Determination of ground displacement of 25 April 2015 Nepal earthquake by GNSS precise point positioning

Vietnam Journal of Earth Sciences, 40(1), 17-25, Doi: 10.15625/0866-7187/40/1/10876 17 (VAST) Vietnam Academy of Science and Technology Vietnam Journal of Earth Sciences Determination of ground displacement of 25 April 2015 Nepal earthquake by GNSS precise point positioning Nguyen Ngoc Lau Ho Chi Minh City University of Technology, Vietnam Received 5 May 2017; Received in revised form 26 October 2017; Accepted 10 November 2017 ABSTRACT The April 2015 Nepal earthquake (known as the Gorkha earthquake) occurred at 06:11:25 (UTC) on the 25th of April, with a magnitude of 7.8Mw. It was the worst natural disaster to strike Nepal since the 1934 Nepal-Bihar earth- quake. Precise determination of ground displacement in this area will provide important information to better under- stand the structure and scope of the earthquake, contributing to faster and more accurate earthquake prediction. In this paper, we use precise point positioning to determine the displacements of 17 GNSS stations around the epicenter for the day of the earthquake. The processing results show that the common displacement direction is close to south- southwest with the largest value being approximately 2 m and the affected area being about 160 km in the southeast direction centered around the earthquake epicenter. However, a detectable GNSS signal was still observed at a station some 647 km away from the epicenter. Keywords: April 2015 Nepal earthquake; GNSS; PPP. ©2018 Vietnam Academy of Science and Technology 1. Introduction1 The Gorkha earthquake killed more than 9,000 people and injured more than 23,000. It occurred at 06:11:25 (UTC) on 25 April 2015, with a magnitude of 7.8Mw or 8.1Ms and a maximum Mercalli Intensity of IX. Its epicen- ter locates at the east of the district of Lam- jung, latitude 28.231°N, longitude 84.731°E and at a depth of approximately 8.2 km. It was the worst natural disaster to strike Nepal since the 1934 Nepal-Bihar earthquake. According to the United States Geological Survey (USGS) (USGS, 2015), the Gorkha earth- quake occurred as the result of thrust faulting on or near the main frontal thrust between the *Corresponding author, Email: nnlau@hcmut.edu.vn subducting Indian plate and the overriding Eurasian plate to the north. At the location of this earthquake, approximately 80 km to the northwest of the Nepalese capital of Kath- mandu, the Indian plate is converging to the Eurasian plate at a rate of 45 mm/year towards the north-northeast, driving the uplift of the Himalayan mountain range. Geophysicists and other experts had warned for decades that Nepal was vulnerable to a deadly earthquake, particularly because of its geology, urbanization, and architecture. For this reason, some scientific organizations had set up instruments and facilities to moni- tor earthquake activity over this region. Uni- versity Navstar Consortium (UNAVCO), a non-profit university-governed consortium, facilitates geoscience research and education Nguyen Ngoc Lau/Vietnam Journal of Earth Sciences 40 (2018) 18 using geodesy, is currently supporting retriev- al of high-rate and standard Global Navigation Satellite System (GNSS) data from stations within Nepal (UNAVCO, 2015). These data can be accessed through the UNAVCO Data Archive as they become available (Figure 1). Figure 1. Epicenter (star) and GNSS station (triangle) displacement vectors With UNAVCO support, we collected some GNSS data from 17 GNSS stations around the epicenter on the day of the earth- quake. These stations are listed in Table 1 and Figure 1. This permanent GNSS station net- work is a favorable condition for applying ex- isting GNSS positioning methods to accurate- ly determine the displacement of the station over time. Data on recent earthquakes have also been collected and processed using GNSS (Ji C. et al., 2004; Yue H. et al., 2013). In Vietnam there is also similar research on the Tohoku earthquake in Japan on March 11, 2011 (Ngu- yen Ngoc Lau, 2012). However, only at the Gorkha event, can scientists first observe earthquakes occurring in an area with many high-rate GNSS stations near the epicenter and covering the affected area completely (Galetzka J. et al., 2015). The displacement of a set of stations over the earthquake area is an important source of information that provides quantitative data for a better understanding of tectonic activity in the area. This can help to make earthquake prediction faster, more accurate, and prevent similar disasters. To accurately determine the displacement of each GNSS measurement station, the coor- dinates of the stations over time are deter- mined by the GNSS processing in relative or absolute terms. If an earthquake occurs and moves the station, we can calculate the dis- placement by comparing its coordinates be- fore and after the earthquake. The GNSS relative (or differential) method was mainly used in the 1980s and early 1990s, when GNSS absolute method had not yet achieved the desired accuracy. The disad-vantage of this method is that it is difficult to provide high positioning accuracy when han- dling long baselines. We have applied the rel- ative method to calculate the ground dis-placement caused by the Tohoku earthquake in Japan on 11-03-2011 (Nguyen Ngoc Lau, 2012). In order to handle the long baselines of up to 1000 km, we had to apply special tech- niques to simultaneously process GPS and GLONASS measurements to get the desired accuracy (Nguyen Ngoc Lau et al., 2011). GNSS Precise Point Positioning (also known as PPP) is being used today to gradual- Vietnam Journal of Earth Sciences, 40(1), 17-25 19 ly replace the relative method. The reason is that its positioning accuracy is increasingly improved and its advantages compared to relative method. PPP is also the method we choose to use for this paper. It will therefore be introduced in more detail in Section 2. 2. GNSS precise point positioning method GNSS PPP is a positioning method that processes phase and code measurements from a single GNSS receiver together with precise GNSS orbit and clock correction products. PPP can provide a common position accuracy of centimeter level in 24h static and decimeter level in kinematic modes (Zumberge J.F. et al., 1997; King M. et al., 2002). The most prestigious organization providing precise GNSS orbit and clock products for civilian users is the International GNSS Service (IGS) (Kouba J., 2009). PPP has an advantage over traditional dif- ferential techniques in that the method re- moves the need for the user to establish a lo- cal base station. Therefore, the spatial operat- ing range limit of differential techniques is negated, as well as the need for simultaneous observations at both rover and base for real- time applications (King M. et al., 2002). In recent years, the accuracy of PPP has improved gradually because the quality of GNSS orbit and clock correction products have been enhanced, and the number of GNSS has increased rapidly. PPP with multi-GNSS potentially can provide an accuracy of better than 1 cm in 24h static and better than 1 decimeter in kinematic modes (Rabbou M.A. et al., 2015; Afifi A. et al., 2016). With such an accuracy, PPP can be used to detect any displacements larger than several decimeters in station coordinates. The current direction of PPP development is real-time positioning and improvement of positioning accuracy. The direction to im- prove accuracy for PPP focuses on resolving ambiguity for carrier phase measurements (Geng J. et al., 2012) and processing mixed measurements from multi-GNSS such as GPS, GLONASS, GALILEO, BEIDOU (Rabbou MA et al., 2015; Afifi A. et al., 2016). In Vietnam, we have researched PPP since 2010 with GPS only (Nguyen Ngoc Lau, 2009; 2010), and then expanded to GPS and GLONASS (Nguyen Ngoc Lau et al., 2012, Nguyen Ngoc Lau, 2013). PPP method has been described in detail in many documents (Nguyen Ngoc Lau, 2009, Nguyen Ngoc Lau et al., 2010; 2012, Nguyen Ngoc Lau, 2013). So in this article, we only mention our self-developed PPP software package, the so-called PPPC. This is the prod- uct of two ministry-level projects chaired by us (Nguyen Ngoc Lau et al., 2010; 2012). PPPC version 3.2 can process code and phase meas- urements from GPS, GLONASS, GALILEO and BEIDOU satellite systems for both static and kinematic modes. Using PPPC to process GPS + GLONASS data at some IGS stations has proven that positioning accuracy is better than 2 cm for 1h static data and better than 1 cm with 24 hours (Nguyen Ngoc Lau, 2013). With particularly advantage, PPPC is able to estimate coordinates before and after an in- dicated epoch. This option is very suitable for precise calculation of station coordinate slips if they have occurred. We use PPPC to pro- cess GNSS data in Table 1. Table 1. GNSS stations are located around the Earth- quake epicenter No. GNSS Station Interval (sec) GNNS satellite systems Distance to the epicenter (km) 1 CHLM 15 GPS 56 2 KKN4 15 GPS 68 3 NAST 15 GPS 81 4 DNSG 15 GPS 94 5 JMSM 15 GPS 119 6 SNDL 15 GPS 137 7 PYUT 15 GPS 169 8 SYBC 15 GPS 202 9 SMKT 15 GPS 348 10 RMTE 15 GPS 227 11 NPGJ 15 GPS 305 12 TPLJ 15 GPS 310 13 BRN2 15 GPS 311 14 LCK3 30 GPS, GLONASS 394 15 LCK4 30 GPS + GLONASS 394 16 DNGD 15 GPS 411 17 LHAZ 30 GPS + GLONASS 647 Nguyen Ngoc Lau/Vietnam Journal of Earth Sciences 40 (2018) 20 3. Results Firstly, we use PPPC to process all of the GNSS stations in the kinematic mode with some options as follows: - Using IGS precise orbit and clock correc- tions; - Using P3 code and L3 carrier phase measurements of GPS and GLONASS (only for LCK3, LCK4 and LHAZ); - Setting the elevation cut off angle as 5; - Estimating one tropospheric zenith delay every 2.5 hours with the Niell mapping func- tion; - Applying IGS08 antenna model and solid Earth model. After screening the processed station coor- dinates epoch by epoch, we detect 4 stations which have large slip values at epoch 6:12:15 (GPST) as shown in Figure 2. Therefore at the time of the earthquake (6:11:25 UTC~ 6:11:41 GPST), the stations were not affected immediately. The shift starts only about 26 seconds later. Figure 2. Station displacements by using PPPC in kinematic mode. Vertical bar indicates the earthquake epoch Vietnam Journal of Earth Sciences, 40(1), 17-25 21 Since GNSS epoch solutions have an accu- racy at the decimeter level, it is not sufficient for precise calculation of station displace- ments. We re-processed the GNSS data by us- ing PPPC in static mode for epochs before and after epoch 6:12:15. As a result, the displace- ment of each station is calculated by subtract- ing two processed station coordinates before and after epoch 6:12:15. These results are giv- en in Table 2. Table 2. Displacements of GNSS stations at epoch 6:12:15 No. GNSS Station Displacements (m) North East Up 1 CHLM -1.380  0.001 -0.220  0.005 -0.590  0.007 2 KKN4 -1.827  0.002 -0.455  0.005 +1.279  0.009 3 NAST -1.293  0.002 -0.318  0.006 +0.623  0.009 4 DNSG +0.004  0.003 +0.006  0.008 +0.003  0.015 5 JMSM -0.006  0.002 +0.003  0.005 +0.004  0.008 6 SNDL -0.220  0.001 +0.045  0.004 +0.064  0.006 7 PYUT +0.001  0.001 -0.005  0.004 +0.012  0.006 8 SYBC -0.015  0.002 -0.010  0.005 +0.003  0.009 9 SMKT +0.001  0.002 -0.002  0.004 -0.002  0.006 10 RMTE -0.000 ±0.001 -0.005±0.004 +0.004 ±0.005 11 NPGJ +0.002 ±0.002 -0.012 ±0.004 +0.008 ±0.006 12 TPLJ +0.002 ±0.001 -0.001 ±0.004 +0.002 ±0.006 13 BRN2 +0.001 ±0.001 -0.003 ±0.005 +0.000 ±0.006 14 LCK3 +0.002±0.001 -0.011 ±0.002 +0.012 ±0.003 15 LCK4 +0.003 ±0.001 -0.011 ±0.002 +0.010 ±0.003 16 DNGD +0.002 0.001 -0.002 0.002 -0.018 0.002 17 LHAZ +0.004 0.001 -0.007 0.002 +0.007 0.003 Table 2 shows that there are only 4 GNSS stations affected by the earthquake including CHLM, KKN4, NAST and SNDL. Where KKN4 was shifted nearly 2 m in the horizon- tal component, SNDL is some 177 km from the epicenter but also moved horizontally more than 0.2 m. Some closer stations, such as DNSG and JMSM distributed in the north- west, are seemingly not affected. This shows that the affected area is stretched in the south- east direction. Figure 3 shows the north, east and up se- ries of 4 GNSS stations distributed east- southeast of the epicenter, including SYBC, RMTE, TPLJ and BRN2. The timing of the earthquake-induced movement is well docu- mented on the charts of the stations. It is not fixed but varies with the distance to the epi- center. The time of movement of the stations SYBC and RMTE about 200 km from the epi- center is 6:13:00 GPST. Stations TPLJ and BRN2, about 300 km from the epicenter, are 6:13:15 GPST. Figure 4 presents the processing results of the farthest GNSS station - the LHAZ (647 km). Watching the sequence of this sta- tion coordinates over time, we can still ob- serve the effects of the earthquake occurring at 6:15:00 GPST, which is about 3 minutes slower than the stations in Figure 2 and almost 2 minutes compared to the stations in Figure 3. We present displacement vectors of the af- fected stations on Figure 1 and see clearly that the common moving direction of GNSS sta- tions is close to the south-southwest. Nguyen Ngoc Lau/Vietnam Journal of Earth Sciences 40 (2018) 22 Figure 3. The coordinate series over time of the 4 GNSS stations are distributed east-southeast of the epicenter 06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00-0.1 -0.05 0 0.05 0.1 Station SYBC No rth (m ) 06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00-0.1 -0.05 0 0.05 0.1 Ea st (m ) 06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00-0.15 -0.1 -0.05 0 0.05 0.1 Up (m ) GPS Time (HH:MM:SS) 06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00-0.3 -0.2 -0.1 0 0.1 Station RMTE No rth (m ) 06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:000 0.05 0.1 0.15 0.2 Ea st (m ) 06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00-0.05 0 0.05 0.1 0.15 Up (m ) GPS Time (HH:MM:SS) 06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00-0.05 0 0.05 0.1 0.15 0.2 Station TPLJ No rth (m ) 06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00-0.04 -0.02 0 0.02 0.04 0.06 Ea st (m ) 06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00-0.1 -0.05 0 0.05 0.1 Up (m ) GPS Time (HH:MM:SS) 06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00-0.05 0 0.05 0.1 0.15 Station BRN2 No rth (m ) 06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00-0.4 -0.2 0 0.2 0.4 Ea st (m ) 06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00-0.1 -0.05 0 0.05 0.1 Up (m ) GPS Time (HH:MM:SS) Vietnam Journal of Earth Sciences, 40(1), 17-25 23 Figure 4. The time series of the LHAZ station. Vertical bar indicates the earthquake epoch 4. Discussions Jianghui Geng in (Geng J., 2015) used GAMIT software to process relatively the GNSS data. His processing results of stations KKN4 and NAST are given in Figure 5. The visual estimates of displacement are -1.8, - 0.45, +1.3 in the north, east and up compo- nents for KKN4 and -1.3, -0.3, +0.6 for NAST. These results agree with our results in Table 2. In (Lemmens M., 2015), Lemmens ana- lyzed the 5Hz GPS data processing results of Galetzka et al., 2015) at two stations KKN4 and NAST. He concluded that the north and eastward movements of the two stations were distinctly different behaviors (Figure 6) be- cause the KKN4 station was located on hard rock, while NAST installed on sediment in the valley Kathmandu. NAST shows prolonged sediment resonance with a sweeping path of almost 2 m. By applying a ScanSAR-based interferom- etry analysis of Advanced Land Observing Satellite 2 (ALOS-2) L-band data, Kobayashi et al. (Kobayashi T. et al., 2015) had similar conclusions that “a major displacement area extends with a length of about 160 km in the east-west direction, and the most concentrated crustal deformation with ground displacement exceeding 1 m is located 20-30 km east of Kathmandu”. However, this technique does not provide precise coordinate displacement values, unlike the GNSS PPP technique. Figure 5. Jianghui Geng‘s processing results of stations KKN4 (left) and NAST (right), accepted from (Geng J., 2015) Nguyen Ngoc Lau/Vietnam Journal of Earth Sciences 40 (2018) 24 Figure 6. Displacements of stations KKN4 (left) and NAST (right), accepted from (Galetzka J., 2015) 5. Conclusions To monitor the effect of the earthquake with the magnitude of 7.8 on 25 April 2015 in Nepal, we collected and processed GNSS data at 17 stations around the epicenter by using the GNSS PPPC method. The GNSS station displacements are calculated precisely with an accuracy of 1cm in the horizontal and the ver- tical components. These displacements show that the affected area stretches about 160 km in the south-east. The common moving direc- tion is close to the south-southeast with the maximum value of 2 m in the horizontal com- ponent. Our results are similar to other studies us- ing different data sources or different pro- cessing methods such as GNSS relative (Geng J., 2015), high rate GNSS PPP (Galetzka J. et al., 2015; Lemmens M., 2015) and ScanSAR (Kobayashi T. et al., 2015). In conclusion, the GNSS PPP method has proven its advantages for monitoring ground movements due to earthquakes such as posi- tion accuracy, large area coverage, availabil- ity, short-term or long-term displacement tracking. In order to have the better determination of ground displacements, our future research di- rection will continue to focus on improving the positioning accuracy of GNSS PPP on the basis of ambiguity resolution of carrier phase measurements, and apply to our PPPC soft- ware. References Akram Afifi, Ahmed El-Rabbany, 2016. Improved Be- tween-Satellite Single-Difference Precise Point Posi- tioning Model Using Triple GNSS Constellations: GPS, Galileo, and Beidou, Positioning, 7, 63-74. Doi.org/10.4236/pos.2016.72006. Galetzka J., Melgar D., Genrich J.F., Geng J., Owen S., Lindsey E.O., Xu X., Bock Y., Avouac J.-P., Adhi- kari L.B., Upreti B.N., Pratt-Sitaula B., Bhattarai T.N., Sitaula B.P., Moore A., Hudnut K.W., Szeliga W., Normandeau J., Fend M., Flouzat M., Bollinger L., Shrestha P., Koirala B., Gautam U., Bhatterai M., Gupta R., Kandel T., Timsina C., Sapkota S.N., Ra- jaure S., Maharjan N., 2015. Slip pulse and reso- nance of the Kathmandu basin during the 2015 Gorkha earthquake, Nepal, Science, 349, 1091-1095. Han Yue, Thorne Lay, Susan Y. Schwartz, Luis Rivera, Marino Protti Timothy H. Dixon, Susan Owen, and Vietnam Journal of Earth Sciences, 40(1), 17-25 25 Andrew V. Newman, 2013, The 5 September 2012 Nicoya, Costa Rica Mw 7.6 earthquake rupture pro- cess from joint inversion of high-rate GPS, strong- motion, and teleseismic P wave data and its relation- ship to adjacent plate boundary interface properties, Journal of Geophysical Research: Solid Earth, 118, 5453-5466. Doi:10.1002/jgrb.50379. Ji C., K. M. Larson, Y. Tan, K. W. Hudnut, and K. Choi, 2004. Slip history of the 2003 San Simeon earth- quake constrained by combining 1-Hz GPS, strong motion, and teleseismic data. Geophysical Research Letters, 31(17), L17608, 10.1029/ 2004GL020448c. Jianghui Geng, 2015. = Lamjung%2C+Nepal. Jianghui Geng, Chuang Shi, Maorong Ge, Alan H. Dod- son, Yidong Lou, Qile Zhao, Jingnan Liu, 2012. Im- proving the estimation of fractional-cycle biases for ambiguity resolution in precise point positioning. Journal of Geodesy, 86, 579-589. Kouba J., 2009, A Guide to using International GNSS Service (IGS) products, Natural Resources Canada, Mahmoud Abd Rabbou, Ahmed El-Rabbany, 2015. PPP Accuracy Enhancement Using GPS/GLONASS Ob- servations in Kinematic Mode, Positioning, 6, 1-6. Mathias Lemmens, 2015. GPS and the 2015 Gorkha Earthquake, GIM International November 2015, 29-31. Matt King, Stuart Edwards and Peter Clarke, 2002. Pre- cise Point Positioning: Breaking the Monopoly of Relative GPS Processing, Engineering Surveying Showcase 10/2002, 33-34. Nguyen Ngoc Lau, 2009. How Accuracy GPS Precise Point Positioning can be achieved? Journal of Sci- ence and Technology Development, 12(18), 25-31. Nguyen Ngoc Lau, 2012. Monitoring the Tohoku earth- quake on 11. March 2011 by using GPS Technolo- gy, International Symposium on Geoinformatics for Spatial-Infrastructure Development in Earth and Al- lied science, Ho Chi Minh - Vietnam, 279-289. Nguyen Ngoc Lau, 2013. Precise Point Positioning us- ing GPS and GLONASS measurements, Journal of Geodesy and Cartography (in Vietnamese), 15, 9-17. Nguyen Ngoc Lau, Duong Minh Au, Nguyen Van Tuan, Dang Van Cong Bang, 2012. Precise point position- ing using GPS and GLONASS measurements. Re- port of ministry-level of project B2010-30-33B2012- 20-33. Nguyen Ngoc Lau, Tran Trong Duc, Duong Tuan Viet, Dang Van Cong Bang, 2010. Automatic GPS precise point processing via internet. Report of ministry- level of project B2010-30-33. Nguyen Ngoc Lau, Ha Minh Hoa, Richard Coleman, 2011. A new technique for processing GLONASS carrier phase measurements over medium length baselines using IGS products, International Journal of Geoinformatics, 7(4), 11-20. Tomokazu Kobayashi, Yu Morishita and Hiroshi Yarai, 2015. Detailed crustal deformation and fault rupture of the 2015 Gorkha earthquake, Nepal, revealed from ScanSAR-based interferograms of ALOS-2, Earth, Planets and Space, 67, 201p. UNAVCO, 2015, https://www.unavco.org/highlights/2015/nepal.html. USCG, 2015. https://earthquake.usgs.gov/research/everyone/nepal 2015/. Zumberge J.F., Heflin M.B., Jefferson D.C., Wattkins M.M. and Webb F.H., 1997. Precise point position- ing for the efficient and robust analysis of GPS data from large networks, Journal of Geophysical Re- search, 102(B3), 5005-5017.