Electrical resistivity tomography and induced polarization study for groundwater exploration in the agricultural development areas of Brunei Darussalam

Electrical resistivity tomography and induced polarization studies were conducted for groundwater exploration in eight selected agricultural sites in Brunei Darussalam. The study was undertaken to meet the growing demands for water supply in the agricultural sector, particularly for paddy field irrigation purposes. This study conducted nineteen surveys with lengths of up to 800 m and a depth of investigation of up to 150 m below ground level. 2D inverted resistivity and chargeability models delineated subsurface geological formations and aquifer zones in the study area. Aquifer zones were detected in all the investigated sites, with resistivity values ranging from 1 to 100 Ω m and chargeability values ranging from 0 to 10 ms. Based on the resistivity and chargeability interpretations, two new groundwater wells were drilled and constructed in two of the investigated sites. Borehole drilling encountered aquifer zones primarily in sand and sandstone layers. Hydraulic tests of the newly drilled boreholes revealed groundwater yields of 4.3 and 288 m3/day. Transmissivity values of the aquifer units estimated using the Cooper–Jacob time-drawdown solutions are 0.53 and 109 m2/day, while their hydraulic conductivity values are 0.05 and 2.75 m/day, respectively. Our findings suggest weak to moderate groundwater yield for withdrawal and distribution for irrigation purposes in the investigated sites. The present study helped decision-makers take suitable measures for placing future groundwater irrigation wells and achieve significant groundwater exploration results in the study area and other areas with similar geological settings.


Introduction
Groundwater is a vital source of freshwater, particularly when surface water resources are limited or of poor quality. Groundwater is mainly used for domestic, industrial, and agricultural purposes. Countries like the United States, China, India, Iran, and Pakistan account for more than 60% of all groundwater abstractions globally each year (Dalin et al. 2017). Groundwater use for agriculture, in particular, has grown exponentially over the past decade, especially in heavily populated areas such as South Asia, Africa and China (Shah et al. 2006). The increased usage in these and other regions has been critical in maintaining agricultural productivity and supplying food for the growing population (Giordano and Villholth 2007). Future population growth is also expected to increase demand for fresh groundwater resources, making them more vulnerable (Essink 2001).
Water resources are under increasing pressure worldwide, both in terms of sustainability and quality. Water shortages, groundwater withdrawal, overexploitation, contamination, ecological degradation and water insecurity due to climatic unpredictability pose a significant challenge to society (e.g., Custodio 2002;Llamas and Martínez-Santoz 2005;Dalin et al. 2017). In countries like China and India, concerns over groundwater quality are rising due to improper disposal of urban wastewater into natural streams or its reuse for irrigation, which ultimately seeps into groundwater (Watto et al. 2018). On the other hand, complex contamination scenarios may need a range of remediation and assessment measures (Marshall et al. 2019). Furthermore, microbial activities can eliminate even the most recalcitrant pollutants (Gӧdeke et al. 2008).
Other studies discussed the impacts of climate change on groundwater recharge (Ranjan et al. 2006;Green et al. 2011;Provenzale and Palazzi 2015). According to Thornton et al. (2014), climate variability causes greater unpredictability in precipitation, including heavier rainfall and drought periods. Therefore, climatic variables can significantly impact groundwater recharge rates, limiting the amount of groundwater available. A recent study in Brunei found that climate changes were evident from an increased rainfall intensity, pointing to the need for careful water management under a changing climate (Gödeke et al. 2020). In addition, surface irrigation practices have been regarded as a significant source of recharge to the nearby groundwater in irrigated agricultural lands (Jia et al. 2020;Tewabe et al. 2021).
Non-intrusive geophysical methods are appropriate exploration and imaging tools for evaluating the spatial variations in different subsurface physical properties from which lithological, hydrogeological, and structural information can be deduced. In particular, the electrical resistivity tomography (ERT) is an established method to produce 2D and 3D estimations of electrical resistivity variations of subsurface materials (Telford et al. 1990;Keller and Frischknecht 1996;Kearey et al. 2002). Studies show that electrical tomography surveys can be effectively used for groundwater exploration (e.g., Ashraf et al. 2018;Aziman et al. 2018;Riwayat et al. 2018;Thiagarajan et al. 2018;Kumar et al. 2020a, b). Other studies show the application of this method in engineering, geotechnical and environmental investigations (e.g., Dahlin 1996;Sudha et al. 2009;Galazoulas et al. 2015;Lech et al. 2020). Variations in resistivities also depend on the type of rock materials, which vary in lithology, porosity, water saturation, and salt concentration (Samouëlian et al. 2005;Hazreek et al. 2015;Annuar and Nordiana 2018).
The induced polarization (IP), as a complementary method to ERT, is often used to identify the electrical chargeability variations of subsurface materials (Telford et al. 1990;Kearey et al. 2002;Binley et al. 2015). The IP method was initially designed to detect mineral ores for mining activity (Telford et al. 1990). However, in the past decade, the IP method has also increasingly been used, jointly with resistivity, to solve complicated geological and hydrological problems (e.g., Goldman and Neubauer 1994;Kumar et al. 2016a, b;Rehman et al. 2016;Defourny et al. 2020). In a study by Alabi et al. 2010, the IP data were effectively used to distinguish between sediments of different lithological compositions, particularly by segmenting subsurface images into relatively clay-free and clay-rich zones. It was found that with careful interpretation, the IP method can improve understanding of the subsurface properties and reduce the ambiguity in interpretation inherent in other geophysical methods such as resistivity (Slater and Lesmes 2002).
In addition to geophysical surveying, this study investigates soil lithology and aquifer characteristics using borehole drilling and testing. The constant rate pumping test and recovery test are the most commonly used methods for determining the soil hydrogeological characteristics (e.g., Shen et al. 2015;Wu et al. 2017;Ashraf et al. 2018;Aziman et al. 2018;Kumar et al. 2020b). Hydraulic parameters such as transmissivity and storativity can be obtained by matching observed time-drawdown data using an analytical method for non-steady flow (Cooper and Jacob 1946).
This study aimed to investigate groundwater potential in the agricultural development areas used for rice farming in tropical Brunei Darussalam. The country relies almost entirely on surface water resources for its agricultural use, which is still insufficient and facing increasing demands from a growing population. Groundwater supplies are being considered to supplement these irrigation needs, but the groundwater potential in these areas is not well studied. This study utilized the ERT and IP methods to delineate subsurface geological formations and aquifer zones. Apart from identifying suitable drilling locations for groundwater wells, the aim was to conduct hydraulic testing and to determine the aquifer characteristics for irrigation purposes.

Location and climate
Brunei Darussalam, or simply known as Brunei, lies on the north coast of Borneo Island in Southeast Asia (Fig. 1). Brunei has a total land area of 5765 km 2 and is divided into four main districts. The Brunei-Muara, Tutong, and Belait Districts make up the country's western half, characterized by hilly lowlands, swampy plains and alluvial valleys. The Temburong District makes up the eastern half, characterized by mountainous terrains (FAO 2011). In the present study, survey locations, sites 1-8, are located within lowland agricultural areas, mainly used for paddy cultivation. Two sites were chosen from the Brunei-Muara District, two sites from the Tutong District, one site from the Belait District and three sites from the Temburong District (Fig. 1).
The climate of Brunei is typical of the equatorial tropics, characterized by high rainfall and temperatures throughout the year. Rainfall shows a seasonal pattern with wet and dry seasons. The wet seasons are from September to January and May to June, and the drier seasons are from February to March and July to August. The average annual rainfall from 2019 was 2909 mm. The temperature is relatively uniform throughout the year, ranging from 23.8 to 32.1 °C (BDMD 2021).

Geological and hydrogeological setting
The country is drained by four main river basins: Brunei, Tutong, Belait, and Temburong rivers (Fig. 1). The Belait river basin is the largest, with an area of 2,130 km 2 . Peat swamp forests dominate the lower catchment areas, with some areas along the river and in the upper catchment area cleared for agricultural development. The Tutong river basin has an area of 930 km 2 , and the basin comprises a floodplain in the lower catchment areas. The upper catchment is forested, with few areas cleared for agricultural development. The Temburong river basin has an area of 840 km 2 with a higher topography than all the other river basins. Finally, the Brunei river basin has an area of 330 km 2 . The upper parts of this river are a primary freshwater source for urban water. The effects of development and urban runoff are also significant in this river basin (Chuan 1992).
The geology of Brunei is closely linked to its neighbouring Malaysian states of Sarawak and Sabah, and many regional geological studies have been conducted (Liechti et al. 1960;Wilford 1961;Sandal 1996;Hutchison 2005). Tectonic events have governed the geological setting in this region since the Cenozoic era (Hall 1997;Hall and Nichols 2002;Baillie et al. 2004). Overall compressional tectonics in the northwest Borneo margin formed deformation zones of mountainous terrains extending through central Borneo. Subsequent uplift and erosion of this mountainous range in the hinterland during the Early and Middle Miocene resulted in rapid sedimentation into basin depocenters, forming major deltaic systems around the island (Hutchison 2005). The Baram delta system of Brunei has been a major source of siliciclastic sedimentation in both the onshore and offshore Brunei areas since the Miocene period (Sandal 1996;Saller and Blake 2003;Torres et al. 2011). Most sediments are gently deformed due to occasional compressional tectonics during the Miocene to Pliocene (Morley et al. 2003). Quaternary rock formations overlay the Liang, Miri, Seria, Lambir, Belait, Setap Shale and Meligan Formations (Fig. 2). The formations are mainly alternating sands and shales, with coal occurrences recorded in the Liang and Belait Formations (Sandal 1996;Osli et al. 2021).
According to Chuan (1992), the most productive groundwater formation is the Liang formation, where the groundwater moves through the lightly consolidated sands. Possible groundwater sources in the study area may be within sandstones and joints in the existing formations. Limited groundwater reserves have been identified in the Liang and Labi areas of the Belait District and Berakas areas of the Brunei-Muara District (Chuan 1992;FAO 2011;Azhar et al. 2019). The groundwater recharge rates in Brunei can reach values over 800 mm/y, and groundwater levels are generally less than 20 m (Moeck et al. 2020).

Water use and irrigation system
In Brunei, the government relies almost entirely on surface water resources, given the relatively sparse population of about 430,000 (DEPS 2021). Surface water accounts for 99.5% of the total water supply, with 0.5% from groundwater resources. Groundwater abstraction is currently limited to the local bottled water industry found within the Liang area of the Belait District (FAO 2011), supplying bottled mineral water from its artesian well for almost three decades. However, industrialization and urbanization are putting more and more pressure on the country's surface and groundwater resources (Suhip et al. 2020).
The country's irrigation system was established in the 1980s and has solely relied on surface water sources. Recent findings show that this water supply is still insufficient to meet the country's irrigation needs, particularly during the dry season. Therefore, farmers must plan their plantings to correspond with the wet season (Wasil 2019). Local farmers also face other challenges, such as unpredictable weather and high soil acidity (Grealish and Fitzpatrick 2013;Wasil 2018). Thus, without proper soil and water management, acidic soils affected water may leach into the shallow aquifer, altering its quality (e.g., Kachi et al. 2016;Jia et al. 2020;Azffri et al. 2022).

Geophysical survey
The ERT and IP surveys are both based on the same acquisition setup and can be performed simultaneously. The surveys were conducted by injecting direct current into the ground through a set of current electrodes (Dahlin 2001). ERT measures the resulting voltage differences through a set of potential electrodes, whereas IP uses the voltage decay characteristics to study the induced polarization, also known as the chargeability (Telford et al. 1990;Kearey et al. 2002;Binley et al. 2015).
The geophysical survey parameters used in this study are given in Table 1. Two sets of data were acquired between the years 2018 and 2020. The first set was carried out in 2018 using the ABEM Terrameter LS2 resistivity meter covering a lateral distance of 800 m with 81 electrodes arranged at an equal length of 10 m (survey lines 1-11). The survey was conducted using the gradient array configuration (Dahlin and Zhou 2006;Aizebeokhai and Oyeyemi 2014).
The second set of data acquisitions was conducted from 2019 through 2020 using the ABEM SAS4000 resistivity meter covering a lateral distance of 400 m (survey lines  Data processing was conducted using the ZONDRES2D software (in DAT format). The Gauss-Newton least-squares inversion method was used to produce the 2D inversion models (Loke and Barker 1996;Loke and Dahlin 2002). The inversion process averaged the resistance measurements to apparent resistivity and the apparent chargeability time window measurements to integral chargeability. The inversion stopping criteria are 5% for Root Mean Square (RMS) errors and 5 for iterations.

Groundwater well drilling and construction
This study conducted the drilling and construction of two new groundwater wells in the study area, namely Well-B1 and Well-L1. Borehole drilling was carried out using a straight rotary drilling method. A 10-inch drill bit attached to a series of hollow drill rods was used to advance into the hole. Water-based drilling fluid flow to the bottom of the hole through the hollow drill rods and passages into the drill 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. Rock cuttings were collected at an interval of 3 m, primarily for lithology identification. Unplasticized polyvinyl chloride (uPVC) pipes of 6-inch diameter were used to construct the wells. Screens with 1.5 mm openings and gravel pack filter were installed in the aquifer layer.

Hydraulic tests
Hydraulic tests were conducted by installing a submersible pump (1.5 HP) inside the newly constructed groundwater wells. As there were no observation wells nearby the newly constructed pumping wells, the time-drawdown data were gathered from the pumping wells. A volume meter was connected to the outlet pipe on the surface to measure the volume of water discharged. The groundwater pumping rate was determined from the volume of water discharged against time. The change in water level (drawdown and recovery) was determined using a depth meter installed inside the well annulus. Using the Cooper-Jacob straight-line time-drawdown solution, the aquifer transmissivity (T) was calculated using Eq. 1 (Cooper and Jacob 1946): where Q is the pumping rate and ho-h is the calculated change in drawdown between initial and final drawdown.
Furthermore, the Cooper-Jacob solution assumes that the aquifer is confined, homogenous, isotropic and of uniform thickness over the pumping area. Aquifer potentiality was classified based on the calculated transmissivity values. Transmissivity values of lower than 50 m 2 /day are considered weak aquifer potential. In contrast, values between 50 to 500 m 2 /day are considered moderate aquifer potential, and values higher than 500 m 2 /day are considered high aquifer potential (De Wiest 1965).
In addition, the apparent horizontal hydraulic conductivity (K) of the aquifer was calculated using Eq. 2 (Mogaji et al. 2011): where T is the calculated aquifer transmissivity, and b is the thickness of the aquifer unit.

Interpretations of 2D resistivity and chargeability inversion models
ERT and IP studies were conducted for groundwater exploration at eight agricultural development areas in Brunei, particularly for paddy field irrigation purposes. The results revealed subsurface resistivity variations in the study area ranging from 1 to 500 Ω m and chargeability values ranging from 0 to 10 ms, with the depth of investigation of up to 150 m below ground level. Resistivity and chargeability inverted models of profile lines 12, 16 and 17 from sites 1, 5 and 8, respectively, are presented in this paper . The selected profiles provide comprehensive representations of the subsurface resistivity and chargeability variations in the study area.
At sites 1 and 8 (Figs. 3a and 5a), the areas are dominated by low resistivity values ranging from 1 to 50 Ω m. In contrast, at site-5, the areas are dominated by higher resistivity values ranging from 5 to 100 Ω m (Fig. 4a). Additionally, sites 1 and 8 were dominated by low chargeability values ranging from 0 to 1 ms ( Fig. 3b and 5b). In contrast, higher chargeability values ranging from 0 to10 m s were observed at site-5 (Fig. 4b).
All the profiles in the study area depicted two distinctive electrical layers. The first uppermost layer was inferred as topsoil, with resistivities ranging from 1 to 500 Ω m. The second layer was inferred as an aquifer zone, with resistivity values ranging from 1 to 100 Ω m. In addition, limited soil lithology information gathered from eight existing boreholes at sites 1-2 and sites 5-7 revealed that the topsoil layer consists predominantly of clay, peat, silty clay and silty sand (Fig. 6). No further information on the deeper geological strata was available in the investigated sites.
The resistivity and chargeability interpretations in the study area are given in Table 2. Resistivity values ranging from 1 to 100 Ω m were deduced as aquifer zone (e.g., Saad et al. 2012;Ashraf et al. 2018;Riwayat et al. 2018). According to Keller and Frischknecht (1996), fresh groundwater yields low resistivities ranging from 10 to 100 Ω m. Furthermore, resistivity values of lower than 10 Ω m are likely associated with higher ion concentrations or clay content (Goldman and Neubauer 1994;Keller and Frischnecht 1996;Samouëlian et al. 2005). Moreover, Saad et al. (2012) found that silts and clays will further reduce resistivity values much lower than groundwater effects.
Chargeability values in the study area range between 0 to 10 ms, with the values ranging from 0 to 1 inferred as a saturated zone. According to Telford et al. (1990), groundwater yields chargeability of 0 ms. Furthermore,  Fig. 6 Lithology information from eight existing boreholes at site-1 (BH 1-2), site 2 (BH 3-4), site-5 (BH 5-6), site-6 (BH-7) and site-7 (BH-8) in the study area chargeability values of specific rock materials are 1 to 4 ms for alluvium, 3 to 9 ms for gravel and 3 to 12 ms for clean sandstone. In contrast, higher chargeability values have been recorded for shales and siltstones, with values ranging from 50 to 500 ms (Telford et al. 1990). Therefore, based on the encountered resistivities and chargeabilities, site-5 was regarded as a more prolific drilling target compared to sites 1 and 8.

Borehole lithology correlation with resistivity datasets
New borehole drilling was conducted at sites 1 and 5 to further investigate the subsurface lithological formations (e.g., Ashraf et al. 2018;Aziman et al. 2018;Kumar et al. 2020b). Boreholes were drilled up to 96 m at site-1 (Well-B1) and 80 m at site-5 (Well-L1). The detailed borehole lithologies are shown in Fig. 7, and their locations on the ERT profile lines are shown in Figs. 3a and 4a.
Groundwater Well-B1 is mainly composed of clay and sand (Fig. 7a). However, the predominance of clay deposits over sands, with traces of peat, may be explained by the deposition of fine-grained materials in alluvial floodplains, specifically in the lacustrine environment. In contrast, Well-L1 is mainly composed of sand and clay, with sandstone and mudstone discovered at deeper depths (Fig. 7b). The upper part of the borehole mainly consists of fluvial deposits (clay, sand and gravel).
Our findings suggest that the higher clay content observed at Well-B1 correlates to the lower resistivity values ranging from 1 to 30 Ω m and lower chargeability values ranging from 0 to 1 ms. In contrast, the higher sand content observed at Well-L1 correlates to the higher resistivity values ranging from 5 to 200 Ω m and higher chargeability values ranging from 0 to 4 ms. Therefore, after drilling the wells at sites 1 and 5, it was decided not to drill another well at site-8 since the resistivities at site-8 were judged to be similar to site-1, where high clay content was discovered.

Hydraulic tests for aquifer characteristics
Hydraulic tests were conducted for the newly drilled groundwater wells to determine the hydraulic parameters and aquifer health (e.g., Shen et al. 2015;Wu et al. 2017). Well-B1 was screened in the saturated clayey sand interval at depths of 84 m to 95 m below ground level, and Well-L1 was screened in the sand and sandstone intervals at depths   Fig. 8. The summary of aquifer hydraulic parameters is given in Table 3.
A preliminary groundwater constant rate pumping test was carried out for four hours at Well-B1, with a steady pumping rate of 0.18 m 3 /h and a maximum drawdown of 3.63 m. The recovery test revealed a complete recovery  period of 2 h (Fig. 8a). However, the aquifer is judged as poor due to the very low pumping rate and slow recovery. Therefore, a full 24 h constant rate pumping test was not conducted. On the other hand, a full 24 h constant rate pumping test was carried out at Well-L1, with a steady pumping rate of 12 m 3 /h and a maximum drawdown of 1.52 m. The recovery test revealed a complete water level recovery 4 min after the stop of pumping (Fig. 8b). The calculated transmissivity values using the Cooper-Jacob time-drawdown solution for the aquifer units at Well-B1 and Well-L1 are 0.53 and 109.8 m 2 /day, respectively. The estimated hydraulic conductivity value of the clayey sand aquifer unit at Well-B1, with an estimated aquifer thickness of 11 m, is 0.05 m/day. The estimated hydraulic conductivity value of the sandy aquifer units at Well-L1, with an estimated aquifer thickness of 40 m, is 2.75 m/day.
Hydraulic test results of Well-B1 at site-1 suggest that a weak aquifer is expected, with an estimated transmissivity value of 0.53 m 2 /day. The low hydraulic conductivity value of 0.05 m/day is typically associated with fine-grained materials such as clay (Spitz and Moreno 1996). In contrast, the hydraulic test results of Well-L1 at site-5 suggest that a moderate aquifer is expected, with an estimated transmissivity value of 109.8 m 2 /day. The calculated hydraulic conductivity value of 2.75 m/day is typically associated with alluvium and fine sand deposits (Spitz and Moreno 1996).
Limited groundwater reserves may be drawn from both groundwater wells for irrigation purposes at the investigated sites 1 and 5. However, additional treatment for groundwater extracted from Well-B1is required to eliminate the clay particles. Future groundwater abstraction rates for irrigation purposes at the site may decline due to declining groundwater levels, fines migration or scaling (Sazali et al. , 2019Peña-Arancibia et al., 2020). Therefore, groundwater monitoring is recommended to ensure the sustainability of groundwater resources for irrigation purposes in the study area.

Conclusion
ERT and IP methods were effectively utilized in this study to detect saturated zones in eight agricultural development areas of Brunei Darussalam. Results revealed subsurface resistivity variations ranging from 1 to 500 Ω m and chargeability values ranging from 0 to 10 ms in the study area. Two electrical layers were deduced. The first uppermost layer was inferred as topsoil, and the second was concluded as an aquifer zone. Aquifer zones were detected in all the investigated sites with resistivity values ranging from 1 to 100 Ω m and chargeability values ranging from 0 to 10 ms. Based on resistivity and chargeability interpretations, two new groundwater wells were drilled in the study area to further investigate the aquifer characteristics. Groundwater Well-B2 were installed to 96 m depth below ground level at site-1, and Well-L1 were installed to 80 m depth below ground level at site-5. Our findings suggest that the higher clay content observed at Well-B1 correlates to the lower resistivity values ranging from 1 to 30 Ω m and lower chargeability values ranging from 0 to 1 ms. In contrast, the higher sand content observed at Well-L1 correlates to the higher resistivity values ranging from 5 to 200 Ω m and higher chargeability values ranging from 0 to 4 ms. Hydraulic test results of Well-B1 suggest a weak aquifer for withdrawal and distribution for irrigation purposes is expected at site-1, with an aquifer transmissivity value of 0.53 m 2 /day and a hydraulic conductivity value of 0.05 m/day. Hydraulic test results of Well-L1 suggest a moderate aquifer for withdrawal and water distribution for irrigation purposes is expected at site-5, with an aquifer transmissivity value of 109.8 m 2 /day and a hydraulic conductivity value of 2.75 m/day. Future studies are necessary to locate irrigation wells with significant groundwater potential. This study provides a base for future groundwater studies in Brunei and may be helpful for other locations with similar geological settings.