The Rio do Peixe Basin is located at the NW boundary of Paraíba (PB) with Ceará (CE) states (Figure 1), encompassing the municipalities of Sousa, Uiraúna, Poço, Brejo das Freiras, Triunfo, Santa Helena and Pombal, more precisely between the meridians 37˚47'00'' and 38˚50'00'' west longitude, and between the parallels 06˚25'00'' and 06˚50'00'' south latitude [4] .

Access to this area is done through federal highways, BR-230 (leaving from Pombal to Sousa), and by BR-405 (leaving from Cajazeiras to Uiraúna), or even on state highways, via PB-393, leaving Marizópolis for São João do Rio do Peixe and PB-391 connecting the city of Sousa to Uiraúna [4] .

The Rio do Peixe Sedimentary Basin is one of the main Cretaceous basins of the northeast of Brazil, containing important paleontological sites. It is located in a set of Mesozoic basins where there are productive areas and good indicators of the existence of petroleum [5] .

More detailed geological surveys of the Rio do Peixe Basin have been carried out in recent years. Geochemical data, in addition to the occurrence of oil exudates indicate good conditions for possible accumulation of hydrocarbons, highlighting the need to consider this and other similar basins in the scenario of exploration activities in the Northeast of Brazil [3] .

According to Silva (2009) [6] , the basin is divided into three sub-basins that correspond to the semi-grabens of Triunfo or Brejo das Freiras, Sousa and Pombal, being crossed by three major faults Tatajuba Fault, Portalegre Fault and Malta Fault (Figure 2 and Figure 3).

Figure 4 presents A and B sections shown in Figure 3. The Rio do Peixe Group is composed by Antenor Navarro (lower), Sousa (intermediate) and Rio Piranhas (upper) formations [3] .

The formations are described by Rocha and Amaral (2006) [8] in Table 1.

The relief of the Rio do Peixe Basin is approximately flat. The altitude varies from 174 m to just over 330 m according to the map made using data from the V16.1 global topography model developed by the Institute of Oceanography of the University of San Diego (Figure 5).

It can be observed that there is a gradual increase in the relief from the Pombal sub-basin to the Brejo das Freiras sub-basin. It is also notable that the basin is almost completely surrounded by hills which reach more than 600 meters.

The magnetometer is an instrument used to measure the intensity of the Earth’s magnetic field. This intensity can vary according to the magnetization of rocks in the Earth’s crust that generates a certain disagreement between the actual value of intensity and the theoretical value. This difference is called Magnetic Anomaly, it is the difference of the real and expected magnitude for a given location [9] .

According to Lowrie (2007) [10] , basically, the magnetometric method consists in obtaining the magnetic anomaly through the following steps:

1) Measurement of the Earth’s magnetic field at predetermined points;

2) Correction of measurements for known changes;

3) Comparison of the resulting field value with the expected value at each measuring station. The expected field value anywhere is considered the international reference geomagnetic field (IGRF). The difference between the observed and expected values is called a magnetic anomaly.

According to Sordi (2008) [11] , magnetometry measures small variations in the intensity of the Earth’s magnetic field from rocks that have variable magnetism distributed in the Earth’s crust above the Curie surface. One can calculate the magnetic field of an isolated surface, using the Equation (1):

AF = MF − ( EF + GF ) (1)

at where:

AF = Anomalous magnetic field (<1%);

MF = Measured field;

EF = External magnetic field (Diurnal―External sources to the planet);

GF = Geomagnetic field―IGRF (99%).

The magnetic susceptibility is presented by each rock in different intensities, being the sedimentary rocks with lower intensity, and the igneous and metamorphic ones with greater intensity. Thus, when magnetic anomalies are observed in environments covered by sediment deposition, they are generated by igneous and metamorphic rocks and/or intrusions. The most common causes of anomalies are dikes, sills, metamorphic or igneous foundations and bodies of magnetite ores [1] .

Blakely (1995) [12] states that derivative filters are techniques where the Magnetic Anomaly undergoes a linear modification, through the derivation of the horizontal and vertical components of the magnetic field. The horizontal derivative in the Fourier domain is given by Equation (2).

F [ d ″ T d x ″ ] = ( i k x ) n F [ T ] and to F [ d ″ T d y ″ ] = ( i k y ) n F [ T ] (2)

at where:

k_x and k_y are the wave numbers in the directions x and y.

These filters change in functions of a linear surface in a derivative of order n, with respect to x and y.

The vertical derivative is obtained through the above equations and the Laplace Equation (3).

∂ 2 T ∂ x 2 + ∂ 2 T ∂ y 2 + ∂ 2 T ∂ z 2 = 0 (3)

In frequency domain:

F ( ∂ 2 T ∂ x 2 ) + F ( ∂ 2 T ∂ y 2 ) + F ( ∂ 2 T ∂ z 2 ) = 0

− ( k x ) 2 F ( T ) − ( k y ) 2 F ( T ) + F ( ∂ 2 T ∂ z 2 ) = 0

F ( ∂ 2 T ∂ z 2 ) = ( k x ) 2 F ( T ) + ( k y ) 2 F (T)

F ( ∂ 2 T ∂ z 2 ) = ( k z ) 2 F (T)

As the nth-order vertical gradient is equal to the Fourier transform of the potential times (k_z)ⁿ, in a similar fashion, the vertical derivative of nth-order with wave number k_z is obtained through Equation (4).

F [ ∂ ″ T ∂ z ″ ] = ( k z ) n F [ T ] (4)

at where:

k_z = wave number in the z direction.

According to Roest (1992) [13] , the amplitude of the 3D analytical signal at (x, y) can be simply derived from the total magnetic field in three orthogonal gradients. The amplitude of the analytical signal can be placed at low altitudes where other techniques generally fail as in the case of reduction to the magnetic pole. The analytical signal is made by combining the vertical and horizontal gradients of the magnetic anomaly. The amplitude is given by Equation (5).

| A ( x , y ) | = ( ∂ T ∂ x ) 2 + ( ∂ T ∂ y ) 2 + ( ∂ T ∂ z ) 2 (5)

at where:

| A ( x , y ) | = amplitude of the analytical signal in (x, y);

( ∂ T ∂ x ) , ( ∂ T ∂ y ) and ( ∂ T ∂ z ) = horizontal derivatives in x and y and vertical derivative in z, respectively;

T = magnetic anomaly.

The analytical signal is able to locate the edges of magnetic bodies, regardless of their geometry and the remaining magnetization. Therefore, it is a good locator of magnetic rocks.

Euler Deconvolution was introduced by Thompson in 1982 to provide rapid analysis of large amounts of magnetic data. It is a technique widely used in the interpretation of magnetic data, since it allows estimating the average depth of sources or magnetic bodies [14] .

This technique uses the Euler equation (Equation (6)) to enable estimation of the position, depth, and type of magnetic sources [15] .

x ∂ T ∂ x + y ∂ T ∂ y + z ∂ T ∂ z = N T (6)

at where:

x, y and z = Cartesian coordinates;

T = Magnetic field;

N = degree of homogeneity.

The method can be used to estimate the location of a wide variety of geological structures such as faults, magnetic contacts, dikes, and sills. These structures are classified by a parameter denominated as a structural index, which is related to the nature of the anomaly and is, according to Barbosa and Silva [16] , the measurement of the magnetic anomaly decay rate with the distance between the source and measurement point. Table 2 compares geological models with the structural index for magnetic data.

Ribeiro, Queiroz, and Santos (2015) [15] detail the structural index (I), where:

I = 0, it is used to identify geological contacts, which are regions where there are contact between layers of different rocks;

I = 1/2, it is used to verify regions of fault, shear, and fracture in layers of rocks;

I = 1, it is used to detect intrusive rocks, such as dikes (vertical intrusions) and sills (horizontal intrusions);

I = 2, it is used to identify cylindrical shaped sources, which may be kimberlite chimneys, for instance;

I = 3, it is used as a model to identify spherical sources, such as iron ore agglutinations.

For Thompson (1982) [14] , the most suitable value for a structural index is one that presents a set of solutions with the lowest possible standard deviation. Therefore, the definition of the structural index depends strictly on the type of source on which the depth is estimated.

For each structural index chosen, it is possible to estimate the geographic coordinates of the source that causes the anomaly by solving the Euler equation in a moving window measured at the point of T i ( x , y , z ) and with its derivatives, in an overdetermined system (Equation (7)).

[ ∂ T 1 ∂ x ∂ T 1 ∂ y ∂ T 1 ∂ z ∂ T 2 ∂ x ∂ T 2 ∂ y ∂ T 2 ∂ z ⋮ ⋮ ⋮ ] [ x − x 0 y − y 0 z − z 0 ] = N [ T 1 T 2 ⋮ ] (7)

After selecting the structural index, the dimensions of the movable window must be determined. This window crosses the mesh, producing the set of solutions. The location of the sources is considered optimal when there is little scattering of the solutions, so different windows can generate solutions in similar positions (Figure 6) [18] .

To achieve more reliable results, the size of the window must be sufficient for the inclusion of the anomaly. The noise can distort the understanding of the magnetic sources, so it is necessary data filtering before using them [19] .