Iraqi Geological

Abstract


Introduction
On the mainland of Sumatra the Great Sumatran Fault (GSF) is a major right-lateral fault system accommodating strike-slip components at the oblique part of convergent zones, caused by the subtraction between the Indo-Australian and Eurasian plates (Sieh and Natawidjaja, 2000;Yanis et al., 2021). The geodynamic process constructs a significant fault along Sumatra island with an estimated length of ~1900 km fault and divided into 20 primary segments, each of which has been geometrically determined to measure 60-200 km long (Genrich et al., 2000;Natawidjaja and Triyoso, 2007). The frequency of ground earthquakes in this region is relatively high because of the tectonic activities from the GSF. Such earthquake events lead to significant damage. For example, the following three The earthquake time series graph shows a full earthquake distribution in 1960 -2020 with some major events followed by significant damage (Gunawan et al., 2018;Muzli et al., 2018).
The GGMPlus has become a fundamental data source for reducing detailed in situ gravity surveys without calculating and determining further reductions (Lewerissa et al., 2020). GGMPlus has been applied to study the fault structure in the Andaman Sea Subduction (Rao et al., 2011), and it has even been reported to be very effective in mapping the geological structure in the Tanimbar basin in East Indonesia . Relative to other satellite gravity data, such as Topex and International Gravimetric Bureau free air, GGMPlus data has good accuracy and correlation with gravity ground data, with the level of correlation being above 95%. In this research, we employed GGMPlus data to map the fault and geological structure in Weh Island. We highlighted the seismic activities felt by local people to support the plan for earthquake mitigation on Weh Island.

Background, Geological Setting, and Seismicity
The tectonic process controls the seismic activities in the northernmost part of Sumatra Island from the oblique subduction between an oceanic plate of Indo-Australia with a slip rate of 5 mm/year that leads under the Eurasian continental plate (Liu et al., 2018). Referring to several previous earthquakes were also followed by a tsunami, such as in Sumatra-Andaman with a magnitude of M 9.0 in 2004 (Lay et al., 2005), Nias with a magnitude of M 8.5 in 2005 (Meltzner et al., 2015), and also in 2012 there was an ocean earthquake with a magnitude of M8.2 and 8.1 in the Wharton Basin, Indian Ocean . Furthermore, several active segments of the GSF system have produced ground earthquakes with a slip rate of 10-20 mm/year (Salman et al., 2020). The tectonic activity produced several earthquakes in Aceh province, such as 1964 in Seulimeum M 6.4, 2013 Bener Meriah with magnitude M 6.3, and2016 in Pidie Jaya with magnitude M 6.5 caused significant building damage and significant fatalities (Muzli et al., 2018).
This process resulted in a series of arcs of nonvolcanic islands, Bukit Barisan mountains with volcanic lines, and the active GSF crossing Sumatra Island from Semangko to Banda Aceh and moving forward to the Andaman Sea (Burton and Hall, 2014;Marwan et al., 2021). On the other hand, the tectonic activity has also formed many volcanoes along the GSF fault in Sumatra (Yanis et al., 2022c), including Jaboi volcanoes on Weh Island (Yanis et al., 2022a), Peut Sagoe (Yanis et al., 2023(Yanis et al., , 2020b and Seulawah Agam Zaini et al., 2022Zaini et al., , 2021. The Weh Island was formed in the depression segments of the Sumatran fault line on the northwest edge of the Sumatra Sea, and this island is categorized as a young volcano located in the Sunda Orogeny area, forming a chain of volcanoes, including Leumoe Matee, Sumereuguh, and Kulam Mountains (Darasutisna and Hasan, 2005). The geological structure in Weh Island has a similar direction to the GSF in the NW-SE. This system then forms segments and volcanic activities, which result in the emergence of volcanoes (Kurnio et al., 2015). In the Weh Island area, volcanoes have appeared in different directions, i.e., Jaboi volcano controlled by several faults, Ceunohot and Leumo Mate (Yanis et al., 2022a). (a) Regional geological map of Weh Island provided by (Bennett et al., 1981). On the map, Weh Island is located at the tip of Sumatra, bordered by a marine area with a distance of ±15 km, so it is very closely related to the GSF to the Andaman Islands.
The volcanic area in Weh Island is a stratovolcano found in the southeast and dominated by volcanic rock, such as andesitic and basaltic rock, and coral reefs and alluvium ( Darasutisna and Hasan, 2005;Kurnio et al., 2016), as shown in Fig. 2 the lithology in Weh Island could be classified into four groups: Miocene-aged Tertiary sedimentary rocks, Quaternary-Tertiary-aged volcanic rocks from Weh Island in the form of lava and pyroclastic fluids, Quaternary-aged volcanic rocks obtained as products from the young volcano chain with NW-SE and north-south directions, and a group of reef limestone (Yanis et al., 2022a;Zarkasyi and Suhanto, 2013). When observed through geological history, Weh Island has experienced nine tectonic periods that have resulted in a local fault structure (Dirasutisna and Hasan, 2005;Weller et al., 2012). The local faults observed in Weh Island were formed because of the Sumatran fault line on the north edge of Sumatra Island (Radhakrishna et al., 2008;Yanis et al., 2020a). The fault zones that emerged on Weh Island are normal faults that shifted horizontally in the north-south and NW-SE directions ( Darasutisna and Hasan, 2005). These local faults are also controllers of surface geothermal manifestation, such as in the Jaboi region, Lhok Pria Laot, Iboih, and Keneukai. Most of the local fault zones in Weh Island are the continuity of the Lam Teuba-Krueng Raya fault located in the northern part with a normal fault structure that appeared in the north-south direction, where the graben field in the west part has a relatively elevated block in the east.

Global Gravity Data
The gravity method is a geophysical method for measuring the difference between Earth's gravity field and density change in the laterally of each material in data measurement (Saibi et al., 2019;Yanis et al., 2022b;Yanis and Marwan, 2019). In terms of physical principles, gravity methods work based on Newton's laws, which describe the force between two anomalies with a particular density separated by the distance of the two materials; the force value between mass particles will be directly proportional to the multiplication of both particles and inversely correlated with the quadratic distance between mass centers (Benyas et al., 2021;Surya et al., 2019). The current study used GGMPlus data developed based on the mathematical calculation of several data combinations, such as GOCE, EGM2008, and GRACE Satellite. (Hirt et al., 2014(Hirt et al., , 2013 stated that GGMPlus provides a complete description of Earth's gravity with a high resolution and wide range and could thus be utilized for geophysical studies. GGMPlus data provide information on gravity acceleration, radial and horizontal field components, and geoid height in a grid format with a space between points (~200 m). The gravity data were obtained from satellite observations in the form of free-air anomaly (ΔgFA). The height anomalies from the satellite observations are computed with eq. 1, while the value of free-air anomalies from GGMPlus is computed using eq. 2.
Where ζ0 and Δg0 are the zero-order terms for the height and free-air anomaly data, respectively; GM is a constant of universal gravity and Earth's mass; aref is the scale parameter for the global gravity model calculated with r as the geometrical distance from the Earth's center; P ̅ nm (θ) is a normalized Legendre function; n and m are the degrees and spherical harmonic order for the GGM satellite, respectively; C ̅ nm and S ̅ nm are the coefficients of the normalized spherical harmonic, and nmax is the value of the maximum degree.

Extracting GGMPlus Data
The GGMPlus data were generated through three processes: synthesis of gravity field to spherical harmonics, forward modeling, and calculation of normal surface gravity to obtain a gravitational model with a high resolution (Hirt et al., 2013;Rexer and Hirt, 2015). The data were obtained from http://murray-lab.caltech.edu/GGMplus/index.html and then extracted in *asci XYZ format using a source code in MATLAB provided by Curtin University. This extraction process generated several data: gravity accelerations, gravity disturbances, geoid undulations, and vertical components from sensor satellite altimetry ( Hirt et al., 2013;Abdullah et al., 2022). The Free-air correction is the first correction applied to the observational gravity data to remove the influence of height in all measured locations. The free-air anomaly is the value of gravity acceleration affected by topography without considering rock mass in the subsurface (Rexer and Hirt, 2015). The depiction of Weh Island from 12.500 datasets was obtained with a high grid resolution. From the extraction of GGMPlus data, further processing was performed for the data of gravity disturbance with topography parameters and a horizontal height of 100 m.

Filtering Technique
In regions with slow tectonic movement, the topography difference on the surface could not be observed well. Hence, Bouguer anomaly data from gravity observations do not show a clear contrast difference because of regional anomalies and residuals from the measured area (Benyas et al., 2021;Xu et al., 2009). The severance of regional and residual anomalies is essential in interpreting gravitational data. It facilitates the identification of local and minor signals commonly associated with shallow subsurface anomalies (Al-Rahim and Abdulrahim, 2021;Dewanto et al., 2022). Residual anomaly is the residue of Bouguer and regional anomaly (Abdulrahim et al., 2022). Mathematically, the separation of residual anomaly is represented in eq.3.
Where G(I, j) is the Gaussian kernel matrix, and c is a constant of the matrix. Meanwhile, i, u, j, and v are components of the matrix, and  is the constant value adjusted with the size of the Gaussian kernel matrix. Furthermore, the kernel matrix is convoluted against the matrix data present in the frequency area. The enhanced technique is also available to clarify surface anomalies in the potential method. Examples of the widely used techniques to identify the borders caused by the fault and contact between rocks in the subsurface are the second vertical derivative (SVD), first horizontal derivative (FHD), and tilt derivative (TDR). The vertical derivative analysis needs to be performed to ensure the presence of a fault structure. The method is usually applied to gravitational data to depict the geological features near the surface and enhance the high wavenumber of the spectral component. The zero value from the vertical derivative of the gravitational data is usually adjusted to the geological limit, as shown in eq. 4. SVD = ∂∆T/ ∂z (4) Where T is the observed gravity data, and ∂z denotes the gravitational data in the vertical direction z. FHD was applied to determine the contact limit of the horizontal density from the gravitational data. This method modifies gravity anomaly from one point to the other. It has sharp characteristics in the form of maximum and minimum values in the contacts of anomaly materials, thus showing the border of a geological structure as a fault (Ibraheem et al., 2019). The total horizontal derivative has been widely used to identify the density contrast of rock correlated with fault and geological structures (Wada et al., 2017). Meanwhile, TDR analysis is also effective in determining the boundary of the anomaly; This is because the tilt method uses two components, which is the ratio between the horizontal derivative and the vertical derivative, as shown in eq. 5.
Where f is the Bouguer anomaly from gravity data at coordinate (x,y). Meanwhile, ( ) is a vertical derivative from the gravity field.

Focal Mechanism
We analyzed five felt earthquakes on Sabang island to complete the gravity modeling. We used a seismic waveform from the seismic network operated by the BMKG Indonesian Agency. First-motion polarities of the seismic waveform are commonly used to determine the focal parameter of lowmagnitude earthquakes (Hardebeck and Shearer, 2002). This research manually repacked the P and S waves with their polarities from all seismic stations that recorded the earthquake. The researchers have completed their theory in determining the polarity of the first P-wave motion was measured manually, which shows the seismic stations in Aceh Province recorded five events of magnitude 3 -4. We analyzed the seismicity data using LocSat as a software part of the SeiscomP3; the software calculated the waveform using a linearized inversion scheme. The focal parameter results from seismic waveform analysis conduct a clear earthquake mechanism near Weh Island that is previously unstudied.

Bouguer Anomaly and Focal Mechanism
The values of the free-air anomaly, as shown in Fig. 3a varied between 27 and 78 mGal, with high values observed on the east and low ones on the central and west sides. The free air distribution suggested several faults contrasted in the NW-SE direction. Meanwhile, the anomaly only represented the topography difference, which could also be applied to fault mapping. In reality, the free-air anomaly cannot describe the subsurface geological condition and thus requires several standardized corrections to generate gravitational data correlated with subsurface rock. Simple Bouguer anomaly data can investigate the subsurface density based on the mass effect of rock in the measured locations. However, these gravity values are still considered accurate because of the field influence of the investigated area with many hills and valleys. Hence, terrain correction should be conducted using the Hummer chart equation (Nowell, 1999) to obtain a Bouguer anomaly that considers the surrounding area's rock mass effect, as shown in Fig. 3b. The values of the Bouguer anomaly of Weh Island varied between 32 and 78 mGal. The anomalies started being centralized in several locations and are not associated with the topography; for example, the east side is dominated by a high Bouguer value (52-78 mGal), possibly due to the response against the basaltic rock with high density (2.84 g/cm 3 ); this characteristic is in line with the regional geological map of Weh Island. As indicated by moderate Bouguer values (45-52 mGal), rhyolites and basalts dominate the southern side.
Meanwhile, on the west side, the values are low and range from 32 mGal to 45 mGal, representing the presence of andesites whose density is low (2.56 g/cm 3 ). Nonetheless, the Bouguer data obtained in this study did not show a detailed distribution of the surface fault, and the data only indicated the difference in the density contrast of the rock. This result was ascribed to the regional influence of satellite observation. Hence, we separated the regional and residual data from the Bouguer anomaly, as depicted in Fig. 3. c and d.  Fig. 3. Results of gravity data processing overlaid with the local fault structure from previously reported works ( Darasutisna and Hasan, 2005) and regional geology (Bennett et al., 1981). (a) Freeair anomaly data obtained from GGMPlus, (b) complete Bouguer anomaly based on the assumption that the rock density is 2.67 g/cm 3 and terrain corrected data, (c) residual anomaly, and (d) regional anomaly from Bouguer data.
In this study, we used Gaussian filtering based on the spectral analysis of amplitude changes in a spatial manner in the form of the cutoff wavenumber (Karunianto et al., 2017). The residual anomaly could reflect the geological structure of the investigated area with local characteristics and be affected by the anomaly sources near the surface, as shown in Fig. 3c. Following the filtering process, the value of gravity anomaly was found to vary between −8 and 10 mGal. The overlaying of the residual anomaly map with the fault location in the NW-SE direction in several locations showed a match, with several faults indicated by the intersection between the high and low anomalies.
Residual anomalies are considered more representative of the measured geological condition than regional anomalies representing inner structures because of their ability to provide the distribution data of other various anomalies (Fig. 3d). According to the regional geological structure, these residual values suggest several rock contrasts intertwined with fault distribution in Weh Island. For instance, gravity contrast was found in the direction of NW-SE on the east side as a response to the Sabang and Seuke faults. Similarly, the Leumomate and Ceunehot faults could be depicted very well by the residual data as the main controller of the Jaboi volcano. Therefore, these data are very efficient to be used as input parameters in the enhancement process to clarify the subsurface geological structure anomalies. We conducted the mechanism for some felt earthquakes near Weh Island to complete the gravity result, as shown in Fig. 4. The focal parameters figured a right-lateral mechanism in the direction of NE -SW following the lineation of the North Seulimeum fault. Five earthquakes figure an actual nodal plane in striking angle 130° -150° while the dips are mostly in NE dipping with a small portion of 80 -90° because of the dominantly horizontal mechanism. The focal results suggest the activities of the Seulimeum fault can control the northern part near Weh Island. Further mitigation can be conceptualized from the focal result because it is an active fault. Historically, some significant earthquakes have occurred in the southern part of the Seulimeum fault, while a long-term absence can be found in the northern part. Regarding (Ito et al., 2012) modeled the Seulimeum fault in the north part with its potential up to M 6.5 and slipped rate of 1 -2 cm/yr. Therefore, spatial and regional further planning with optimal mitigation concepts in the Weh Island can be studied with an update with geophysical modeling and assessing an active fault from this research.

4.2.Boundary and Edge Detection
The analysis and interpretation of gravitational data herein were based on the spatial variations characterized by the horizontal and vertical derivatives. Gradients or descendants of a gravity field could be computed through the spatial or wavenumber domains. Edge detection or digital filtering techniques have been used regularly in the graphical interpretation of gravity method in detecting geological structures, such as faults, rock formations, and basins. The results of the geophysical anomaly filter can provide direct information on the presence of subsurface structures. However, it remains to be determined whether the lineament corresponds to the geological anomaly; in this study we clarify the anomaly by overlaying fault information from previous studies. As a part of data enhancement, an analytical method for horizontal and vertical gradients was used to identify the border anomalies representing the subsurface geological structure. We applied a horizontal filter, vertical filter, or a ratio of both methods to reveal the anomaly contrast from the fault of Weh Island. Fig. 5 compares the vertical response and TDR from the residual anomaly. The data of the vertical anomaly varied between −7 and 5 mGal/m2, thus suggesting the presence of faults in several locations (indicated by the black line in the figure). The low derivative value supports this stipulation (<1 mGal/m2) and is surrounded by high values (3-5 mGal/m2). For example, on the east side, the derivative value is low and is correlated with the faults of Sabang. Moreover, the Labu Bau fault is present in the NW-SE direction in the central part. On the southern side is the active Jaboi volcano, suggesting that the derivative data could indicate the Ceunohot fault contrast in the NW-SE direction and Leumomate in the SW-NE direction. Similarly, the west side shows several low derivative values corresponding to the Kulam faults. In addition, this vertical derivative data can also provide a clear contrast to several other local faults on the west side such as the Balohan and Seuku faults, and this information is also supported by some seismic activity obtained from the area with a magnitude of M3.1-3.2, as well as on East area where there is the Pria Laot fault with relatively high seismic activity, with a magnitude of M4.3 -5.
In general, the vertical data of the derivative indicate the distribution of several local faults controlling the tectonic mechanisms from Weh Island. The tilt derivative data were also applied against residual gravity for complementary results, as shown in Fig. 6. Overall, the tilt derivative varied between −1.15 and 1.5 rad; the value is positive if the position is above the anomaly surface, zero if the position is on the edge of the geological structure, and the negative value is represented as outside the object. Hence, we only conducted data plotting above 0 rad. Regarding (Cooper and Cowan, 2006), this technique is beneficial owing to its simplicity in detecting the geological structure and host around an anomaly with the assumption that the fault is the border between one geological structure and another, the fault trace could be represented by the tilt value of rad. According to the tilt maps, the fault distribution was not clear enough. For instance, the data cannot contrast on the east side where the Sabang and Seuke faults are located. Similar results were obtained in Lhok Jeumpa, and Pria Laot faults in the NW-SE direction. However, in several locations of Weh Island, the vertical and tilt derivative data cannot show a fault structure that has been overlaid with previous research. For example on the Pria Laot, Lhok Jeumpa, and Sabang faults in different directions, it caused the sensitivity of global gravity, which cannot detect faults with slow tectonic rates, where the topographical differences above the surface in this case are relatively low. One of the advantages of using vertical and tilt derivative filters is that they can clarify the presence of faults previously not mapped by the Bouguer gravity anomaly. Nevertheless, several faults were observable above the surface because of topography changes; thus, the TDR data could exhibit anomaly contrast very well. For example, the Labu Bau fault on the central side, the Kulam fault on the west side, and Ceunehot and Leumomate are the volcanic mechanisms of the Jaboi volcano. Furthermore, the anomalies in the tilt derivative data are also very diverse due to the detailed data resolution of GGM+ (200m/px), which may represent near-surface anomalies. Still, these various anomalies are mostly focused on several faults in Weh Island. Generally, the anomaly of the vertical and tilt derivatives commonly shows the existence of faults. Still, the filter only uses the vertical component data without considering the mass effect from the horizontal direction. Thus, the horizontal derivative filter was calculated against the residual anomaly, as shown in Fig. 6. The first horizontal derivative was used to determine the contact border of the horizontal density from the gravitational data; this technique is based on the change of gravity value from one point to another, as characterized by the maximum or minimum value on the anomaly trace, including the faults Earthquake events, USGS andBMKG data from 1980 -2022 and subsurface geological structures (Wada et al., 2017). According to (Nasuti et al., 2012), the maximum value could drop if the fault structure has a slope position (not approaching a vertical position) or is close to other fault structures. The values of horizontal derivative varied between −5.3 and 5.4 mGal/m, and several exciting anomalies parallel to several Weh Island faults were noted in the SW-NE direction. The observation indicated that the anomalies shown by the horizontal derivatives correspond to geological contacts, such as faults. For instance, on the edge of the northern side of Weh Island, these horizontal derivative data can show the anomaly contrast in the NW-SE direction and parallel to the fault information from the regional geology.
On the east side, the filtering technique was found to show the Balogan fault despite the presence of the Sabang and Labu Bau faults. The result was ascribed to faults that are too localized and close to other faults in the same NW-SE direction. The existence of the Ceunohot fault and the Leumomate fault, which are also controllers for the Jaboi volcano can be demonstrated. However, because this fault is local, the direction of this fault cannot be described properly with regional gravity data. Similar results were noted on the west side, thereby supporting the claim that this filter could exhibit the different contrasts of Pria Laot and Kulam faults in different directions. This filter could clarify the local fault structures that become the main controller of the tectonic setting in Weh Island based on the transformation analysis. After applying the horizontal derivative filter to the global gravity data, the map visualized faults more evident than the previous filter. Still, some faults that are not readable with the previous filter cannot be mapped from the horizontal derivative. This is due to slow tectonic rates on several faults that make topographical differences relatively low. Besides, the gravity data used in the research is involved from regional satellites, so the presence of local faults such as on Weh Island is relatively difficult to detect such as on the Lhok Jeumpa, Sabang, and Leumomate Fault. However, the study of local faults with gravity satellites can provide an initial map of the existence of the geological structure affiliated with faults, and this method can be carried out free and quickly compared to conventional gravity methods, which generally spend a lot of time and finances on a large scale.

Conclusions
In the current research, we provided a detailed structure of the fault distribution in Weh Island by utilizing the latest gravity observation, GGMPlus, with a spatial resolution of ~200 m/px. The data obtained from this satellite were free-air anomalies varying from 27 mGal to 78 mGal and were only associated with the topographical changes from the investigated area. We calculated the Bouguer anomaly by assuming a density of 2.67 g/cm3 from the volcanic rocks and applied terrain correction from the Hummer chart, with the Bouguer data varying from 25 mGal to 78 mGal. A high Bouguer value dominated the east side due to the dense basaltic rock (2.84 g/cm3). On the west side, the obtained Bouguer values were low (32-45 mGal) and indicated the presence of andesites with low density (2.56 g/cm3). The transformation analysis of vertical and tilt derivatives also showed several previously unclear faults in the gravity residual data, including the Lhok Jeumpa and Balohan faults in the NW-SE direction. Meanwhile, the horizontal derivative filter showed differences in anomaly contacts suspected as faults, especially in fault distribution with different directions. The faults included the Ceunohot, Leumomate fault in the SW-NE direction and several other local faults that controlled the Jaboi geothermal system. According to the latest gravity observation analysis, the data could map the fault structure in Weh Island. These data could be applied as an initial method to analyze the presence of faults, especially in developing countries with high topography (such as Indonesia), contributing to the difficult instrument installation in gravity ground surveys. The focal parameter results in figure five earthquakes with an actual nodal plane in striking angle 130° -150°, while the dips are mostly in NE dipping with a small portion of 80 -90° because of the dominantly horizontal mechanism. The focal results suggest the activities of the Seulimeum fault can control the northern part near Weh Island.