IAG/USP TEST SITE: A NEAR SURFACE GEOPHYSICS TEACHING AND RESEARCH LABORATORY

. This work shows the construction project of the Geophysical Test Site (Sítio Controlado de Geofísica Rasa, SCGR-I) of the Institute of Astronomy, Geophysics and Atmospheric Sciences (IAG), of the University of São Paulo (USP) and its impact on teaching and research in Geophysics. The IAG/USP test site (SCGR-I) has 1500 m 2 , being characterized by 7 studies lines with 30 m length in the NS direction. Targets, such as metallic pipes and tanks, plastic pipes and tanks, concrete tubes, ceramic pots, among others, with different geometries and physical properties were buried at depths from 0.5 to 2 m in relation to the surface. A metallic guide pipe of 3.8 cm of diameter was buried at the 15 m position along the EW direction, crossing all 7 lines. Targets simulate objects found in archaeological studies, geotechnical and urban planning studies and environmental studies. In this work, comparative analyzes between real and synthetic GPR results on metallic and plastic tanks are shown, as well as EM38 results on metallic tanks. The SCGR-I proved to be an important tool for teaching and research related to the applications of geophysical methods for near surface investigations and could be a motivation to build more test sites.


INTRODUCTION
Since 1993, São Paulo campus of the University of São Paulo (USP) has been used as a applied geophysics laboratory, but only in 1997 the area in front of the Institute of Astronomy, Geophysics and Atmospheric Sciences (IAG/USP) began to be systematically used as a laboratory for practical activities by undergraduate and graduate students in geophysics.
In these systematic surveys, several geophysical methods were used, such as GPR-Ground Penetrating The importance of this test site is that the geophysical signatures of targets whose physical and geometric properties are known can be used as standard responses for each type of material and can be extrapolated to areas where subsurface information is not available.
SCGR-I constitutes an important tool for teaching and research in geophysics, and will be of great importance to our community, consisting of a new underground laboratory. With the installation of the SCGR-I, an important step was taken to improve the knowledge regarding the geophysical responses of targets found in environmental, engineering and archaeology studies.
To illustrate it, numerical modeling results are presented by the FDTD -Finite Differences in Time Domain method, which simulates the responses from the GPR reflections on the metallic tanks installed on Line 4 and the plastic tanks installed on Line 5 of SCGR-I, as well as the GPR results on the same targets.
Additionally, the results obtained with the inductive electromagnetic method using the EM38 equipment on the metallic tanks are also presented.
The present work summarizes the construction project of the IAG/USP Test Site (SCGR-I), shows the comparative results of numerical modeling GPR 2D and real data, EM38 results over metallic tanks and ends with the impacts on teaching and research activities in near surface geophysics. Geological information for the São Paulo basin in the SCGR-I area was obtained through the lithology of three wells for geological and geophysical research that were drilled in the study area (Porsani et al., 2004).  (Borges, 2007).

SCGR-I of the IAG/USP: Constructive Project
The constructive project of the SCGR-I is presented, aiming to serve as an inspiration for the construction of    Line 2 ( Figure 4) is constituted by brown High Density Polyethylene (PAD -Polietileno de Alta Densidade) pipes with a diameter of 11 cm and 2 m in length. PAD pipes simulate the transport of drinking water to homes, and they are often found in large cities. These pipes are used by the Basic Sanitation Company of the State of São Paulo (SABESP). Figure 5 shows the targets buried in Line 3 which is characterized by concrete tubes of 26, 48 and 70 cm in diameter. The tubes simulate rainwater channeling galleries and sewage drainage.
Line 4 ( Figure 6) is characterized by 200 liter metallic tanks that were arranged both horizontally and vertically, individually and in pairs. All tanks were buried empty to avoid corrosion problems. This line    aims to simulate environmental studies, whose goal is the location and determination of their depths.  The wave field was simulated using an "exploding reflector" source, in which waves are generated simultaneously from the target and sent to the surface (Yilmaz, 1987;Daniels, 1996). This procedure corresponds to the repositioning of the diffraction hyperbolas in the targets, collapsing the energy to the apex of the hyperbola, being a common procedure in the step of GPR and seismic data migration.       It is observed that the top of the metallic tanks is characterized by strong hyperbolic reflections, which was expected, as shown in the numerical modeling result (Figure 10a). Note a hyperbolic reflection at position 19.5 m and arranged at a depth of 1.5 m, highlighted in Figure 10b by an arrow. This reflection, called "artifact", corresponds to a constructive interference of the reflection of the GPR signal between the tank at the 19 m position and the tank at the 20 m position. A detailed discussion of the identification and removal of this artifact through effective processing of the GPR data can be found in Porsani and Sauck (2007). Also note three other hyperbolic reflections ("artifacts") under positions 24, 25 and 25.5 m, being related to voids in the subsoil due to poor soil compaction. These three anomalies were confirmed by means of auger boreholes.

GPR Profiles
The guide metal pipe arranged at 15 m position and at 0.5 m depth is characterized by a tighter hyperbolic reflection. Note that from 2.5 m depth, the GPR signal is attenuated due to the conductive characteristics of the sediments of the São Paulo basin (Porsani et al., 2004). Figure 11 shows the comparison between the results of the 2D GPR numerical modeling and the 200 MHz GPR profile on line 5 of the SCGR-I consisting of plastic tanks. Figure 11a shows the results of the GPR numerical modeling for 150 MHz. It is observed that the plastic tanks and the guide metal pipe are characterized by hyperbolic reflections generated at the top of the targets, whose apex indicates their underground positions. It is noted that the tanks filled with water are characterized by reflections generated at the top and bottom. Additionally, it is also observed that the tanks filled with water and brine present reflections with inverted polarity compared to the top of the empty tanks. Figure 11b shows For the tanks filled with water arranged at positions 4, 17 and 23 m, two reflections are observed at different times.
The first reflector characterizes the top and the second reflector is related to the base of the tank. Note also that the reflectors at the top of the tanks present an inversion of polarity in relation to the reflections generated at the top of the empty tanks due to the high impedance contrast between the clayey soil and the water. A more detailed discussion on identifying the polarity change of the GPR signal can be found in Rodrigues and Porsani (2006).
The half-filled tanks with brine arranged at positions 7 and 26 m were characterized by reflections generated at the soil/plastic/air interface. The upper limit of the brine is not detected, due to the overlap of the reflections at the top of the tank (empty part) and the top of the brine, similar to the empty tanks. The base of these tanks is also not determined due to the high electrical conductivity of brine, causing a high attenuation of the GPR signal.
Tanks filled with brine arranged at positions 10, 13 and 29 m are characterized by reflections with reversed signal polarity generated at the top of the targets, similar to tanks filled with water. Note that the base of the tanks is not detected, due to the attenuation of the electromagnetic wave in brine which is very conductive. The guide metal pipe arranged at a 15 m position and at a depth of 0.5 m, served as a reference target for all seven lines of studies installed at SCGR-I. The top of the metallic pipe is characterized by a strong reflection due to the high electrical conductivity of the metal, causing a total reflection of the GPR signal.

EM38 Profile
Geophysical methods have been important in detecting underground objects such as communication cables, pipes and other public infrastructure. Electromagnetic systems operating in the frequency domain have been shown to be suitable and efficient in detecting buried metallic objects, with several examples in studies of unexploded ordnance detection, urban interference and archaeology (Nelson et al., 2007;Qu et al., 2017).
GPR has been successfully used to detect buried objects and interference networks underground.
However, it has limitations to detect objects buried in conductive saturated clayey soils. On the other hand, the electromagnetic frequency domain system (EM38) does not suffer this limitation in the case of metallic objects buried in conductive environments. Therefore, the integrated application of GPR and EM38 can be complementary, improving the ability to detect metallic targets arranged in conductive soils.
In this work, Geonics EM38 equipment was used to detect metallic tanks buried in SCGR-I. A profile of 30 meters in length was acquired with measurements spaced of 1 meter. The equipment allows a maximum theoretical investigation depth of 1.5 meters, being indicated only for mapping shallow targets. The measurements were performed with the coils in the vertical and horizontal positions, so that two theoretical depths were obtained at each station. Data collected at two investigation depths were entered into inversion program which provided a 2D section of the conductivity profile.
For the interpretation of the EM38 profile on the metallic tanks, the inversion program called EM34-2D was used (Monteiro Santos, 2004). This program uses the non-linear inversion algorithm presented in Sasaki (1989). The algorithm uses a smoothness constraint regularized inversion technique for electromagnetic data acquired along profiles. The algorithm corresponds to a modified 1D inversion with 2D smoothness constraints between adjacent 1D models. Thus, it is possible to obtain the answer in terms of the variation of electrical conductivity and real depths for the measurement points that, interpolated, allow the creation of a 2D image.  vertically at the depths where the equipment has reach were clearly detected by anomalies of high electrical conductivity values. Tanks that are installed up to 1.5 m deep were detected with good accuracy. This shows that EM38 can be very efficient in detecting buried metallic objects up to 1 meter deep, and its application integrated with GPR is interesting, especially in conductive soils. Note that the guide metal pipe installed at 0.5 m depth was not clearly detected. This fact is due to its small dimensions, i.e., 3.8 cm in diameter, which is below the detection limit of the EM38 equipment. Among the studies published in the literature are: Porsani et al. (2004Porsani et al. ( , 2006Porsani et al. ( , 2010Porsani et al. ( , 2017Porsani et al. ( , 2018; Rodrigues (2004); Lima (2006); Rodrigues and Porsani (2006); Porsani and Sauck (2007)