Elsevier

Journal of Asian Earth Sciences

Volume 147, 1 October 2017, Pages 210-221
Journal of Asian Earth Sciences

Full length article
3-D magnetotelluric imaging of the Phayao Fault Zone, Northern Thailand: Evidence for saline fluid in the source region of the 2014 Chiang Rai earthquake

https://doi.org/10.1016/j.jseaes.2017.07.034Get rights and content

Highlights

  • The Phayao Fault Zone (PFZ) in Thailand is imaged via a 3-D magnetotelluric survey.

  • The shallow 3-D resistivity structure matches well with the surface geology.

  • A deep conductor related to the saline fluid is observed beneath the PFZ.

  • A high fault plane instability and saline fluid govern the slip of the PFZ.

Abstract

Seismicity in Thailand had been relatively low for decades prior to the Mw 6.5 earthquake of 5 May 2014 which came as a surprise and was followed by thousands of aftershocks. Most of the epicenters were located in the transition region between the Mae Lao Segment (MLS) and the Pan Segment (PS) of the Phayao Fault Zone (PFZ). We conducted a 3-D magnetotelluric (MT) survey (31 sites) to image the deep PFZ structure. The shallow 3-D resistivity structure matches very well with the surface geology, while the deeper structures disclose many interesting resistive and conductive anomalies. However, the most interesting feature of this study is the large conductive anomaly (ML) located at a depth of 4 km to the mid-crust beneath the MLS near the seismogenic zone. Our current hypothesis is that the ML conductor has a highly interconnected aqueous fluid content and also plays crucial role in the earthquake sequence of the 5 May 2014 event. As our previous seismic waveform study revealed that the MLS has a relatively high fault plane instability, the fluid within the fractured fault would further reduce the fault strength. The accumulated pre-existing tectonic stress from the north can therefore overcome the maximum frictional strength of the MLS, and hence cause it to slip and produce the main shock. With the local structural heterogeneities and fluid in the fractured fault zones, the aftershocks then occurred on both the PS and MLS. This is in contrast to the Mae Chan Fault Zone (MCFZ) in the north which many scientists expected to generate a larger magnitude earthquake than any other faults. Since instrumental record, it has only generated a few Mw 4 earthquakes. Some of our MT stations were located within the MCFZ. However, there is no deep conductor as the conductor lies beneath the MLS. A lack of interconnected fluid within the deep fault beneath the MCFZ might be one of the reasons for the lower seismic activity from the MCFZ. Other geophysical methods, such as seismic tomography, are necessary in order to confirm the presence of the fluid beneath the MLS and also the lack of a deep conductor beneath the MCFZ.

Introduction

Most of the seismicity in Thailand is localized in the north as there are many fault zones that cut through the region (Fig. 1a). The number and magnitude of earthquakes is generally lower than neighboring countries like Myanmar. Nevertheless, an unexpected moderate earthquake occurred on 5 May 2014 in Pan District, Chiang Rai province, northern Thailand (Fig. 1a and b) with thousands of aftershocks (Fig. 1b). The local magnitude (Ml) of the main shock was reported as 6.3 by the Thai Meteorological Department (TMD) with a hypocenter depth of 7 km, but later regional moment tensor inversion studies by Noisagool et al. (2016) suggested Mw 6.5 with a depth of 14 km. The main event was the largest earthquake in Thailand since the installation of the seismometer in 1974. The epicenters of the main shock and aftershocks were mostly in the Phayao Fault Zone (PFZ, Fig. 1b) which runs from south to north and is clearly separated into three major segments (Fig. 1b; Uttamo et al., 2003, DMR, 2007, DMR, 2014). The two segments involved with the recent events are the left–lateral strike–slip Mae Lao Segment (MLS) and the right lateral strike–slip Pan Segment (PS) (Fig. 1b). Interestingly, most of the epicenters are in the transition area between the MLS and the PS (Fig. 1b). The focal mechanisms (Noisagool et al., 2016) of most events (Ml > 4) indicate strike–slip motions in agreement with the characteristics of these fault segments.

It is interesting that the second largest recent earthquake in northern Thailand (Mw 5.2, 15 km depth, on 11 September 1994; USGS) also occurred in the PFZ (Fig. 1b). The focal mechanism of this earthquake was the normal motion (ISC, 2012; Fig. 1b). However, Noisagool et al. (2016) pointed out that its focal mechanism should be similar to the main event of the 5 May 2014. In fact, its epicenter is still in doubt as in 1994 there were no seismometers installed nearby. Since the PFZ is classified as a tentatively active fault (Charusiri et al., 1999) due to the lack of the data, even with the 1994 event (M5) and a few M4 events between 1994 and 2014 in this area, no scientists expected the PFZ to have enough energy to rupture and produce a Mw 6.5 earthquake so soon. This might be because of its very slow slip rate of about 0.1 mm/yr (Fenton et al., 2003), and relatively short length (45 km). In addition, Noisagool et al. (2016) has suggested a very crucial point about the regional stress and the fault plane instability revealed from the inversion of derived focal mechanisms from the 5 May 2014 earthquake sequences. They found that the Mae Lao Segment (MLS) has a higher shear stress than the other segments of the PFZ. With the stress from the north (≈N18E; Noisagool et al., 2016), the MLS is likely to be the fault that produced the main shock. Its movement has induced the Pan Segment (PS) which has a lower shear stress to slip. Noisagool et al. (2016) showed that there were mixtures of all three kinds of focal mechanisms: strike–slip, normal, and thrust motions. These mixed mechanisms result from the motions of both the MLS and PS and the structural heterogeneities between them.

In contrast to the PFZ, the east-trending Mae Chan Fault Zone (MCFZ, Fig. 1a) in the north of the PFZ has a slightly faster slip rate and is longer. It was classified as an active fault in the study of Charusiri et al. (1999). Most of the scientists then believed that the MCFZ would be the fault that produces large events based on its paleo-seismicity (Nutalaya et al., 1985, Charusiri et al., 2000, Fenton et al., 2003). This has led to many geological and geophysical studies on the MCFZ (e.g., Wood, 1995, Wood, 2001, Hinthong, 1995, Rymer et al., 1997, Kosuwan et al., 1998, Kosuwan et al., 2000, Kosuwan et al., 2003, Kosuwan and Lumjuan, 1999, Fenton et al., 2003, Wood et al., 2004, Wood et al., 2015, Phodee et al., 2014) in the past decades and little attention on the PFZ. In the end, only a few M4 earthquakes occurred in the MCFZ since the installation of first seismometer. With the accumulated tectonic stress coming from the north to both faults, this leads to the intriguing question of why the PFZ slipped instead of the MCFZ. Answers probably lie in the characteristics of these two faults, e.g. the fault plane, the frictional coefficient, and the interconnected fluid in the fault.

The occurrence of the largest earthquake in Thailand history on the PFZ is a good reason to conduct geological and geophysical investigations to probe the characteristic of the PFZ. One of the geophysical methods conducted worldwide is the magnetotelluric (MT) method. A magnetotelluric survey yields the resistivity structure which can be used to image faults and to study the fault properties and characteristics and its relation with earthquakes (Bedrosian et al., 2002, Becken and Ritter, 2012, Rao et al., 2004, Patro et al., 2005, Wannamaker et al., 2014; among many others). In many regions around the world, MT results reveal that high conductivity anomalies are often associated with the faults and seismicity of the region. In the San Andrea Fault, seismicity along each fault segment was linked to resistivity anomalies probed by MT surveys (Bedrosian et al., 2002, Bedrosian et al., 2004, Unsworth and Bedrosian, 2004a, Unsworth and Bedrosian, 2004b, Becken and Ritter, 2012). For the creeping segment of the fault, the high conductivity zone was observed within the fault indicating the presence of saline fluid. In contrast, in the locked fault segment, there was no high conductivity zone within the fractured fault. In the tectonic studies of the Central Indian Tectonic Zone, many conductors were found at different depths, e.g., from upper crust to mid-crust, and/or from mid- to lower crust, in association with the faults where seismicity also took place (e.g., Rao et al., 2004, Patro et al., 2005, Naganjaneyulu and Santosh, 2010, Abdul Azeez et al., 2013). Their interpretations of the low resistivity bodies were mainly based on either the mafic magmatic underplating and/or the presence of the aqueous fluid either in the fractures or intruding from the deep. In the Cascadia subduction system, USA (Wannamaker et al., 2014), the relatively high resistivity anomalies were interpreted as a lack of fluid and sediment indicating a plate locking zone, while the low resistivity anomalies were associated with the no plate lock zone implying with the presence of shallow sediment, and the fluid at great depth. In an arc-continent collision of Taiwan, Bertrand et al. (2012) found conductive zones at great depth and these have some relationship with the crustal seismicity. They interpreted the low resistivity zones as the interconnected saline fluid. A deep crustal conductor associated with the fluid was also observed in the New Zealand Southern Alps (Wannamaker et al., 2002) and in Japan (Mitsuhata et al., 2001, Ogawa and Honkura, 2004). The low resistivity anomalies were also found to be related to the low velocity zone in the lower crust beneath the Tangshan area indicating the existence of fluids which help to weaken the upper and middle crustal seismogenic layers and so cause large earthquakes (Wang et al., 2013). Similar scenarios were also found in the 1995 Kobe earthquake, Japan (Zhao et al., 1996), the 2001 Bhuj earthquake, India (Mishra and Zhao, 2003), the 2000 western Tottori earthquake, western Japan (Zhao et al., 2004) and the 2008 Iwate-Miyagi inland earthquake, Japan (Ichihara et al., 2011), among many others. With the correlation of low velocity and low resistivity, one of the explanations for the earthquake occurrence in these regions is the crustal heterogeneities, rather than just the stress condition alone (Zhao et al., 1996, Zhao et al., 2004, Mishra and Zhao, 2003, Wang et al., 2013)

Many of these studies have shown the importance of MT method for studying fault zones and earthquake mechanisms. Here, to study the characteristic of the PFZ, we conducted a 3-D MT survey covering the PFZ with more stations in the region where the Pan Segment intersects the Mae Lao Segment which is the area where most of the epicenters were located. In the absence of any geological and geophysical studies, our resistivity model obtained from the MT survey is therefore the first providing detailed information in the region around the PFZ.

Section snippets

Regional geologic framework of the Phayao Fault Zone and Northern Thailand

The tectonic setting and the regional geologic map around the Phayao Fault Zone (PFZ) are shown in Fig. 2a and b, respectively. According to the tectono-stratigraphic zone (Metcalfe, 2013, Morley et al., 2011, DMR, 2014), the tectonics of the northern Thailand is divided into the Inthanon zone on the west (west of the Mae Tha Fault Zone, MTFZ, in Fig. 2a), the north – south trend Sukhothai fold belt in the center which covers most of our study area (Fig. 2a), and the Indochina terrane on the

Magnetotelluric survey and the three-dimensional inversion

In November 2015, 29 MT stations were deployed covering a 140 km × 140 km area (Fig. 1b and Fig. 2) to ensure that the MT technique can “image” the near surface down to the lower crust. Uniform coverage of the stations is impossible as some of the land was not accessible due to lack of navigable terrain and/or permission from the owner. Most of the MT sites were above sedimentary rocks (Fig. 2b) in rice and corn farms. Only a few sites were placed over the granitic rocks. The main interest of this

Interpretation and discussion

As stated earlier, there are two main reasons why we prefer the inverted model in Fig. 5, Fig. 6 as our final model. First, it can produce the responses that fit the observed data with an RMS of 1.97 (Fig. 3 and Fig. 4). Second, the surface geology (Fig. 7a) corresponds closely with the final inverted model at 50 m depth (Fig. 7b). West of the PFZ, the Triassic granite rock (Tgr) corresponds well with high resistivity bodies (>100 Ω m). Most of the Quaternary alluvial deposit (Qa; Fig. 7a) within

Conclusion

  • A total of 31 magnetotelluric (MT) stations were deployed within a 140 km × 140 km area to investigate the Phayao Fault Zone (PFZ) rupture area in northern Thailand.

  • We found conductive anomalies beneath the Mae Lao Segment (MLS) and partially beneath the Pan Segment (PS) from a depth of 4 km to mid-crust where both MLS and PS are part of the PFZ. They were interpreted as saline fluid rich zones.

  • The saline fluid rich zones strongly support our assumption of previous studies (Noisagool et al., 2016)

Acknowledgement

This research has been supported by the Development and Promotion of Science and Technology Research Grant 037/2557, the Thailand Center of Excellence in Physics (ThEP), and the Thailand Research Fund (RSA5780010). We would like to thank Dr. Michael Allen for editing the English of this manuscript, and the two reviewers who provide many good comments to improve the manuscript.

References (72)

  • W. Siripunvaraporn et al.

    Three-dimensional Magnetotelluric inversion: data-space method

    Phys. Earth Planet. Inter.

    (2005)
  • W. Siripunvaraporn et al.

    WSINV3DMT: vertical magnetic field transfer function inversion and parallel implementation

    Phys. Earth Planet. Inter.

    (2009)
  • M. Sone et al.

    Parallel Tetyan sutures in mainland Southeast Asia: new insights for Palaeo-Tethys closure and implications for the Indosinian Orogeny

    C.R. Geosci.

    (2008)
  • M. Sone et al.

    The Chanthaburi terrane of southeastern Thailand: stratigraphic confirmation as a disrupted segment of the Sukhothai Arc

    J. Asian Earth Sci.

    (2012)
  • M.R.P. Tingay et al.

    Present-day stress orientation in Thailand’s basins

    J. Struct. Geol.

    (2010)
  • J. Wang et al.

    Crustal and uppermost mantle structure and seismotectonics of North China Craton

    Tectonophysics

    (2013)
  • D.P. Zhao et al.

    Crustal heterogeneity in the 2000 western Tottori earthquake region: effect of fluids from slab dehydration

    Phys. Earth Planet. Inter.

    (2004)
  • K.K. Abdul Azeez et al.

    Resistivity structure of the Central Indian Tectonic Zone (CITZ) from Multiple Magnetotelluric (MT) profiles and tectonic implications

    Pure Appl. Geophys.

    (2013)
  • S. Aoki et al.

    Three-dimensional distribution of S wave reflectors in the northern Kinki district, southwestern Japan, Earth

    Planets Space

    (2016)
  • M. Becken et al.

    Magnetotelluric studies at the San Andreas fault Zone: implications for the role of fluids

    Surv. Geophys.

    (2012)
  • P.A. Bedrosian et al.

    Magnetotelluric imaging of the creeping segment of the San Andreas Fault near Hollister

    Geophys. Res. Lett.

    (2002)
  • E.A. Bertrand et al.

    Magnetotelluric imaging beneath the Taiwan orogen: an arc-continent collision

    J. Geophys. Res.

    (2012)
  • P. Charusiri et al.

    Review of active faults and seismicity in Thailand. GEOSEA’98

    Bull. Geol. Soc. Malays.

    (1999)
  • Charusiri, P., Kosuwan, S., Daorerk, W., Wechbunthung, B., Kuthranon, S., 2000. Earthquake in Thailand and Southeast...
  • P. Charusiri et al.

    Cenozoic tectonic evolution of major sedimentary Basins in Central, Northern, and the Gulf of Thailand

    Bull. Earth Sci. Technol.

    (2009)
  • F. Cox

    Fluid flow in mid-to deep crustal shear systems: experimental constraints, observations on exhumed high fluid flux shear systems, and implications for seismogenic processes

    Earth, Planets and Space

    (2002)
  • Department of Mineral Resource

    Geology and Resource Chiang Rai Province Classification and Management

    (2007)
  • Department of Mineral Resource, 2014. Report on a Study of 5th May 2014, 6.3 Earthquake. Department of Mineral...
  • G.D. Egbert

    Robust multiple-station magnetotelluric data processing

    Geophys. J. Int.

    (1997)
  • C.H. Fenton et al.

    Recent paleoseismic investigations in Northern and Western Thailand

    Ann. Geophys.

    (2003)
  • K. Fujimoto et al.

    Water-rock interaction observed in the brittle-plastic transition zone

    Earth Planets Space

    (2002)
  • Hinthong C, 1995. The study of active faults in Thailand. Proceedings of the technical conference on the progression...
  • R.D. Hyndman et al.

    Water in the lower continentalcrust: Modeling magneto-telluric and seismic reflection results

    Geophys. J. Int.

    (1989)
  • H. Ichihara et al.

    Resistivity and density modeling in the 1938 Kutcharo earthquake source area along a large caldera boundary

    Earth, Planets Space

    (2009)
  • H. Ichihara et al.

    A fault-zone conductor beneath a compressional inversion zone, northeastern Honshu, Japan

    Geophys. Res. Lett.

    (2011)
  • International Seismological Centre, 2012, On-line Bulletin. Internatl. Seis. Cent., Thatcham, United Kingdom...
  • Cited by (7)

    • Structural interpretation of Mae Suai Basin, Chiang Rai Province, based on gravity data analysis and modelling

      2019, Heliyon
      Citation Excerpt :

      The NNE-SSW striking faults are probably a predominantly oblique-normal slip fault, while Fault F3 is probably an oblique strike-slip fault with a normal displacement component. Relocation of earthquake epicenters of the 2014 Earthquakes by Noisagool et al. (2016) and Pananont et al. (2017) (Fig. 2b) and the larger scale high conductive fault zone of the Magnetotellurics inversion solution by Boonchaisuk et al. (2017) are also compatible with the ENE-WSE orientation of Fault F3 (Figs. 7 and 15). Interpretations of the moment tensor solutions for the 2014 earthquakes by Pananont et al. (2017) indicate Fault F3 fits with a sinistral motion model (Fig. 2b).

    • Evaluating the state of stress and seismic hazard in Thailand and vicinity through finite element modeling

      2018, Journal of Asian Earth Sciences
      Citation Excerpt :

      Studies of these mapped faults and seismicity Mw ≤ 4.0 were used to identify this region as having a 10% probability of exceeding a peak ground acceleration (PGA) of 0.15 g for a 50-year time window (Ornthammarath et al., 2011). On 5 May 2014, the Mw 6.2 Mae Lao earthquake occurred in this region of mapped faults, within an area inferred to have high fluid content (Boonchaisuk et al., 2017). The event produced a maximum PGA of 0.30 g measured on top of a dam 15 km from the epicenter and 0.13 g at a ground-based station 27 km from the epicenter (Ornthammarath and Warnitchai, 2015), consistent with the previously estimated hazard probabilities.

    View all citing articles on Scopus
    View full text