Magnetotelluric images of magma distribution beneath Volcán Uturuncu, Bolivia: Implications for magma dynamics Matthew J. Comeau1, Martyn J. Unsworth1*, Faustino Ticona2, and Mayel Sunagua3 1Department of Physics, University of Alberta, Edmonton, Alberta T6E 2E1, Canada 2Universidad Mayor de San Andrés, Casilla 10077 Correo Central, La Paz, Bolivia 3Empresa Kawsaqi, Correo Central, La Paz, Bolivia ABSTRACT and therefore electromagnetic geophysical The Altiplano-Puna volcanic complex in the central Andes records a history of major cal- methods are useful for investigations of subsur- dera-forming eruptions over the past 10 m.y. Geophysical and geodetic data indicate that face magma distribution. magma is still present, and perhaps mobile, in the crust. Broadband magnetotelluric data were used to generate two-dimensional and three-dimensional electrical resistivity models of MT DATA COLLECTION AND ANALYSIS the Altiplano-Puna magma body (APMB) with a focus on the zone of inflation around Volcán Previous long-period MT data from the Uturuncu in southern Bolivia. Low electrical resistivities (<3 Wm) at a depth of ~15 km below APVC showed a high conductivity zone in the sea level are interpreted as being due to the presence of andesitic melts of the APMB and mid-crust (Schilling et al., 2006), but the large require a melt fraction >20%. The upper surface of the APMB is shallowest beneath Uturuncu interstation spacing (~15 km) could not pro- and the geometry is consistent with geodynamic models that require the upward movement duce detailed images of crustal magma bodies. of a melt layer at this location. The shallower resistivity structure is characterized by discrete An array of 180 broadband MT stations was electrically conductive bodies, oriented east-west near sea level (depth of 5 km), which are collected from 2011 to 2013 to investigate the interpreted as a combination of partial melt and fluids. crustal magma distribution below the APVC (Fig. 1). The MT data were of high quality and INTRODUCTION mation, geophysical data are needed to image indicate the presence of a multilayer resistivity The Altiplano-Puna (South America) is the subsurface structure and understand the magma structure (see the GSA Data Repository1). only modern example of a high plateau that dynamics. The dimensionality of the MT data was inves- has formed above a subduction zone, and con- Magnetotelluric (MT) data are useful because tigated, using the methods of McNeice and Jones vergence has produced a crustal thickness that they map subsurface electrical resistivity using (2001) and Caldwell et al. (2004), to determine reaches 75 km (e.g., Beck et al., 2014). In addi- natural electromagnetic signals (Chave and if a two-dimensional (2-D) or 3-D approach was tion to ongoing arc magmatism, the region Jones, 2012). The resistivity of a rock is sensi- required. The short-period MT data (0.003–1 s) underwent a major ignimbrite flare-up during the tive to the quantity and composition of magma, sample the near-surface structure and indicate past 10 m.y. that formed the Altiplano-Puna vol- canic complex (APVC) (de Silva, 1989). Geo- physical studies have shown that the APVC is 21.5˚ S underlain by anomalously low seismic velocities Bolivia N (Chmielowski et al., 1999; Zandt et al., 2003), high seismic attenuation (Haberland, 2003), and low electrical resistivities (e.g., Schilling et al., W E 2006), and is associated with a negative Bouguer anomaly (del Potro et al., 2013). Together, these 22.0˚ S anomalies are evidence of a major mid-crustal A S magma body, the Altiplano-Puna magma body U (APMB), and make this region a key location for studying potential granite emplacement and plu- Q ton formation, geological processes that are not B N c fully understood (e.g., Brown, 2013). 22.5˚ S Geodetic data have shown that a region of the APVC in southern Bolivia is being uplifted C at 10–15 mm/yr (e.g., Pritchard and Simons, ar olcanic 2004), while a surrounding ring is subsiding V Chile Argentina 20 km (Fig. 1) (Fialko and Pearse, 2012; Henderson 23.0˚ S and Pritchard, 2013). The deformation is cen- 68.0˚ W 67.5˚ W 67.0˚ W 66.5˚ W 66.0˚ W 65.5˚ W tered on Volcán Uturuncu, a composite volcano last active 270 k.y. ago (Sparks et al., 2008), and Figure 1. Study area. Magnetotelluric stations are denoted by small circles. Those in black gives strong evidence for magma motion in the were used in the regional two-dimensional inversion. U—Volcán Uturuncu; Q—Volcán Quetena. Large gray circles denote the limits of the inflation and subsidence (Henderson crust. It is significant that Uturuncu is located and Pritchard, 2013). White box shows the extent of the three-dimensional inversion. Rose close to the center of the APVC and above the diagram shows the geoelectric strike direction for the period range 10–3000 s for all stations. APMB, but there is no consensus on the mecha- Points A, B, and C define the profile shown in Figure 2A. nisms responsible for the uplift. Given the nonu- niqueness in the interpretation of surface defor- 1GSA Data Repository item 2015087, supporting documentation, related figures, and an explanations of meth- ods, is available online at www.geosociety .org/pubs /ft2015.htm, or on request from [email protected] or *E-mail: [email protected] Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301, USA. GEOLOGY, March 2015; v. 43; no. 3; p. 243–246; Data Repository item 2015087 | doi:10.1130/G36258.1 | Published online 5 February 2015 GEOLOGY© 2015 Geological | Volume Society 43 | ofNumber America. 3 For| www.gsapubs.org permission to copy, contact [email protected]. 243 a 1-D resistivity structure. Longer period data sions. The 3-D inversion of Siripunvaraporn et Shallow Conductor (1–3000 s) sample to mid-crustal depths and al. (2005) was applied to a set of 73 stations Conductor C1 is <1 km thick and has highly detected a geoelectric strike direction of N30°E surrounding Uturuncu. variable properties. It is likely caused by hydro- on profile A-B that changes to N0°E on profile thermal alteration of the regionally extensive B-C (Fig. 1; Fig. DR3 in the Data Repository). RESULTS AND DISCUSSION ignimbrite layer. Ignimbrites undergo low-tem- Dimensionality analysis showed that a 2-D The 2-D and 3-D resistivity models (Fig. 2) perature alteration to form highly conductive analysis was valid on a regional scale, but that show three distinct layers: (1) a shallow surface clay within a time scale of 1 m.y. (Bibby et al., 3-D resistivity structure was locally present. conductor (C1) <1 km thick; (2) a deeper, spa- 2005). Both 2-D and 3-D inversions were applied to tially uniform conductor (C2) that begins at a the MT data. The 2-D inversion used the algo- depth of 15–20 km below sea level (BSL); and Deeper Conductor rithm of Rodi and Mackie (2001), with data (3) an intermediate layer with spatially variable The prominent conductor C2 is shown in both rotated to N30°E and N0°E, for the profiles resistivity that contains a number of discrete the 2-D and 3-D models. The western margin A-B and B-C, respectively. Many combina- low-resistivity zones (C3–C7), particularly the of C2 is located beneath the volcanic arc and it tions of inversion parameters were investigated area near Uturuncu. Low resistivity in a volcanic extends ~170 km to the east. If the top is defined and the main resistivity features of the models environment can be caused by aqueous fluids, as the 3 Wm contour, then this conductor is did not depend on any specific parameters. No partial melt, or hydrothermal alteration, so inter- located at 15–20 km BSL in both models. This significant changes in the resistivity model pretation must use additional geophysical data contour was chosen based on the synthetic inver- were observed when strike angles of N20°E or and laboratory studies of fluids and melts to dis- sions, as it corresponds to the resistivity change N40°E were used on profile A-B. Model reso- tinguish between these possibilities (Chave and at the top of C2 (see Fig. DR12). Complications lution was investigated using synthetic inver- Jones, 2012). arise from the fact that the 2-D inversions cannot APVC Subsidence Uplift Subsidence ABQVU C C A C1 C1 C1 Figure 2. A: Regional two-dimensional (2-D) 0 C5 R1 C4 resistivity model obtained from the inver- R2 10 C3 C7 sion of transverse electric and transverse C3 C6 20 magnetic mode magnetotelluric (MT) data at 73 stations. A root mean square (r.m.s.) 30 C2 misfit of 1.5 was achieved. Red triangle in- C2 40 dicates the volcanic arc. White box shows the extent of the 3-D model. U—Volcán Utu- 50 runcu; Q—Volcán Quetena; VC—Vilama Depth below sea level (km) 20 40 60 80 100 120 20 40 60 80 100 120 140 Distance along profile (km) Caldera; APVC—Altiplano-Puna Volcanic Complex. Hachure pattern indicates minimal U U -5 300 resolution by showing approximate limit of B C penetration of electromagnetic signals with 0 C4 C4 a period of 1000 s. B: East-west (EW) vertical 100 section through the 3-D resistivity model. 5 Misfit error (r.m.s.) of 0.95 was achieved. m) Black dots are earthquake hypocenters from 10 Ω Jay et al. (2012). The locations of each sec- 30 ( tion or slice are indicated by white lines. C: 15 North-south (NS) vertical sections through the 3-D resistivity model.