Tectonic Model of Bali Island Inferred From GPS Data

DOI: 10.17014/ijog.5.1.81-91Seven periods of GPS campaign have been conducted for three years since March 2013 - October 2015 on fourteen GPS sites across Bali Island. The GAMIT/GLOBK 10.6 version was used to compute data with respect for thirteen reference sites of International Terrestrial Reference Frame (ITRF) 2008 surrounding Bali. The result shows that horizontal displacement varies between 1.93 and 22.53 mm/yr dominantly northeastward. Vertical displacement ranges at -184.34 to 33.79 mm/yr. The result of modeling using Coulomb 3.3 version indicates the deformation in Bali was mostly contributed by subduction at the southern part, West and East Flores Back-Arc Thrust at the north, Lombok Strait Fault and a fault at the eastern coast of Bali with the estimation maximum magnitude of 7.1, 6.6, 6.8, 5.8, and 5.2, respectively.


Introduction
Since 1818, Bali Island has experienced sixteen destructive earthquakes (Supartoyo et al., 2014). A strong earthquake occurred in 1818 with the intensity of VII MMI causing 1500 people lost their lives. Another destructive earthquake with magnitude of 6.1 struck Seririt in 1976 killing over than 559 people. The most recent destructive one occurred on October 13, 2011 creating some damages (Sulaeman, 2011).
Tectonically, Bali Island and Nusa Tenggara are part of Lesser Sunda Islands, which are defined as a group of small islands situated between east of Java and Banda Islands (Figure 1). The tectonic of this region is dominated by the col-lision between Australia and Eurasia Plates, as an earthquake source in the south of Bali. The Australia Plate subducting beneath Eurasia Plate was at a rate of 67 ± 7 mm/yr in N11°E ± 4°direction (Tregoning et al., 1994). Bock et al. (2003) revealed that Sunda Shelf block is estimated to be moving 6 ± 3 mm/yr SE relative to Eurasia. Another earthquake source which affected Bali originates from Flores Back-Arc Thrust extending from the north of Bali on the west and East Nusa Tenggara on the east side. Flores Back-Arc thrusting has been recognized by Hamilton (1979) to be behind Alor and Pantar Islands in the east and from central Flores to central Sumbawa in the west. Silver et al. (1983) suggested that the Flores thrust zone disappears beneath Bali Basin. In contrast, Koulali et al. (2016) reveals Flores Back-Arc continues westward until Kendeng Thrust. A seismotectonic study (Soehaimi, 2015) found a fault scarp at N 50° E direction at Lombok Strait, eastern Bali, parallel to the eastern coastal zone.
Earthquake source characteristics are important in order to define an earthquake hazard level in a certain area (Figure 1). Source characteristic information (location, geometry, slip rate, and maximum magnitude) can be obtained from geological and geodetic methods. In this study, a geodetic method using GPS measurement was used to identify the characteristic of the earthquake source. The survey had been conducted for three years since 2013 to 2015. This paper presents its result and discusses the displacement pattern as well as performs a model describing earthquake sources affect Bali and surrounding areas.

GPS Survey And Analysis
GPS survey is a method which is often used to analyze deformation associated with crustal fault activities by observing the displacement pattern. A GPS survey can be carried out either with episodic or continuous measurements. In contrast to a continuous observation approach, an episodic method, also known as a campaign survey, is conducted by careful observing changes with a specified interval, for instance, once or twice a year. A deformation rate requires a high precision level which is generally at the level of mm/year. GPS surveys in Bali were conducted seven times from 2013 to 2015: in March 2013, June Raw data in RINEX format were processed using GAMIT-GLOBK software 10.6 version (Herring et al., 2015)  For each day measurement, the dual-frequency carrier phase and pseudo-range observations from all sites were taken into account in processing the data. The International Terrestrial Reference Frame (ITRF) 2008 was used as a reference system. The ITRF 2008 is the realization of International Terrestrial Reference System (ITRS) where its centre is located on geocentric (the centre of mass of earth) with axes oriented consistently towards BIH (the Bureau International de l'Heure) at epoch 1984.0, and its length is defined in meters (Kuncoro, 2013).
To find out the displacement at each GPS site, the coordinate in the geocentric system was converted to topocentric with respect for an epoch on a specific reference. In this case, the first   (Kuncoro, 2013). Furthermore, in order to find the best fit model of the earthquake source to the available velocity data, the Coulomb 3.3 software (Toda et al., 2011) was performed.

Results and Analysis
Data obtained from the GPS processing are displacement of position which was visualized in a time series. It describes the movement of a GPS site within a specific time interval. From the time series of position changes, the velocity vector value of each GPS movement site can be estimated using the least square method. For example, displacement (north, east, and up) for SSUT site during the period of March 2013 -October 2015 is shown in Figure 3. The obtained displacement are -1.39 mm/yr, 28.25 mm/yr, and 1.41 mm/yr for component of north, east, and up, respectively.
The velocity for each GPS site is shown in Table 1 and Figure 4. Normalized root mean square (nrms) was estimated from differences between the observations and calculated phase and the value of the weighted root mean square residual (wrms) representing short-term correlation within data, as shown in Figure 4 and Table  1. The horizontal displacement of Bali tends to east-southeast ( Figure 4) before being corrected by Sunda Block, except NSDA site.
According to Table 1 and Figure 4 the velocity values vary 22.09 mm/yr to 33 mm/yr with the error bar ranging from 0.99 mm/yr to 20.08 mm/yr for north component, and 1.26 mm/yr to 23.72 mm/yr for the east one. The lowest velocity was recorded at BTUR (22.09 mm/yr) and the highest at KLKG (49.29 mm/yr). Meanwhile, the vertical velocity varies between -184.34 mm/yr (downward) to 33.79 mm/yr (upward). The lowest downward was recorded at KLKG (-184 mm/yr), and the highest upward at PMTR (39 mm/yr). The error bar for horizontal measurement relatively seems to be better compared to the vertical one. The large error bar recorded at KLKG for both vertical and horizontal could be due to vibration noise from the stream since the site is located just on the river bank.
The Sunda Block moves 27 mm/yr in SE direction as seen in Figure 5. Figure 6 and        Table 3). Those sources are considered to affect most deformation across Bali. The result of modeling using the best fit approach between the observed and calculated data can be seen in Table 4. The five source models produce GPS velocity at 0.81 mm/yr to 13.51 mm/ yr northeastward (Figure 8). This direction is in G -20.52 mm/yr and 6.33 mm/yr for east and north components, respectively. Other sites show 0.19 mm/yr to 9.65 mm/yr for the east component, and 0.52 mm/yr to -16.52 mm/yr for the north one. Figure 1 shows earthquake sources around Bali. In Bali and Nusa Tenggara, most of the deformation of the island arc can be attributed to collision with the Australian Continent. At the southern part, the earthquakes are dominated by reverse mechanism. This shows that the earthquakes are associated with the subduction zone, such as destructive earthquakes in Buleleng (1862) and the most recent in Nusa Dua (2011) (Figure 9a, b). Earthquakes at the northern part of Bali and Nusa Tenggara are also dominated by thrusting mechanism (Figure 1). According to McCaffrey and Nabelek (1987)       are associated with the back-arc-thrust zone that has been found deeper than 50 km (Figure 9b). Flores Back-Arc is believed to contribute to the 1976 Seririt Earthquake (Figure 9a, b) and the 1992 Flores earthquake generating tsunami. Meanwhile, twenty active faults and four potential active ones were found in Bali (Soehaemi, 2015), and the maximum magnitude of the faults are estimated to be around 6.5 Mw. The GPS velocities are obtained from observations at fourteen GPS station networks across Bali Island. Campaign GPS sites have been surveyed irregularly from 2013 to 2015. Using the ITRF 2008 reference system, globally the results show that the velocity of GPS sites tends to SE direction which is similar to a previous study by Bock et al. (2003). However, when using the movement of Sunda Block to correct the result, the direction change to NE as Koulali et al. (2016)'s. It seems that deformation in Bali is generated from the subduction activity in the south of Bali which can host a maximum of 7.0 Mw earthquake, followed by Flores Back-Arc Thrust at the north of Bali with 6.8 Mw. The local  (Soehaimi, 2015) can generate a 5.8 Mw earthquake. The subduction and Flores Back-Arc Thrust have the reverse fault mechanism, where the Bali Island acts as a hanging wall. The position of the island is closer to the nodal plane of the Flores-Back-Arc Thrust than to Indian-Australian Plate subduction that may explain the direction of Bali deformation to the northeast.

Conclusion
The horizontal velocity of Bali GPS sites varies 1.93 mm/yr to 22.53 mm/yr in NE direction, except NSDA, BTUR, and KASM to the north, and KLKG to the east. PMTR site has the smallest velocity at 1.93 mm/yr and the highest one (KLKG) at 22.53 mm/yr. While the vertical velocity ranges between -184.34 mm/yr to 33.79 mm/yr. The relative velocity may associate with the tectonic activity due to the subduction at southern Bali, West Flores Back-Arc Thrust, and East Flores Back-Arc Thrust at the northern part; and Lombok Strait Fault and a fault at eastern Bali with magnitude (Mw) 7.1, 6.6, 6.8, 5.8, and 5.2, respectively.