- Email: [email protected]
3-D velocity structure of Damavand volcano, Iran, from local earthquake tomography
Journal of Asian Earth Sciences 42 (2011) 1091–1096
Contents lists available at ScienceDirect
Journal of Asian Earth Sciences journal homepage: www.elsevier.com/locate/jseaes
3-D velocity structure of Damavand volcano, Iran, from local earthquake tomography Akram Mostafanejad a,b,⇑, Z. Hossein Shomali b,c, Ali A. Mottaghi b a
Center for Earthquake Research and Information, University of Memphis, 3890 Central Ave. Memphis, TN 38152, USA Institute of Geophysics, University of Tehran, 14155-6466, Iran c Department of Earth Sciences, Uppsala University, Uppsala 75236, Sweden b
a r t i c l e
i n f o
Article history: Received 21 May 2010 Received in revised form 4 March 2011 Accepted 9 March 2011 Available online 21 March 2011 Keywords: Local earthquake tomography Magma chamber Crustal structure Damavand volcano
a b s t r a c t Damavand volcano is a large intraplate Quaternary composite cone overlying the active fold and thrust belt of the Central Alborz Mountains in northern Iran. In this study, we present the first 3-D P-wave velocity model of the upper crust for the Central Alborz region using local earthquake data provided by the Iranian Seismic Telemetry Network. The final P-wave velocity model reveals several high and low velocity anomalies in the upper 30 km of the crust. A low velocity zone parallel to the main faulting system is imaged at a depth of 15 km. A relatively high velocity body to the north of the Damavand summit down to a depth of 20 km is resolved, which may represent the elderly crystallized magma chamber. Right below the Damavand cone, a low velocity area to the depth of 7 km can be interpreted as a shallow magma chamber. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Much progress has been made in recent years in resolving the physical and chemical characteristics of volcanic fields. Improved datasets and methods have enabled 3-D imaging of volcanic areas with great detail (e.g. Lees, 2007; Salah and Seno, 2008; Waite and Moran, 2009; Zhao et al., 2011). High velocity anomalies observed in these areas are generally addressed to as older, consolidated magmatic bodies, and low velocity anomalies are regarded as the presence of molten zone or fluids (Lees, 2007). The Alborz Mountain range in the southern margin of the Caspian Sea is a part of the Alpine–Himalayan orogenic belt. In the central part of this compressional ranges, Damavand volcano has been constructed in Quaternary (Allenbach, 1966; Davidson et al., 2004). Formation of such a large but isolated volcano, with regard to the overall compressional regime in the Alborz Mountains, has long been subjected to discussions. We performed a local earthquake tomography in this region for the first time, which aims at elucidating some undetermined aspects of this complex geological region, particularly internal structure of the Damavand volcano. 2. Geological setting Damavand volcano, a young, dormant strato-volcano situated 50 km north of the metropolitan Tehran, is a large intraplate ⇑ Corresponding author at: Center for Earthquake Research and Information, University of Memphis, 3890 Central Ave. Memphis, TN 38152, USA. E-mail address: [email protected] (A. Mostafanejad). 1367-9120/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jseaes.2011.03.011
Quaternary composite cone of trachyandesite lava and pyroclastic deposits overlying the active fold and thrust belt of the Central Alborz Mountains. It has an altitude of 5670 m, which corresponds to the highest peak in the Middle East. Isotope dating, and geological studies, have revealed that the present cone (young Damavand) has been constructed over the last 600 k.y. a little to the south and on an older, eroded edifice of the old Damavand (Davidson et al., 2004). The Alborz Mountains, with an arc shape fold and thrust belt borders the Caspian Sea in the south (Stöcklin, 1974; Axen et al., 2001; Davidson et al., 2004). According to gravity modeling the crustal thickness beneath the Alborz Mountains is about 35 km (Dehghani and Makris, 1984). The central part of the Alborz Mountains is characterized by a ‘‘V-shaped’’ structure (at about 50E to 54E) separating folds and faults with a NW–SE trend to the west from a NE–SW trend in the east (Berberian, 1976) (Fig. 1). The geology of the Alborz Mountains is controlled largely by major thrust faults. The Mosha fault (Fig. 1) is the most prominent structure in the southern part of the Central Alborz region (Moinabadi and Yassaghi, 2007). It is an active, high dip angle reverse fault with an approximate length of 400 km (Berberian et al., 1993), which had experienced several destructive earthquakes with magnitude greater than 6.5 (Berberian and Yeats, 2001). The maximum displacement along the Mosha fault is up to 4 km (Allenbach, 1966; Tchalenko et al., 1974). The North Tehran fault is another major feature in this region. 20 km to the east of Tehran, the North Tehran fault diverges from the Mosha fault and forms a north dipping, active, left lateral oblique thrust fault (Allen et al., 2003). Guest et al. (2006) suggested that the North Tehran thrust merges with
1092
A. Mostafanejad et al. / Journal of Asian Earth Sciences 42 (2011) 1091–1096
Fig. 1. (a) Map of Central Alborz Mountains showing stations (green squares) and local earthquakes (from 1996 to 2006, purple circles) used in this study. Yellow triangle: Damavand volcano, black lines: main faults in the region, diamond: Tehran, white dashed lines: profile along Damavand volcano and 10 km to the east and west side, cross: grid node. Node spacing is 20 km. Fault plane solutions of large events (Mw > 5.0) are from Harvard University (2009). Inset map shows location of studied area in the Middle East. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Mosha fault at depth. Damavand volcano is situated 20 km northwards from the Mosha active fault and lies unconformably on folds and thrusts (Liotard et al., 2008). Tectonic activity in the Alborz Mountains is due both to the convergence of central Iran toward stable Eurasia, and to motion of the South Caspian Basin, with respect to Eurasia. Coupled with the NS convergence of central Iran (52 mm/yr), the south-westward motion of the South Caspian Basin, with respect to central Iran, leads to a NNE-SSW transpressional regime in the Alborz (Vernant et al., 2004; Ritz et al., 2006). Ritz et al. (2006) documented a transition from transpression to active transtension in very recent time in the Central Alborz region. According to their estimation, the transtension started between 1 and 1.5 Ma. This is contemporaneous with the Damavand volcanic activity, which Davidson et al. (2004) dated as occurring between 1.8 Ma and 7 Ka.
3. Data and method Fig. 1 shows the locations of the seismic stations and epicenters of relocated earthquakes. The seismic array consists of 19 permanent stations of the Iranian Seismic Telemetry Network (IRSC) equipped with 3-component short period (SS1) seismometers. The location of earthquakes was improved through using the hypoDD method (Waldhauser and Ellsworth, 2000). HypoDD determines relative locations within clusters using the doubledifference algorithm. This algorithm improves relative location accuracy by removing effects due to an un-modeled velocity structure through the propagation path. A significant improvement in
hypocenter locations is observed by comparing the relocated earthquakes using hypoDD and those routinely determined by IRSC (Mottaghi et al., 2010). We selected 895 relocated earthquakes between 1996 and 2006 based on the following criteria: (1) all the local earthquakes used are located within the seismic network with azimuthal gap less than 180°, (2) the uncertainties of the hypocentral locations (both horizontal and vertical) are, on average, less than 6 km during relocation with the 1-D background model and (3) all the selected earthquakes are recorded by at least four stations. Our final dataset is composed of 5559 P-wave observations, giving an average of 6 P-wave observations per event (Figs. 1 and 2). For the initial 1-D velocity model, we used the model developed by Ashtari et al. (2005), which is based on 1-D inversion of the arrival times for local events (Kissling, 1988). We used a damped least-squared iterative method (SIMULPS14), (Thurber, 1983; Eberhart-Phillips, 1990; Evans et al., 1994) to solve the non-linear tomography problem (Haslinger and Kissling, 2001; Husen et al., 2003). Hypocenter locations are included in the inversion as unknowns, due to the coupling of hypocenter locations and velocities (Thurber, 1992). Travel times through the velocity model are calculated using full 3-D shooting ray tracing (Virieux and Fara, 1991; Haslinger and Kissling, 2001). Having considered the ray coverage of the dataset (Fig. 2), we chose a horizontal model parameterization of 20 20 km for the regional model. This model parameterization represents the finest possible model grid spacing without showing a strongly heterogeneous pattern of the derivative weight sum (DWS) (Husen et al., 2000, 2003). Grid node spacing with depth is 5 km. Following the method introduced by Eberhart-Phillips (1986) for estimation of an optimum damping
A. Mostafanejad et al. / Journal of Asian Earth Sciences 42 (2011) 1091–1096
1093
Fig. 2. Location of the 895 selected local events (gray dots), and grid nodes position (gray crosses). Horizontal ray coverage of the dataset is depicted in black solid lines. Gray squares: stations, yellow triangle: Damavand volcano. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
value, we performed a large range of single iteration inversions with damping values ranging from 1 to 1000. The optimum damping value of 50 was obtained based on the trade-off curve between data variance and model length. The trade-off curve shows similar trends, both for inversions with and without station corrections (see Fig. 3). A strong increase in data variance with decreasing damping can be avoided by including station delays in the inversion process, as depicted in Fig. 3. Station delays are usually used to compensate for the near-surface structure beneath the seismic stations (e.g., Husen et al., 2003). The effect of stations delays is
Fig. 3. Trade-off curve for single iteration results with (squares) and without (circles) station correction. The selected value of damping, 50, is shown as a black square.
more pronounced as the degree of velocity heterogeneity increases as a result of lowered damping. 4. Inversion results A significant problem in seismic tomography is to optimize and evaluate the spatial resolution of the estimated model. Model resolution is limited by various factors including data errors and the three-dimensional geometry of the crossing rays through the volume. In order to evaluate the resolving power of the dataset, we incorporated various measures including a checkerboard test, assessing DWS, Resolution Diagonal Elements (RDE), and hit count. An alternating 50 50 km high and low velocity (checkerboard) of 10%, relative to the 1-D initial velocity model, was produced using the real ray geometry of the data. The synthetic dataset is then inverted in the same way as real data. The result of the synthetic checkerboard test is shown in Fig. 4a. It indicates that the resolution around the Damavand volcano in the central part of the model is fairly good with the least smearing between adjacent nodes. DWS is a measure of ray density in the vicinity of a model parameter (Toomey and Foulger, 1989) that weighs the importance of each ray segment by its distance to the model node. RDE is yet a better estimate of the model resolution. In an ideal situation, zero value of RDE implies no resolution, while 1 stands for perfect resolution. However, the amount of RDE strongly depends on different factors like model parameterization and damping value. Increasing the density of model parameters, although increases the fidelity of the model, it decreases RDE (Toomey and Foulger, 1989). Fig. 5 shows the amount of RDE and DWS at the depth sections of 5, 10 and 15 km as well as vertical cross sections of DWS along the
1094
A. Mostafanejad et al. / Journal of Asian Earth Sciences 42 (2011) 1091–1096
Fig. 4. (a) Horizontal slices showing a synthetic checkerboard model test with 10% velocity anomaly relative to 1-D initial model. Yellow triangle: Damavand volcano. (b) Horizontal slices of real inversion results after 3 iterations. (c) Vertical slices through Damavand volcano (middle panel) and 10 km to the west and east of the volcano (the locations of the cross-section are shown in Fig. 1). Black circles show hypocenters of earthquakes. Black contour lines show absolute velocity. Pink contour outlines area with diagonal elements of the resolution matrix > 0.1. The topography variations along the profiles are also shown on top of each panel respectively. The position of Damavand volcano and surface trace of the Mosha fault are also denoted by arrows. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
volcano and 10 km to the east and west sides of it. Based on the result of synthetic testing, we chose areas with RDE greater than 0.1 as the region with reliable resolution. This value, while low, is consistent with previous studies using this inversion method (e.g. Haslinger et al., 1999; Waite and Moran, 2009). In summary, our extensive synthetic tests indicate that regularisation parameters used for inversion enhance the resolution, especially in the central part of the model. The final P-wave velocity model reveals several high and low velocity anomalies in the upper 30 km of the crust in the Central Alborz region (Fig. 4b and c). The first striking feature of the resulted images is the almost perfect correlation between the pronounced V-shape structure of the Alborz Mountain range and the linear low velocity anomaly along the main faulting system at 15 km depth (Fig. 4b). We relate this low velocity anomaly to the fractured zone caused by the Mosha and the North Tehran faults (see Fig. 1). Based on stratigraphy of the region, the sedimentary cover in the Alborz Mountains is up to 6 to 8 km thick (Stöcklin, 1974) and is underlain by crystalline basement (Dedual, 1967). As a result, it is very probable that the fracturing pattern, as well
as P-wave velocity, vary considerably, which eventually causes the linear low velocity along the Mosha and North Tehran fault to be so conspicuous only in depth of 15 km and not in shallower depths. In the vertical section across the volcano a relatively high velocity zone is clearly imaged in the upper 20 km of the model on the north side of the Damavand volcano (Fig. 4c). We interpret this relatively high velocity body as the elderly crystallized magma chamber. The position of the hypothetical conduit related to this old chamber is in accordance with the old Damavand cone, which is slightly to the north of the present cone (Davidson et al., 2004) (Fig. 4c). A relatively low velocity body is detected from near surface to the depth of 7 km, right below the young Damavand cone. We interpret this relatively low velocity region as being the magma chamber of Damavand volcano. It is worth mentioning that a geothermal reservoir of 550 km2 with a mean temperature of 190 °C at depths of 2–3 km and estimated thermal energy of 5.11 1018 J is investigated in Damavand area (Ghobadian et al., 2009). The 3-D velocity model strongly reduces travel time residuals by 58% after three iterations (corresponds to RMS residuals from
A. Mostafanejad et al. / Journal of Asian Earth Sciences 42 (2011) 1091–1096
1095
Fig. 5. (a) Resolution Diagonal Elements (RDE), and (b) Derivative weight sum (DWS) for depth sections 5, 10 and 15 km (c) DWS for vertical slices through Damavand volcano and 10 km to the east and west of the volcano.
1.2 to 0.7 s). However, the variance reduction depends on number of regularization parameters, including the damping factor. Thus, we consider the model presented as a smoothed model describing the major structural features resolved by the inversion. It is worth noting that the remnant residuals of the final model are high (0.7 s). This relatively high value should be considered next to the relatively high phase picking error, 0.2 s, of the dataset. It can be assumed that the remnant residuals partly reflect some features of the true velocity anomalies in the earth that cannot be taken into account by our model. The most important one might be anisotropy, which can be strong in this highly heterogeneous area. The other factor can be related to the existence of some small-scale velocity anomalies in the uppermost crust, which are below the resolution capacity of our inversion scheme. A detailed understanding of the geological setting of the studied area requires the deployment of a dense seismic network, which is clearly the ambition of further research in this area. 5. Conclusion For the first time, a 3-D velocity model of the Central Alborz region, including the Damavand volcano, is presented for the upper
30 km of the crust using the available local earthquake dataset. The 3-D model reveals a number of interesting features including: 1. A low velocity anomaly along the main structural features at the depth of 15 km which is due to the fractured zone caused by the Mosha and North Tehran faults. 2. A relatively high velocity body is resolved in the northern side of the Damavand volcano almost in the same location as the old Damavand caldera. We consider this to be the elderly crystallized magma chamber. 3. There is a pronounced localized low velocity zone beneath the volcano to the depth of 7 km that is considered to be the magma chamber of the Damavand volcano.
Acknowledgments We are grateful to the Iranian Seismological Center (http://irsc.ut.ac.ir/) for the access to the earthquake data base. We thank M. Berberian for his early comments and encouragements, and B. Darvishzad for invaluable discussions and help. Roland Roberts, from Uppsala University, is especially thanked for his valuable comments. We appreciate H. DeShon and C. Powell for their
1096
A. Mostafanejad et al. / Journal of Asian Earth Sciences 42 (2011) 1091–1096
insightful comments and suggestions during the review process. We also thank two anonymous reviewers for their constructive suggestions which helped further improvements of the manuscript. Figures are produced using the GMT software (Wessel and Smith, 1998).
References Allen, M., Ghassemi, M.R., Shahrabi, M., Qorashi, M., 2003. Accommodation of late Cenozoic oblique shortening in the Alborz range, northern Iran. J. Struct. Geol. 25, 659–672. Allenbach, P., 1966. Geologie und petrographie des Damavand und seiner umgeurg (Zentral Elburz). Geologisches Institut, ETH-Zurich, Mitteilung Nr. 63, Iran, 114 p. (In german). Ashtari, M., Hatzfeld, D., Kamalian, N., 2005. Microseismicity in the region of Tehran. Tectonophysics 395, 193–208. Axen, G.J., Lam, P.S., Grove, M., Stockli, D.F., Hassanzadeh, J., 2001. Exhumation of the west-central Alborz Mountains, Iran, Caspian subsidence, and collisionrelated tectonics. Geology 29, 559–562. Berberian, M., 1976. An explanatory note on the first seismotectonic map of Iran; A seismo-tectonic review of the country. Geol. Surv. Iran. 39, 7–141. Berberian, M., Yeats, R.S., 2001. Contribution of archaeological data to studies of earthquake history in the Iranian plateau. J. Struct. Geol. 23, 563–584. Berberian, M., Qorashi,M., Argang Ravesh, B., Mohajer Ashjaie, A., 1993. Seismotectonics and earthquake-fault hazard investigation in the Tehran Region, contribution to the seismotectonics of Iran, Report. 56, Geol. Surv. Iran. Iran. Davidson, J., Hassanzadeh, J., Berzins, R., Stockli, D.F., Bashukooh, B., Turrin, B., Pandamouz, A., 2004. The geology of Damavand volcano, Alborz Mountains, northern Iran. GSA Bull. 116, 16–29. Dedual, E., 1967. Zur Geologic des mittleren und unteren Karaj-Tales Zentral-Elburz (lran). Mitt. Geol. Inst ETH u. Univ. Zurich, n.s., 76 (In German) 123p. Dehghani, G.A., Makris, J., 1984. The gravity field and crustal structure of Iran: Neues Jahrbuch für Geologie und Palaontologie Abhandlungen. 168, 215–229. Eberhart-Phillips, D., 1986. Three-dimensional velocity structure in Northern California Coast Ranges from inversion of local earthquake arrival times. Bull. Seismol. Soc. Am. 76, 1025–1052. Eberhart-Phillips, D., 1990. Three-dimensional P and S velocity structure in the Coalinga region, California. J. Geophys. Res. 95, 343–363. Evans, J.R., Eberhart-Phillips, D., Thurber, C.H., 1994. User’s manual for SIMULPS12 for imaging Vp and Vp/Vs: a derivative of the ‘‘Thurber’’ tomographic inversion SIMUL3 for local earthquakes and explosions. Open-File Rep 94-431. U.S. Geol. Surv., Reston, VA. 101pp. Ghobadian, B., Najafi, G., Rahimi, H., Yusaf, T.F., 2009. Future of renewable energies in Iran. Renew. Sust. Energy Rev. 13, 689–695. Guest, B., Axen, G.J., Lam, P.S., Hassanzadeh, J., 2006. Late Cenozoic shortening in the west-central Alborz Mountains, northern Iran, by combined conjugate strikeslip and thin-skinned deformation. Geosphere 2, 35–52. Harvard University (2009), Global CMT catalog,, (accessed October 2009). Haslinger, F., Kissling, E., 2001. Investigating effects of 3-D ray tracing methods in local earthquake tomography. Phys. Earth Planet. Int. 123, 103–114. Haslinger, F., Kissling, E., Ansorge, J., Hatzfeld, D., Papadimitriou, E., Karakostas, V., Makropoulos, K., Kahle, H.-G., Peter, Y., 1999. 3D crustal structure from local
earthquake tomography around the Gulf of Arta (Ionian region, NW Greece). Tectonophysics 304, 201–218. Husen, S., Kissling, E., Flueh, E.R., 2000. Local earthquake tomography of shallow subduction in north Chile: a combined onshore and offshore study. J. Geophys. Res. 105, 28183–28198. Husen, S., Kissling, E., Deichmann, N., Wiemer, S., Giardini, D., Baer, M., 2003. Probabilistic earthquake location in complex 3-D velocity models: application to Switzerland. J. Geophys. Res. 108, 2077–2096. Kissling, E., 1988. Geotomography with local earthquake data. Rev. Geophys. 26, 659–698. Lees, J.M., 2007. Seismic tomography of magmatic systems. J. Volcanol. Geotherm. Res. 167, 37–56. Liotard, J.M., Dautria, J.M., Bosch, D., Condomines, M., Mehdizadeh, M., Ritz, J.F., 2008. Origin of the absarokite–banakite association of the Damavand volcano (Iran): trace elements and Sr, Nd, Pb isotope constraints. Int. J. Earth Sci. (Geol Rundsch). 97, 89–102. Moinabadi, E.M., Yassaghi, A., 2007. Geometry and kinematics of the Mosha Fault, south central Alborz Range, Iran: an example of basement involved thrusting. J. Asian Earth Sci. 29, 928–938. Mottaghi, A., Rezapour, M., Yaminifard, F., 2010. Double-difference relocation of earthquake hypocenters along the southern flank of the Central Alborz (Iran). Bull. Seismol. Soc. Am. 100, 2014–2023. Ritz, J.F., Nazari, H., Ghassemi, A., Salamati, R., Shafei, A., Solaymani, S., Vernant, P., 2006. Active transtension inside central Alborz: a new insight into northern Iran–southern Caspian geodynamics. Geology 34, 477–480. Salah, M.K., Seno, T., 2008. Imaging of Vp, Vs, and Poisson’s ratio anomalies beneath Kyushu, southwest Japan: implications for volcanism and forearc mantle wedge serpentinization. J. Asian Earth Sci. 31, 404–428. Stöcklin, J., 1974. Northern Iran: Alborz Mountains. In: Spencer, A.M. (Ed.), Mesozoic–Cenozoic Orogenic Belts data for Orogenic Studies; Alpin— Himalayan Orogens. Geological Society Special Publication 4, London, pp. 212–234. Tchalenko, J. S., Berberian, M., Iranmanesh, H., Bailly, M., Arsovsky, M., 1974. Tectonic framework of the Tehran region. Geological Survey of Iran, Report no. 29. Thurber, C.H., 1983. Earthquake locations and three-dimensional crustal structure in the Coyote Lake area, central California. J. Geophys. Res. 88, 8226–8236. Thurber, C.H., 1992. Hypocenter-velocity structure coupling in local earthquake tomography. Phys. Earth Planet. Inter. 75, 55–62. Toomey, D.R., Foulger, G.R., 1989. Tomographic inversion of local earthquake data from the Hengill–Grensdalur central volcano complex, Iceland. J. Geophys. Res. 94, 17497–17510. Vernant, P., Nilforoushan, F., Chéry, J., Bayer, R., Djamour, Y., Masson, F., Nankali, H., Ritz, J.F., Sedighi, M., Tavakoli, F., 2004. Deciphering oblique shortening of central Alborz in Iran using geodetic data. Earth Planet. Sci. Lett. 223, 177–185. Virieux, J., Fara, V., 1991. Ray tracing for earthquake location in laterally heterogeneous media. J. Geophys. Res. 93, 6585–6599. Waite, G.P., Moran, S.C., 2009. VP structure of Mount St. Helens, Washington, USA, imaged with local earthquake tomography. J. Volcanol. Geotherm. Res. 182, 113–122. Waldhauser, F., Ellsworth, W.L., 2000. A double-difference earthquake location algorithm: method and application to the northern Hayward Fault, California, Bull. Seismol. Soc. Am. 90, 1353–1368. Wessel, P., Smith, W.H.F., 1998. New Improved Version of Generic Mapping Tools Released, vol. 79. EOS Trans. AGU, 579. Zhao, D., Yu, S., Ohtani, E., 2011. East Asia: seismotectonics, magmatism and mantle dynamics. J. Asian Earth Sci. 40, 689–709.
Contents lists available at ScienceDirect
Journal of Asian Earth Sciences journal homepage: www.elsevier.com/locate/jseaes
3-D velocity structure of Damavand volcano, Iran, from local earthquake tomography Akram Mostafanejad a,b,⇑, Z. Hossein Shomali b,c, Ali A. Mottaghi b a
Center for Earthquake Research and Information, University of Memphis, 3890 Central Ave. Memphis, TN 38152, USA Institute of Geophysics, University of Tehran, 14155-6466, Iran c Department of Earth Sciences, Uppsala University, Uppsala 75236, Sweden b
a r t i c l e
i n f o
Article history: Received 21 May 2010 Received in revised form 4 March 2011 Accepted 9 March 2011 Available online 21 March 2011 Keywords: Local earthquake tomography Magma chamber Crustal structure Damavand volcano
a b s t r a c t Damavand volcano is a large intraplate Quaternary composite cone overlying the active fold and thrust belt of the Central Alborz Mountains in northern Iran. In this study, we present the first 3-D P-wave velocity model of the upper crust for the Central Alborz region using local earthquake data provided by the Iranian Seismic Telemetry Network. The final P-wave velocity model reveals several high and low velocity anomalies in the upper 30 km of the crust. A low velocity zone parallel to the main faulting system is imaged at a depth of 15 km. A relatively high velocity body to the north of the Damavand summit down to a depth of 20 km is resolved, which may represent the elderly crystallized magma chamber. Right below the Damavand cone, a low velocity area to the depth of 7 km can be interpreted as a shallow magma chamber. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Much progress has been made in recent years in resolving the physical and chemical characteristics of volcanic fields. Improved datasets and methods have enabled 3-D imaging of volcanic areas with great detail (e.g. Lees, 2007; Salah and Seno, 2008; Waite and Moran, 2009; Zhao et al., 2011). High velocity anomalies observed in these areas are generally addressed to as older, consolidated magmatic bodies, and low velocity anomalies are regarded as the presence of molten zone or fluids (Lees, 2007). The Alborz Mountain range in the southern margin of the Caspian Sea is a part of the Alpine–Himalayan orogenic belt. In the central part of this compressional ranges, Damavand volcano has been constructed in Quaternary (Allenbach, 1966; Davidson et al., 2004). Formation of such a large but isolated volcano, with regard to the overall compressional regime in the Alborz Mountains, has long been subjected to discussions. We performed a local earthquake tomography in this region for the first time, which aims at elucidating some undetermined aspects of this complex geological region, particularly internal structure of the Damavand volcano. 2. Geological setting Damavand volcano, a young, dormant strato-volcano situated 50 km north of the metropolitan Tehran, is a large intraplate ⇑ Corresponding author at: Center for Earthquake Research and Information, University of Memphis, 3890 Central Ave. Memphis, TN 38152, USA. E-mail address: [email protected] (A. Mostafanejad). 1367-9120/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jseaes.2011.03.011
Quaternary composite cone of trachyandesite lava and pyroclastic deposits overlying the active fold and thrust belt of the Central Alborz Mountains. It has an altitude of 5670 m, which corresponds to the highest peak in the Middle East. Isotope dating, and geological studies, have revealed that the present cone (young Damavand) has been constructed over the last 600 k.y. a little to the south and on an older, eroded edifice of the old Damavand (Davidson et al., 2004). The Alborz Mountains, with an arc shape fold and thrust belt borders the Caspian Sea in the south (Stöcklin, 1974; Axen et al., 2001; Davidson et al., 2004). According to gravity modeling the crustal thickness beneath the Alborz Mountains is about 35 km (Dehghani and Makris, 1984). The central part of the Alborz Mountains is characterized by a ‘‘V-shaped’’ structure (at about 50E to 54E) separating folds and faults with a NW–SE trend to the west from a NE–SW trend in the east (Berberian, 1976) (Fig. 1). The geology of the Alborz Mountains is controlled largely by major thrust faults. The Mosha fault (Fig. 1) is the most prominent structure in the southern part of the Central Alborz region (Moinabadi and Yassaghi, 2007). It is an active, high dip angle reverse fault with an approximate length of 400 km (Berberian et al., 1993), which had experienced several destructive earthquakes with magnitude greater than 6.5 (Berberian and Yeats, 2001). The maximum displacement along the Mosha fault is up to 4 km (Allenbach, 1966; Tchalenko et al., 1974). The North Tehran fault is another major feature in this region. 20 km to the east of Tehran, the North Tehran fault diverges from the Mosha fault and forms a north dipping, active, left lateral oblique thrust fault (Allen et al., 2003). Guest et al. (2006) suggested that the North Tehran thrust merges with
1092
A. Mostafanejad et al. / Journal of Asian Earth Sciences 42 (2011) 1091–1096
Fig. 1. (a) Map of Central Alborz Mountains showing stations (green squares) and local earthquakes (from 1996 to 2006, purple circles) used in this study. Yellow triangle: Damavand volcano, black lines: main faults in the region, diamond: Tehran, white dashed lines: profile along Damavand volcano and 10 km to the east and west side, cross: grid node. Node spacing is 20 km. Fault plane solutions of large events (Mw > 5.0) are from Harvard University (2009). Inset map shows location of studied area in the Middle East. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Mosha fault at depth. Damavand volcano is situated 20 km northwards from the Mosha active fault and lies unconformably on folds and thrusts (Liotard et al., 2008). Tectonic activity in the Alborz Mountains is due both to the convergence of central Iran toward stable Eurasia, and to motion of the South Caspian Basin, with respect to Eurasia. Coupled with the NS convergence of central Iran (52 mm/yr), the south-westward motion of the South Caspian Basin, with respect to central Iran, leads to a NNE-SSW transpressional regime in the Alborz (Vernant et al., 2004; Ritz et al., 2006). Ritz et al. (2006) documented a transition from transpression to active transtension in very recent time in the Central Alborz region. According to their estimation, the transtension started between 1 and 1.5 Ma. This is contemporaneous with the Damavand volcanic activity, which Davidson et al. (2004) dated as occurring between 1.8 Ma and 7 Ka.
3. Data and method Fig. 1 shows the locations of the seismic stations and epicenters of relocated earthquakes. The seismic array consists of 19 permanent stations of the Iranian Seismic Telemetry Network (IRSC) equipped with 3-component short period (SS1) seismometers. The location of earthquakes was improved through using the hypoDD method (Waldhauser and Ellsworth, 2000). HypoDD determines relative locations within clusters using the doubledifference algorithm. This algorithm improves relative location accuracy by removing effects due to an un-modeled velocity structure through the propagation path. A significant improvement in
hypocenter locations is observed by comparing the relocated earthquakes using hypoDD and those routinely determined by IRSC (Mottaghi et al., 2010). We selected 895 relocated earthquakes between 1996 and 2006 based on the following criteria: (1) all the local earthquakes used are located within the seismic network with azimuthal gap less than 180°, (2) the uncertainties of the hypocentral locations (both horizontal and vertical) are, on average, less than 6 km during relocation with the 1-D background model and (3) all the selected earthquakes are recorded by at least four stations. Our final dataset is composed of 5559 P-wave observations, giving an average of 6 P-wave observations per event (Figs. 1 and 2). For the initial 1-D velocity model, we used the model developed by Ashtari et al. (2005), which is based on 1-D inversion of the arrival times for local events (Kissling, 1988). We used a damped least-squared iterative method (SIMULPS14), (Thurber, 1983; Eberhart-Phillips, 1990; Evans et al., 1994) to solve the non-linear tomography problem (Haslinger and Kissling, 2001; Husen et al., 2003). Hypocenter locations are included in the inversion as unknowns, due to the coupling of hypocenter locations and velocities (Thurber, 1992). Travel times through the velocity model are calculated using full 3-D shooting ray tracing (Virieux and Fara, 1991; Haslinger and Kissling, 2001). Having considered the ray coverage of the dataset (Fig. 2), we chose a horizontal model parameterization of 20 20 km for the regional model. This model parameterization represents the finest possible model grid spacing without showing a strongly heterogeneous pattern of the derivative weight sum (DWS) (Husen et al., 2000, 2003). Grid node spacing with depth is 5 km. Following the method introduced by Eberhart-Phillips (1986) for estimation of an optimum damping
A. Mostafanejad et al. / Journal of Asian Earth Sciences 42 (2011) 1091–1096
1093
Fig. 2. Location of the 895 selected local events (gray dots), and grid nodes position (gray crosses). Horizontal ray coverage of the dataset is depicted in black solid lines. Gray squares: stations, yellow triangle: Damavand volcano. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
value, we performed a large range of single iteration inversions with damping values ranging from 1 to 1000. The optimum damping value of 50 was obtained based on the trade-off curve between data variance and model length. The trade-off curve shows similar trends, both for inversions with and without station corrections (see Fig. 3). A strong increase in data variance with decreasing damping can be avoided by including station delays in the inversion process, as depicted in Fig. 3. Station delays are usually used to compensate for the near-surface structure beneath the seismic stations (e.g., Husen et al., 2003). The effect of stations delays is
Fig. 3. Trade-off curve for single iteration results with (squares) and without (circles) station correction. The selected value of damping, 50, is shown as a black square.
more pronounced as the degree of velocity heterogeneity increases as a result of lowered damping. 4. Inversion results A significant problem in seismic tomography is to optimize and evaluate the spatial resolution of the estimated model. Model resolution is limited by various factors including data errors and the three-dimensional geometry of the crossing rays through the volume. In order to evaluate the resolving power of the dataset, we incorporated various measures including a checkerboard test, assessing DWS, Resolution Diagonal Elements (RDE), and hit count. An alternating 50 50 km high and low velocity (checkerboard) of 10%, relative to the 1-D initial velocity model, was produced using the real ray geometry of the data. The synthetic dataset is then inverted in the same way as real data. The result of the synthetic checkerboard test is shown in Fig. 4a. It indicates that the resolution around the Damavand volcano in the central part of the model is fairly good with the least smearing between adjacent nodes. DWS is a measure of ray density in the vicinity of a model parameter (Toomey and Foulger, 1989) that weighs the importance of each ray segment by its distance to the model node. RDE is yet a better estimate of the model resolution. In an ideal situation, zero value of RDE implies no resolution, while 1 stands for perfect resolution. However, the amount of RDE strongly depends on different factors like model parameterization and damping value. Increasing the density of model parameters, although increases the fidelity of the model, it decreases RDE (Toomey and Foulger, 1989). Fig. 5 shows the amount of RDE and DWS at the depth sections of 5, 10 and 15 km as well as vertical cross sections of DWS along the
1094
A. Mostafanejad et al. / Journal of Asian Earth Sciences 42 (2011) 1091–1096
Fig. 4. (a) Horizontal slices showing a synthetic checkerboard model test with 10% velocity anomaly relative to 1-D initial model. Yellow triangle: Damavand volcano. (b) Horizontal slices of real inversion results after 3 iterations. (c) Vertical slices through Damavand volcano (middle panel) and 10 km to the west and east of the volcano (the locations of the cross-section are shown in Fig. 1). Black circles show hypocenters of earthquakes. Black contour lines show absolute velocity. Pink contour outlines area with diagonal elements of the resolution matrix > 0.1. The topography variations along the profiles are also shown on top of each panel respectively. The position of Damavand volcano and surface trace of the Mosha fault are also denoted by arrows. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
volcano and 10 km to the east and west sides of it. Based on the result of synthetic testing, we chose areas with RDE greater than 0.1 as the region with reliable resolution. This value, while low, is consistent with previous studies using this inversion method (e.g. Haslinger et al., 1999; Waite and Moran, 2009). In summary, our extensive synthetic tests indicate that regularisation parameters used for inversion enhance the resolution, especially in the central part of the model. The final P-wave velocity model reveals several high and low velocity anomalies in the upper 30 km of the crust in the Central Alborz region (Fig. 4b and c). The first striking feature of the resulted images is the almost perfect correlation between the pronounced V-shape structure of the Alborz Mountain range and the linear low velocity anomaly along the main faulting system at 15 km depth (Fig. 4b). We relate this low velocity anomaly to the fractured zone caused by the Mosha and the North Tehran faults (see Fig. 1). Based on stratigraphy of the region, the sedimentary cover in the Alborz Mountains is up to 6 to 8 km thick (Stöcklin, 1974) and is underlain by crystalline basement (Dedual, 1967). As a result, it is very probable that the fracturing pattern, as well
as P-wave velocity, vary considerably, which eventually causes the linear low velocity along the Mosha and North Tehran fault to be so conspicuous only in depth of 15 km and not in shallower depths. In the vertical section across the volcano a relatively high velocity zone is clearly imaged in the upper 20 km of the model on the north side of the Damavand volcano (Fig. 4c). We interpret this relatively high velocity body as the elderly crystallized magma chamber. The position of the hypothetical conduit related to this old chamber is in accordance with the old Damavand cone, which is slightly to the north of the present cone (Davidson et al., 2004) (Fig. 4c). A relatively low velocity body is detected from near surface to the depth of 7 km, right below the young Damavand cone. We interpret this relatively low velocity region as being the magma chamber of Damavand volcano. It is worth mentioning that a geothermal reservoir of 550 km2 with a mean temperature of 190 °C at depths of 2–3 km and estimated thermal energy of 5.11 1018 J is investigated in Damavand area (Ghobadian et al., 2009). The 3-D velocity model strongly reduces travel time residuals by 58% after three iterations (corresponds to RMS residuals from
A. Mostafanejad et al. / Journal of Asian Earth Sciences 42 (2011) 1091–1096
1095
Fig. 5. (a) Resolution Diagonal Elements (RDE), and (b) Derivative weight sum (DWS) for depth sections 5, 10 and 15 km (c) DWS for vertical slices through Damavand volcano and 10 km to the east and west of the volcano.
1.2 to 0.7 s). However, the variance reduction depends on number of regularization parameters, including the damping factor. Thus, we consider the model presented as a smoothed model describing the major structural features resolved by the inversion. It is worth noting that the remnant residuals of the final model are high (0.7 s). This relatively high value should be considered next to the relatively high phase picking error, 0.2 s, of the dataset. It can be assumed that the remnant residuals partly reflect some features of the true velocity anomalies in the earth that cannot be taken into account by our model. The most important one might be anisotropy, which can be strong in this highly heterogeneous area. The other factor can be related to the existence of some small-scale velocity anomalies in the uppermost crust, which are below the resolution capacity of our inversion scheme. A detailed understanding of the geological setting of the studied area requires the deployment of a dense seismic network, which is clearly the ambition of further research in this area. 5. Conclusion For the first time, a 3-D velocity model of the Central Alborz region, including the Damavand volcano, is presented for the upper
30 km of the crust using the available local earthquake dataset. The 3-D model reveals a number of interesting features including: 1. A low velocity anomaly along the main structural features at the depth of 15 km which is due to the fractured zone caused by the Mosha and North Tehran faults. 2. A relatively high velocity body is resolved in the northern side of the Damavand volcano almost in the same location as the old Damavand caldera. We consider this to be the elderly crystallized magma chamber. 3. There is a pronounced localized low velocity zone beneath the volcano to the depth of 7 km that is considered to be the magma chamber of the Damavand volcano.
Acknowledgments We are grateful to the Iranian Seismological Center (http://irsc.ut.ac.ir/) for the access to the earthquake data base. We thank M. Berberian for his early comments and encouragements, and B. Darvishzad for invaluable discussions and help. Roland Roberts, from Uppsala University, is especially thanked for his valuable comments. We appreciate H. DeShon and C. Powell for their
1096
A. Mostafanejad et al. / Journal of Asian Earth Sciences 42 (2011) 1091–1096
insightful comments and suggestions during the review process. We also thank two anonymous reviewers for their constructive suggestions which helped further improvements of the manuscript. Figures are produced using the GMT software (Wessel and Smith, 1998).
References Allen, M., Ghassemi, M.R., Shahrabi, M., Qorashi, M., 2003. Accommodation of late Cenozoic oblique shortening in the Alborz range, northern Iran. J. Struct. Geol. 25, 659–672. Allenbach, P., 1966. Geologie und petrographie des Damavand und seiner umgeurg (Zentral Elburz). Geologisches Institut, ETH-Zurich, Mitteilung Nr. 63, Iran, 114 p. (In german). Ashtari, M., Hatzfeld, D., Kamalian, N., 2005. Microseismicity in the region of Tehran. Tectonophysics 395, 193–208. Axen, G.J., Lam, P.S., Grove, M., Stockli, D.F., Hassanzadeh, J., 2001. Exhumation of the west-central Alborz Mountains, Iran, Caspian subsidence, and collisionrelated tectonics. Geology 29, 559–562. Berberian, M., 1976. An explanatory note on the first seismotectonic map of Iran; A seismo-tectonic review of the country. Geol. Surv. Iran. 39, 7–141. Berberian, M., Yeats, R.S., 2001. Contribution of archaeological data to studies of earthquake history in the Iranian plateau. J. Struct. Geol. 23, 563–584. Berberian, M., Qorashi,M., Argang Ravesh, B., Mohajer Ashjaie, A., 1993. Seismotectonics and earthquake-fault hazard investigation in the Tehran Region, contribution to the seismotectonics of Iran, Report. 56, Geol. Surv. Iran. Iran. Davidson, J., Hassanzadeh, J., Berzins, R., Stockli, D.F., Bashukooh, B., Turrin, B., Pandamouz, A., 2004. The geology of Damavand volcano, Alborz Mountains, northern Iran. GSA Bull. 116, 16–29. Dedual, E., 1967. Zur Geologic des mittleren und unteren Karaj-Tales Zentral-Elburz (lran). Mitt. Geol. Inst ETH u. Univ. Zurich, n.s., 76 (In German) 123p. Dehghani, G.A., Makris, J., 1984. The gravity field and crustal structure of Iran: Neues Jahrbuch für Geologie und Palaontologie Abhandlungen. 168, 215–229. Eberhart-Phillips, D., 1986. Three-dimensional velocity structure in Northern California Coast Ranges from inversion of local earthquake arrival times. Bull. Seismol. Soc. Am. 76, 1025–1052. Eberhart-Phillips, D., 1990. Three-dimensional P and S velocity structure in the Coalinga region, California. J. Geophys. Res. 95, 343–363. Evans, J.R., Eberhart-Phillips, D., Thurber, C.H., 1994. User’s manual for SIMULPS12 for imaging Vp and Vp/Vs: a derivative of the ‘‘Thurber’’ tomographic inversion SIMUL3 for local earthquakes and explosions. Open-File Rep 94-431. U.S. Geol. Surv., Reston, VA. 101pp. Ghobadian, B., Najafi, G., Rahimi, H., Yusaf, T.F., 2009. Future of renewable energies in Iran. Renew. Sust. Energy Rev. 13, 689–695. Guest, B., Axen, G.J., Lam, P.S., Hassanzadeh, J., 2006. Late Cenozoic shortening in the west-central Alborz Mountains, northern Iran, by combined conjugate strikeslip and thin-skinned deformation. Geosphere 2, 35–52. Harvard University (2009), Global CMT catalog,
earthquake tomography around the Gulf of Arta (Ionian region, NW Greece). Tectonophysics 304, 201–218. Husen, S., Kissling, E., Flueh, E.R., 2000. Local earthquake tomography of shallow subduction in north Chile: a combined onshore and offshore study. J. Geophys. Res. 105, 28183–28198. Husen, S., Kissling, E., Deichmann, N., Wiemer, S., Giardini, D., Baer, M., 2003. Probabilistic earthquake location in complex 3-D velocity models: application to Switzerland. J. Geophys. Res. 108, 2077–2096. Kissling, E., 1988. Geotomography with local earthquake data. Rev. Geophys. 26, 659–698. Lees, J.M., 2007. Seismic tomography of magmatic systems. J. Volcanol. Geotherm. Res. 167, 37–56. Liotard, J.M., Dautria, J.M., Bosch, D., Condomines, M., Mehdizadeh, M., Ritz, J.F., 2008. Origin of the absarokite–banakite association of the Damavand volcano (Iran): trace elements and Sr, Nd, Pb isotope constraints. Int. J. Earth Sci. (Geol Rundsch). 97, 89–102. Moinabadi, E.M., Yassaghi, A., 2007. Geometry and kinematics of the Mosha Fault, south central Alborz Range, Iran: an example of basement involved thrusting. J. Asian Earth Sci. 29, 928–938. Mottaghi, A., Rezapour, M., Yaminifard, F., 2010. Double-difference relocation of earthquake hypocenters along the southern flank of the Central Alborz (Iran). Bull. Seismol. Soc. Am. 100, 2014–2023. Ritz, J.F., Nazari, H., Ghassemi, A., Salamati, R., Shafei, A., Solaymani, S., Vernant, P., 2006. Active transtension inside central Alborz: a new insight into northern Iran–southern Caspian geodynamics. Geology 34, 477–480. Salah, M.K., Seno, T., 2008. Imaging of Vp, Vs, and Poisson’s ratio anomalies beneath Kyushu, southwest Japan: implications for volcanism and forearc mantle wedge serpentinization. J. Asian Earth Sci. 31, 404–428. Stöcklin, J., 1974. Northern Iran: Alborz Mountains. In: Spencer, A.M. (Ed.), Mesozoic–Cenozoic Orogenic Belts data for Orogenic Studies; Alpin— Himalayan Orogens. Geological Society Special Publication 4, London, pp. 212–234. Tchalenko, J. S., Berberian, M., Iranmanesh, H., Bailly, M., Arsovsky, M., 1974. Tectonic framework of the Tehran region. Geological Survey of Iran, Report no. 29. Thurber, C.H., 1983. Earthquake locations and three-dimensional crustal structure in the Coyote Lake area, central California. J. Geophys. Res. 88, 8226–8236. Thurber, C.H., 1992. Hypocenter-velocity structure coupling in local earthquake tomography. Phys. Earth Planet. Inter. 75, 55–62. Toomey, D.R., Foulger, G.R., 1989. Tomographic inversion of local earthquake data from the Hengill–Grensdalur central volcano complex, Iceland. J. Geophys. Res. 94, 17497–17510. Vernant, P., Nilforoushan, F., Chéry, J., Bayer, R., Djamour, Y., Masson, F., Nankali, H., Ritz, J.F., Sedighi, M., Tavakoli, F., 2004. Deciphering oblique shortening of central Alborz in Iran using geodetic data. Earth Planet. Sci. Lett. 223, 177–185. Virieux, J., Fara, V., 1991. Ray tracing for earthquake location in laterally heterogeneous media. J. Geophys. Res. 93, 6585–6599. Waite, G.P., Moran, S.C., 2009. VP structure of Mount St. Helens, Washington, USA, imaged with local earthquake tomography. J. Volcanol. Geotherm. Res. 182, 113–122. Waldhauser, F., Ellsworth, W.L., 2000. A double-difference earthquake location algorithm: method and application to the northern Hayward Fault, California, Bull. Seismol. Soc. Am. 90, 1353–1368. Wessel, P., Smith, W.H.F., 1998. New Improved Version of Generic Mapping Tools Released, vol. 79. EOS Trans. AGU, 579. Zhao, D., Yu, S., Ohtani, E., 2011. East Asia: seismotectonics, magmatism and mantle dynamics. J. Asian Earth Sci. 40, 689–709.