1D Audio Magnetotelluric Modelling for Deep Aquifer Identification in the Lava Fan Area of Haruman Peak, Malabar Mountains

The Malabar Mountains area acts as a catchment and infiltration zone for rainwater. Haruman Peak is the location of one of these areas. Information on the well-preserved depth of the shallow aquifer at 40 meters reinforces this. This research reviews the results of subsurface 1D resistivity structures from AMT data to obtain information on the depth of deep aquifers on the western slope of the Haruman Peak Lava Fan, Haruman Mountains. 1D modelling shows an aquifer at depth of 140.56-2080.07 meters with resistivity ranging from 5.25-68.08 Ωm. At depths of 453.32 m (HR02), 530.8 m (HR03), 1464.97 m (HR01), and 2080.07 m (HR02), interbedded tuff-andesite with minor pumice identifies the deep aquifer. In addition, a depth of 140.56 m (HR02) with a resistivity value of 68.08 Ωm indicates a shallow aquifer. Looking at the elevation of the Bandung Basin, water from aquifers located at elevations > 700 meters above sea level will flow into the Bandung Basin.


Introduction
Mount Malabar is a type B volcano located south of Bandung and forms the boundary between the Bandung plateau (700 m) to the north and the Pangalengan plateau (1400 m) to the south.The Malabar Mountain complex is a crucial catchment and rainwater infiltration zone.Haruman Peak is one of the main infiltration zones of the Bandung-Soreang Groundwater Basin located on the western side of the Malabar Mountain range (Harja et al., 2021).The mountain slopes that lead into the Bandung Basin support this claim, as stated by Suhari and Siebenhüner (1993).Furthermore, recent studies by Harja et al. (2021Harja et al. ( , 2023) ) have provided clear and precise information on the depth of shallow aquifers at 40 meters, which are still wellpreserved.
Information on the existence of deeper aquifers is not yet clearly known.Therefore, a deep investigation of geophysical measurement is needed, one of which uses the Audio Magnetotelluric (AMT) method that can detect resistivity variations from tens of meters to tens of kilometers (Gomo, 2023).This method utilizes lightning activity in the ionosphere as its source with a frequency range from 0.1 Hz-10 kHz (Constable, 2016), which is both environmentally friendly and easy to use.The AMT method is used in hydrology to determine aquifer thickness and depth, to define the deepest geological strata, and to characterize groundwater migration (Li et al, 2021;Montahaei, 2022;Tripathi et al., 2019;Xu et al., 2020;2023;Zaher et al., 2021).The expected result of using AMT method in this research is to determine the presence of deeper aquifers based on the contrast of rock resistivity values modelled in 1D.

Geology of the research area
The research was focused on the western slope of Haruman Peak, Malabar Mountains, West Java, with three points including HR01 (7°8'20.10"S and 107°34'48.36"E),HR02 (7°8'24.78"S and 107°34'51.48"E),and HR03 (7°8'28.98"S and 107°34'50.14"E)(Figure 1).The description of stratigraphy and lithology in the research area is based on the Geological Map of Garut and Pameungpeuk Quadrangle, Java (Alzwar et al., 1992).The geology of the measuring region is separated into volcanic rock units and surface sediment units, with Tertiary volcanic rocks overlain by Quaternary volcanic rocks that deposit surface sediment units.The rock units exposed around the AMT research area are Miocene to Holocene age rocks, including the Beser Formation (Tmb), with a thickness of about 500-1000 m consisting of tuffaceous breccia and lavas of andesitic to basaltic composition and containing pumice fragments; and the Waringin-Bedil Andesite, Old Malabar (Qwb), with a thickness of about 600 m consisting alternation of lavas, breccia, and tuffs.

Research Methods
AMT method measurements with the recorder main unit MTU-5A were carried out at three points on the western slope of the Malabar Mountains, precisely in Warjabakti Village, Cimaung District, Bandung Regency, West Java.The distance between HR01 and HR02 is about 175 meters, while the distance between HR02 and HR03 is about 135 meters.The data generated from the AMT acquisition is in the form of magnetic and electric field intensity variation data in the time domain (Hidayat et al., 2021).Each point is measured using vector measurement mode for 30 minutes.The length of the AMT measurement makes the data obtained in the frequency range of 0.35 Hz to 10400 Hz (Wu et al., 2023).The data is then processed to obtain an apparent resistivity curve, which is calculated based on the ratio 1D Audio Magnetotelluric Modelling for Deep Aquifer … of the size of the electric field (E) and the magnetic field (H) (Figure 2).The comparison is called the Cagniard equation (Equation 1) (Cagniard, 1953).
The steps of the procedure are Fast Fourier Transform, Robust Processing, and crosspower selection (XPR) (Kurniawan et al., 2019).Fast Fourier Transform is a method used to transform data from the time domain to the frequency domain.The next step is robust processing, which tries to detect, eliminate, and decrease outliers.This step performs iterative weighting of residuals to lower the weight of outliers.
The robust process considers the coherence of each robust parameter's results.
Coherence is a real number with a dimension ranging from 0 to 1, where 1 indicates a perfectly coherent transmission.
The robust process has three types of crosspower parameters, such as No Weightdoing the same weighting at each frequency, Rho Variance-doing high weighting on the type resistance that has as many snippets as the desired cross-power value, and Ordinary Coherence-doing high weighting at frequencies with good coherence between the H and E components (Junursyah et al., 2020).The Ordinary Coherence parameter has the highest coherence value among the three cross-power parameters, with coherence greater than 70%.The apparent resistivity and phase curves generated in the robust process usually have random patterns due to noise in the recorded data, therefore XPR is carried out to smoothen the curve.The XPR is carried out by selecting high-weight XPR data and eliminating low-weight XPR data (Figures 2, 3, and 4).
Curve trend analysis is the next step that removes data far from the curve trend (outliers).The deviant curve data is affected by noise in the initial data, robust results, and XPR.The maximum depth is then calculated based on the skin depth using Equation 2.
After the curve trend analysis, the data is modeled in 1D with Occam inversion to calculate the maximum depth based on resistivity variation.The Occam inversion serves to minimize the roughness level of the resulting model.The results from the inversion calculation of the model will be smoother, establishing the fitting process of field data (Constable et al., 1987).
Furthermore, the data is analyzed to determine the dimensionality of the subsurface resistivity structure, known as skewness analysis.This analysis to determine if there are deviations or distortions in the impedance tensor in magnetotelluric data (Junursyah, 2022).
There are two techniques for skewness analysis: Swift skew and Bahr skew.Swift skew is the ratio between the diagonal components (Zxx and Zyy) and the nondiagonal components (Zxy and Zyx) of the impedance tensor (Equation 3).If the skewSwift value > 0.3 implies 3D or noise.If the skewSwift value < 0.3, it suggests 1D or 2D (Pranata et al., 2017).Swift skew is a one-dimensional data analysis for determining whether the underlying electrical structure is distorted or not.The skewness value will look deviated if the impedance tensor has telluric distortion (Xiao et al., 2011;Tripathi et al., 2019).
Bahr skew is a calculation developed on the value of phase sensitivity skewness (η) or regional skew by considering telluric distortion in the impedance tensor (Bahr, 1991).The phase sensitivity slope can be determined using Equation (4).A value of η < 0.3 implies the response of 2D structure (Xiao et al., 2011).A value of less than 0.3 indicates a regionally 2D structure, but it is not sufficient (Tripathi et al., 2019).In this research, the technique used in dimensionality analysis is Bahr skew to clarify the subsurface structure.
Research flow chart can be seen in Figure 6.

Results and Discussion
Data was processed in 1D inversion modelling (Figure 7) using invariant mode (combined mode TE and TM).The analysis and interpretation of 1D modelling results will be correlated with stratigraphic models (Figure 8) regarding the relationship between rock resistivity values in general (Figure 9) and formation unit information on the Geological Map of Garut and Pameungpeuk Quadrangle, Java.According to Figure 8, the increase or decrease in resistivity values as depth increases shows the presence of different lithological formations or arrangements.filled with fluid, whereas layers 1-4 and 6-8 have impermeable qualities, as shown by a higher resistivity value than layer 5.The higher the resistivity value, the more massive or dense the rock.Hence, layer 5, at the HR01 has the potential to be an aquifer comprised of tuff-andesite and a minor of pumice.When viewed from the lithological properties and types of watercarrying layers (Kodoatie, 2021), the aquifer at the HR01 is identified as confined aquifer with impermeable upper and lower layers.The results of modelling the 1D resistivity structure shown in Figure 7c., obtained a depth of up to 1488.83 meters with a variation in resistivity values between 5.25 to 157470 Ωm.Based on the correlation of resistivity and depth values in Figure 8, layers 1-4 are predicted as Malabar Tilu Volcano Rock Unit (Qmt) based on the correlation of resistivity and depth data in Figure 8.These comprised andesitic lithic tuffs with a minor amount of pumice, breccia, and lavas.In addition, layers 5-8 are predicted as Waringin-Bedil Andesite Units, an Old Malabar (Qwb) formation comprised of lavas, breccia, and tuff rocks.
Layer 4 has a low resistivity (5.25 Ωm) compared to other layers, indicating it is an aquifer layer comprised of tuff-andesite rock with a minor pumice content.When viewed from the lithological properties and types of water-carrying layers (Kodoatie, 2021), the aquifer at the HR03 is a type of confined aquifer.Another step is dimensionality analysis, which aims to clarify the subsurface resistivity structure under research.Based on the results of the dimensionality analysis of AMT data, the Bahr Skew has value variation of 0.01 -0.13 (Figure 10).The skewness value of the aquifer layer at HR01 ranges from 0.01 -0.02 (light blue), the aquifer layer at HR02 has two skewness values ranging from 0.01 -0.02 (light blue) and 0.02 -0.03 (light green), and the skewness value of the aquifer layer at HR03 ranges from 0 to 0.01 (dark blue).The effect of 3D dimensionality on data is defined by a skewness value (η) > 0.3, while η < 0.3 implies the 2D dimensionality effect.The ideal condition of 1D-2D electrical structure is indicated by η = 0 (Xiao, 2011).Hence, the aquifer layer in the 1D model interpretation indicates that it is 1D Audio Magnetotelluric Modelling for Deep Aquifer … unaffected by the dimensionality effect.In the skewness cross-section, there is a slight dimensionality effect in the subsurface of measuring points HR01 and HR03.Geological features are thought to be the cause of this; however, it does not affect the surrounding aquifers.

Conclusion
Deep aquifer (confined aquifers) with resistivity values of 5.25 -40.65 Ωm were identified at depths of 453.32 m (HR02), 530.8 m (HR03), 1464.97 m (HR01), and 2080.07 m (HR02) with lithology in the form of andesite-tuff with minor of pumice.In addition, an unconfined aquifer was also identified at the HR02 point at a depth of 140.56 m with a resistivity value of 68.08 Ωm.Bandung Basin is located between 650 and 700 meters above sea level (m.a.s.l.), and water from aquifers that have an altitude >700 meters above sea level is expected to flow into the Bandung Basin.Water from aquifers located at depths of 140.56 m (HR02), 453.32 m (HR02), and 530.8 m (HR03) has the potential to flow into the Bandung Basin.

Figure 2 .
Figure 2. Comparison of the apparent resistivity curve and phase at HR01 before (left) and after XPR (right).

Figure 3 .
Figure 3.Comparison of the apparent resistivity curve and phase at HR02 before (left) and after XPR (right).

Figure 4 .Figure 5 .
Figure 4. Comparison of the apparent resistivity curve and phase at HR03 before (left) and after XPR (right).

Figure 8 .
Figure 8. Stratigraphic Model of Haruman Peak, Malabar Mountains Area and Its Surroundings Based on the Range of Rock Resistivity Values in Figure 9 and Formation Information on the Geological Map of Garut and Pameungpeuk Quadrangle, Java.

Figure 9 .
Figure 9. Range of Rock Resistivity Values (Alviyanda et al., 2014; 2020).Based on Figure 7b., 1D inversion modelling obtained depths of up to 2080.07 meters with variations in resistivity values of 34.2 to 259.31 Ωm.The resistivity value in layers 1 and 2 increased from 114.82 Ωm to 259.31 Ωm at a depth of 80.32 meters.When viewed from the correlation of resistivity values and depth in Figure 8, layers 1 and 2 can be predicted as soil layers formed due to the physical-chemical weathering process of rocks, biological activities of organisms, and a long-term climate (Kalev & Toor, 2018).Different things happen in layers 3 to 8, where resistivity values fluctuate as depth increases.The resistivity value in layer 3 reaches 68.08 Ωm, increases significantly to 195.06 Ωm in layer 4, and decreases sharply to 34.2 Ωm in layer 5. Furthermore, resistivity rises again in layer 6 with a value of 59.94 Ωm but falls back in layer 7 to 40.65 Ωm before finally reaching 65.44 Ωm in layer 8.The resistivity value of rocks that fluctuate with increasing depth illustrates According toOktariadi et al. (2021), the Bandung Basin is located between 650 and 700 meters above sea level.Thus, water from aquifers at 140.56 m (HR02), 453.32 m (HR02), and 530.8 m (HR03) are predicted to flow into the Bandung Basin.