Theory
Geoelectrical resistivity methods have been widely used in groundwater exploration, engineering and environmental applications (Dahlin 1996; Keller and Frischknecht 1996; Sudha et al. 2009; Galazoulas et al. 2015; Lech et al. 2020). Electrical resistivity tomography (ERT) is one of the most commonly used geoelectrical resistivity methods to map subsurface electrical resistivity (Griffiths and Barker, 1993; Dahlin, 2001), which subsequently can be interpreted from a hydrological perspective (e.g., Saad et al. 2012; Ashraf et al. 2018; Aziman et al. 2018; Riwayat et al. 2018; Thiagarajan et al. 2018; Kumar et al. 2020a, b).
The electrical resistivity method estimates the variation in the ground's resistivity by injecting direct current into the ground (Fig. 4) through a set of current electrodes (C1 and C2) and measuring the resulting voltage differences at the potential electrodes (P1 and P2). The resistivity is determined from Ohm’s law using the voltage differences for a known current and correcting the current geometrical pathway through the earth. From the current (I), voltage (V) and geometric factor (k), the apparent resistivity (pa) was calculated using the following Eq. 1 (Telford et al. 1990; Dahlin 2001; Binley et al. 2015):
(1)
Resistivity interpretations have proved helpful for detecting geological units of unconsolidated sediments and groundwater prospects (e.g., Saad et al. 2012; Ashraf et al. 2018; Aziman et al. 2018; Riwayat et al. 2018; Thiagarajan et al. 2018; Kumar et al. 2020a, b). Igneous and metamorphic rocks, depending on the degree of fracturing and the percentage of the fractures filled with groundwater, generally have higher resistivities than sedimentary rocks, which are usually more porous and have higher water content. For example, the resistivity of granite ranges between 5,000 to 10,000 ohm-m depending on the degree of fracturing and moisture content (e.g., Kumar et al. 2020a, b), whereas the resistivity of sand and clay materials ranges within 1 to 1,000 ohm-m (e.g., Saad et al. 2012; Ashraf et al. 2018). Furthermore, clay has significantly lower resistivity values than sand. The resistivity values of clay range between 1 to 100 ohm-m, whereas the resistivity values of sand range between 60 to 1000 ohm-m (Keller and Frischknecht, 1996). The degree of overlap in different resistivities of different types of materials and waters is dependent on several factors such as porosity, degree of water saturation and concentration of dissolved salts (Samouëlian et al. 2005; Hazreek et al. 2015; Annuar and Nordiana 2018). Keller and Frischknecht (1996) found that groundwater yields low resistivity values ranging from 10 to 100 ohm-m depending on the concentration of dissolved salts.
Due to the ambiguity of resistivity values with overlapping resistivity ranges, several researchers have used an integrated geophysical approach combining resistivity with the induced polarization (IP) technique to solve complicated geological and hydrological problems (e.g., Goldman and Neubauer 1994; Amaya et al. 2016; Kumar et al. 2016a, b; Rehman et al. 2016). The IP method uses the voltage decay characteristics to study the soil’s induced polarization, also known as the chargeability (Telford et al. 1990; Keller and Frischknecht 1996; Kearey et al. 2002). Subsurface chargeability measurements have been used to determine high clay content in sedimentary settings. The clay particles have a negative charge that can attract positive ions from the electrolyte contained in the cavities of rocks (Telford et al. 1990). High chargeability measurements have been observed for fine-grained sediments such as clay, while deposits with larger particle size like sand and gravel typically yield lower chargeability values (Keller and Frischknecht 1996; Slater and Lesmes 2002; Alabi et al. 2010; Amaya et al. 2016).
Hydraulic tests of groundwater wells are crucial for understanding the aquifer potential in any hydrogeological setting (Ashraf et al. 2018; Aziman et al. 2018; Kumar et al. 2020b). Transmissivity and hydraulic conductivity are two key hydraulic parameters that can help determine the aquifer characteristics (e.g., Mogaji et al. 2011; Shen et al. 2015; Wu et al. 2017). Furthermore, the records of time-drawdown data have been used to evaluate the aquifer transmissivity and hydraulic conductivity. The Cooper-Jacob solution assumes that the aquifer is confined, homogenous, isotropic and of uniform thickness over the area of pumping. The assumption discussed in Fetter (2001) for determining aquifer parameters from time-drawdown data is that the pumping well is screened throughout the entire thickness of the aquifer being tested. Using the Cooper-Jacob straight-line time-drawdown method, the aquifer transmissivity was calculated using Eq. 2 (Cooper and Jacob 1946):

(2)
Where Q is the pumping rate and ho - h is the calculated change in drawdown between initial and final drawdown.
From the calculated transmissivity (T) and the aquifer thickness (b), the hydraulic conductivity (K) was calculated using Eq. 3:

(3)
Data acquisition and processing
Data acquisition was conducted with a combined measurement of resistivity and chargeability. A total of 19 survey lines from eight agricultural sites were carried out in the study area (Table 1). Information such as aerial, topographical and geological maps is required when considering resistivity surveying suitability (to select profile lines). The selection of survey locations was predominantly planned based on area availability and accessibility. The survey locations were generally free from obstacles, such as houses, crops or fences. In some areas, permission from landowners was needed to perform the surveys on their land.
Table 1 Survey parameters for each resistivity survey line (sites 1-8).
Survey location
|
Survey line
|
Electrode
configuration
|
Electrode spacing (m)
|
Survey length
(m)
|
Site 1
|
1 - 3
|
Gradient
|
10
|
800
|
12
|
Pole-dipole
|
5, 10
|
400
|
19
|
Pole-dipole
|
5, 10
|
400
|
Site 2
|
4 - 8
|
Gradient
|
10
|
800
|
13
|
Pole-dipole
|
5, 10
|
400
|
Site 3
|
14
|
Pole-dipole
|
5, 10
|
400
|
Site 4
|
15
|
Pole-dipole
|
5, 10
|
400
|
Site 5
|
16
|
Pole-dipole
|
5, 10
|
400
|
Site 6
|
9
|
Gradient
|
10
|
800
|
Site 7
|
10 - 11
|
Gradient
|
10
|
800
|
Site 8
|
17 - 18
|
Pole-dipole
|
5, 10
|
400
|
Two sets of data were acquired between the years 2018 and 2020. The first set of data acquisition was carried out in 2018 using the ABEM Terrameter LS2 resistivity meter covering a lateral distance of 800 m with eighty stainless-steel electrodes arranged at an equal length of 10 m (Line 1–11). This set of data acquisition was conducted using the gradient array configuration. The gradient array configuration uses two current electrodes and two potential electrodes, placed with equal spacings (Fig. 4). The gradient method is suitable for multichannel acquisition due to dense and fast data point collection (Dahlin and Zhou 2006; Loke 2012; Aizebeokhai and Oyeyemi 2014). With the multichannel acquisition, resistivity measurements continue down the row of electrodes until the whole survey line is measured. The second set of data acquisition was conducted from 2019 through to 2020 using the ABEM SAS4000 resistivity meter covering a lateral distance of 400 m (Line 12–19). The line uses sixty-one electrodes with 5 m spacings for the inner cables and 10 m spacings for the outer cables. This set of data acquisition was conducted using the pole-dipole array configuration. For the pole-dipole array configuration, one transmitting current electrode, also known as the infinity electrode, was moved to an effective infinity distance, approximately five times the survey depth (Fig. 4). Simultaneously, the other current electrode is placed in the vicinity of the two potential electrodes. The pole-dipole method is suitable for deep earth investigation, making it a popular option among researchers (Saad et al. 2012; Annuar and Nordiana 2018; Ashraf et al. 2018; Kumar et al. 2020a).
Raw data were processed and inverted using the ZONDRES2D software (in DAT format). Resistivity and chargeability inversion models were obtained from the inversion process (Loke and Barker 1996). The inversion process averaged the resistance measurements to apparent resistivity values and the apparent chargeability time window measurements to integral chargeability values. The Gauss-Newton inversion method was used to determine the appropriate resistivity values (Griffith and Barker 1993; Loke and Barker 1996; Dahlin 2001; Loke and Dahlin 2002). 2D pseudo-sections were generated to help delineate subsurface geological structures, formations and aquifer zones in the surveyed areas. All the pseudo-sections displayed present RMS errors of not more than 5%, which indicates the measured data are fitted with the computed apparent resistivity; the number of iterations for each survey was ten.
Groundwater well drilling and hydraulic tests
Groundwater wells were drilled at site-1 in the Brunei-Muara District (Well-B1) and site-5 in the Belait District (Well-L1) through the aquifer zones inferred from resistivity and chargeability interpretations. Borehole drilling was conducted using a straight rotary method. The hole was advanced by rotating a drill string consisting of a series of hollow drill rods to the bottom attached to a 10-inch drill bit. Water-based drilling fluid under pressure was introduced into the bottom of the hole through the hollow drill rods and passages into the bit. The drilling fluid served the dual function of cooling the rotating bit as it entered the borehole and removing the rock cuttings from the bottom of the hole. The rock cuttings move through the annular space between the drill rods and the walls of the hole as they returned to the surface. Rock cuttings were collected at an interval of 3 m, primarily for soil identification purposes. 6-inch diameter UPVC casings and screens with 1.5 mm openings were used to construct the groundwater pumping wells. A gravel pack filter was installed between the aquifer and UPVC screens.
A 4-inch submersible pump was installed inside the well to continuously pump water out from the well through a 2-inch riser pipe for hydraulic testing. The pumping well responses in terms of the water discharge and changes in water depth were recorded using a volume meter connected at the outlet pipe on the surface and a water depth meter installed inside the well annulus. None of the newly drilled water wells had yet a nearby observation well that could have been used for time-drawdown observation due to pumping. Hydraulic tests of the newly drilled wells were investigated using time-drawdown data gathered from the pumping well. A five-step drawdown test with different flow rates, constant discharge test and recovery test was carried out in this investigation.