophysical data acquisition.